Migrating D-Bus to GitLab

[Simon McVittie]

On Tue, 14 Aug 2018 at 11:13:43 +0100, Daniel Stone wrote:
> In the short term, we would like to move your Git repository. All this
> means is that you will have to activate your gitlab.freedesktop.org
> account (separate to the old SSH account), and that you will only be
> able to push to GitLab from now on; anongit and cgit will still
> function as read-only mirrors so your users do not need to change
> anything.

For dbus-glib and dbus-python I've been the only committer for some
time. Please go ahead with migrating git hosting for these, and it's
fine to migrate bugs from Bugzilla into Gitlab issues too if you have
a mechanism to do so (they would make good test projects).

For dbus I think the Gitlab migration for git hosting (but not bugs yet!)
is also fine. Ideally Ralf and Thiago should also reclaim their accounts
first, but I'm either a reviewer or a committer for most bugs, so I'm
already on the critical path there.

dbus-doc, dbus-mono, dbus-qt, dbus-qt3 seem to be dead, so I think they
should be migrated but immediately archived. dbus-doc has had one commit,
ever; dbus-mono has not had any activity in a decade and has been
superseded by <https://github.com/mono/dbus-sharp>; and dbus-qt{,3} have
been superseded by the QtDBus module inside Qt itself.

dbus-java and dbus-test are Matthew Johnson's projects. I think it would
be OK to migrate them anyway, if Matthew does not object soon; it's not
as if they've had any commits this decade.

We also seem to have bug tracking (but no source code,
and no open bugs) for the Perl binding, Net::DBus. That can
probably be closed? It also has alternative bug tracking at
https://rt.cpan.org/Public/Dist/Display.html?Name=Net-DBus
(with open bugs) and source code hosting at
<https://gitorious.org/net-dbus/net-dbus>.

> Longer term, we are also hoping to move away from Bugzilla.

For dbus, I would prefer not to rush into this; we have a lot of bugs
tracked, and we have used the embargoed bugs feature quite a lot in
the past, so I would need to be sure that this still works.

For dbus-python and dbus-glib I think we can just go ahead though,
and that would make a good test for this process.

Thanks,
    smcv