[GHC] #16263: Rework GHC's treatment of constraints in kinds (was: GHC accepts illegal constraint in kind)

[GHC]

#16263: Rework GHC's treatment of constraints in kinds
-------------------------------------+-------------------------------------
        Reporter:  goldfire          |                Owner:  (none)
            Type:  bug               |               Status:  patch
        Priority:  normal            |            Milestone:  8.10.1
       Component:  Compiler          |              Version:  8.6.3
      Resolution:                    |             Keywords:  TypeInType,
                                     |  newcomer
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |  https://gitlab.haskell.org/ghc/ghc/merge_requests/128
-------------------------------------+-------------------------------------
Changes (by RyanGlScott):

 * status:  new => patch
 * milestone:  8.8.1 => 8.10.1
 * differential:   => https://gitlab.haskell.org/ghc/ghc/merge_requests/128


Comment:

 Replying to [comment:7 simonpj]:
 > Thanks for the ping.  Fixing this will fit right into my work on #16185
 (see `wip/T16185`) where I have already written new Notes on constraints
 in kinds.

 OK, I didn't realize that you were directly working on fixing this. In
 that case, I'll retitle this ticket to reflect your more ambitious goals.

 > I'll adjust `TcValidity` to reject this kind up-front, as you suggest.

 By "this kind", I assume you mean `Eq a => Type` from the original
 example? If so, yes, rejecting that (and adding a test case for it) would
 be wonderful.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/16263#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler