Metrics for tracking Safe Browsing hit patterns for the Safe Browsing lists Subresource Filter.

components/subresource_filter/content/browser/content_subresource_filter_driver_factory.h

Index: components/subresource_filter/content/browser/content_subresource_filter_driver_factory.h
diff --git a/components/subresource_filter/content/browser/content_subresource_filter_driver_factory.h b/components/subresource_filter/content/browser/content_subresource_filter_driver_factory.h
index 2e1056e39988ce7fcba95a7b8a5bebd1eecae79a..4fc56d8dc23b433514beb5d4a7c4d67aa6996a9d 100644
--- a/components/subresource_filter/content/browser/content_subresource_filter_driver_factory.h
+++ b/components/subresource_filter/content/browser/content_subresource_filter_driver_factory.h
@@ -24,7 +24,7 @@ class RenderFrameHost;
 
 namespace subresource_filter {
 
-using HostSet = std::set<std::string>;
+using HostPathSet = std::set<std::string>;
 
 class ContentSubresourceFilterDriver;
 class SubresourceFilterClient;
@@ -61,8 +61,6 @@ class ContentSubresourceFilterDriverFactory
   // filtering for the lifetime of this WebContents.
   void AddHostOfURLToWhitelistSet(const GURL& url);
 
-  void AddToActivationHitsSet(const GURL& url);
-
   // Called when Safe Browsing detects that the |url| corresponding to the load
   // of the main frame belongs to the blacklist with |threat_type|. If the
   // blacklist is the Safe Browsing Social Engineering ads landing, then |url|
@@ -76,9 +74,12 @@ class ContentSubresourceFilterDriverFactory
   // Reloads the page and inserts the url to the whitelist.
   void OnReloadRequested();
 
-  const HostSet& safe_browsing_blacklisted_patterns_set() const {
+  const HostPathSet& safe_browsing_blacklisted_patterns_set() const {
     return safe_browsing_blacklisted_patterns_;
   }
+
+  const HostPathSet& whitelisted_set() const { return whitelisted_hosts_; }
+
   ActivationState activation_state() { return activation_state_; }
 
  private:
@@ -99,10 +100,12 @@ class ContentSubresourceFilterDriverFactory
   void OnFirstSubresourceLoadDisallowed();
 
   // content::WebContentsObserver:
-  void RenderFrameCreated(content::RenderFrameHost* render_frame_host) override;
-  void RenderFrameDeleted(content::RenderFrameHost* render_frame_host) override;
   void DidStartNavigation(
       content::NavigationHandle* navigation_handle) override;
+  void DidRedirectNavigation(
+      content::NavigationHandle* navigation_handle) override;
+  void RenderFrameCreated(content::RenderFrameHost* render_frame_host) override;
+  void RenderFrameDeleted(content::RenderFrameHost* render_frame_host) override;
   void ReadyToCommitNavigation(
       content::NavigationHandle* navigation_handle) override;
   bool OnMessageReceived(const IPC::Message& message,
@@ -120,24 +123,39 @@ class ContentSubresourceFilterDriverFactory
       content::RenderFrameHost* render_frame_host,
       const GURL& url);
 
-  bool IsHit(const GURL& url) const;
+  bool IsBlacklistedBySafeBrowsing(const GURL& url) const;
+  bool IsInActivationSet(const GURL& url) const;
+
+  void AddToBlacklistedBySafebrowsing(const GURL& url);
+  void AddToActivationSet(const GURL& url);
+
+  void RecordRedirectChainMatchPattern() const;
 
   static const char kWebContentsUserDataKey[];
 
   FrameHostToOwnedDriverMap frame_drivers_;
   std::unique_ptr<SubresourceFilterClient> client_;
 
-  HostSet whitelisted_hosts_;
+  HostPathSet whitelisted_hosts_;
 
   // Host+path list of the URLs, where the Safe Browsing detected hit to the
   // threat list of interest. When the navigation is commited
   // |safe_browsing_blacklisted_patterns_| is used to determine whenever
   // the activation signal should be sent. All entities are deleted from the
-  // list on navigation commit event.
-  HostSet safe_browsing_blacklisted_patterns_;
+  // list on navigation start. The list is used to collect metrics for the
+  // activation patterns.
+  HostPathSet safe_browsing_blacklisted_patterns_;
+
+  // Host+path list of the URLs, where the Safe Browsing detected hit to the
+  // list, where the activation is enabled. The list is used to make activation
+  // decision. All entities are deleted from the list on navigation start.
+  HostPathSet activation_patterns_;

[engedy, 2016/10/28 10:58:07]

As discussed, let's combine this member, |navigation_chain_|, and
|sb_blacklisted_patterns| into one container that stores all URLs in the
navigation/redirect chain, and the list of SB blacklists that each of them
matches.

Then we can calculate the histogram samples, as well as the activation state, in
ReadyToCommitNavigation.

[melandory, 2016/11/16 15:59:35]

Done.

 
   ActivationState activation_state_;
 
+  // The first URL in the navigation.
+  std::vector<GURL> navigation_chain_;
+
   DISALLOW_COPY_AND_ASSIGN(ContentSubresourceFilterDriverFactory);
 };