[GHC] #15868: Standard deriving should be less conservative when `UndecidableInstances` is enabled

[GHC]

#15868: Standard deriving should be less conservative when `UndecidableInstances`
is enabled
-------------------------------------+-------------------------------------
        Reporter:  edsko             |                Owner:  (none)
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.6.1
      Resolution:                    |             Keywords:  deriving
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #11008            |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by RyanGlScott):

 Replying to [comment:5 edsko]:
 > Would simply refusing to add constraints without any type variables be
 sufficient to rule out examples like the one you are worried about?

 Not necessarily. Imagine if `NotAShowInstance` had a type parameter, for
 instance. Under your proposed scheme, we'd still allow `data Foo a = MkFoo
 Int (NotAShowInstance a) deriving Show`.

 I think the right way to frame the discussion is to permit constraints
 that involve non–type constructors (such as type families or higher-kinded
 type variables) in certain ways. That's still not anything resembling a
 specification, but it's closer.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15868#comment:6>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler