Code review comment for lp://staging/~chris.macnaughton/charm-helpers/pids-can-be-a-list

Thanks for the unit test coverage on this. That shows that the behavior is as intended. If we are to omit a sync/amulet test, we should also add negative unit tests here, where you pass bad scenarios and assert that the validation method has returned something other than None.

Can you add 1-per? That would make this rock solid.