From mboxrd@z Thu Jan  1 00:00:00 1970
Received: (from weis@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id NAA09943 for caml-redistribution; Fri, 17 Sep 1999 13:12:12 +0200 (MET DST)
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 PAA25910 for <caml-list@pauillac.inria.fr>; Wed, 15 Sep 1999 15:30:32 +0200 (MET DST)
Received: from isil.maya.com (isil.maya.com [192.70.254.5])
	by nez-perce.inria.fr (8.8.7/8.8.7) with ESMTP id PAA14522
	for <caml-list@inria.fr>; Wed, 15 Sep 1999 15:30:30 +0200 (MET DST)
Received: (from prevost@localhost)
	by isil.maya.com (8.9.3/8.9.3) id JAA25996;
	Wed, 15 Sep 1999 09:33:33 -0400
Sender: weis
To: Steve Stevenson <steve@cs.clemson.edu>
Cc: caml-list@inria.fr
Subject: Re: Imperative list operations
References: <14302.41638.913957.615588@merlin.cs.clemson.edu>
From: John Prevost <prevost@maya.com>
Date: 15 Sep 1999 09:33:33 -0400
In-Reply-To: Steve Stevenson's message of "Tue, 14 Sep 1999 15:36:18 -0400 (EDT)"
Message-ID: <m3iu5cb8pe.fsf@isil.maya.com>
User-Agent: Gnus/5.070096 (Pterodactyl Gnus v0.96) Emacs/20.4
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

Steve Stevenson <steve@cs.clemson.edu> writes:

> 	I need a double-ended queue implementation. {...imperative...}

> 	I've tried all the naïve type declarations --- all of which
> don't seem to work.  I've tried the age old tricks. What am I not
> understanding? or doing right? I'm not fussy: tuples, records or
> objects are fine.

It'd help to know what you've tried.  The following work for me:

type 'a dlist_node =
  | Nil
  | Node of 'a dlist_node ref * 'a * 'a dlist_node ref

Or better:

type 'a dlist_node =
  { mutable dlist_prev : 'a dlist_node option;
            dlist_val  : 'a;
    mutable dlist_next : 'a dlist_node option }

Note that in the first case, you need to use value constructors, since:

type 'a dlist_node = 'a dlist_node option ref * 'a * 'a dlist_node option ref

is recursive, and O'Caml doesn't allow type aliases (as opposed to
union types or record types, which "create" new types) to be
recursive.

John.