public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "cvs-commit at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/104477] [C++23] Implement P2255R2, type trait to detect reference binding to temporary
Date: Fri, 15 Jul 2022 15:33:12 +0000	[thread overview]
Message-ID: <bug-104477-4-ROaOcYOeYd@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-104477-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104477

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:9a15d3beace26d68561cb3481b70b0bbcb122ca5

commit r13-1714-g9a15d3beace26d68561cb3481b70b0bbcb122ca5
Author: Marek Polacek <polacek@redhat.com>
Date:   Wed Jun 29 19:00:54 2022 -0400

    c++: Add __reference_con{struc,ver}ts_from_temporary [PR104477]

    This patch implements C++23 P2255R2, which adds two new type traits to
    detect reference binding to a temporary.  They can be used to detect code
    like

      std::tuple<const std::string&> t("meow");

    which is incorrect because it always creates a dangling reference, because
    the std::string temporary is created inside the selected constructor of
    std::tuple, and not outside it.

    There are two new compiler builtins, __reference_constructs_from_temporary
    and __reference_converts_from_temporary.  The former is used to simulate
    direct- and the latter copy-initialization context.  But I had a hard time
    finding a test where there's actually a difference.  Under DR 2267, both
    of these are invalid:

      struct A { } a;
      struct B { explicit B(const A&); };
      const B &b1{a};
      const B &b2(a);

    so I had to peruse [over.match.ref], and eventually realized that the
    difference can be seen here:

      struct G {
        operator int(); // #1
        explicit operator int&&(); // #2
      };

    int&& r1(G{}); // use #2 (no temporary)
    int&& r2 = G{}; // use #1 (a temporary is created to be bound to int&&)

    The implementation itself was rather straightforward because we already
    have the conv_binds_ref_to_prvalue function.  The main function here is
    ref_xes_from_temporary.
    I've changed the return type of ref_conv_binds_directly to tristate,
because
    previously the function didn't distinguish between an invalid conversion
and
    one that binds to a prvalue.  Since it no longer returns a bool, I removed
    the _p suffix.

    The patch also adds the relevant class and variable templates to
<type_traits>.

            PR c++/104477

    gcc/c-family/ChangeLog:

            * c-common.cc (c_common_reswords): Add
            __reference_constructs_from_temporary and
            __reference_converts_from_temporary.
            * c-common.h (enum rid): Add RID_REF_CONSTRUCTS_FROM_TEMPORARY and
            RID_REF_CONVERTS_FROM_TEMPORARY.

    gcc/cp/ChangeLog:

            * call.cc (ref_conv_binds_directly_p): Rename to ...
            (ref_conv_binds_directly): ... this.  Add a new bool parameter. 
Change
            the return type to tristate.
            * constraint.cc (diagnose_trait_expr): Handle
            CPTK_REF_CONSTRUCTS_FROM_TEMPORARY and
CPTK_REF_CONVERTS_FROM_TEMPORARY.
            * cp-tree.h: Include "tristate.h".
            (enum cp_trait_kind): Add CPTK_REF_CONSTRUCTS_FROM_TEMPORARY
            and CPTK_REF_CONVERTS_FROM_TEMPORARY.
            (ref_conv_binds_directly_p): Rename to ...
            (ref_conv_binds_directly): ... this.
            (ref_xes_from_temporary): Declare.
            * cxx-pretty-print.cc (pp_cxx_trait_expression): Handle
            CPTK_REF_CONSTRUCTS_FROM_TEMPORARY and
CPTK_REF_CONVERTS_FROM_TEMPORARY.
            * method.cc (ref_xes_from_temporary): New.
            * parser.cc (cp_parser_primary_expression): Handle
            RID_REF_CONSTRUCTS_FROM_TEMPORARY and
RID_REF_CONVERTS_FROM_TEMPORARY.
            (cp_parser_trait_expr): Likewise.
            (warn_for_range_copy): Adjust to call ref_conv_binds_directly.
            * semantics.cc (trait_expr_value): Handle
            CPTK_REF_CONSTRUCTS_FROM_TEMPORARY and
CPTK_REF_CONVERTS_FROM_TEMPORARY.
            (finish_trait_expr): Likewise.

    libstdc++-v3/ChangeLog:

            * include/std/type_traits (reference_constructs_from_temporary,
            reference_converts_from_temporary): New class templates.
            (reference_constructs_from_temporary_v,
            reference_converts_from_temporary_v): New variable templates.
            (__cpp_lib_reference_from_temporary): Define for C++23.
            * include/std/version (__cpp_lib_reference_from_temporary): Define
for
            C++23.
            * testsuite/20_util/variable_templates_for_traits.cc: Test
            reference_constructs_from_temporary_v and
            reference_converts_from_temporary_v.
            * testsuite/20_util/reference_from_temporary/value.cc: New test.
            * testsuite/20_util/reference_from_temporary/value2.cc: New test.
            * testsuite/20_util/reference_from_temporary/version.cc: New test.

    gcc/testsuite/ChangeLog:

            * g++.dg/ext/reference_constructs_from_temporary1.C: New test.
            * g++.dg/ext/reference_converts_from_temporary1.C: New test.

  parent reply	other threads:[~2022-07-15 15:33 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-02-09 19:46 [Bug c++/104477] New: " jason at gcc dot gnu.org
2022-02-09 19:47 ` [Bug c++/104477] " jason at gcc dot gnu.org
2022-05-27 15:04 ` mpolacek at gcc dot gnu.org
2022-05-27 15:24 ` jakub at gcc dot gnu.org
2022-06-29 23:20 ` mpolacek at gcc dot gnu.org
2022-06-29 23:23 ` pinskia at gcc dot gnu.org
2022-07-01 11:27 ` redi at gcc dot gnu.org
2022-07-15 15:33 ` cvs-commit at gcc dot gnu.org [this message]
2022-07-15 15:35 ` mpolacek 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-104477-4-ROaOcYOeYd@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).