Move browser initiated navigation from RenderViewHost to RenderFrameHost.

chrome/test/chromedriver/chrome/navigation_tracker.cc

Index: chrome/test/chromedriver/chrome/navigation_tracker.cc
diff --git a/chrome/test/chromedriver/chrome/navigation_tracker.cc b/chrome/test/chromedriver/chrome/navigation_tracker.cc
index 5f5fb3a705b77027d95c9ce3dbf8d3b47379fca2..3dffe43fb53192a8d0cef54ceb1ba2f9aa89e3ac 100644
--- a/chrome/test/chromedriver/chrome/navigation_tracker.cc
+++ b/chrome/test/chromedriver/chrome/navigation_tracker.cc
@@ -131,20 +131,22 @@ Status NavigationTracker::OnCommandSuccess(DevToolsClient* client,
     // it is unknown what will happen.
     //
     // There are a few cases (perhaps more):
-    // 1 The RenderViewHost has already queued ViewMsg_Navigate and loading
+    // 1 The RenderFrameHost has already queued FrameMsg_Navigate and loading
     //   will start shortly.
-    // 2 The RenderViewHost has already queued ViewMsg_Navigate and loading
+    // 2 The RenderFrameHost has already queued FrameMsg_Navigate and loading
     //   will never start because it is just an in-page fragment navigation.
-    // 3 The RenderViewHost is suspended and hasn't queued ViewMsg_Navigate
-    //   yet. This happens for cross-site navigations. The RenderViewHost
-    //   will not queue ViewMsg_Navigate until it is ready to unload the
+    // 3 The RenderFrameHost is suspended and hasn't queued FrameMsg_Navigate
+    //   yet. This happens for cross-site navigations. The RenderFrameHost
+    //   will not queue FrameMsg_Navigate until it is ready to unload the
     //   previous page (after running unload handlers and such).
+    // TODO(nasko): Revisit case 3, since now unload handlers are run in the

[Charlie Reis, 2014/02/13 01:23:53]

Good point.  Can you put a link to Camille's bug here for context?

[nasko, 2014/02/13 17:05:27]

Done.

+    // background.
     //
     // To determine whether a load is expected, do a round trip to the
     // renderer to ask what the URL is.
     // If case #1, by the time the command returns, the frame started to load
     // event will also have been received, since the DevTools command will
-    // be queued behind ViewMsg_Navigate.
+    // be queued behind FrameMsg_Navigate.
     // If case #2, by the time the command returns, the navigation will
     // have already happened, although no frame start/stop events will have
     // been received.