We were saving up the access check perfectly well, but then
inadvertently throwing it away because of a call to
pop_deferring_access_checks (which discards them) rather than
pop_to_parent_deferring_access_checks (which adds them to the parent
context).
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 93fa63b45d657409ab35823b824f3acaa3e8b1ae
Author: Jason Merrill <ja...@redhat.com>
Date: Thu Mar 14 17:14:07 2013 -0400
PR c++/45917
* parser.c (cp_parser_template_id): Don't forget access checks.
diff --git a/gcc/cp/parser.c b/gcc/cp/parser.c
index b0df636..f4a4278 100644
--- a/gcc/cp/parser.c
+++ b/gcc/cp/parser.c
@@ -12818,7 +12818,7 @@ cp_parser_template_id (cp_parser *parser,
error_at (token->location, "parse error in template argument list");
}
- pop_deferring_access_checks ();
+ pop_to_parent_deferring_access_checks ();
return template_id;
}
diff --git a/gcc/testsuite/g++.dg/template/access26.C b/gcc/testsuite/g++.dg/template/access26.C
new file mode 100644
index 0000000..1c5de9a
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/access26.C
@@ -0,0 +1,6 @@
+// PR c++/45917
+
+template < typename T >
+struct A { static int i; };
+class B { typedef int X; }; // { dg-error "private" }
+void f() { A<B::X>::i = 0; } // { dg-error "this context" }