Migrating Wayland & Weston to GitLab
daniel at fooishbar.org
Tue May 29 10:43:43 UTC 2018
Thanks for the reply!
On 29 May 2018 at 11:20, Pekka Paalanen <ppaalanen at gmail.com> wrote:
> On Tue, 29 May 2018 10:59:52 +0100
> Daniel Stone <daniel at fooishbar.org> wrote:
>> But even once those are done, we need to clean up the bugs we're
>> importing. The plan is to only import open bugs: closed bugs will stay
>> in Bugzilla/Phabricator forever as a read-only archive, and GitLab
>> will only have new active issues. I think a sensible transition plan
>> would be for us to aim to do the import at the end of June, which
>> means sweeping through all our open bugs before then, closing them if
>> they're no longer useful or just cleaning up titles/etc to be helpful
>> in future.
> The sweep is to reduce the work you need to do in the migration, right?
> Aside from the obvious maintenance work we should have been doing all
> these years but didn't.
Partly. The migration does involve some manual legwork: I do an import
into a local test instance first, check it to make sure the bugs
haven't been completely mangled, and only then do the real import.
There are also some random issues with Bugzilla's export API affecting
some but not all bugs; fewer bugs, fewer chances to hit those.
Probably most important is the spam issue: the bugs get closed on
Bugzilla with a link to the newly-opened GitLab issue. Sending people
a Bugzilla mail saying their bug has been moved, followed later by a
GitLab mail saying it's been closed after triage, isn't going to make
us popular at all.
But yeah, it's also a good moment for us to clear out and actually
have a usable issue tracker again. :) Trying to figure out a good
workflow with things like labels, milestones, todos, assignments, etc,
is far easier when we're starting with a clean(ish) slate of (mostly)
relevant issues, rather than five years of irrelevant garbage.
More information about the wayland-devel