Extracting page shingle hashes for similarity detection.

chrome/renderer/safe_browsing/phishing_classifier.h

Index: chrome/renderer/safe_browsing/phishing_classifier.h
diff --git a/chrome/renderer/safe_browsing/phishing_classifier.h b/chrome/renderer/safe_browsing/phishing_classifier.h
index 66e59695e4eb66d93af52ea5716174632d80acc2..2b17b1d4e60466f46a391839d7c237148a7a4e10 100644
--- a/chrome/renderer/safe_browsing/phishing_classifier.h
+++ b/chrome/renderer/safe_browsing/phishing_classifier.h
@@ -18,6 +18,8 @@
 #ifndef CHROME_RENDERER_SAFE_BROWSING_PHISHING_CLASSIFIER_H_
 #define CHROME_RENDERER_SAFE_BROWSING_PHISHING_CLASSIFIER_H_
 
+#include <set>
+
 #include "base/basictypes.h"
 #include "base/callback.h"
 #include "base/memory/scoped_ptr.h"
@@ -135,6 +137,7 @@ class PhishingClassifier {
 
   // State for any in-progress extraction.
   scoped_ptr<FeatureMap> features_;
+  scoped_ptr<std::set<uint32> > shingle_hashes_;
   const base::string16* page_text_;  // owned by the caller
   DoneCallback done_callback_;