Haskell version is broken - there's a case where n=99999 (n is supposed to be 20000 max).

Corrected.
Thank you for your feedback ;-)

I think there's a bug with the random tests in Haskell. The expected value and actual value that it gives seem to be flipped.

My solution generates SKI from lambda as an EDSL so it works :)

https://www.codewars.com/kata/reviews/5a03e576882dca839200284e/groups/5b7fc96437a27f177800180e

There's a link to an explanation of the trick. I'm not copying it here, so that this comment is spoiler-free.

didn't do those katas

But they're fun! :]

new four

"one" is approved, four has description conflicts. I cannot approve the others, sorry (didn't do those katas)

OK, can I hit you up for these in exchange? :]

• one
• two
• three
• four
• five

You're about four years late. I don't think this kata is going to be changed anymore.

OK, so it was fixed before it was raised, it just had to be approved.

Closing.

yes, I know. I forked the translation, it's awaiting for an approval (description conflicts... :/)

Python version:

A sort order is not specified, but one is required by the sample tests because Test.assert_equals is used.

Possible fixes:

• Specify the sort order. (It appears to want what sorted(parens, reverse=True) would return.)
• Sort the answer and the solution and compare.
• Compare element by element.
• Use sets.

I liked the problem, but it should require a DP solution IMO.