Re: [Haskell-cafe] Rigid type-var unification failure in existentials used with parametrically polymorphic functions