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 KAA03509 for caml-red; Mon, 2 Oct 2000 10:19:35 +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 RAA02491 for <weis@pauillac.inria.fr>; Sat, 30 Sep 2000 17:19:46 +0200 (MET DST)
Received: from miss.wu-wien.ac.at (miss.wu-wien.ac.at [137.208.107.17])
	by nez-perce.inria.fr (8.10.0/8.10.0) with ESMTP id e8UFJkX05192;
	Sat, 30 Sep 2000 17:19:46 +0200 (MET DST)
Received: (from mottl@localhost)
	by miss.wu-wien.ac.at (8.9.0/8.9.0) id RAA23911;
	Sat, 30 Sep 2000 17:19:39 +0200 (MET DST)
Date: Sat, 30 Sep 2000 17:19:39 +0200
From: Markus Mottl <mottl@miss.wu-wien.ac.at>
To: Pierre Weis <Pierre.Weis@inria.fr>
Cc: Julian Assange <proff@iq.org>, caml-list@inria.fr
Subject: Re: bottom types and threaded exits
Message-ID: <20000930171939.B9945@miss.wu-wien.ac.at>
References: <wxzokqtf8r.fsf@suburbia.net> <200009301425.QAA17425@pauillac.inria.fr>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2i
In-Reply-To: <200009301425.QAA17425@pauillac.inria.fr>; from Pierre.Weis@inria.fr on Sat, Sep 30, 2000 at 16:25:32 +0200
Sender: weis@pauillac.inria.fr

On Sat, 30 Sep 2000, Pierre Weis wrote:
> Once more I would prefer a more general recursive function, that would
> loop for any (even non looping) function. For instance:
> 
> # let rec bottom f = f (); bottom f;;     
> val bottom : (unit -> 'a) -> 'b = <fun>
> 
> and I consider the 'b in the type of bottom as clearly indicating the
> looping for ever nature of any application of the bottom function.

The name "bottom" might be a bit confusing in this context, because "f"
could be a function with observable side effects (e.g. I/O), whereas
"bottom" is usually used to indicate non-termination only, which cannot be
observed.

But this function (with a different name) and others would come handy more
often than once, e.g.:

  let rec forever f = f (); forever f
  let try_forever e f = try forever f with exn -> if e <> exn then raise exn

This would allow:

  let _ =
    try_forever End_of_file (fun () -> print_endline (read_line ()));
    print_endline "Done!"

It is much more tedious (and error-prone if one forgets the "begin ...
end") to write the following:

  let _ =
    begin
      try while true do print_endline (read_line ()) done
      with End_of_file -> ()
    end;
    print_endline "Done!"

Maybe functions similar to "forever" and "try_forever" could make it into
the standard library? Such loops come up quite often with e.g. servers or
I/O-loops in general.

Best regards,
Markus Mottl

-- 
Markus Mottl, mottl@miss.wu-wien.ac.at, http://miss.wu-wien.ac.at/~mottl