William F Mitchell <[log in to unmask]> writes:
> I guess I'll have to come up with another way of handling
> GLUNULLFUNC. Sigh.
I have to shoot from the hip, because I haven't looked at in in detail
yet, but why don't you just make the NO_NULLFUNC case the default?
This violates the `1-to-1 translation from the C API' paradigm, but
isn't it as intuitive to give no argument at all, as is it is to give
NULL? Currently one has to change the application when moving from on
system to the other :-(.
BTW: as far as I can seen, you export the specific OpenGL procedures
(e.g. glVertex3d, glVertex3dv, glVertex3f, glVertex3fv, etc), but not
a generic procedure (glVertex in that case). Wouldn't that be the
natural thing to do? This can be hairy, because you have to know at
compile time which of you gl* kinds will be different. But it would
make the programmer's job much easier, because he/she doesn't have to
append the explicit kinds to each literal constant. Or are you afraid
that people will write
glVertex (1, 1, 1)
when they mean
glVertex (1.0, 1.0, 1.0)
which will give very different results, of course ...
Cheers,
-Thorsten
PS:
> And from Malcolm Cohen we have:
>
> > (1) procedureness is not a disambiguating characteristic for a
> > generic [though on thinking about it, maybe it ought to be allowed]
>
> That sounds like a good idea.
Yes! First class functions in Fortran: first implementation shipping
in 2021 ...
--
Thorsten Ohl, Physics Department, TU Darmstadt -- [log in to unmask]
http://crunch.ikp.physik.tu-darmstadt.de/~ohl/ [<=== PGP public key here]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|