public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
From: plaflamme@konova.com
To: gcc-gnats@gcc.gnu.org
Cc: lletourneau@konova.com
Subject: c++/7707: GCC 3.1.1 Optimization incomplete/incorrect : leaks temporary object
Date: Fri, 23 Aug 2002 16:36:00 -0000	[thread overview]
Message-ID: <20020823230136.15407.qmail@sources.redhat.com> (raw)


>Number:         7707
>Category:       c++
>Synopsis:       GCC 3.1.1 Optimization incomplete/incorrect : leaks temporary object
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          wrong-code
>Submitter-Id:   net
>Arrival-Date:   Fri Aug 23 16:06:00 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     Konova Solutions Inc.
>Release:        GCC 3.1.1
>Organization:
>Environment:
RedHat Linux 7.3
Dual Intel Pentium III 833 MHz
GCC compiled with:
--prefix=/usr/gcc-3.1.1 --enable-threads=posix --enable-languages=c,c++
>Description:
g++ seems not to optimize temporary objects copy correctly and results in a memory leak. The cpp file attached provides a example of how g++ fails to destruct a temporary object.

The output of the example shows that 2 temporary objects are created (using copy ctor) while only one of them is destructed.

A temporary object is created on the stack in method 
operator+(const foo&, const foo&). The result of the operator+ is passed to another method taking a const foo as parameter. Doing so, g++ calls the copy constructor. 

It is up to the compiler to actually call the copy ctor or not. g++ does not seem to optimize this copy, but then fails to destroy the object.

The attachment contains the cpp and ii files.
>How-To-Repeat:

>Fix:
Modifing the leakAFoo method to take a const foo& instead.

g++ is still to blame. Either the optimization is incomplete, or some logic is erronous when destroying  temporary objects.
>Release-Note:
>Audit-Trail:
>Unformatted:


             reply	other threads:[~2002-08-23 23:06 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-08-23 16:36 plaflamme [this message]
2002-08-23 23:06 Phil Edwards
2002-08-26  7:16 Philippe Laflamme
2002-08-30  2:26 Philippe Laflamme
2002-09-13 14:59 nathan

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=20020823230136.15407.qmail@sources.redhat.com \
    --to=plaflamme@konova.com \
    --cc=gcc-gnats@gcc.gnu.org \
    --cc=lletourneau@konova.com \
    /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).