[SciPy-dev] What code base to use?

Robert Kern rkern at ucsd.edu
Wed Nov 9 01:00:19 CST 2005


Travis Oliphant wrote:
> Robert Kern wrote:
> 
>>Scipy dev's: what do you think about moving the current trunks into
>>branches (or tags) and making the "new" branches the trunks ASAP?
> 
> I think that is a great idea.
> 
> I'm not sure exactly what that means for everybody's tree and if they 
> will have to check-out again.  I know there a few people still working 
> on some parts and haven't checked in what they have.
> 
> I would suggest that people check in their code as soon as possible. 

I'm pretty sure we'll have to check out again (if nothing else, that's
probably the most obvious, least complicated method). We're going to
have to pay the cost eventually, so we might as well do it now. I
suggest laying down a firm timetable, e.g. declaring a code freeze
beginning at such-and-such hour on Thursday.

I'll also take the opportunity to plug SVK (http://svk.elixus.org/)
while I'm at it. SVK is a form of distributed RCS that uses SVN's
database underneath. It's also why my checkin messages look funny. I use
SVK to mirror the main repository and make a local branch. I can do all
of my work locally and make checkins to the local branch without needing
to talk to the main repository.

When you switch the directories around, I'll have to checkout the new
trunks as new mirrors

  $ svk mirror http://svn.scipy.org/svn/scipy_core/trunk //mirror/scipy_core
  $ svk sync --skipto HEAD //mirror/scipy_core
  $ svk cp //mirror/scipy_core //local/scipy_core

but then I can switch my working copy while keeping all of my uncommited
changes

  $ cd ~/svk-projects/scipy_core
  $ svk switch //local/scipy_core

One downside to SVK for working with scipy is that it doesn't use the
.svn/ directories, so I have to forge __svn_version__.py files in order
to build. Natch.

-- 
Robert Kern
rkern at ucsd.edu

"In the fields of hell where the grass grows high
 Are the graves of dreams allowed to die."
  -- Richard Harter




More information about the Scipy-dev mailing list