geode-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ernest Burghardt <eburgha...@pivotal.io>
Subject Re: [DISCUSS] Blocking merge button in PR
Date Fri, 18 Oct 2019 19:46:44 GMT
I had one recently that was Approved and I merged pre-maturely and had to
be reverted: d63638e4654bc6c71a232838b745dec6ef476ec9

Subsequently I have run into some test flakiness, but if a PR submitter has
a pre-checkin failure it could be tricky to tell that its a Flaky
situation... In my last go at a Flaky failure in pre-checkin, I was able to
search the Geode Jira and found the failure was a known flaky like this one
<https://issues.apache.org/jira/browse/GEODE-6324>

I'd prefer to trust our committers to perform their due diligence and make
good choices.

EB

On Fri, Oct 18, 2019 at 12:18 PM Owen Nichols <onichols@pivotal.io> wrote:

> Do you have a recent example of a PR that was merged despite failed PR
> checks, which then broke the build?
>
> At last discussion, one concern raised was providing a way that anyone in
> the community could re-trigger a failed PR check if it hit an unrelated
> flaky failure.
>
> Let’s be sure we've identified the problem before assuming the solution.
> Apache values people over process.
>
> > On Oct 18, 2019, at 11:48 AM, Nabarun Nag <nnag@apache.org> wrote:
> >
> > Hi devs,
> >
> > A few months ago a proposal was brought up regarding blocking the merge
> > button on the github PR page in case of failing tests in the precheck.
> >
> > What is the sentiment regarding this now? Do we feel that it should be
> > implemented?
> >
> > Or at least take the minimal step of not allowing merge till all tests
> are
> > done?
> >
> >
> > Regards
> > Nabarun Nag
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message