From: Jean-Christophe Filliatre <Jean-Christophe.Filliatre@lri.fr>
To: Michel Pastor <K4ng@codefx.org>
Cc: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] Infinite loop when catching an exception
Date: Fri, 10 Dec 2004 09:21:49 +0100 [thread overview]
Message-ID: <16825.23709.530721.745815@gargle.gargle.HOWL> (raw)
In-Reply-To: <41B91A9F.8050400@codefx.org>
Michel Pastor writes:
>
> I have a problem with a little piece of code but
> I attached the whole thing since i cannot reproduce
> the "bug" in reduced environement with the same values.
> I'm using the num library for doing exact maths.
>
> [...]
>
> If I try to catch the exception like that the process
> wont even go beyond the "try with" block
> I tested printing something instead of trying to exit
> but nothing is writed to the console and the cpu is at 100%
> All I can do is to kill the process.
> But if I don't try catch the Failure all works fine.
> I don't think doing unsafe things since there is no C code,
> no black magic, only some maths.
This is a ``syntax'' problem: you are not parsing the code the same
way the ocaml compiler does. When you insert the try/with like this:
> (************************)
> (* HERE IS THE PROBLEM *)
> (************************)
>
> let draw_curve color =
> begin
> set_color color;
> let index = ref !x_min in
> let ival = (!x_max -/ !x_min) // num_of_int (graph_width * 2) in
> while !index <=/ !x_max do
> let yval = poly_apply poly !index in
> (* uncomment the "try with" block to trigger the bug *)
> try
> begin
> print_endline ("ici " ^ string_of_num yval);
> let (graph_x, graph_y) = scale_point (!index, yval) in
> begin
> printf "%d %d\n" graph_x graph_y;
> fill_circle graph_x graph_y 1;
> end;
> print_endline "la";
> end;
> with _ -> failwith "ici maintenant";
> print_endline "et encore la";
> index := !index +/ ival;
> done
> end
then the three lines "failwith ...; print_endline ...; index := ..."
are all part of the exception handler. Since in your code no
exception is raised, these lines are never evaluated, including the
update of the index, which explains the infinite loop.
If you really need this try/with, you need to rewrite your code this
way (for instance):
while ... do
...
begin
try
...
with _ ->
...
end;
...
done
To avoid such pitfalls, it is convenient to use an editor with
automatic indentation of the code which is consistent with the ocaml
syntax, such as the Tuareg mode under Emacs. At some point, you become
so familiar with ocaml syntax that you don't need it anymore, but
meanwhile it helps.
Best regards,
--
Jean-Christophe
next prev parent reply other threads:[~2004-12-10 8:21 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-12-10 3:40 Michel Pastor
2004-12-10 8:21 ` Jean-Christophe Filliatre [this message]
2004-12-10 8:51 ` [Caml-list] " skaller
2004-12-10 10:32 ` Radu Grigore
2004-12-10 13:49 ` Michel Pastor
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=16825.23709.530721.745815@gargle.gargle.HOWL \
--to=jean-christophe.filliatre@lri.fr \
--cc=K4ng@codefx.org \
--cc=caml-list@yquem.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