On Mon, Oct 25, 1999 at 07:57:10AM -0500, Stephen J Baker wrote:
> On Fri, 22 Oct 1999 [EMAIL PROTECTED] wrote:
> > Oh, no! Steve's gone over to the dark side!
>
> I think this is beyond debate now - we are voting.
I think you're wildly optimistic about "voting" == "end of debate".
However, FWIW my preference at this point is for the context-independent
pointers, and moving on to finish this sometime before the millennium.
There are a whole bunch of other issues which need to be resolved. I
propose starting with the library naming conventions and these two points:
(a) the version number of a particular implementation cannot be part
of the mandated name of a library destined for multiple
implementations.
(b) to minimize confusion during the conversion, it seems preferable
to use a name that nobody is using today - someone can always
symlink an existing name to the new chosen name as an interim
measure.
Thus the libGL.1.2.lsb.1 name, albeit ugly, is IMO the way to go.
Stuart, is this in fact the proper name under the LSB conventions? Is
there anything else relevant about LSB library naming/placement?
> OpenGL 1.2 (and later) functions might as well be treated as extensions
> since you can't guarantee that all implementations will have them.
For our purposes, we can guarantee that all the 1.2 entry points
exist statically. I believe we should do so, along with
ARB_multitexture.
> We *MUST* come up with
> a final decision - and it needs to be before Mesa 3.1 hits the
> streets.
Rather than being gated by the release schedule of some particular
implementation, I want to make this into an ARB extension. This means
(a) the extension specification must be complete by November 8th and (b)
it will not be finalized until the ARB gets to vote on it December 7-8.
Jon Leech
SGI
P.S. And if people must say something else about *GetProcAddress in
response to this, please change the subject line back.
