Re: [Haskell-cafe] Confusing type specialisation in ghci