pg_regress allows controlling amount of test parallelism by separating groups of tests
e.g. a schedule such as
test: t0
test: t1 t2
test: t3 t4 t5
would run 3 groups of parallel tests where each group runs in isolation (no other group runs when one group runs)
Our bash scripts currently do not handle this well and collapse the schedule above into a single group of 6 tests.
This needs to be addressed. Particularly important for parallel and update tests as we want to run various test batches in distinct groups in there
pg_regress allows controlling amount of test parallelism by separating groups of tests
e.g. a schedule such as
would run 3 groups of parallel tests where each group runs in isolation (no other group runs when one group runs) Our bash scripts currently do not handle this well and collapse the schedule above into a single group of 6 tests.
This needs to be addressed. Particularly important for parallel and update tests as we want to run various test batches in distinct groups in there