Despite #3884, we still get random unit test failures related to the plane stress calculation. #3027 #3856 are issues and PRs related to this issue occurring before.
I believe due to the random nature of failures (on the CI, failure or lack of failure is deterministic for a give system) that this could be related to software versions (e.g. libmath) being slightly different. We should find out what software is causing this and update the skipping criteria added in #3884 to reflect the actual cause.
================================ tests coverage ================================
_______________ coverage: platform linux, python 3.10.19-final-0 _______________
Coverage XML written to file coverage.xml
=========================== short test summary info ============================
FAILED tests/unit/test_tfcoil.py::test_plane_stress[planestressparam2] - assert array([ 1.953...71252995e+05]) == approx([9.769...4 ± 0.171253])
comparison failed. Mismatched elements: 1 / 300:
Max absolute difference: 9.769733861957293e-08
Max relative difference: 0.5
Index | Obtained | Expected
0 | 1.9539467723914586e-07 | 9.769733861957293e-08 ± 1.0e-12
=========== 1 failed, 647 passed, 103 warnings in 104.20s (0:01:44) ============
Despite #3884, we still get random unit test failures related to the plane stress calculation. #3027 #3856 are issues and PRs related to this issue occurring before.
I believe due to the random nature of failures (on the CI, failure or lack of failure is deterministic for a give system) that this could be related to software versions (e.g.
libmath) being slightly different. We should find out what software is causing this and update the skipping criteria added in #3884 to reflect the actual cause.