Re: The solution to Mach64 problems on Lombard
On Fri, Feb 09, 2001 at 11:37:07AM +0100, Bastien Nocera wrote:
> Quoting Sven LUTHER <email@example.com>:
> > On Fri, Feb 09, 2001 at 11:20:46AM +0100, Michel Dänzer wrote:
> > > Sergio Brandano wrote:
> > > >
> > > > I am posting the solution to the above problem, in the hope that
> > the
> > > > package maintainers will actually see this mail and will update the
> > binary
> > > > trees overnight.
> > >
> > > Dream on. It's not Branden's fault that Ani can't seem to get the
> > > modifications in his tree integrated into XFree86. Dan even tried to
> > integrate
> > > them into the Debian packages a few months ago but gave up as they
> > broke with
> > > every new version. It's a pain that nobody can be expected to bear.
> > The best bet is to have it integrated into the Xfree86 xf-4_0_2-branch
> > if
> > possible, so Branden will have no choice but to accept them when the
> > 4.0.3
> > gets released.
> It's always the same kind of problem, be it with the kernel or XFree86 or any
> other big project of this kind.
> If changes don't make it into the main tree, you end up chasing the right tree
> with diffs of a couple of megs to the main tree and life is a nightmare for
> users and package maintainers.
> I wish I could use ac patches to test out the new functionalities, or the ati
> tree from linuxvideo.org, but I can't... (well, I can, but not without spending
> time hacking the sources to make it even compile)
I understand you, but anyway, ...
The problem is that branden is maybe not having the time to check it and
incorporate it, or maybe he simply don't care, don't know.
That said, there will now be stable Xfree releases and developpment Xfree
I think branden will no more track the developpment Xfree releases, but the
stable + bugfix releases, which is where your fixes should go.
Also if you want an historical view of the situation, in the 3.3.x days, there
were lot of patches against the debian package, but not all of them were
included in upstream, which made them not very happy about it, and also made
live thougher for the maintainer, as he had to recheck every patch at each new
The idea is that now, the patches should migrate into upstream more quickly,
and that upstream will release stable releases more often. (well at least i
hope this is the idea, it somewhat makes sense, but maybe i am wrong anddid
In the long run, this will be better for now, and if you have a patch, it
should go to upstream first, since this is the main repository,
debian'spatches being just a fork.
And anyway, personnaly, i had more luck going for upstream than sending stuff
to branden (who don't likes me :))) and look at the loads of bug reports in