Hi Andreas,

You can switch off extensions using
osg::setGLExtensionDisableString(std::string) that you'll find in
include/osg/GLExtensions.  Have a look at the header for docs on the
method and the associated env var.

Robert.

On 10/27/06, Andreas Goebel <[EMAIL PROTECTED]> wrote:
Hi,

as Robert said that there had been fixes checked in because of
NVidia-Driver-bugs, I have tried the latest cvs with mingw.

The crash when setting up an fbo is still there.

I will paste my bug-analysis down there again, maybe it got lost during
the highland-gathering. Personally I still suspect a problem with the
mingw-openGL libs, but I didn´t manage to replace them with the MS-Libs
(I do suspect that because the MSVC7.1 Version by Mike works).

If nobody has any idea how to fix this (I have tried, but I ran out of
ideas, and bertrands fix didn´t work for me), I would like to know if
there is a way to turn off the use of fbo system-wide.

Regards,

Andreas

-------------------------------------------------------------------------------
Here my bug-analysis, which uses stone-age-methods (cout the linenumber
the program is at) as I didn´t get a stacktrace from gdb:


What I found out is _very_ strange. When setting up the FBO this is done
in RenderStage.cpp.

In Line 331 (I hope the number is correct, I might have changed the
linenumbers by inserting something) we have

osg::notify(osg::INFO)<<"Setting up
osg::CameraNode::FRAME_BUFFER_OBJECT"<<std::endl << std::flush;
          which was the last line of debug-information I get before the
crash.

Then the last line that is being executed is

fbo->apply(state);

(which is nr. 384 in my code).

The whole function apply(state) is being executed (I printet out
linenumbers there, too), but the line after the function-call is never
reached. (I have a cout after the fbo->apply(state) which doesn´t get
called).

So the program crashes while returning from that function, can this
happen? Or could it be that the program crashes in a different thread?


So at least it is clear that this problem occurs on all
MinGW-installations whose users have responded to my call. Maybe Robert
could comment on this?

I tried my best to localize the problem, sorry I don´t get a stacktrace.
I could not reproduce the program working in debug mode, and I could not
reproduce the program working when commenting out the pointsprite-thing.

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

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

Reply via email to