On Tue, Oct 15, 2024 at 03:11:19AM +0200, Santiago Vila wrote: > El 15/10/24 a las 2:16, Gregor Riepl escribió: > > > > The following packages have unmet dependencies: > > > > sbuild-build-depends-main-dummy : Depends: python3-skimage but it is > > > > not installable > > > > E: Unable to correct problems, you have held broken packages. > > > > apt-get failed. > > > > E: Package installation failed > > > > > > Broken build dependencies are often FTBFS in a build dependency, > > > and it creates extra work having to manually merge such bugs. > > > > > > Please check why a build dependency is no longer installable before > > > opening such bugs. > > > > I'm confused by what happened here exactly, and what I could have done to > > help the situation. Reading through #1084781 now to see the context... > > In this particular case, there was nothing to do on your side, that's why > Adrian reassigned > the bug, meaning "not really an issue for you".
But if I hadn't done this, there would have been a maintainer confused and wasting time on this - one maintainer per bug. > In the general case, however, a build failure due to a build-dependency which > may not > be installed might indicate that the build-dependency is obsolete, in which > case the > maintainer should decide what to do. >... The people who have routine in finding quickly whether this is true, and who might already know what the problem is, are the people who are having an overview of the whole archive because they are doing archive-wide QA. The average maintainer does not have any clue about archive-wide breakages, or issues due to ongoing transitions. On Mon, Oct 14, 2024 at 12:44:05AM +0200, Santiago Vila wrote: >... > During a rebuild of all packages in unstable, your package failed to build: >... Uninstallable skimage should make the build dependencies of around 30 packages uninstallable. The sane approach is that one of the people who are doing regular archive-wide QA checks what the problem is, and then either sends one bug for the root cause, or describes in the 30 bugs what changed. That's pretty fast when you have routine doing it. The insane approach is to send 30 RC bugs with the raw error message, and then 30 maintainers are wasting time on that issue. If you would have tried to check why the package is not installable, you would have quickly gotten to [1] which lists the FTBFS bug there since I added the affects to #1084781 on Saturday. skimage has been workarounded, right now we have an openmpi FTBFS and a perl transition - each breaking plenty build dependencies. That's normal in unstable. It is not helpful to file bugs against random packages that are just victims of whatever the breakage of the day is in unstable. > Thanks. cu Adrian [1] https://buildd.debian.org/status/package.php?p=skimage
