Re: Solving the linux-2.6 firmware issue
On Thu, Jan 18, 2007 at 08:16:50PM +0100, Andreas Barth wrote:
> * Bastian Blank (firstname.lastname@example.org) [070116 15:06]:
> > On Sun, Jan 14, 2007 at 06:05:44PM +0100, Andreas Barth wrote:
> > > * Bastian Blank (email@example.com) [070114 15:21]:
> > > > This fix and the prefered patch (attached) needs testing to make sure
> > > > the following things works fine:
> > > > - linux-patch-debian-*/linux-tree-*
> > > > - linux-modules-*
> > > What do you mean with this? The second I guess that linux-modules-*
> > > still build correctly?
> > Exact.
> I compiled the "regular" linux-2.6-package, and got the diff below. I
> think based on that we can commit your latest patch now, and sort any
> reminders out later on. (I used + instead of . as version number - but
> that shouldn't make a difference anyways.)
> (In other words, I guess we are safe now with an upload, but of course
> need to continue checking what happens.)
I understood that what Bastian wanted was a regression test for the modules
packages and the linux-tree-* virtual packages, the latter of which relies
heavily on the patch handling in linux-2.6. I'm not sure how important the
latter is; I've previously heard concerns expressed within the kernel team
that extracting the right patch level from linux-tree-* was already broken,
but if it works today we wouldn't want to break it as it has possible GPL
implications (i.e., if the linux-2.6 package revs a patchlevel, is it still
possible to extract the version of the source that was used to build the
kernels used by d-i?).
Since this is the sort of breakage that wouldn't be picked up in user
testing post-upload, I'm going to see what I can do about regression-testing
it before committing the patch to svn.
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.