[GHC] #663: Confusion about types

[GHC]

#663: Confusion about types
----------------------------------------+-----------------------------------
    Reporter:  guest                    |        Owner:         
        Type:  bug                      |       Status:  new    
    Priority:  normal                   |    Milestone:         
   Component:  Compiler (Type checker)  |      Version:  6.4.1  
    Severity:  normal                   |     Keywords:         
          Os:  Unknown                  |   Difficulty:  Unknown
Architecture:  Unknown                  |  
----------------------------------------+-----------------------------------
In GHC 6.0.1 the following data type definition (for a binary search tree)
 seems to be invalid.
 I know I've used this kind of thing before, so whats up.  This is the
 entire file, by the
 way:

 data BST a = Ord a => Leaf | Node a (BST a) (BST a)

 Gives an error message :

     All of the type variables in the constraint `Ord a' are already in
 scope
         (at least one must be universally quantified here)
     When checking the existential context of constructor `Leaf'
     In the data type declaration for `BST'

 I'm confused - where would I put a universal quantification?   Does Leaf
 now need a type declaration and if so what would it be?

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/663>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler