#14155: GHC mentions unlifted types out of the blue (to me anyway) -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: (none) Type: bug | Status: new Priority: highest | Milestone: 8.4.1 Component: Compiler | Version: 8.2.1 Resolution: | Keywords: TypeInType Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Changes (by simonpj): * keywords: => TypeInType * priority: normal => highest * milestone: => 8.4.1 Comment: Aha. Based on comment:1 I very rapidly found that the constraint solver was going wrong thus. We have {{{ m :: k0 -> k0 -> * -> * i :: k0 }}} and a wanted constraint {{{ [W] (m i :: k0 -> * -> *) ~# ((->) LR LR :: * -> * -> *) }}} where `LR` is `LiftedRep`. This much is fine. But then the canonicaliser decomposes it thus {{{ [W] (i :: k0) ~# (LR :: *) [W] (m :: k0 -> k0 -> * -> *) ~# ((->) LR :: forall r. * -> TYPE r -> *) }}} which is utterly bogus. Presumably we need an extra guard in the `can_eq_app` that checks that neither function kind is a forall? Richard please comment. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14155#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler