Re: kernel-package for t-p-u
On Wed, Sep 28, 2005 at 08:31:12PM -0600, dann frazier wrote:
> On Thu, 2005-09-29 at 04:15 +0200, Sven Luther wrote:
> > On Wed, Sep 28, 2005 at 01:49:49PM -0600, dann frazier wrote:
> > > On Wed, 2005-09-28 at 21:21 +0200, Sven Luther wrote:
> > > > On Wed, Sep 28, 2005 at 08:32:17AM -0500, Manoj Srivastava wrote:
> > > > > Hi,
> > > > >
> > > > > So we are taking the version in Sarge, and just making a
> > > > > couple of changes to it an uploading? Whynot just take the very
> > > > > latest version of kernel-package and put that in t-p-u?
> > > >
> > > > Because we are aiming volatile, which should build with sarge alone, so we are
> > > > planing on uploading to s-p-u and not t-p-u, and have it go into a point
> > > > release.
> > >
> > > I'm confused about what this thread is all about. The subject seems to
> > > suggest that we're wanting to push something into testing, but I get the
> > > feeling the intent is to propose 2.6.12 be included in a point release.
> > >
> > > I like the idea of getting linux-2.6 into volatile, with all necessary
> > > build-deps, but I'd be surprised if the stable release manager was
> > > willing to consider upgrading the kernel in a point release. If that is
> > > indeed the goal, has anyone talked to Joey about it?
> > The point is to either :
> > 1) upload a patched kernel-package to s-p-u, in order build linux-2.6 out of
> > sarge+s-p-u, and thus allow for a linux-2.6 upload to volatile with the
> > current rules.
> > 2) bend the volatile rules, and allow building linux-2.6 with a
> > kernel-package also uploaded to volatile.
> > 3) forget about volatile, and us the backport.org infrastructure instead.
> > 4) upload dilinger's packages to volatile, and disable powerpc support in
> > those, and provide powerpc kernels from another source. We will not be able
> > to provide powerpc64 packages anyway in any of the above solutions.
> What about:
> 5) Upload kernel-package backport and linux-2.6 backport to volatile
> Is this somehow not permitted by volatile policy?
The crux of the volatile problem is.
a) volatile packages must build on an umodified Sarge
b) linux-2.6 requires an updated kernel-package, and
Sven advises that without this powerpc kernels can't be built.
There is also another problem, which affects all architectures,
but Dilinger has a solution to that issue that doesn't require