[Dovecot] Dovecot raw backtrace when copying to folder

[Bernd Kuhls]

Timo Sirainen wrote:
> On Sat, 2007-09-29 at 20:06 +0200, Bernd Kuhls wrote:
>> dovecot: 2007-09-25 16:54:08 Panic: Lost login process PID 420
> ..
>> Dovecot is started through inetd.
> 
> Running from inetd hasn't been tested all that much. I'll see if I can
> get this fixed, but you could avoid this by not using inetd.
> 

Hi,

unless you can provide another way to restrict access to IMAP and POP3 
to our local IP range this is not an option for us. Debian Etchs 
tcpwrapper has been patched 
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=377154
to allow a syntax like this:

# imap
143: 127.0.0.1, x.y.z.
# imaps
993: ALL

# pop3
110: 127.0.0.1, x.y.z.
# pop3s
995: ALL

which requires starting Dovecot using inetd.

I re-compiled Dovecot 1.0 yesterday using the latest hg checkout:

# hg parents
changeset:   5419:f85adea35b74
tag:         tip
user:        Timo Sirainen <tss at iki.fi>
date:        Mon Sep 24 11:15:48 2007 +0300
summary:     Don't crash with empty messagesets.

removing dh_strip from debian/rules to provide a better backtrace:

#!/bin/sh
cd /home/kuhls
rm -rf ./dovecot_compile
mkdir ./dovecot_compile
cd ./dovecot_compile
hg clone http://hg.dovecot.org/dovecot-1.0/
wget http://ftp.debian.org/debian/pool/main/d/dovecot/dovecot_1.0.5-1.dsc
wget 
http://ftp.debian.org/debian/pool/main/d/dovecot/dovecot_1.0.5-1.diff.gz
cd dovecot-1.0
zcat ../dovecot_1.0.5-1.diff.gz | patch -p1 -E
chown -R root *
grep -v dh_strip ./debian/rules > ./debian/rules2
mv ./debian/rules2 ./debian/rules
chmod +x ./debian/rules
./autogen.sh
dpkg-buildpackage -uc -us -b

# build snarf-plugin -> 
http://www.mail-archive.com/dovecot%40dovecot.org/msg02174.html
export DOVECOT=/home/kuhls/dovecot_compile/dovecot-1.0
wget http://dovecot.org/patches/1.0/mbox-snarf-plugin.c
gcc -fPIC -shared -Wall -I$DOVECOT -I$DOVECOT/src/lib \
-I$DOVECOT/src/lib-storage -I$DOVECOT/src/lib-mail \
-I$DOVECOT/src/lib-imap -DHAVE_CONFIG_H \
mbox-snarf-plugin.c -o ../mbox_snarf_plugin.so

I hope this is sufficient to provide the necessary informations.
Until now no new crash in the logs. I will try to reproduce the bug.

Greetings, Bernd