On Wed, Sep 15, 1999 at 09:44:00AM -0500, Stephen J Baker wrote:
> On Sat, 11 Sep 1999, Bernd Kreimeier wrote:
> > There has to be a good technical reason then why
> > glxGetProcAddress isn't sufficient?
>
> Yes, two reasons I think:
>
> 1) A program may choose not to link to glu. Hence
> glxGetProcAddress cannot tell the address of
> a glu function by having code like:
>
> if ( strcmp ( s, "gluEndCurve" ) ) return gluEndCurve ;
>
> ...so glu functions have to be interrogated by gluGetFuncAddressEXT
I question whether enough people care about extending GLU on a
per-driver basis that we need bother with it at all. Maybe so, of course
I can think of weird cases (driver supports new format/types as an
extension; must enhance gluBuild2DMipmaps() to understand them for
completeness). But hopefully not.
Jon Leech
SGI
