Remove page id from "page capture".

chrome/renderer/safe_browsing/phishing_classifier_delegate.cc

 // Copyright (c) 2011 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.
 
 #include "chrome/renderer/safe_browsing/phishing_classifier_delegate.h"
 
 #include <set>
 
 #include "base/bind.h"
 #include "base/callback.h"
@@ skipping 126 matching lines @@
 
 void PhishingClassifierDelegate::OnStartPhishingDetection(const GURL& url) {
   last_url_received_from_browser_ = StripRef(url);
   // Start classifying the current page if all conditions are met.
   // See MaybeStartClassification() for details.
   MaybeStartClassification();
 }
 
 void PhishingClassifierDelegate::DidCommitProvisionalLoad(
     blink::WebLocalFrame* frame, bool is_new_navigation) {
-  // A new page is starting to load, so cancel classificaiton.
-  //
-  // TODO(bryner): We shouldn't need to cancel classification if the navigation
-  // is within the same page.  However, if we let classification continue in
-  // this case, we need to properly deal with the fact that PageCaptured will
-  // be called again for the in-page navigation.  We need to be sure not to
-  // swap out the page text while the term feature extractor is still running.
-  DocumentState* document_state = DocumentState::FromDataSource(
-      frame->dataSource());
-  NavigationState* navigation_state = document_state->navigation_state();
-  CancelPendingClassification(navigation_state->was_within_same_page() ?
-                              NAVIGATE_WITHIN_PAGE : NAVIGATE_AWAY);
   if (frame == render_view()->GetWebView()->mainFrame()) {
+    DocumentState* document_state = DocumentState::FromDataSource(
+        frame->dataSource());
+    NavigationState* navigation_state = document_state->navigation_state();
     last_main_frame_transition_ = navigation_state->transition_type();
   }
 }
 
 void PhishingClassifierDelegate::PageCaptured(base::string16* page_text,
                                               bool preliminary_capture) {
-  if (preliminary_capture) {
+  if (preliminary_capture || *page_text != classifier_page_text_)

[mattm, 2014/07/08 23:35:50]

should that be "==" ?

[Avi (use Gerrit), 2014/07/08 23:43:18]

D'oh. Yes.

     return;
-  }
+
   // Make sure there's no classification in progress.  We don't want to swap
   // out the page text string from underneath the term feature extractor.
   //
   // Note: Currently, if the url hasn't changed, we won't restart
   // classification in this case.  We may want to adjust this.
   CancelPendingClassification(PAGE_RECAPTURED);
   last_finished_load_url_ = GetToplevelUrl();
   classifier_page_text_.swap(*page_text);
   have_page_text_ = true;
   MaybeStartClassification();
@@ skipping 104 matching lines @@
   VLOG(2) << "Starting classification for " << last_finished_load_url_;
   last_url_sent_to_classifier_ = last_finished_load_url_;
   is_classifying_ = true;
   classifier_->BeginClassification(
       &classifier_page_text_,
       base::Bind(&PhishingClassifierDelegate::ClassificationDone,
                  base::Unretained(this)));
 }
 
 }  // namespace safe_browsing