Add SafeBrowsingNavigationObserver to listen to navigation events

chrome/browser/safe_browsing/safe_browsing_navigation_observer.h

Index: chrome/browser/safe_browsing/safe_browsing_navigation_observer.h
diff --git a/chrome/browser/safe_browsing/safe_browsing_navigation_observer.h b/chrome/browser/safe_browsing/safe_browsing_navigation_observer.h
new file mode 100644
index 0000000000000000000000000000000000000000..07a98b16e0e7bc440a3fb6c910c16e3d605f9cbf
--- /dev/null
+++ b/chrome/browser/safe_browsing/safe_browsing_navigation_observer.h
@@ -0,0 +1,104 @@
+// Copyright 2016 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 CHROME_BROWSER_SAFE_BROWSING_SAFE_BROWSING_NAVIGATION_OBSERVER_H_
+#define CHROME_BROWSER_SAFE_BROWSING_SAFE_BROWSING_NAVIGATION_OBSERVER_H_
+
+#include "content/public/browser/navigation_handle.h"
+#include "content/public/browser/notification_observer.h"
+#include "content/public/browser/notification_registrar.h"
+#include "content/public/browser/web_contents_observer.h"
+#include "url/gurl.h"
+
+struct RetargetingDetails;
+
+namespace content {
+class NavigationHandle;
+class NotificationRegistrar;
+class RenderFrameHost;
+struct ResourceRedirectDetails;
+}
+
+namespace safe_browsing {
+
+// Observes the navigation events that might lead to a download.
+class SafeBrowsingNavigationObserver : public content::NotificationObserver,
+                                       public content::WebContentsObserver {
+ public:
+  struct NavigationEvent {
+    NavigationEvent();
+    NavigationEvent(const GURL& source_url,

[Nathan Parker, 2016/09/03 00:21:35]

Is there away to avoid the many-args-of-similar-type issue here?  They're easy
to mix up w/o the compiler noticing.

Not sure if it'd make things super wordy, but you could just remove this
constructor and require callers set each field.

[Jialiu Lin, 2016/09/07 00:42:41]

Agree. Removed this constructor. 

+                    int source_tab_id,
+                    const GURL& target_url,
+                    int target_tab_id,
+                    int frame_id,
+                    const GURL& main_frame_url,
+                    double timestamp,
+                    bool has_user_gesture,
+                    bool has_committed);
+    explicit NavigationEvent(const NavigationEvent& nav_event);
+    ~NavigationEvent();
+    GURL source_url;
+    int source_tab_id;
+
+    GURL target_url;
+    int target_tab_id;
+
+    int frame_id;         // Frame tree node id.
+    GURL main_frame_url;  // Main frame url of the source.
+
+    double timestamp;  // Timestamp of last time this object is updated.
+
+    bool is_user_initiated;
+    bool has_committed;
+    bool is_server_redirect;

[Nathan Parker, 2016/09/03 00:21:35]

It's is possible for is_user_initiated and is_server_redirect both to be true? 
(and if not, would an enum work better?)

[Jialiu Lin, 2016/09/07 00:42:41]

Yes, it's possible. User clicks on a link (url_a), and this link resolves to a
server redirection (url_b). In that case, both bools are true. And this counts
as one navigation event, since they shares a single NavigationHandle.

+    bool is_finished;
+    GURL server_redirect_url;
+  };
+
+  // NavigationMap keyed on target url.
+  typedef std::map<GURL, std::vector<NavigationEvent>> NavigationMap;
+  typedef std::map<content::NavigationHandle*, NavigationEvent>

[Nathan Parker, 2016/09/03 00:21:35]

Who owns these pointers? Add a comment, or down below where they're used.

[Jialiu Lin, 2016/09/07 00:42:41]

done.

+      NavigationHandleMap;
+
+  SafeBrowsingNavigationObserver();
+
+  ~SafeBrowsingNavigationObserver() override;
+
+  NavigationMap* navigation_map();
+
+  void ClearNavigationMap();
+
+  // TODO(jialiul): more functions coming for managing navigation_map_.
+
+ protected:
+  void RecordNavigationPendingEntry(
+      const content::NotificationSource& source,
+      const content::NotificationDetails& details);
+
+  void RecordRetargeting(const content::NotificationDetails& details);

[Nathan Parker, 2016/09/03 00:21:35]

What does retargeting mean?

[Jialiu Lin, 2016/09/07 00:42:41]

Retargeting means navigation triggered in one tab and finished in another, e.g.
open link in a new tab/window either by user or by code. 

+
+  // content::NotificationObserver:
+  void Observe(int type,
+               const content::NotificationSource& source,
+               const content::NotificationDetails& details) override;
+
+  // content::WebContentsObserver:
+  void DidStartNavigation(
+      content::NavigationHandle* navigation_handle) override;
+
+  void DidRedirectNavigation(
+      content::NavigationHandle* navigation_handle) override;
+
+  void DidFinishNavigation(
+      content::NavigationHandle* navigation_handle) override;
+
+  NavigationHandleMap navigation_handle_map_;

[Nathan Parker, 2016/09/03 00:21:35]

Add some comments on what these are used for, or how the lifetime of entries
relate to each other.

[Jialiu Lin, 2016/09/07 00:42:41]

done.

+  content::NotificationRegistrar registrar_;
+  NavigationMap navigation_map_;
+  DISALLOW_COPY_AND_ASSIGN(SafeBrowsingNavigationObserver);
+};
+
+}  // namespace safe_browsing
+
+#endif  // CHROME_BROWSER_SAFE_BROWSING_SAFE_BROWSING_NAVIGATION_OBSERVER_H_