[gnu.andrew at redhat.com: Re: OpenJDK 8u and Backport Bugs]

[Langer, Christoph]

Hi,

I'm also all in for stopping the label pollution ��

So, first of all, I fully agree that a backport bug should have the person doing the backport as the assignee. Currently this can only be set manually, e.g. at the time of manually creating a backport item but also any time later by manual intervention. hgupdater currently can't do that since current practice is mostly to submit a backport patch with the original authors/reviewers. This, however, will be solved by the Skara tooling when we switch there. Skara backporting will honor the backport authors (and reviewers) and also reflect that in the backport items in JBS. You can watch this already in real life in 13u. So this change in tooling and process is really one of the things I like about Skara and why I think we should go there for 11u rather sooner than later.

But I do not want to have an obligation to manually create backport issues while the tooling can do it for you. People who want to do it - fine. But the machinery can do this for me. And I want to rely on it since it means less effort and thus process overhead for me. And, finally with Skara tooling in place, backport statistics via backport items will be fully correct.

As for maintainer approvals: We currently do this via labels/comments on the main bug. I can imagine moving this process to GitHub issues eventually, after Skara was rolled out. E.g. backport authors would start a backport PR with all the reasoning in the PR comments. Skara could offer some command like "/approve"/"disapprove" for maintainers. This could become a prerequisite an "/integrate". But let's do one thing after the other...

Best regards
Christoph


> -----Original Message-----
> From: Severin Gehwolf <sgehwolf at redhat.com>
> Sent: Donnerstag, 17. Dezember 2020 19:48
> To: Hohensee, Paul <hohensee at amazon.com>; David Holmes
> <david.holmes at oracle.com>; Langer, Christoph
> <christoph.langer at sap.com>; Andrew Hughes <gnu.andrew at redhat.com>;
> jdk-updates-dev at openjdk.java.net
> Cc: jdk8u-dev <jdk8u-dev at openjdk.java.net>
> Subject: Re: [gnu.andrew at redhat.com: Re: OpenJDK 8u and Backport Bugs]
> 
> Thanks for the clarification, Paul.
> 
> On Thu, 2020-12-17 at 18:20 +0000, Hohensee, Paul wrote:
> > Manually created backports can of course be assigned to anyone, but
> > assigning them to anyone other the original patch author is a change
> > from current practice.
> 
> This is a feature, not a bug, IMO. It more accurately reflects who did
> the backport.
> 
> Thanks,
> Severin