#9725: Constraint deduction failure -------------------------------------+------------------------------------- Reporter: heisenbug | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.8.3 (Type checker) | Keywords: Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: GHC | Related Tickets: rejects valid program | Test Case: | Blocking: | Differential Revisions: | -------------------------------------+------------------------------------- Comment (by simonpj): No it's not just kind inference. Do you want that poly-kinded type for `test`? Sounds as if the answer is "yes". Well, then `k` is a skolem variable, and if it is is locally known to be equal to `Bool`, that's a local kind equality (a la GADTs). From what you say, in some branches of `h`, you have a local equality `k ~ Bool` and in others you have `k ~ Nat`. This is just what Richard is working on! Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/9725#comment:8 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler