Hello, IMO it is high time to revisit this issue.
On 2020-03-11 14:33, Matthias Klose wrote: > openbabel autopkg tests areq quiet interesting. Working around the timeout on > arm64 hides the fact that two other tests fail on every non-amd64 > architecture. I agree that the way the autopkgtest is written hides some of the failures. Your patch does indeed un-hide them. > Also looking at the test output, you see that one test succeeds despites the > python3 binary segfaulting. Test failures are ignored for openbabel now (pinging Michael who did that [1]). If we really aim at catching all test failures, I guess we should not ignore them. However, this may result in openbabel being RMed from non-amd64 architectures and I got the impression that the upstream may not have enough resources to debug on non-amd64 architectures [2]. [1] https://salsa.debian.org/debichem-team/openbabel/-/commit/eb17fde215d2ec6d3b37f365ff42250cde3700e7 [2] https://github.com/openbabel/openbabel/issues/2417#issuecomment-942098541 Best, Andrius
