On Wed, Feb 11, 2026 at 11:55:21AM -0500, Jeremy Bícha wrote: > pkgconf 2.5.1-3 didn't seem to work. lowdown and zarchive's > autopkgtests are still failing with the new pkgconf. > > https://qa.debian.org/excuses.php?package=pkgconf
Indeed... 8b3dd835aee113ede402e98a502a0e9f1b7b1132 seems to also be in play here. So basically: * == 2.3.0 = good * == 2.4.0 = bad * == 2.4.0 + a79952a reverted = good * >= 2.4.1 + a79952a reverted = bad * >= 2.4.1 + a79952a + reverted + 8b3dd83 reverted = good (good/bad in bisect jargon; >= 2.4.1 means up to and including 2.5.1) That said, 8b3dd83 is literally the only change between 2.4.0 and 2.4.1. I would be hesitant in just reverting it. If what it fixed was important enough to release a point release for, then I wonder what bugs will get reintroduced by undoing it. I wonder if you've brought this broader bug up with upstream? These are just educated guesses on my end based on bisect and commit messages, and are not expressed with any confidence. I really do not know much about pkgconf's internals. Faidon
