This issue serves to track CI test cases we should include to seml.
[ ] Source Code Stashing: Test that the experiment actually uses the stashed source code. Proposal: Add experiments, then change the source code to include a raise RuntimeError(...) statement, then run the code and check that all experiments are COMPLETED eventually.
[ ] Test that experiments_per_job actually runs code in parallel: Devise a python script with wait times as dummy and ensure that they start roughly synchronised by checking their outputs.
[ ] Ensure that the src layout is correctly supported by source code stashing (how?)
Add further test cases when you encounter weird behaviour that needs to be fixed.
This issue serves to track CI test cases we should include to seml.
raise RuntimeError(...)statement, then run the code and check that all experiments areCOMPLETEDeventually.experiments_per_jobactually runs code in parallel: Devise a python script with wait times as dummy and ensure that they start roughly synchronised by checking their outputs.Add further test cases when you encounter weird behaviour that needs to be fixed.