I just noticed that the test runner does not actually fail on any failures.
I just noticed that the test runner does not actually fail on any failures.