[GHC] #10343: Make Typeable track kind information better

[GHC]

#10343: Make Typeable track kind information better
-------------------------------------+-------------------------------------
        Reporter:  oerjan            |                   Owner:
            Type:  feature request   |                  Status:  new
        Priority:  normal            |               Milestone:
       Component:  Compiler          |                 Version:  7.10.1
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  None/Unknown      |  Unknown/Multiple
      Blocked By:                    |               Test Case:  typeOf ::
 Related Tickets:  #9858             |  Typeable (a::k) => Proxy a ->
                                     |  TypeRep
                                     |                Blocking:
                                     |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by oerjan):

 I am not sure I can be called a client, I don't have actual useful code
 hinging on this.  As a hobbyist, I spend more time ''thinking'' about
 Haskell than actually programming in it.

 I think that simply adding `kindRep :: TypeRep -> TypeRep` as a field of
 `TypeRep` would make the necessary information available at runtime.  This
 ought to fit just as well into the current design as into the proposed
 `TTypeRep` on the wiki.  I also have a hunch this might make the already
 existing list of `KindRep`s redundant.

 The question of how to make the `Typeable` solver make use of that
 information for building kind evidence seems to me like the more
 complicated part of this.

 Just to prove it is possible in principle, I'll restate my mock
 representation from back in [comment:48:ticket:9858]:

 {{{
 class (KindableParts a, a ~ Proxy) => Kindable (a :: k -> *) where
     type KindableParts a :: Constraint
     kindRep :: KindRep a

 class Kindable (Proxy :: k -> *) => Typeable (a :: k) where
     typeRep :: proxy a -> TypeRep
 }}}

 This worked well enough to be a basis for
 [http://oerjan.nvg.org/haskell/TypeableDesign/ running my kitchen sink
 test case] in 7.8.3.  However, it doesn't support Richard's ideas for
 avoiding exposing "`Kindable`" constraints when doing type inference.  And
 of course the representation's made for reusing already existing type
 features rather than for efficiency.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10343#comment:7>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler