Hi,

We also observed this problem in Wengo, and the difficulty of debugging
what happens after a suspend got in the way of solving the problem. gdb
does not give particularly useful information.

The only way I can think of to debug these kinds of threading issues is
an old method I cann "pepper printfs everywhere". You need to identify
key entry points, have printfs for when you enter & leave, from which
thread, and see where your assumptions are wrong.

I am interested in what makes you think that the majority of users are
laptop users.

Cheers,
Dave.

Roland Alton-Scheidl wrote:
> We have done some long-term observations both with kvats and qutecom.
> 
> In most cases, freeze happens after a longer suspend period. This annoys all 
> people, who run kvats or qutecom on a laptop, and this will be the majority 
> of 
> users. 
> 
> After some discussions with Laurent in Krems, it seems like a deadlock 
> between 
> boost and Qt.
> 
> How can we debug this? Neither gdb does not throw any useful information nor 
> do we have an indicator in the log files with OWLOGGER_DEFAULT set to 0. Can 
> we turn on debugging for the semaphores?
> 
> Our observation is, that a network change does most time not result in a 
> freeze, however, SIP does not re-connect properly in most cases.
> 
> --RAS
> 

-- 
maemo.org docsmaster
Email: [EMAIL PROTECTED]
Jabber: [EMAIL PROTECTED]

_______________________________________________
QuteCom-dev mailing list
[email protected]
http://lists.qutecom.org/mailman/listinfo/qutecom-dev

Reply via email to