Hi,
I have a GLJPanel in a JInternalFrame - works well.
Normally, when I close the JInternalFrame, there are no problems. I can open and close JInternalFrames for an hour or more without problems. Then, all of a sudden, I will get this error:
The result of the error is that I lose the JInternalFrame rendering, but the GLJPanel is still fully working, with my mouse Arcball and events and everything.
Any advice would be most welcome.
TIA,
Pinky
This is pretty clearly a bug in your vendor’s OpenGL drivers. What graphics card do you have? Have you downloaded and installed their latest driver set?
My card is GeFORCE MX 4000 - nVidia driver version 6.14.10.9371 which came out on the 22 Oct 2006.
I will try new drivers and let you know result. I should experience the bug sometime today if it’s still there. It’s an elusive creature, unfortunately. The worst kind of bug 
EDIT
The driver 6.14.10.9371 is the actual current driver. So this problem is still unsolved.
Sorry, I don’t have any good suggestions. If you can boil things down into a test case (for example, the JRefract demo has a stress mode where it repeatedly opens and closes a JInternalFrame – does this eventually fail for you?) I can try to reproduce the issue here. However this is pretty clearly a bug in the pbuffer support for your particular graphics card. NVidia is the right place to report this bug though if it’s only intermittent there’s very little chance it’s going to be fixed. Sorry. You should probably spring for a new graphics card anyway, which will probably improve reliability.
Yes, the ‘Loop Gears demo’ in the JRefract example eventually fails in the same manner, after about 2-3 minutes of running.
Hehe - I intentionally use a low-end gfx card to develop with, to maximise the number of machines our code will work on, and to experience these sorts of problems before our customers do! But yeah, an MX 4000 is probably a little bit too low end.
Thanks again for your help. 
Cheers,
Pinky