On Oct 2, 2008, at 1:01 AM, Rene Luria wrote:
Le 1 oct. 08 à 23:47, Rene Luria a écrit :
Here is a patch against 1.1.3 solving this issue
I forgot the link :)
http://www.luria.ch/9e93676668e453b739ee3e085434e086/dovecot-lib-mail.diff
I actually did several error handling fixes to message-
Le 1 oct. 08 à 23:47, Rene Luria a écrit :
Here is a patch against 1.1.3 solving this issue
I forgot the link :)
http://www.luria.ch/9e93676668e453b739ee3e085434e086/dovecot-lib-mail.diff
smime.p7s
Description: S/MIME cryptographic signature
Ok, found the problem.
Here is a patch against 1.1.3 solving this issue
it comes from the "undisclosed-recipients:;" string
and incrementing ctx->parser.data going after the end of the buffer
maybe there are other issues like this one because in many other
places in message-address.c the poin
Ralf Hildebrandt a écrit :
>> (the problematic email is anonimized, tell-me if you need the originalone)
> #3 0x080c86e6 in message_address_parse (pool=0x8139270, data=0x813d668
> "undisclosed-recipients:;, Postkorb\" <[EMAIL PROTECTED]>il.net>",
> That header looks fairly broken. But with all
* Rene Luria <[EMAIL PROTECTED]>:
> (the problematic email is anonimized, tell-me if you need the originalone)
#3 0x080c86e6 in message_address_parse (pool=0x8139270, data=0x813d668
"undisclosed-recipients:;, Postkorb\" <[EMAIL PROTECTED]>il.net>",
That header looks fairly broken. But with all
Dovecot dies with signal 11 (segfault) when doing some commands with a
specific message
you'll find relevant information stored here:
http://www.luria.ch/9e93676668e453b739ee3e085434e086/
the email:
msg.anon.txt
input rawlog:
20081001-120137-24371.in
dovecot -n:
dovecot-n.txt
the gdb backtrace
