Re: [Caml-list] About polymorphic methods and recursive classes

[Sebastien Briais <sebastien.briais@epfl.ch>]

Jacques Garrigue wrote:

>From: Sebastien Briais <sebastien.briais@epfl.ch>
>  
>
>>>># class virtual ['a] visitor =
>>>>  object
>>>>    method virtual caseNil : 'a
>>>>  end
>>>>  and virtual int_list =
>>>>  object
>>>>    method virtual visit : 'a.('a visitor -> 'a)
>>>>  end
>>>>  ;;
>>>>class virtual ['a] visitor : object method virtual caseNil : 'a end
>>>>class virtual int_list : object method virtual visit : 'a visitor ->
>>>>'a end
>>>>        
>>>>
>>I am still puzzled by the answer of the interpreter which says me that
>>"v has type 'a visitor; It has no method caseNil"
>>although it has inferred just before that
>>"class virtual ['a] visitor : object method virtual caseNil : 'a end"
>>
>>Moreover, it seems odd to me that it works with class types but not without.
>>These class types should be inferred by Ocaml, shouldn't they ? (and 
>>they seems
>>to be correctly inferred in my example (in the first part of it))
>>    
>>
>
>Actually no. Polymorphic methods are not inferred, but declared.
>And polymorphic type variables should not be used as parameters to
>simultaneously defined types.
>Actually, your mutually recursive definition should not be accepted to
>begin with. It clearly results in a wrong internal type.
>  
>
So, if I understand you well, it is not possible to write something like 
that in ocaml:

class virtual ['a,'b] list_visitor =
object
  method virtual caseNil : 'a
  method virtual caseCons : 'b -> 'b list -> 'a
end
and virtual ['b] list =
object
  method virtual visit : 'a.(('a,'b) list_visitor -> 'a)
end

class ['a] nil =
object
  inherit ['a] list
  method visit v = v#caseNil
end

class ['a] cons = fun h t ->
object
  inherit ['a] list
  method visit v = v#caseCons h t
end

and especially, the first definition should be considered as incorrect 
since the
polymorphic variable 'a of list is used as a parameter of the list_visitor.

But what is the reason of this limitation ? 
I do not see the reasons since for example, in Generic Java, such a 
definition is correct

abstract class Visitor<a, b> {
  abstract a caseNil();
  abstract a caseCons(b x, List<b> xs);
}

abstract class List<b> {
  abstract <a> a visit(Visitor<a, b> v);
}

Can you explain in more details your answer ?
Does it lead to type insafety or indecidability of typing to authorize 
polymorphic
type variables to be used as parameters to simulteanously defined types ?


Regards

Sébastien


-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners