From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <dave@ramenlabs.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
Received: from concorde.inria.fr (concorde.inria.fr [192.93.2.39])
	by yquem.inria.fr (Postfix) with ESMTP id 3A5D7BC69
	for <caml-list@yquem.inria.fr>; Thu, 23 Aug 2007 06:31:07 +0200 (CEST)
Received: from tenhost.net (tenhost.net [64.246.28.52])
	by concorde.inria.fr (8.13.6/8.13.6) with SMTP id l7N4V5b4002266
	for <caml-list@yquem.inria.fr>; Thu, 23 Aug 2007 06:31:06 +0200
Received: (qmail 24523 invoked by uid 502); 22 Aug 2007 23:03:42 -0500
Received: from localhost (sendmail-bs@127.0.0.1)
  by localhost with SMTP; 22 Aug 2007 23:03:42 -0500
Date: Wed, 22 Aug 2007 23:03:42 -0500 (CDT)
From: Dave Benjamin <dave@ramenlabs.com>
X-X-Sender: ramen@tenhost.net
To: Eric Cooper <ecc@cmu.edu>
Cc: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] commands.getoutput () in ocaml?
In-Reply-To: <20070822230418.GC10505@stratocaster.home>
Message-ID: <Pine.LNX.4.63.0708222301001.24186@tenhost.net>
References: <28fa90930708221256t497e8356k84861abd6b2e91b8@mail.gmail.com>
 <Pine.LNX.4.63.0708221509200.13687@tenhost.net>
 <95513600708221535t7508e900h192554da73d37fcf@mail.gmail.com>
 <20070822230418.GC10505@stratocaster.home>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
X-Miltered: at concorde with ID 46CD0D89.000 by Joe's j-chkmail (http://j-chkmail . ensmp . fr)!
X-Spam: no; 0.00; ocaml:01 0200,:01 andrieu:01 buffer:01 buffer:01 2048:01 2007,:98 23,:98 fires:98 wrote:01 wrote:01 unix:01 unix:01 caml-list:01 benjamin:01 

On Wed, 22 Aug 2007, Eric Cooper wrote:

> On Thu, Aug 23, 2007 at 12:35:14AM +0200, Olivier Andrieu wrote:
>> you could use Buffer.add_substring instead of add_string, that will
>> avoid an intermediate copy of the string
>
> And you can use Buffer.add_channel to avoid intermediate strings
> altogether (until the final Buffer.to_string).

I'm not quite sure how to do this, actually. Here's what I've got so far:

let read_process command =
   let buffer_size = 2048 in
   let buffer = Buffer.create buffer_size in
   let in_channel = Unix.open_process_in command in
   begin
     try
       while true do
         Buffer.add_channel buffer in_channel buffer_size;
       done
     with
       | End_of_file -> ()
       | e ->
           ignore (Unix.close_process_in in_channel);
           raise e;
   end;
   ignore (Unix.close_process_in in_channel);
   Buffer.contents buffer

However, this doesn't work unless I set buffer_size to 1. Otherwise, the 
End_of_file fires before all the data is read in, so I either get nothing 
or an incomplete result.

Any advice?

Thanks,
Dave