From: Jason Merrill <jason@redhat.com>
To: Patrick Palka <ppalka@redhat.com>, gcc-patches@gcc.gnu.org
Subject: Re: [PATCH] c++: unqual lookup performed twice w/ template-id ADL [PR102670]
Date: Wed, 17 Nov 2021 23:17:50 -0500 [thread overview]
Message-ID: <aa3aaeca-2441-8aaa-eed3-7faba7ee5ff9@redhat.com> (raw)
In-Reply-To: <20211103160419.1472321-1-ppalka@redhat.com>
On 11/3/21 12:04, Patrick Palka wrote:
> Here we're incorrectly performing unqualified lookup of 'adl' again at
> substitution time for the call adl<I>(t) (for which name lookup at parse
> time found nothing) which causes us to reject the testcase because the
> second unqualified lookup finds the later-declared variable template
> 'adl', leading to confusion. Fixed thusly.
>
> The testcase concepts-recursive-sat1.C needed to be adjusted use ADL
> proper instead of relying on this incorrect behavior.
>
> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
> trunk and perhaps 11 given it's a C++20 bugfix?
OK for trunk. Not for 11, I think, as it also affects non-C++20 code.
> PR c++/102670
>
> gcc/cp/ChangeLog:
>
> * pt.c (tsubst_copy_and_build) <case CALL_EXPR>: When looking
> for an identifier callee in the koenig_p case, also look through
> TEMPLATE_ID_EXPR. Use tsubst_copy to substitute through the
> template arguments of the template-id.
>
> gcc/testsuite/ChangeLog:
>
> * g++.dg/cpp2a/concepts-recursive-sat1.C:
> * g++.dg/cpp2a/fn-template23.C: New test.
> ---
> gcc/cp/pt.c | 11 +++++-
> .../g++.dg/cpp2a/concepts-recursive-sat1.C | 15 +++++---
> gcc/testsuite/g++.dg/cpp2a/fn-template23.C | 36 +++++++++++++++++++
> 3 files changed, 56 insertions(+), 6 deletions(-)
> create mode 100644 gcc/testsuite/g++.dg/cpp2a/fn-template23.C
>
> diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
> index 66040035b2f..40f84648ed2 100644
> --- a/gcc/cp/pt.c
> +++ b/gcc/cp/pt.c
> @@ -20256,7 +20256,10 @@ tsubst_copy_and_build (tree t,
> /*done=*/false,
> /*address_p=*/false);
> }
> - else if (koenig_p && identifier_p (function))
> + else if (koenig_p
> + && (identifier_p (function)
> + || (TREE_CODE (function) == TEMPLATE_ID_EXPR
> + && identifier_p (TREE_OPERAND (function, 0)))))
> {
> /* Do nothing; calling tsubst_copy_and_build on an identifier
> would incorrectly perform unqualified lookup again.
> @@ -20269,6 +20272,12 @@ tsubst_copy_and_build (tree t,
> FIXME but doing that causes c++/15272, so we need to stop
> using IDENTIFIER_NODE in that situation. */
> qualified_p = false;
> +
> + if (TREE_CODE (function) == TEMPLATE_ID_EXPR)
> + /* Use tsubst_copy to substitute through the template arguments
> + of the template-id without performing unqualified lookup on
> + the template name. */
> + function = tsubst_copy (function, args, complain, in_decl);
> }
> else
> {
> diff --git a/gcc/testsuite/g++.dg/cpp2a/concepts-recursive-sat1.C b/gcc/testsuite/g++.dg/cpp2a/concepts-recursive-sat1.C
> index 22696c30d81..4c178b77946 100644
> --- a/gcc/testsuite/g++.dg/cpp2a/concepts-recursive-sat1.C
> +++ b/gcc/testsuite/g++.dg/cpp2a/concepts-recursive-sat1.C
> @@ -3,16 +3,21 @@
> template<int N, typename T>
> concept Foo = requires(T t) { foo<N + 1>(t); }; // { dg-error "template instantiation depth" }
>
> -template<int N = 1, typename T = int>
> - requires Foo<N, T>
> -int foo(T t)
> +namespace ns
> {
> - return foo<N + 1>(t);
> + struct S { };
> +
> + template<int N, typename T>
> + requires Foo<N, T>
> + int foo(T t)
> + {
> + return foo<N + 1>(t);
> + }
> }
>
> int main(int, char**)
> {
> - return foo<1>(1);
> + return ns::foo<1>(ns::S{});
> }
>
> // { dg-prune-output "compilation terminated" }
> diff --git a/gcc/testsuite/g++.dg/cpp2a/fn-template23.C b/gcc/testsuite/g++.dg/cpp2a/fn-template23.C
> new file mode 100644
> index 00000000000..b85d4c96dab
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp2a/fn-template23.C
> @@ -0,0 +1,36 @@
> +// PR c++/102670
> +// { dg-do compile { target c++20 } }
> +
> +namespace ns {
> + struct S { };
> +
> + template<int I>
> + constexpr int adl(const S &) {
> + return I;
> + }
> +}
> +
> +namespace redirect {
> + template<class T, int I>
> + concept can_call_adl = requires(T t) {
> + adl<I>(t);
> + };
> +
> + template<int I>
> + struct adl_fn {
> + template<can_call_adl<I> T>
> + constexpr decltype(auto) operator()(T t) const {
> + return adl<I>(t);
> + }
> + };
> +
> + namespace {
> + template<int I>
> + constexpr inline adl_fn<I> adl{};
> + }
> +}
> +
> +int main() {
> + static_assert(redirect::can_call_adl<ns::S, 3>);
> + redirect::adl<3>(ns::S{});
> +}
>
prev parent reply other threads:[~2021-11-18 4:18 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-03 16:04 Patrick Palka
2021-11-10 16:53 ` [PATCH] c++: template-id ADL and partial instantiation [PR99911] Patrick Palka
2021-11-18 5:15 ` Jason Merrill
2021-11-18 14:45 ` Patrick Palka
2021-11-18 15:31 ` Patrick Palka
2021-11-18 16:02 ` Jason Merrill
2021-11-18 4:17 ` Jason Merrill [this message]
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=aa3aaeca-2441-8aaa-eed3-7faba7ee5ff9@redhat.com \
--to=jason@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=ppalka@redhat.com \
/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).