From: Gabriel Dos Reis <gdr@integrable-solutions.net>
To: Joe Buck <Joe.Buck@synopsys.COM>
Cc: Michael Veksler <VEKSLER@il.ibm.com>, gcc@gcc.gnu.org
Subject: Re: signed is undefined and has been since 1992 (in GCC)
Date: Tue, 28 Jun 2005 17:03:00 -0000 [thread overview]
Message-ID: <m3d5q6ifd2.fsf@uniton.integrable-solutions.net> (raw)
In-Reply-To: <20050628163249.GB9524@synopsys.com>
Joe Buck <Joe.Buck@synopsys.COM> writes:
| On Tue, Jun 28, 2005 at 10:23:51AM +0300, Michael Veksler wrote:
|
|
| On Jun 28, 2005, at 1:12 AM, Gabriel Dos Reis wrote:
| > > > So,
| > > > please, do refrain from reasoning like "since we did X for Y and Y was
| > > > undefined behaviour, we should do the same for Z." "Undefined
| > > > behaviour" isn't a 0 or 1 thingy, even though it is about computers.
| > > > You need to evaluate them on case-by-case basis.
|
| Andrew Pinski wrote on 28/06/2005 08:34:25:
I think there is a slight misattribution in your message. The example
was given my Michael.
[...]
| Consider a processor whose integer addition instruction wraps. Then
| the cheapest implementation for examples 1 and 2 above that cover the
| defined cases is to eliminate the loop in case 1, and produce a modulo
| result in case 2. You worried about interaction between the two
| constructs. Consider
|
| /* int a, b, c; */
| if (b > 0) {
| a = b + c;
| int count;
| for (int i = c; i <= a; i++)
| count++;
| some_func(count);
| }
|
| Since behavior on integer overflow is undefined, we can optimize assuming
| that overflow has not occurred. Then a > c, so the for loop always
| executes b+1 times, and we end up with
|
| if (b > 0)
| some_func(b+1);
|
| Any attempt to assign meaning to integer overflow would prevent this
| optimization.
We document that
a = (int) ((unsigned) b + c)
is well-defined and given by the wrapping semantics. Does the current
optimizer takes that into account or will it assume b+1 execution times?
If the optimizer takes that into account, then the question becomes
when do we consider breaking the ABI to switch numeric_limits<signed
type>::is_modulo back to old behaviour.
-- Gaby
next prev parent reply other threads:[~2005-06-28 17:03 UTC|newest]
Thread overview: 119+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-06-28 4:08 Andrew Pinski
2005-06-28 4:20 ` Michael Veksler
2005-06-28 9:49 ` Robert Dewar
2005-06-28 4:34 ` Gabriel Dos Reis
2005-06-28 4:50 ` Andrew Pinski
2005-06-28 5:13 ` Gabriel Dos Reis
2005-06-28 5:34 ` Andrew Pinski
2005-06-28 6:01 ` Gabriel Dos Reis
2005-06-28 9:18 ` Robert Dewar
2005-06-28 11:50 ` Gabriel Dos Reis
2005-06-28 12:07 ` Robert Dewar
2005-06-28 12:33 ` Gabriel Dos Reis
2005-06-28 12:57 ` Robert Dewar
2005-06-28 13:19 ` Gabriel Dos Reis
2005-06-28 22:58 ` Georg Bauhaus
2005-06-28 23:53 ` Gabriel Dos Reis
2005-06-29 0:27 ` Robert Dewar
2005-06-29 0:43 ` Gabriel Dos Reis
2005-06-29 0:48 ` Robert Dewar
2005-06-29 1:14 ` Gabriel Dos Reis
2005-06-29 1:21 ` Diego Novillo
2005-06-29 2:19 ` Marcin Dalecki
2005-06-29 3:13 ` Scott Robert Ladd
2005-06-28 14:24 ` Olivier Galibert
2005-06-28 14:28 ` Jonathan Wilson
2005-06-28 14:42 ` Olivier Galibert
2005-06-28 14:39 ` Dave Korn
2005-06-28 14:52 ` Olivier Galibert
2005-06-28 15:01 ` Robert Dewar
2005-06-28 15:04 ` Andrew Haley
2005-06-28 17:18 ` Olivier Galibert
2005-06-28 17:36 ` Dave Korn
2005-06-28 18:02 ` Olivier Galibert
2005-06-28 18:36 ` Dave Korn
2005-06-28 18:56 ` Gabriel Dos Reis
2005-06-28 19:10 ` Olivier Galibert
2005-06-28 19:13 ` Andrew Pinski
2005-06-28 19:20 ` Robert Dewar
2005-06-28 21:48 ` Joe Buck
2005-06-28 19:25 ` Gabriel Dos Reis
2005-06-28 19:32 ` Robert Dewar
2005-06-28 19:48 ` Gabriel Dos Reis
2005-06-28 20:37 ` Robert Dewar
2005-06-28 20:58 ` Gabriel Dos Reis
2005-06-28 21:57 ` Robert Dewar
2005-06-28 21:44 ` Joe Buck
2005-06-28 21:50 ` Olivier Galibert
2005-06-28 21:59 ` Robert Dewar
2005-06-28 18:52 ` Robert Dewar
2005-06-28 19:17 ` Olivier Galibert
2005-06-28 19:21 ` Robert Dewar
2005-06-28 20:18 ` Paul Koning
2005-06-28 20:24 ` Robert Dewar
2005-06-28 21:41 ` Joe Buck
2005-06-28 21:53 ` Michael Veksler
2005-06-28 23:05 ` Michael Veksler
2005-07-02 17:15 ` Florian Weimer
2005-07-02 18:59 ` Gabriel Dos Reis
2005-07-02 23:20 ` Robert Dewar
2005-07-03 0:07 ` Gabriel Dos Reis
2005-07-03 9:49 ` Robert Dewar
2005-07-02 23:12 ` Nicholas Nethercote
2005-07-02 23:20 ` Robert Dewar
2005-07-03 0:13 ` Gabriel Dos Reis
2005-07-03 9:54 ` Robert Dewar
2005-07-03 10:02 ` Florian Weimer
2005-07-03 10:10 ` Robert Dewar
2005-07-03 12:01 ` Gabriel Dos Reis
2005-07-14 7:21 ` Marc Espie
2005-07-02 17:06 ` Florian Weimer
2005-06-28 17:51 ` Joe Buck
2005-06-28 18:21 ` Gabriel Dos Reis
2005-06-28 18:53 ` Robert Dewar
2005-06-28 18:28 ` Olivier Galibert
2005-06-28 18:38 ` Dave Korn
2005-06-28 18:50 ` Robert Dewar
2005-06-28 19:02 ` Gabriel Dos Reis
2005-06-28 19:17 ` Robert Dewar
2005-06-28 19:43 ` Gabriel Dos Reis
2005-06-28 20:31 ` Robert Dewar
2005-06-28 20:51 ` Gabriel Dos Reis
2005-06-28 20:59 ` Robert Dewar
2005-06-28 21:20 ` Gabriel Dos Reis
2005-06-28 21:27 ` Paul Koning
2005-06-28 21:39 ` Andreas Schwab
2005-06-28 21:35 ` Joe Buck
2005-06-28 22:09 ` Joseph S. Myers
2005-06-28 22:16 ` Falk Hueffner
2005-06-29 6:59 ` Eric Botcazou
2005-06-28 22:19 ` Robert Dewar
2005-06-28 16:42 ` Joe Buck
2005-06-28 17:10 ` Dave Korn
2005-06-28 17:21 ` Joe Buck
2005-06-28 22:41 ` Georg Bauhaus
2005-06-28 14:47 ` Gabriel Dos Reis
2005-06-28 16:38 ` Joe Buck
2005-06-28 21:59 ` Mike Stump
2005-06-28 13:47 ` Gabriel Paubert
2005-06-28 13:52 ` Andrew Pinski
2005-06-28 14:33 ` Robert Dewar
2005-06-28 12:08 ` Robert Dewar
2005-06-28 12:34 ` Gabriel Dos Reis
2005-06-28 7:25 ` Michael Veksler
2005-06-28 16:32 ` Joe Buck
2005-06-28 16:56 ` Joe Buck
2005-06-28 17:03 ` Gabriel Dos Reis [this message]
2005-06-28 17:34 ` Joe Buck
2005-06-28 18:09 ` Gabriel Dos Reis
2005-06-28 17:35 ` Diego Novillo
2005-06-28 6:55 ` Steven Bosscher
2005-06-28 7:20 ` Michael Veksler
2005-06-28 7:39 ` Falk Hueffner
2005-06-28 12:08 ` Gabriel Dos Reis
2005-06-28 12:01 ` Gabriel Dos Reis
2005-06-28 16:59 Morten Welinder
2005-06-28 17:23 ` Olivier Galibert
2005-06-28 18:44 ` Michael Veksler
2005-06-28 17:41 Paul Schlie
[not found] <2382433.1119938227627.JavaMail.root@dtm1eusosrv72.dtm.ops.eu.uu.net>
2005-06-28 19:44 ` Toon Moene
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=m3d5q6ifd2.fsf@uniton.integrable-solutions.net \
--to=gdr@integrable-solutions.net \
--cc=Joe.Buck@synopsys.COM \
--cc=VEKSLER@il.ibm.com \
--cc=gcc@gcc.gnu.org \
/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 read-only IMAP folder(s) and NNTP newsgroup(s).