From: Harald Nordgren Date: Tue, 09 Jun 2026 13:20:04 GMT Subject: Re: [PATCH v13 2/6] branch: let delete_branches warn instead of error on bulk refusal Message-ID: In-Reply-To: On Tue, Jun 9, 2026 at 2:38 PM Junio C Hamano wrote: > > Harald Nordgren writes: > > > The GitHub CI has been broken for some time, maybe I should have told > > you about this earlier, but it coincided with a period where other > > open source projects I worked on also had mass CI failures, so I > > chalked it up to upstream issues (GitHub, Linux, etc). But it seems to > > have not gone away. > > > > All of my GitHub pull requests have broken tests (see e.g. which a > > quite minimal change: https://github.com/git/git/pull/2313). This > > makes it harder to detect actual issues. But of course it's not an > > excuse. > > FWIW, the breakage was observed in my local testing, and that is why > I found it so disturbing. Apparently you didn't see such breakages > that can be detected so easily during your local testing (otherwise > you wouldn't have pushed it out to update your GitHub pull request), > which may mean something in the test are platform dependent? No, it was broken on my local machine as well. I was sloppy when I pushed out v13 and didn't run tests locally. Usually I will push to my GitHub PR incrementally as I work on a new version. I don’t necessarily keep the GitHub PR clean between submitting versions. I will diff against my latest version tag to see my own progress, and if I mess it up I can always hard reset to the latest version tag to start over from there. Normally, I would never push out a new version unless the GitHub CI passes. But it’s been broken for a month. Harald