Re: VBO glBufferDataARB fix

On Thu, Apr 05, 2007 at 12:47:31PM -0400, Mike C. Fletcher wrote:
> Thank you very much Andreas!  I've applied the patch to my workstation's 
> CVS.  Just need to get back home and write a test to be sure it's 
> working as applied.  If you happen to have a simple test case I can use 
> that shows the failure and the success it would be great.  If not I'll 
> just see if I can get my old test case under OpenGLContext running again.

Sorry, I don't have an isolated test case around.  Originally I found
the possible reason for the bug by experimenting with padding source
arrays to multiples of their size.  Plus the one blog post of yours
where you mentioned that some parts seem to be missing in the display of
some VBO demo code.

However I found out that testing the fix can get a bit peculiar.  The
driver (NVidia in my case) can reliably assign the same buffer memory
for subsequent runs, apparantly with enough padding/alignment so that
multiple buffers are separated despite varying sizes.  This means
running the fix can leave data where a subsequent unfixed run can pick
it up and appear to work correctly even though it didn't actually upload
the complete data.

-- 
Andreas Bombe <[email protected]>    GPG key 0x04880A44

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
PyOpenGL Homepage
http://pyopengl.sourceforge.net