Comment 4 for bug 1219486

> 1. GNOME 3.8 has only been tested with gnome-settings-daemon 3.8 and there's no guarantee that things work correctly when we mix and match pieces.

the pieces that are the most matched are gnome-settings-daemon and gnome-control-center, which are both in sync on 3.6 at the moment. Does that mean you plan to FFe gnome-control-center as well (saying the number of changes in g-c-c 3.8 I doubt it's a good idea to do that at the end of the cycle)

> 2. Since we did the gsettings-desktop-schemas update, there are gsettings keys that won't actually do anything without the updated gnome-settings-daemon backend.

Do we have those exposed to users anywhere or is that a dconf editor issue? We might also want to just drop those keys from the schemas if you think that really confuses users

> 3. We need to update this cycle because we need to fix bug 1201679 which is a High priority regression in Unity (GNOME was unaffected).

Seems like that should be a one commit to backport if we need that fixes

> 4. There are numerous other fixes and improvements if you read through the NEWS file. The update isn't just GNOME dropping Fallback support and we've patched back in the fallback pieces that Unity currently needs.

Right, there are lot of changes in there, which makes me cautious, gnome-settings-daemon has little automated testing and it's a somewhat fragile component.

> 5. Tim did a lot of work on this update with the assumption that it could make it into Saucy. We were delayed several weeks by
> indicator-keyboard taking longer to land than expected. I tried to help indicator-keyboard along as well as I could and
> complained that the longer the delay, the more difficult it would be to land gnome-settings-daemon 3.8 by Feature Freeze.

Right, thanks Tim for the work! The work is not wasted in any case, even if that only lands next cycle. Things being less trivial than expected is not a real reason to grant a FFe though, as you pointed we still have fallout from 3.6 ibus migration we didn't sort out, it seems like what we need is extra work on stabilization rather than changing that much late in the cycle

> There was one final regression that Tim fixed over the weekend. Ubuntu GNOME shouldn't be penalized for taking care not to
> cause regressions and waiting for Unity work.

It's not a matter to be "penalized", what is the direct impact at the moment on Ubuntu GNOME (out of not having the latests upstream version)? The argument can go the other way around, Ubuntu GNOME wanted to be on the current version is not a reason to create instability from Unity/Ubuntu...

> I've been running Unity all weekend to make sure that the update didn't cause any new problems. Many people have installed
> the GNOME3 Staging PPA; some of those people use Unity.

That's good to know!

To summarize my opinion: from the past record I would be careful to update gnome-settings-daemon that late in the cycle, that's a component that drives many settings in the desktop and which is not easy to test.

Having gnome-settings-daemon/gnome-control-center out of sync is something we avoided so far, because it has extra potential to create issue, and I'm still unsure that's something we should try...

Anyway, I'm going to let the release team review the request. I've a small request, if the ffe is approved it should be under the condition that those doing the update are going to deal with the bugs and regressions it's going to create (not like previous cycle with nautilus)

Thanks