If Rust is okay with it, the math must be correct. Perhaps we've entered another dimension?

Apparently back then if I'm understanding correctly, the kata expected "int[]" to be returned instead of "HashMap<String, int>" despite saying that it needs to be an HashMap. So now, it's fixed but there's still backwards compatibility for "int[]" solutions.

Worth noting that this takes O(n^2) time, when O(n) is possible.

Yeah, mine asked for a HashMap too. And we have the same solutions. I would say we are good. :)

This is weird.. My kata asked for a HashMap as return value.
Maybe I'm watching the solutions of all forks? Anyone knows?

But this solution does not account for negative values... 🤔

I really like this solution

I like how much faster the strconv.Itoa operation is than fmt.Sprint (according to benchmark at https://gist.github.com/evalphobia/caee1602969a640a4530)