Note that this probably has nlogn rather than logn performance, as in Haskell's collections. It can probably be improved very significantly. Would like to add some benchmarks (https://github.com/purescript/purescript-maps/issues/101) before doing so.
I think this is good, but we should probably try to implement it by walking the tree, which would be more efficient. This is great for a first pass though, thanks!
addresses https://github.com/purescript/purescript-maps/issues/26
Note that this probably has nlogn rather than logn performance, as in Haskell's collections. It can probably be improved very significantly. Would like to add some benchmarks (https://github.com/purescript/purescript-maps/issues/101) before doing so.