First step to move common accessibility code out of content.

content/public/common/ax_node.h

Index: content/public/common/ax_node.h
diff --git a/content/public/common/ax_node.h b/content/public/common/ax_node.h
new file mode 100644
index 0000000000000000000000000000000000000000..ce0f0113dbda760329e9b6911e177a20f591f633
--- /dev/null
+++ b/content/public/common/ax_node.h
@@ -0,0 +1,33 @@
+// Copyright (c) 2013 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef CONTENT_PUBLIC_COMMON_AX_NODE_H_
+#define CONTENT_PUBLIC_COMMON_AX_NODE_H_
+
+#include "content/public/common/ax_node_data.h"
+
+namespace content {
+
+class AXTree;
+
+// One node in an AXTree. The accessibility information is all in
+// data(); the rest of the interface is to walk the tree.
+class CONTENT_EXPORT AXNode {
+ public:
+  // Accessors.

[David Tseng, 2013/11/11 18:43:51]

The traversal API mirrors some of the same functionality provided by AXTree. Is
this intentional; if so, why?

Should the caller traverse using AXNode or AXtree?

[dmazzoni, 2013/11/12 00:03:04]

Hmmm, good point.

The only reason AXTree provides this is because of the AXSourceTree abstraction.

Perhaps AXSourceTree should be an adaptor rather than an actualy class? That
might simplify things...let me know what you think once you understand why I did
it this way.

+  virtual int32 GetId() const = 0;
+  virtual AXNode* GetParent() const = 0;
+  virtual int GetChildCount() const = 0;
+  virtual AXNode* ChildAtIndex(int index) const = 0;
+  virtual const AXNodeData& data() const = 0;
+
+ protected:
+  AXNode();
+  virtual ~AXNode();
+};
+
+
+}  // namespace content
+
+#endif  // CONTENT_PUBLIC_COMMON_AX_NODE_H_