Failure marker by arihant2math · Pull Request #5695 · RustPython/RustPython

How do you run this code and what does it output?

Currently I need to patch it a little ... but the idea is that the workflow is the following:

1. Make some changes to Lib/ and Lib/test/
2. Find which tests are failing
3. Run python ./scripts/fix_test.py --test test_venv --path ./Lib/test/test_venv.py or equivalent for the test from the project root.
4. Actually fix the test.

This works best for platform-independent tests.

The limitations are:

1. Can't handle unexpected successes (I could fix this, but this is rare enough, I think it's fine
2. Can't handle entire test file crashes (this is another thing the user has to handle themselves)
3. Can't handle subclasses and nested classes etc. (I'm working on a fix for this though)

Having a working --junit-xml would help with finding failing tests, instead of parsing stdout.
Because some tests like printing out to stdout, ruining the parser.

#4072 is the blocking issue.