[visionegg] Re: New User, Seg Fault

  • From: Andrew Straw <andrew.straw@xxxxxxxxxxxxxxx>
  • To: visionegg@xxxxxxxxxxxxx
  • Date: Wed, 17 Sep 2003 10:47:23 +0930

Dear Asif,

That being said, I've already encountered a problem running the demos.  I've gone through the step-by-step Windows installation guide (I'm running XP), and everything seems to have proceeded without a hitch.  check-config.py also confirms that everything is OK, with the exception of the optional modules Pyro and winioport which failed.

Pygame Parachute Traceback:
  File "C:\Python22\Lib\site-packages\VisionEgg\PlatformDependent.py", line 192, in sync_swap_with_vbl_post_gl_init
Fatal Python error: (pygame parachute) Segmentation Fault

The problem seems to occur when trying to synchronize buffer swaps with vertical sync. To get the demos running, turn off this feature in the GUI console. (The Windows Display settings control panel may allow you to set this feature anyway.)


Here's some advice about isolating it further. There's not a lot I can do, since this works on the myriad Windows computers I have access to.

It appears that the offending line (192) in PlatformDependent.py is the if statement below:

import OpenGL.WGL.EXT.swap_control
if OpenGL.WGL.EXT.swap_control.wglInitSwapControlARB(): # Returns 1 if it's working
OpenGL.WGL.EXT.swap_control.wglSwapIntervalEXT(1)


Try running the above lines direct from the Python interpreter and see what happens. It looks like either a PyOpenGL or possibly a driver problem to me. Perhaps you can update your OpenGL drivers? (It looks like this in the onboard video of an Intel 845 mainboard? Perhaps you need to update your mainboard drivers?) If you can isolate any further, we can submit a question or bug report to the PyOpenGL team.

Hmm, looking at the output of check-config.py again, it has listed:

VISIONEGG_USER_DIR = H:\VisionEgg

which doesn't exist; everything has been installed to the C: directory.  Might this have something to do with the problem?  If so, how do I change this?

That's strange... That means that os.path.expanduser("~") returns "H:". If what you say is true, this sounds a bug in that Python function, which is possible but unlikely. Is there an H: drive at all?


Cheers!
Andrew

Other related posts: