Hello Carsten,
I just wanted to confirm that this issue is fixed for us.
I found out that a colleaque of me had compiled an additional field container
type into our main application a long time ago.
When I added that extension library unloading that old field container inside
the main app seemed to cause that crash during osgExit().
I simply moved it to our extension library and everything seems to run fine
again!
Thanks for your support,
Michael
-------- Original-Nachricht --------
> Datum: Mon, 16 May 2011 14:11:48 -0500
> Von: Carsten Neumann <carsten_neum...@gmx.net>
> An: opensg-users@lists.sourceforge.net
> Betreff: Re: [Opensg-users] OpenSG1.8 - Crash in osgExit()
> Hello Michael,
>
> On 05/16/2011 01:14 PM, Michael Raab wrote:
> > I'm calling osgExit() explicitly before shutting down our application.
> > I don't see any reason why our extension library should be already
> > unloaded at this moment.
> > Do you have an idea how I can track down the field container type causes
> > that problem.
>
> only the two more or less obvious choices of running everything under a
> debugger and digging backwards from the crash's stack trace and perhaps
> putting a printf() into FieldContainerFactory::terminate(), inside the
> for loop to at least get the type id it crashes on.
>
> > Which function is called for each FC type during shutdown?
>
> FieldContainerType::terminate().
>
> > Are there any log messages are generated?
>
> not that I've seen it, looking over the code.
>
> However, I did see that osgExit() does this:
>
> SharedObjectHandler::the()->terminate();
>
> if(osgSystemExitFunctions != NULL)
> {
> for(Int32 i = osgSystemExitFunctions->size() - 1; i >= 0; i--)
> {
> returnValue &= (*osgSystemExitFunctions)[i]();
>
> if(returnValue == false)
> break;
> }
> }
>
> Can you try moving the call to SharedObjectHandler::terminate() after
> the "if"? Or use
> SharedObjectHandler::the()->removeSharedObject("your_library_name");
> before calling osgExit() to unload your extension library.
>
> Normally unloading a shared lib would result in the types deregistering
> themselves, but if GlobalSystemState == Shutdown (one of the first
> things osgExit does is setting this), this does not happen - so this is
> a bug. I'm not sure yet if adjusting osgExit as above is the correct way
> to fix this.
>
> Cheers,
> Carsten
>
> ------------------------------------------------------------------------------
> Achieve unprecedented app performance and reliability
> What every C/C++ and Fortran developer should know.
> Learn how Intel has extended the reach of its next-generation tools
> to help boost performance applications - inlcuding clusters.
> http://p.sf.net/sfu/intel-dev2devmay
> _______________________________________________
> Opensg-users mailing list
> Opensg-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/opensg-users
--
NEU: FreePhone - kostenlos mobil telefonieren und surfen!
Jetzt informieren: http://www.gmx.net/de/go/freephone
------------------------------------------------------------------------------
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay
_______________________________________________
Opensg-users mailing list
Opensg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensg-users
