containAllInAnyOrder fails to detect mismatch in element counts, as long as the overall counts of elements match

[AlexCue987]

latest from master
Presumably containAllInAnyOrder should detect mismatch in element counts, and sometimes it does, as follows:

        "detect count mismatch" {
            sequenceOf("apple", "apple", "orange") should containAllInAnyOrder(
                sequenceOf("apple", "apple", "orange", "orange")
            )
        }
-----------
Sequence should contain the values of [apple, apple, orange, orange] in any order, but was [apple, apple, orange]

However, the following test passes, which seems counter-intuitive to me:

        "fail to detect count mismatch" {
            sequenceOf("apple", "apple", "orange") should containAllInAnyOrder(
                sequenceOf("apple", "orange", "orange")
            )
        }

Interestingly, when I tried to fix the issue by comparing counts of elements, it broke two existing tests, as follows:

        succeed("for subset, same count with nulls") {
            sampleData.sparse.shouldContainAllInAnyOrder(sampleData.nulls)
         }
----
Sequence should contain the values of [null, null, null, null] in any order, but was [null, null, null, 3].
----
         fail("for subset, same count with nulls") {
            sampleData.sparse.shouldNotContainAllInAnyOrder(sampleData.nulls)
         }
------
Expected exception java.lang.AssertionError but no exception was thrown.
// this test matches [null, null, null, null] with [null, null, null, 3] - these seems different to me and as such the test should pass

My understanding is that these two tests are not validating correct behavior, and as such they should be fixed.
My understanding is that these two tests should not be passing in the first place.
If this is correct, I have a fix and can raise a PR.

[LeoColman]

I believe you're right. Feel free to submit a PR :)

[AlexCue987]

I believe you're right. Feel free to submit a PR :)

there you go:

#3907