Taral wrote: > Ah, the dreaded $ with existential types problem. $ is not quite > equivalent to application -- the type checker does something funny > with forall types. Just take out the $ and you'll be fine. > Is this a ghc bug, or some subtlety of the type system that I don't understand?