Jon Leech wrote:
>
> On Tue, Mar 21, 2000 at 08:32:01AM -0500, John R. Dennis wrote:
> > > I don't really understand why you object so strongly to GL_HAS_GLEXT.
>
> I don't object to the symbol. I object to mandating that it be in
> gl.h. Apps will have to generate the symbol in their Makefiles for a
> non-oglbase environment anyway, and when in an oglbase environment the
> symbol is pointless because the very fact of being in that environment
> guarantees that the header exists.
>
> > I'd like to agree with Mark and Brian, I think this is a simple solution to a
> > real problem with very few downsides.
> >
> > I'd also like to mention that when the ARB designed the extension mechanism for
> > OpenGL avoiding compile time problems was a major goal. To meet that goal we
> > mandated that every extension had to define an "Extension Is Present" symbol so
> > source code could test for the existence of the symbol and avoid any compile
> > time problems.
>
> Unfortunately, that no longer obtains. The preprocessor symbols
> cannot tell you whether or not the symbol can be statically linked
> against at link time, or must be obtained via wglGetProcAddress /
> glXGetProcAddress at run time. All they're good for now is telling you
> whether or not the interfaces are defined at compile time.
While function entrypoints shouldn't be directly used, the compile-time
"extension is present" symbol is useful for testing for extension tokens.
-Brian
