array: generalize S.take and S.drop

[davidchambers]

The code is more complex than I would like it to be, but I believe the complexity is necessary. Please correct me if I'm wrong about this! :)

[Bradcomp]

I think it looks OK. It is a little hard to follow what is going on, but once I sat with it a bit I got what was going on.

I think it should be possible to do it in one pass, but right now I am struggling to figure out how, so I could be wrong.

[davidchambers]

I think it should be possible to do it in one pass

Could you elaborate, Brad? I believe we're traversing the structure just once, with Z.reduce.

[Bradcomp]

Ugh, I am tired today, but I see what is happening now. reduce is returning a Just Pair (Int (f a)), and we are rejecting if the int is > 0 because it means the foldable wasn't long enough. Otherwise we take the resultant array.

I had things inside out in my thinking.

Looks good to me!

[masaeedu]

Hi @davidchambers. I talked to @Bradcomp for a bit about this PR, I think maybe we don't really need the Applicative constraint on here.

[davidchambers]

The Applicative constraint is not obvious. It does not arise from _takeDrop itself, but from the generalCase functions provided to _takeDrop in the definitions of take and drop:

var take = _takeDrop (
  function(n, xs) { return xs.slice (0, n); },
  function(n, xs, x) { return n > 0 ? Z.append (x, xs) : xs; }
  //                                  ^^^^^^^^^^^^^^^^
);

var drop = _takeDrop (
  function(n, xs) { return xs.slice (n); },
  function(n, xs, x) { return n > 0 ? xs : Z.append (x, xs); }
  //                                       ^^^^^^^^^^^^^^^^
);