[GHC] #6074: Instance inference failure with GADTs

[GHC]

#6074: Instance inference failure with GADTs
---------------------------------+------------------------------------------
    Reporter:  goldfire          |       Owner:                  
        Type:  bug               |      Status:  new             
    Priority:  normal            |   Milestone:                  
   Component:  Compiler          |     Version:  7.5             
    Keywords:  GADTs             |          Os:  Unknown/Multiple
Architecture:  Unknown/Multiple  |     Failure:  None/Unknown    
  Difficulty:  Unknown           |    Testcase:                  
   Blockedby:                    |    Blocking:                  
     Related:                    |  
---------------------------------+------------------------------------------
Changes (by simonpj):

  * difficulty:  => Unknown


Comment:

 I think this is a non-starter.  Consider the translation of the orignal
 code into FC:
 {{{
 mkString' :: GADT a -> String
 mkString' g = case g of
   GInt  -> toString Int dStringableInt g
   GBool -> toString Bool dStringableBool g
 }}}
 Remember that `toString :: forall a. Stringable a => a -> String`.  So it
 takes a value argument that is a `Stringable` dictionary.  And the two
 calls may LOOK the same to you, but they aren't really the same: we must
 pass two different dictionaries (in this case gotten from the two instance
 declarations.  (And there's a different type parameter too.)

 Another way to put it is: what is the FC translation of this?
 {{{
 mkString :: GADT a -> String
 mkString g = toString g
 }}}
 So I don't see where to go with this one.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/6074#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler