public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "redi at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/111685] Segfault while sorting on array element address
Date: Wed, 04 Oct 2023 09:48:25 +0000	[thread overview]
Message-ID: <bug-111685-4-IzgTlnNUuF@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-111685-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111685

--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> ---
The standard only defines sorting in terms of comparisons on "every iterator i
pointing to the sequence", which seems to preclude using a temporary object on
the stack that is outside the sequence.

That said, the comparison object seems like nonsense and I don't see any great
need to support this case.

We could define the linear insertion in terms of swaps, something like:

  template<typename _RandomAccessIterator, typename _Compare>
    _GLIBCXX20_CONSTEXPR
    void
    __unguarded_linear_insert(_RandomAccessIterator __last,
                              _Compare __comp)
    {
      _RandomAccessIterator __next = __last;
      --__next;
      while (__comp(*__last, __next))
        {
          std::iter_swap(__last, __next);
          __last = __next;
          --__next;
        }
    }

However, that would perform 3N copies, where the current code does N.

Alternatively, find the partition point first and then do N copies, something
like:

  template<typename _RandomAccessIterator, typename _Compare>
    _GLIBCXX20_CONSTEXPR
    void
    __unguarded_linear_insert(_RandomAccessIterator __last,
                              _Compare __comp)
    {
      _RandomAccessIterator __next = __last;
      --__next;
      while (__comp(*__last, __next))
        --__next;
      typename iterator_traits<_RandomAccessIterator>::value_type
        __val = _GLIBCXX_MOVE(*__last);
      _RandomAccessIterator __prev = __last;
      --__prev;
      while (__prev != __next)
        {
          *__last = _GLIBCXX_MOVE(*__prev);
          __last = __prev;
        }
      *__last = _GLIBCXX_MOVE(__val);
    }

But this traverses the sequence twice, where the original does it in one
traversal.

  parent reply	other threads:[~2023-10-04  9:48 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-03 18:57 [Bug c++/111685] New: " knoepfel at fnal dot gov
2023-10-03 19:06 ` [Bug libstdc++/111685] " pinskia at gcc dot gnu.org
2023-10-03 19:10 ` pinskia at gcc dot gnu.org
2023-10-03 19:14 ` knoepfel at fnal dot gov
2023-10-03 19:19 ` knoepfel at fnal dot gov
2023-10-03 21:24 ` pinskia at gcc dot gnu.org
2023-10-03 21:37 ` pinskia at gcc dot gnu.org
2023-10-04  1:35 ` xry111 at gcc dot gnu.org
2023-10-04  9:48 ` redi at gcc dot gnu.org [this message]
2023-10-04 10:06 ` redi at gcc dot gnu.org
2023-10-04 17:46 ` fchelnokov at gmail dot com
2023-10-05 19:27 ` knoepfel at fnal dot gov
2023-10-06 10:10 ` fchelnokov at gmail dot com

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-111685-4-IzgTlnNUuF@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).