hi, after some thought i would like to discuss adding a generic testdir, so for each test different plugins could store data/metadata
currently only the tempdir plugin creates a session dir on first use of the tmpdir funcarg/pytest.mktemp i'd rather have to possibility to store for each test * full test id * stdout/err * logfile for stdlib logging * tmpdir for plugin * some kind of serialized full taceback, optionally with some locals/globals * coverage reports * profiler reports since all of those are a misfit for the tmpdir functionality, i'd like to propose a extension of the mechanism the basic tmpdir would move to the "tmpdir" subdir of the testdir, stdout/err, log, would in turn be files in the testdir it would also be easily possible to keep those in sync with xdist+rsync should the need arise, just make a new testdir for the remote id, and sync the directory content. further possible uses include using coverage data and test results of previous runs to decide what tests can likely be skipped without causing too much of a problem or providing relative progress information on the collection status (based on the number of test id's in the last run). Please discuss additional uses and possible problems. Regards Ronny
signature.asc
Description: This is a digitally signed message part
_______________________________________________ py-dev mailing list py-dev@codespeak.net http://codespeak.net/mailman/listinfo/py-dev
