From: Jason Merrill <jason@redhat.com>
To: Patrick Palka <ppalka@redhat.com>, gcc-patches@gcc.gnu.org
Subject: Re: [PATCH 2/2] c++: redeclaring member of constrained class template [PR96830]
Date: Tue, 14 Mar 2023 17:21:42 -0400 [thread overview]
Message-ID: <24e806d0-e329-1317-249c-a9aaba0e295a@redhat.com> (raw)
In-Reply-To: <20230314164146.1470993-2-ppalka@redhat.com>
On 3/14/23 12:41, Patrick Palka wrote:
> When redeclaring a member of a constrained class template, we need to
> repeat the class's constraints, but it turns out we don't verify
> anywhere that the repeated constraints match the class's constraints.
> A safe place to do so seems to be in push_template_decl, nearby a
> similar consistency check for the template parameter list length of a
> redeclaration.
>
> Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK for
> trunk/12?
Both patches are OK for trunk, but not 12; we tend not to backport fixes
for accepts-invalid bugs, especially if they aren't regressions. We
don't want to add new diagnostics to an active release branch.
> PR c++/96830
>
> gcc/cp/ChangeLog:
>
> * pt.cc (push_inline_template_parms_recursive): Set
> TEMPLATE_PARMS_CONSTRAINTS.
> (push_template_decl): For a redeclaration, verify constraints
> for each enclosing template scope match those of the original
> template declaratation.
>
> gcc/testsuite/ChangeLog:
>
> * g++.dg/cpp2a/concepts-class5.C: New test.
> * g++.dg/cpp2a/concepts-class5a.C: New test.
> ---
> gcc/cp/pt.cc | 26 +++++++++++++
> gcc/testsuite/g++.dg/cpp2a/concepts-class5.C | 37 ++++++++++++++++++
> gcc/testsuite/g++.dg/cpp2a/concepts-class5a.C | 38 +++++++++++++++++++
> 3 files changed, 101 insertions(+)
> create mode 100644 gcc/testsuite/g++.dg/cpp2a/concepts-class5.C
> create mode 100644 gcc/testsuite/g++.dg/cpp2a/concepts-class5a.C
>
> diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
> index 053cc52b285..370a4f70f08 100644
> --- a/gcc/cp/pt.cc
> +++ b/gcc/cp/pt.cc
> @@ -471,6 +471,8 @@ push_inline_template_parms_recursive (tree parmlist, int levels)
> current_template_parms
> = tree_cons (size_int (current_template_depth + 1),
> parms, current_template_parms);
> + TEMPLATE_PARMS_CONSTRAINTS (current_template_parms)
> + = TEMPLATE_PARMS_CONSTRAINTS (parmlist);
> TEMPLATE_PARMS_FOR_INLINE (current_template_parms) = 1;
>
> begin_scope (TREE_VEC_LENGTH (parms) ? sk_template_parms : sk_template_spec,
> @@ -6134,6 +6136,30 @@ push_template_decl (tree decl, bool is_friend)
> DECL_INTERFACE_KNOWN (decl) = 1;
> return error_mark_node;
> }
> +
> + /* Check that the constraints for each enclosing template scope are
> + consistent with the original template declaration. */
> + if (flag_concepts)
> + {
> + tree decl_parms = DECL_TEMPLATE_PARMS (tmpl);
> + tree scope_parms = current_template_parms;
> + if (PRIMARY_TEMPLATE_P (tmpl))
> + {
> + decl_parms = TREE_CHAIN (decl_parms);
> + scope_parms = TREE_CHAIN (scope_parms);
> + }
> + while (decl_parms)
> + {
> + if (!template_requirements_equivalent_p (decl_parms, scope_parms))
> + {
> + error ("redeclaration of %qD with different constraints",
> + TPARMS_PRIMARY_TEMPLATE (TREE_VALUE (decl_parms)));
> + break;
> + }
> + decl_parms = TREE_CHAIN (decl_parms);
> + scope_parms = TREE_CHAIN (scope_parms);
> + }
> + }
> }
>
> gcc_checking_assert (!tmpl || DECL_TEMPLATE_RESULT (tmpl) == decl);
> diff --git a/gcc/testsuite/g++.dg/cpp2a/concepts-class5.C b/gcc/testsuite/g++.dg/cpp2a/concepts-class5.C
> new file mode 100644
> index 00000000000..5d893d9f6a7
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp2a/concepts-class5.C
> @@ -0,0 +1,37 @@
> +// PR c++/96830
> +// { dg-do compile { target c++20 } }
> +
> +template<class T> concept C = true;
> +
> +template<C T> requires true
> +struct A {
> + void f();
> + template<class U> void g();
> +
> + struct B;
> + template<class U> struct C;
> +
> + static int D;
> + template<class U> static int E;
> +};
> +
> +template<C T>
> +void A<T>::f() { } // { dg-error "different constraints" }
> +
> +template<class T> requires true
> +template<class U>
> +void A<T>::g() { } // { dg-error "different constraints" }
> +
> +template<C T> requires (!!true)
> +struct A<T>::B { }; // { dg-error "different constraints" }
> +
> +template<class T> requires true
> +template<class U>
> +struct A<T>::C { }; // { dg-error "different constraints" }
> +
> +template<C T>
> +int A<T>::D = 0; // { dg-error "different constraints" }
> +
> +template<class T> requires true
> +template<class U>
> +int A<T>::E = 0; // { dg-error "different constraints" }
> diff --git a/gcc/testsuite/g++.dg/cpp2a/concepts-class5a.C b/gcc/testsuite/g++.dg/cpp2a/concepts-class5a.C
> new file mode 100644
> index 00000000000..a6a3684c258
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/cpp2a/concepts-class5a.C
> @@ -0,0 +1,38 @@
> +// PR c++/96830
> +// A valid version of concepts-class5.C.
> +// { dg-do compile { target c++20 } }
> +
> +template<class T> concept C = true;
> +
> +template<C T> requires true
> +struct A {
> + void f();
> + template<class U> void g();
> +
> + struct B;
> + template<class U> struct C;
> +
> + static int D;
> + template<class U> static int E;
> +};
> +
> +template<C T> requires true
> +void A<T>::f() { }
> +
> +template<C T> requires true
> +template<class U>
> +void A<T>::g() { }
> +
> +template<C T> requires true
> +struct A<T>::B { };
> +
> +template<C T> requires true
> +template<class U>
> +struct A<T>::C { };
> +
> +template<C T> requires true
> +int A<T>::D = 0;
> +
> +template<C T> requires true
> +template<class U>
> +int A<T>::E = 0;
prev parent reply other threads:[~2023-03-14 21:21 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-03-14 16:41 [PATCH 1/2] c++: constrained template friend class matching [PR96830] Patrick Palka
2023-03-14 16:41 ` [PATCH 2/2] c++: redeclaring member of constrained class template [PR96830] Patrick Palka
2023-03-14 21:21 ` 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=24e806d0-e329-1317-249c-a9aaba0e295a@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).