Time based releases?

Dan Meltzer parallelgrapefruit at gmail.com
Fri Mar 27 21:21:03 CET 2009


On Fri, Mar 27, 2009 at 4:17 PM, Leonardo Franchi <lfranchi at kde.org> wrote:
> On Friday 27 March 2009 19:58:08 Jeff Mitchell wrote:
>> Mark Kretschmann wrote:
>> > 10:03 < leinir> that whole branch thing - being able to develop new
>> > stuff without trashing the whole app for months at a time :)
>>
>> Another thread we should revive.
>>
>> Everything is basically ready for whatever we decide to do until
>> git.kde.org is ready (which we all know is more of a long-term thing
>> because of scripty headaches and the like).  So until then, if people
>> want to use git, I propose that we either:
>>
>> 1) Have one master git-svn checkout; everyone does their work on git;
>> periodically we push back up (especially with string changes)
>> 2) Keep using SVN for trunk, but have git available for people to
>> individually store feature branches on (similar to how it was working
>> with http on my server before, except properly).
>>
>
> I think it's only worth doing (1) if we do it at all. If we still use git-svn
> then we can't do all the cool things git lets us do like sharing branches and
> the like. So i don't see the point of (2).
>
> I would push for (1) but then we have issues---we need to do 2-way syncing
> (scripty commits etc). Also then do we require new contributors to use git?
> That kind of raises the bar (although if KDE is going to git in the long run,
> we may as well start.

Two way syncing is going to cause issues I think.  I really feel we
should just stick with svn until git.kde.org is set up (and efforts
should be focused on making that a reality, rather than on a temporary
solution).  SVN has worked fine for us for the entire 2.* development
process, and I don't think that having git would have made more
features appear (and definataly would not have made more bugs be
fixed, which is more important at this point...)

Dan,
>
> How exactly would this "master" git-svn server work? Would all the individual
> commit data get retained transparently?
>
> leo
>
>> Regardless of which of these we use, as far as Git itself goes, the
>> following are essentially ready to go, depending only on picking one and
>> getting public SSH keys from people:
>>
>> 1) GitHub -- heard more nays than yeas about this (from those that
>> responded) as it's not OSS.  It's actually not much more than gitosis
>> with a few scripts and a non GitWeb frontend.
>> 2) Gitorious -- not much different than Git + GitWeb, except we don't
>> host it ourselves
>> 3) Hosting Git + GitWeb ourselves with Gitosis (I already have this set
>> up, and would be normal git protocol, not http)
>
> I don't really care, but i think the simplest way is the way to go...
>
> leo
>
> --
> -----
> lfranchi at kde.org                Tufts  University 2010
> leonardo.franchi at tufts.edu                The KDE Project
> _______________________________________________
> Amarok-devel mailing list
> Amarok-devel at kde.org
> https://mail.kde.org/mailman/listinfo/amarok-devel
>


More information about the Amarok-devel mailing list