[GHC] #8883: FlexibleContexts checking should happen also on inferred signature

[GHC]

#8883: FlexibleContexts checking should happen also on inferred signature
------------------------------------------------+--------------------------
        Reporter:  Blaisorblade                 |            Owner:
            Type:  bug                          |  jstolarek
        Priority:  normal                       |           Status:  closed
       Component:  Compiler                     |        Milestone:
      Resolution:  fixed                        |          Version:  7.6.3
Operating System:  Unknown/Multiple             |         Keywords:
 Type of failure:  None/Unknown                 |     Architecture:
       Test Case:  typecheck/should_fail/T8883  |  Unknown/Multiple
        Blocking:                               |       Difficulty:
                                                |  Unknown
                                                |       Blocked By:
                                                |  Related Tickets:
------------------------------------------------+--------------------------

Comment (by Lemming):

 Replying to [comment:17 simonpj]:

 > I can't get terribly excited about this decision; but I can see the
 logic, and I'll gladly change the current behaviour if there is support
 for doing so.  (This would affect both inferred and declared types, of
 course.)

 As I said, you cannot eliminate all `FlexibleContexts` using an auxiliary
 variable.
 You cannot convert
 {{{class (Functor (PF a)) => C a where}}}
 to
 {{{class (Functor b, b ~ PF a) => C a where}}}
 because you cannot introduce the variable `b` in the superclass
 constraint.

 Thus features provided by `FlexibleContexts` cannot always be achieved by
 `TypeFamilies`, and thus `FlexibleContexts` should not be a "superfeature"
 of `TypeFamilies`.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8883#comment:19>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler