Control: tags -1 + upstream pending # technically fixed upstream, fixed locally, and confirmed good # for 7/7 attempts on abel.debian.org's armhf schroot
Dear Paul, Paul Gevers <elb...@debian.org> writes: > Source: vorta > Version: 0.7.3-1 > X-Debbugs-CC: debian...@lists.debian.org > Severity: serious > User: debian...@lists.debian.org > Usertags: regression timeout > > Dear maintainer(s), > > With a recent upload of vorta the autopkgtest of vorta fails in testing > on armhf when that autopkgtest is run with the binary packages of vorta > from unstable. It passes when run with only packages from testing. In > tabular form: > > pass fail > vorta from testing 0.7.3-1 > all others from testing from testing > > I copied some of the output at the bottom of this report, but as it > times out (after 2 hours and 47 minutes, where it used to take only > around 2 minutes) in autopkgtest, you can only get the last successfully > output log. > > Currently this regression is blocking the migration to testing [1]. Can > you please investigate the situation and fix it? > Thank you for filing this bug. I've been investigating armhf issues for some time, while collaborating with upstream about corner cases and races and locking issues other archs were encountering. At this point I think we finally have a working solution! It was very inspiring to see how to community came together upstream. The solution was adding four commits onto our 0.7.4+1 (cherry picked commit). The delta between 0.7.1-4 and 0.7.4+5 (cherry picked commits, future 0.7.5) might reasonably be considered "medium"; although it solves a *number* of RC bugs, and a lot of "million papercuts" user issues. Consequently I believe that Debian bullseye should have 0.7.4+5. Would you please advise me how to proceed with something like a checklist before contacting the release-team? My upload of 0.7.4-1 was rushed, and I didn't include a full list of commits relative to 0.7.1-4. Should I put the list of commits in the next changelog entry? I believe anyone reading the upstream commit history will agree that we should have 0.7.5 (=0.7.4+5commits) is higher quality than 0.7.1-4, and that that's the version we should have in Bullseye. I hope that my failure to reply to this bug until now doesn't count against the case--upstream can confirm I've been talking about armhf CI failures since before this bug was filed ;-) Kind regards, Nicholas
signature.asc
Description: PGP signature
