Oh, another one: the byte length of my strings is the same as the character length of my strings. That one is so nasty, especially in a UTF-8 setting where if you naively test with ASCII strings the tests will all be written with that assumption in mind.