[GHC DevOps Group] The future of Phabricator

[Ben Gamari]

Manuel Chakravarty <manuel.chakravarty at tweag.io> writes:

> Hi Ben,
>
> With my DevOps Group Chair hat on, it seems that GitLab (as proposed
> by you) provides a suitable road forward (both in technical terms and
> in terms of being a compromise that most people are reasonably happy
> with). So, let’s do it!
>
Yes, I agree!

> Could you please make a list of all the tasks that need to be done and
> attach approximate effort/timelines where possible? (Maybe as a wiki
> page?)
>
Yes, I have started a Wiki page here [1].

On the whole, I would like to proceed fairly quickly. For one, CI is
currently unacceptably flaky. Secondly, I would like to be able to
avoid having to start paying for our Rackspace boxes when support ends
at the end of the year.

However, there is then the question of scope. In principle, GitLab can
supplant three distinct services pretty well:

 a. Phabricator for code review
 b. gitolite, for repository hosting
 c. Trac, for issue tracking and the Wiki
 d. What remains of the Harbormaster CI

This discussion so far has largely focused on (a). I discuss the
remaining items in some depth on the wiki page.

Cheers,

- Ben


[1] https://ghc.haskell.org/trac/ghc/wiki/GitLabMigration?action=edit
[2] https://github.com/gitlabhq/gitlabhq/blob/667c0a909bde1cf71f21d8ec9768e98b1c489030/doc/hooks/custom_hooks.md
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 487 bytes
Desc: not available
URL: <http://mail.haskell.org/pipermail/ghc-devops-group/attachments/20181106/b0cceda5/attachment.sig>