Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes: > Partly this is cleaning up a mess after an existing test, but here > there's no reason we can't use test_when_finished() for all the new > tests to make them clean up after themselves: I do not mind if somebody wants to send in a janitorial patch after the dust settles, but adding "test_when_finished reset --hard" after each "refs --hard" at the beginning of each test is not something I would expect to see. Such a patch should first choose between "each test cleans after itself" and "expect previous ones may have left a mess, so each test clears the slate sufficiently before it starts" and then stick to the approach, not mixture of both, I would think. Thanks.
