[muscle] Re: QMutex::lock() segfaults - SOLVED

  • From: Marius Kjeldahl <marius@xxxxxxxxxxxx>
  • To: muscle@xxxxxxxxxxxxx
  • Date: Tue, 11 Nov 2003 13:11:16 +0100

After recompiling Qt with debug information (which did not give a lot of 
additional information), I also attempted to rebuild glibc at my end, but 
that did not solve it.

I took a big breath and stepped back and upgraded to 2.45. In addition, I 
tried adding the MUSCLE .cpp file dependencies directly, so that they got 
compiled into my application instead of linked in through a static library. 
After doing this, it works. 

In short, I am not sure why, but at least I got it working. I would speculate 
that different compiler options when compiling MUSCLE on its own and stuffing 
the object files into a library file that was then linked to my Qt based 
client is to blame.

Thanks for the help anyway.


Marius K.

On Monday 10 November 2003 23:30, Jeremy Friesner wrote:
> On Monday 10 November 2003 02:02 pm, you wrote:
> > I'm getting segfault in QMutex::lock(). Is anybody else having trouble =
> with
> > these? I'm running qt-copy (from KDE cvs) and MUSCLE 2.44 on a fairly
> > updated Gentoo Intel system.
> >
> > Any help or hints on where to search further would be appreciated.
> One possibility I can think of is that perhaps you forgot to declare a=20
> CompleteSystemSetup object at the top of your main() function... your mai=
> n()=20
> should look something like this:
> int main(int argc, char ** argv)
> {
>    muscle::CompleteSetupSystem css;
> [...rest of program here...]
> }
> Another possibility is that maybe you need to declare your QApplication o=
> bject=20
> earlier on... perhaps it does some behind-the-scenes setup that QMutex=20
> requires.
> Let me know if neither of those are it!
> Jeremy

Other related posts:

  • » [muscle] Re: QMutex::lock() segfaults - SOLVED