By using the same exit code in all my tests, I didn't catch that the internal state of the exit evaluator is not being reset.
By using the same exit code in all my tests, I didn't catch that the internal state of the exit evaluator is not being reset.