Do not create proxy hosts in the subtree of navigating frame.

content/browser/accessibility/site_per_process_accessibility_browsertest.cc

Index: content/browser/accessibility/site_per_process_accessibility_browsertest.cc
diff --git a/content/browser/accessibility/site_per_process_accessibility_browsertest.cc b/content/browser/accessibility/site_per_process_accessibility_browsertest.cc
index fdc0dc6a8a7417d7475c2c878a4d38aa5c79d9a4..12cdc89376c779a873b2c3968a6bbfb16e75e59f 100644
--- a/content/browser/accessibility/site_per_process_accessibility_browsertest.cc
+++ b/content/browser/accessibility/site_per_process_accessibility_browsertest.cc
@@ -80,7 +80,7 @@ IN_PROC_BROWSER_TEST_F(SitePerProcessAccessibilityBrowserTest,
   NavigateFrameToURL(root->child_at(0), cross_site_url);
 
   // Ensure that we have created a new process for the subframe.
-  ASSERT_EQ(1U, root->child_count());
+  ASSERT_EQ(2U, root->child_count());
   SiteInstance* site_instance = child->current_frame_host()->GetSiteInstance();
   EXPECT_NE(shell()->web_contents()->GetSiteInstance(), site_instance);
 
@@ -113,7 +113,7 @@ IN_PROC_BROWSER_TEST_F(SitePerProcessAccessibilityBrowserTest,
 
   BrowserAccessibility* ax_group = ax_root->PlatformGetChild(0);
   ASSERT_EQ(ui::AX_ROLE_GROUP, ax_group->GetRole());
-  ASSERT_EQ(1U, ax_group->PlatformChildCount());
+  ASSERT_EQ(2U, ax_group->PlatformChildCount());

[dmazzoni, 2014/09/11 15:09:31]

Could you either access ax_group->PlatformGetChild(1) below and assert what it
is, or else comment what it is and why we should ignore it?

Use the VLOGs above for a nice text dump of the accessibility tree so you
shouldn't have to guess the role.

[nasko, 2014/09/11 17:10:36]

As discussed, I'll update this test in a separate CL to check for the entire
frame tree.

 
   BrowserAccessibility* ax_iframe = ax_group->PlatformGetChild(0);
   ASSERT_EQ(ui::AX_ROLE_IFRAME, ax_iframe->GetRole());