PlzNavigate: implement CommitNavigation on the browser side

content/common/frame_messages.h

Index: content/common/frame_messages.h
diff --git a/content/common/frame_messages.h b/content/common/frame_messages.h
index d9aca27058cc49c017d423b5357fa2cc16eaa94e..3ce361e1d7ce20394c4969844c90c06a97844562 100644
--- a/content/common/frame_messages.h
+++ b/content/common/frame_messages.h
@@ -305,6 +305,62 @@ IPC_STRUCT_BEGIN(FrameHostMsg_BeginNavigation_Params)
   IPC_STRUCT_MEMBER(bool, allow_download)
 IPC_STRUCT_END()
 
+IPC_STRUCT_BEGIN(FrameMsg_CommitNavigation_Params)
+  // The URL to commit.
+  IPC_STRUCT_MEMBER(GURL, url)
+
+  // The URL to request to the browser to get access to the stream of data.
+  IPC_STRUCT_MEMBER(GURL, stream_url)
+
+  // The page_id for this navigation, or -1 if it is a new navigation.  Back,
+  // Forward, and Reload navigations should have a valid page_id.  If the load
+  // succeeds, then this page_id will be reflected in the resultant
+  // FrameHostMsg_DidCommitProvisionalLoad message.
+  IPC_STRUCT_MEMBER(int32, page_id)
+
+  // If page_id is -1, then pending_history_list_offset will also be -1.
+  // Otherwise, it contains the offset into the history list corresponding to
+  // the current navigation.
+  IPC_STRUCT_MEMBER(int, pending_history_list_offset)

[(Do not use) nasko, 2014/08/28 16:39:09]

Do we really need to send history parameters down to the renderer?

[clamy, 2014/09/02 18:25:19]

I don't know. Right now the history gets updated in the handler of
FrameMsg_Navigate. In the new logic, the history may not have been updated,
because the renderer never got the equivalent of FrameMsg_Navigate.

+
+  // Informs the RenderView of where its current page contents reside in
+  // session history and the total size of the session history list.
+  IPC_STRUCT_MEMBER(int, current_history_list_offset)
+  IPC_STRUCT_MEMBER(int, current_history_list_length)
+
+  // Informs the RenderView the session history should be cleared. In that
+  // case, the RenderView needs to notify the browser that the clearing was
+  // succesful when the navigation commits.
+  IPC_STRUCT_MEMBER(bool, should_clear_history_list)
+
+  // TODO(clamy): Investigate whether we should support data urls here.
+
+  // The URL to send in the "Referer" header field. Can be empty if there is
+  // no referrer.
+  IPC_STRUCT_MEMBER(content::Referrer, referrer)
+
+  // TODO(clamy): Investigate whether we should send the chain of redirects
+  // along.
+
+  // The type of transition.
+  IPC_STRUCT_MEMBER(content::PageTransition, transition)
+
+  // Opaque history state (received by ViewHostMsg_UpdateState).
+  IPC_STRUCT_MEMBER(content::PageState, page_state)
+
+  // Type of navigation.
+  IPC_STRUCT_MEMBER(FrameMsg_Navigate_Type::Value, navigation_type)
+
+  // Whether or not the user agent override string should be used.
+  IPC_STRUCT_MEMBER(bool, is_overriding_user_agent)
+
+  // The navigationStart time to expose through the Navigation Timing API to JS.
+  IPC_STRUCT_MEMBER(base::TimeTicks, browser_navigation_start)
+
+  // TODO(clamy): Add other timing values needed by the Navigation Timing API
+  // (redirects, etc...).
+IPC_STRUCT_END()
+
 #if defined(OS_MACOSX) || defined(OS_ANDROID)
 // This message is used for supporting popup menus on Mac OS X and Android using
 // native controls. See the FrameHostMsg_ShowPopup message.
@@ -478,6 +534,10 @@ IPC_MESSAGE_ROUTED1(FrameMsg_SelectPopupMenuItem,
 
 #endif
 
+// Tells the renderer to commit the specific navigation, as part of the
+// browser-side navigation refactoring.
+IPC_MESSAGE_ROUTED1(FrameMsg_CommitNavigation, FrameMsg_CommitNavigation_Params)

[(Do not use) nasko, 2014/08/28 16:39:09]

If we are adding an IPC, let's add a handler in the renderer process that either
does a NOTREACHED or just logs something. Leaving unhandled IPCs is not a good
practice.

[clamy, 2014/09/02 18:25:19]

Done.

+
 // -----------------------------------------------------------------------------
 // Messages sent from the renderer to the browser.