From: Marek Polacek <polacek@redhat.com>
To: Jonathan Wakely <jwakely@redhat.com>
Cc: gcc-patches@gcc.gnu.org, libstdc++@gcc.gnu.org,
Jason Merrill <jason@redhat.com>
Subject: Re: [PATCH] c++: Make __is_{,nothrow_}convertible SFINAE on access [PR107049]
Date: Tue, 27 Sep 2022 08:18:36 -0400 [thread overview]
Message-ID: <YzLqHMMW8q1zWWN6@redhat.com> (raw)
In-Reply-To: <20220927103510.2321453-1-jwakely@redhat.com>
On Tue, Sep 27, 2022 at 11:35:10AM +0100, Jonathan Wakely wrote:
> Tested powerpc64le-linux. OK for trunk?
>
> -- >8 --
>
> The is_convertible built-ins should return false if the conversion fails
> an access check, not report an error.
Ah, so we do need that sentinel after all.
Patch looks good, thanks.
> PR c++/107049
>
> gcc/cp/ChangeLog:
>
> * method.cc (is_convertible_helper): Use access check sentinel.
>
> gcc/testsuite/ChangeLog:
>
> * g++.dg/ext/is_convertible4.C: New test.
> * g++.dg/ext/is_nothrow_convertible4.C: New test.
>
> libstdc++-v3/ChangeLog:
>
> * testsuite/20_util/is_convertible/requirements/access.cc: New
> test.
> ---
> gcc/cp/method.cc | 1 +
> gcc/testsuite/g++.dg/ext/is_convertible4.C | 33 +++++++++++++++++++
> .../g++.dg/ext/is_nothrow_convertible4.C | 33 +++++++++++++++++++
> .../is_convertible/requirements/access.cc | 18 ++++++++++
> 4 files changed, 85 insertions(+)
> create mode 100644 gcc/testsuite/g++.dg/ext/is_convertible4.C
> create mode 100644 gcc/testsuite/g++.dg/ext/is_nothrow_convertible4.C
> create mode 100644 libstdc++-v3/testsuite/20_util/is_convertible/requirements/access.cc
>
> diff --git a/gcc/cp/method.cc b/gcc/cp/method.cc
> index 9f917f13134..55af5c43c18 100644
> --- a/gcc/cp/method.cc
> +++ b/gcc/cp/method.cc
> @@ -2246,6 +2246,7 @@ is_convertible_helper (tree from, tree to)
> return integer_one_node;
> cp_unevaluated u;
> tree expr = build_stub_object (from);
> + deferring_access_check_sentinel acs (dk_no_deferred);
> return perform_implicit_conversion (to, expr, tf_none);
> }
>
> diff --git a/gcc/testsuite/g++.dg/ext/is_convertible4.C b/gcc/testsuite/g++.dg/ext/is_convertible4.C
> new file mode 100644
> index 00000000000..8a7724c5852
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/ext/is_convertible4.C
> @@ -0,0 +1,33 @@
> +// PR c++/107049
> +// { dg-do compile { target c++11 } }
> +// Failed access check should be a substitution failure, not an error.
> +
> +template<bool B>
> +struct bool_constant { static constexpr bool value = B; };
> +
> +template<typename From, typename To>
> +struct is_convertible
> +: public bool_constant<__is_convertible(From, To)>
> +{ };
> +
> +#if __cpp_variable_templates
> +template<typename From, typename To>
> +constexpr bool is_convertible_v = __is_convertible(From, To);
> +#endif
> +
> +class Private
> +{
> + operator int() const
> + {
> + static_assert( not is_convertible<Private, int>::value, "" );
> +#if __cpp_variable_templates
> + static_assert( not is_convertible_v<Private, int>, "" );
> +#endif
> + return 0;
> + }
> +};
> +
> +static_assert( not is_convertible<Private, int>::value, "" );
> +#if __cpp_variable_templates
> +static_assert( not is_convertible_v<Private, int>, "" );
> +#endif
> diff --git a/gcc/testsuite/g++.dg/ext/is_nothrow_convertible4.C b/gcc/testsuite/g++.dg/ext/is_nothrow_convertible4.C
> new file mode 100644
> index 00000000000..f81b5944ca2
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/ext/is_nothrow_convertible4.C
> @@ -0,0 +1,33 @@
> +// PR c++/107049
> +// { dg-do compile { target c++11 } }
> +// Failed access check should be a substitution failure, not an error.
> +
> +template<bool B>
> +struct bool_constant { static constexpr bool value = B; };
> +
> +template<typename From, typename To>
> +struct is_nt_convertible
> +: public bool_constant<__is_nothrow_convertible(From, To)>
> +{ };
> +
> +#if __cpp_variable_templates
> +template<typename From, typename To>
> +constexpr bool is_nt_convertible_v = __is_nothrow_convertible(From, To);
> +#endif
> +
> +class Private
> +{
> + operator int() const
> + {
> + static_assert( not is_nt_convertible<Private, int>::value, "" );
> +#if __cpp_variable_templates
> + static_assert( not is_nt_convertible_v<Private, int>, "" );
> +#endif
> + return 0;
> + }
> +};
> +
> +static_assert( not is_nt_convertible<Private, int>::value, "" );
> +#if __cpp_variable_templates
> +static_assert( not is_nt_convertible_v<Private, int>, "" );
> +#endif
> diff --git a/libstdc++-v3/testsuite/20_util/is_convertible/requirements/access.cc b/libstdc++-v3/testsuite/20_util/is_convertible/requirements/access.cc
> new file mode 100644
> index 00000000000..04a8c525961
> --- /dev/null
> +++ b/libstdc++-v3/testsuite/20_util/is_convertible/requirements/access.cc
> @@ -0,0 +1,18 @@
> +// { dg-do compile { target c++11 } }
> +// PR c++/107049
> +
> +#include <type_traits>
> +
> +class Private
> +{
> + operator int() const
> + {
> + static_assert( not std::is_convertible<Private, int>::value, "" );
> +#if __cpp_lib_type_trait_variable_templates
> + static_assert( not std::is_convertible_v<Private, int>, "" );
> +#endif
> + return 0;
> + }
> +};
> +
> +static_assert( not std::is_convertible<Private, int>::value, "" );
> --
> 2.37.3
>
Marek
next prev parent reply other threads:[~2022-09-27 12:18 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-27 10:35 Jonathan Wakely
2022-09-27 12:18 ` Marek Polacek [this message]
2022-09-27 12:53 ` Jason Merrill
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=YzLqHMMW8q1zWWN6@redhat.com \
--to=polacek@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jason@redhat.com \
--cc=jwakely@redhat.com \
--cc=libstdc++@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).