Re: [email@example.com: Re: The powerpc port should be removed from etch release candidates ...]
On 5/17/06, Sven Luther <firstname.lastname@example.org> wrote:
On Wed, May 17, 2006 at 12:18:28AM -0400, Albert Cahalan wrote:
> On 5/12/06, Sven Luther <email@example.com> wrote:
> >There was an easy way not to problong the discussion. Restore the svn
> >acces, which you could have done all those weeks ago if you had not been
> >proud and afraid to lose face.
> Better: do like Linus, and take away access from all but one person.
> BSD has always had nasty fights over commit access. Commit bits
> are greatly political in nature. They can not be removed without hurt.
> Thus the existance of OpenBSD and DragonflyBSD.
> Linux doesn't suffer this way. The closest thing ever was the IDE
> maintainer being changed a couple times. (so avoid MAINTAINERS
> files too) Feelings can't get hurt all that much if there isn't any
> status to revoke.
> Both SVN and CVS have a server-centric model that ultimately leads
> to nasty poltics. The alternatives are git, Mercurial, and monotone.
It does mean forking and fragmentation of the code base, which would not be
best for d-i and debian. But yes, having a distributed revision system would
be helpful in these cases, and if people don't come to their sense and this
issue be solved, i will be left only to create a svk-based duplicate of the
d-i svn repo, and make this one the authoritative version for the packages i
upload or changes i make. Imagine the mess this will cause :)
Forking and fragmentation vs. BSD-style fights? No contest.
Linux has been forked about a zillion times, but the developers don't
mind the "mess" at all. People just merge every which way using git.
If people come to their sense, they'll never again grant commit bits
to more than one person.
BTW, the Wine project has a nice comparison. It seems that git is
much more compact than svk. (big surprise for me) It's well-known
for being very fast.