Ah, I see. Thanks for explaining. That makes more sense! On Fri, Apr 5, 2019 at 10:29 PM 张铎(Duo Zhang) <[email protected]> wrote:
> Oh, Zheng Hu is a committer so he has the permission to merge... What I > said is that, he should approve first before merging... > > Misty Linville <[email protected]> 于2019年4月6日周六 下午1:19写道： > > > Yes, but you were doing the merge, unless they were a committer. I > > understood (perhaps incorrectly) Duo was describing a situation where a > > chang was merged by someone who shouldn’t have been able to do so > > otherwise. > > > > On Fri, Apr 5, 2019 at 9:52 PM Sean Busbey <[email protected]> wrote: > > > > > > > > This seems like a difference in ease compared to jira and not > > > something wildly different. There have certainly been times where a > > > committer posted a patch to jira for review and I merged it at a part > > > of giving my +1. > > > > > > We should make sure things default to squash-and-rebase instead of > > > merge for PRs in the UI, but I think we did that already. > > > > > > On Fri, Apr 5, 2019 at 10:54 PM 张铎(Duo Zhang) <[email protected]> > > > wrote: > > > > > > > > IIRC we have filed an infra ticket to disable several operations > > related > > > to > > > > PR, and for merging, I think we should only allow committers to merge > > > PRs. > > > > > > > > Misty Linville <[email protected]> 于2019年4月6日周六 上午10:11写道： > > > > > > > > > Can we protect the GitHub branches from direct merges? That’s a > > > repo-level > > > > > setting and we may not be able to change it. It seems potentially > > > dangerous > > > > > for people to be able to merge their own changes especially if it > > only > > > > > takes one successful reviewer. Other communities use mechanisms > like > > > Prow > > > > > [1] for this kind of thing. I imagine it requires some > infrastructure > > > > > though. > > > > > > > > > > [1] https://github.com/kubernetes/test-infra/tree/master/prow > > > > > > > > > > On Fri, Apr 5, 2019 at 7:04 PM 张铎(Duo Zhang) < > [email protected]> > > > > > wrote: > > > > > > > > > > > Yes, at least there should be a relevant JIRA issue. > > > > > > > > > > > > And on the retesting, we need to find a way to re-trigger the > > > webhook. > > > > > But > > > > > > anyway, we can fall back to use the old pre commit way, just > > > checkout the > > > > > > branch and make a patch and upload it to the jira issue... > > > > > > > > > > > > I'm trying to make use of GitHub in the recent works. And > > yesterday, > > > I > > > > > > added Zheng Hu as a reviewer for the addendum of HBASE-22152, and > > he > > > > > posted > > > > > > a LGTM and then just merged the PR... In fact I just want him to > > > approve > > > > > > the PR, this is the correct way to '+1' on GitHub. So I think we > > > need to > > > > > > write something done in the tell committers how to make use of > the > > > GitHub > > > > > > PR... > > > > > > > > > > > > > > > > > > > > > > > > Sean Busbey <[email protected]> 于2019年4月6日周六 上午9:43写道： > > > > > > > > > > > > > Excellent to see Duo! > > > > > > > > > > > > > > Do we have any guidelines for committers in the ref guide? I > > think > > > we > > > > > had > > > > > > > previously discussed calling out that they should make sure > > > there's a > > > > > > JIRA > > > > > > > for anything merged? > > > > > > > > > > > > > > Does retesting work from the github UI or is it like before > where > > > one > > > > > > > resubmits the jenkins job? > > > > > > > > > > > > > > On 2019/04/04 06:15:39, 张铎(Duo Zhang) <[email protected]> > > > wrote: > > > > > > > > Please see here > > > > > > > > > > > > > > > > https://github.com/apache/hbase/pull/110 > > > > > > > > > > > > > > > > Still need to polish the jenkinsfile so we can keep the same > > > > > experience > > > > > > > > with the old hadoop QA, but anyway, it basically works. > > > > > > > > > > > > > > > > So I think it is time to set up our github based workflow. > Need > > > to > > > > > > > discuss > > > > > > > > how to work together with our jira. > > > > > > > > > > > > > > > > Thanks. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
