#14331: Overzealous free-floating kind check causes deriving clause to be rejected -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: merge Priority: normal | Milestone: 8.2.2 Component: Compiler (Type | Version: 8.2.1 checker) | Resolution: | Keywords: deriving Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: GHC rejects | Test Case: valid program | deriving/should_compile/T14331 Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by goldfire): But we can only get `ty` after dropping `n` arguments. Relating this to the existing code in `deriveTyData`, what I've called `ki2` is `cls_arg_kind` and what I've called `n` is `n_args_to_drop`. `n_args_to_drop` is computed by counting the number of args spun off by a call to `splitFunTys cls_arg_kind`. For a unification variable `kappa`, `splitFunTys` will return no arguments -- just like it would when `cls_arg_kind` is `Type`. So GHC currently behaves like my algorithm; I just have more checks here. However, calling `splitFunTys` on a unification variable is bogus, because that variable could potentially stand for anything, including a function. This bogus use happens to work out, but that's more luck than science in this case, I think. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14331#comment:35 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler