OK, now that I have my shiny new NVidia SDK installed, I'm working on making my standard old OpenGL engine do some new tricks.
I'm working on making some of the simple SDK demos build without all of the SDK helper utilities such as glh etc. Once I get this far, I will build some extended graphics utilities in my engine. But, I have found that this is a bit of a pain.
I have distilled one of the demos down to needing only nvparse, which will not compile easily in a striped down environment.
I would like to know if I could also remove nvparse also in my distiling process. ?
Just what exactly does nvparse do?
nvparse does several things, such as taking strings like this:
discard = expand(tex0) . expand(tex1);
discard = expand(tex2);
spare0 = sum();
out.rgb = spare0.rgb;
out.a = 1.0;
and translates them into the equivalent register combiner state settings. It also handles setting NV_texture_shader state, can convert DX8 pixel shaders to register combiner+texture shader settings, can pinpoint load problems with NV_vertex_program, etc...
If the demo uses texture shader or register combiner scripts through nvparse, then you'll have to manually translate them to equivalent OpenGL calls if you want to remove nvparse.
[This message has been edited by jra101 (edited 01-16-2003).]
Looking further into this it seems to be a lot of work to do what nvparse is doing (+ cg compiler). It's much better to just go with the flow
I've now managed to compile in the NV SDK libs just like in the demos -- it takes a little hacking, which I don't like but hey! Now I can get on with playing with the new features...