On Wed, Nov 10, 2010 at 3:41 AM, Thiago Macieira <[email protected]> wrote: > On Wednesday, 10 de November de 2010 09:32:48 Dawit A wrote: >> sigh I got everthing crossed up! Indeed since this is the system bus >> a sendmsg here is response back to the client. For some reason this >> response does not seem to make it to the client or at least cause the >> client to react until 25 secs later. Now it is even more weird... >> >> > That is actually showing that the target replied immediately. >> >> hmm... should I then trace the client and see what it is waiting on ? > > Yes. > > If the client was given the reply but only reacted to it 25 seconds later, it > probably means it was stuck (in any thread) waiting for another reply thta > didn't come. > > You'll need to start the program with strace now, so you can use -f.
This issue seems to be caused by some sort of race condition because when I build dbus 1.4.0 in debug mode and install it, I do not see the problem anylonger. As a result I cannot even generate any backtrace into dbus to see what is going on... If I then re-install the version without debug symbols then viola the problem comes back. Oh well... On a side note, I see a lot of the following error message in my ~/.xsession-errors file: QObject: Cannot create children for a parent that is in a different thread. (Parent is Solid::Backends::UDisks::UDisksDevice(0xf04820), parent's thread is QThread(0xc90d90), current thread is ThreadWeaver::Thread(0x111fb60) No clue which application/library is causing that to be printed...
