Hi,
shame on me. I think the patch almost qualifies as obvious. Tested
x86_64-linux.
Thanks,
Paolo.
//////////////////
/cp
2012-05-10 Paolo Carlini <paolo.carl...@oracle.com>
PR c++/53301
* decl.c (check_default_argument): Fix typo (POINTER_TYPE_P
instead of TYPE_PTR_P) in zero-as-null-pointer-constant warning.
/testsuite
2012-05-10 Paolo Carlini <paolo.carl...@oracle.com>
PR c++/53301
* g++.dg/warn/Wzero-as-null-pointer-constant-6.C: New.
Index: testsuite/g++.dg/warn/Wzero-as-null-pointer-constant-6.C
===================================================================
--- testsuite/g++.dg/warn/Wzero-as-null-pointer-constant-6.C (revision 0)
+++ testsuite/g++.dg/warn/Wzero-as-null-pointer-constant-6.C (revision 0)
@@ -0,0 +1,6 @@
+// PR c++/53301
+// { dg-options "-Wzero-as-null-pointer-constant" }
+
+class x { public: x(int v) {} };
+
+void foo(const x& = 0);
Index: cp/decl.c
===================================================================
--- cp/decl.c (revision 187335)
+++ cp/decl.c (working copy)
@@ -10619,7 +10619,7 @@ check_default_argument (tree decl, tree arg)
if (warn_zero_as_null_pointer_constant
&& c_inhibit_evaluation_warnings == 0
- && (POINTER_TYPE_P (decl_type) || TYPE_PTR_TO_MEMBER_P (decl_type))
+ && (TYPE_PTR_P (decl_type) || TYPE_PTR_TO_MEMBER_P (decl_type))
&& null_ptr_cst_p (arg)
&& !NULLPTR_TYPE_P (TREE_TYPE (arg)))
{