[GHC] #12670: Representation polymorphism validity check is too strict

[GHC]

#12670: Representation polymorphism validity check is too strict
-------------------------------------+-------------------------------------
           Reporter:  bgamari        |             Owner:
               Type:  bug            |            Status:  new
           Priority:  normal         |         Milestone:  8.2.1
          Component:  Compiler       |           Version:  8.0.1
  (Type checker)                     |
           Keywords:  typeable       |  Operating System:  Unknown/Multiple
       Architecture:                 |   Type of failure:  None/Unknown
  Unknown/Multiple                   |
          Test Case:                 |        Blocked By:
           Blocking:                 |   Related Tickets:
Differential Rev(s):                 |         Wiki Page:
-------------------------------------+-------------------------------------
 As of 465c6c5d15f8fb54afb78408f3a79e75e74d2cd4 GHC rejects this program,

 {{{#!hs
 {-# LANGUAGE GADTs #-}
 {-# LANGUAGE TypeInType #-}
 {-# LANGUAGE RankNTypes #-}
 module Hi where
 import GHC.Exts
 data TypeRep (a :: k) where
     TrFun :: forall (r1 :: RuntimeRep) (r2 :: RuntimeRep) (a :: TYPE r1)
 (b :: TYPE r2).
              TypeRep a
           -> TypeRep b
           -> TypeRep (a -> b)
 }}}

 Claiming that it has inappropriate polymorphism over `RuntimeRep`,
 {{{
 $ ghc hi2.hs
 [1 of 1] Compiling Hi               ( hi2.hs, hi2.o )

 hi2.hs:7:5: error:
     • A representation-polymorphic type is not allowed here:
         Type: a
         Kind: TYPE r1
     • In the definition of data constructor ‘TrFun’
       In the data type declaration for ‘TypeRep’
 }}}

 I suspect this check is unnecessarily strict since we never need to store
 a value of type `a` to represent `TypeRep a` at runtime.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12670>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler