Opened 4 years ago
Last modified 4 years ago
#32609 closed Cleanup/optimization
runtests.py setup should use DiscoverRunner's test label logic — at Version 1
| Reported by: | Chris Jerdonek | Owned by: | nobody |
|---|---|---|---|
| Component: | Testing framework | Version: | dev |
| Severity: | Normal | Keywords: | DiscoverRunner, runtests |
| Cc: | Triage Stage: | Ready for checkin | |
| Has patch: | yes | Needs documentation: | no |
| Needs tests: | no | Patch needs improvement: | no |
| Easy pickings: | no | UI/UX: | no |
Description (last modified by )
I noticed that runtests.py does its own rudimentary "parsing" of the provided test labels here:
https://github.com/django/django/blob/548dce50cf548e777a0c34b5485a146a0606ae73/tests/runtests.py#L128-L132
However, it would be better if it used the same logic as DiscoverRunner.build_suite():
https://github.com/django/django/blob/548dce50cf548e777a0c34b5485a146a0606ae73/django/test/runner.py#L612
There are a few reasons for this. First, it would let runtests take into account the test tags when determining which app modules apply. Second, it would centralize the test label logic, which should simplify maintenance. (For example, I was previously unaware of this code path, which explains why some things mysteriously weren't working before.) Third, it might even permit test labels to be directory paths when used with runtests.py, as a free side effect. (Currently, directory paths don't seem to work with runtests.py, I believe for this reason.) A fourth is that runtests.py will fail faster if a bad test label is passed.