When will CVS be replaced by modern version control system?

Jeremy Katz katzj at redhat.com
Mon Nov 12 14:02:26 UTC 2007


On Sat, 2007-11-10 at 10:31 -0600, Josh Boyer wrote:
> I think it's because you have two classes of users.  Developers, and
> packagers.

I think you hit the nail _exactly_ on the head here.  We continue to
have this discussion moving around in circles but when you step back a
little bit, this starts to be more clear.

So the next question I'd end up having is how do we think about what
we're trying to enable such that we continue to make things "easy" for
the packagers while also allowing more complicated interactions for the
developer case.  

The current best I've got (which isn't really fleshed out at all, but
maybe someone with less on their plate than me is interested? :) is that
we really want to have "the central repository of spec files + patches"
like we do now[1] but that we then also want an easy way to get from
that to something more developer friendly.  I suspect that the more
developer friendly format is probably more like exploded trees with
patch series on top so that you can easily import new sources, rediff,
etc.  And then from that repo, you can have a script to get back to the
spec file + patches on top of the pristine source.

Jeremy

[1] And maybe there are convincing reasons to move that off of CVS, but
I think that we need to have a better idea of the full picture before we
can say that definitively.  And I say that as a now-relatively-rabid
"git for the win" user :-)




More information about the fedora-devel-list mailing list