> I don't know how to check that mesa isn't interfering, but I'm pretty > sure the other two points are okay.
It was a long shot; I though perhaps some libGL*.so was being loaded from somewhere else on the search path; but it's unlikely. > > > A good start in tracking down the problem would be to check that the > > arguments passed to glXCreateWindow seem reasonable (i.e., not null). > > These are in pyglet/window/xlib/__init__.py. > > These arguments are not null, but I didn't did any deeper into their > values. I'm happy to do so if you think there's anything interesting. No, it looks to me like a driver bug. I just noticed that glxgears doesn't use glXCreateWindow... let me know if you happen to have a GL program that does installed and whether or not it works (running `strings` on the executable should let you know if it uses glXCreateWindow). > > > It might also be worth trying the GLX 1.1 path, which doesn't use > > glXCreateWindow -- patch the have_version() method in > > pyglet/gl/glx_info.py to always return False. > > This seemed to work for me. FWIW -- the versions here are 1.4 for > client and server. > > I ran tests/test.py, and everything seems to pass now. However, by > disabling glx, does that mean I am not getting hardware acceleration? > The opengl.py torus example, for instance, seems much slower than I > would normally expect from this machine... pyglet checks to see if GLX 1.3 or later is available, and if so, uses the new glXCreateWindow API. If the version is <= 1.2, pyglet uses the older glXCreateContext API. This difference is only during the setup of the rendering context, and may affect which options are selected. If you run tools/gl_info.py you can see what options in the context have been created; it's possible that you're not getting a hardware context (but unlikely, in my experience with the nvidia driver). Be aware that the torus example may just be running slowly relative to a C program -- Python incurs a large overhead over the runloop, so the bottleneck might not be in graphics. I'll add a workaround into pyglet to force usage of GLX 1.1 when a Quadro 550 is detected unless you or someone else with this hardware can get some more information. Alex. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "pyglet-users" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pyglet-users?hl=en -~----------~----~----~----~------~----~------~--~---
