james tittle wrote:
On Oct 5, 2005, at 10:53 AM, IOhannes m zmoelnig wrote:
does anybody know, why apple's openGL headerfiles behave like that ? why don't they follow SGI's reference implementation ? can this be thought of as a "bug" ?
...does SGI's reference implementation suggest that everything must exist as a lib?
no, i was unclear here. it has nothing to do withthe actual implementation. the question was rather, why the header-structure is different from SGI's "reference headers" (if they can be called such)
I'd view this simply as a non-portable missuse of autoconf...
it is not really an autoconf problem; everything that uses headers (that are different in structure across various platforms) ought to have the same problems.
...but I'd also like to suggest (again, first posted in april) that we sidestep this whole morass and adopt glew (or something similar, tho I've happily used glew on many other projects), and move all this silliness of compile-time extension checking to runtime checking...
...it's just a header and a coupla c files, that can either be compiled into a lib or just absorbed into gem like we've done with other necessary libs...the only problem is that this is a big task, and probably affects every file in GEM...other than that, I would've done it already ;-)
whatcha think?
i think it is a good idea (have to check with w32 first)
fmg.asdr. IOhannes