[SciPy-Dev] Van der Waerden test in Scipy?

Fernando Perez fperez.net@gmail....
Fri Jun 3 00:33:20 CDT 2011


On Wed, Jun 1, 2011 at 2:07 PM, Ralf Gommers
<ralf.gommers@googlemail.com> wrote:
> Pull requests should be for code that is ready to be merged imho, not for
> using it as a poor man's Trac. For numpy we have kept the number of open
> pull requests below ten for almost a year now because we haven't been using
> it for unfinished contributions like this. Please let's try to keep it like
> that.

+1

While work is in-development, I'd suggest that a ticket be opened in
the scipy repo, with a link to a branch in the person's own fork.  The
discussion can then happen on the scipy repo, but the work will evolve
in that fork possibly changing a lot, maybe even being dropped and
with a new branch started, perhaps a different person will come up
with a better approach elsewhere, etc.

Once something gets baked out enough to be truly pulled, then such an
action should be requested.  Hence the name, pull request :)

Cheers,

f


More information about the SciPy-Dev mailing list