Fix some unit_tests under PlzNavigate

content/public/test/test_renderer_host.h

 // Copyright (c) 2012 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 CONTENT_PUBLIC_TEST_TEST_RENDERER_HOST_H_
 #define CONTENT_PUBLIC_TEST_TEST_RENDERER_HOST_H_
 
 #include <stdint.h>
 
 #include <memory>
@@ skipping 78 matching lines @@
 
   // Simulates a navigation to |url| failing with the error code |error_code|.
   virtual void SimulateNavigationError(const GURL& url, int error_code) = 0;
 
   // Simulates the commit of an error page following a navigation failure.
   virtual void SimulateNavigationErrorPageCommit() = 0;
 
   // Simulates a navigation stopping in the RenderFrameHost.
   virtual void SimulateNavigationStop() = 0;
 
+  // If we are doing a cross-site navigation, this simulates the current RFH

[clamy, 2016/08/30 23:08:38]

nit: s/RFH/RenderFrameHost here and below.

[scottmg, 2016/08/30 23:55:30]

Done.

+  // notifying that it has unloaded so the pending RFH is resumed and can
+  // navigate.
+  // PlzNavigate: the pending RFH is not created before the navigation commit,

[clamy, 2016/08/30 23:08:38]

This comment from "the pending RFH" to "deleted in the meantime" actually
belongs to where it was before. It describes why in this specific function we
need to call PrepareForCommit before trying to access the main or pending RFH.

For the comment, I would suggest getting the one from
TestRenderFrameHost::PrepareForCommit, and keeping most of the last paragraph.

[scottmg, 2016/08/30 23:55:30]

Done.

+  // so it is necessary to simulate the IO thread response here to commit in the
+  // proper renderer. It is necessary to call PrepareForCommit before getting
+  // the main and the pending frame because when we are trying to navigate to a
+  // webui from a new tab, a RenderFrameHost is created to display it that is
+  // committed immediately (since it is a new tab). Therefore the main frame is
+  // replaced without a pending frame being created, and we don't get the right
+  // values for the RFH to navigate: we try to use the old one that has been
+  // deleted in the meantime.
+  // Note that for some synchronous navigations (about:blank, javascript
+  // urls, etc.) there will be no NavigationRequest, and no simulation of the

[clamy, 2016/08/30 23:08:38]

Please remove the mention of NavigationRequest, since this is not a class we
expose outside of content/

[scottmg, 2016/08/30 23:55:30]

Done.

+  // network stack is required.
+  virtual void SimulateUnloading() = 0;
+
   // Calls OnDidCommitProvisionalLoad on the RenderFrameHost with the given
   // information with various sets of parameters. These are helper functions for
   // simulating the most common types of loads.
   //
   // Guidance for calling these:
   // - nav_entry_id should be 0 if simulating a renderer-initiated navigation;
   //   if simulating a browser-initiated one, pass the GetUniqueID() value of
   //   the NavigationController's PendingEntry.
   // - did_create_new_entry should be true if simulating a navigation that
   //   created a new navigation entry; false for history navigations, reloads,
@@ skipping 173 matching lines @@
   std::unique_ptr<aura::test::AuraTestHelper> aura_test_helper_;
 #endif
   RenderViewHostTestEnabler rvh_test_enabler_;
 
   DISALLOW_COPY_AND_ASSIGN(RenderViewHostTestHarness);
 };
 
 }  // namespace content
 
 #endif  // CONTENT_PUBLIC_TEST_TEST_RENDERER_HOST_H_