Migrating Wayland & Weston to GitLab

[Daniel Stone]

Hi Erik,

On 31 May 2018 at 09:36, Erik De Rijcke <derijcke.erik at gmail.com> wrote:
> First of all I'd like to say that the move to Gitlab makes me really happy
> \o/! It will definitely  lower the contribution barrier for a lot of people
> (including me!) as things are now far more accessible, visible and overall
> easier to manage.
>
> Which brings me to a remark/question on how merge requests are done. Is
> there any plan to allow/move merge requests to Gitlab? Having a central
> point of all things code related would really make things clear and visible,
> and overall easier to contribute.
> Being able to utilize Gitlab, manage your account, create your own fork and
> then having to do a git-send-email would really defeat the point of the
> whole move to Gitlab imho.
>
> Anyway, just my 2 cents. Very glad to see this Gitlab thing moving forward!

Personally, I'd like to use MRs for at least Weston development. I'm
much happier reviewing them there than mail, and although the workflow
isn't perfect, mail certainly isn't either.

Some other people said they preferred a mail workflow for
wayland-protocols. That does make a little more sense to me, though if
Weston moves to GitLab, then it would make wayland-protocols the odd
one out for protocol development: AGL using Gerrit review,
Enlightenment/EFL using Phabricator review, GENIVI using Gerrit
review, Mutter/GTK+ using GNOME GitLab MRs, KDE using Phabricator
review, Qt using Gerrit review, Tizen using a mix of Gerrit and
Phabricator, Weston using fd.o GitLab MRs, and wlc/wlroots using
GitHub review. But our volume of protocol review is small enough that
it's probably not a massive deal.

Similarly, I have a preference for using MRs for the core Wayland
repo, but again we don't have a super high volume of patches right
now.

Using MRs would also allow us to hook up CI pipelines so we could get
fast feedback on whether the basic build and checks succeeded, which I
think is pretty helpful given the number of times we've broken
distcheck lately.

What do others think?

Cheers,
Daniel