How to contribute? (pull requests or non-publishing repo)
Martin Geisler
martin at geisler.net
Fri Apr 25 13:08:46 UTC 2014
Greg Ward <greg at gerg.ca> writes:
> Pierre-Yves --
>
> during the PyCon sprints last week, you said the best way to
> contribute is via pull requests. To me, "pull request" pretty much
> implies "immutable published history", at least with the current state
> of bitbucket (and github for that matter).
Slightly off-topic, not hopefully not too much. This isn't quite true:
GitHub works very well when you 'git push -f' to re-push your branch
after responding to review comments.
This one little and pretty simple feature makes a huge difference in how
you use the two tools. I've used Git since December and I consider that
to be the (only) cool feature :-)
I know that evolve will bring a more robust version of this to
Mercurial, but I just wanted to remark that I'm sad that we haven't
implemented a destructive push feature a long time ago. It really is the
key feature that makes Git popular, IMO.
I even went so far as implementing a small client that would login to
Bitbucket and strip a commit by driving their web interface like a
browser would -- turned out that this doesn't help much since a pull
request breaks in weird ways if you strip the commits involved. GitHub
pull requests just deal with the branch head moving around (I tried
simulating this with a bookmark on Bitbucket, but no luck).
--
Martin Geisler
https://plus.google.com/+MartinGeisler/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial-evolve-testers/attachments/20140425/13b71d80/attachment.asc>
More information about the Evolve-testers
mailing list