Gnome plans for the lenny cycle

[Loïc Minier]

        Hi,

 (sorry for not getting back to this earlier)

On Thu, Apr 12, 2007, Gustavo Franco wrote:
> On 4/12/07, Marc 'HE' Brockschmidt <he at debian.org> wrote:
> >We would like to know which major upstream versions of Gnome are
> >expected to be released in the next 24 months and how much time you
> >expect them to need to get stable enough for a Debian stable release.
> 2.22 - 1st half 2008
> 2.24 - 2nd half 2008
> >Our current, very rough plans would mean a release in 18 months with some
> >padding in both directions, which would lead to a lenny release around
> >October 2008. We expect to shuffle this a bit around to fit everyone's
> >needs, so please tell us if this date works for you.
> I would be glad to see Lenny with GNOME 2.22 and released in the 1st
> half of 2008, or postponed for late 2nd half (and not october) with
> 2.24.

 Yes, it's the same feeling here, it's hard to tell with "late 2008"
 whether we will have the time of pushing 2.24 or not.  Nowadays,
 upstream releases of GNOME are timely and quite predictible (who knows
 in 18 months though), but sometimes very buggy, especially the .0
 release.  The .1 release, which happens something like 3 or 4 weeks
 after the .0 release, is way more usable.

 We will have to decide about 1 month before the .0 release, past the
 API freeze, whether we're pushing a particular GNOME version or not.
 If we are, then all SONAME changes etc. could be prepared about two
 weeks before the .0 release of GNOME, and uploaded straight to
 unstable, which should leave us just enough time for all important
 changes to migrate to testing before the GNOME .0 release.


 I have followed other discussions about a proposed Upstream Version
 Freeze, and I must note that I think GNOME modules should be exempted
 from such a freeze (I think they are exempted in Ubuntu BTW), but
 reviewed depending on the amount of code they change, just like these
 were hinted for etch.

> lool, could you please elaborate on what we discussed on IRC some days
> ago about why we shipped 2.14 with just some 2.16 modules in Etch? The
> "december" factor and stuff like that? TIA.

 Sure, in short:
 - we had to decide of 2.14 versus 2.16 in october or so, for a
   scheduled release in december
 - there were plenty of transitions for 2.16:
   * the usual bunch of SONAME changes
   * gtk 2.10, with a new module ABI which means coordination of 20
     uploads by > 15 maintainers with sourceful changes and risks for
     g-i (gtk directfb based)
   * bonobo/gnome-vfs had to be uploaded together and transition
     together IIRC
   * icon-theme
 - gtk 2.10 was very buggy, it has not completely healed, but it's
   better nowadays
 - .0 releases of GNOME are relatively buggy
 - 2.16 was not fully complete in experimental (some minor modules were
   still 2.14 only, some modules were 2.15.9x only)
 - the GNOME team was not particularly rich in manpower

 The biggest "problem" we now have in the face of the world is that the
 release ended up being 8 months after the 2.16 release instead of 2 as
 initially planned.  On the other hand, we had more time to polish our
 desktop until the very last hours before the release.

 I expect less things will be in the bag of issues before the release of
 lenny, and if we keep a stable release date, it should be
 natural to pick the most logical GNOME release which meets our quality
 standards.

   Bye,
-- 
Loïc Minier