[GHC] #8631: Need ImpredicativeTypes for GeneralizedNewtypeDeriving?

[GHC]

#8631: Need ImpredicativeTypes for GeneralizedNewtypeDeriving?
-------------------------------------+------------------------------------
        Reporter:  goldfire          |            Owner:
            Type:  bug               |           Status:  new
        Priority:  normal            |        Milestone:
       Component:  Compiler          |          Version:  7.7
      Resolution:                    |         Keywords:
Operating System:  Unknown/Multiple  |     Architecture:  Unknown/Multiple
 Type of failure:  None/Unknown      |       Difficulty:  Unknown
       Test Case:                    |       Blocked By:
        Blocking:                    |  Related Tickets:
-------------------------------------+------------------------------------
Changes (by simonpj):

 * cc: sweirich@… (added)


Comment:

 We really want explicit type application here (#4466, #5296). The point is
 that we want to say "instantiate `coerce` at these particular types",
 which happen in this case to be polymorphic. (There is a good reason to be
 suspicious about filling in unification variables with polytypes, but
 explicit type application would mean that didn't happen.)

 Stephanie has a student working on this, I believe.

 Simon

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8631#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler