Hi all, >> But it lacks an explanation why the flaky flag was removed. > > I think the explanation is the other changelog entry: > >> * Make autopkgtest more resilient when circuits are missing. >> Closes: 909275 > > FWIW there's a bit more info on the corresponding bug report.
Yes. If no circuit can be built for the test run (which may be due to the lack of control about the CI server's network situation), then the test failed in the past. However, it does indeed look like I missed something while adding checks for this case. > But indeed, given the tests are still flaky despite these > improvements, IMO they should be marked as such until someone has time > to make them more robust :) I think this is easily fixed. I will address this in an upcoming upload. BTW, is it intended to also have Tails releases in the Debian git repo with upstream versions newer than the one in Debian? If so, I would be glad if you could make sure to also include a pristine-tar tarball for the new upstream version so we don't work off different upstream sources. I have now imported the orig tarball from Tails [1] to make sure we're in sync for now. Thanks! Cheers Sascha [1] https://deb.tails.boum.org/pool/main/o/onioncircuits/onioncircuits_0.6.orig.tar.xz
