Paul Miller wrote:
>
> A lot of discussion here has been on the side of the driver-writer. As
> an application developer interested in multi-head OpenGL applications,
> I'm wondering what would happen in a (contrived) scenario like this:
>
> I have two graphics accelerators installed. The "main" one is a typical
> AGP 2D/3D card like the TNT2 or a Diamond FireGL 1. It can do
> 1280x1024x32 with 32-bit Z and 8-bit stencil. My second head is attached
> to a PCI-based board that can do 1600x1200x64 (16-bits per component)
> with a 21-bit log Z and 8-bit stencil, plus it can do multitexture and
> some other extensions.
>
> My application is smart enough to put the typical controls on the
> "typical" card, and the fancy super-high-res deep-pixel image processing
> windows on the "fancy" card. Naturally I'm interested in whatever
> extensions are present in the fancy card such as MultiTexture and the
> future SeparateAlphaBlendMode extension.
>
> Since I have these different windows, I would expect to be able to get
> different features/extensions from the different contexts. What would
> happen when I query the extensions with context-independent function
> pointers? Assuming I get a pointer, what happens when I call it?
I would envision that if you call a pointer to an unsupported extension,
nothing would happen (i.e. the same thing that happens if you call a gl*
function with no currently bound context). I think it would be a Bad
Thing to simply call the behavior undefined.
Cheers!
--
Brett Johnson <[EMAIL PROTECTED]>
Workstation Systems Lab
Hewlett-Packard Company
"Politicians, like diapers, should be changed regularly,
and for the same reason."
