caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: "S. Ted Sandler" <tsandler@seas.upenn.edu>
To: caml-list@inria.fr
Subject: [Caml-list] strange tail recursion behavior by caml compiler
Date: Wed, 25 Aug 2004 15:57:04 -0400	[thread overview]
Message-ID: <20040825195704.GA11359@red.seas.upenn.edu> (raw)

Here's a question to those who are "in the know"
about when the caml compiler eliminates tail-calls.

I wrote the following tail-recursive function
"listbest", which is supposed to return the "best"
element of a list (where "best" is determined by the 
behavior of the "compare_fn" that is passed as an
argument).

Just for background, know that the "compare_fn"
should return +1 if arg1 is better than arg2, -1 if
arg2 is better than arg1, or 0 if they are equally
good.

So here's the problem.  When I write listbest like
this, it's not tail recursive.  I get stack overflow
exceptions on long lists:

  let listbest = fun compare_fn lst ->
    let rec listbest_rec = fun best lst_ ->
      match lst_ with
        | [] -> best
        | head :: tail ->
            let cmp = compare_fn head best in
              if cmp > 0
              then listbest_rec head tail
              else listbest_rec best tail
    in
      match lst with
        | [] -> raise
           (Invalid_argument "error: list is empty")
        | head :: tail ->
           listbest_rec head tail



BUT when I write "listbest" like this, it is:

  let listbest = fun compare_fn lst ->
    let rec listbest_rec = fun cmp_fn best lst_ ->
      match lst_ with
        | [] -> best
        | head :: tail ->
            let cmp = cmp_fn head best in
              if cmp > 0
              then listbest_rec cmp_fn head tail
              else listbest_rec cmp_fn best tail
    in
      match lst with
       | [] -> raise
          (Invalid_argument "error: list is empty")
       | head :: tail ->
          listbest_rec compare_fn head tail



The difference btwn these 2 defns doesn't seem like
one that should affect whether they are tail recursive,
does it?  Thanks in advance for shedding light on this.

-ted


PS I am using the ocaml-3.08 ocamlopt compiler.

PPS btw, "List.map", as defined in the ocaml stdlib,
is not tail recursive either.  Someone might want to
change that.

 let rec map f = function
     [] -> []
   | a ::l -> let r = f a in r :: map f l

-------------------
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:[~2004-08-25 19:57 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-08-25 19:57 S. Ted Sandler [this message]
2004-08-26 12:36 ` Damien Doligez
2004-08-26 15:44 ` Xavier Leroy

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=20040825195704.GA11359@red.seas.upenn.edu \
    --to=tsandler@seas.upenn.edu \
    --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).