On 2026-05-05, Peter Blackman wrote: > Looks to me like a systemic issue > with CI reproducibility. > > 2nd runs are failing across the board with logs like > > I: Adding a custom variable just for the fun of it... > I: Changing /bin/sh to bash > '/bin/sh' -> '/bin/bash' > lrwxrwxrwx 1 root root 9 Jun 7 00:21 /bin/sh -> /bin/bash > I: Setting pbuilder2's login shell to /bin/bash > /tmp/hooks/D01_modify_environment: 29: usermod: not found
Sorry about that! At one point it affected over 40% of the build results for unstable/amd64, and started affecting the builds for forky recently as well... I noticed it a little over a week ago in unstable, but sometimes glitches like this "sort themselves out" ... this obviously did not! Should be worked around as of yesterday when this commit was deployed: https://salsa.debian.org/qa/jenkins.debian.net/-/commit/a741f156dff8dd2f2a0b6531756865c958ef7fd7 It will take a while for things to get back to normal... but future builds should start working... Something, not sure what exactly, caused "passwd" to no longer be part of the base tarball, possibly the indirect results of aiming for a smaller and smaller base set, and the last thing depending on "adduser" probably dropped it... There has been a lot of focus and excitement on the "new" and important reproduce.debian.net, which actually verifies against the archive, which we have wanted to be able to do for so many years! It really is great! The older tests.reproducible-builds.org serves other purposes (e.g. testing toolchain regressions and several important variations such as systematic time variation) and still needs some attention now and again! live well, vagrant
signature.asc
Description: PGP signature
_______________________________________________ Reproducible-builds mailing list [email protected] https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/reproducible-builds
