On Tue, Dec 19, 2023 at 10:47 PM Dmitry Kazakov <dimul...@gmail.com> wrote:
> The main problem for me is that I want to have the patch merged asap, to > be able to test/resolve issues of some other patches in combination with > this one. That's why I just merge that without rebase-pipeline completion. > > Speaking truly, I'm not very happy with our 'always fast-forward merge' > policy [0], but given GitLab doesn't provide an option to switch policy > on-the-fly, it should be fine for now. > That global policy is in place for good reason - and not only because people want a clean history. It is also there because merge commits cannot be easily reverted and you do not want to try reverting a merge commit (it ends very very badly - we've had to force push repositories out of that mess before) Cheers, Ben > > [0] - this policy is good for small patches, but for bigger patchset I > would prefer to see an explicit merge commit > > On Tue, Dec 19, 2023 at 10:25 AM Ben Cooksley <bcooks...@kde.org> wrote: > >> On Tue, Dec 19, 2023 at 10:16 PM Dmitry Kazakov <dimul...@gmail.com> >> wrote: >> >>> Hi, Ben! >>> >>> I think we have kind of forgotten about that. I usually build the MRs >>> locally or check if the pipeline has succeeded for the current revision, >>> then I just "rebase without pipeline" and do the merge. It might be that I >>> should use the merge service for that, I'm not sure... >>> >> >> That is fine. Given it doesn't seem to be too well loved i'm going to >> decommission the service. >> >> The initial issues we had with it's reliability had been resolved however >> I think that happened not too long before people stopped using it. >> >> Cheers, >> Ben >> >> >>> >>> On Sat, Dec 16, 2023 at 2:41 AM Ben Cooksley <bcooks...@kde.org> wrote: >>> >>>> Hi all, >>>> >>>> I've just been reviewing services we're looking after and while doing >>>> so have noticed that Marge Bot (invent.kde.org/merge-service) doesn't >>>> seem to have done anything for 4 months. >>>> >>>> As the two projects with it enabled, is this something you're still >>>> using as it doesn't look like it. >>>> >>>> Thanks, >>>> Ben >>>> >>> >>> >>> -- >>> Dmitry Kazakov >>> >> > > -- > Dmitry Kazakov >