Hi.
Thanks a lot for such comprehensive analysis of the issue!
El 7/4/25 a las 7:31, Simon Josefsson escribió:
At this point, I'm not sure if it makes sense to re-upload
golang-github-openpubkey-openpubkey with that Breaks, since nobody
should ever attempt to build opkssh 0.3.0-1 with
golang-github-openpubkey-openpubkey 0.7.3-1 again, right? There is a
implicit temporal build dependency ordering in each archive. I'm
tagging this bug with that information though.
In general, it's a good thing that we care about those versioned
relations, because otherwise it could happen that some packages
propagate to testing prematurely and they become unbuildable.
In this case I agree that it's too late and does not worth the effort.
- Yes, I know that this built ok in the buildd merely 6 hours ago,
but they have additional packages in the chroot and they disable
network access. My theories to explain this would be a) Maybe this package
is trying to access Internet and b) Maybe there are missing build-depends.
- In either case, weird cases like this one are the reason I always offer
a VM to test in case it's necessary.
Maybe you could add 'c) Maybe you uploaded a new version of a build
dependency that triggered this failure and forgot to add Breaks' to this
list of reasonable explanations.
Yes, this is also an option to consider so I will try to mention it
in similar future cases should they arise.
Thanks a lot!
