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