Yeah, I totally agree it's an upstream bug, the bug exists in not only this version but also the master branch so I can't backport a fix.
I did have a quick look at the test code, and it's not obviously due to that, so it must be related to how they store or retrieve the offset in this case. I'll leave this bug open until we have an actual fix. But I think that preventing random failures from Debian build infrastructure is the best I can do for the time being. Thanks for adding the forwarded tag, I'm still learning and I forget these things sometimes. Chris Lamb <la...@debian.org> writes: > forwarded 841532 https://github.com/libgit2/libgit2/issues/3970 > thanks > > Russell Sim wrote: > >> I have forwarded this bug report upstream >> https://github.com/libgit2/libgit2/issues/3970 in the mean time I'll add >> a fix to the existing package to force tests to run in GMT timezone. > > Be careful; it could mean that you are masking underlying bugs by forcing > the timezone - the code (outside of the tests) itself could be buggy! > > > Regards, -- Cheers, Russell