Yea I agree with Sergio, my only thought is the HiveQA is currently
dependent on the patch upload system, its just the way it was built, and
would take work to change to use the PR system along with the concerns
Sergio brought up.
So if we decide to go with this, we would have keep to the upload p
Github pull requests are significantly better than review board:
* They are tied to a git branch and thus reference the base commit. It is
trivial to review changes to branches other than master.
* Updating them is as easy as pushing an update to github.
* The pull request and comments automatical
I understand that some projects are using pull request. However, I'm not
sure why we need to do the same just because that works for those projects.
Hive has its own working model. Unless it brings significant benefits, I
don't really see the point of switching. On the other hand, I will see a
lear
I liked the idea of having pull requests as an option for review :). This
way we could have our own local repository, and commit our changes to a
personal branch instead of creating patches to RB for every new change.
Just one question, is this going to affect the way the precommit tests
work? I t
Thanks for the link to the previous discussion. Much of the previous
discussion was about the discussion about git versus subversion. Obviously,
we decided to go forward with that. We already have pull requests on
github. See the list at https://github.com/apache/hive/pulls
Without the Apache inte
I personally am a big fan of pull requests which is primarily the reason
for a similar proposal that I made almost a year and half ago[1] :). I
think the consensus we reached at the time was that to move the primary
source code from svn to git(which we did) but still use patches submitted
to JIRAs