I'm currently reading the Testing Swift book, and so far I haven't seen any mention in using random values within Unit Tests. I can see that doing so can (potentially) break the "R" (repeatable) of "FIRST", but at the same time I see so much value in doing such tests, e.g.:

struct Superhero {
      name: String

      func isFemale() -> String {
            return name.contains("woman")
      }
}

class ConverterTests: XCTestCase {
      func test_superheroIsNotFemale_whenItsNameDoesNotEndWithWoman() {
            let sut = Superhero(name: "\(RandomString)man"

            let isFemale = sut.isFemale()

            XCTAssertFalse(isFemale)
      }
}

Ok, now this is quite a silly example, but it helps illustrait my point. Clearly the implementation of isFemale has a bug, if we ever come across with a superhero named "Womaniterman", it will say he is a Female, but clearly he is not. Using random strings in the unit test this could eventually be detected (yes... probably not cause the chances to get the word "woman" out of a random string are super low, but you get my point).

So, how do you guys feel about using random values in unit tests? I know they can not be used in all situation, but should they be completelly forbbiden, or do you also see value in using them?