From: Brian P. <br...@tu...> - 2003-06-17 22:16:54
|
Steve Baker wrote: > Fay John F Contr AAC/WMG wrote: > >> (1) Everybody gets his changes into CVS >> (2) Everybody checks out the new CVS version and runs all his >> pet programs using it >> (3) Somebody creates a release tarball > > > A "release candidate" to use the trendy jargon. > >> (4) Everybody checks out the release tarball and makes sure it >> runs properly on his system >> (5) Steve Baker announces a new release. > > > Hmmm - I feel that a fanfare should be item 4.5 :-) > >> While I have been in open source projects for a while, I have never >> overseen this sort of thing before. I am certainly open to hints, >> suggestions, or even an outright transfer of control to somebody >> else. I just don't want to see the ball dropped. > > > How do you feel about sticking in the glutGetProcAddress thing? > > It really just needs to be a wrapper around glXGetProcAddress and > wglGetProcAddress, > with a test to see if the first four letters of the string are "glut" - > and a > hash-lookup into a table of all of the freeglut functions. I didn't bother with a hash table. It's just a loop over an array of all the GLUT function names. glutGetProcAddress() doesn't have to be high-performance. > The semantics are ever so slightly messy because glXGetProcAddress > returns a result that's specific to the current rendering context and > wglGetProcAddress returns the same result for all rendering contexts. Actually, it's just the opposite way; GLX's pointers are context independent. > We can solve that by simply saying that freeglut works the X way, so > you have to call it in each rendering context...this will be a little > redundant in Windows - but it'll at least be consistent. We'll say it works the WGL way; call glutGetProcAddress for each window/context to be safe. -Brian |