- dovecot - dovecot.org

[Dovecot] deliver mail with postfix to dovecot for mailinglists
by Sven Kirmess 25 Nov '03

25 Nov '03

This is a bit off topic on that list. But the question is required because of a limitation of dovecot. It cannot have shared mailboxes into which one would collect mailing list mails. I assume one has already solved that problem and therefore ask this question on that list. I try to do the following: I would like to setup an alias dovecot(a)kzone.ch and subscribe this alias to the dovecot mailing list. Then I would like to create an alias to run maildrop with the correct parameter to put it into the directory ~ski/Maildir/.dovecot/. Of course, there is no user dovecot on my system and if there would be one she couldn't write to a directory owned by me. Is there a trick to do the following: ski -> ~ski/Maildir dovecot -> ~ski/Maildir/.dovecot postfix -> ~ski/Maildir/.postfix Any other idea? Any good documentation on that? Sven

3 3

[Dovecot] Compiling dovecot
by Ted S. Garaventa 24 Nov '03

24 Nov '03

Still trying to compile dovecot to run on solaris8 x86. The program compiles without errors but does not run. The truss out put I get running the program is illegal system call by program: /usr/local/sbin/dovecot Any help is welcome.

4 3

[Dovecot] entourage + dovecot bug?
by Payton 24 Nov '03

24 Nov '03

I have dovecot set up to run IMAP services without SSL (I have tested many other configurations as well, but this reproduces the problem the simplest). When attempting to connect to the imap server using Entourage for OS X on a mac, I receive the following error: Mail could not be received at this time. The server returned the following error . Dovecot reports a successful login and nothing else in the mail log. All other email clients I have tried (mutt, Mail.app) work fine with the IMAP server. I have also tried this on multiple clients with the same results. All this leads me to believe a bug exists in Entourage, however: I also have a Red Hat 7.3 server running qmail and dovecot in nearly the same manner (except that it is using dovecot 0.99.10 and shadow for password authentication, instead of dovecot 0.99.10.2 and ldap password authentication). Entourage (and all other clients) works just fine with this setup. I'm not certain if the problem lies in some tweak between 0.99.10 and 0.99.10.2 or if it is something with the red hat system itself, but I can't find anything. I am using dovecot 0.99.10.2 on a Red Hat 9 system using qmail 1.04 (netqmail) and checking with the latest version (update 10.1.5) of Entourage on Mac OS 10.2.8. Thanks for any help! -Payton

6 8

[Dovecot] IMAP subscribe/unsubscribe broken in 0.99.10.2
by Scott Heavner 23 Nov '03

23 Nov '03

I saw this problem with 0.99.10.2 and had to revert back to 0.99.9.1. As I recall the subscription stuff was not respecting the end-of-line as a field separator. It would just cut the old text and drop the new in it's place. Or it might cut an inappropriate entry that should have been a partial string match. FolderOneOhNine FolderOne FolderThree Trying to delete FolderOne would probably [i didn't save any botched subscription files] result in this OhNine FolderOne FolderThree -scott

1 0

[Dovecot] 0.99.10.1 released
by Timo Sirainen 23 Nov '03

23 Nov '03

Since release candidate I added fix for CRLF + partial BODY[] fetching. I hope everything works :) I've been running it with mbox for a while and it seems to work fine. v0.99.10.1 2003-11-10 Timo Sirainen <tss(a)iki.fi> * mbox: \Draft and \Deleted flags used opposite flag chars in X-Status header. We were incompatible with other mbox accessing software. WARNING: Upgrading from previous version doesn't automatically swap the flags, so be careful not to accidentally expunge messages that had their \Draft flag changed to \Deleted. * Configuration file changes: - Whitespace at end of line is stripped, use quotes if you need it - # comments are supported after key=value lines. if you need '#' character, quote the value - Both " and ' quotes are supported. If you need to use them, '\' can be used for escaping. - mbox: COPY into same mailbox didn't work and could have corrupted the mailbox - Using Dovecot without index files would crash after using a while - Partial BODY[header] or BODY[part] fetches were buggy if client requested more data than was available in the header/part. - Partial BODY[...] fetches were buggy with messages that had CRLFs - Some BODY and BODYSTRUCTURE replies missed data for message/rfc822 MIME parts causing clients to break - SORT (SUBJECT) was buggy - Timezone fixes with Date-header

11 26

[Dovecot] Compiling Dovecot
by Ted S. Garaventa 21 Nov '03

21 Nov '03

Just found out about Dovecot and it seems to work great. My problem is that it compiles and runs on Redhat just fine. On the other hand it compiles on suse and solaris but will not work. When I start dovecot it silently stops running. No entries in the log or anything. Are there any specific instructions for compiling it on solaris x86 using gcc. Thanks

2 1

[Dovecot] SSL and certificate authorities.
by James Tyson 20 Nov '03

20 Nov '03

Hi all. Usually for in-house use and SSL I would just generate a self-signed certificate because most clients either ignore it or only ask the first time the account is configured. In terms of offering the service to our customers is there any value of getting someone like thawte or instantssl to sign a certificate for imaps/pop3s/smtp? Also, is there a configuration directive for dovecot to add the issuers ca bundle similar to apache's SSLCACertificateFile? And thanks for writing such a kick-arse imap server. It blows courier out of the water! --- James Tyson Director, Giant Robot Ltd http://www.giantrobot.co.nz/

4 6

[Dovecot] patch and bug report (from an other in fact)
by DINH Viet Hoa 20 Nov '03

20 Nov '03

When messages marked \Recent are deleted and expunged, the patch now sends the new message count and the new recent messages count (although the message count is not really necessary). >From RFC 3501 errata (ftp://ftp.cac.washington.edu/imap/rfc3501-errata) << Section 6.4.3, page 49: old: The EXPUNGE command permanently removes all messages that have the \Deleted flag set from the currently selected mailbox. Before returning an OK to the client, an untagged EXPUNGE response is sent for each message that is removed. new: The EXPUNGE command permanently removes all messages that have the \Deleted flag set from the currently selected mailbox. Before returning an OK to the client, an untagged EXPUNGE response is sent for each message that is removed. Note that if any messages with the \Recent flag set are expunged, an untagged RECENT response is sent after the untagged EXPUNGE(s) to update the client's count of RECENT messages. >> The bug is that when we connect to the IMAP server through a command (ssh /usr/libexec/imap for example). \Recent flags are not removed when we have finish the access to the mailbox. (I have not yet investigated about that) Besides that, did you found a solution for the message parser ? (the bug in FETCH ENVELOPE that I noticed personnaly to Timo) -- DINH V. Hoa, "s/^$\(\\.\|[^\[]\|\[\(\^.\|[^^]$[^]]*\]\)*\)$\\[^\*[]$/\1\\\4/" -- Stéphane CHAZELAS

2 1

[Dovecot] run POP3S + IMAPS simultaneously?
by darren david 19 Nov '03

19 Nov '03

a quick one: is it possible to run POP3S and IMAPS services through dovecot at the same time? would i just uncomment both "login = imap" and "login = pop3" in /etc/dovecot.conf? or can i only run one at a time? apologies if the answer is obvious, i'm just still trying to wrap my head around all of this. thanks, darren david

2 2

[Dovecot] error msgs about missing ssl/tls support
by Kai Vehmanen 18 Nov '03

18 Nov '03

Hi, I've been just testing 0.99.10.2 and had a really hard time getting the new Dovecot to start properly. First I thought it was an error in my old dovecot.conf, but the actual problem turned out to be simply that Dovecot was compiled without SSL/TLS support (I hadn't noticed this), but dovecot.conf on the other hand specified SSL/TLS ports. The error messages might be a bit more informative in this case. Currently I only got this to my logs: --cut-- Nov 18 18:17:00 dovecot: Dovecot starting up Nov 18 18:17:01 imap-login: BUG: SSL initialization parameters not given while they should have been Nov 18 18:17:01 imap-login: BUG: SSL initialization parameters not given while they should have been Nov 18 18:17:01 pop3-login: BUG: SSL initialization parameters not given while they should have been Nov 18 18:17:01 imap-login: BUG: SSL initialization parameters not given while they should have been Nov 18 18:17:01 dovecot: Login process died too early - shutting down Nov 18 18:17:01 dovecot: child 18482 (login) returned error 89 Nov 18 18:17:01 dovecot: child 18481 (login) returned error 89 Nov 18 18:17:01 dovecot: child 18480 (login) returned error 89 Nov 18 18:17:01 dovecot: child 18479 (login) returned error 89 --cut-- ... and Dovecot would exit. Running Dovecot under gdb would just exit (even with -F), and segfault if ran under strace. So it took a while before I realized that configure had not even detected SSL/TLS... Oh well, not a big thing, but still something to keep in mind. -- http://www.eca.cx Audio software for Linux!

1 0

Re: [Dovecot] Problem with DIGEST-MD5 authentication and plainte xt passwords
by Matthew Reimer 18 Nov '03

18 Nov '03

Maikel Verheijen wrote: > For what it's worth: > > We use postgres userlookups, and use an older version of dovecot (an old > cvs version of around 0.99.9.1), and we DO use digest-md5 password > encryption in the database and realms. It seems to work flawlessly in > our situation, I used kmail, evolution, apple-mail, entourage, outlook, > outlook-express and mutt. > > > Have there been changes in the digest-md5 code since 0.99.9.1 ? > > > Kind regards, > > > Maikel Verheijen > Ladot Nederland BV. The problem is not with passwords that are stored in the db in digest-md5 format (i.e., as an md5 hash of user:realm:password), but with dovecot's password_generate() that takes the username and password and hashes it on the fly. Matt

2 1

Re: [Dovecot] Problem with DIGEST-MD5 authentication and plaintext passwords
by Matthew Reimer 18 Nov '03

18 Nov '03

Timo Sirainen wrote: > On Wed, 2003-11-12 at 03:09, Matthew Reimer wrote: > >>I wasn't able to get DIGEST-MD5 authentication working with passwords >>stored as plaintext until I tweaked password_generate() to make it not >>use the domain portion of user(a)domain.com as the realm. Both evolution >>and kmail send the email address as the username with no realm; when >>dovecot tries to convert the username to user + realm, the hash is >>different and so DIGEST-MD5 authentication fails. Having >>password_generate() hash user@domain.com::password instead of >>user:domain.com:password fixes it. I need this to work because I need to >>support all of PLAIN, CRAM-MD5, and DIGEST-MD5 at the same time. >> >>Timo, do you think this is a proper fix, or are kmail and evolution broken? > > > Well .. Looks like most clients don't actually use realms, but I rather > wouldn't completely drop support for them which your change would do. > > Correct fix would be to separate realm and "@domain" handling in code, > that should work with both cases. I'll see if I can get that done. Thanks Timo. Send me a patch when you're done and I'll test it. Matt

1 0

[Dovecot] Children, HasChildren
by Miquel van Smoorenburg 17 Nov '03

17 Nov '03

Hello, I'm using dovecot 0.99.10.2 on a debian box. Mailstore is maildir. If I list the IMAP hierarchy in mutt (debian, 1.5.4+20031024-1) it shows all my plain mailboxes as having subfolders - while none of them have. I straced mutt and this is the IMAP exchange (a bit reformatted for readability): write(6, "a0001 LSUB \"\" \"%\"\r\n", 19) = 19 read(3, " * LSUB () \".\" \"Drafts\" * LSUB () \".\" \"INBOX\" * LSUB () \".\" \"Sent\" * LSUB () \".\" \"Trash\" * LSUB (\\Noselect \\Children) \".\" \"broadband\" * LSUB (\\Noselect \\Children) \".\" \"cistron\" * LSUB (\\Noselect \\Children) \".\" \"debian\" * LSUB (\\Noselect \\Children) \".\" \"ip\" * LSUB (\\Noselect \\Children) \".\" \"lists\" * LSUB (\\Noselect \\Children) \".\" \"prive\" a0001 OK Lsub completed. ", 1024) = 374 Now this is strange - I read RFC3348 and shouldn't \\Children be \\HasChildren, also shouldn't the plain mailboxes (first four) be listed with \\HasNoChildren ? Mike.

2 3

[Dovecot] Corrupted index files
by Tomi Hakala 17 Nov '03

17 Nov '03

Hi, I'm playing with Dovecot in a lab environment and got it in state where it crashes on every incoming connection. I moved mail spool from ext2 file system to a ReiserFS on a another disk, mount point changed from /var/vmail to /spool/vmail. After that I managed to open mailbox and I moved about 2000 messages from INBOX to a subfolder. Then I tried to open mailbox with POP3 and it didn't succeed, from that point pop3d and imapd died every time I logged in until I removed all index files by hand. This is what I have in logs. imap-login: Nov 17 13:30:53 Info: Login: tomi.hakala(a)xx.fi [xx.xx.xx.xx] pop3-login: Nov 17 13:32:04 Info: Login: tomi.hakala(a)xx.fi [xx.xx.xx.xx] pop3(tomi.hakala(a)xx.fi): Nov 17 13:32:04 Error: Corrupted index file /spool/vmail/xx.fi/tomi.hakala(a)xx.fi/Maildir/.INBOX/.imap.index: Sequence 1848 not found from binary tree (5398 msgs says header) pop3-login: Nov 17 13:32:12 Info: Login: tomi.hakala(a)xx.fi [xx.xx.xx.xx] pop3(tomi.hakala(a)xx.fi): Nov 17 13:32:21 Error: Couldn't lock created modify log file /spool/vmail/xx.fi/tomi.hakala(a)xx.fi/Maildir/.INBOX/.imap.index.log imap(tomi.hakala(a)xx.fi): Nov 17 14:09:32 Error: IndexID mismatch for modify log file /spool/vmail/xx.fi/tomi.hakala(a)xx.fi/Maildir/.INBOX/.imap.index.log imap(tomi.hakala(a)xx.fi): Nov 17 14:09:33 Panic: file mail-tree-redblack.c: line 187 (rb_left_rotate): assertion failed: (node[x].right != RBNULL) dovecot: Nov 17 14:09:33 Error: child 19984 (imap) killed with signal 6 imap-login: Nov 17 14:27:24 Info: Login: tomi.hakala(a)xx.fi [xx.xx.xx.xx] imap(tomi.hakala(a)xx.fi): Nov 17 14:27:24 Panic: file mail-tree-redblack.c: line 232 (rb_right_rotate): assertion failed: (node[y].left != RBNULL) dovecot: Nov 17 14:27:24 Error: child 10478 (imap) killed with signal 6 -- Tomi Hakala

1 0

[Dovecot] Problem with DIGEST-MD5 authentication and plaintext passwords
by Matthew Reimer 17 Nov '03

17 Nov '03

I wasn't able to get DIGEST-MD5 authentication working with passwords stored as plaintext until I tweaked password_generate() to make it not use the domain portion of user(a)domain.com as the realm. Both evolution and kmail send the email address as the username with no realm; when dovecot tries to convert the username to user + realm, the hash is different and so DIGEST-MD5 authentication fails. Having password_generate() hash user@domain.com::password instead of user:domain.com:password fixes it. I need this to work because I need to support all of PLAIN, CRAM-MD5, and DIGEST-MD5 at the same time. Timo, do you think this is a proper fix, or are kmail and evolution broken? Matt

3 2

[Dovecot] MySQL patch + Debian
by James Tyson 17 Nov '03

17 Nov '03

Hi folks. Sorry to post here about a patch that hasn't been officially integrated into Dovecot, however I am trying to get it going and seem to have hit a dead end. I downloaded the Debianised source for Dovecot (0.99.10-9) from sarge: > wedge:/usr/local/src# apt-get -t testing source dovecot then applied the latest MySQL patch that I could find on the list: > wedge:/usr/local/src/dovecot-0.99.10# cat ../dovecot-mysql.patch | > patch -p 2 and manually fixed a merge issue with src/auth/master-connection.c, fiddled with the debian changelog to indicate the patch and then used dpkg-buildpackage -b to compile the package. Everything looked to build alright and I grepped the output to make sure that it compiled userdb-mysql.o and passdb-mysql.o into dovecot-auth. Problem is that dovecot-auth says "Unknown userdb type 'mysql'" and an ldd of /src/auth/dovecot-auth shows no libmysqlclient10: > wedge:/usr/local/src/dovecot-0.99.10# ldd src/auth/dovecot-auth > libpam.so.0 => /lib/libpam.so.0 (0x4001a000) > libldap.so.2 => /usr/lib/libldap.so.2 (0x40022000) > libpq.so.3 => /usr/lib/libpq.so.3 (0x4005a000) > libcrypt.so.1 => /lib/libcrypt.so.1 (0x4006e000) > libdl.so.2 => /lib/libdl.so.2 (0x4009c000) > libc.so.6 => /lib/libc.so.6 (0x4009f000) > liblber.so.2 => /usr/lib/liblber.so.2 (0x401cd000) > libresolv.so.2 => /lib/libresolv.so.2 (0x401da000) > libsasl2.so.2 => /usr/lib/libsasl2.so.2 (0x401ec000) > libgnutls.so.7 => /usr/lib/libgnutls.so.7 (0x401fe000) > libssl.so.0.9.7 => /usr/lib/i686/cmov/libssl.so.0.9.7 > (0x40231000) > libcrypto.so.0.9.7 => /usr/lib/i686/cmov/libcrypto.so.0.9.7 > (0x40260000) > libkrb5.so.17 => /usr/lib/libkrb5.so.17 (0x40351000) > libnsl.so.1 => /lib/libnsl.so.1 (0x40387000) > /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000) > libtasn1.so.0 => /usr/lib/libtasn1.so.0 (0x4039c000) > libgcrypt.so.1 => /usr/lib/libgcrypt.so.1 (0x403aa000) > libz.so.1 => /usr/lib/libz.so.1 (0x403e8000) > libcom_err.so.1 => /usr/lib/libcom_err.so.1 (0x403f7000) > libkrb.so.1 => /usr/lib/libkrb.so.1 (0x403f9000) > libasn1.so.6 => /usr/lib/libasn1.so.6 (0x40415000) > libroken.so.16 => /usr/lib/libroken.so.16 (0x40435000) > libdb3.so.3 => /usr/lib/libdb3.so.3 (0x40445000) Anyone know what I'm missing? --- James Tyson Director, Giant Robot Ltd http://www.giantrobot.co.nz/

2 2

[Dovecot] Migrate from Courier-imap
by Jerry Chiu 16 Nov '03

16 Nov '03

Hi, I have a redhat server running courier-imap + squirrelmail webmail. I am migrating to dovecot... the problem is - the existing sub-folders are all disappear from the folder list. What can I do to fix this problem? Thanks, Jerry Chiu

2 1

[Dovecot] namesapce patch
by Farkas Levente 16 Nov '03

16 Nov '03

hi, where can I find the namspace patch? I remember there was an email on this list, that if I'd not like to open all sufolder's in mozilla's mailer on startup, but I'd like to start mozilla to open all subfolder I've to apply the namespace patch. now it getting more and more anoying so I'd like to apply, but I can't find it... thanks. -- Levente "Si vis pacem para bellum!"

2 1

[Dovecot] dovecot vs cyrus, uw, etc.
by Scott Klein 13 Nov '03

13 Nov '03

I've been doing research on switching our current e-mail server (qpopper, sendmail) to imap. The decision on which server to use is essentially down to Cyrus and Dovecot -- I like Cyrus' approach to a lot of things, but the "blackbox" nature of it makes some niceties like using spamassassin and procmail difficult, or at least counterintuitive. Dovecot seems to play nicer with other apps. I'm a bit concerned with stability, though. Although we don't have a lot of users (about 60) they're all very big e-mail users -- my users rely on their e-mail more than on their phones. We can't afford any downtime, and with our pop3 server, we haven't had any. Does anybody on this list use Dovecot to serve mission critical e-mail to an entire company? What kind of uptime can I expect? Is it genuinely ready for prime time? Has anybody here migrated from a recent release of Cyrus and can compare apples-to-apples? Many thanks, Scott Klein -- Scott Klein Web Publisher/Director of Technology The Nation Magazine http://www.thenation.com

8 8

[Dovecot] 0.99.10.2 for debian
by Jaldhar H. Vyas 12 Nov '03

12 Nov '03

I just uploaded it to sid. A woody backport is available from http://www.braincells.com/open/ I came across an error: giving --without-cyrus-sasl2 to the configure script doesn't work i.e. it looks for -lsasl anyway. Other than that, it seems to be working fine. -- Jaldhar H. Vyas <jaldhar(a)debian.org> La Salle Debain - http://www.braincells.com/debian/

1 0

Re: [Dovecot] Bug with partial IMAP fetches
by Matthew Reimer 11 Nov '03

11 Nov '03

--- message-send.c.orig Sun Jun 22 18:09:48 2003 +++ message-send.c Tue Nov 4 12:53:04 2003 @@ -31,12 +31,11 @@ if (msg_size->physical_size == msg_size->virtual_size && !fix_nuls) { /* no need to kludge with CRs, we can use sendfile() */ - i_stream_skip(input, virtual_skip); old_limit = input->v_limit; limit = input->v_offset + max_virtual_size; i_stream_set_read_limit(input, I_MIN(limit, old_limit)); - ret = o_stream_send_istream(output, input) > 0; + ret = o_stream_send_istream(output, input); i_stream_set_read_limit(input, old_limit); return ret;

2 2

[Dovecot] PATCH: CRAM-MD5 for Dovecot
by Joshua Goodall 11 Nov '03

11 Nov '03

Hi Timo, Mozilla 1.5 doesn't understand DIGEST-MD5; it only speaks CRAM-MD5 or PLAIN. Maybe other clients have the same problem. My users like Mozilla, so I wrote new code for Dovecot to speak CRAM-MD5, using your mech-digest-md5.c as a reference. Attached diff in two formats i) for current CVS ii) against last release both tested and working with Mozilla and KMail's CRAM-MD5. I don't know if I've caught precisely the coding style you use for Dovecot. I hope so, but I'm unsure if I've used your string and buffer libraries properly. Please enlighten me if you have time. Would you consider including this in the next release? Regards, Joshua -- Joshua Goodall joshua(a)roughtrade.net "Your object hit ratio is weak, old man" "If you cache me now, I will dump more core than you can possibly imagine"

2 4

[Dovecot] Problems with Linux 2.6
by Aredridel 10 Nov '03

10 Nov '03

I'm running Dovecot under PLD GNU/Linux, on two servers. Under kernel 2.4, it works wonderfully. Under 2.6, I get this in the log: Nov 9 21:36:38 polis dovecot: Dovecot starting up Nov 9 21:36:39 polis imap-login: setuid(97) failed: Resource temporarily unavailable Nov 9 21:36:39 polis imap-login: setuid(97) failed: Resource temporarily unavailable Nov 9 21:36:39 polis pop3-login: setuid(97) failed: Resource temporarily unavailable Nov 9 21:36:39 polis pop3-login: setuid(97) failed: Resource temporarily unavailable Nov 9 21:36:39 polis dovecot: Login process died too early - shutting down Nov 9 21:36:39 polis dovecot: child 32404 (login) returned error 89 Nov 9 21:36:39 polis dovecot: child 32405 (login) returned error 89 Nov 9 21:36:39 polis dovecot: child 32406 (login) returned error 89 Nov 9 21:36:39 polis dovecot: child 32408 (login) returned error 89 Nov 9 21:36:39 polis pop3-login: fd_send(-1) failed: Broken pipe Any ideas? the setuid32() call is returning EAGAIN, though I have no resource limits on user nobody (nor does it work as a dovecot user). Aredridel

2 1

[Dovecot] openssl decryption failed
by Tom Allison 09 Nov '03

09 Nov '03

I've got dovecot working on imap But when I try to run imaps I get: imap-login: Nove 08 19:11:19 Warning: SSL_accept() failed: error:1408F455:SSL routines:SSL3_GET_RECORD:decryption failed or bad record mac [192.168.1.103] Now I'm not sure if theres any difference here, but I don't have string for 'login = imaps' or 'login = pop3s' to match with my protocols string I also don't have any specific imaps_listen interface defined, but since there's something in the logs, I assume that this is not a problem. SETUP fluffy:~ # egrep "^\w" /usr/local/etc/dovecot.conf base_dir = /var/run/dovecot/ protocols = imap imaps pop3s log_path = /var/log/dovecot.log info_log_path = /var/log/dovecot.info login = imap login = pop3 verbose_ssl = yes auth = default auth_mechanisms = plain auth_userdb = passwd auth_passdb = pam auth_user = root auth_verbose = yes

2 2

[Dovecot] openssl config.
by Tom Allison 09 Nov '03

09 Nov '03

I checked and I do have the openssl-devel rpm installed SuSE 8.2 openssl-devel-0.9.6i-12 openssl-0.9.61-19 dovecot 0.99.10 ./configure --with-ssl-openssl At the end of the configuration it says ssl support == 'no' :(... I was able to get pam support by picking up the pam-devel modules as suggested. But I got stuck here. After reading up a bit on imap, it seems reasonable to have both imap and imap-ssl support. imap for localhost connections (like squirrelmail) and imap-ssl for everything else.

1 2

[Dovecot] 0.99.10.1 RC
by Timo Sirainen 08 Nov '03

08 Nov '03

Available from http://dovecot.fi/rc/ While still waiting for .11 with index file rewrites to stabilize, here's a release candidate for 0.99.10.1 which is basically 0.99.10 patched with most of the simple bugfixes from CVS. Some bugfixes such as LIST's children-flags were done with larger rewrites so I didn't try to apply them. All the changes are listed in ChangeLog. The important changes are: * mbox: \Draft and \Deleted flags used opposite flag chars in X-Status header. We were incompatible with other mbox accessing software. WARNING: Upgrading from previous version doesn't automatically swap the flags, so be careful not to accidentally expunge messages that had their \Draft flag changed to \Deleted. * Configuration file changes: - Whitespace at end of line is stripped, use quotes if you need it - # comments are supported after key=value lines. if you need '#' character, quote the value - Both " and ' quotes are supported. If you need to use them, '\' can be used for escaping. - mbox: COPY into same mailbox didn't work and could have corrupted the mailbox - Using Dovecot without index files would crash after using a while - Partial BODY[header] or BODY[part] fetches were buggy if client requested more data than was available in the header/part. - Partial BODY[...] fetches were buggy with messages that had CRLFs - Some BODY and BODYSTRUCTURE replies missed data for message/rfc822 MIME parts causing clients to break - SORT (SUBJECT) was buggy - Timezone fixes with Date-header

1 0

[Dovecot] [POP3] RFC2449 conformance?
by Matthias Andree 08 Nov '03

08 Nov '03

RFC2449 Section 5 ("Capabilities available in the AUTHORIZATION state MUST be announced in both states.") requires that POP3 extensions offered in Authentication state MUST be offered in Transaction state as well, this affects the "SASL" extension. This is a CVS version that is a couple of days old (the old 0.99.10 wouldn't run on Linux 2.6 because it has tighter resource limiting): +OK dovecot ready. capa +OK CAPA TOP USER UIDL RESP-CODES SASL PLAIN . user SECRET +OK pass VERYSECRET +OK Logged in. capa +OK CAPA TOP USER UIDL RESP-CODES . quit +OK Logging out. -- Matthias Andree

2 3

[Dovecot] evolution - folders not displayed correctly
by Terry 08 Nov '03

08 Nov '03

Hello, First, I used thunderbird, and everything looks great. I tried my account in evolution and all the folders appear in one level and not in several as they were in thunderbird. Inbox Inbox.foo Inbox.bar instead of Inbox foo bar What could be the problem? Thanks! ===== Terry __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com

2 7

[Dovecot] pop3s fails, imaps works great
by Darren David 08 Nov '03

08 Nov '03

hey all- i'm running dovecot-0.99.10p2 on OpenBSD 3.4-snapshot (installed from ports). i have had 0 problems with imaps -- it's been running like a charm! however, i'm trying to get pop3s set up for some of my users, but i can't seem to connect from ANY client. in /etc/dovecot.conf, i've got: protocols = imaps pop3s and i've also got pop3s_listen set up with the same IP as imaps. i'm using vpopmail for auth. nmap shows that port 995 is open, and netstat does show an established connection to pop3s when i attempt to connect, but i see no activity in the dovecot log, and the client just churns and churns and eventually i have to quit it since i get no response. any ideas what i could have wrong, or how to test? thanks in advance, darren david

1 0

[Dovecot] PATCH: make local IP address available to auth modules
by Matthew Reimer 04 Nov '03

04 Nov '03

The attached patch makes the local IP address to which the client connected available to the authentication modules; i.e., the local IP address is available for substitution as %i for the mysql and pgsql modules. We needed this feature to support thousands of our legacy accounts which are authenticated by username/local_part (not the full email address) and IP address (one per domain). Timo, would this be acceptable to commit? One further improvement would be to make the remote IP address available as well, in case one wanted to implement a tcpwrapper-like functionality. Matt diff -u -r work.patched/dovecot-0.99.10/src/auth/auth-login-interface.h work/dovecot-0.99.10/src/auth/auth-login-interface.h --- work.patched/dovecot-0.99.10/src/auth/auth-login-interface.h Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/auth-login-interface.h Fri Oct 24 13:57:18 2003 @@ -1,6 +1,8 @@ #ifndef __AUTH_LOGIN_INTERFACE_H #define __AUTH_LOGIN_INTERFACE_H +#include "network.h" + /* max. size for auth_login_request_continue.data[] */ #define AUTH_LOGIN_MAX_REQUEST_DATA_SIZE 4096 @@ -47,6 +49,7 @@ struct auth_login_request_new { enum auth_login_request_type type; /* AUTH_LOGIN_REQUEST_NEW */ unsigned int id; /* unique ID for the request */ + struct ip_addr ip_local; enum auth_mech mech; enum auth_protocol protocol; @@ -56,6 +59,7 @@ struct auth_login_request_continue { enum auth_login_request_type type; /* AUTH_LOGIN_REQUEST_CONTINUE */ unsigned int id; + struct ip_addr ip_local; size_t data_size; /* unsigned char data[]; */ diff -u -r work.patched/dovecot-0.99.10/src/auth/master-connection.c work/dovecot-0.99.10/src/auth/master-connection.c --- work.patched/dovecot-0.99.10/src/auth/master-connection.c Fri Aug 22 17:56:18 2003 +++ work/dovecot-0.99.10/src/auth/master-connection.c Fri Oct 24 13:57:35 2003 @@ -133,8 +133,9 @@ } send_reply(&failure_reply, sizeof(failure_reply), request->tag); } else { - userdb->lookup(auth_request->user, userdb_callback, - POINTER_CAST(request->tag)); + userdb->lookup(auth_request->user, + net_ip2addr(&auth_request->ip_local), + userdb_callback, POINTER_CAST(request->tag)); mech_request_free(login_conn, auth_request, request->id); } } diff -u -r work.patched/dovecot-0.99.10/src/auth/mech.c work/dovecot-0.99.10/src/auth/mech.c --- work.patched/dovecot-0.99.10/src/auth/mech.c Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/mech.c Fri Oct 24 13:57:51 2003 @@ -97,6 +97,7 @@ auth_request->created = ioloop_time; auth_request->conn = conn; auth_request->id = request->id; + auth_request->ip_local = request->ip_local; auth_request->protocol = request->protocol; hash_insert(conn->auth_requests, POINTER_CAST(request->id), diff -u -r work.patched/dovecot-0.99.10/src/auth/mech.h work/dovecot-0.99.10/src/auth/mech.h --- work.patched/dovecot-0.99.10/src/auth/mech.h Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/mech.h Fri Oct 24 13:58:02 2003 @@ -14,6 +14,7 @@ struct login_connection *conn; unsigned int id; + struct ip_addr ip_local; time_t created; enum auth_protocol protocol; diff -u -r work.patched/dovecot-0.99.10/src/auth/passdb-ldap.c work/dovecot-0.99.10/src/auth/passdb-ldap.c --- work.patched/dovecot-0.99.10/src/auth/passdb-ldap.c Thu Mar 20 09:55:57 2003 +++ work/dovecot-0.99.10/src/auth/passdb-ldap.c Fri Oct 24 13:58:20 2003 @@ -145,7 +145,8 @@ passdb_ldap_conn->attr_names[ATTR_VIRTUAL_USER], user); } else { str = t_str_new(512); - var_expand(str, conn->set.pass_filter, user, NULL); + var_expand(str, conn->set.pass_filter, user, NULL, + net_ip2addr(&auth_request->ip_local)); filter = str_c(str); } diff -u -r work.patched/dovecot-0.99.10/src/auth/passdb-mysql.c work/dovecot-0.99.10/src/auth/passdb-mysql.c --- work.patched/dovecot-0.99.10/src/auth/passdb-mysql.c Fri Aug 29 17:47:24 2003 +++ work/dovecot-0.99.10/src/auth/passdb-mysql.c Fri Oct 24 13:58:31 2003 @@ -109,7 +109,8 @@ str = t_str_new(512); var_expand(str, conn->set.password_query, - str_escape(auth_request->user), NULL); + str_escape(auth_request->user), NULL, + net_ip2addr(&auth_request->ip_local)); query = str_c(str); mysql_request->callback = mysql_handle_request; diff -u -r work.patched/dovecot-0.99.10/src/auth/passdb-pgsql.c work/dovecot-0.99.10/src/auth/passdb-pgsql.c --- work.patched/dovecot-0.99.10/src/auth/passdb-pgsql.c Fri Apr 4 06:17:25 2003 +++ work/dovecot-0.99.10/src/auth/passdb-pgsql.c Fri Oct 24 14:01:08 2003 @@ -105,7 +105,8 @@ str = t_str_new(512); var_expand(str, conn->set.password_query, - str_escape(auth_request->user), NULL); + str_escape(auth_request->user), NULL, + net_ip2addr(&auth_request->ip_local)); query = str_c(str); pgsql_request->callback = pgsql_handle_request; diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-ldap.c work/dovecot-0.99.10/src/auth/userdb-ldap.c --- work.patched/dovecot-0.99.10/src/auth/userdb-ldap.c Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/userdb-ldap.c Fri Oct 24 15:46:21 2003 @@ -142,8 +142,8 @@ t_pop(); } -static void userdb_ldap_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void userdb_ldap_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { struct ldap_connection *conn = userdb_ldap_conn->conn; struct userdb_ldap_request *request; @@ -156,7 +156,7 @@ userdb_ldap_conn->attr_names[ATTR_VIRTUAL_USER], user); } else { str = t_str_new(512); - var_expand(str, conn->set.user_filter, user, NULL); + var_expand(str, conn->set.user_filter, user, NULL, ip_local); filter = str_c(str); } diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-mysql.c work/dovecot-0.99.10/src/auth/userdb-mysql.c --- work.patched/dovecot-0.99.10/src/auth/userdb-mysql.c Fri Oct 24 15:07:12 2003 +++ work/dovecot-0.99.10/src/auth/userdb-mysql.c Fri Oct 24 14:58:08 2003 @@ -115,8 +115,8 @@ } } -static void userdb_mysql_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void userdb_mysql_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { struct mysql_connection *conn = userdb_mysql_conn->conn; struct userdb_mysql_request *request; @@ -124,7 +124,8 @@ string_t *str; str = t_str_new(512); - var_expand(str, conn->set.user_query, str_escape(user), NULL); + var_expand(str, conn->set.user_query, str_escape(user), NULL, + ip_local); query = str_c(str); request = i_malloc(sizeof(struct userdb_mysql_request) + strlen(user)); diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-passwd-file.c work/dovecot-0.99.10/src/auth/userdb-passwd-file.c --- work.patched/dovecot-0.99.10/src/auth/userdb-passwd-file.c Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/userdb-passwd-file.c Fri Oct 24 15:28:09 2003 @@ -11,8 +11,8 @@ struct passwd_file *userdb_pwf = NULL; -static void passwd_file_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void passwd_file_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { struct user_data data; struct passwd_user *pu; diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-passwd.c work/dovecot-0.99.10/src/auth/userdb-passwd.c --- work.patched/dovecot-0.99.10/src/auth/userdb-passwd.c Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/userdb-passwd.c Fri Oct 24 15:28:26 2003 @@ -10,8 +10,8 @@ #include <pwd.h> -static void passwd_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void passwd_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { struct user_data data; struct passwd *pw; diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-pgsql.c work/dovecot-0.99.10/src/auth/userdb-pgsql.c --- work.patched/dovecot-0.99.10/src/auth/userdb-pgsql.c Sun May 18 07:26:28 2003 +++ work/dovecot-0.99.10/src/auth/userdb-pgsql.c Fri Oct 24 15:45:51 2003 @@ -84,8 +84,8 @@ } } -static void userdb_pgsql_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void userdb_pgsql_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { struct pgsql_connection *conn = userdb_pgsql_conn->conn; struct userdb_pgsql_request *request; @@ -93,7 +93,7 @@ string_t *str; str = t_str_new(512); - var_expand(str, conn->set.user_query, str_escape(user), NULL); + var_expand(str, conn->set.user_query, str_escape(user), NULL, ip_local); query = str_c(str); request = i_malloc(sizeof(struct userdb_pgsql_request) + strlen(user)); diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-static.c work/dovecot-0.99.10/src/auth/userdb-static.c --- work.patched/dovecot-0.99.10/src/auth/userdb-static.c Wed Feb 19 18:46:51 2003 +++ work/dovecot-0.99.10/src/auth/userdb-static.c Fri Oct 24 15:30:40 2003 @@ -16,8 +16,8 @@ static gid_t static_gid; static char *static_home_template; -static void static_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void static_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { struct user_data data; string_t *str; @@ -29,7 +29,7 @@ data.virtual_user = data.system_user = user; str = t_str_new(256); - var_expand(str, static_home_template, user, NULL); + var_expand(str, static_home_template, user, NULL, ip_local); data.home = str_c(str); callback(&data, context); diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb-vpopmail.c work/dovecot-0.99.10/src/auth/userdb-vpopmail.c --- work.patched/dovecot-0.99.10/src/auth/userdb-vpopmail.c Thu Mar 20 09:46:33 2003 +++ work/dovecot-0.99.10/src/auth/userdb-vpopmail.c Fri Oct 24 15:29:24 2003 @@ -46,8 +46,8 @@ #ifdef USERDB_VPOPMAIL -static void vpopmail_lookup(const char *user, userdb_callback_t *callback, - void *context) +static void vpopmail_lookup(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context) { char vpop_user[VPOPMAIL_LIMIT], vpop_domain[VPOPMAIL_LIMIT]; struct vqpasswd *vpw; diff -u -r work.patched/dovecot-0.99.10/src/auth/userdb.h work/dovecot-0.99.10/src/auth/userdb.h --- work.patched/dovecot-0.99.10/src/auth/userdb.h Fri Aug 22 13:07:53 2003 +++ work/dovecot-0.99.10/src/auth/userdb.h Fri Oct 24 15:47:11 2003 @@ -17,8 +17,8 @@ void (*init)(const char *args); void (*deinit)(void); - void (*lookup)(const char *user, userdb_callback_t *callback, - void *context); + void (*lookup)(const char *user, const char *ip_local, + userdb_callback_t *callback, void *context); }; extern struct userdb_module *userdb; diff -u -r work.patched/dovecot-0.99.10/src/imap-login/client.c work/dovecot-0.99.10/src/imap-login/client.c --- work.patched/dovecot-0.99.10/src/imap-login/client.c Wed Jun 25 20:02:04 2003 +++ work/dovecot-0.99.10/src/imap-login/client.c Fri Oct 24 14:21:46 2003 @@ -49,7 +49,7 @@ if (!verbose_proctitle || !process_per_connection) return; - addr = net_ip2addr(&client->common.ip); + addr = net_ip2addr(&client->common.ip_remote); if (addr == NULL) addr = "??"; @@ -126,7 +126,7 @@ client->common.io = NULL; } - fd_ssl = ssl_proxy_new(client->common.fd, &client->common.ip); + fd_ssl = ssl_proxy_new(client->common.fd, &client->common.ip_remote); if (fd_ssl != -1) { client->tls = TRUE; client_set_title(client); @@ -338,7 +338,8 @@ } } -struct client *client_create(int fd, struct ip_addr *ip, int ssl) +struct client *client_create(int fd, struct ip_addr *ip_remote, + struct ip_addr *ip_local, int ssl) { struct imap_client *client; @@ -357,7 +358,8 @@ client->refcount = 1; client->tls = ssl; - client->common.ip = *ip; + client->common.ip_remote = *ip_remote; + client->common.ip_local = *ip_local; client->common.fd = fd; client_open_streams(client, fd); @@ -439,7 +441,7 @@ { const char *addr; - addr = net_ip2addr(&client->common.ip); + addr = net_ip2addr(&client->common.ip_remote); if (addr == NULL) addr = "??"; diff -u -r work.patched/dovecot-0.99.10/src/imap-login/client.h work/dovecot-0.99.10/src/imap-login/client.h --- work.patched/dovecot-0.99.10/src/imap-login/client.h Tue Mar 4 19:43:04 2003 +++ work/dovecot-0.99.10/src/imap-login/client.h Fri Oct 24 14:23:02 2003 @@ -30,7 +30,8 @@ unsigned int destroyed:1; }; -struct client *client_create(int fd, struct ip_addr *ip, int ssl); +struct client *client_create(int fd, struct ip_addr *ip_remote, + struct ip_addr *ip_local, int ssl); void client_destroy(struct imap_client *client, const char *reason); void client_ref(struct imap_client *client); diff -u -r work.patched/dovecot-0.99.10/src/lib/var-expand.c work/dovecot-0.99.10/src/lib/var-expand.c --- work.patched/dovecot-0.99.10/src/lib/var-expand.c Sun Jan 26 19:27:51 2003 +++ work/dovecot-0.99.10/src/lib/var-expand.c Fri Oct 17 13:12:51 2003 @@ -5,7 +5,7 @@ #include "var-expand.h" void var_expand(string_t *dest, const char *str, - const char *user, const char *home) + const char *user, const char *home, const char *ip) { const char *var; unsigned int width; @@ -36,6 +36,9 @@ case 'd': var = strchr(user, '@'); if (var != NULL) var++; + break; + case 'i': + var = ip; break; default: str_append_c(dest, '%'); diff -u -r work.patched/dovecot-0.99.10/src/lib/var-expand.h work/dovecot-0.99.10/src/lib/var-expand.h --- work.patched/dovecot-0.99.10/src/lib/var-expand.h Sun Jan 26 19:27:51 2003 +++ work/dovecot-0.99.10/src/lib/var-expand.h Fri Oct 17 13:13:08 2003 @@ -6,8 +6,9 @@ %u user or user@domain %h home %n user - %d domain */ + %d domain + %i ip */ void var_expand(string_t *dest, const char *str, - const char *user, const char *home); + const char *user, const char *home, const char *ip); #endif diff -u -r work.patched/dovecot-0.99.10/src/login-common/auth-connection.c work/dovecot-0.99.10/src/login-common/auth-connection.c --- work.patched/dovecot-0.99.10/src/login-common/auth-connection.c Mon May 26 10:27:13 2003 +++ work/dovecot-0.99.10/src/login-common/auth-connection.c Fri Oct 24 14:47:17 2003 @@ -325,6 +325,7 @@ auth_request.protocol = protocol; auth_request.mech = request->mech; auth_request.id = request->id; + auth_request.ip_local = ((struct client *) request->context)->ip_local; if (o_stream_send(request->conn->output, &auth_request, sizeof(auth_request)) < 0) { errno = request->conn->output->stream_errno; @@ -342,6 +343,7 @@ /* send continued request to auth */ auth_request.type = AUTH_LOGIN_REQUEST_CONTINUE; auth_request.id = request->id; + auth_request.ip_local = ((struct client *) request->context)->ip_local; auth_request.data_size = data_size; if (o_stream_send(request->conn->output, &auth_request, diff -u -r work.patched/dovecot-0.99.10/src/login-common/client-common.h work/dovecot-0.99.10/src/login-common/client-common.h --- work.patched/dovecot-0.99.10/src/login-common/client-common.h Tue Mar 4 19:43:04 2003 +++ work/dovecot-0.99.10/src/login-common/client-common.h Fri Oct 24 14:23:58 2003 @@ -5,7 +5,8 @@ #include "master.h" struct client { - struct ip_addr ip; + struct ip_addr ip_remote; + struct ip_addr ip_local; int fd; struct io *io; @@ -17,7 +18,8 @@ /* ... */ }; -struct client *client_create(int fd, struct ip_addr *ip, int ssl); +struct client *client_create(int fd, struct ip_addr *ip_remote, + struct ip_addr *ip_local, int ssl); unsigned int clients_get_count(void); void clients_notify_auth_process(void); diff -u -r work.patched/dovecot-0.99.10/src/login-common/main.c work/dovecot-0.99.10/src/login-common/main.c --- work.patched/dovecot-0.99.10/src/login-common/main.c Sun Jun 22 18:09:48 2003 +++ work/dovecot-0.99.10/src/login-common/main.c Fri Oct 24 14:25:02 2003 @@ -75,28 +75,30 @@ static void login_accept(void *context __attr_unused__) { - struct ip_addr ip; + struct ip_addr ip_remote, ip_local; int fd; - fd = net_accept(LOGIN_LISTEN_FD, &ip, NULL); + fd = net_accept(LOGIN_LISTEN_FD, &ip_remote, NULL); if (fd < 0) { if (fd < -1) i_fatal("accept() failed: %m"); return; } + net_getsockname(fd, &ip_local, NULL); + if (process_per_connection) main_close_listen(); - (void)client_create(fd, &ip, FALSE); + (void)client_create(fd, &ip_remote, &ip_local, FALSE); } static void login_accept_ssl(void *context __attr_unused__) { - struct ip_addr ip; + struct ip_addr ip_remote, ip_local; int fd, fd_ssl; - fd = net_accept(LOGIN_SSL_LISTEN_FD, &ip, NULL); + fd = net_accept(LOGIN_SSL_LISTEN_FD, &ip_remote, NULL); if (fd < 0) { if (fd < -1) i_fatal("accept() failed: %m"); @@ -106,11 +108,12 @@ if (process_per_connection) main_close_listen(); - fd_ssl = ssl_proxy_new(fd, &ip); + fd_ssl = ssl_proxy_new(fd, &ip_remote); if (fd_ssl == -1) net_disconnect(fd); - else - (void)client_create(fd_ssl, &ip, TRUE); + + net_getsockname(fd_ssl, &ip_local, NULL); + (void)client_create(fd_ssl, &ip_remote, &ip_local, TRUE); } static void open_logfile(const char *name) @@ -219,7 +222,7 @@ int main(int argc __attr_unused__, char *argv[], char *envp[]) { const char *name, *group_name; - struct ip_addr ip; + struct ip_addr ip_remote, ip_local; int i, fd = -1, master_fd = -1; is_inetd = getenv("DOVECOT_MASTER") == NULL; @@ -257,7 +260,7 @@ main_init(); if (is_inetd) { - if (net_getsockname(1, &ip, NULL) < 0) { + if (net_getsockname(1, &ip_remote, NULL) < 0) { i_fatal("%s can be started only through dovecot " "master process, inetd or equilevant", argv[0]); } @@ -265,7 +268,7 @@ fd = 1; for (i = 1; i < argc; i++) { if (strcmp(argv[i], "--ssl") == 0) { - fd = ssl_proxy_new(fd, &ip); + fd = ssl_proxy_new(fd, &ip_remote); if (fd == -1) i_fatal("SSL initialization failed"); } else if (strncmp(argv[i], "--group=", 8) != 0) @@ -276,8 +279,10 @@ closing_down = TRUE; } - if (fd != -1) - (void)client_create(fd, &ip, TRUE); + if (fd != -1) { + net_getsockname(fd, &ip_local, NULL); + (void)client_create(fd, &ip_remote, &ip_local, TRUE); + } io_loop_run(ioloop); main_deinit(); diff -u -r work.patched/dovecot-0.99.10/src/login-common/master.c work/dovecot-0.99.10/src/login-common/master.c --- work.patched/dovecot-0.99.10/src/login-common/master.c Tue Mar 4 19:43:04 2003 +++ work/dovecot-0.99.10/src/login-common/master.c Fri Oct 24 14:25:57 2003 @@ -44,7 +44,7 @@ req.tag = client->fd; req.auth_pid = auth_pid; req.auth_id = auth_id; - req.ip = client->ip; + req.ip = client->ip_remote; if (fd_send(master_fd, client->fd, &req, sizeof(req)) != sizeof(req)) i_fatal("fd_send(%d) failed: %m", client->fd); diff -u -r work.patched/dovecot-0.99.10/src/master/mail-process.c work/dovecot-0.99.10/src/master/mail-process.c --- work.patched/dovecot-0.99.10/src/master/mail-process.c Fri Aug 22 18:04:27 2003 +++ work/dovecot-0.99.10/src/master/mail-process.c Fri Oct 17 13:25:54 2003 @@ -91,7 +91,7 @@ } /* expand %vars */ - var_expand(str, env, user, home); + var_expand(str, env, user, home, NULL); return str_c(str); } diff -u -r work.patched/dovecot-0.99.10/src/pop3-login/client.c work/dovecot-0.99.10/src/pop3-login/client.c --- work.patched/dovecot-0.99.10/src/pop3-login/client.c Mon May 26 10:27:13 2003 +++ work/dovecot-0.99.10/src/pop3-login/client.c Fri Oct 24 14:27:50 2003 @@ -42,7 +42,7 @@ if (!verbose_proctitle || !process_per_connection) return; - addr = net_ip2addr(&client->common.ip); + addr = net_ip2addr(&client->common.ip_remote); if (addr == NULL) addr = "??"; @@ -80,7 +80,7 @@ client->common.io = NULL; } - fd_ssl = ssl_proxy_new(client->common.fd, &client->common.ip); + fd_ssl = ssl_proxy_new(client->common.fd, &client->common.ip_remote); if (fd_ssl != -1) { client->tls = TRUE; client_set_title(client); @@ -231,7 +231,8 @@ } } -struct client *client_create(int fd, struct ip_addr *ip, int ssl) +struct client *client_create(int fd, struct ip_addr *ip_remote, + struct ip_addr *ip_local, int ssl) { struct pop3_client *client; @@ -250,7 +251,8 @@ client->refcount = 1; client->tls = ssl; - client->common.ip = *ip; + client->common.ip_remote = *ip_remote; + client->common.ip_local = *ip_local; client->common.fd = fd; client->common.io = io_add(fd, IO_READ, client_input, client); client_open_streams(client, fd); @@ -322,7 +324,7 @@ { const char *addr; - addr = net_ip2addr(&client->common.ip); + addr = net_ip2addr(&client->common.ip_remote); if (addr == NULL) addr = "??"; diff -u -r work.patched/dovecot-0.99.10/src/pop3-login/client.h work/dovecot-0.99.10/src/pop3-login/client.h --- work.patched/dovecot-0.99.10/src/pop3-login/client.h Tue Mar 4 19:43:04 2003 +++ work/dovecot-0.99.10/src/pop3-login/client.h Fri Oct 24 14:28:05 2003 @@ -24,7 +24,8 @@ unsigned int destroyed:1; }; -struct client *client_create(int fd, struct ip_addr *ip, int ssl); +struct client *client_create(int fd, struct ip_addr *ip_remote, + struct ip_addr *ip_local, int ssl); void client_destroy(struct pop3_client *client, const char *reason); void client_ref(struct pop3_client *client);

1 0

Re: [Dovecot] dovecot vs cyrus, uw, etc.
by Scott Klein 04 Nov '03

04 Nov '03

Thanks, all. I'm definitely leaning toward Dovecot. Cyrus is just too big a pain for very little payoff.

1 0

[Dovecot] current CVS seems to have index problems.
by Andreas Jaekel 04 Nov '03

04 Nov '03

Hi Timo. I've checked out the current CVS version and tested it. There seems to be a problem when dovecot accesses mailboxes for the first time: I have a mailbox with 4 mails, and no special IMAP headers. (no X-IMAPBase, X-UID etc.) and it's in mbox format. I log in, SELECT the folder and do FETCH 2 BODY[] and receive only a snipplet of the mail from somewhere in the middle of it. After some debugging it seems like dovecot: 1. creates an index, for mbox_index_rewrite(), 2. rewrites the mbox, but then 3. forgets to update the index after the rewrite. and ends up in a state with a pre-rewrite index but a post-rewrite folder. Now, there are many ways to make dovecot rebuild the index after rewriting the folder, but some are ugly and some are not, some need extra code to avoid endless loops. I don't know how you would want this fixed. I attached a patch that worked for me. After that patch, I have more (new?) problems, which may or may not be related. Now, when I SELECT a folder and do a COPY within the folder (*smile*) I get this: A COPY 1 INBOX * BYE Mailbox is in inconsistent state, please relogin. Connection closed by foreign host. ... which is what I'm gonna start looking for right now. :) Regards, Andy

1 1

[Dovecot] -test8 & Mail.app
by Dan Sully 04 Nov '03

04 Nov '03

A user just tried to use OSX 10.3's Mail.app via imap. It caused the process to start eating CPU, and this in the logs. Any idea what could be happening here? What information should I get on his setup? imap-login: Nov 03 14:28:37 Info: Login: fenner [12.106.35.5] imap(fenner): Nov 03 14:30:37 Error: Timeout while waiting for release of exclusive fcntl() lock for index file /home/f/fenner/M>aildir/.INBOX/.imap.index imap(fenner): Nov 03 14:32:37 Error: Timeout while waiting for release of exclusive fcntl() lock for index file /home/f/fenner/M>aildir/.INBOX/.imap.index imap(fenner): Nov 03 14:32:37 Panic: file index-messageset.c: line 346 (index_messageset_next): assertion failed: (!(ctx->expung>es->uid1 <= mail->rec->uid && ctx->expunges->uid2 >= mail->rec->uid)) dovecot: Nov 03 14:32:37 Error: child 26371 (imap) killed with signal 6 Thanks. -D -- Ooo-Eee-Ooo-Ah-Ah-Ting-Tang-Walla-Walla-Bing-Bang

1 0

[Dovecot] dies on start-up: Auth process
by Tom Allison 04 Nov '03

04 Nov '03

I was digging through some of the archives and ran into someone who was having the same problem, but for completely different reasons....i think. dovecot: Nov 02 07:58:51 Info: Dovecot starting up dovecot: Nov 02 07:58:52 Error: Auth process died too early - shutting down dovecot: Nov 02 07:58:52 Error: child 27673 (login) killed with signal 11 dovecot: Nov 02 07:58:52 Error: child 27672 (login) killed with signal 11 dovecot: Nov 02 07:58:52 Error: child 27671 (login) killed with signal 11 dovecot: Nov 02 07:58:52 Error: child 27670 (auth) killed with signal 11 So I took the advice posted and ran 'strace dovecot' and this is what I got. Unfortunately I'm not exactly good enough with this to even know what I'm looking at. But I suspect the notation about "/dev/null" might have something to do with it. my user, dovecot as a shell of /bin/bash and a $HOME of /usr/local/lib/dovecot By all rights, he's almost a regular user, except that his UID is <100 (I built user:dovecot as a system uid under SuSE 8.2) open("/var/run/dovecot/", O_RDONLY|O_LARGEFILE) = 3 fstat64(3, {st_mode=S_IFDIR|0700, st_size=4096, ...}) = 0 close(3) = 0 lstat64("/var/run/dovecot/", {st_mode=S_IFDIR|0700, st_size=4096, ...}) = 0 open(".", O_RDONLY|O_LARGEFILE) = 3 open("/var/run/dovecot//login", O_RDONLY|O_LARGEFILE|O_NOFOLLOW) = 4 fchdir(4) = 0 open("/dev/null", O_RDONLY|O_NONBLOCK|O_DIRECTORY) = -1 ENOTDIR (Not a directory) open(".", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 6 fstat64(6, {st_mode=S_IFDIR|0750, st_size=4096, ...}) = 0 fcntl64(6, F_SETFD, FD_CLOEXEC) = 0 brk(0) = 0x8061000 brk(0x8062000) = 0x8062000 getdents64(6, /* 2 entries */, 4096) = 48 getdents64(6, /* 0 entries */, 4096) = 0 close(4) = 0 close(6) = 0 fchdir(3) = 0 close(3) = 0 geteuid32() = 0 lstat64("/var/run/dovecot//login", {st_mode=S_IFDIR|0750, st_size=4096, ...}) = 0 open("/var/run/dovecot//login", O_RDONLY|O_LARGEFILE) = 3 fstat64(3, {st_mode=S_IFDIR|0750, st_size=4096, ...}) = 0 close(3) = 0 lstat64("/var/run/dovecot//login", {st_mode=S_IFDIR|0750, st_size=4096, ...}) = 0 access("/usr/local/libexec/dovecot/dovecot-auth", X_OK) = 0 open("/dev/null", O_RDONLY|O_LARGEFILE) = 3 fcntl64(3, F_GETFD) = 0 fcntl64(3, F_SETFD, FD_CLOEXEC) = 0 dup(3) = 4 fcntl64(4, F_GETFD) = 0 fcntl64(4, F_SETFD, FD_CLOEXEC) = 0 socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 6 setsockopt(6, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 setsockopt(6, SOL_SOCKET, SO_KEEPALIVE, [1], 4) = 0 bind(6, {sa_family=AF_INET, sin_port=htons(143), sin_addr=inet_addr("0.0.0.0")}, 16) = 0 getsockname(6, {sa_family=AF_INET, sin_port=htons(143), sin_addr=inet_addr("0.0.0.0")}, [16]) = 0 listen(6, 8) = 0 fcntl64(6, F_SETFL, O_RDONLY|O_NONBLOCK) = 0 fcntl64(6, F_GETFD) = 0 fcntl64(6, F_SETFD, FD_CLOEXEC) = 0 dup(4) = 7 fcntl64(7, F_GETFD) = 0 fcntl64(7, F_SETFD, FD_CLOEXEC) = 0 dup(4) = 8 fcntl64(8, F_GETFD) = 0 fcntl64(8, F_SETFD, FD_CLOEXEC) = 0 dup(4) = 9 fcntl64(9, F_GETFD) = 0 fcntl64(9, F_SETFD, FD_CLOEXEC) = 0 dup2(4, 0) = 0 dup2(4, 1) = 1 dup2(4, 2) = 2 fork() = 27706 exit_group(0) = ?

4 10

Re: [Dovecot] Re: Bug with partial IMAP fetches
by Matthew Reimer 03 Nov '03

03 Nov '03

Timo Sirainen wrote: > On Fri, Oct 31, 2003 at 04:43:55PM -0600, Matthew Reimer wrote: > >>A little more information: >> >>This bug results in Mozilla and pine never rendering the message, since >>dovecot lies about how much data was really sent, as seen in the >>following log in which only 394 bytes (rather than the 10240 dovecot is >>claiming to send) are returned: > > > Does this patch fix it? http://dovecot.fi/fetch-body.patch No; I'm already using all available 0.99.10 patches. Sorry for not mentioning that before. Matt

1 0

[Dovecot] Re: Bug with partial IMAP fetches
by Matthew Reimer 03 Nov '03

03 Nov '03

A little more information: This bug results in Mozilla and pine never rendering the message, since dovecot lies about how much data was really sent, as seen in the following log in which only 394 bytes (rather than the 10240 dovecot is claiming to send) are returned: ---- 8 UID fetch 94 (UID RFC822.SIZE BODY[]<10240.10240>) * 24 FETCH (UID 94 RFC822.SIZE 20873 BODY[]<10240> {10240} G SRC="/world/image/nav/contents_bottom.gif" WIDTH="49" HEIGHT="15" BORDER="0" VSPACE="10"></A></TD></TR></TABLE>  <script type="text/javascript"> if(parent.frames.length>0&&!isContents) top.location.replace(document.location) </script> </BODY> </HTML>  --------------010405000308030207040500-- ) 8 OK Fetch completed. ---- Matt

2 1

[Dovecot] Debian bug #201444: More verbose logging
by Jaldhar H. Vyas 03 Nov '03

03 Nov '03

I brought this up a while ago but the submitter is wondering if there has been any progress on this issue. He says: It would be extremely nice to have a little bit more detailed logging, and I believe it doesn't really require that much code. I use now a daemon which logs login (name, ip; as dovecot does), and logout (with name, logout reason if any special) with the numbers of "read new mail number and size" and "mail left on server number and size". Would be nice to have something like that in dovecot. Could something like this be added for the next version? -- Jaldhar H. Vyas <jaldhar(a)debian.org> La Salle Debain - http://www.braincells.com/debian/

3 4

[Dovecot] Problems finding information on passwd-file
by David A. Desrosiers 03 Nov '03

03 Nov '03

I've recently converted over to dovecot to uw-imap, and am having a heck of a time trying to find actual docs that describe how to set up virtual users properly without using ldap or posgres.. Basically I need to know what goes in auth_userdb and auth_passdb for a system with *NO* local user accounts in /etc/passwd and /etc/shadow. I will only create mbox files for these users in /var/mail/%u and alternately a home directory where they can store their own mail folders under /home/%u. How can I do this, given the current dovecot codebase? I've googled for a few hours, talked to many people on irc, and checked the mailing lists (of what I could find) related to this, and found nothing whatsoever describing how to achive this. Dovecot apparently supports "virtual users", but I can't find any supporting documentation that describes exactly HOW to set that up. Thanks in advance. d.

2 1

Re: [Dovecot] evolution - folders not displayed correctly
by Terry 03 Nov '03

03 Nov '03

--- Jonas Smedegaard <dr(a)jones.dk> wrote: > s�n, 2003-11-02 kl. 22:54 skrev Terry: > ... > > I tried my account in evolution and all the > folders > > appear in one level and not in several as they > were in > > thunderbird. > ... > > What could be the problem? Thanks! > > The documentation of offlineimap seems to reveil the > cause of the > problem: Evolution uses "/" as delimiter, not "." as > dovecot. > > It seems to not be configurable in Evolution, so > access the files > through dovecot instead of directly. > > - Jonas I am accessing my files via dovecot. That is why I am mailing the dovecot list. Maybe I misunderstood your response. I will check around to see that deliter...probably in the xml configs somewhere.... ===== Terry __________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com

2 1

[Dovecot] Bug with partial IMAP fetches
by Matthew Reimer 01 Nov '03

01 Nov '03

X-UIDL: 61c08992b287bf24b6fff123164620b2 Return-Path: <mreimer(a)vpop.net> Delivered-To: krista(a)vpop.net Received: from vpop.net (bilbo.vpop.net [65.103.33.41]) by boromir.vpop.net (Postfix) with ESMTP id 6B7D43A60F6 for <krista(a)vpop.net>; Sun, 6 Oct 2002 21:16:38 -0700 (PDT) Message-ID: <3DA10B4B.50405(a)vpop.net> Date: Sun, 06 Oct 2002 23:19:23 -0500 From: Matthew Reimer <mreimer(a)vpop.net> Organization: VPOP Technologies, Inc. User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.1) Gecko/20020829 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Krista Reimer <krista(a)vpop.net> Subject: WORLD October 12, 2002: Eight principles Content-Type: multipart/mixed; boundary="------------010405000308030207040500" This is a multi-part message in MIME format. --------------010405000308030207040500 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Lemme know if this doesn't come through right. http://www.worldmag.com/world/issue/10-12-02/closing_1.asp Matt --------------010405000308030207040500 Content-Type: text/html; charset=WINDOWS\252; name="www.worldmag.com/world/issue/10-12-02/closing_1.asp" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="www.worldmag.com/world/issue/10-12-02/closing_1.asp" Content-Base: "http://www.worldmag.com/world/issue/10 -12-02/closing_1.asp" Content-Location: "http://www.worldmag.com/world/issue/10 -12-02/closing_1.asp" <DOCTYPE HTML PUBLIC "-//IETF//DTD HTML3.0//EN" "html.dtd"> <HTML><HEAD><TITLE>WORLD October 12, 2002: Eight principles</TITLE> <META HTTP-EQUIV="Content-type" CONTENT="text/html; charset=Windows\252"> <META NAME="ISSUE" CONTENT="2002-10-12"> <META NAME="ISSUEFOLDER" CONTENT="10-12-02"> <META NAME="DESCRIPTION" CONTENT="How Christians should relate to those of other faiths : 1. We should not conceal aspects of our faith so as to avoid criticism or disapproval, and we should not ask others to do that. We believe it is just and loving to point out publicly the errors of other faiths, provided this is done with suf..."> <META NAME="ARTICLEAUTHOR" CONTENT=""></HEAD> <BODY BGCOLOR=ffffff BACKGROUND=/world/image/bg.gif LINK=#000099 VLINK=#6699CC ALINK=#0000FF> <A NAME="top">  <script type="text/javascript" src="/world/world_ad.js"></script> <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0" WIDTH="520"> <TR><TD WIDTH=20 VALIGN=TOP ALIGN=LEFT><IMG SRC="/world/image/dot_red.gif" WIDTH="20" HEIGHT="56" BORDER="0"></TD><TD WIDTH=500 VALIGN=MIDDLE ALIGN=LEFT><A TARGET="_top" HREF="/world/adredir.asp?url=http://stats.schwa.com/ss%3Fclick%26world%263a5f25c8&image=/world/image/ads/maiad.jpg" ><IMG SRC="/world/image/ads/maiad.jpg" ALT="Medical Ambassadors International" WIDTH=468 HEIGHT=60 BORDER=1></A> <BR></TD></TR></TABLE> <BR> <script type="text/javascript"> var isContents=false </script> <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0" WIDTH="720"> <TR><TD ALIGN=LEFT VALIGN=TOP WIDTH="4"><IMG SRC="/world/image/dot_red.gif" WIDTH="4" HEIGHT="1" BORDER="0"><BR></TD><TD ALIGN=LEFT VALIGN=MIDDLE WIDTH="110"><FONT SIZE="1" FACE="arial, helvetica">Call 1-800-951-NEWS (6397) to subscribe today! <BR></TD><TD ALIGN=LEFT VALIGN=MIDDLE WIDTH="38"><BR></TD><TD WIDTH="280" VALIGN=BOTTOM ALIGN=LEFT><IMG SRC="/world/image/world_header.gif" WIDTH="276" HEIGHT="51" BORDER="0"> </TD><TD WIDTH=98 VALIGN=MIDDLE ALIGN=LEFT><FONT SIZE="1" FACE="arial, helvetica">October 12, 2002<BR> Volume 17<BR> Number 39 </FONT><BR> </TD> <TD width="190"> <a href="http://www.weeklystandard.com/world"><img src="/world/image/side_ads/TWS-Wrld-2.gif" width="180" height="100"></a> </TD></TR></TABLE> <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0" WIDTH="720"><TR>  <TD ALIGN=LEFT VALIGN=TOP ROWSPAN="1" WIDTH="4"><IMG SRC="/world/image/dot_red.gif" WIDTH="4" HEIGHT="1" BORDER="0"></TD>   <TD ALIGN=LEFT VALIGN=TOP WIDTH="136"> <IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="11" BORDER="0"><BR><A HREF="home.asp" TARGET="_top"><IMG SRC="/world/image/nav/contents.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="opening_1.asp" TARGET="_top"><IMG SRC="/world/image/nav/opening.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="cover_1.asp" TARGET="_top"><IMG SRC="/world/image/nav/cover.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="national_1.asp" TARGET="_top"><IMG SRC="/world/image/nav/national.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="international_1.asp" TARGET="_top"><IMG SRC="/world/image/nav/international.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="cultural_1.asp" TARGET="_top"><IMG SRC="/world/image/nav/cultural.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><IMG SRC="/world/image/nav/closing_v.gif" WIDTH="131" HEIGHT="18" BORDER="0" VSPACE="0"><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="20" BORDER="0"><BR><A HREF="/world/meta/search/" TARGET="_top"><IMG SRC="/world/image/nav/search.gif" WIDTH="106" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="/world/meta/subscribe/" TARGET="_top"><IMG SRC="/world/image/nav/subscribe.gif" WIDTH="106" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="https://secure.palmcoastd.com/pcd/document?ikey=057FCIREN" TARGET="_top"><IMG SRC="/world/image/nav/renew.gif" WIDTH="106" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="/world/meta/archives/" TARGET="_top"><IMG SRC="/world/image/nav/archives.gif" WIDTH="106" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="/world/classifieds/" TARGET="_top"><IMG SRC="/world/image/nav/classifieds.gif" WIDTH="106" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR><IMG SRC="/world/image/dot_red.gif" WIDTH="1" HEIGHT="7" BORDER="0"><BR><A HREF="/world/meta/adrates/" TARGET="_top"><IMG SRC="/world/image/nav/adinfo.gif" WIDTH="106" HEIGHT="18" BORDER="0" VSPACE="0"></A><BR> <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0" WIDTH="110"><TR><TD ALIGN=LEFT VALIGN=TOP WIDTH="110"><FONT SIZE="1" FACE="arial, helvetica"><STYLE> A.side {color:white;} </STYLE> <P><A class=side HREF=/world/issue/08-17-02/home.asp TARGET=_top><IMG SRC=/world/issue/08-17-02/image/cover_thumb.jpg WIDTH=72 HEIGHT=97 HSPACE=15 VSPACE=10 BORDER=0><BR><center>REMEMBERING 9/11 </center></A> <P>Read <A TARGET=_top HREF=/world/olasky/centralideas/index.html class=side>Central Ideas in the Development of American Journalism</a>, <A TARGET=_top HREF=/world/olasky/truth1.html class=side>Telling the Truth</A> and <A TARGET=_top HREF=/world/olasky/Prodigal/index.html class=side>Prodigal Press</a>, three books on journalism by Marvin Olasky.  <P> <A TARGET=_top HREF=http://www.gwbc.com/godsworld/worldad.asp class=side>Add Olasky, Belz, and other great thinkers to your reading list!</A> <P> <A class=side HREF=/world/meta/wear/ TARGET=_top> <IMG SRC=/world/image/world_wear_ad.gif WIDTH=99 HEIGHT=99 HSPACE=0 VSPACE=0 BORDER=1></A> <P> <A class=side HREF=/world/meta/subscribe/sub_standard.asp TARGET=_top> <IMG SRC=/world/image/bennett_book.gif WIDTH=75 HEIGHT=101 HSPACE=15 VSPACE=0 BORDER=1> <center>Free Book!</center></A> <P>Find out <A TARGET=_top HREF=/world/meta/contact/ class=side>how to contact us</A>. <P>Not familiar with WORLD Magazine? <A HREF=/world/meta/about/pr.asp TARGET=_top class=side>Find out more about us.</A> <P>Read our <A HREF=/world/meta/about/mission.asp TARGET=_top class=side>Mission Statement.</A> <P>If you are already a WORLD reader, head to <A HREF=/world/meta/help/ TARGET=_top class=side>Help Us Out</A> to find out how to help us grow.  <P>Find out <A HREF=/world/meta/about/about.asp TARGET=_top class=side>about this site</A>. <P>Excellence in Home Education starts here. Visit the new God's World Book Club Online! <A HREF=http://www.gwbc.com/godsworld TARGET=_top class=side>www.gwbc.com</A>. <p>Check out the World Journalism Institute website at <a href=http://www.worldji.com TARGET=_top class=side>www.worldji.com</a> and apply online! <p>Since 1981, God's World News has been helping children to understand and interpret the world around them. Now you can visit God's World News Online at <A HREF=http://www.gwnews.com TARGET=_top class=side>GWNews.com</a></font></body> <P></FONT> </TD> </TR> </TABLE> </TD>   <TD ALIGN=LEFT VALIGN=TOP WIDTH="16"> <IMG SRC="/world/image/dot_clr.gif" WIDTH="16" HEIGHT="1" BORDER="0"><BR></TD>   <TD ALIGN=LEFT VALIGN=TOP WIDTH="364">   <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0" BORDER="0" HEIGHT="45"> <TR VALIGN="MIDDLE"> <TD ALIGN="LEFT"><FORM METHOD="POST" ACTION="/world/redirect.asp"> <INPUT TYPE=HIDDEN NAME="folder" VALUE="10-12-02"> <SELECT NAME="where"> <OPTION VALUE="closing_1.asp">Also in CLOSING THOUGHTS <OPTION VALUE="closing_2.asp">What's the difference? </SELECT> </TD> <TD ALIGN="LEFT"> <INPUT TYPE=SUBMIT VALUE="go"></FORM> </TD></TR></TABLE>  <FONT SIZE="5" COLOR="CC0000" FACE="arial, helvetica"><B>JOHN PIPER</B></FONT><BR><IMG SRC="/world/image/line_364.gif" WIDTH="364" HEIGHT="1" BORDER="0"><BR> <H2>Eight principles</H2> <H3><I>How Christians should relate to those of other faiths</I></H3>   <P>1. We should not conceal aspects of our faith so as to avoid criticism or disapproval, and we should not ask others to do that. We believe it is just and loving to point out publicly the errors of other faiths, provided this is done with sufficient evidence that the sacred writings or representative spokesmen of those faiths do indeed express these errors. We must strive not to misrepresent other faiths, as that is not only disrespectful but also undermining of our own credibility.<P> 2. We should acknowledge our sin and desperate need of salvation by a crucified and risen Savior, and not posture ourselves as worthy of salvation as if we had superior intellect or wisdom or goodness. We are beggars who have, by grace, found the life-giving bread of truth, forgiveness, and joy. As we expose the errors of other religions, we should feel and express sorrow and compassion for those who do not embrace Christ so as to be saved.<P> 3. We should not expect a fair fight in a secular world that is hostile to God and uncomfortable around the truth of Christ. Therefore, our response to abuse, distortion, or slander should not be angry resentment, but patient witness to the truth, in the hope and with the prayer that returning good for evil may open hearts to the truth. We must recognize that persecution is normal and that much of the protection Christians have in America is abnormal in history and in the world. Our witness will be best advanced by kind, steadfast statements and reasonable defenses of the truth.<P> 4. We should renounce all violence as a means of spreading our faith. Christians spread biblical faith by suffering, not by causing suffering. Authentic Christianity cannot be coerced by force or manipulation. We should acknowledge and proclaim that Christ will, at his personal appearing, punish those who have rejected him; his violence at the end of the age is a decisive reason we should not and may not exert violence against anyone because of his beliefs. This is Christ's right, not ours.<P> 5. We believe that God has given to civil government, not individuals or the church, the duty to bear the sword for justice and safety. Civil authorities should not punish beliefs, but some behaviors rooted in beliefssuch as killing, assault, or theftmay be outside the law and therefore punishable by the civil authority. Which behaviors are legally prohibited in a society based on freedom of belief and religion will be determined in a process of persuasion, debate, and election of representative lawmakers, with checks and balances provided by the executive and judicial branches and by constitutional safeguards for the rights of minorities.<P> 6. We should distinguish between a just war of defense against aggression and a religious war against people because of their beliefs. We should acknowledge that this distinction will probably not be recognized by certain religions that define their beliefs to include the right of cultural domination by force. But we should insist on this distinction rather than accept the claim of aggressors that our resistance is a religious attack on their faith. We should argue that the ground of such national defense is the civil right to freedom (of religion and speech and press and assembly), not disapproval of the religion underlying the attack.<P> 7. We should distinguish between the right to express criticism of erroneous beliefs and sinful behaviors, on the one hand, and the false inference some draw from this criticism that proponents of the criticized beliefs can therefore legitimately be mistreated. We also should not accept the claim that being criticized or denounced as mistaken or as sinners is a form of mistreatment. It is not a crime (hate crime or otherwise) to say publicly that someone's belief is wrong and harmful, or to call someone's behavior sinful and destructive. A necessary part of all debate concerning beliefs, behaviors, or proposals is the argument that some are wrong, ill-founded, and have deleterious effects.<P> 8. We believe that every religion, worldview, or philosophy of life may freely endeavor to influence and shape our culture. We renounce the use of force or bribery or deceit in this culture-shaping effort. We affirm preaching the gospel, publishing truth, and modeling love and justice, as well as prayer, persuasion, and participation in the political process. We recognize that all laws "impose" some group's behavioral conviction on all, so it is not a compelling criticism to say that a law that governs behavior is bad because it "imposes someone's morality" on society. It is therefore particularly vital to support principles, laws, and policies that protect the legal freedoms of minorities that do not have the numbers to sway lawmaking processes.<P> These principles are further developed at the website http://www.desiringgod.org/library/topics/culture/tolerance_principles.html  <BR></TD><TD width="200" valign="top"> <link rel="Stylesheet" href="/world/right-bar-basic.css" type="text/css"></link> <style type="text/css"> @import: "/world/right-bar-advanced.css"; </style> <div id="rightbar"> <div id="listenlinks"> <h4><a target="_top" href="/world/listen/">Listen to WORLD</a></h4> <div>Daily radio shorts by Nick Eicher. <a target="_top" href="/world/listen/">Read more...</a><br /> <p><img src="/world/image/speaker.gif" height="11" width="12" /> <a target="_top" href="/world/listen/home.asp?file=000000201.mp3">10/7/02 download MP3</a></p> <p><img src="/world/image/speaker.gif" height="11" width="12" /> <a target="_top" href="/world/listen/home.asp?file=000000199.mp3">10/8/02 download MP3</a></p> <p><img src="/world/image/speaker.gif" height="11" width="12" /> <a target="_top" href="/world/listen/home.asp?file=000000198.mp3">10/9/02 download MP3</a></p> <p><img src="/world/image/speaker.gif" height="11" width="12" /> <a target="_top" href="/world/listen/home.asp?file=000000197.mp3">10/10/02 download MP3</a></p> <p><img src="/world/image/speaker.gif" height="11" width="12" /> <a target="_top" href="/world/listen/home.asp?file=000000196.mp3">10/11/02 download MP3</a></p> </div> </div> <div id="recentlinks"> <h4><a target="_top" href="/world/meta/archives/">From Previous Weeks</a></h4> <div> <p><a target="_top" href="/world/issue/10-05-02/"><img src="/world/issue/10-05-02/image/cover_thumb.jpg" width="72" height="97" /><br /> October 5, 2002</a></p> <p><a target="_top" href="/world/issue/09-28-02/"><img src="/world/issue/09-28-02/image/cover_thumb.jpg" width="72" height="97" /><br /> September 28, 2002</a></p> <p><a target="_top" href="/world/issue/09-21-02/"><img src="/world/issue/09-21-02/image/cover_thumb.jpg" width="72" height="97" /><br /> September 21, 2002</a></p> </div> <br /> <a href="http://worldmag.com/world/classifieds/"><img src="/world/image/side_ads/Searching_Ad_100x180.gif" width="180" height="100"></a> </div> </div> </TD></TR> </TABLE> <BR> <TABLE BORDER="0" WIDTH=520><TR><TD WIDTH=110 ALIGN=right VALIGN=top><FONT COLOR=ffffff><B>Search for more<BR>articles about:</B></FONT><BR></TD><TD WIDTH=40><BR></TD><TD WIDTH=370 VALIGN=top ALIGN=left BGCOLOR="#CCCC99"> <TABLE WIDTH="370" BORDER="0" CELLPADDING="8" CELLSPACING="0"><TR><TR><TD WIDTH="370" ALIGN="LEFT" VALIGN="top"><form name="searchForm" action="/world/meta/search/search.asp" onsubmit=""> <input type="hidden" name="performSearch" value="y"> <INPUT TYPE="text" NAME="target" SIZE="30" MAXLENGTH="100"><br> <input type=checkbox name="stype" value="phrase" checked> exact phrase </TD><TD ALIGN="LEFT" VALIGN="TOP"> <INPUT TYPE=SUBMIT VALUE="search"></FORM> </TD></TR> </TABLE> </TD></TR></TABLE> <TABLE BORDER="0" CELLPADDING="0" CELLSPACING="0" WIDTH="500"><TR><TD WIDTH="156"><BR></TD><TD ALIGN=LEFT WIDTH="253"><FONT SIZE="1"><A HREF="/world/meta/copyright.asp" TARGET="_top">© 1996 - 2002 WORLD</A> Magazine. <A HREF="/world/meta/contact/" TARGET="_top">mailbag(a)worldmag.com</A></FONT></TD><TD COLSPAN="1" ALIGN=RIGHT WIDTH="91"><A HREF="#top"><IMG SRC="/world/image/nav/top_bottom.gif" WIDTH="22" HEIGHT="15" HSPACE="10" VSPACE="10" BORDER="0"></A><A HREF="home.asp" TARGET="_top"><IMG SRC="/world/image/nav/contents_bottom.gif" WIDTH="49" HEIGHT="15" BORDER="0" VSPACE="10"></A></TD></TR></TABLE>  <script type="text/javascript"> if(parent.frames.length>0&&!isContents) top.location.replace(document.location) </script> </BODY> </HTML>  --------------010405000308030207040500--

1 0

Re: [Dovecot] SSL Certificate Chains?
by Matthew Reimer 29 Oct '03

29 Oct '03

Timo Sirainen wrote: > On Mon, 2003-10-27 at 15:01, Steve Gare wrote: > >>Just a quicky, I can't see anything about how Dovecot handles >>certificate chains. >> >>1, Does it handle them? >>2, How, if if does? >> >>(Just got a cert from www.instantssl.com and I need there intermediate >>cert do validate it.., Works fine with Apache and Exim ) > > > Dovecot doesn't check client's certificate. It's client's job to check > server's certificate and check the chains. Or at least I don't think > there's anything special Dovecot should do with them.. I had trouble with an instantssl cert, and found that what I needed to do was to also include all the certs up the chain, in a certain order, to keep the client from complaining about an invalid certificate. The first certificate in the pem file should be the the server certificate, followed by its chain starting from the root certificate down. Matt

1 0

[Dovecot] Copy from INBOX to INBOX corrupts file
by Andreas Jaekel 29 Oct '03

29 Oct '03

Aloha, again, Timo. Concerning the problem of copying within one folder... I'm still not sure it's not a problem I created myself, but I think it's caused by the use of the buffered IO routines in ostream-file.c ... you use io_add() to schedule write operations for later while simultanously reading new chunks from the very same file. This seems to confuse dovecot and make it write the copy to the strangest locations. I think when working within one folder I should use either blocking IO or a lot of calls to o_stream_flush()... what do you think? Greetings, Andy

2 3

[Dovecot] SSL Certificate Chains?
by Steve Gare 29 Oct '03

29 Oct '03

Hi, Just a quicky, I can't see anything about how Dovecot handles certificate chains. 1, Does it handle them? 2, How, if if does? (Just got a cert from www.instantssl.com and I need there intermediate cert do validate it.., Works fine with Apache and Exim ) Cheers, Steve Gare - Xantara.Net -- 0845 123 5707

2 1

[Dovecot] root privilidge issues?
by lists＠netxp.com.au 24 Oct '03

24 Oct '03

hi all, i am migrating from imap-uw to dovecot on a freebsd machine and i am getting this error message: imap(phil): Oct 22 19:24:01 Fatal: We couldn't drop root group privileges imap-login: Oct 22 19:24:01 Info: Login: phil [220.240.217.2] dovecot: Oct 22 19:24:01 Error: child 1019 (imap) returned error 89 when it trie manually logining in this is what i see telnet localhost 143 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. * OK dovecot ready. 1 login phil password 1 OK Logged in. Connection closed by foreign host. any ideas? thansk in advance phil

2 1

[Dovecot] Shared Folders
by D Canfield 21 Oct '03

21 Oct '03

I was just wondering if any thought has been given to shared folders and how they might be implemented yet? True shared folders are compelling enough to our organization that I've been looking into moving to Cyrus from Courier, but transparent folder migration looks to be nearly impossible for more than a handful of accounts... so that leaves me back at hoping courier or dovecot can eventually do what we want. We have a lot of organizations on campus that have a generic mail address such as "maintenance" of "physics." We don't like the idea of shared accounts (invariably either the mailbox gets forgotten about when a secretary changes, the password gets lost, or everyone on campus ends up with the password), so our policies require such generic accounts to be forwarded to a folder in a "real person's" account. This works well for us, but the users are clamoring for the ability to have more than one person access those folders. The problem with Courier's shared folders is that one person owns the folder and while others can see that mailbox, only the owner has write access. This kills most of the utility of shared folders, except for use as announcement areas. As I understand it, the reason Courier does it this way is that it uses the underlying unix permissions to control access to the mailbox. I'm thinking that even if dovecot uses the same concept for shared folders, it should be possible to allow multiple users to write to the folders if filesystem ACL support is used. At least ext3 and XFS both support ACL's, so there is a reasonable amount of support for them, and they would allow dovecot to give multiple users write access to the folders without having to manage it's own security system. The only issue would be that the delivery agent would have to make sure that that ACLs for a folder are kept consistent... I'm not even sure dovecot has to be ACL-aware. Anyway, if there is a completely different plan in the works, that's great too, but I wanted to at least voice an opinion and offer a suggestion of how to make it work. Thanks DC

4 5

[Dovecot] A view quirks
by Andreas Jaekel 21 Oct '03

21 Oct '03

Aloha Timo. I found a few odd things while playing with dovecot. At least one of them bugs me, but it might be a result of my own patches - I haven't testet it with mbox or maildir, yet. The details :) 1. * OK dovecot ready. A = (<return> Connection closed by foreign host. 2. * OK dovecot ready. A001 LOGIN xxxxxx xxxxxx A001 OK Logged in. A002 COPY 1 INBOX Connection closed by foreign host. (segfault) 3. [...] A003 SEARCH SUBJECT "test" * SEARCH 1 A003 OK Search completed. A004 SEARCH SUBJECT {6} + OK "test" * SEARCH A004 OK Search completed. I think these two searches should give identical results. 4. (my current nemesis) A002 CREATE bla A003 SELECT INBOX A004 COPY 1 bla A005 SELECT bla A006 COPY 1 bla * NO Internal error [2003-10-16 16:07:21] A006 OK Copy completed. If I copy from the current folder to the current folder I get a scrambled mailbox and an error. This might be selfmade, though. I'm currently trying to find this bug, but the various layers of [_oi]streams and callback and function pointer structs do not help :) I think the stream routines hide the mstream(istream(realstream(iostream))) and the ostream(_ostream(iostream)) behind the same file descriptor and the data gets written to the wrong position. BTW: what exactly is "max_pos"? Greetings and thanks! Andy

2 6

[Dovecot] move from imap-uw
by Sergey Matveychuk 18 Oct '03

18 Oct '03

Hello! I've just subscribed and I'd like to ask. Is there any troubles if I change my imap-uw with davecot? Will davecot pick up imap-uw created folders (mboxes)? I use all protocols: imap, imaps, pop and pops. I've seen some problems with mbox was mention on this maillist. What kind of problem? -- Sem.

2 1

[Dovecot] Re: Darwin Build (cvs)
by Jesse Peterson 16 Oct '03

16 Oct '03

On Thursday, October 9, 2003, at 02:00 AM, dovecot-request(a)procontrol.fi wrote: > From: Timo Sirainen <tss(a)iki.fi> > Subject: Re: [Dovecot] Darwin Build (cvs) > To: dovecot(a)procontrol.fi > Message-ID: <1065657988.1377.43.camel@hurina> > Content-Type: text/plain > > On Wed, 2003-10-08 at 17:23, Jesse Peterson wrote: >> Here's the deal: OpenDarwin 6.6 (closest relative being Mac OS X >> 10.2.6) won't build from CVS. Specifically libtoolize doesn't seem to >> exist (does it even come with libtool?) and the other parts of the >> "autogen.sh" fail miserably. > > So, the solution is to get libtoolize :) Mine seems to have come from > libtool14 package in Fink (although it is libtool v1.5, weird). Thought so. I wasn't sure if it came with libtool or not (OpenDarwin does ship with libtool). > > Also it looks like you're missing iconv.m4. It's in gettext package in > Fink. Ahh, I'll grab that as well. > > Did OpenDarwin use Fink? > Yes, although they've switched from 'DarwinPorts' to MetaPkg. Not sure the status on either of those, though. > Anyway, building from CVS requires some more packages to install which > aren't needed if you stick with tarballs. I could also build a new test > tarball if you need :) > This would be great. Maybe an automatically posted daily snapshot? Thanks Timo.

2 1

[Dovecot] problem getting mail from mbox
by Gavin Estey 16 Oct '03

16 Oct '03

I'm trying to set up dovecot and I'm running into a few problems. I just want to use IMAP for one user, everyone else gets POP3 from the existing (working server). $MAIL is set to /var/spool/mail/gavin_imap In my /var/spool/mail directory I now have a problem: -rw------- 1 gavin_im mail 1392 Oct 14 22:28 gavin_imap -rw------- 1 gavin_im popusers 0 Oct 14 21:54 gavin_imap When I try to get my mail, I guess it's checking the wrong file as I can't download anything. I also can't delete the file (which I find confusing as I am root): [root@chicken mail]# rm gavin_imap rm: cannot remove `gavin_imap': No such file or directory I actually have the same problem with the dovecot log file. I have this in my config: info_log_path = /tmp/dovecot -rw------- 1 root root 4436 Oct 14 22:40 dovecot [root@chicken tmp]# cat dovecot cat: dovecot: No such file or directory I don't get any obvious errors if I telnet in: [root@chicken tmp]# telnet localhost 143 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. * OK dovecot ready. 1 LOGIN gavin_imap <pass> 1 OK Logged in. 1 SELECT Inbox * FLAGS (\Answered \Flagged \Deleted \Seen \Draft) * OK [PERMANENTFLAGS (\Answered \Flagged \Deleted \Seen \Draft \*)] Flags permitted. * 0 EXISTS * 0 RECENT * OK [UIDVALIDITY 1066184985] UIDs valid * OK [UIDNEXT 1] Predicted next UID 1 OK [READ-WRITE] Select completed. 1 LOGOUT * BYE Logging out 1 OK Logout completed. Connection closed by foreign host. This is confusing the hell out of me. Is there anything obvious that I've screwed up? Thanks! Gavin.

2 3

[Dovecot] exchange, dovecot, mutt, evolution
by Krammer Clemens 16 Oct '03

16 Oct '03

Dear all, When I "migrate" my outlook folders to imapd (dovecot) it works well from a Windoze perspective, but if I open a migrated folder via imap (mutt, evolution) a cannot see the Senders. If i open the folder locally via mutt, the Senders are displayed correctly. By examining one folder with vi I recognised that there are no valid From headers but Return-Paths. Is there a common way to see the senders when connecting with imap? Best Regards, Clemens

2 1