Even more interesting is the fact that whenever the test does not fail, the build log is like this:
============================= test session starts ============================== platform linux -- Python 3.14.5, pytest-9.0.3, pluggy-1.6.0 rootdir: /<<PKGBUILDDIR>> configfile: pyproject.toml collected 1 item tests/test_smoke.py . [100%] ============================== 1 passed in 0.04s =============================== and not like this: ============================= test session starts ============================== platform linux -- Python 3.14.5, pytest-9.0.3, pluggy-1.6.0 rootdir: /<<PKGBUILDDIR>> configfile: pyproject.toml collected 669 items tests/test_all_fixers.py . [ 0%] tests/test_fixers.py ................................................... [ 7%] ........................................................................ [ 18%] ........................................................................ [ 29%] ........................................................................ [ 40%] ........................................................................ [ 50%] ........................................................................ [ 61%] ........................................................... [ 70%] tests/test_main.py xxx. [ 71%] tests/test_parser.py ..x.F.............................................. [ 78%] ............................................... [ 85%] tests/test_pytree.py .................................... [ 91%] tests/test_refactor.py .................... [ 94%] tests/test_util.py ........................................ [100%] ------------------------------------------------------------------------------- So, it would seem that the tests would actually always fail if they would be always executed, and there is the additional problem that they are executed only half of the time. This is indeed a mystery. Thanks.
