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 JAA20023 for caml-redistribution; Fri, 19 Feb 1999 09:27:38 +0100 (MET)
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 UAA04634 for <caml-list@pauillac.inria.fr>; Thu, 18 Feb 1999 20:40:29 +0100 (MET)
Received: from mail12.svr.pol.co.uk (mail12.svr.pol.co.uk [195.92.193.215])
	by nez-perce.inria.fr (8.8.7/8.8.7) with ESMTP id UAA11958
	for <caml-list@inria.fr>; Thu, 18 Feb 1999 20:40:26 +0100 (MET)
Received: from modem-96.amantadine.dialup.pol.co.uk
	([62.136.49.224] helo=toy.william.bogus ident=williamc)
	by mail12.svr.pol.co.uk with esmtp (Exim 2.10 #1)
	id 10DZJB-0008FW-00
	for caml-list@inria.fr; Thu, 18 Feb 1999 19:40:25 +0000
Received: (from williamc@localhost)
	by toy.william.bogus (8.8.7/8.8.7) id TAA09359;
	Thu, 18 Feb 1999 19:12:55 GMT
Date: Thu, 18 Feb 1999 19:12:55 GMT
Message-Id: <199902181912.TAA09359@toy.william.bogus>
From: William Chesters <williamc@dai.ed.ac.uk>
To: caml-list@inria.fr
Subject: Anonymous classes
Sender: weis

   Could we have anonymous classes, as a natural counterpart to
anonymous functions?

	let x =
	  object
	    method foo = 17
	  end
	in
	...

They would sometimes be useful for specifying the little bits of `glue'
one needs to interface a library routine with one's current problem.

   On the other hand maybe this kind of sugar isn't strictly
necessary, since I found this pretty easy to add with camlp4 :-),
using a dummy local module.

	expr: LEVEL "simple"
	  [ [ "object"; cspo = OPT class_self_patt;
                        cf = class_structure; "end" ->
		<:expr< let module Dummy__ =
			  struct
			    class c () = object $cspo$ $list:cf$ end ;
			  end
			 in
			 new Dummy__.c () >> ] ] ;

(except that the definitions of class_structure and class_self_patt
need to be copied from pa_o.ml)