> -----Ursprüngliche Nachricht----- > Von: Joe Orton <[email protected]> > Gesendet: Dienstag, 11. Februar 2020 11:18 > An: [email protected] > Betreff: Re: Use of [skip ci] in commit messages to avoid Travis builds > > On Mon, Feb 10, 2020 at 07:22:45AM -0500, Eric Covener wrote: > > Is there anything possible in SVN like a pre-commit hook that can see > > what changed and append [skip ci] unless some other incantantation is > > present? > > Seems like a neat idea, I think that could work *if* infra actually > allows us to do that. I can imagine they might not want to since the > pre-commit hooks are global to the asf repos. > > An alternative I found would be to check the diff in the travis jobs and > pass if there are non-docs changes: > > https://reflectoring.io/skip-ci-build/ > > con: we'd probably still hit the transient failures at apt install > level, though I we could/should also move that to before_script level > anyway > > con: we'd have lots of "false" passes in the travis history, and having > a pass even if the tree is genuinely broken by prior non-docs commits is > bad. > > (... ideally the travis jobs could report "skip" rather than pass/fail > but I can't see a way to do that either.) > > pro: we can easily control what to skip w/o needing infra > > pro: doesn't need asf-wide repos changes & doesn't modify commits > (possibly fragile) > > Not sure about this. Maybe status quo is best and everyone should learn > to use [skip ci] until travis implement filtering. >
Given the above, especially the "false" passes I would tend to let everything as is and see if an increased awareness on [skip ci] already helps. Regards Rüdiger
