Terry Reedy writes:
I agree that complete 100.000% test coverage is a nice ideal, but sometimes the last percent can take hours to accomplish, if it is indeed sensibly possible.
100% test coverage is an ideal. Reports *claiming* 100% coverage, however, are of practical benefit. The point is to identify a regression. It's best to have 100% coverage, because it's possible that improvements in the environment allow a test that wasn't reliable (== deterministic) or maybe not even feasible before to become feasible and reliable, and a coveragerc that says "oh, that line is OK" will obscure that possibility. But a *claim* (albeit somewhat undermined by a non-trivial coveragerc) of 100% coverage means it's *easy to identify regressions in coverage*. I think that's a bigger deal, at least at this time.
I have asked more that once what .coveragerc file is being used by CI and whether we can somehow properly customize it for CPython.
Seconded. Thank you for pushing this, and for all the other efforts you're making here. Steve