[GHC] #9123: Need for higher kinded roles

[GHC]

#9123: Need for higher kinded roles
-------------------------------------+------------------------------------
        Reporter:  simonpj           |            Owner:
            Type:  bug               |           Status:  new
        Priority:  normal            |        Milestone:
       Component:  Compiler          |          Version:  7.8.2
      Resolution:                    |         Keywords:
Operating System:  Unknown/Multiple  |     Architecture:  Unknown/Multiple
 Type of failure:  None/Unknown      |       Difficulty:  Unknown
       Test Case:                    |       Blocked By:
        Blocking:                    |  Related Tickets:
-------------------------------------+------------------------------------

Comment (by goldfire):

 Now, the question is whether we can derive `Coercible (s -> p a s) (s -> p
 b s)` from `(Rep p, Coercible a b)`, right? Once again, we assume that the
 only way we can use `Rep` is the rule (*) in comment:15.

 1. Decompose `(->)` to get that we need to show `Coercible (p a s) (p b
 s)`.

 2. Use the eta rule requested at the top of #9117 to reduce the goal to
 `Coercible (p a) (p b)`.

 3. Use rule (*) to reduce the goal to `(Rep p, Coercible a b)`.

 4. We are done by assumption.

 This was indeed somewhat harder than the derivation in comment:15 in that
 it requires the constraint solver to do something it currently cannot (the
 eta reduction in #9117), but it still doesn't require an asymmetrical rule
 for `Rep`. Note that the eta rule is easily expressible in Core -- it's
 just that the constraint solver doesn't know to use it.

 Does this address your concern, Edward?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9123#comment:17>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler