Keith Harrison wrote:
>
> ----- Original Message -----
> From: Brian Paul <[EMAIL PROTECTED]>
> To: mesa-dev <[EMAIL PROTECTED]>
> Sent: Monday, June 21, 1999 3:16 PM
> Subject: Re: [Mesa-dev] Changes on experimental branch
>
> > Keith Whitwell wrote:
> > >
> > > Hi
> > >
> > > I've made a couple of changes on the experimental branch. The main one
> > > is the removal of SGIS multitexture.
> >
> > Did that clean things up much? Since the experimental branch probably
> > will not go into 3.1-final, perhaps SGIS multitexture should be removed
> > from the main branch too. I'd like to retire this extension right away
> > unless somebody has good reason not to.
>
> It's interesting to note that nVidia still expose SGIS_multitexture - in
> addition to ARB_multitexture - in their latest TNT drivers, despite Michael
> Gold (of nVidia) stating many months ago that they were going to remove it
> once ARB_multitexture was ratified.
>
> This is not the official word, but the reason is that multitexture can be a
> big win on TNT hardware, therefore removing SGIS_multitexture will hurt the
> performance of at least one well-known and widespread OpenGL game - one that
> does not support ARB_multitexture and (almost certainly) never will.
>
> The removal of SGIS_multitexture is IMHO premature, even though I too would
> like "to see the back of it".
>
Well - there has been a very long-standing bug in our support of SGIS
multitexture, at least as far as q2 is concerned. The enum values we
they use to specify texture sets are off by one from the ones we
expect.
Additionally Q2 doesn't seem to get a big perf. boost (on Mesa) from
multitexture - I'm not sure why this is - maybe just my testing.
Anyhow, nobody seems to have complained, and I'm not sure how useful it
is to q2, so I have no real qualms about moving on.
Keith
_______________________________________________
Mesa-dev maillist - [EMAIL PROTECTED]
http://lists.mesa3d.org/mailman/listinfo/mesa-dev
