- dovecot - dovecot.org

[dovecot] Some experiences
by Amelia A.Lewis 06 Jan '03

06 Jan '03

Heyas ... I'm trying to use dovecot to set up a 'closed server' (actually, mixed) sort of environment, for a box to be deployed co-located. I wanted to talk about some of that, and see if some of my experience qualifies as 'bug'. Note that I'm notoriously verbose, so if that bothers you, read no further .... Since it's a more-or-less closed server, I didn't want to use pam (well, I wouldn't mind, for users with shell accounts, but I can't see a way to mix pam and passwd-file ... is there a way to do that?). It turns out that digest-md5 support for clients is, well, not very widespread. I expect this box to be used by various windoze clients, eventually, and while I'm willing to lay down the law a little, I can't find enough digest-md5 support to lay down that law. So, regretfully, I abandoned that plan. This kinda bothers me, since I think TLS/STARTTLS is absolutely the *wrong* solution for mailbox access (or for anything that deals with a store-and-forward transport). I therefore tried md5 passwords in a passwd-file, with STARTTLS enabled. Didn't work. I turned off SSL, and reenabled plain-text, and watched the login go by. Very sniffable, of course. But correct username and password ... failed. There are instructions for creating digest-md5 style secrets in auth.txt, but none for md5 passwords; I used openssl passwd -1 [password] (and cut and paste). It seems odd to me that this didn't work; does that command use a different algorithm than dovecot? Falling back ... since it *shouldn't* matter if the passwords are weak, as long as they're inside TLS (along with the entire rest of the transaction, ugh), I used mkpasswd to generate a standard unix passwd (and cut and paste). This worked. It worked with TLS, and it worked without (as long as I remembered to allow sniffable logins). One more oddity: I originally had a realm in the passwd-file, which also established the mail location and 'home' directory (most of the users can't actually log in and don't have accounts in /etc/passwd, so it's not a real home directory). There was an odd message in syslog, approximately like so (where 'user' represents the username, and 'domain.realm' represents both my domain name and the realm in the passwd-file): imap(user:@omain.realm): Failed to create storage with data: domain.realm Note the missing first letter after user:@ And I don't know what it was trying to do; the maildirs already exist. /etc/dovecot.conf sets the default mail location to /home/users/%n/Maildir (non-shell users are all inside /home/users; shell users have mail there, with symlinks to ~/Maildir); the passwd-file establishes the location of the INBOX pretty definitively, I would think. The solution was to remove the realm (which isn't useful except for digest md5, anway, right?). Hey, it's working! Wahoo! *laugh* Anyway, if someone can offer suggestions on what's up with the pieces that *aren't* working, I'd be interested to know. It might give me a greater range of options. Amy! -- Amelia A. Lewis amyzing {at} talsever.com The less I seek my source for some definitive, the closer I am to fine. -- Indigo Girls

6 12

[dovecot] Util for pop3 to dovecot?
by Brad Figg 04 Jan '03

04 Jan '03

What I want to do is pop my email off of a Pop3 server and give it to dovecot for imap serving. Does anyone know of a utility that will allow me to do this? I looked at fetchmail but it seems to want to talk to an SMTP server. Thanks, Brad -- Brad Figg - brad(a)tpi.com 503.601.0235 ext. 105 TriplePoint, Inc. - http://www.tpi.com

3 2

[dovecot] maildirmake, shared folders
by Amelia A.Lewis 03 Jan '03

03 Jan '03

Does dovecot support shared folders, maildir style? I've set it up, using the maildirmake from debian maildrop to create folders. I have a number of as-yet unresolved issues, so I haven't been able to test with many clients. It's interesting that dovecot manages to present folders as siblings (rather than children) of INBOX. But mutt, at least, doesn't seem to see the shared maildir (created per instructions in the maildirmake man page). I find in the debian archives that courier is unable to see messages delivered to new in a shared maildir. This doesn't seem to be the same set of symptoms, though. It appears, instead, that dovecot isn't recognizing (and then advertising) the shared-maildirs file (a text file that describes the name and the location). Is that the case? I don't think it's a high priority, mind, just wondering. Another question: can the dovecot distribution include a maildirmake utility? The documentation is pretty blunt about the significant advantages of maildirs, but one has to go to the courier distribution to find a utility to pre-create these (which dovecot seems to want, since it's recognizing the mail type by looking for the dir). This is most significant when setting up new users, of course. Possibly this is inappropriate, and I should instead ask the debian package maintainer to 'recommend' or 'suggest' the maildrop package in order to get the required utility. Nice stuff, though. I really like the ease of administration; dovecot looks to me like a sort of equivalent to postfix, for IMAP instead of SMTP. A request for the website: perhaps developers could supply information on clients used with the server, and what oddities are encountered? I know that IMAP client support is spotty, at best (for instance, I can't find a client that can do digest-md5; mutt is supposedly capable, but my installation doesn't even admit that that's a valid authenticator). This is partly significant because setting up a system that uses options for which there is little or no client support is, perhaps, an exercise in futility. Amy! -- Amelia A. Lewis amyzing {at} talsever.com Never imagine yourself not to be otherwise than what it might appear to others that what you were or might have been was not otherwise than what you had been would have appeared to them to be otherwise. -- The Duchess [Lewis Carroll]

3 2

[dovecot] 0.99.5 released
by Timo Sirainen 02 Jan '03

02 Jan '03

v0.99.5 2003-01-02 Timo Sirainen <tss(a)iki.fi> * This release includes a lot of code cleanups, especially related to security. Direct buffer modifying was replaced in several places with a new buffer API, which provides a safe way to do it. Code that looks unsafe contains @UNSAFE tag to make auditing easier. + Support for SORT extension. Originally I thought about not implementing any extensions before 1.0, but too many people want webmails which want SORT. THREAD is another extension which they want, but we don't support it yet. + imap_listen and imaps_listen settings now accept "*" and "::" to specify if we want to listen in IPv4 or IPv6 addresses. "::" may also listen in all IPv4 addresses depending on the OS (Linux does, BSD doesn't) + New setting: default_mail_env can be used to specify where to find users mailbox. Default is still to use autodetection. + New setting: imap_log_path to log logins etc. informational messages to different file. + We support now separate mbox file for INBOX folder, no need for symlink kludging anymore. + Support for keeping index files in different location than actual mailboxes. ? Disabled mailbox_check_interval setting by default, it breaks Evolution. - SEARCH was still somewhat buggy, especially with laggy networks. Also body/header searches might have crashed or at least used memory too much - Deleting messages in the middle of mbox caused dovecot to reindex the following messages as new messages (ie. change UIDs and set \Recent flag). - Digest-MD5 auth: Initial server challenge packet was missing a comma, which might have made it incompatible with some implementations. - Some more fixes to unnecessarily high memory usage - SELECT and EXAMINE often printed UNSEEN reply or maybe complained about corrupted indexes. Happened usually only with mbox. - FETCH BODYSTRUCTURE gave incorrect reply, breaking pine - LIST was pretty buggy with mbox subfolders - CHECK command returned just "missing parameters" error - DELETE didn't work with mbox folders - CREATE mailbox<hierarchy separator> failed always. - CREATE and RENAME didn't create required hierarchies with mbox - RFC822 date parser didn't handle single digit days correctly. - login_process_per_connection = yes didn't work with imaps port connections which is exactly where it was mostly wanted. - ssl_disabled = yes didn't disable listening in imaps port - process limiting code didn't compile everywhere (eg. FreeBSD). - Linux sendfile() was never detected - We didn't work at all with gcc/PowerPC

1 0

[dovecot] Re: Maildir format
by Timo Sirainen 30 Dec '02

30 Dec '02

On Mon, 2002-12-30 at 09:25, Gareth J. Greenaway wrote: > On Mon, Dec 30, 2002 at 08:51:21AM +0200, Timo Sirainen wrote: > > Dovecot wants the folders directory names to begin with a dot, so name > > it ".test". Subfolders would be named as ".test.subfolder". This is > > compatible with Courier's Maildir++ format. > > Right. that is what I have as im using the Maildir folders from > Courier but I still do not see mail showing up in them. The folder > itself shows up but no mail. Oh, I think then you're just using mutt a bit wrong :) Enter-key goes inside a subfolder, space-key selects the folder. A folder can itself contain mail but it can also contain subfolders, so this is needed. With Courier you probably got away with just enter-key because you didn't have such subfolders and Courier gives mutt hints about if folders have subfolders. Dovecot doesn't yet support that extension so mutt always assumes they have subfolders.

1 0

[dovecot] Maildir format
by gareth＠wiked.org 30 Dec '02

30 Dec '02

Started using dovecot and I am very impressed by its speed. I am confused by what it is expecting the format of the Maildir directory to be in. I currently have ~/Maildir which contains new, cur, and tmp all containing email that should be in my inbox and it pops up correctly as such when I connect to dovecot using Mutt. After the connection a new .INBOX is created under ~/Maildir with sym links back to new, cur, and tmp. The problem arrises when I want to view email from another folder, I have a folder called "test" which contains some email also in Maildir format, however when I attempt to view this email I only get an empty listing with the ".." to go one directory up. Thanks. Gareth Greenaway -- "My friends say I should act my age, What's my age again?" Gareth J. Greenaway | gareth(a)wiked.org

2 1

[dovecot] passwd-file; configure args..
by Jesse Peterson 28 Dec '02

28 Dec '02

I couldn't reproduce this again (because i can't get imap-master running due to problem below..) but.. When compiling without shadow, passwd, or pam, when logging in I get after a successful login (in my syslog): Dec 28 22:39:38 badcase imap(username): MAIL environment missing and autodetection failed (home /some/home/directory) Now I thought this was covered in a previous fix before (ChangeLog: "2002-12-17 03:00 Timo Sirainen <tss(a)iki.fi>") but for some reason it's broken again..? The configure args problem has to do with --localstatedir. If i leave it default (PREFIX/var) it forces me (even with non-relative pathnames in the config file) for the socket-files (ssl_parameters_file, login_dir) that should go in /var/run to be put under PREFIX, (/usr/local for NetBSD 1.6). Not too bad, if I could override it in the config file, which I can't. So I ./configure'd with --localstatedir=/, and when I run imap-master I now get: Fatal: Can't create directory //run/dovecot: No such file or directory Somewhat odd. Any clues on either of these would be helpful, thanks. -- Jesse Peterson erage(a)softhome.net

2 7

[dovecot] o.99rc2 works on PowerPC - sort of...
by Jonas Smedegaard 27 Dec '02

27 Dec '02

Hi all, The new release starts on PowerPC now. On to the next problem... :-) When accessing my folders through Pine (setup to always go through IMAP, even local mboxes) I get "kicked out" it seems, pretty often. Here's what shows in the logfile: imap(jonas): file imap-bodystructure.c: line 342 (part_write_body): assertion failed: (part->children->next == NULL) - Jonas -- Debian GNU/Linux developer http://www.debian.org/ Jonas Smedegaard +45 40843136 http://dr.jones.dk/ Consulting for machines and humans: Spiff ApS

2 1

[dovecot] RLIMIT_AS patch for FreeBSD
by Dominic Marks 27 Dec '02

27 Dec '02

Hey, I've just updated the FreeBSD dovecot port to the 0.99.4 beta release, the following patch was needed as FreeBSD has setrlimit(2) but does not define RLIMIT_AS (I believe). I switched the code the require both, alternatively find a parallel option in the FreeBSD setrlimit(2) manual page. ( http://www.freebsd.org/cgi/man.cgi?setrlimit ) You may be interested in integrating this patch. --- src/lib/restrict-process-size.c Fri Dec 27 18:19:41 2002 +++ src/lib/restrict-process-size.c-new Fri Dec 27 18:19:22 2002 @@ -32,7 +32,8 @@ void restrict_process_size(unsigned int size __attr_unused__) { -#ifdef HAVE_SETRLIMIT +#if defined(HAVE_SETRLIMIT) && defined(RLIMIT_AS) +/* FreeBSD has setrlimit(2) but does not define RLIMIT_AS */ struct rlimit rlim; rlim.rlim_max = rlim.rlim_cur = Thanks, -- Dominic <dom at cus.org.uk> <d.marks at student.umist.ac.uk>

2 1

[dovecot] Dovecot splited fetch problem
by Hon 27 Dec '02

27 Dec '02

Hello, When I was trying out the IMAP server (using Mozilla as client), I have no success of receving mail with attachments. Looking it down further I've found that Mozilla fetches the message in 10k blocks, and the server seems missed a couple of bytes at where the split occurs... Mozilla start the fetch by saying: * 413 FETCH (UID 827 RFC822.SIZE 106371 BODY[]<0> {10240} -- snip to place of split -- zmmAWdUxYA6Aobczv4BguxSZGqdYs6ELNIpgjiPUdYLxsHy) 7 OK Fetch completed. 8 UID fetch 827 (UID RFC822.SIZE BODY[]<10240.10240>) * 413 FETCH (UID 827 RFC822.SIZE 106371 BODY[]<10240> {10240} 6BilWYuUJGA4YeNYFJbI1gL The resulting line (saved from Mozilla view source) is: zmmAWdUxYA6Aobczv4BguxSZGqdYs6ELNIpgjiPUdYLxsHy6BilWYuUJGA4YeNYFJbI1gL ^^ Two bytes missing Where the original is: zmmAWdUxYA6Aobczv4BguxSZGqdYs6ELNIpgjiPUdYLxsHygZ6BilWYuUJGA4YeNYFJbI1gL ^^ I have tried this using Debian packaged (0.99.4-1) and self compiled 0.99.5rc1. Both exhibit this problem.

2 1