[GHC] #11471: Kind polymorphism and unboxed types: bad things are happening

[GHC]

#11471: Kind polymorphism and unboxed types: bad things are happening
-------------------------------------+-------------------------------------
        Reporter:  bgamari           |                Owner:
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  7.10.3
  checker)                           |             Keywords:  TypeInType,
      Resolution:                    |  LevityPolymorphism
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by goldfire):

 Re comment:2: Try with `-fprint-explicit-coercions`. Here's what I think
 is happening, though I don't have a fresh enough build locally to check:

 `foo :: Proxy a -> F a -> F a` becomes `foo :: forall k2 (a :: *). Proxy *
 a -> F k2 a -> F k2 a`. Note that `F` is not polymorphic in the kind of
 `a`, which defaults to `*`. We also know that `k2` must be either `*` or
 `#`, because `F k2 a :: k2` is an argument of an arrow. GHC wisely
 defaults it to `*`, leading to the second error.

 Once we clean up kind errors with coercions, I think the error messages
 will improve. There's also clearly the word `kind` appearing where `type`
 should in the first message. But these errors are legitimate.

 As for the larger issue, we'll need to discuss.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11471#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler