On Saturday 17 January 2009, Casey Link <unnamedramb...@gmail.com> wrote about 'Re: [Kde-scm-interest] On Amarok Switching to Git': >Leo and I brought this topic up at a break a few minutes ago during >Camp KDE, and immediately Chani and Thomas jumped in. We started >discussing technical and social hurdles that have to be crossed before >all of KDE can ditch svn for good.
Well, Thiago went into a few of them in Message-id: <200901180017.06567.thi...@kde.org>. First, of all Amarok developers need to purge their tree of "svn:extern"s, handling them is non-trivial in git and I don't think anyone is really interested in that work. Beyond that, I think all the other issues can be resolved "in conference" with the Amarok developers. I know Thiago isn't volunteering, but I *think* there are others on the list that may have time to discuss (and possibly implement) this. I work a normal, US-Central, 40 hour week, so I could only kick in 8-10 hours a week regularly, irregular hours, on the weekend. I think the Amarok team would like administrativa to have a faster turn-around time *AND* I'm not willing to do the work unless I have a couple of people I trust to add more hours. I feel setting up a model KDE git infrastructure is "full-time work" for at least a few months until things get ironed out. >The more I think about it, the more I like the idea of Amarok >switching to git initially before the rest of KDE. There are two >primary advantages to Amarok switching to git: > >1) The selfish reason: Amarok devs can finally share branches, and >fully utilize the power of git. >2) The egalitarian reason: We can blaze the trail for the rest of KDE, >by squashing bugs, annoyances, and generally pioneering the mass >svn->git switch. I see both reasons as valid. However, I thought being "part of KDE" meant being maintained under the same source control. That said, Amarok survived fine being not "part of KDE" in the past and I think it could flourish there again. If I git<->svn bridge needs to be built between a Amarok git tree and a KDE svn tree, that could be a problem. The stock "git svn" doesn't *really* support this. All the history would have to be linearized regularly; even modern subversion doesn't handle branch merges the way git does. This means that all git-using Amarok developers would have to regularly rebase. However, Amarok developers using "git svn" already have to do this a lot; if the majority have already switched to git, this might actually reduce the number of rebases. I have some ideas about building the bridge, but they aren't solidified, and would require the SVN repository to be the "canonical" version. Since client-side SVN doesn't have any method of "rewriting history" it's pretty impossible to "rebase" a svn commit on top of a git commit. That, coupled with the lack of svn:commit hooks that fail (AFAIK), makes it very difficult to operate with the git repository being "canonical". >There has been lots of discussion on this list (scm-interest) and >actual work done regarding the switch, why don't we attempt an actual >switch using only a small subset of KDE (Amarok)? Are you volunteering? I'm trying to get some local support here (I live with 1 other developer and 2 system administrators [CC'd]) but all of us would be unknowns in the KDE development community, so I think Amarok might be better supported with at least a few volunteers that are already known (trusted, really) in the community. -- Boyd Stephen Smith Jr. ,= ,-_-. =. b...@iguanasuicide.net ((_/)o o(\_)) ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-' http://iguanasuicide.net/ \_/
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Kde-scm-interest mailing list Kde-scm-interest@kde.org https://mail.kde.org/mailman/listinfo/kde-scm-interest
