[GHC] #16391: "Quantified type's kind mentions quantified type variable" error with fancy-kinded GADT

[GHC]

#16391: "Quantified type's kind mentions quantified type variable" error with
fancy-kinded GADT
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  8.6.3
  checker)                           |
      Resolution:                    |             Keywords:  TypeInType
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  GHC rejects       |  Unknown/Multiple
  valid program                      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 The other bad thing is that the error message is a mess, becuase it
 mentions 'a' in one place and 'a1' in the other.  Very confusing.

 Reason:
 {{{
                    (forAllEscapeErr env' ty tau_kind)
 }}}
 we pass `env'` (the result of tidying tvs) to `forAllEscapeErr`, which
 then prints `ty` in this tidy-env, thereby tidying the same tyvars again.
 This would not be hard to clear up, perhpas not by printing `ty` afresh,
 but by passing `tvs'` and `phi` to `forAllEscapeErr`, and printing the
 foralls manually, as it were.

 Does that make sense?  Also I wonder if if might be easier to understand
 if we said
 {{{
    • Quantified type's kind mentions quantified type variable
            type: forall a1. T
      where the body of the forall has this kind
            T :: Const * a
 }}}

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16391#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler