Use WalkingDistance as cost/heuristic

[ephphatha]

Since diagonal steps are the same framecount as cardinal steps there's no reason to treat them any differently. Lets us simplify the cost and heuristic functions.

This isn't a proper fix for #2404, just something I was playing around with while finding test cases for glebm's pathfinding refactor. It doesn't appear to be any worse than vanilla's logic and lets max length paths occur more often so might be useful as a temp measure?