Hello, Am 15.08.25 um 10:15 schrieb Alexandru Mihail:
Thanks for the tips Soren. Some funky business indeed, it does NOT fail in pbuilder --twice. Works like a charm here. So we're on the second branch of your hypothesis. Something about the env in Salsa..
reprotest isn't about simply building a package two times in a row, it's about reproducibility with a different environment.
So yes, of course the build environment is different. https://salsa.debian.org/reproducible-builds/reprotest The first run has for example this setting beside others: INFO:reprotest.build:TIMEZONE variation: TZ=GMT+12 And the second build uses this. INFO:reprotest.build:TIMEZONE variation: TZ=GMT-14I personally don't look that much at this test and silence it if needed as there is no helpful output where to look. There is also the option that the core environment is misbehaving ao the run itself can't be successful. The people from reproducible build should able to help here.
-- Regards Carsten