Move RenderViewHost management out of WebContents into a new

RenderViewHostManager object.

The goal for this patch is to change no logic or APIs, just move the code. So
there are some not very clean callback functions and no new unit tests for this
file (although is is still covered by the same WebContents unit tests). This
should make the actual cleanup in a later pass much easier to follow.

I changed the ordering of only a few operations (like WebContents shutdown), and checked that this shouldn't matter. I had to change the "source" for several notifications since they are no longer sent from the WebContents.


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@716 0039d316-1c4b-4281-b951-d872f2087c98

1 files changed, 3 insertions, 2 deletions

diff --git a/chrome/browser/tabs/tab_strip_model.cc b/chrome/browser/tabs/tab_strip_model.cc
index c71c089d..2828d51 100644
--- a/chrome/browser/tabs/tab_strip_model.cc
+++ b/chrome/browser/tabs/tab_strip_model.cc
@@ -271,14 +271,15 @@ bool TabStripModel::TabHasUnloadListener(int index) {
   //             in testing and then provide better test coverage for features
   //             like "close other tabs".
   WebContents* web_contents = GetContentsAt(index)->AsWebContents();
-  if (web_contents)
+  if (web_contents) {
     // If the WebContents is not connected yet, then there's no unload
     // handler we can fire even if the WebContents has an unload listener.
     // One case where we hit this is in a tab that has an infinite loop 
     // before load.
     return web_contents->notify_disconnection() && 
-        !web_contents->IsShowingInterstitialPage() &&
+        !web_contents->showing_interstitial_page() &&
         web_contents->render_view_host()->HasUnloadListener();
+  }
   return false;
 }