Merge 102541 - Change PhishingDOMFeatureExtractor to cache the WebDocument rather than a WebFrame...

chrome/renderer/safe_browsing/phishing_dom_feature_extractor.h

 // Copyright (c) 2011 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.
 //
 // PhishingDOMFeatureExtractor handles computing DOM-based features for the
 // client-side phishing detection model.  These include the presence of various
 // types of elements, ratios of external and secure links, and tokens for
 // external domains linked to.
 
 #ifndef CHROME_RENDERER_SAFE_BROWSING_PHISHING_DOM_FEATURE_EXTRACTOR_H_
 #define CHROME_RENDERER_SAFE_BROWSING_PHISHING_DOM_FEATURE_EXTRACTOR_H_
 
 #include <string>
 
 #include "base/basictypes.h"
 #include "base/callback_old.h"
 #include "base/memory/scoped_ptr.h"
 #include "base/task.h"
+#include "third_party/WebKit/Source/WebKit/chromium/public/WebDocument.h"
 
 class GURL;
 class RenderView;
 
 namespace WebKit {
 class WebElement;
-class WebFrame;
 }
 
 namespace safe_browsing {
 class FeatureExtractorClock;
 class FeatureMap;
 
 class PhishingDOMFeatureExtractor {
  public:
   // Callback to be run when feature extraction finishes.  The callback
   // argument is true if extraction was successful, false otherwise.
@@ skipping 59 matching lines @@
   // debug builds if the state is not as expected.  This is a no-op in release
   // builds.
   void CheckNoPendingExtraction();
 
   // Runs |done_callback_| and then clears all internal state.
   void RunCallback(bool success);
 
   // Clears all internal feature extraction state.
   void Clear();
 
-  // Called after advancing |cur_frame_| to update the state in
+  // Called after advancing |cur_document_| to update the state in
   // |cur_frame_data_|.  Returns true if the state was updated successfully.
   bool ResetFrameData();
 
+  // Returns the next document in frame-traversal order from cur_document_.
+  // If there are no more documents, returns a null WebDocument.
+  WebKit::WebDocument GetNextDocument();
+
   // Given a URL, checks whether the domain is different from the domain of
   // the current frame's URL.  If so, stores the domain in |domain| and returns
   // true, otherwise returns false.
   bool IsExternalDomain(const GURL& url, std::string* domain) const;
 
   // Called once all frames have been processed to compute features from the
   // PageFeatureState and add them to |features_|.  See features.h for a
   // description of which features are computed.
   void InsertFeatures();
 
   // Non-owned pointer to the view that we will extract features from.
   RenderView* render_view_;
 
   // Non-owned pointer to our clock.
   FeatureExtractorClock* clock_;
 
   // The output parameters from the most recent call to ExtractFeatures().
   FeatureMap* features_;  // The caller keeps ownership of this.
   scoped_ptr<DoneCallback> done_callback_;
 
-  // Non-owned pointer to the current frame that we are processing.
-  WebKit::WebFrame* cur_frame_;
+  // The current (sub-)document that we are processing.  May be a null document
+  // (isNull()) if we are not currently extracting features.
+  WebKit::WebDocument cur_document_;
 
-  // Stores extra state for |cur_frame_| that will be persisted until we
+  // Stores extra state for |cur_document_| that will be persisted until we
   // advance to the next frame.
   scoped_ptr<FrameData> cur_frame_data_;
 
   // Stores the intermediate data used to create features.  This data is
   // accumulated across all frames in the RenderView.
   scoped_ptr<PageFeatureState> page_feature_state_;
 
   // Used to create ExtractFeaturesWithTimeout tasks.
   // These tasks are revoked if extraction is cancelled.
   ScopedRunnableMethodFactory<PhishingDOMFeatureExtractor> method_factory_;
 
   DISALLOW_COPY_AND_ASSIGN(PhishingDOMFeatureExtractor);
 };
 
 }  // namespace safe_browsing
 
 #endif  // CHROME_RENDERER_SAFE_BROWSING_PHISHING_DOM_FEATURE_EXTRACTOR_H_