Closed jgfouca closed 9 months ago
I feel like it really should not be there - shouldn't we add a call to case.setup to the tests in question?
case.setup should be called in the test, for example, here
@jedwards4b , there are a few approaches here. We should seek a solution that minimizes the chances of a mistake for developers working on SystemTests.
I'm wondering why system_tests_compare_two does not call case_setup for case2 in _setup_cases
:
# This assures that case one namelists are populated
# and creates the case.test script
self._case.case_setup(test_mode=False, reset=True)
fix_single_exe_case(self._case)
# Set up case 2
with self._case2:
self._activate_case2()
self._common_setup()
self._case_two_setup()
It does not seem like good design for developers to have to remember to call case_setup in case_two_setup.
I would be fine with that change.
@jedwards4b , I pushed another commit and I'm much happier now. I think this is the best solution.
It's removal in #4546 seems to have broken a number of less-common test types (ERR, MCC, IRT, PRE).
Test suite: the test cases above, by hand Test baseline: Test namelist changes: Test status: bit for bit
Fixes [CIME Github issue #]
User interface changes?:
Update gh-pages html (Y/N)?: