public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "personal@e-maxx.ru" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/61751] New: Empty brace-initializer causes double destruction of unique_ptr
Date: Tue, 08 Jul 2014 16:34:00 -0000	[thread overview]
Message-ID: <bug-61751-4@http.gcc.gnu.org/bugzilla/> (raw)

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61751

            Bug ID: 61751
           Summary: Empty brace-initializer causes double destruction of
                    unique_ptr
           Product: gcc
           Version: 4.8.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: personal@e-maxx.ru

Created attachment 33091
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33091&action=edit
Full example.

Use of empty brace-initializer in default function argument causes strange
effects and, finally, crashes.

The minimal code is:


#include <memory>

class A {
    std::unique_ptr<int> ptr_;
public:
    A() : ptr_(new int(123))
    { }
    A(A&& other) : ptr_(std::move(other.ptr_))
    { }
};

void f(A a)
{ }

void g(A a = {}) // replace "{}" with "A()" makes it work
{ f(std::move(a)); }

int main()
{ g(); }


There is a more detailed example in the attachment, which produces some debug
output, e.g.:

A() called [this=0x7ffffbac34f0, constructed unique_ptr=0xe18010]
A(A&&) called [this=0x7ffffbac34b0, other=0x7ffffbac34d0]
~A() called [this=0x7ffffbac34b0, unique_ptr=0xe18010]
~A() called [this=0x7ffffbac34f0, unique_ptr=0xe18010]
*** glibc detected *** ./a.out: double free or corruption (fasttop):
0x0000000000e18010 ***

On the contrary, using some correct version of compiler (I tried 4.6.3 and
4.9.0) we get:

A() called [this=0x7fff77c52810, constructed unique_ptr=0xe1c010]
A(A&&) called [this=0x7fff77c527e0, other=0x7fff77c52810]
~A() called [this=0x7fff77c527e0, unique_ptr=0xe1c010]
~A() called [this=0x7fff77c52810, unique_ptr=0]

As it can be seen, the difference is that the bogus version moves from object
that has never been constructed.


             reply	other threads:[~2014-07-08 16:34 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-08 16:34 personal@e-maxx.ru [this message]
2014-07-08 17:32 ` [Bug c++/61751] " redi at gcc dot gnu.org
2014-07-08 17:33 ` redi at gcc dot gnu.org

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=bug-61751-4@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@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).