Thank You,

I noticed that in Windows changing Threading Optimization to ON in NVidia
driver properties does remove the issue in OSG 1.2. Unfortuanately curent
SVN version is still crashing. So suspicion that it may be NVidia seems well
founded.

Big thanks again,
Wojtek


-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Craig, Joel S.
(SED/TMI)
Sent: Friday, April 20, 2007 3:57 PM
To: osg users
Subject: RE: [osg-users] Threading model + multi stage texture.
Bug(UNCLASSIFIED)


Classification:  UNCLASSIFIED
Caveats: NONE

I changed to viewer.realize(Producer::CameraGroup::SingleThreaded) and it
does not crash.
I changed back to viewer.realize(Producer::CameraGroup::ThreadPerCamera) and
I get SIG32 signal.

Unfortunately, I am using Nvidia, so I can't rule out the driver
possibility, but it does crash on Red Hat Linux. Actually, outside of the
debugger, it hangs on exit rather than crashes. The screen remains blue and
I have to CTRL-ALT-DEL to interrupt the window manager (GDM).

Joel

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Wojciech
Lewandowski
Sent: Friday, April 20, 2007 12:29 AM
To: osg users
Subject: Re: [osg-users] Threading model + multi stage texture.
Bug?(UNCLASSIFIED)

Thanks, application does just that - its textured redish cube. But all i
wanted to know if it crashes on exit in Linux. And it looks like it does ?
Similar thing happens in Windows

Can you test once again and check if changing viewer.realize to (
osg::Viewer::SingleThreaded ) instead of ( osg::Viewer::ThreadPerCamera )
removes the problem ?

Big Big Thanks,

Wojtek Lewandowski

----- Original Message -----
From: "Craig, Joel S. (SED/TMI)" <[EMAIL PROTECTED]>
To: "osg users" <[email protected]>
Sent: Thursday, April 19, 2007 5:54 PM
Subject: Re: [osg-users] Threading model + multi stage texture. Bug
?(UNCLASSIFIED)


Classification:  UNCLASSIFIED
Caveats: NONE

Greetings,
    I'm not sure what is supposed to happen with your app, but when I ran
it, I got a SIG32 on the "return 0" statement. I used the 1.2 code.

I'm running on a Dell x86-64 dual core processor with Nvidia Quadro FX-1400,
driver version 1.0-8178 with Red Hat Linux.

When I run the app, the display shows the cube. I spin it, then press ESC.
It removes the cube, but sometimes doesn't return control back to the window
manager. Sometimes it does return to the window manager. I ran it with gdb
compiled with "-g". I tried using the breakpoint supplied in the code, but
that's not a valid address in my executable.

(gdb) run
[Thread debugging using libthread_db enabled] [New Thread 182...] [New
Thread 108...] [New Thread 109...] [New Thread 110...] [Thread 110...
exited]

Program received signal SIG32, Real-time event 32 [Switching to thread
108...]
(gdb) backtrace
#0 0x..... in __select_nocancel()
#1 0x..... in XEnq()
#2 0x..... in _XRead()
#3 0x..... in _XReadEvents()
#4 0x..... in XNextEvent()
#5 0x..... in Producer::RenderSurface::_checkEvents()
#6 0x..... in Producer::RenderSurface::run()
#7 0x..... in OpenThreads::ThreadPrivateActions::StartThread()
#8 0x..... in start_thread()
#9 0x..... in clone
#10 0x00000000 in ?? ()

The other threads were in sched_yield() called from
Producer::RenderSurface::~RenderSurface$delete()

Hope that helps a bit...
Joel
Classification:  UNCLASSIFIED
Caveats: NONE

_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/

-------------------------------------------------------------
This mail was scanned by BitDefender
For more informations please visit http://www.bitdefender.com


-------------------------------------------------------------



-------------------------------------------------------------
This mail was scanned by BitDefender
For more informations please visit http://www.bitdefender.com


-------------------------------------------------------------


_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
Classification:  UNCLASSIFIED
Caveats: NONE



-------------------------------------------------------------
This mail was scanned by BitDefender
For more informations please visit http://www.bitdefender.com


-------------------------------------------------------------


_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/

Reply via email to