Me plenty of hours ago: "There's a regression in a forked project's version that introduced one new downstream commit and about 5000 commits from upstream. That one commit looks obviously unrelated, so I have to bisect these 5000."
...guess which commit it turned out to be ๐
It appeared so unrelated that when I eventually ended up with a known-good upstream version that now seemingly turned bad and realized that the trees I used differed just by that one commit, the first thing I checked was whether I used the same compiler toolchain as CI did. And yet... ๐