Lesson learned: vm.runInNewContext is slow in js. :) Wonder if it was the context switches.

Might be worth updating the node version here or explicitly marking it in the task description. Node 0.10 is considered very out of date by now.

Thanks for re-enabling require and sorry this turned to be a source of controversy.

I think the question to ask when disabling things is if helps with what the Kata is trying to teach? Sometimes you do want to teach people how the tools they're using under the hood work and these make for great taske.

In this case however, the task is not made easier as there's no library for doing exhaustive search. Using libraries here however allows users to write cleaner code, better reflecting what they might write the real world.

Anyways, just my opinion. YMMV. Thanks again for working on this task! :)

Great task, had fun solving it!

Some notes:

• For whatever reason requiring outside javascript modules fails with "TypeError: require is not a function" for this particular task. Using lodash would have removed some utils I needed to create.
• A rule that will not be obvious for people who have not played the game: will falling lower than height 1 end the game. From experimentation with the task, the answer is yes.

Note: You can set maxsize to None to have the cache be unlimited.

Fun task, thanks.

A single nitpick would be that the description could be improved: I understood the task that in addition to doing what needed to be done, each instance should have a unique_class_method and unique_instance_method methods which return the real (unique) methods accordingly.

That's an artifact of me copying and pasting. Since I passed all the other tests that submission and all of them on re-submitting I don't think that's the reason it was rejected.

Looking at the expected solution, hate and have positions are reversed from what I take they should be based on my guesswork of what rule 2 means. I'm thinking there might be a subtle bug in the reference solution, worth checking?

Perfect, thanks!

Re 3rd: is this a valid puzzle (spaces denoted with _)?

__foobar
__o
too

Note the missing whitespace on the right-hand side for 2nd and 3rd row.

Couple of things that might be useful to add to the description:

• Maximum dimensions of puzzle
• If words are guaranteed to be unique
• If each row of puzzle of equal length. Some tasks strip extra whitespace which makes for a nasty surprise.

This test is still there and buggy as from this board state X can only win.

Really curious about what people think of this solution. The functools library is fantastic for solving these sorts of tasks efficiently in python 3.