Re: Scheduling 2.6.17-1

[Sven Luther <sven.luther@wanadoo.fr>]

On Sun, Jun 18, 2006 at 11:39:16PM +0200, Frederik Schueler wrote:
> Hello,
> 
> On Sun, Jun 18, 2006 at 11:29:32AM -0600, dann frazier wrote:
> > Is the plan to do a "linux-2.6.16" upload at the same time, or does
> > this imply a decision to use >= 2.6.17 in etch?
> 
> We have not yet decided about what to use for etch. There has not been
> any more commitment towards a long-term support for 2.6.16 on LKML, but
> maybe the last word is not spoken here yet. Anyone who knows more about
> it?
> 
> Here two proposals on how to go on, satisfying the need of a stable
> 2.6.16 for testing and the need of 2.6.17 for the architectures broken
> in 2.6.16 (at least sparc, mipsel):
> 
> 1. we upload 2 source packages: linux-2.6.16 version 2.6.16-15 and 
> linux-2.6 version 2.6.17-1. This will allow us to upload 2.6.17 ASAP,
> and to continue 2.6.16 support until the etch release kernel decision is
> made.
> 
> 2. we upload linux-2.6 2.6.16-15 on Monday with urgency=high and have it
> added to testing ASAP, and upload linux-2.6.17-1 on Friday (5 days
> later), probably with XEN images in place, and one round of NEW less. 
> With this option, taking care of 2.6.16 in testing could be done through
> t-p-u uploads.
> 
> 
> I personally would like to follow option #1: We could backport
> sparc/niagara and smp-alt support if 2.6.16 gets long-term support, we
> would have a stable kernel for etch until 2.6.17 stabilizes, and we had
> enough time to decide which kernel to release with.

What about :

  We upload linux-2.6 2.6.17 ASAP, and if 2.6.16 is finally the way to go, we
  upload linux-2.6.16 or linux-2.6.etch, and propagate this one to testing.

Friendly,

Sven Luther