[Biopython-dev] Subversion Repository

[Chris Lasher]

On 3/21/07, Marc Colosimo <mcolosimo at mitre.org> wrote:
> I've found that svn is more useable than cvs. I especially like the
> move command (svn mv), which moves the file(s) and all the associated
> information. Thus, you can undo a move and see all the change history
> for the file(s). In addition, it handles binary files automagically.

Good points. Another +1 for Subversion, it tracks changes
concommitantly and maintains a revision across the entire repository,
not on a file-by-file basis. The Subversion Book explains this concept
better than I can. See <http://tinyurl.com/dhebm> and
<http://tinyurl.com/2wq3ne>.

> The one thing that I don't like at times is that you need to
> explicitly set keyword tags (like $Id$).

That's true. One thing you can do, however, is configure Subversion to
always set keyword substitution on all files of a filetype (e.g.,
always set keyword substitution for $Revision$ for all ".py" files).

> Also, some places block webDAV commands through proxies.
> So, keeping a read only cvs would be good.

Hmm, I hadn't thought about this. How common is this practice? It's
certainly a good argument in favor of maintaining CVS. (I'm assuming
CVS does not do WebDAV).

Chris