[Santoshkumarpuppala]

Description of changes

python/ctsm/run_ctsm_py_tests.py dropped the return value of pytest.main(), so ./run_ctsm_py_tests exited 0 on failures and the python-unit-tests job stayed green. The runner now calls sys.exit(pytest.main(pytest_args)).

That alone would turn the job red at collection. The checkout had no cime, so 8 test modules failed to import CIME and pytest stopped before running a test, and the .nc inputs were LFS pointers. The job now checks out with lfs: true and runs bin/git-fleximod update cime, the same pattern as docs-common.yml:24-32 (which updates all submodules with -o). Background in #4230.

Until #4231 is merged, this job fails on test_inputdata_setup_files_gswp3_error (it looks for /glade). Merging #4231 first avoids that on b4b-dev.

Specific notes

Contributors other than yourself, if any:

• None

CTSM issues resolved or otherwise addressed, if any:

• Resolves CI for python tests maybe broken/missing? #4230 (the unit-test job)
• Some work on Python system tests are Python system tests aren't portable #3236)

Any user interface changes (namelist or namelist defaults changes)?
No namelist changes. ./run_ctsm_py_tests and make utest now exit non-zero when a test fails.

Testing planned or performed, if any:
Local only: a depth-1 checkout of this branch on b4b-dev 8e3f959, pip venv with ctsm_pylib's test-dependency pins (Python 3.13, pytest 8.3.5), network blocked. No conda, and the workflow itself has not run.

• Stub test: failing exits 0 before this change and 1 after; passing exits 0. make utest with the failing stub: 0 before, 2 after.
• The CI command with today's checkout: collected 558 items / 8 errors, exit 0 before, 2 after.
• bin/git-fleximod update cime as written: cime at 92ae0fd (cime6.5.5), no other submodule checked out.
• cime plus the LFS .nc inputs (sha256-checked against the pointers): 602 collected, 5 deselected (below), 596 passed, 1 failed, the gswp3 test. With test(subset_data): make unit test work outside NCAR cluster, like its siblings do #4231's line applied, 597 passed, exit 0. Without the LFS inputs, 35 tests fail on pointer files.
• pylint 3.3.4 (10.00/10) and black 25.1.0 --check on run_ctsm_py_tests.py.
• Not run: the 5 tests that download from NEON (test_check_neon_time, test_get_neon, test_get_neon_false_site, test_check_neon_listing, test_check_neon_listing_misspelled).

Requirements before merge:

• I have followed the CTSM contribution guidelines.
• The code in this PR branch builds with no errors. Not applicable: Python test runner and CI only; CTSM itself was not built.
• The code in this PR branch runs with no errors. Not ticked: one test fails until test(subset_data): make unit test work outside NCAR cluster, like its siblings do #4231 is merged (noted above). Briefly describe tested configuration(s): macOS arm64, pip venv as above.
• This either (a) does not change answers, (b) it only changes answers at roundoff level, or (c) I have performed a scientific evaluation of the answer changes. Which?: (a)
• I have reviewed relevant parts of the CLM documentation Tech Note or User's Guide to determine if anything needs to be changed or added. If it does, describe: Nothing under doc/ covers the Python test runner or this job.
• This PR either (a) does not create a need to update the documentation or (b) includes required documentation updates (see guidelines for contributing documentation). Which?: (a)

[ekluzek]

LGTM

[samsrabin]

@Santoshkumarpuppala Thanks for your contribution! Would you mind emailing ctsm-support@ucar.edu so we can discuss possible future contributions?