Deal with half-baked effects: Coroutine, CutFalse and NonDet

[trskop]

IMHO, these effects aren't production ready. This was decided on today's session where @liskin, @xkollar, @Siprj, and I participated.

[trskop]

Alternative to removing an effect is to find a more practical definition. For example Coroutine could be useful, but its current definition is a proof of concept, and not a practical effect.

[xkollar]

Also NonDet does not feel right...

[trskop]

We should look into extensible-effects, they seem to have figured some things out already.

[trskop]

I think that PR #25 made Coroutines much more useful for real applications.

[spacekitteh]

@xkollar how does NonDet not feel right?

[xkollar]

@spacekitteh: I would expect something with that name to behave more like list. Or do I miss something obvious?

[spacekitteh]

It's an implementation of a non-biased nondeterminism monad. See here.