Is there a way to mark a test as using external APIs/outside of pants information, so it's cached result can't be trusted? Essentially I want to never cache the execution result of a test, but still cache all the pytest/venv builds for it

Not currently but it should be straightforward to implement. The Pants backends which implement the

test

goal know how to apply the --test-force option by modifying the "cache scope" of a test execution. So the work to implement this idea would mainly be in deciding the UX for it and then just threading that value to the places setting cache scope.

Actually for my usecase it's not that big of a headache to have a separate

pants test

call for this test (since you need to setup these external dependency access anyway) so the goal level setting works for me. Thanks!

Just a note, I think caching for pytest is a file-level cache (but I don't recall if there are plugins to do it at the test level)

Actually for my usecase it's not that big of a headache to have a separate

pants test

call for this test (since you need to setup these external dependency access anyway) so the goal level setting works for me. Thanks!

Even simpler then!