The rewritten test is not super clean with the manual cleanup etc.. If this is a
one-off it's not a big deal. However, if in subsequent rewrites I spot more
occurrences of this pattern, then I'll refactor.