Remove allowed_failure from Python 3.12 #5815
Conversation
and disable profiling tests there, since it doesn't look like we'll be able to fix them. Knowing about other regressions is probably more important than testing profiling...
and disable profiling tests there, since it doesn't look like we'll be able to fix them. Knowing about other regressions is probably more important than testing profiling...
|
In 3.0.x in f71205d |
|
Looks like we used the "in" operator just above to disable tests. Seems simpler than something as obscure as "find() != - 1"
|
I've gone back and fixed that. |
|
I missed that there's a Windows Py3.12 failure. I'll have a look at that this evening. Possibly by just turning Windows 3.12 to an allowed failure for now. |
|
It seems master is failing in more than just windows. The master run (after the merge of #5819) seems to be failing on python3.13, ubuntu-20.04, pypy-3.9-v7.3.12, the latter with a failure in |
|
#5826 is the fixup for CyCache on Windows. @mattip - I think that trashcan test may be "flakey" in PyPy rather a permanent failure (which I suspect we can live with) - it seems OK on other recent runs. That probably makes sense - it's looks like something that might work best with deterministic destructors. |
and disable profiling tests there, since it doesn't look like we'll be able to fix them.
Knowing about other regressions is probably more important than testing profiling...