Hi all,

On Tue, 21 Jan 2003 18:37:29 -0500, I sent the message below.  I didn't
get any feedback whatsoever, so I over the next couple of days I'll go
over the diff and start committing these changes.  Unless, of course,
somebody foresees any problems with this...

-Jan



Jan Schaumann <jschauma@netmeister.org> wrote:
> Hi,
> 
> As discussed previously, I've been working on the Mesa update to 5.0.
> 
> This entailed:
> 
> o  creating a clean chroot for a clean pkgsrc tree
> o  apply patches from PR pkg/19302 to update Mesa and friends to 5.0
> o  move Mesa,glu,glut to LOCALBASE, so they can be installed regardless
>    of XF version
> o  bump PKGREVISION on all packages including Mesa*,glu,glut buildlinks
> o  only include Mesa/buildlink2.mk if the package requires all three of
>    MesaLib, glu and glut - else only include the necessary buildlink2s.
> o  verify they compile/run with both Mesa version 3.4.2 and 5.0
> o  verify that they only depend on Mesa-5.0 et al if MESA_REQD=5.0
> 
> The patch at
> http://guinness.cs.stevens-tech.edu/~jschauma/netbsd/mesa.diff is the
> result.
> 
> Comments:
> All packages build and function fine with either version.  The GL, GLU
> and GLUT APIs seem to be fully backwards compatible, which is good news.
> 
> I was unable to test emulators/rewind -- I don't have any windows apps.
> I was unable to test emulators/wine -- I don't have any windows apps.
> I was unable to test games/quake -- I don't have the quake maps.
> I was unable to test games/quake6 -- I don't have the quake apps.
> They all compile fine, though, with either version.
> 
> misc/molden compiles fine, but I can't test it, due to the license.
> 
> graphics/gtkglext doesn't compile with either version currently.
> 
> 
> The Mesa*,glu buildlinks should check for MESA_REQD >= 3.4.2.  I was
> unable to determine an efficient way to match the version numbers, so I
> just match for 5.x.
> 
> I'd like to have as many people as possible review and test this patch.
> All suggestions, corrections and improvements are appreciated.
> 
> Thanks,
> -Jan
> 
> -- 
> I always said there was something fundamentally wrong with the universe.

-- 
A common mistake that people make when trying to design something completely
foolproof is to underestimate the ingenuity of complete fools.