Re: [Haskell-cafe] A type signature inferred by GHCi that is rejected when written explicitly