Explanation of tests in anagram

[erip]

I'm working on the anagram problem and my code is failing on these three tests:

#[test]
fn test_does_not_detect_a_word_as_its_own_anagram() {
    let inputs = ["banana"];
    let outputs: Vec<&str> = vec![];
    assert_eq!(anagram::anagrams_for("banana", &inputs), outputs);
}

#[test]
fn test_does_not_detect_a_differently_cased_word_as_its_own_anagram() {
    let inputs = ["bAnana"];
    let outputs: Vec<&str> = vec![];
    assert_eq!(anagram::anagrams_for("banana", &inputs), outputs);
}

#[test]
fn test_does_not_detect_a_differently_cased_unicode_word_as_its_own_anagram() {
    let inputs = ["ΑΒγ"];
    let outputs: Vec<&str> = vec![];
    assert_eq!(anagram::anagrams_for("ΑΒΓ", &inputs), outputs);
}

I think these tests should actually be panic because these assertions are not true if considering case insensitive code for the latter two and simply because there's no output for the trivial case of a self-describing anagram.

Perhaps I'm missing something, but I was hoping these could be addressed somehow.

[petertseng]

Hmm, I think the main question here is whether a word is its own anagram. We can ask whether the no-op rearrangement is a rearrangement. In mathematics, the identity permutation is a permutation, but that is not the domain we're working in. I definitely see there could be arguments for either way, but at some point Exercism had to decide, and we made the choice that a word isn't its own anagram.

This is in agreement with the common test inputs/outputs at https://github.com/exercism/x-common/blob/master/anagram.json that all Exercism tracks are using.

So that explains why test_does_not_detect_a_word_as_its_own_anagram expects the output to be the empty vector. And since we know from an earlier test that the anagram check should be case-insensitive, the last two tests are just a combination of the "word isn't own anagram" and case insensitivity tests.

[erip]

@petertseng Excellent explanation. After some sleep it's becoming clearer why a word isn't its own anagram. I'll close this. Thanks!

[BartMassey]

@petertseng Both the case-insensitivity of anagrams and the rule that a word is not its own anagram should be mentioned in the problem description. The tests are fine: the description is missing important detail.

[petertseng]

You are correct that the problem description https://github.com/exercism/problem-specifications/blob/master/exercises/anagram/description.md does not say either of those things, and I agree with you that it would be an improvement (clearer to the student solving the problem) if it would say those two things.