Re: [Haskell-cafe] Re: Existentially-quantified constructors: Hugs is fine, GHC is not?