Closed xylar closed 1 year ago
When I run a custom suite with:
ocean/spherical/icos/geostrophic
ocean/spherical/icos/geostrophic/with_viz
and with the convergence tolerance set to 2.0 (so analysis fails), I see the expected failures with this branch, whereas for a similar test with #104, I was seeing:
Traceback (most recent call last):
File "/gpfs/fs1/home/ac.xylar/e3sm_work/polaris/ocn-port-sphere-transport/polaris/run/serial.py", line 326, in _log_and_run_task
baselines_passed = _run_task(task, available_resources)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/gpfs/fs1/home/ac.xylar/e3sm_work/polaris/ocn-port-sphere-transport/polaris/run/serial.py", line 399, in _run_task
config_filename = os.path.join(step.base_work_dir,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "<frozen posixpath>", line 90, in join
File "<frozen genericpath>", line 152, in _check_arg_types
TypeError: join() argument must be str, bytes, or os.PathLike object, not 'NoneType'
Before this merge, we were not setting a config's
filepath
attribute after loading the config options from a file. This causes the step to fail for cryptic reasons if it is rerun later as part of a test suite (because it failed the first time).Checklist
Testing
comment in the PR documents testing used to verify the changes