[openstack-dev] [TripleO]: landing code faster
ayoung at redhat.com
Wed Apr 20 16:27:34 UTC 2016
On 04/20/2016 11:44 AM, Dan Prince wrote:
> We've had a run of really spotty CI in TripleO. This is making it
> really hard to land patches if reviewers aren't online. Specifically we
> seem to get better CI results when the queue is less full (nights and
> weekends)... often when core reviewers aren't around.
> One thing that would help is if core reviews would +2 instead of +1'ing
> a patches. If you buy the approach of a gerrit review, the code looks
> good, etc. then go on and +2 it. Don't wait for CI to pass before
> coming back around to add your final stamp of approval. We all agree
> that the tripleo-check jobs should be passing (or have passed once
> collectively) before making any final +A to the patch.
Agreed. The rationale for +1 is only if you are a contributor to the
patch, and someone else has made changes. +1 Indicates that you are
happy with the other person's changes.
> The case for a core reviewer to +1 a patch is rare I think. If you have
> some comments to add but don't want to +2 it then perhaps add those
> comments with a +0 (or -1 if you think it needs fixed). Sure there are
> some edge cases where +1's are helpful. But if our goal is to land good
> code faster I think it would be more helpful to go ahead and +2 and let
> the CI results fall where they may.
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
More information about the OpenStack-dev