[wp-hackers] Switching from SVN
aoberoi at gmail.com
Fri Dec 10 05:59:09 UTC 2010
i write better code if i can experiment more.
i experiment more if its easy to branch
i branch if its easy to merge
its easier to merge in a distributed system because of the features it has
(rebasing, submodules, tracking changes across files, "smarter" diff-ing
through ancestry, n-way merges)
if i write better code, you (as a maintainer) get less crappy stuff to deal
with, more time for beer.
On Fri, Dec 10, 2010 at 12:53 AM, Doug Stewart <zamoose at gmail.com> wrote:
> SVN has always stunk at merging, and the Collab.net guys (and now Apache
> Foundation, I suppose) will readily admit that. It is a difficult matter.
> The problem is that you're still coming at this from an RCS/CVS/SVN "repo
> is king" mindset. I'm on my mobile, so I'll let the following link describe
> exactly what I mean:
> It's a mindjack at first, especially if you're used to an orderly hierarchy
> for changes, but once you glom onto the distributed mindset, it's really a
> Doug Stewart
> On Dec 10, 2010, at 12:44 AM, Otto <otto at ottodestruct.com> wrote:
> > On Thu, Dec 9, 2010 at 11:38 PM, scribu <mail at scribu.net> wrote:
> >> That's where your SVN mentality comes in, where merges are hard.
> > No. Merges are hard because *merges are hard*.
> > If two people change the same line, then somebody has to fix the line.
> > Simple as that. A version control system can't fix that. This is code,
> > not magic.
> > Even SVN can merge two different changesets that don't touch the same
> > lines (but do touch the same files). I use this all the time. It works
> > fine.
> >> No, as previously mentioned by someone else, a merge that doesn't work
> >> of the box would simply be marked as needs-refresh.
> > Sooo... how is this different then?
> > On Thu, Dec 9, 2010 at 11:41 PM, scribu <mail at scribu.net> wrote:
> >> It's easier when you move code from one file to another, for example,
> >> because git keeps track of that.
> >> I had that problem when working on my GSoC project, which involved a lot
> >> moving code around.
> > I admit that that is a great improvement over SVN's purely single file
> > based system. However, it seems like a fairly thin and uncommon case
> > to me. Furthermore, this is basically an incremental improvement that
> > is not based on the distributed nature of the version control system.
> > How does making the version control *distributed* improve the ability to
> > -Otto
> > _______________________________________________
> > wp-hackers mailing list
> > wp-hackers at lists.automattic.com
> > http://lists.automattic.com/mailman/listinfo/wp-hackers
> wp-hackers mailing list
> wp-hackers at lists.automattic.com
More information about the wp-hackers