OCaml translation

Add the reference bad solutions you have to JS/Python tests and sure.

Yep, or just above the random test parameters with a note saying that it should always fail. I published an ocaml fork as an example of how to do that.

I would suggest including them within the tests somewhere for translators to access.

Can I get an example of the exact solution you dont want to pass? My implementations of naive brute force were failing even length 100.

done

Because it won't lex correctly otherwise. Fork the translation and try it.

So your rationale is that you wanna make the test output less readable for no reason?

Your current random spray doesn't really produce any actually interesting cases fwiw. Targeted generation is definitely a better idea (and you'd be able to ensure you're getting 50/50 true/false generation).

I'll change OCaml once you've explained how exactly you want your tests generated and the rationale behind it as well as expected time complexity, because right now I'm somewhat unconvinced the random test generation produces much edge case coverage.

Temporary issue to prevent approval whilst requirements are standardised across languages.

You can approve the fork (has your new specs as well), but why the change in requirements? It doesn't add any extra performance constraints and makes the messages even more unwieldy to debug.

Also, do you want it to be explicitly strings for all languages? I didn't use them since they're less nice to destructure.

Scala translation

fixed (probably)

Python random test generation is insufficient to prevent incorrect solutions passing.