[GHC] #10635: -fwarn-redundant-constraints should not be part of -Wall

[GHC]

#10635: -fwarn-redundant-constraints should not be part of -Wall
-------------------------------------+-------------------------------------
        Reporter:  Lemming           |                Owner:
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  7.11
  checker)                           |
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Incorrect         |  Unknown/Multiple
  warning at compile-time            |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #9939, #9973,     |  Differential Rev(s):
  #10100, #10183, #11370             |
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 The proposal on the thread is

 * `-Wredundant-constraints` reports when a user writes a constraint that
 is fully equivalent to some other, strictly smaller constraint, like
 suggesting simplifying (Eq a, Ord a) to (Ord a).

 * `-Wtype-overly-specific` warns about any type signature that's more
 specific than it needs to be.  That would include both

   (a) When a constraint is specified but not used
 {{{
 f :: Eq a => a -> a
 f x = x
 }}}
   (b) When a type variable is instantiated to something more specific than
 necessary
 {{{
 g :: [a] -> [a]
 g x = x
 }}}
 That sounds attractive but (b) isn't an easy fit with the current
 implementation.  GHC treats a type signature as normative and "pushes it
 inward".  So when type checking
 {{{
 g :: [a] -> [a]
 g x = x && True
 }}}
 we immediately given an error "can't unify [a] with Bool".  We do ''not''
 infer the most general type of g, namely `g :: Bool -> Bool`, and then
 compare with the signature.

 So I don't know an easy way to do (b).  But (a) would be fine.  Patches
 welcome.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10635#comment:19>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler