Re: Advancing to Patch::review

[Jonas Hahnfeld]

Am Mittwoch, dem 02.12.2020 um 12:37 +0100 schrieb Michael Käppler:
> Am 02.12.2020 um 12:27 schrieb Michael Käppler:
> > Am 02.12.2020 um 07:36 schrieb James Lowe:
> > > On 01/12/2020 19:10, Jonas Hahnfeld wrote:
> > > > FYI: 'make check' is running in pipelines for merge requests since this
> > > > morning 🙂
> > > > 
> > > > One thing I forgot in that lengthy email two weeks ago was how MRs get
> > > > to Patch::review. Traditionally James posted "Passes make check" and
> > > > updated the labels accordingly. As 'make check' is now happening
> > > > automatically, I think it makes sense to distribute this job: Basically
> > > > once the pipeline is green and 'make check' produced *something* to
> > > > look at, every member of the group passing by may update the labels. My
> > > > gut feeling is that this should be somebody else than the MR author,
> > > > but I'm not sold on that. Furthermore, it might be nice to make a
> > > > screenshot of the visual difference and post that and the link for
> > > > discussion, which makes it much easier to spot (tree.gittxt and the
> > > > spurious errors about "Parsed object should be dead" can be skipped).
> > > > 
> > > > WDYT? For the record, moving to Patch::countdown and ::push as well as
> > > > composing the countdown messages would remain on James as the Patch
> > > > Meister.
> > > > 
> > > > Jonas
> > > 
> > > Seems that MR #533 is the first.
> > > 
> > > I'll manually test MR 531 as that seemed to have no quite made the cut.
> > > 
> > > James
> > Hmm... it seems that on #534 the 'test' stage was run, not 'check'.
> > I'm completely lost, did I do something wrong?
> Ah, okay, I did not rebase and my branch was based on an outdated master
> that did not have the "Make check for merge requests" commit, right?

Yes, the CI can only run the definitions present in .gitlab-ci.yml.

Circling back to my original proposal:
> My gut feeling is that this should be somebody else than the MR author

Do I interpret your actions that you disagree with this? To elaborate a
bit, this tries to keep the pleasant effect that somebody else at least
opens your MR and nobody is tempted to change labels because "I'm sure
this will pass testing".

Jonas

signature.asc
Description: This is a digitally signed message part