No random tests

The tests have a typo in them ( 9e19 /= 9Em ), the reference solution is bugged and doesn't catch the typo, the tests are inadequate and don't catch the reference solution. Boy, what a mess.

Was confused for a second reading your function definitions... specifically the first arguments to encode and decode. Thought it was some fancy pattern matching because of the as syntax in defining the functions. :'(

Upvoted but it's probably best practice to avoid overshadowing keywords in function arguments in general :')

Hm, that sounds like a bug. You probably want to report it.

I seem to be unable to publish the one labelled as haskell as it is a fork of the "coffeescript" one, and it complains that the solution code is not unique from parent. I'll try some renaming to see what I can do. Also, the unpublish button does nothing.

It seems that you have two translations: one that's published (and mistakenly filed as CoffeeScript), and one that's still a draft (and correctly filed as Haskell).

5 is only in the name, and I think it was due to a connection problem while trying to save it. When clicking "add translation", Haskell is now unavailable, but CoffeeScript is available.

Hm, that's strange. On the left side, I can see that you have a Haskell translation draft. Did you check your drafts?

• 1 and 3 were leftovers from me trying to debug the quickcheck (error was that I used == instead of ===); that's fixed in the most recent fork
• 2 and 4 are now fixed.
• 5 is only in the name, and I think it was due to a connection problem while trying to save it. When clicking "add translation", Haskell is now unavailable, but CoffeeScript is available.

I did indeed test it, though there were indeed a few mistakes based on my inexperience with QuickCheck. I've responded further to your comment on the translation directly. No clue why it filed as coffeescript, though! I had some server connection issues while trying to save it the first time, which will not have helped.

• unecessary return () at the end of hspec
• missing describe "meters" around it "..."
• no need for {-# LANGUAGE ScopedTypeVariables #-} in the test
• importing all functions from Metric (which can lead to unexpected clashes and frustration
• filed as CoffeeScript translation

You didn't event test your translation (on Codewars). It's filed as a CoffeeScript one!

Translated into Haskell, if you would like to approve it

How should the state of (x=3) + (x=2) look after evaluation? {'x':2}? Or just left as undefined behaviour as it is in C?

Unknown error is normally caused by a website error, rather than an error with your code found by the test suite. Try resubmitting!