On Thu, 10 Feb 2000, Bernd Kreimeier wrote:
> > I'm not proposing this should be part of the "OpenGL ABI for Linux"
> > standard.
>
> Is there a disadvantage is doing so? I.e. not generic enough?
I think that's the general thought. :)
> > "[Mesa | SI] [DRI | UtahGLX] <card> <version> <card-specific>"
>
> "Mesa" | "SI" would be a special case of <package>?
> <implementation>? Which has to be used in connection
> with GL_VENDOR (Brian's Mesa, a Mesa derivative, SGI's SI,
> an SI derivative from MetroLink, Xi?).
> [ DRI | "UtahGLX" ] - <architecture>? Why is this necessary?
> If there are different builts from the same version of the
> same implementation, then the version string seems inadequate?
> I understand there won't be DRI for VG/V2, but the cards won't
> go away, so we need a string for non-DRI, non-GLX builts of Mesa?
> And one for software-only GLX? And fake-GLX? It seems open
> ended, and I am tempted to defer this into the <version> specifier.
As Brian mentioned this is just what they're planning on doing, and it's
helpful to distinguish both the core codebase and the driver
package/implementation. So Metrolink's SI derivative might be "SI XI"
Re Thomas Roell's complaints about the version string, I find the
datestamp version idea a little limiting as well. We'd like to be able to,
for instance distinguish binaries build for a release tarball from those
build from cvs. see
http://lists.sourceforge.net/pipermail/utah-glx-dev/2000-February/000878.html
for a suggestion about that.
I'm not sure we can come up with something sufficiently broad to make
everyone happy though, so I tend to agree with just making it an informal
suggestion. And for my part, I'm happy to use it for UtahGLX.
-ralph
