Re: Advice for Nautilus to gtk4

[Carlos Soriano <csoriano gnome org>]

Thanks Iain!

I think the status re gtk4 is that core apps that are close to its development are good to move, so API changes can be followed up closely. Not many projects have the resources to be the initial ones helping out gtk4, the core GNOME are the most likely the best positioned and Nautilus probably the most likely of those. But you are right that the balance between making things stable but helping out pushing the platform forward is being hard to get it right, so any testing/comments/feedback around this is appreciated.

but I am actually looking at upgrading to 3.30 at the minute, and possibly to 3.32 in turn

 

It would be nice to have an updated Nautilus in distributions, I think Nautilus 3.30 would be a fine option if a distribution lacks gtk4 and we've decided 3.32 with gtk4 is good enough for distributions. The key here for me is where the line is in "it's good enough" from an upstream quality perspective.

On Wed, 14 Nov 2018 at 16:59, Iain Lane <iain orangesquash org uk> wrote:

On Mon, Nov 12, 2018 at 01:28:16PM +0100, Carlos Soriano wrote:
> Hello all,
>
> As part of the big work of moving Nautilus to gtk4

This is interesting. I hadn't been following nautilus upstream
development very closely, but I thought the general feeling for gtk4 was
that it was OK to try to move things over but only as an experiment, not
to actually make releases that depend on it. At least for GNOME core
stuff, while the API/ABI is still in flux. Do I misread/misremember the
situation?

> I wonder what's the best approach for distributors:
> - Merge gtk4 branch now, and the contingency plan is to warn here and
> recommend to not ship Nautilus 3.32 closer to release in very stable
> distributions in case things cannot be worked out. Fedora Rawhide, Arch,
> Tumbleweed, etc. can still benefit us from testing.
> - Branch 3.32 already, merge gtk4 branch in master and have 1'5 releases of
> time to work things out.

The above notwithstanding, here's my personal opinion from the Ubuntu
perspective. I think that if Nautilus (or anything probably) were to
depend on GTK4 then we'd probably have to hold off taking the upgrade
for 19.04. We haven't started looking at things like themes or any other
required integration yet (e.g. do we need to port any of our distro
patches from gtk3 over? [hopefully not!]) and I don't think we'd be able
to take that on this cycle.

I guess that us potentially being behind by a release wouldn't be the
end of the world (clearly, since we already are), just wanted to give
you the situation.

(Partially unrelated, but a reason why this opinion is not completely
irrelevant: I know we're way behind on 3.26 right now - but I am
actually looking at upgrading to 3.30 at the minute, and possibly to
3.32 in turn. After the actual distro patch rebase itself is done the
next step will be to check out the desktop icon extension and possibly
contribute there if anything's needed.)

Cheers,

--
Iain Lane                                  [ iain orangesquash org uk ]
Debian Developer                                   [ laney debian org ]
Ubuntu Developer                                   [ laney ubuntu com ]