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++/97912] New: Get rid of location-invariant requirement in std::function small object optimisation
Date: Thu, 19 Nov 2020 17:29:17 +0000	[thread overview]
Message-ID: <bug-97912-4@http.gcc.gnu.org/bugzilla/> (raw)

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97912

            Bug ID: 97912
           Summary: Get rid of location-invariant requirement in
                    std::function small object optimisation
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: redi at gcc dot gnu.org
                CC: barry.revzin at gmail dot com
  Target Milestone: ---

std::function will only store the target object in an internal buffer (rather
than on the heap) if it is "location invariant", which means trivially
copyable.

It would be good to remove the trivially copyable requirement, and store any
type of suitable size and alignment to fit in the buffer.

The difficulty is that std::function::swap does a byte-wise copy of the
_Any_data buffer, which is why we require trivially copyable. Changing that
might be an ABI break (existing code will have inlined that function::swap
definition, and will still do a byte-wise copy rather than whatever the new
code does to support non-trivially copyable types).

This testcase fails if we remove the trivially copyable requirement:

#include <functional>
#include <cassert>

struct selfobsessed
{
  selfobsessed() : me(this) { }

  selfobsessed(const selfobsessed&) : me(this) { }
  selfobsessed& operator=(const selfobsessed&) { return *this; }

  void operator()() const { assert(me == this); }

  selfobsessed* me;
};

int main()
{
  std::function<void()> f1( selfobsessed{} );
  std::function<void()> f2;
  swap(f1, f2);
  f2();
}

             reply	other threads:[~2020-11-19 17:29 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-11-19 17:29 redi at gcc dot gnu.org [this message]
2020-11-19 17:31 ` [Bug libstdc++/97912] " redi at gcc dot gnu.org
2021-08-31 16:13 ` 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-97912-4@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).