Re: Fwd: Re: GLsizei type

Am Donnerstag, 28. Juli 2005 21:38 schrieb James Tomaschke:
>  From my experience of using GL I can say that the GLsizei has never
> been an issue for me.  I have seen more problems with people using ints
> instead of intptr_t for address values which caused portability issues.[...]

... and OpenGL added GLsizeiptr and GLintptr for just these purposes. Both map 
to ptrdiff_t on the systems I know of, BTW. If size really matters, we should 
probably add something similar to OpenAL, too.

> If I could rephrase what the GL spec is doing, GLsizei is intended to
> transport positive integers while keeping it's type signed.  The
> Viewport example shows this. So even though the compiler will allow
> negative values the API will reject it though internal <0 tests.  I'm
> not sure not all implementations do this though.

I think they always *have* to do this, because the OpenGL spec is quite clear 
about this issue (sec. 2.5) and I'll bet the conformance test checks this:

   "If a negative number is provided where an argument of type sizei is
   specified, the error INVALID VALUE is generated."

Again, I propose to follow OpenGL here and add this as a clarification to our 
OpenAL spec.

Cheers,
   S.