On 2017-12-06, Jelmer Vernooij wrote: > On Wed, Dec 06, 2017 at 05:45:11PM -0800, Vagrant Cascadian wrote: > Can you reproduce the issue once you set LC_ALL=C.UTF-8 ? > > The delta between -8 and -9 is minimal and not related to the test > that failed, so I'm wondering if this is a spurious failure (timing?) of some > sort or perhaps a regression introduced by a dependency.
Well, I managed to reproduce it on amd64 too, so I think this is a different issue entirely than what caused the original armel build to fail (which might have just been a transient failure)... The following link should have two build logs, one with the C.UTF-8 patch applied to debian/rules, and one with without it, built with sbuild: https://cascadia.debian.net/~vagrant/bzr/patched.build.gz https://cascadia.debian.net/~vagrant/bzr/unpatched.build.gz The patched build succeeds, the unpatched build fails. I'm guessing something in the toolchain changed the way it handles UTF-8 since the builds almost a month ago... Hope that's helpful. live well, vagrant p.s. I noticed no almost no arch:all buildd logs, and very few recent amd64 buildd logs... maybe save some bandwith for yourself and start doing source-only uploads. :)
signature.asc
Description: PGP signature