Re: [Caml-list] Making a polymorphic type non-polymorphic to comply with original signature

[Hugo Ferreira <hmf@inescporto.pt>]

Thomas Gazagnaire wrote:
> Or you can also functorize your own piece of code :
> 
> module Make (A : sig type t end) =
> struct
> 
>     module rec H :
>     sig
>         type 'a node =
>             | Node of 'a node J.t
>             | Leaf of 'a
>                  
>         type t = A.t node
>         val equal : t -> t -> bool
>         val hash : t -> int
>     end =
>     struct
>         type 'a node =
>             | Node of 'a node J.t
>             | Leaf of 'a
>                  
>         type t = A.t node
>         let equal = (==)
>         let hash = Hashtbl.hash
>     end
>        
>     and J : Hashtbl.S with type key = A.t H.node = Hashtbl.Make( H )
>    
> end
>

Incredible! You have just provided a excellent solution.
Works "out of the box":

# module N = Make(struct type t = int end ) ;;

# let jls = J.create 13 ;;
val jls : '_a N.J.t = <abstr>

# let r = H.Node(jls) ;;
val r : '_a N.H.node = N.H.Node <abstr>

# let _ = J.add jls r r ;;
- : unit = ()

# let x = J.find jls r ;;
val x : int N.H.node = N.H.Node <abstr>

# let r = (x == r ) ;;
val r : bool = true

Hmmm... now I wonder if that node type can also be pulled in by the
functor. If so we have a very general solution to the problem of
equality checking of polymorphic types. Anyone care to comment?

Once again Thomas,
thank you.
Hugo F.


> 2009/1/20 Hugo Ferreira <hmf@inescporto.pt <mailto:hmf@inescporto.pt>>
> 
>     David Teller wrote:
> 
>         It's probably feasible without copy & paste by building a
>         functor on top
>         of the defunctorized hashtable in Batteries. Or by just using the
>         defunctorized hashtable of Batteries directly, although it's not
>         as safe
>         as the functorized version, due to the absence of existential types.
> 
> 
>     If I understand you correctly I would have to redefine equivalents for:
>     - HashedType
>     - S
>     - Make(H: HashedType)
> 
>     Basically copy & paste these and change the type.
>     Doable although not to my liking.
> 
>     TIA,
>     Hugo F.
> 
> 
> 
>         Cheers,
>          David
> 
>         On Tue, 2009-01-20 at 12:24 +0100, Daniel Bünzli wrote:
> 
>             Le 20 janv. 09 à 11:59, Hugo Ferreira a écrit :
> 
>                 Is it possible to make H comply with Hashtbl.HashedType
>                 i.e: make
>                 J.Key = 'a H.node ?
> 
>             This issue is well known (e.g. see here [1]). Your are
>             running into  limitations of the standard library. The only
>             unsatisfying answer is  to copy the code from the standard
>             library and add the parameter  yourself.
> 
>             Best,
> 
>             Daniel
> 
>             [1]
>             http://groups.google.com/group/fa.caml/browse_thread/thread/f2acb593da91553c?hl=fr&ie=UTF-8&q=type+var+in+functor+fa.caml
>             <http://groups.google.com/group/fa.caml/browse_thread/thread/f2acb593da91553c?hl=fr&ie=UTF-8&q=type+var+in+functor+fa.caml>
> 
>             ____________________________________________ugs
> 
> 
>     _______________________________________________
>     Caml-list mailing list. Subscription management:
>     http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
>     Archives: http://caml.inria.fr
>     Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
>     Bug reports: http://caml.inria.fr/bin/caml-bugs
> 
>