From kde-commits  Fri Jun 02 06:01:16 2006
From: Ralf Habacker <ralf.habacker () freenet ! de>
Date: Fri, 02 Jun 2006 06:01:16 +0000
To: kde-commits
Subject: Re: kdesupport/qt-dbus
Message-Id: <447FD42C.9050903 () freenet ! de>
X-MARC-Message: https://marc.info/?l=kde-commits&m=114922827718227

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thiago Macieira schrieb:
> Ralf Habacker wrote:
>> The problem i believe is, that QMutex:unlock() is called with a mutex
>> already unlocked and that QMutex:unlock() does not detect this case.
> 
> Yes, it does. That's why it prints the warning in the first place.

But the message
"ASSERT failure in QMutex::unlock(): "A mutex must be unlocked in the
same thread that locked it.", file thread\qmutex.cpp, line 248" looks a
bit misleading. It suggests that there is a thread problem. Instead
there is a repeated unlock.

> But it freezes the application because you were not supposed to unlock 
> anything at that point.
> 
> So I ask again: which mutex is being unlocked and at which point in the 
> code?
Just took a look into qt-dbus source and saw that there is no mutex
usage, so it must be in dbus-daemon.

I found one additional DBUS_UNLOCK() call in _dbus_close(), removing of
this fix this issue, but I'm wondering why this unlock results in a
QMutex problem.

Thanks for your help.

BTW: Do you have found any hints relating to the other test case failures ?

Regards
Ralf

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEf9QsoHh+5t8EXncRAg+7AKCJuQb3n+niexR01aUqE8LTVoZr6ACfTjJk
TGmlHCEfigcJNGV1HKjiBkY=
=Q/tb
-----END PGP SIGNATURE-----