From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <jon@ffconsultancy.com>
X-Spam-Checker-Version: SpamAssassin 3.1.3 (2006-06-01) on yquem.inria.fr
X-Spam-Level: 
X-Spam-Status: No, score=0.0 required=5.0 tests=AWL autolearn=disabled 
	version=3.1.3
X-Original-To: caml-list@yquem.inria.fr
Delivered-To: caml-list@yquem.inria.fr
Received: from mail2-relais-roc.national.inria.fr (mail2-relais-roc.national.inria.fr [192.134.164.83])
	by yquem.inria.fr (Postfix) with ESMTP id DC69DBB84
	for <caml-list@yquem.inria.fr>; Tue, 17 Jun 2008 16:40:16 +0200 (CEST)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AugEAKBpV0jUnw6DfGdsb2JhbACCOY9uAQELBQIECREDnBc
X-IronPort-AV: E=Sophos;i="4.27,658,1204498800"; 
   d="scan'208";a="12175780"
Received: from pih-relay04.plus.net ([212.159.14.131])
  by mail2-smtp-roc.national.inria.fr with ESMTP; 17 Jun 2008 16:40:16 +0200
Received: from [80.229.56.224] (helo=beast.local)
	 by pih-relay04.plus.net with esmtp (Exim) id 1K8cM3-0007Rl-EH
	for caml-list@yquem.inria.fr; Tue, 17 Jun 2008 15:40:15 +0100
From: Jon Harrop <jon@ffconsultancy.com>
Organization: Flying Frog Consultancy Ltd.
To: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] optimization of sequence of List.map and inlining
Date: Tue, 17 Jun 2008 15:36:31 +0100
User-Agent: KMail/1.9.9
References: <676aba050806101201x526f03b1lf1fdbed665ee2e3f@mail.gmail.com>
In-Reply-To: <676aba050806101201x526f03b1lf1fdbed665ee2e3f@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-15"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200806171536.31220.jon@ffconsultancy.com>
X-Plusnet-Relay: d102069457dcd7e00aea67125fb623a5
X-Spam: no; 0.00; inlining:01 mls:01 inlining:01 seq:01 sig:01 val:01 val:01 struct:01 seq:01 abstr:01 abstr:01 compiler:01 showstopper:01 uglier:01 camlp:01 

On Tuesday 10 June 2008 20:01:12 Charles Hymans wrote:
> Thank you for your answer.
> 
>> In MLs, you deforest by hand. You might like to use a function
>> composition
> 
> Unfortunately, in the case from which my example is extracted, I
> believe I can't do that:
> The several List.maps, even though they are executed in sequence, are
> not in the same module. In order to deforest by hand, I have to break
> the module interface, do some function inlining and then do the
> function composition.
> Something I don't want to do, because it would make the code difficult
> to read and to maintain.

Perhaps you could use something like the following abstract lazy sequence 
implementation:

  module Seq : sig
    type ('a, 'b) t
    val of_list : 'a list -> ('a, 'a) t
    val map : ('a -> 'b) -> ('c, 'a) t -> ('c, 'b) t
    val to_list : ('a, 'b) t -> 'b list
  end = struct
    type ('a, 'b) t = ('a -> 'b) * 'a list
    let of_list list = (fun x -> x), list
    let map f (g, list) = (fun x -> f(g x)), list
    let to_list (f, list) = List.map f list
  end

Where your existing implementation exposes an 'a list you now expose an 
('b, 'a) Seq.t instead.

# let a = Seq.of_list [1];;
val a : (int, int) Seq.t = <abstr>
# let b = Seq.map float a;;
val b : (int, float) Seq.t = <abstr>
# let c = Seq.map string_of_float b;;
val c : (int, string) Seq.t = <abstr>
# Seq.to_list c;;
- : string list = ["1."]

> That's why I wondered if there was a way to tell the compiler to do it
> for me. Or maybe some incantation with ocamlp4?

I would not recommend pursuing that line of thought under any circumstances. 
Even if this problem is a critical showstopper for you, I would always 
recommend uglier code over camlp4 hacks.

-- 
Dr Jon D Harrop, Flying Frog Consultancy Ltd.
http://www.ffconsultancy.com/products/?e