Stephen J Baker wrote:
>
> On Wed, 15 Sep 1999, Brian Paul wrote:
>
> > > One place where this would be of immediate and enormous use
> > > would be in detecting the XMesaSetFXmode call...but that doesn't
> > > obviously fall into any of the scope of this stuff.
> > >
> > > Could this call now be considered a GLX extension so I can
> > > test for it sensibly?
> > >
> > > There is currently no way to check for the existance of this -
> > > even at compile time!
> >
> > The XMesaSetFXMode() function was added in Mesa 2.6. You can
> > test the XMESA_MAJOR_VERSION and XMESA_MINOR_VERSION preprocessor
> > symbols at compile time to see if you have a version >= 2.6.
>
> OK - but those are on for NON-3Dfx Mesa too - so when the
> target hardware is (say) an nVidia board, the function gets compiled
> into my code anyway.
>
> The GL/gl.h file is identical on 3Dfx and nVidia equipped machines.
>
> > Similarly, at runtime you can call XMesaGetString(XMESA_VERSION)
> > to get the runtime version number.
>
> ...and I never get to runtime because the function isn't in
> the library for nVidia/Mesa and the runtime linker barfs.
>
> I can't even search for the function at runtime using the
> shiney new gl{u,X,}GetProcAddressEXT stuff because it's
> not a glu/glX/gl function - and even if it were, I'd
> still be required to search glGetString(GL_EXTENSIONS)
> to see if it was valid.
>
> > I suppose an XMesa extension would have been better.
>
> A GLX extension would make most sense I think.
Ah, I guess I missed your point the first time.
I'll look into this.
-Brian
