The examples in the kata description are misleading because they only contain a very limited subset of test cases (where text is either only added or only deleted, not both). You should add some human readable examples with text both added and deleted to clearly show what you are asking for. For example something like: findDifferent("Hello John, how are you?", "Hello Jane, how are you?"). I would also suggest adding a few reasonable human readable strings to the test cases rather than using only purely random strings or concatenations of single words.

Riddle me this random test:

expected        { index: 1
                , addedText:  ' '
                , deletedText: ''
                }
to deeply equal { index: 1
                , deletedText: '22222222222222222'
                , addedText:  ' 22222222222222222'
                }

Why does the reference solution insist on pointlessly removing and adding all those 2s? That's not "the most efficient, succinct way".

It is not explained how to handle cases with multiple possbile answers.

{index: -1, addedText: "", deletedText: ""} makes no sense. If nothing was changed, the output should be null.

• still no random tests
• B1ts's solution still passing while still using a wrong logic => the edge case I was asking for in the previous version is still lacking
• you forgot to remove the wrong test from the description

edit: btw, ti seems you ddidn't get how the ranks are working on cw: 1 kyu isn't the easiest difficulty, it's the hardest.

hi,

• edge cases are missing:
  • this solution should never have passed
  • case with no changes needs to be tested AND described
• need sample tests (not a suggestion, an issue!)

No random tests.

( A gazillion fixed tests is a worthy effort, but not a replacement for random tests. )

What would be the expected return value for identical strings?

How do you expect ambiguities to be resolved?

In general, do not rely on examples - specify that sh*t.