caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: John Prevost <j.prevost@gmail.com>
To: Ocaml Mailing List <caml-list@inria.fr>
Subject: Re: [Caml-list] looping recursion
Date: Tue, 27 Jul 2004 20:38:48 -0400	[thread overview]
Message-ID: <d849ad2a04072717381ea7b0ee@mail.gmail.com> (raw)
In-Reply-To: <d849ad2a04072717275c865d9f@mail.gmail.com>

Oops.  Hit the send button before I meant to...

Anyway, that little loop with Boom example should show why the try
prevents the inner call from being in tail position: because the try
block's expressions can refer to variables in the scope of the
original function call, there's a possible continuation (you raise an
exception) that brings those variables back into scope.

The compiler *could*, I suppose, do some magic and notice when there
are no locally bound variables in the exception code path, but even
just thinking about how that might be implemented, I think it could
get messy.  What if it's not always the same try block that's around
the call that would otherwise be a tail call?  What if the fact that
there are multiple exception handlers is important, like:

exception Boom2 of int

let rec loop2 x =
  try
    if x < 100 then loop2 (x + x) else raise (Boom2 0)
  with Boom2 y -> raise (Boom2 (y + 1))

let use_loop2 x = try loop2 x with Boom2 y -> y

In this second example, use_loop2 is using the exceptions in loop2 to
calculate the result.  (Admittedly, this is a bizarre little setup.)

So figuring out if it's "safe" to make it a tail call is hard--not to
mention it would probably end up confusing the user even more if it
did work only part of the time.  (How do I change this so it's
recongnized as tail recursive again?)

Hope this helps.  :)  I tried to find an FAQ entry on this, because I
thought there was one, but was unable to find anything.

John.

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners


  reply	other threads:[~2004-07-28  0:38 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-07-27 23:43 briand
2004-07-28  0:27 ` John Prevost
2004-07-28  0:38   ` John Prevost [this message]
2004-07-28  1:17     ` skaller
2004-07-28  1:05   ` briand
2004-07-28  1:43     ` Brian Hurt
2004-07-28  2:49       ` briand
2004-07-28  3:12         ` Brian Hurt
2004-07-28  3:20         ` Brian Hurt
2004-07-28  5:54         ` brogoff
2004-07-28  7:22           ` Alex Baretta
2004-07-28 16:38             ` brogoff
2004-07-28 19:40               ` Jon Harrop
2004-07-28 20:18                 ` Brandon J. Van Every
2004-07-29  6:01                   ` Alex Baretta
2004-07-28 21:22                 ` brogoff
2004-07-29  9:13                   ` Daniel Andor
2004-07-29  9:25                     ` Keith Wansbrough
2004-07-29  9:41                       ` Nicolas Cannasse
2004-07-29  9:57                       ` Xavier Leroy
2004-07-29 10:44                         ` Daniel Andor
2004-07-29 12:56                           ` brogoff
2004-07-29 10:11                     ` skaller
2004-07-29 12:41                     ` brogoff
2004-07-29  6:28               ` Alex Baretta
2004-07-29 14:58                 ` brogoff
2004-07-29 16:12                   ` Brian Hurt
2004-07-29 17:49                     ` james woodyatt
2004-07-29 19:25                       ` Brian Hurt
2004-07-29 20:01                         ` brogoff
2004-07-30  4:42                           ` james woodyatt
2004-07-29 17:44                   ` james woodyatt
2004-07-29 23:12                     ` skaller
2004-07-29 22:42                   ` Alex Baretta
2004-07-30  2:38                     ` Corey O'Connor
     [not found]                     ` <200407300136.14042.jon@jdh30.plus.com>
2004-07-30 12:45                       ` Alex Baretta
2004-07-30 17:07                     ` brogoff
2004-07-30 18:25                       ` [Caml-list] kaplan-okasaki-tarjan deque (was "looping recursion") james woodyatt
2004-07-30 21:20                         ` brogoff
2004-07-31  5:37                           ` james woodyatt
2004-07-28  7:27       ` [Caml-list] looping recursion skaller
2004-07-28 14:36         ` Brian Hurt
2004-07-28 22:05           ` skaller
2004-07-28  0:37 ` skaller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=d849ad2a04072717381ea7b0ee@mail.gmail.com \
    --to=j.prevost@gmail.com \
    --cc=caml-list@inria.fr \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).