#5051: Typechecker behaviour change
---------------------------------------------+------------------------------
  Reporter:  igloo                           |          Owner:                  
      Type:  bug                             |         Status:  new             
  Priority:  high                            |      Milestone:  7.2.1           
 Component:  Compiler                        |        Version:  7.0.2           
Resolution:                                  |       Keywords:                  
        Os:  Unknown/Multiple                |   Architecture:  Unknown/Multiple
   Failure:  None/Unknown                    |     Difficulty:  Unknown         
  Testcase:  typecheck/should_compile/T5051  |      Blockedby:                  
  Blocking:                                  |        Related:                  
---------------------------------------------+------------------------------
Changes (by simonpj):

  * owner:  simonpj =>
  * difficulty:  => Unknown
  * status:  closed => new
  * resolution:  fixed =>


Comment:

 I think we have to re-introduce silent superclass paramters (see #5751).
 And indeed, looking at the example above, arguably it ''should'' complain
 about overlap.  Just suppose that the `Ord` instance for lists was
 declared like this:
 {{{
 instance Ord a => Ord [a] where
   x >= y == x>y || x==y
 }}}
 With this implementation, at type `Ord [T]` we really should be using the
 `Eq [T]` instance, so the overlap complaint for `foo` is absolutely right.

 '''Serge:''' how bad would it be if we re-introduced the behaviour you
 reported in http://www.haskell.org/pipermail/glasgow-haskell-
 users/2011-March/020116.html?  I now think (a) it's the right behaviour
 (b) it's necessary to cure #5751.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5051#comment:5>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler

_______________________________________________
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Reply via email to