[GHC] #14961: QuantifiedConstraints: introducing classes through equality constraints fails

[GHC]

#14961: QuantifiedConstraints: introducing classes through equality constraints
fails
-------------------------------------+-------------------------------------
        Reporter:  mrkgnao           |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.5
      Resolution:                    |             Keywords:
                                     |  QuantifiedConstraints wipT2893
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  GHC rejects       |  Unknown/Multiple
  valid program                      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #14860            |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 I get different results for `wip/T2893` branch.  First I need
 `FlexibleContexts`.  Second, all the definitions fail.  For the first one
 I get
 {{{
 T14961.hs:16:8: error:
     • Could not deduce: c0 a
       from the context: (c ~=> Monoid, c a)
         bound by the type signature for:
                    foo :: forall (c :: * -> Constraint) a. (c ~=> Monoid,
 c a) => a
         at T14961.hs:16:8-45
     • In the ambiguity check for ‘foo’
       To defer the ambiguity check to use sites, enable
 AllowAmbiguousTypes
       In the type signature:
         foo :: forall c a. c ~=> Monoid => (c a => a)
    |
 16 | foo :: forall c a. c ~=> Monoid => (c a => a) -- ok
    |        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 }}}
 And the same happens if I simplify the type a bit more to
 {{{
 foo :: forall c a. (forall x. c x => Monoid x) => (c a => a) -- ok
 }}}
 Sure enough, this is an ambiguous type!  In a call, how do you expect `c`
 to be instantiated??

 I'm at a loss for what you are trying to achieve here.  Before we can look
 at substituting equalities, we need to work out these simpler questions.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14961#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler