From mboxrd@z Thu Jan  1 00:00:00 1970
Received: (from majordomo@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id RAA15874; Fri, 19 Sep 2003 17:49:38 +0200 (MET DST)
X-Authentication-Warning: pauillac.inria.fr: majordomo set sender to owner-caml-list@pauillac.inria.fr using -f
Received: from nez-perce.inria.fr (nez-perce.inria.fr [192.93.2.78]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id RAA08056 for <caml-list@pauillac.inria.fr>; Fri, 19 Sep 2003 17:49:37 +0200 (MET DST)
Received: from mail.speakeasy.net (mail11.speakeasy.net [216.254.0.211])
	by nez-perce.inria.fr (8.11.1/8.11.1) with ESMTP id h8JFna507078
	for <caml-list@inria.fr>; Fri, 19 Sep 2003 17:49:36 +0200 (MET DST)
Received: (qmail 14740 invoked from network); 19 Sep 2003 15:49:34 -0000
Received: from unknown (HELO apprentice.genxnet.com) ([64.81.145.152])
          (envelope-sender <ayerkes@speakeasy.net>)
          by mail11.speakeasy.net (qmail-ldap-1.03) with SMTP
          for <caml-list@inria.fr>; 19 Sep 2003 15:49:34 -0000
Date: Fri, 19 Sep 2003 10:58:52 -0500
From: art yerkes <ayerkes@speakeasy.net>
To: caml-list@inria.fr
Subject: Re: [Caml-list] Subclasses and pattern matching
Message-Id: <20030919105852.24b1ea03.ayerkes@speakeasy.net>
In-Reply-To: <20030919152252.GA30989@henchmonkey.org>
References: <20030919142156.GA30187@henchmonkey.org>
	<20030919145220.GD4205@redhat.com>
	<20030919152252.GA30989@henchmonkey.org>
X-Mailer: Sylpheed version 0.8.11 (GTK+ 1.2.10; i686-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
X-Loop: caml-list@inria.fr
X-Spam: no; 0.00; yerkes:01 ayerkes:01 caml-list:01 model:01 extern:01 extern:01 ints:01 inherit:01 inherit:01 speakeasy:01 int:01 int:01 sep:01 match:02 match:02 
Sender: owner-caml-list@pauillac.inria.fr
Precedence: bulk

On Fri, 19 Sep 2003 11:22:52 -0400
katre <katre@henchmonkey.org> wrote:

> Richard Jones wrote:
> > 
> > Possibly I'm being dumb here, but wouldn't it be better to use
> > a union type, eg:
> > 
> > type num = Int of int | Float of float
> > 
> 
> Yes, except for two facts:
> 
> a) Each different type has different data available (some are ints, some
> strings, some more complex)
> b) Each different type has a lot of data.  I don't really want to
> specify a 10-member tuple for each constructor :)
> c) Where the data is used, I want to match two or three different types,
> and then call a default method with anything left over.  With a union, I
> have to deal with every possible type, or do nothing with the data.
> 
> To clarify, using your type above, I can't do
> 
> match n with
>     Int(i) -> print_int i
>     | _ -> print_int some_data_from_n
> 
> Remember, all of my structures have many more than 1 piece of data.  A
> class inheritance really is the best way to model the data structure, I
> just want to be able to pattern match on it.
> 
> katre
> 

You could wrap your items in a union:

exception NoExternalRepForBase

type 'a base_extern_t = 
        Derived1 of 'a | Derived2 of 'a

class virtual base =
object
	method virtual as_external : unit -> base base_extern_t
end

class derived1 =
object(self)
	inherit base
	method as_external () = Derived1 (self :> base)
end

class derived2 =
object(self)
	inherit base
	method as_external () = Derived2 (self :> base)
end

let l = [ (new derived2 # as_external()) ; (new derived1 # as_external()) ]
-- 
"Should array indices start at 0 or 1? My compromise of 0.5 was rejected
 without, I thought, proper consideration."
   - S. Kelly-Bootle

-------------------
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