From 9f26e34a5a9614a5b66f146752ecef9ea67b3e2d Mon Sep 17 00:00:00 2001
From: Patrick Palka <ppalka@redhat.com>
Date: Sat, 26 Jun 2021 11:05:54 -0400
Subject: [PATCH] c++: access scope during partial spec matching [PR96204]
Here, when determining whether the partial specialization matches
has_type_member<Child>, we do so from the scope of where the template-id
appears rather than from the scope of the specialization, and this
causes us to select the partial specialization (since Child::type is
accessible from Parent). When we later instantiate this partial
specialization, we've entered the scope of the specialization and so
substitution into e.g. the DECL_CONTEXT of has_type_member::value fails
with access errors since the friend declaration that we relied on to
choose the partial specialization no longer applies.
It seems the appropriate access scope from which to perform partial
specialization matching is the specialization itself (similar to how
we check access of base-clauses), which is what this patch implements.
PR c++/96204
gcc/cp/ChangeLog:
* pt.c (instantiate_class_template_1): Enter the scope of the
type when calling most_specialized_partial_spec.
gcc/testsuite/ChangeLog:
* g++.dg/template/access40.C: New test.
* g++.dg/template/access40a.C: New test.
---
gcc/cp/pt.c | 5 +++-
gcc/testsuite/g++.dg/template/access40.C | 28 +++++++++++++++++++++++
gcc/testsuite/g++.dg/template/access40a.C | 28 +++++++++++++++++++++++
3 files changed, 60 insertions(+), 1 deletion(-)
create mode 100644 gcc/testsuite/g++.dg/template/access40.C
create mode 100644 gcc/testsuite/g++.dg/template/access40a.C
diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
index e5a2a2cd5255..f2039e09cd7c 100644
--- a/gcc/cp/pt.c
+++ b/gcc/cp/pt.c
@@ -11769,8 +11769,11 @@ instantiate_class_template_1 (tree type)
deferring_access_check_sentinel acs (dk_no_deferred);
/* Determine what specialization of the original template to
- instantiate. */
+ instantiate; do this relative to the scope of the class for
+ sake of access checking. */
+ push_nested_class (type);
t = most_specialized_partial_spec (type, tf_warning_or_error);
+ pop_nested_class ();
if (t == error_mark_node)
return error_mark_node;
else if (t)
diff --git a/gcc/testsuite/g++.dg/template/access40.C b/gcc/testsuite/g++.dg/template/access40.C
new file mode 100644
index 000000000000..d035e99e462f
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/access40.C
@@ -0,0 +1,28 @@
+// PR c++/96204
+
+template<class, class = void>
+struct has_type_member {
+ static const bool value = false;
+};
+
+template<class T>
+struct has_type_member<T, typename T::type> {
+ static const bool value = true;
+};
+
+struct Parent;
+
+struct Child {
+private:
+ friend struct Parent;
+ typedef void type;
+};
+
+struct Parent {
+ static void f() {
+ // The partial specialization does not match despite Child::type
+ // being accessible from the current scope.
+ extern int x[1];
+ extern int x[!has_type_member<Child>::value];
+ }
+};
diff --git a/gcc/testsuite/g++.dg/template/access40a.C b/gcc/testsuite/g++.dg/template/access40a.C
new file mode 100644
index 000000000000..94025c513b78
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/access40a.C
@@ -0,0 +1,28 @@
+// PR c++/96204
+
+template<class, class = void>
+struct has_type_member {
+ static const bool value = false;
+};
+
+template<class T>
+struct has_type_member<T, typename T::type> {
+ static const bool value = true;
+};
+
+struct Parent;
+
+struct Child {
+private:
+ friend struct has_type_member<Child>;
+ typedef void type;
+};
+
+struct Parent {
+ static void f() {
+ // The partial specialization matches because has_type_member<Child>
+ // is a friend of Child.
+ extern int x[1];
+ extern int x[has_type_member<Child>::value];
+ }
+};
--
GitLab