Closed gmaze closed 11 months ago
I note that:
with core-free-py3.8: ========= 676 passed, 300 skipped, 53 warnings in 1091.14s (0:18:11) ==========
and with core-free-py3.9: ======= 676 passed, 300 skipped, 130526 warnings in 1578.13s (0:26:18) ========
We see a huge number of warnings ! but quite similar execution time in 20/25 mins. Under linux, with core-free-py.39 we have: ======== 676 passed, 300 skipped, 130531 warnings in 436.34s (0:07:16) ========= it's much faster and with similar number of warnings, so that's not the issue
Tests are failed because the execution time exceeds the 45mins limit. And it looks like all tests are passed, but that it's later on that the process takes too much time.
so, what was the change between python 3.8 and 3.9 under windows that makes tests to take forever ?
actually it's the tests that take forever to run, but the end of the pytest command, probably the generation of the jsonl report.
Will try to run tests without report generation
For a few weeks now, upstream tests under Windows have been systematically failing
This occurred between run 639 and 647
Need to investigate the cause.
Upstream tests are passer under core-free-py3.8 but fail under core-free-py3.9