Thanks! Though I’m not sure we want to be bring the “old” behavior back.
In all test runners I’ve ever used, running cool-test-runner path/to/some.file
only runs the tests in that file, which is nice when debugging a failing tests for a certain module. If we also run tests in files that are import
ed that command could run a lot more tests than one would expect.
I have just looked at this quickly, but the impression I got from the PRs/issues leading up to the revision3 release was that 0.19.1 and 0.19.1-revision2 were regressions from 0.18 – the run tests via import
s (if that’s even how it works, I haven’t dug into it) was an accident and never how it was supposed to work.
I need to dig deeper to know for sure, though.