- dovecot - dovecot.org

[Dovecot] gzipped files in Mail directory
by Olivier Tharan 09 Oct '03

09 Oct '03

Hi, I have been testing dovecot on my own machine before a bigger deployment. Basically I have all my mbox mailboxes in ~/Mail, delivered by procmail. Some mbox's are gzipped because they are archives which I can grep into, but do not want to look at often. When I first launched an IMAP client, dovecot created the indexes (I think) and gave several errors when finding a gzipped file, such as: ,---- | Oct 1 19:53:21 mafate imap(olive): File isn't in mbox format: /home/olive/Mail//Archives/daily.d/mafate.2003.gz `---- After 19 gzipped mbox's, I had the following crash: ,---- | Oct 1 19:53:21 mafate dovecot: child 4923 (imap) killed with signal 10 `---- Does dovecot not like gzipped mbox's? For the moment, all gzipped files are in ~/Mail/.Archives and I can live with it. -- olive

1 0

[Dovecot] Line length limitation in dovecot leads to out-of-control process?
by Amelia A. Lewis 09 Oct '03

09 Oct '03

Heylas! I've been copying mailboxes from one dovecot server (on an alpha running debian testing, 0.99.10-9) to another (on an i686 running debian unstable, 0.99.10-9). I encountered a weird problem, which I seem to have resolved by moving a message with a To: line from hell out of the mailbox. Both machines are running maildir. They're not exactly identical, but they're fairly similar, apart from differences in distribution and architecture; they're running the same version of dovecot, supposedly. Specifically, when I copied this folder over to the new location, dovecot went into CPU-feeding frenzy (up to 99.7% of CPU), and stayed there. If I killed the client and tried to look at the same mailbox using a different client (mutt instead of sylpheed), without killing the imap processor on the server, the indexes would be locked, mutt would fall back to reading headers instead, and then would send its imap process into CPU-feeding frenzy when it got to one particular message. After some experimentation, I found the message. Once I moved it out of cur, I could access the mailbox without problems. The odd thing is that I can look at the message on the original machine, but on the machine it was copied to, its mere existence makes the mailbox inaccessible. After a little more exploration (grep ^To name-of-message | wc, on both machines) I seem to have found the issue: the original message has a To: line that is 16504 characters long. In the copy, it's 8192 characters long, and ends in the middle of a <email@address> (line in copy ends: user name <emai, chopped right there). I suspect that it is the *latter* that's causing dovecot to die a messy death. Or, actually, to turn into a very *rude* process that *doesn't* die. It's a really stupid trick to stuff that much into a header line (it's from helixcode in 2000, btw, notifying people that the promised CD is late), and perfectly reasonable to truncate (even to truncate to 1000 characters, per 282[12]). However, I kinda thing dovecot shouldn't choke on the results of truncation, which is what I think is the problem. Amy! -- Amelia A. Lewis amyzing {at} talsever.com I have spent nights with matches and knives, leaning over ledges, only two flights up. Cutting my heart, burning my soul. Nothing left to hold. Nothing left, but blood and fire. -- Indigo Girls

2 2

[Dovecot] Darwin Build (cvs)
by Jesse Peterson 09 Oct '03

09 Oct '03

Oops, I apologize for my short, breif message that didn't even include what I said it should inlcude =). 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. This time (really) attached is a gzip of the log of autogen.sh: "./autogen.sh >& autogen.log". Thanks for any help! - Jesse

2 1

[Dovecot]
by Jesse Peterson 08 Oct '03

08 Oct '03

% uname -a Darwin xxx 6.6 Darwin Kernel Version 6.6: Sun Jul 6 05:12:51 PDT 2003; root:tmp/xnu-6.6.2-4-root.obj/RELEASE_PPC Power Macintosh powerpc Attached is a gzipped "sh autogen.sh >& autogen.log" for the above machine. What do you think? Jesse Peterson <jesse(a)mactechs.com> Office 206.441.7443 / Fax 206.441.8078 MacTechs / Macintosh System and Network Support 2121 First Avenue, Suite 103 / Seattle, Washington 98121 http://www.mactechs.com/

1 0

[Dovecot] those assertions
by dovecot＠spam.turbolink.net 06 Oct '03

06 Oct '03

Timo, Just FYI while you're working on the mbox code, I discovered a bug that occasionally causes dovecot to insert a single blank line at the beginning of the mbox, I'm not sure when/how/why. A single blank line seems to really confuse dovecot, and causes the server to crash when somebody connects to an afflicted mbox. ("Error indexing mbox" in the log, and disconects the user right away). I fixed it with lots of ugly code at our site (basically just remember that we skipped an intial blank line when doing other operations, instead of skipping it then forgetting later). I suspect there's a more elegant way to do it. - Mike

1 0

[Dovecot] SSL Client Certificate Support
by Bert Koelewijn 06 Oct '03

06 Oct '03

Dear Timo, most modern enterprises make use of a Public Key Infrastructure. It would be nice to have dovecot check a client certificate instead of a password. This makes life much easier and more secure. Mail clients like Mozilla and MS Outlook do support this. What do you think of the following feature request: - Client authenticates with a certificate via SSL. (Like stunnel can) - Dovecot looks the username up in a table with (public key, username) - The mailclient gives a name and password, but dovecot ignores them - Dovecot gives the client access by the username found in the table This way existing mail clients can use this system and you can save your username with an empty password. Best Regards, Bert Koelewijn

5 10

[Dovecot] those assertions
by Bernard Blackham 05 Oct '03

05 Oct '03

Hi, I did a bit of digging into the assertions from mbox-rewrite.c (hdr_parsed_size.physical_size == hdr_size). Every time it happens, hdr_size is 0, while hdr_parsed_size.physical_size is non-zero (presumably the correct value). Does this help track it down? I can't manage to get hold of an mbox that causes this though. Another infrequent assertion that's popped up: file message-address.c: line 262 (message_address_write): assertion failed: (addr->domain != NULL) And one that's only ever been spotted once in the last few weeks: file imap-bodystructure.c: line 278 (part_parse_headers): assertion failed: (part->physical_pos >= input->v_offset - start_offset) I'm waiting for them to happen again to provide some more info... Regards, Bernard. -- Bernard Blackham bernard at blackham dot com dot au

2 1

[Dovecot] another Dovecot problem
by Tim Miller 05 Oct '03

05 Oct '03

Hi All-- First, thanks to Timo for solving my last issue. I got dovecot up and running on one of my machines, and it does exactly what I want. Now I'm trying to install dovecot on an older box (glibc 2.0.7 Linux kernel 2.0.36), and encountering some problems: -- firstly autoconf/automake does not seem to correctly deal with the madvise function on this system. It detects that we don't have madvise, yet it's prototyped in /usr/include/sys/mman.h. This causes the compilation to fail when dovecot tries to redefine this function. -- I attempted to solve this problem by renaming the function dovecot prototypes to dc_madvise and then changed all the calls in the code (and the implementation) from madvise to dc_madvise. This allows the code to compile correctly. -- However, once the server is installed, it does not appear to work correctly. I can login, but when I try to list my folders or select any folder (one that exists or one that doesn't exist) the connection times out and the following entry is written to the log file: dovecot: Oct 03 05:38:56 Error: login: received another "not listening" notification (if you can't login at all, see src/lib/fdpass.c) I tracked this message down in the source, but didn't see any obvious way to trace back the problem. Does anyone have experience building on this older configuration (the glibc is a Red Hat version, if it makes any difference)? Any ideas on what might be causing this issue? Best, Tim Miller

2 5

mbox/maildir/directories [was: Re: [Dovecot] still getting weird .{inbox} folder creations with latest -test7]
by Miquel van Smoorenburg 05 Oct '03

05 Oct '03

This is a bit of a late reply, but still: On Thu, 2003-09-04 at 15:06, Timo Sirainen wrote: > CVS has been optimized to assume all files beginning with a dot (except > .subscriptions currently) are mailboxes. This is faster since it doesn't > have to stat() each file to figure out if it's a directory or not. > Perhaps that should be optional. New Linux systems (maybe others too) > wouldn't actually even need the stat().. > > I'll probably some day also add support for dotless directory-structured > mailbox locations, but it's not really high priority. Well, on our systems the mailbox is in /var/mail in maildir format, and in ~/Mail users have split up their folders in subdirs, in mixed maildir and mbox formats. So if dovecot could support both dotless directory-structured mailbox locations and maildir/mbox mix-n-match in the same tree, that would be excellent for us. Mbox is a better format for archiving. Now if Dovecot could also access (read-only ofcourse) compressed mbox archives ... Food for thought. Just put it on the wishlist ;) Mike.

2 1

[Dovecot] Patches
by Jaldhar H. Vyas 05 Oct '03

05 Oct '03

As the freeze for Debian sarge slowly approaches I want to make sure the Dovecot packages are in as good condition as possible. I see there have been a number of patches since 0.99.10. I have added the following patches: * segfault when user home directory is empty * Proper PAM service name * Make suid work on 2.6 kernels Any other patches thatI ought to add. Or better yet, will there be a 0.99.11 soon? -- Jaldhar H. Vyas <jaldhar(a)debian.org> La Salle Debain - http://www.braincells.com/debian/

5 4

[Dovecot] Why does dovecot confuse mutt?
by Amelia A Lewis 04 Oct '03

04 Oct '03

Heylas, I just switched my primary mail server from courier imap to dovecot (I've been waiting to do this; dovecot showed up in debian testing for alpha, at last). So, using the same set of maildirs, there are slightly different behaviors. First, of course, dovecot displays the hierarchy as I thought it ought to be, such that INBOX is a sibling of all other top-level boxes, not the parent of everything. This is A Good Thing[tm]. That works as expected. However, as far as mutt is concerned, *every* mailbox has child folders. My muttrc contains the line: set folder=imap://localhost/. (and the home folder is set to imap://localhost/INBOX; the idea is to do all mail munging through the server, never directly touching the maildirs with an MUA). Oh, and I'm using maildir (~/Maildir). In mutt, from the currently-selected mailbox, one can use 'c' to change boxes, and ? displays a navigable folder hierarchy. If a folder contains only mail (no child folders), it shows up as "FolderName", if it has child folders (whether or not it has mail), it shows up as "FolderName.". You can display the mail in a folder that has child folders, but it requires a different keystroke than the default (enter selects a folder, which means showing child folders if there are any, showing mail if there aren't; to display mail in a folder that has child folders, space). This is tripping me up a little, because of course the commands are mostly bred into my fingertips, and it's a difficult adjustment. I hit return, and I see a folder that contains the "child" folder "../". Is this a glitch in mutt's handling of information that I ought to report to mutt maintainers? Is it a glitch in dovecot's presentation of information that I ought to report to ... err, well. *laugh* My other primary mail client seems to cope beautifully with the information that's supplied (sylpheed, if it matters). But then, sylpheed also coped perfectly when courier was supplying information. I gather that courier and dovecot are reporting something slightly different, somehow, but I don't really know what, and I don't know which should be considered "correct". Amy!

5 5

[Dovecot] mailbox sorting issue
by Paul 04 Oct '03

04 Oct '03

Thanks for the "junk at end" help. Now I'm having one more perplexing issue: When I sort a Dovecot mbox in Pine, the most recent messages tend to linger near the bottom of the message list, regardless of how they're sorted. Here's a snippet of the bottom of the list, sorted by arrival time: 448 2003-09-24 20:31 sprightly@e... Re: *hoopoe 449 2003-09-24 21:05 mlescaut@w... Re: yes! 450 2003-09-24 23:57 jackulysses@y... Re: oh my lord 451 2003-09-25 01:43 sprightly@e... attachments==take a look 452 2003-09-25 01:29 avantguild@m... Your Membership is Almost Up 453 2003-09-25 10:40 heidi@p... xhtml 454 2003-09-25 11:04 dembo@s. Re: bridge 455 2003-09-25 11:54 JKesner@m... Bombay 456 2003-09-25 12:05 cm2144@c... Re: bridge at linda's And here is a snippet of the same mbox list, sorted by From address: 449 2003-09-25 01:43 sprightly@e... attachments==take a look 450 2003-09-10 16:06 staff@c... Tribute in Light Shines 451 2003-09-08 18:44 starpiekat@y... Re: Sorry! 452 2003-09-23 22:01 starry@p... but... 453 2003-09-25 12:05 cm2144@c... Re: bridge at linda's 454 2003-09-17 17:07 wahrman@s... RE: bridge etc. 455 2003-09-19 13:11 wahrman@s... RE: bridge etc. 456 2003-09-15 23:39 waycross@u... waycross - the eagle Notice that message from cm2144 in both lists, way out of order in the second! I'm not 100% sure that this is a Dovecot issue, but I suspect it is, since this doesn't happen if I open the same mailbox locally or with UW-IMAP. I know that Pine makes use of IMAP server-based sorting if it is available. I tried deleting .imap-index* files, just in case, but that didn't help. thanks, Paul

3 2

Re: [Dovecot] PATCH: mysql authentication
by Matthew Reimer 03 Oct '03

03 Oct '03

> On Fri, 2003-10-03 at 19:54, Matthew Reimer wrote: > >Here's another patch. This one isn't much different from the last patch, >except it tries a little harder to make sure that the required fields >are returned by the query (like the pgsql module does). BTW, the patch lacks configure magic to detect mysql includes and libs, and to define USERDB_MYSQL and PASSDB_MYSQL in config.h. Matt

1 0

Re: [Dovecot] PATCH: mysql authentication
by Matthew Reimer 03 Oct '03

03 Oct '03

Timo Sirainen wrote: > On Fri, 2003-10-03 at 19:54, Matthew Reimer wrote: >>One question about how t_strdup works. How and who should memory >>returned by t_strdup be freed? My patch has a function my_get_str() >>which returns the value of t_strdup, and then I do this: >> >> user.uid = atoi(my_get_str(res, row, "uid")); /* XXX leak */ >> user.gid = atoi(my_get_str(res, row, "gid")); /* XXX leak */ >> >>Will this result in a memory leak? > > > No. > > Memory allocated by t_*() functions are freed "later". There's first > t_push(), then you allocate memory, then t_pop() frees it all. So they > never leak memory, but they could allocate more memory than necessary if > called inside loop. That can be fixed by adding t_push/t_pop calls > there. Nice! Thanks for the answer Timo. Matt

1 0

Re: [Dovecot] PATCH: mysql authentication
by Matthew Reimer 03 Oct '03

03 Oct '03

Here's another patch. This one isn't much different from the last patch, except it tries a little harder to make sure that the required fields are returned by the query (like the pgsql module does). One question about how t_strdup works. How and who should memory returned by t_strdup be freed? My patch has a function my_get_str() which returns the value of t_strdup, and then I do this: user.uid = atoi(my_get_str(res, row, "uid")); /* XXX leak */ user.gid = atoi(my_get_str(res, row, "gid")); /* XXX leak */ Will this result in a memory leak? Matt diff -u -r --new-file work/dovecot-0.99.10/doc/dovecot-mysql.conf work.patched/dovecot-0.99.10/doc/dovecot-mysql.conf --- work/dovecot-0.99.10/doc/dovecot-mysql.conf Wed Dec 31 16:00:00 1969 +++ work.patched/dovecot-0.99.10/doc/dovecot-mysql.conf Tue Sep 30 12:24:00 2003 @@ -0,0 +1,70 @@ +# For the mysql passdb module, you'll need a database with a table that +# contains fields for at least the userid and password. If you want to +# use the user@domain syntax, you might want to have a separate domain +# field as well. +# +# If your users all have the same uig/gid, and have predictable home +# directories, you can use the static userdb module to generate the home +# dir based on the userid and domain. In this case, you won't need fields +# for home, uid, or gid in the database. +# +# If you prefer to use the mysql userdb module, you'll want to add fields +# for home, uid, and gid. Here is an example table: +# +# CREATE TABLE users ( +# userid VARCHAR(128) NOT NULL, +# password VARCHAR(64) NOT NULL, +# home VARCHAR(256) NOT NULL, +# uid INTEGER NOT NULL, +# gid INTEGER NOT NULL, +# active CHAR(1) DEFAULT 'Y' NOT NULL +# ); + +db_host = localhost +db_port = 3306 +#db_unix_socket = /var/tmp/mysql.sock +db = users +db_user = dovecot-db +db_passwd = opensesame +db_client_flags = 0 + +# Default password scheme. +# +# Currently supported schemes include PLAIN, PLAIN-MD5, DIGEST-MD5, and CRYPT. +# +#default_pass_scheme = PLAIN-MD5 + +# Query to retrieve the password. +# +# The query should return one row, one column. If more than one row or column +# is returned, authentication will automatically fail. +# +# Available substitutions: +# %u = entire userid +# %n = user part of user@domain +# %d = domain part of user@domain +# +# Example: +# password_query = SELECT password FROM users WHERE userid = '%n' AND domain = '%d' +# password_query = SELECT password FROM users WHERE userid = '%u' AND active = 'Y' +# +#password_query = SELECT password FROM users WHERE userid = '%u' + +# Query to retrieve the user information. +# +# The query must return only one row. The columns to return are: +# home - Home directory +# mail - MAIL environment +# system_user - System user name (for initgroups()) +# uid - System UID +# gid - System GID +# +# Either home or mail is required. uid and gid are required. If more than one +# row is returned or there's missing fields, login will automatically fail. +# +# Examples +# user_query = SELECT home, uid, gid FROM users WHERE userid = '%n' AND domain = '%d' +# user_query = SELECT dir AS home, user AS uid, group AS gid FROM users where userid = '%u' +# user_query = SELECT home, 501 AS uid, 501 AS gid FROM users WHERE userid = '%u' +# +#user_query = SELECT home, uid, gid FROM users WHERE userid = '%u' diff -u -r --new-file work/dovecot-0.99.10/src/auth/Makefile.in work.patched/dovecot-0.99.10/src/auth/Makefile.in --- work/dovecot-0.99.10/src/auth/Makefile.in Thu Jun 26 09:15:11 2003 +++ work.patched/dovecot-0.99.10/src/auth/Makefile.in Tue Sep 30 12:24:00 2003 @@ -99,10 +99,10 @@ dovecot_auth_LDADD = ../lib-settings/libsettings.a ../lib/liblib.a $(AUTH_LIBS) $(RAND_LIBS) $(MODULE_LIBS) -dovecot_auth_SOURCES = auth-module.c db-ldap.c db-pgsql.c db-passwd-file.c login-connection.c main.c master-connection.c md5crypt.c mech.c mech-anonymous.c mech-cyrus-sasl2.c mech-plain.c mech-digest-md5.c mycrypt.c passdb.c passdb-ldap.c passdb-passwd.c passdb-passwd-file.c passdb-pam.c passdb-shadow.c passdb-vpopmail.c passdb-pgsql.c password-scheme.c userdb.c userdb-ldap.c userdb-passwd.c userdb-passwd-file.c userdb-static.c userdb-vpopmail.c userdb-pgsql.c +dovecot_auth_SOURCES = auth-module.c db-ldap.c db-pgsql.c db-mysql.c db-passwd-file.c login-connection.c main.c master-connection.c md5crypt.c mech.c mech-anonymous.c mech-cyrus-sasl2.c mech-plain.c mech-digest-md5.c mycrypt.c passdb.c passdb-ldap.c passdb-passwd.c passdb-passwd-file.c passdb-pam.c passdb-shadow.c passdb-vpopmail.c passdb-pgsql.c passdb-mysql.c password-scheme.c userdb.c userdb-ldap.c userdb-passwd.c userdb-passwd-file.c userdb-static.c userdb-vpopmail.c userdb-pgsql.c userdb-mysql.c -noinst_HEADERS = auth-login-interface.h auth-master-interface.h auth-mech-desc.h auth-module.h db-ldap.h db-pgsql.h db-passwd-file.h common.h login-connection.h master-connection.h md5crypt.h mech.h mycrypt.h passdb.h password-scheme.h userdb.h userdb-vpopmail.h +noinst_HEADERS = auth-login-interface.h auth-master-interface.h auth-mech-desc.h auth-module.h db-ldap.h db-pgsql.h db-mysql.h db-passwd-file.h common.h login-connection.h master-connection.h md5crypt.h mech.h mycrypt.h passdb.h password-scheme.h userdb.h userdb-vpopmail.h mkinstalldirs = $(SHELL) $(top_srcdir)/mkinstalldirs CONFIG_HEADER = ../../config.h @@ -116,17 +116,17 @@ LDFLAGS = @LDFLAGS@ LIBS = @LIBS@ dovecot_auth_OBJECTS = auth-module.$(OBJEXT) db-ldap.$(OBJEXT) \ -db-pgsql.$(OBJEXT) db-passwd-file.$(OBJEXT) login-connection.$(OBJEXT) \ +db-pgsql.$(OBJEXT) db-mysql.$(OBJEXT) db-passwd-file.$(OBJEXT) login-connection.$(OBJEXT) \ main.$(OBJEXT) master-connection.$(OBJEXT) md5crypt.$(OBJEXT) \ mech.$(OBJEXT) mech-anonymous.$(OBJEXT) mech-cyrus-sasl2.$(OBJEXT) \ mech-plain.$(OBJEXT) mech-digest-md5.$(OBJEXT) mycrypt.$(OBJEXT) \ passdb.$(OBJEXT) passdb-ldap.$(OBJEXT) passdb-passwd.$(OBJEXT) \ passdb-passwd-file.$(OBJEXT) passdb-pam.$(OBJEXT) \ passdb-shadow.$(OBJEXT) passdb-vpopmail.$(OBJEXT) \ -passdb-pgsql.$(OBJEXT) password-scheme.$(OBJEXT) userdb.$(OBJEXT) \ +passdb-pgsql.$(OBJEXT) passdb-mysql.$(OBJEXT) password-scheme.$(OBJEXT) userdb.$(OBJEXT) \ userdb-ldap.$(OBJEXT) userdb-passwd.$(OBJEXT) \ userdb-passwd-file.$(OBJEXT) userdb-static.$(OBJEXT) \ -userdb-vpopmail.$(OBJEXT) userdb-pgsql.$(OBJEXT) +userdb-vpopmail.$(OBJEXT) userdb-pgsql.$(OBJEXT) userdb-mysql.$(OBJEXT) dovecot_auth_DEPENDENCIES = ../lib-settings/libsettings.a \ ../lib/liblib.a dovecot_auth_LDFLAGS = @@ -295,6 +295,13 @@ ../../src/lib/strfuncs.h ../../src/lib/network.h \ ../../src/lib/str.h ../../src/lib-settings/settings.h \ db-pgsql.h +db-mysql.o: db-mysql.c ../../config.h common.h ../../src/lib/lib.h \ + ../../src/lib/compat.h ../../src/lib/macros.h \ + ../../src/lib/failures.h ../../src/lib/data-stack.h \ + ../../src/lib/mempool.h ../../src/lib/imem.h \ + ../../src/lib/strfuncs.h ../../src/lib/network.h \ + ../../src/lib/str.h ../../src/lib-settings/settings.h \ + db-mysql.h login-connection.o: login-connection.c common.h ../../src/lib/lib.h \ ../../config.h ../../src/lib/compat.h ../../src/lib/macros.h \ ../../src/lib/failures.h ../../src/lib/data-stack.h \ @@ -388,6 +395,14 @@ ../../src/lib/str.h ../../src/lib/strescape.h \ ../../src/lib/var-expand.h password-scheme.h db-pgsql.h \ passdb.h mech.h auth-login-interface.h +passdb-mysql.o: passdb-mysql.c ../../config.h common.h \ + ../../src/lib/lib.h ../../src/lib/compat.h \ + ../../src/lib/macros.h ../../src/lib/failures.h \ + ../../src/lib/data-stack.h ../../src/lib/mempool.h \ + ../../src/lib/imem.h ../../src/lib/strfuncs.h \ + ../../src/lib/str.h ../../src/lib/strescape.h \ + ../../src/lib/var-expand.h password-scheme.h db-mysql.h \ + passdb.h mech.h auth-login-interface.h passdb-shadow.o: passdb-shadow.c ../../config.h common.h \ ../../src/lib/lib.h ../../src/lib/compat.h \ ../../src/lib/macros.h ../../src/lib/failures.h \ @@ -428,6 +443,13 @@ ../../src/lib/imem.h ../../src/lib/strfuncs.h \ ../../src/lib/str.h ../../src/lib/strescape.h \ ../../src/lib/var-expand.h db-pgsql.h userdb.h +userdb-mysql.o: userdb-mysql.c ../../config.h common.h \ + ../../src/lib/lib.h ../../src/lib/compat.h \ + ../../src/lib/macros.h ../../src/lib/failures.h \ + ../../src/lib/data-stack.h ../../src/lib/mempool.h \ + ../../src/lib/imem.h ../../src/lib/strfuncs.h \ + ../../src/lib/str.h ../../src/lib/strescape.h \ + ../../src/lib/var-expand.h db-mysql.h userdb.h userdb-static.o: userdb-static.c ../../config.h common.h \ ../../src/lib/lib.h ../../src/lib/compat.h \ ../../src/lib/macros.h ../../src/lib/failures.h \ diff -u -r --new-file work/dovecot-0.99.10/src/auth/db-mysql.c work.patched/dovecot-0.99.10/src/auth/db-mysql.c --- work/dovecot-0.99.10/src/auth/db-mysql.c Wed Dec 31 16:00:00 1969 +++ work.patched/dovecot-0.99.10/src/auth/db-mysql.c Tue Sep 30 12:24:00 2003 @@ -0,0 +1,181 @@ +/* Copyright (C) 2003 Alex Howansky, Timo Sirainen */ + +#include "config.h" +#undef HAVE_CONFIG_H + +#if defined(PASSDB_MYSQL) || defined(USERDB_MYSQL) + +#include "common.h" +#include "network.h" +#include "str.h" +#include "settings.h" +#include "db-mysql.h" + +#include <limits.h> +#include <stddef.h> +#include <stdlib.h> + +#define DEF(type, name) { type, #name, offsetof(struct mysql_settings, name) } + +static struct setting_def setting_defs[] = { + DEF(SET_STR, db_host), + DEF(SET_STR, db_port), + DEF(SET_STR, db_unix_socket), + DEF(SET_STR, db), + DEF(SET_STR, db_user), + DEF(SET_STR, db_passwd), + DEF(SET_STR, db_client_flags), + DEF(SET_STR, password_query), + DEF(SET_STR, user_query), + DEF(SET_STR, default_pass_scheme) +}; + +struct mysql_settings default_mysql_settings = { + MEMBER(db_host) "localhost", + MEMBER(db_port) "0", + MEMBER(db_unix_socket) "/var/tmp/mysql.sock", + MEMBER(db) "email_accounts", + MEMBER(db_user) "dovecot", + MEMBER(db_passwd) "changeme", + MEMBER(db_client_flags) "0", + MEMBER(password_query) "SELECT password FROM users WHERE userid = '%u'", + MEMBER(user_query) "SELECT home, uid, gid FROM users WHERE userid = '%u'", + MEMBER(default_pass_scheme) "PLAIN-MD5" +}; + +static struct mysql_connection *mysql_connections = NULL; + +static int mysql_conn_open(struct mysql_connection *conn); +static void mysql_conn_close(struct mysql_connection *conn); + +void db_mysql_query(struct mysql_connection *conn, const char *query, + struct mysql_request *request) +{ + MYSQL_RES *res; + int failed; + + if (!conn->connected) { + if (!mysql_conn_open(conn)) { + request->callback(conn, request, NULL); + return; + } + } + + if (verbose_debug) + i_info("MYSQL: Performing query: %s", query); + + if (mysql_query(conn->mysql, query)) + i_info("MYSQL: Error executing query \"%s\": %s", query, + mysql_error(conn->mysql)); + + if ((res = mysql_store_result(conn->mysql))) + failed = FALSE; + else { + i_info("MYSQL: Error retrieving results: %s", + mysql_error(conn->mysql)); + failed = TRUE; + } + + request->callback(conn, request, failed ? NULL : res); + mysql_free_result(res); + i_free(request); +} + +static int mysql_conn_open(struct mysql_connection *conn) +{ + if (conn->connected) + return TRUE; + + if (conn->mysql == NULL) { + conn->mysql = mysql_init(NULL); + if (conn->mysql == NULL) { + i_error("MYSQL: mysql_init failed"); + return FALSE; + } + + if (!mysql_real_connect(conn->mysql, conn->set.db_host, + conn->set.db_user, conn->set.db_passwd, + conn->set.db, + atoi(conn->set.db_port), + conn->set.db_unix_socket, + strtoul(conn->set.db_client_flags, + NULL, 10))) { + i_error("MYSQL: Can't connect to database %s: %s", + conn->set.db, mysql_error(conn->mysql)); + return FALSE; + } + } + + conn->connected = TRUE; + return TRUE; +} + +static void mysql_conn_close(struct mysql_connection *conn) +{ + conn->connected = FALSE; + + if (conn->mysql != NULL) { + mysql_close(conn->mysql); + conn->mysql = NULL; + } +} + +static struct mysql_connection *mysql_conn_find(const char *config_path) +{ + struct mysql_connection *conn; + + for (conn = mysql_connections; conn != NULL; conn = conn->next) { + if (strcmp(conn->config_path, config_path) == 0) + return conn; + } + + return NULL; +} + +static const char *parse_setting(const char *key, const char *value, + void *context) +{ + struct mysql_connection *conn = context; + + return parse_setting_from_defs(conn->pool, setting_defs, + &conn->set, key, value); +} + +struct mysql_connection *db_mysql_init(const char *config_path) +{ + struct mysql_connection *conn; + pool_t pool; + + conn = mysql_conn_find(config_path); + if (conn != NULL) { + conn->refcount++; + return conn; + } + + pool = pool_alloconly_create("mysql_connection", 1024); + conn = p_new(pool, struct mysql_connection, 1); + conn->pool = pool; + + conn->refcount = 1; + + conn->config_path = p_strdup(pool, config_path); + conn->set = default_mysql_settings; + settings_read(config_path, parse_setting, conn); + + (void)mysql_conn_open(conn); + + conn->next = mysql_connections; + mysql_connections = conn; + return conn; +} + +void db_mysql_unref(struct mysql_connection *conn) +{ + if (--conn->refcount > 0) + return; + + mysql_conn_close(conn); + pool_unref(conn->pool); +} + +#endif diff -u -r --new-file work/dovecot-0.99.10/src/auth/db-mysql.h work.patched/dovecot-0.99.10/src/auth/db-mysql.h --- work/dovecot-0.99.10/src/auth/db-mysql.h Wed Dec 31 16:00:00 1969 +++ work.patched/dovecot-0.99.10/src/auth/db-mysql.h Tue Sep 30 12:24:00 2003 @@ -0,0 +1,51 @@ +#ifndef __DB_MYSQL_H +#define __DB_MYSQL_H + +#include <mysql.h> + +struct mysql_connection; +struct mysql_request; + +typedef void mysql_query_callback_t(struct mysql_connection *conn, + struct mysql_request *request, + MYSQL_RES *res); + +struct mysql_settings { + const char *db_host; + const char *db_port; + const char *db_unix_socket; + const char *db; + const char *db_user; + const char *db_passwd; + const char *db_client_flags; + const char *password_query; + const char *user_query; + const char *default_pass_scheme; +}; + +struct mysql_connection { + struct mysql_connection *next; + + pool_t pool; + int refcount; + + char *config_path; + struct mysql_settings set; + + MYSQL *mysql; + + unsigned int connected:1; +}; + +struct mysql_request { + mysql_query_callback_t *callback; + void *context; +}; + +void db_mysql_query(struct mysql_connection *conn, const char *query, + struct mysql_request *request); + +struct mysql_connection *db_mysql_init(const char *config_path); +void db_mysql_unref(struct mysql_connection *conn); + +#endif diff -u -r --new-file work/dovecot-0.99.10/src/auth/master-connection.c work.patched/dovecot-0.99.10/src/auth/master-connection.c --- work/dovecot-0.99.10/src/auth/master-connection.c Mon May 26 08:27:13 2003 +++ work.patched/dovecot-0.99.10/src/auth/master-connection.c Tue Sep 30 12:24:00 2003 @@ -55,7 +55,7 @@ reply.virtual_user_idx = reply_add(buf, user->virtual_user); reply.mail_idx = reply_add(buf, user->mail); - p = strstr(user->home, "/./"); + p = user->home ? strstr(user->home, "/./") : NULL; if (p == NULL) { reply.home_idx = reply_add(buf, user->home); reply.chroot_idx = reply_add(buf, NULL); diff -u -r --new-file work/dovecot-0.99.10/src/auth/passdb-mysql.c work.patched/dovecot-0.99.10/src/auth/passdb-mysql.c --- work/dovecot-0.99.10/src/auth/passdb-mysql.c Wed Dec 31 16:00:00 1969 +++ work.patched/dovecot-0.99.10/src/auth/passdb-mysql.c Tue Sep 30 12:24:00 2003 @@ -0,0 +1,171 @@ +/* Copyright (C) 2003 Alex Howansky, Timo Sirainen */ + +#include "config.h" +#undef HAVE_CONFIG_H + +#ifdef PASSDB_MYSQL + +#include "common.h" +#include "str.h" +#include "strescape.h" +#include "var-expand.h" +#include "password-scheme.h" +#include "db-mysql.h" +#include "passdb.h" + +#include <mysql.h> +#include <stdlib.h> +#include <string.h> + +struct passdb_mysql_connection { + struct mysql_connection *conn; +}; + +struct passdb_mysql_request { + struct mysql_request request; + + enum passdb_credentials credentials; + union { + verify_plain_callback_t *verify_plain; + lookup_credentials_callback_t *lookup_credentials; + } callback; + + char password[1]; +}; + +static struct passdb_mysql_connection *passdb_mysql_conn; + +static void mysql_handle_request(struct mysql_connection *conn, + struct mysql_request *request, MYSQL_RES *res) +{ + struct passdb_mysql_request *mysql_request = + (struct passdb_mysql_request *) request; + struct auth_request *auth_request = request->context; + const char *user, *password, *scheme; + int ret = 0; + + user = auth_request->user; + password = NULL; + + if (res != NULL) { + if (mysql_num_rows(res) == 0) { + if (verbose) + i_info("mysql(%s): Unknown user", user); + } else if (mysql_num_rows(res) > 1) { + i_error("mysql(%s): Multiple matches for user", user); + } else if (mysql_num_fields(res) != 1) { + i_error("mysql(%s): Password query returned " + "more than one field", user); + } else { + MYSQL_ROW row; + + row = mysql_fetch_row(res); + if (row) + password = t_strdup(row[0]); + } + } + + scheme = password_get_scheme(&password); + if (scheme == NULL) { + scheme = conn->set.default_pass_scheme; + i_assert(scheme != NULL); + } + + if (mysql_request->credentials != -1) { + passdb_handle_credentials(mysql_request->credentials, + user, password, scheme, + mysql_request->callback.lookup_credentials, + auth_request); + return; + } + + /* verify plain */ + if (password == NULL) { + mysql_request->callback.verify_plain(PASSDB_RESULT_USER_UNKNOWN, + auth_request); + return; + } + + ret = password_verify(mysql_request->password, password, + scheme, user); + if (ret < 0) + i_error("mysql(%s): Unknown password scheme %s", user, scheme); + else if (ret == 0) { + if (verbose) + i_info("mysql(%s): Password mismatch", user); + } + + mysql_request->callback.verify_plain(ret > 0 ? PASSDB_RESULT_OK : + PASSDB_RESULT_PASSWORD_MISMATCH, + auth_request); +} + +static void mysql_lookup_pass(struct auth_request *auth_request, + struct mysql_request *mysql_request) +{ + struct mysql_connection *conn = passdb_mysql_conn->conn; + const char *query; + string_t *str; + + str = t_str_new(512); + var_expand(str, conn->set.password_query, + str_escape(auth_request->user), NULL); + query = str_c(str); + + mysql_request->callback = mysql_handle_request; + mysql_request->context = auth_request; + + db_mysql_query(conn, query, mysql_request); +} + +static void +mysql_verify_plain(struct auth_request *request, const char *password, + verify_plain_callback_t *callback) +{ + struct passdb_mysql_request *mysql_request; + + mysql_request = i_malloc(sizeof(struct passdb_mysql_request) + + strlen(password)); + mysql_request->credentials = -1; + mysql_request->callback.verify_plain = callback; + strcpy(mysql_request->password, password); + + mysql_lookup_pass(request, &mysql_request->request); +} + +static void mysql_lookup_credentials(struct auth_request *request, + enum passdb_credentials credentials, + lookup_credentials_callback_t *callback) +{ + struct passdb_mysql_request *mysql_request; + + mysql_request = i_new(struct passdb_mysql_request, 1); + mysql_request->credentials = credentials; + mysql_request->callback.lookup_credentials = callback; + + mysql_lookup_pass(request, &mysql_request->request); +} + +static void passdb_mysql_init(const char *args) +{ + struct mysql_connection *conn; + + passdb_mysql_conn = i_new(struct passdb_mysql_connection, 1); + passdb_mysql_conn->conn = conn = db_mysql_init(args); +} + +static void passdb_mysql_deinit(void) +{ + db_mysql_unref(passdb_mysql_conn->conn); + i_free(passdb_mysql_conn); +} + +struct passdb_module passdb_mysql = { + passdb_mysql_init, + passdb_mysql_deinit, + + mysql_verify_plain, + mysql_lookup_credentials +}; + +#endif diff -u -r --new-file work/dovecot-0.99.10/src/auth/passdb.c work.patched/dovecot-0.99.10/src/auth/passdb.c --- work/dovecot-0.99.10/src/auth/passdb.c Sun May 18 05:26:28 2003 +++ work.patched/dovecot-0.99.10/src/auth/passdb.c Tue Sep 30 12:24:00 2003 @@ -110,6 +110,10 @@ if (strcasecmp(name, "pgsql") == 0) passdb = &passdb_pgsql; #endif +#ifdef PASSDB_MYSQL + if (strcasecmp(name, "mysql") == 0) + passdb = &passdb_mysql; +#endif #ifdef HAVE_MODULES passdb_module = passdb != NULL ? NULL : auth_module_open(name); if (passdb_module != NULL) { diff -u -r --new-file work/dovecot-0.99.10/src/auth/passdb.h work.patched/dovecot-0.99.10/src/auth/passdb.h --- work/dovecot-0.99.10/src/auth/passdb.h Thu Mar 20 07:46:33 2003 +++ work.patched/dovecot-0.99.10/src/auth/passdb.h Tue Sep 30 12:24:00 2003 @@ -58,6 +58,7 @@ extern struct passdb_module passdb_vpopmail; extern struct passdb_module passdb_ldap; extern struct passdb_module passdb_pgsql; +extern struct passdb_module passdb_mysql; void passdb_init(void); void passdb_deinit(void); diff -u -r --new-file work/dovecot-0.99.10/src/auth/userdb-mysql.c work.patched/dovecot-0.99.10/src/auth/userdb-mysql.c --- work/dovecot-0.99.10/src/auth/userdb-mysql.c Wed Dec 31 16:00:00 1969 +++ work.patched/dovecot-0.99.10/src/auth/userdb-mysql.c Tue Sep 30 12:24:00 2003 @@ -0,0 +1,160 @@ +/* Copyright (C) 2003 Alex Howansky, Timo Sirainen */ + +#include "config.h" +#undef HAVE_CONFIG_H + +#ifdef USERDB_MYSQL + +#include "common.h" +#include "str.h" +#include "strescape.h" +#include "var-expand.h" +#include "db-mysql.h" +#include "userdb.h" + +#include <mysql.h> +#include <stdlib.h> +#include <string.h> + +struct userdb_mysql_connection { + struct mysql_connection *conn; +}; + +struct userdb_mysql_request { + struct mysql_request request; + userdb_callback_t *userdb_callback; + + char username[1]; /* variable width */ +}; + +static struct userdb_mysql_connection *userdb_mysql_conn; + +static int is_result_valid(MYSQL_RES *res) +{ + int i, n_fields, found; + MYSQL_FIELD *fields; + + if (res == NULL) { + i_error("MYSQL: Query failed"); + return FALSE; + } + + if (mysql_num_rows(res) == 0) { + if (verbose) + i_error("MYSQL: Authenticated user not found"); + return FALSE; + } + + n_fields = mysql_num_fields(res); + fields = mysql_fetch_fields(res); + + /* Make sure the 'uid' field exists. */ + for (found = 0, i = 0; i < n_fields; i++) + if (strcmp("uid", fields[i].name) == 0) { + found = 1; + break; + } + + if (!found) { + i_error("MYSQL: User query did not return 'uid' field"); + return FALSE; + } + + /* Make sure the 'gid' field exists. */ + for (found = 0, i = 0; i < n_fields; i++) + if (strcmp("gid", fields[i].name) == 0) { + found = 1; + break; + } + + if (!found) { + i_error("MYSQL: User query did not return 'gid' field"); + return FALSE; + } + + return TRUE; +} + +static const char *my_get_str(MYSQL_RES *res, MYSQL_ROW row, const char *field) +{ + int i, n_fields; + unsigned long *lengths; + MYSQL_FIELD *fields; + + n_fields = mysql_num_fields(res); + lengths = mysql_fetch_lengths(res); + fields = mysql_fetch_fields(res); + for (i = 0; i < n_fields; i++) + if (strcmp(field, fields[i].name) == 0) + return (const char *) lengths[i] == 0 ? + NULL : t_strdup(row[i]); + + return NULL; +} + +static void mysql_handle_request(struct mysql_connection *conn __attr_unused__, + struct mysql_request *request, MYSQL_RES *res) +{ + struct userdb_mysql_request *urequest = + (struct userdb_mysql_request *) request; + struct user_data user; + MYSQL_ROW row; + + if (res != NULL && is_result_valid(res) && + (row = mysql_fetch_row(res))) { + memset(&user, 0, sizeof(user)); + user.virtual_user = urequest->username; + user.system_user = my_get_str(res, row, "system_user"); + user.home = my_get_str(res, row, "home"); + user.mail = my_get_str(res, row, "mail"); + user.uid = atoi(my_get_str(res, row, "uid")); /* XXX leak */ + user.gid = atoi(my_get_str(res, row, "gid")); /* XXX leak */ + urequest->userdb_callback(&user, request->context); + } else { + urequest->userdb_callback(NULL, request->context); + } +} + +static void userdb_mysql_lookup(const char *user, userdb_callback_t *callback, + void *context) +{ + struct mysql_connection *conn = userdb_mysql_conn->conn; + struct userdb_mysql_request *request; + const char *query; + string_t *str; + + str = t_str_new(512); + var_expand(str, conn->set.user_query, str_escape(user), NULL); + query = str_c(str); + + request = i_malloc(sizeof(struct userdb_mysql_request) + strlen(user)); + request->request.callback = mysql_handle_request; + request->request.context = context; + request->userdb_callback = callback; + strcpy(request->username, user); + + db_mysql_query(conn, query, &request->request); +} + +static void userdb_mysql_init(const char *args) +{ + struct mysql_connection *conn; + + userdb_mysql_conn = i_new(struct userdb_mysql_connection, 1); + userdb_mysql_conn->conn = conn = db_mysql_init(args); +} + +static void userdb_mysql_deinit(void) +{ + db_mysql_unref(userdb_mysql_conn->conn); + i_free(userdb_mysql_conn); +} + +struct userdb_module userdb_mysql = { + userdb_mysql_init, + userdb_mysql_deinit, + + userdb_mysql_lookup +}; + +#endif diff -u -r --new-file work/dovecot-0.99.10/src/auth/userdb.c work.patched/dovecot-0.99.10/src/auth/userdb.c --- work/dovecot-0.99.10/src/auth/userdb.c Sun May 18 05:26:28 2003 +++ work.patched/dovecot-0.99.10/src/auth/userdb.c Tue Sep 30 12:24:00 2003 @@ -49,6 +49,10 @@ if (strcasecmp(name, "pgsql") == 0) userdb = &userdb_pgsql; #endif +#ifdef USERDB_MYSQL + if (strcasecmp(name, "mysql") == 0) + userdb = &userdb_mysql; +#endif #ifdef HAVE_MODULES userdb_module = userdb != NULL ? NULL : auth_module_open(name); if (userdb_module != NULL) { diff -u -r --new-file work/dovecot-0.99.10/src/auth/userdb.h work.patched/dovecot-0.99.10/src/auth/userdb.h --- work/dovecot-0.99.10/src/auth/userdb.h Sun May 18 05:26:28 2003 +++ work.patched/dovecot-0.99.10/src/auth/userdb.h Tue Sep 30 12:24:00 2003 @@ -29,6 +29,7 @@ extern struct userdb_module userdb_vpopmail; extern struct userdb_module userdb_ldap; extern struct userdb_module userdb_pgsql; +extern struct userdb_module userdb_mysql; void userdb_init(void); void userdb_deinit(void);

2 1

Re: [Dovecot] PATCH: mysql authentication
by Matthew Reimer 03 Oct '03

03 Oct '03

Timo Sirainen wrote: > > On Saturday, Aug 30, 2003, at 02:16 Europe/Helsinki, Matthew Reimer > wrote: > >> Here's a patch that implements mysql authentication. I started with >> the pgsql files and tweaked them to use mysql instead. It works for >> me, but there might be a couple of memory leaks. I'm welcome to >> suggestions on how to clean it up so it can be committed. > > Thank you. I'll look at it more closely later, but it looked fine with > a quick look. > > It would be nice to be able to use asynchronous database lookups. I'm > not sure how easy that is with MySQL, of if it's possible at all. With > PostgreSQL it looked annoyingly difficult so I haven't done it yet. > > Synchronous calls anyway mean that there's only one SQL statement > executing at a time and that may slow down the authentication if > there's a _lot_ of users logging in constantly. If that's a problem, > growing auth_count should help at least some. It specifies the number > of auth processes that respond to the authentication queries. > > And some day I'll probably move the sql stuff into separate lib-sql > and have only a single sql authenticator.. I'm not sure what you mean by asynchronous database lookups. Do you mean you'd like to be able to fire off a db lookup and then come back later to get the results? Like blocking vs non-blocking? I've pasted what may be relevant mysql documentation below in case it might help answer the question. N db handles could be allocated, giving us N simultaneous lookups. Matt --- From http://www.mysql.com/doc/en/C_API_function_overview.html: There are two ways for a client to process result sets. One way is to retrieve the entire result set all at once by calling mysql_store_result(). This function acquires from the server all the rows returned by the query and stores them in the client. The second way is for the client to initiate a row-by-row result set retrieval by calling mysql_use_result(). This function initialises the retrieval, but does not actually get any rows from the server. In both cases, you access rows by calling mysql_fetch_row(). With mysql_store_result(), mysql_fetch_row() accesses rows that have already been fetched from the server. With mysql_use_result(), mysql_fetch_row() actually retrieves the row from the server. Information about the size of the data in each row is available by calling mysql_fetch_lengths(). After you are done with a result set, call mysql_free_result() to free the memory used for it. The two retrieval mechanisms are complementary. Client programs should choose the approach that is most appropriate for their requirements. In practice, clients tend to use mysql_store_result() more commonly. An advantage of mysql_store_result() is that because the rows have all been fetched to the client, you not only can access rows sequentially, you can move back and forth in the result set using mysql_data_seek() or mysql_row_seek() to change the current row position within the result set. You can also find out how many rows there are by calling mysql_num_rows(). On the other hand, the memory requirements for mysql_store_result() may be very high for large result sets and you are more likely to encounter out-of-memory conditions. An advantage of mysql_use_result() is that the client requires less memory for the result set because it maintains only one row at a time (and because there is less allocation overhead, mysql_use_result() can be faster). Disadvantages are that you must process each row quickly to avoid tying up the server, you don't have random access to rows within the result set (you can only access rows sequentially), and you don't know how many rows are in the result set until you have retrieved them all. Furthermore, you must retrieve all the rows even if you determine in mid-retrieval that you've found the information you were looking for. --- end

1 0

[Dovecot] can't read attachments using javamail with dovecot
by Jonas Bosson 03 Oct '03

03 Oct '03

Dear folks, this might be a java-bug, but it seems to occur only when connecting to dovecot-IMAP. UW and Cyrus-ditto works, while when I try to read attachments (base64), I get slapped in the face with this: java.lang.NullPointerException at com.sun.mail.iap.Response.parse(Response.java:99) at com.sun.mail.iap.Response.<init>(Response.java:68) at com.sun.mail.imap.protocol.IMAPResponse.<init>(IMAPResponse.java:31) at com.sun.mail.imap.protocol.IMAPResponse.readResponse(IMAPResponse.java:105) at com.sun.mail.imap.protocol.IMAPProtocol.readResponse(IMAPProtocol.java:153) at com.sun.mail.iap.Protocol.command(Protocol.java:215) at com.sun.mail.imap.protocol.IMAPProtocol.fetch(IMAPProtocol.java:882) at com.sun.mail.imap.protocol.IMAPProtocol.fetch(IMAPProtocol.java:874) at com.sun.mail.imap.protocol.IMAPProtocol.fetchBody(IMAPProtocol.java:677) at com.sun.mail.imap.IMAPInputStream.fill(IMAPInputStream.java:74) at com.sun.mail.imap.IMAPInputStream.read(IMAPInputStream.java:105) at com.sun.mail.util.BASE64DecoderStream.decode(BASE64DecoderStream.java:143) at com.sun.mail.util.BASE64DecoderStream.read(BASE64DecoderStream.java:54) at se.illuminet.httpd.mail.MailXMLService.serviceXml(MailXMLService.java:361) This occurs in booth javamail 1.2 and 1.3(.1) using dovecot. Of course this unfriendly/unhandled response is a problem with javamail while biting the dust. Further analysis points to that the reading works OK up until 2^16 bytes, so it fails either in reading the tail of the attachment or resuming after the first buffered data is consumed. I tried the BODYSTRCUTURE-patch (booth #1 and then #2), but it didn't help. Now I am fresh out of ideas. The java-bug has a three week delay to surface on "bug-parade". But I still have to explain why it works in UW/Cyrus and not with dovecot. Next step would be to add trace-lines in dovecot. Any ideas? /Jonas

2 6

[Dovecot] pop3s in Outlook
by Antonin Karasek 03 Oct '03

03 Oct '03

Hi, I'm using Dovecot for pop3-ssl. Unofortunatelly, most of my users are using MS Outlook. I want them to be able to remove standard mesage: "The certificat is not signed by...", but I don't know how :-( In the dialog isn't a button "Trust this certificate forever"... Is there anybody, who know, how to solve this? Many thanks.

2 1

[Dovecot] Handling of X-Status flags
by Bernard Blackham 29 Sep '03

29 Sep '03

Hi all, I've done a bit of searching and haven't seen anybody else coming across this problem yet: I noticed that Dovecot uses the opposite flags for marking messages as deleted and draft to most other MUAs I've seen or used: With PINE, mutt and uw-imapd at least, messages are marked with X-Status 'D' for Deleted, and 'T' for drafT. Dovecot reverses this behaviour, using 'D' for Draft, and 'T' for Trashed. Was there any reason behind this? Or have I gone loopy? At our site, we have users reading mail with a mix of PINE locally, and through dovecot via WebMail and there seems to arise a bit of confusion between the two flags. Attached is a patch that reverses this behaviour (prepared against 0.99.10 but applies with a little fuzz to cvs). I'd like to hear people's comments on the issue... is it just me? :) Thanks, Bernard. -- Bernard Blackham bernard at blackham dot com dot au

2 2

[Dovecot] UIDPLUS?
by Keith Edmunds 29 Sep '03

29 Sep '03

Any plans to implement UIDPLUS in Dovecot? Keith

2 1

[Dovecot] Odd problem with dovecot segfaulting
by Tim Miller 29 Sep '03

29 Sep '03

Hi All, I installed dovecot on one of my servers the other day, and I want to use it in a vhosting environment. To this end, I have set it up to use password file authentication with a passwd file specified by me. This file will contain both system users and virtual accounts. To test it, I set up two accounts in the dovecot passwd file. One is a system user, I give his home dir in the correct field, along with the MD5 hash of his password in its field, and everything seems to work OK (he can login, read mail out of /var/spool/mail/user etc.). The seconmd account is a vhosted account, and I set the uid and gid to that under which the vhosting account exists. I did not set a home dir, but I set a mail dir in /var/spool/vmail/imap/vhost/user, and made this dir group writeable by dovecot. The user can login, but immediately upon authentication, dovecot segfaults (the log reports an "Internal login failure" and the authentication process is killed with signal 11, i.e. a segfault). Has anyone encountered this before? Is this even the right way to set up vhosting (I don't want to go with LDAP or SQL authentication right at the moment, but I may move to that later on)? Are there any things I should look for (this is probably due to something I've done wrong) or make dovecot be a bit more verbose in reporting errors to its log? Thanks in advance for any help. From what I've seen so far, dovecot looks like an excellent product. Best, Tim Miller

2 1

[Dovecot] pine and mozilla at the same time
by Gasha 26 Sep '03

26 Sep '03

Hi all, what is the best way to read e-mail from server using "pine" ?? I would like to use graphical IMAP client, mozilla for example. And sometimes run pine from shell. mbox or Maildir? Or maybe setup pine to use IMAP server on "localhost" ?? Gasha

2 1

[Dovecot] can't read attachments using javamail with dovecot
by Jonas Bosson 25 Sep '03

25 Sep '03

Dear folks, this might be a javamail-bug, but it seems to occur only when connecting to dovecot-IMAP. UW and Cyrus-ditto works, while when I try to read attachments (base64), I get slapped in the face with this: java.lang.NullPointerException at com.sun.mail.iap.Response.parse(Response.java:99) at com.sun.mail.iap.Response.<init>(Response.java:68) at com.sun.mail.imap.protocol.IMAPResponse.<init>(IMAPResponse.java:31) at com.sun.mail.imap.protocol.IMAPResponse.readResponse(IMAPResponse.java:105) at com.sun.mail.imap.protocol.IMAPProtocol.readResponse(IMAPProtocol.java:153) at com.sun.mail.iap.Protocol.command(Protocol.java:215) at com.sun.mail.imap.protocol.IMAPProtocol.fetch(IMAPProtocol.java:882) at com.sun.mail.imap.protocol.IMAPProtocol.fetch(IMAPProtocol.java:874) at com.sun.mail.imap.protocol.IMAPProtocol.fetchBody(IMAPProtocol.java:677) at com.sun.mail.imap.IMAPInputStream.fill(IMAPInputStream.java:74) at com.sun.mail.imap.IMAPInputStream.read(IMAPInputStream.java:105) at com.sun.mail.util.BASE64DecoderStream.decode(BASE64DecoderStream.java:143) at com.sun.mail.util.BASE64DecoderStream.read(BASE64DecoderStream.java:54) at se.illuminet.httpd.mail.MailXMLService.serviceXml(MailXMLService.java:361) This occurs in booth javamail 1.2 and 1.3(.1) using dovecot. Of course this unfriendly/unhandled response is a problem with javamail while it bites the dust. Further analysis points to that the reading works OK up until 2^16 bytes, so it fails either in reading the tail of the attachment or resuming after the first buffered data is consumed. I tried the BODYSTRCUTURE-patch (booth #1 and then #2), but it didn't help. Now I am fresh out of ideas. The java-bug has a three week delay to surface on "bug-parade". But I still have to explain why it works in UW/Cyrus and not with dovecot. Next step would be to add trace-lines in dovecot. Any ideas? /Jonas

1 0

[Dovecot] Multiple instances of .customflags staying open
by Ron 25 Sep '03

25 Sep '03

# cat /proc/version Linux version 2.4.20-18.9 (bhcompile(a)daffy.perf.redhat.com) (gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)) #1 Thu May 29 07:09:43 EDT 2003 # dovecot --version 0.99.10 I've just recently noticed a serious problem with my system. After about an hour or so I get errors in /var/log/maillog such as this; Sep 21 06:40:16 the-system imap(ray): opendir() failed with file \ \/home/ray/Maildir//new: Too many open files I investigated with 'lsof' and found that many copies of '.customflags' were open for each and every folder. ie; Each folder's '.customflags' file was open 15 times! I suspect that postfix is not able to store newly collected email while this is occuring, therefore I am likely losing messages. I've tried with multiple mail clients (Sylpheed & Evolution) and multiple users, always with the same result. Further investigation indicates that each time the client scans for new mail another instance of each and every '.customflags' is opened. As I have the client set for checking at five minute intervals then it was probably 75 minutes until the client barfed. Less, if I manualy scan for new mail a few times. If I shutdown the client and restart it then everything is fine for about an hour. Running 'lsof /home/ray/Maildir/.customflags' on the server I can see the list of files increasing each time the client checks for new mail. Any help would be appreciated!

4 14

[Dovecot] Re: Bug: Message time gets changed during move/copy
by Timo Sirainen 25 Sep '03

25 Sep '03

On Sat, 2003-09-13 at 13:10, Akhilesh Agarwal wrote: > SERVER > > * Redhat 9.0 stable (with all up2date updates installed) > * Sendmail-8.12.8-6.90 > * Dovecot 0.99.10 release > * On the workstation running Outlook Express, there should be > some existing messages in Local Folders. (example: message > downloaded from a POP3 server) > * Change the time zone of the workstation machine to non-GMT > time zone. In my case GMT +0530 (Indian Standard Time) > * Create a test IMAP account on the server > * Configure Outlook Express to connect to the IMAP account. > * Now select a message in the Local Folders and copy it to a > folder on the IMAP server. > * If you check the message time on the IMAP account, you will > find that +0530 hrs was added to the message time. (+0530 is > apparently the time zone of my workstation as configure above) Sorry for taking this long to reply, it always takes so much effort to turn on my Windows computer :) I cannot reproduce this. I have my server and workstation set to GMT+3 and copying message from Dovecot to OE6 local folder and back keeps the message date the same. Or actually it looks like IMAP -> local folder copies using the message's Date field rather than IMAP's INTERNALDATE which Received-field shows. I'm not sure if this could be Dovecot's or OE6's fault .. Has anyone else seen timestamps going wrong when copying? > Can anyone please confirm this and let me know, as I am not subscribed > to the list. It actually never made it to list because it was decided to be spam (mostly due to being HTML).

1 0