Re: numpy 1.2.1, switching to git?
On Tue, 23 Dec 2008 20:01:06 +0100 Pietro Battiston <toobaz@email.it> wrote:
>Il giorno mar, 23/12/2008 alle 11.41 -0500, Scott Kitterman ha scritto:
>> On Tue, 23 Dec 2008 15:08:03 +0100 Loïc Minier <lool@dooz.org> wrote:
>> >On Mon, Dec 08, 2008, Ondrej Certik wrote:
>> >> P.S. bzed, POX, isn't it time to move our packaging to git? So that I
>> >> can just commit such patches in a branch and also so that we don't
>> >> have to mess with the orig.tar.gz, svn-uscan and other things -- i.e.
>> >> everything will be in one git repo, so users can just download, hit
>> >> one command and they have a working package (as opposed to the current
>> >> scheme, were they need to download svn, then setup some tarball
>> >> directories, then setup svn-uscan, then execute it and only then they
>> >> can actually build the package, so it's very annoying for casual users
>> >> to setup the environment to contribute to the packaging)
>> >
>> > WARNING bikeshedding VCS discussion spotted!
>> >
>> > Executive suggestion: we wont be able to use the same VCS for all
>> > packages; use whatever the top contributors prefer, not what all
>> > contributors to all packages prefer.
>> >
>> I'll argue we want something different. We want VCS that will maximize participation. That means both keeping top contributors happpy and keeping it accessible to newcomers.
>>
>> I don't think hg, bzr, or git obviously qualify as accesible. My vote, fwiw, is to stay with svn until we have a documented, accessible workflow with tool support.
>
>Just my two cents (I certainly have no right/reason to express a vote, but I followed you discussion with interest): as of now, git may maximize participation of newcomers because it's seen as "cool" (whether or not it is) and more and more adopted, while bzr may maximize participation of newcomers because it's popular _now_ (Launchpad and Ubuntu are; I don't think I'm the only maintainer that became it because of being an Ubuntu user).
>
>I frankly don't see how svn can maximize participation of newcomers. Svn
>users are tipically long-time "versioners" who probably tried at least
>once some more recent versioning tool, probably bzr, or git, or hg (lots
>of projects switched from svn to one of those), while bzr users often
>have no idea of what svn is.
>
>Then one can say "let's stick to svn because it's a pain to change"; I
>just think the "participation" argument is misleading.
>
Anyone who's used VCS for more than a couple of years has used CVS or SVN (and using either is very similar). Yes all the cool kids are using Git these days, but even in Debian it's much less used than SVN (accelerating though). In Ubuntu, not many developers outside of a group of core-devs uses bzr (note: I'm an Ubuntu core-dev, so I have more than a passing familiarity with the subject). The biggest kick against BZR is that it's little used outside of Ubuntu. OTOH, you can, if you want, use BZR just like you would SVN (bzr co, bzr ci -m ' .... ').
Everyone on the team has a workflow for SVN. Not true for the others. We have a working system and we ought not move off of it until we have a approach that is easily accessible and well documented.
Scott K
Reply to: