RE: [Haskell] A question about a possible bug in GHC regarding GADTs and type classes