[jdk16u] Are we ignoring jdk16u-fix-* protocol?
Langer, Christoph
christoph.langer at sap.com
Mon Feb 15 22:54:56 UTC 2021
Hi,
I personally think that maintainer approvals, checked by "the system" would be a desirable feature for the future.
It doesn't necessarily have to be reading labels from JBS - it could also be entering an approval command in a PR or the like. (Similar to /sponsor for people who aren't committer in a project). This can be discussed in more detail once we get there.
But for the moment I agree to focus on moving the processes to GitHub/Skara and make sure we fix bugs and get things smooth without transforming the process beyond the minimum that's necessary to adopt Skara. So the jdk*u-fix protocol is what should be followed for the time being.
Best regards
Christoph
> -----Original Message-----
> From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.java.net> On
> Behalf Of Robert Mckenna
> Sent: Montag, 15. Februar 2021 19:53
> To: Kevin Rushforth <kevin.rushforth at oracle.com>
> Cc: Aleksey Shipilev <shade at redhat.com>; Attila Szegedi
> <szegedia at gmail.com>; jdk-updates-dev at openjdk.java.net
> Subject: Re: [jdk16u] Are we ignoring jdk16u-fix-* protocol?
>
> Sorry folks, I’m actually ooto atm so I haven’t been paying as much attention
> as I should be. (I have a cron job that should be warning me about things like
> this but I hadn’t enabled it since we added branch support to a library we
> use, I need to get that back up and running.)
>
> Certainly at some point deploying a check via skara bot would be helpful. It
> will need some discussion first however and we certainly don’t want to make
> changes like that for the first skara-based update.
>
> I will contact the fix authors if there are problems with retroactive approval.
>
> -Rob
>
> > On 15 Feb 2021, at 17:32, Kevin Rushforth <Kevin.Rushforth at oracle.com>
> wrote:
> >
> > With so many different processes and subtle rules for different projects at
> different points in time, it seems unwise for Skara to start down the slippery
> slope of looking at JBS labels (and maybe bug priority and issuetype (bug vs
> RFE)?) to decide whether it is OK to integrate a particular bug fix to a
> particular repo at a particular point in time.
> >
> > FWIW, calling the absence of this level of checking a "dealbreaker" seems
> like hyperbole when you don't have any such checks today in hg.
> >
> > Perhaps Rob McKenna can chime in.
> >
> > -- Kevin
> >
> >
> >> On 2/15/2021 5:32 AM, Aleksey Shipilev wrote:
> >> Hi Attila,
> >>
> >>> On 2/15/21 2:26 PM, Attila Szegedi wrote:
> >>> None of this is a justification for what happened, just an explanation
> how I got to screw up the process.
> >>
> >> That is not your fault.
> >>
> >> Really, that's a process bug: the bots should not have allowed to integrate
> without the approval. To me, the existence of such easy opportunity to miss
> the crucial step looks like a dealbreaker for adopting Skara for 11u and 8u
> projects.
> >>
> >>> I profoundly apologize about it. I’ll strive to do better. I’m fine with my
> changes being
> >>> reverted, adding the request tag, and resubmitting if that’s the
> reasonable way forward.
> >>> Alternatively, if it gets approved after the fact, I’ll graciously accept that
> too while
> >>> acknowledging that this isn’t the right way to go about it. I just added the
> 16u request tag to
> >>> the issue in JBS.
> >>
> >> No problem here. Retroactive approvals happen from time to time. My
> concern was not with the quality of the backport, but with the fact that 16u
> maintainers did not acknowledge it, while they should actually be in full
> control about what is going in.
> >>
> >
More information about the jdk-updates-dev
mailing list