[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-scm-interest
Subject:    Re: [Kde-scm-interest] KDE Git hosting status update
From:       Eike Hein <hein () kde ! org>
Date:       2010-06-10 16:17:03
Message-ID: 4C110FFF.4010307 () kde ! org
[Download RAW message or body]

On 06/10/2010 09:44 AM, Esben Mose Hansen wrote:
> On Thursday 10 June 2010 02:26:06 Eike Hein wrote:
>> We're sharing it with you now - see the attachments to this mail -
>> and hope you will agree that we have found a solid way forward for
>> the KDE Git migration efforts.
> 
> Thank you. The attached document shows an impressive effort.
> 
> One worry: The document repeatedly says "Gitolite developer". As in singular. 
> Doesn't that make us more than a little vulnerable in case of this developer 
> taking sick or loosing interest? If we go this route, I think we should at 
> least attempt to take some steps to mitigate this risk.

Good thinking. We feel this is not a big concern because:

* gitolite has an significant number of serious production de-
  ployments and a pretty active community right now. This is
  not an obscure piece of software, you can find it documented
  in Git literature at this point:

  http://progit.org/book/ch4-8.html

  We feel that if the main developer does get hit by a bus, we
  won't be left alone and a new upstream should arise. And
  socially-speaking gitolite is used by the sort of people who
  get open source, so hopefully it would be a decent upstream.

* The stack we've decided to run with is a lot more modular
  than the rather monolithic Gitorious software. If I remem-
  ber my rough numbers right, Gitorious is a 250000 line Ruby
  app. gitolite is a 1500 line Perl script (don't worry, it's
  readable Perl). If we absolutely had to maintain it oursel-
  ves, we feel a lot better about being able to do that than
  Gitorious just in terms of scale and scope.

* Should things with gitolite ever go seriously awry after all,
  there are conceptually similar tools like gitosis that would
  hopefully pick up the slack and could be integrated as re-
  placements. (We've chosen gitolite because it is currently
  more powerful than gitosis, however.)


-- 
Best regards,
Eike Hein
_______________________________________________
Kde-scm-interest mailing list
Kde-scm-interest@kde.org
https://mail.kde.org/mailman/listinfo/kde-scm-interest
[prev in list] [next in list] [prev in thread] [next in thread]