[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 (Type    |              Version:  8.0.1
  checker)                           |
      Resolution:                    |             Keywords:  typeable
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 simonpj):

 Very good point. We can't make `Dyn res (f x)` unless `f x :: Type`, I
 agree.

 Let's see what Richard has to say.

 What is `Fun`?  You're referring to some code I can't see.

 In any case, I'd remark that we don't necessarily need the full generality
 of `typeRepKind`; we only need to know the arguments to `(->)` when we
 take `rx` apart.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12670#comment:13>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler