C++ PATCH for c++/51889 (problem overloading function using in template)

C++ PATCH for c++/51889 (problem overloading function using in template)

[Jason Merrill]

[-- Attachment #1: Type: text/plain, Size: 698 bytes --]

When we have a class-scope using-declaration that nominates functions, 
we want to insert those functions into the derived class' 
CLASSTYPE_METHOD_VEC.  In non-template code we do this in 
handle_using_decl, which is called from check_bases_and_members.  But we 
don't call that function for a class template definition.  Previously, 
for templates we were inserting them from finish_member_declaration, but 
that leads to name collisions if a function with the same signature is 
declared after the using-declaration, as in this testcase.  So this 
patch delays inserting the used function until finish_struct time, to 
match the non-template case.

Tested x86_64-pc-linux-gnu, applied to trunk.

[-- Attachment #2: 51889.patch --]
[-- Type: text/x-patch, Size: 2406 bytes --]

commit 542a826b421991226a5083f769fd1054339b4ff6
Author: Jason Merrill <jason@redhat.com>
Date:   Thu Jan 19 00:53:45 2012 -0500

    	PR c++/51889
    	* class.c (finish_struct): Call add_method here for function usings.
    	* semantics.c (finish_member_declaration): Not here.

diff --git a/gcc/cp/class.c b/gcc/cp/class.c
index 9b957fe..e6f33fe 100644
--- a/gcc/cp/class.c
+++ b/gcc/cp/class.c
@@ -6195,6 +6195,18 @@ finish_struct (tree t, tree attributes)
 	if (DECL_PURE_VIRTUAL_P (x))
 	  VEC_safe_push (tree, gc, CLASSTYPE_PURE_VIRTUALS (t), x);
       complete_vars (t);
+      /* We need to add the target functions to the CLASSTYPE_METHOD_VEC if
+	 an enclosing scope is a template class, so that this function be
+	 found by lookup_fnfields_1 when the using declaration is not
+	 instantiated yet.  */
+      for (x = TYPE_FIELDS (t); x; x = DECL_CHAIN (x))
+	if (TREE_CODE (x) == USING_DECL)
+	  {
+	    tree fn = strip_using_decl (x);
+	    if (is_overloaded_fn (fn))
+	      for (; fn; fn = OVL_NEXT (fn))
+		add_method (t, OVL_CURRENT (fn), x);
+	  }
 
       /* Remember current #pragma pack value.  */
       TYPE_PRECISION (t) = maximum_field_alignment;
diff --git a/gcc/cp/semantics.c b/gcc/cp/semantics.c
index 40676b6..a5a10d0 100644
--- a/gcc/cp/semantics.c
+++ b/gcc/cp/semantics.c
@@ -2671,20 +2671,6 @@ finish_member_declaration (tree decl)
     {
       if (TREE_CODE (decl) == USING_DECL)
 	{
-	  /* We need to add the target functions to the
-	     CLASSTYPE_METHOD_VEC if an enclosing scope is a template
-	     class, so that this function be found by lookup_fnfields_1
-	     when the using declaration is not instantiated yet.  */
-
-	  tree target_decl = strip_using_decl (decl);
-	  if (dependent_type_p (current_class_type)
-	      && is_overloaded_fn (target_decl))
-	    {
-	      tree t = target_decl;
-	      for (; t; t = OVL_NEXT (t))
-		add_method (current_class_type, OVL_CURRENT (t), decl);
-	    }
-
 	  /* For now, ignore class-scope USING_DECLS, so that
 	     debugging backends do not see them. */
 	  DECL_IGNORED_P (decl) = 1;
diff --git a/gcc/testsuite/g++.dg/inherit/using7.C b/gcc/testsuite/g++.dg/inherit/using7.C
new file mode 100644
index 0000000..de177c9
--- /dev/null
+++ b/gcc/testsuite/g++.dg/inherit/using7.C
@@ -0,0 +1,12 @@
+// PR c++/51889
+
+struct A {
+  void f();
+};
+
+template <class T>
+struct B: A
+{
+  using A::f;
+  void f();
+};

Re: C++ PATCH for c++/51889 (problem overloading function using in template)

[Fabien Chêne]

2012/1/19 Jason Merrill <jason@redhat.com>:
> When we have a class-scope using-declaration that nominates functions, we
> want to insert those functions into the derived class' CLASSTYPE_METHOD_VEC.
>  In non-template code we do this in handle_using_decl, which is called from
> check_bases_and_members.  But we don't call that function for a class
> template definition.  Previously, for templates we were inserting them from
> finish_member_declaration, but that leads to name collisions if a function
> with the same signature is declared after the using-declaration, as in this
> testcase.  So this patch delays inserting the used function until
> finish_struct time, to match the non-template case.

Thanks for fixing it !

-- 
Fabien