Comment 78 for bug 219630

On Sat, Jun 28, 2008 at 05:41:06AM -0000, Martin-Éric Racine wrote:
> Guys, please don't do this. This has been abundantly tested enough as it
> is, by myself and others. The symlink causes too many problems. It was
> removed for good reasons. Do NOT re-introduce it!

The problem is that we have a very real risk of regression for users who
have static xorg.conf, and still no clear understanding of what the reasons
are that necessitated the removal of the symlink; i.e., we have the
statement that it broke LTSP, but on the other hand we don't have any logs
that show what broke, we have a statement of the X maintainer that it
shouldn't have broken, and we have Oliver's tests which show it didn't break
for him.

Can we please pinpoint what's different between the systems where it broke
for you, and the system where Oliver found that it worked fine with the
symlink? When it appeared that the symlink broke the use of the driver for
all or most users, I was willing to allow this as an SRU; now it appears
that others who have tested the package can't reproduce the problem, which
both lowers the benefit of the SRU (since it's not broken for everyone) and
increases the severity of the regression (since users who have Driver "amd"
set don't necessarily have unusable systems).

> However, if you REALLY insist on keeping the symbolic link, then at
> least make it purgeable by putting it in the -amd transitional package,
> not in the main -geode package.

Oh yes, that would be fine with me; I don't understand why the compat
symlink wasn't put there in the first place. If you think that's the best
option, I can do an upload of this tomorrow.

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
<email address hidden> <email address hidden>