Re: [edk2-devel] failed Pr

["Ard Biesheuvel" <ardb@kernel.org>]

On Tue, 30 May 2023 at 17:36, Kinney, Michael D
<michael.d.kinney@intel.com> wrote:
>
> After looking at GitHub Action features the best approach is likely
> a PR comment that provides a request to rerun all failed jobs.
>
> We can define a comment format to indicate a command to a tianocore bot.
>

Sounds reasonable, although doesn't that imply that anyone who can
comment can trigger this action?


>
> > -----Original Message-----
> > From: Ard Biesheuvel <ardb@kernel.org>
> > Sent: Tuesday, May 30, 2023 7:36 AM
> > To: devel@edk2.groups.io; Kinney, Michael D <michael.d.kinney@intel.com>
> > Cc: Sean Brogan <spbrogan@outlook.com>; Rebecca Cran <rebecca@bsdio.com>;
> > Kubacki, Michael <michael.kubacki@microsoft.com>; Gao, Liming
> > <gaoliming@byosoft.com.cn>
> > Subject: Re: [edk2-devel] failed Pr
> >
> > On Mon, 29 May 2023 at 18:45, Michael D Kinney
> > <michael.d.kinney@intel.com> wrote:
> > >
> > > Hi Sean,
> > >
> > >
> > >
> > > I see the “rerun” button on the “checks” page of GitHub from my login.
> > >
> > >
> > >
> > > We have a mix of both Azure Pipelines and GitHub Actions today, so I
> > think it is important to see the “rerun” from GitHub view for Maintainers.
> > >
> >
> > Yes, that would be lovely.
> >
> > >
> > >
> > > This github page describes the feature, but not the permissions required.
> > >
> > >
> > >
> > > https://docs.github.com/en/actions/managing-workflow-runs/re-running-
> > workflows-and-jobs
> > >
> > >
> > >
> > > We currently have EDK II Maintainer Team set to “Triage Role”
> > permissions.  From the table below, only “Write” and above are allowed to
> > re-run.
> > >
> > >
> > >
> > > https://docs.github.com/en/organizations/managing-user-access-to-your-
> > organizations-repositories/repository-roles-for-an-
> > organization#permissions-for-each-role
> > >
> > >
> > >
> > > “Write” permissions add a lot of options, so I think we should discuss
> > tradeoffs in the next Tools/CI meeting.
> > >
> > >
> > >
> > > One option is to add a bot based on a new label to request failed jobs to
> > be re-run.  The bot can use a token with permissions to re-run.
> > >
> >
> > This would work for me - as long as I don't have to make changes to
> > the branch, I'm happy with whatever we come up with.