Remove page id from "page capture".

chrome/renderer/chrome_render_view_observer.cc

 // 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.
 
 #include "chrome/renderer/chrome_render_view_observer.h"
 
 #include "base/bind.h"
 #include "base/bind_helpers.h"
 #include "base/command_line.h"
 #include "base/debug/trace_event.h"
@@ skipping 78 matching lines @@
 
 // maximum number of characters in the document to index, any text beyond this
 // point will be clipped
 static const size_t kMaxIndexChars = 65535;
 
 // Constants for UMA statistic collection.
 static const char kTranslateCaptureText[] = "Translate.CaptureText";
 
 namespace {
 
-GURL StripRef(const GURL& url) {
-  GURL::Replacements replacements;
-  replacements.ClearRef();
-  return url.ReplaceComponents(replacements);
-}
-
 #if defined(OS_ANDROID)
 // Parses the DOM for a <meta> tag with a particular name.
 // |meta_tag_content| is set to the contents of the 'content' attribute.
 // |found_tag| is set to true if the tag was successfully found.
 // Returns true if the document was parsed without errors.
 bool RetrieveMetaTagContent(const WebFrame* main_frame,
                             const GURL& expected_url,
                             const std::string& meta_tag_name,
                             bool* found_tag,
                             std::string* meta_tag_content) {
@@ skipping 39 matching lines @@
 
 }  // namespace
 
 ChromeRenderViewObserver::ChromeRenderViewObserver(
     content::RenderView* render_view,
     ChromeRenderProcessObserver* chrome_render_process_observer)
     : content::RenderViewObserver(render_view),
       chrome_render_process_observer_(chrome_render_process_observer),
       translate_helper_(new TranslateHelper(render_view)),
       phishing_classifier_(NULL),
-      last_indexed_page_id_(-1),
       capture_timer_(false, false) {
   const CommandLine& command_line = *CommandLine::ForCurrentProcess();
   if (!command_line.HasSwitch(switches::kDisableClientSidePhishingDetection))
     OnSetClientSidePhishingDetection(true);
 }
 
 ChromeRenderViewObserver::~ChromeRenderViewObserver() {
 }
 
 bool ChromeRenderViewObserver::OnMessageReceived(const IPC::Message& message) {
@@ skipping 171 matching lines @@
     Send(new ChromeViewHostMsg_PageHasOSDD(
         routing_id(), main_frame->document().url(), osdd_url,
         search_provider::AUTODETECTED_PROVIDER));
   }
 
   // Don't capture pages including refresh meta tag.
   if (HasRefreshMetaTag(main_frame))
     return;
 
   CapturePageInfoLater(
-      render_view()->GetPageId(),
       false,  // preliminary_capture
       base::TimeDelta::FromMilliseconds(
           render_view()->GetContentStateImmediately() ?
               0 : kDelayForCaptureMs));
 }
 
 void ChromeRenderViewObserver::DidCommitProvisionalLoad(
     WebLocalFrame* frame, bool is_new_navigation) {
   // Don't capture pages being not new, or including refresh meta tag.
-  if (!is_new_navigation || HasRefreshMetaTag(frame))
+  if (!is_new_navigation || HasRefreshMetaTag(frame)) {
+    capture_timer_.Stop();
     return;
+  }
 
   CapturePageInfoLater(
-      render_view()->GetPageId(),
       true,  // preliminary_capture
       base::TimeDelta::FromMilliseconds(kDelayForForcedCaptureMs));
 }
 
-void ChromeRenderViewObserver::CapturePageInfoLater(int page_id,
-                                                    bool preliminary_capture,
+void ChromeRenderViewObserver::CapturePageInfoLater(bool preliminary_capture,
                                                     base::TimeDelta delay) {
   capture_timer_.Start(
       FROM_HERE,
       delay,
       base::Bind(&ChromeRenderViewObserver::CapturePageInfo,
                  base::Unretained(this),
-                 page_id,
                  preliminary_capture));
 }
 
-void ChromeRenderViewObserver::CapturePageInfo(int page_id,
-                                               bool preliminary_capture) {
-  // If |page_id| is obsolete, we should stop indexing and capturing a page.
-  if (render_view()->GetPageId() != page_id)
-    return;
-
+void ChromeRenderViewObserver::CapturePageInfo(bool preliminary_capture) {
   if (!render_view()->GetWebView())
     return;
 
   WebFrame* main_frame = render_view()->GetWebView()->mainFrame();
   if (!main_frame)
     return;
 
   // Don't index/capture pages that are in view source mode.
   if (main_frame->isViewSourceModeEnabled())
     return;
@@ skipping 13 matching lines @@
   // Retrieve the frame's full text (up to kMaxIndexChars), and pass it to the
   // translate helper for language detection and possible translation.
   base::string16 contents;
   base::TimeTicks capture_begin_time = base::TimeTicks::Now();
   CaptureText(main_frame, &contents);
   UMA_HISTOGRAM_TIMES(kTranslateCaptureText,
                       base::TimeTicks::Now() - capture_begin_time);
   if (translate_helper_)
     translate_helper_->PageCaptured(contents);
 
-  // TODO(shess): Is indexing "Full text search" indexing?  In that
-  // case more of this can go.
-  // Skip indexing if this is not a new load.  Note that the case where
-  // page_id == last_indexed_page_id_ is more complicated, since we need to
-  // reindex if the toplevel URL has changed (such as from a redirect), even
-  // though this may not cause the page id to be incremented.
-  if (page_id < last_indexed_page_id_)
-    return;
-
-  bool same_page_id = last_indexed_page_id_ == page_id;
-  if (!preliminary_capture)
-    last_indexed_page_id_ = page_id;
-
-  // Get the URL for this page.
-  GURL url(main_frame->document().url());
-  if (url.is_empty()) {
-    if (!preliminary_capture)
-      last_indexed_url_ = GURL();
-    return;
-  }
-
-  // If the page id is unchanged, check whether the URL (ignoring fragments)
-  // has changed.  If so, we need to reindex.  Otherwise, assume this is a
-  // reload, in-page navigation, or some other load type where we don't want to
-  // reindex.  Note: subframe navigations after onload increment the page id,
-  // so these will trigger a reindex.
-  GURL stripped_url(StripRef(url));
-  if (same_page_id && stripped_url == last_indexed_url_)
-    return;
-
-  if (!preliminary_capture)
-    last_indexed_url_ = stripped_url;
-
   TRACE_EVENT0("renderer", "ChromeRenderViewObserver::CapturePageInfo");
 
 #if defined(FULL_SAFE_BROWSING)
   // Will swap out the string.
   if (phishing_classifier_)
     phishing_classifier_->PageCaptured(&contents, preliminary_capture);
 #endif
 }
 
 void ChromeRenderViewObserver::CaptureText(WebFrame* frame,
@@ skipping 45 matching lines @@
     WebElement element = node.to<WebElement>();
     if (!element.hasHTMLTagName(tag_name))
       continue;
     WebString value = element.getAttribute(attribute_name);
     if (value.isNull() || !LowerCaseEqualsASCII(value, "refresh"))
       continue;
     return true;
   }
   return false;
 }