[Desktop12.04-Topic] GNOME Version for the LTS

[Sebastien Bacher]

Le mardi 04 octobre 2011 à 09:53 +0200, Sebastien Bacher a écrit :
> Hey, 
> 
> So first easy one, discuss what GNOME version we will track for the
> LTS.
> It's basically a discussion we have at every UDS ;-) 

Hey again,

So when I wrote "easy" it was rather "obvious topic for UDS", the choice
is not that easy and I would like everybody to think on going for GNOME
3.2 or 3.4 will impact on what they are doing, keeping in mind that our
first goal for the LTS is quality.

Let me bootstrap the discussion and share some of the concerns I have:

- glib is going though quite some refactoring [1], it should stay
compatible but it change how glib work and could have bug, that seems
risky for the lts

- while we usually keep up with GNOME updates they take some time and
the .0 quality is never "great", .1 tend to be better and the schedule
should allow us to get .1 in for the LTS but still, time spent on
tracking features is not spent on fixing bugs

- new glib and gtk will deprecate apis as usual, that's not an issue but
lead to work on the archive to fix that stop building due to that

- seems like udisk2 (udisk rewrite) should land in GNOME 3.4, having a
rewrite landing in the lts cycle is risky, gnome-disk-utility, gvfs and
some other bits in GNOME are being ported to it and will not keep an
udisk1 backend so 3.4 will likely force that choice on us


Now there are good reasons for going for GNOME 3.4

- GNOME 3.2 is still early in the 3 serie and still has issues, they
will keep improving this cycle and some of the improvements would be
good to have

- we like to have at least an uptodate platform for a lts (i.e glib,
gtk), out of the fact that glib itself is the issue this time

- some users will want GNOME 3.4


Some of my thoughs on the challenges:

- glib 2.31 will likely get support for their improvement "application
menu", which means it's likely most applications will depend on the new
glib, which is going to make hard to update some components only

- if we don't go for GNOME 3.4 it's likely the new version will be
packaged in a ppa as GNOME3 was for natty, that's ok but if we do it
because we think glib is going to create issues it's really suboptimal
to let a ppa ship a new version of glib

- some of the goals for next cycle are to improve the control center, we
will really want most of g-s-d and g-c-c 3.4 I think, we could probably
play some backporting game there or get 3.4 to work on GNOME 3.2...


Thanks for reading and please share you though on the topic, we can
start the discussion on the list to include extra people and collect
extra arguments before the UDS session

Cheers,

Sebastien Bacher

[1] http://mail.gnome.org/archives/gtk-devel-list/2011-September "GLib
"next cycle" update"