Revert 35735 - Relanding the language detection code....

chrome/renderer/render_view.cc

 // Copyright (c) 2009 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/render_view.h"
 
 #include <algorithm>
 #include <string>
 #include <vector>
 
@@ skipping 243 matching lines @@
       history_forward_list_count_(0),
       has_unload_listener_(false),
       decrement_shared_popup_at_destruction_(false),
       autofill_query_id_(0),
       popup_notification_visible_(false),
       spelling_panel_visible_(false),
       send_content_state_immediately_(false),
       send_preferred_size_changes_(false),
       ALLOW_THIS_IN_INITIALIZER_LIST(
           notification_provider_(new NotificationProvider(this))),
+      determine_page_text_after_loading_stops_(false),
       view_type_(ViewType::INVALID),
       browser_window_id_(-1),
       last_top_level_navigation_page_id_(-1),
 #if defined(OS_MACOSX)
       has_document_tag_(false),
 #endif
       document_tag_(0),
       webkit_preferences_(webkit_preferences) {
 }
 
@@ skipping 164 matching lines @@
     IPC_MESSAGE_HANDLER(ViewMsg_Replace, OnReplace)
     IPC_MESSAGE_HANDLER(ViewMsg_ToggleSpellPanel, OnToggleSpellPanel)
     IPC_MESSAGE_HANDLER(ViewMsg_AdvanceToNextMisspelling,
                         OnAdvanceToNextMisspelling)
     IPC_MESSAGE_HANDLER(ViewMsg_ToggleSpellCheck, OnToggleSpellCheck)
     IPC_MESSAGE_HANDLER(ViewMsg_Delete, OnDelete)
     IPC_MESSAGE_HANDLER(ViewMsg_SelectAll, OnSelectAll)
     IPC_MESSAGE_HANDLER(ViewMsg_CopyImageAt, OnCopyImageAt)
     IPC_MESSAGE_HANDLER(ViewMsg_ExecuteEditCommand, OnExecuteEditCommand)
     IPC_MESSAGE_HANDLER(ViewMsg_Find, OnFind)
+    IPC_MESSAGE_HANDLER(ViewMsg_DeterminePageText, OnDeterminePageText)
     IPC_MESSAGE_HANDLER(ViewMsg_Zoom, OnZoom)
     IPC_MESSAGE_HANDLER(ViewMsg_SetZoomLevelForLoadingHost,
                         OnSetZoomLevelForLoadingHost)
     IPC_MESSAGE_HANDLER(ViewMsg_SetPageEncoding, OnSetPageEncoding)
     IPC_MESSAGE_HANDLER(ViewMsg_ResetPageEncodingToDefault,
                         OnResetPageEncodingToDefault)
     IPC_MESSAGE_HANDLER(ViewMsg_SetupDevToolsClient, OnSetupDevToolsClient)
     IPC_MESSAGE_HANDLER(ViewMsg_DownloadFavIcon, OnDownloadFavIcon)
     IPC_MESSAGE_HANDLER(ViewMsg_ScriptEvalRequest, OnScriptEvalRequest)
     IPC_MESSAGE_HANDLER(ViewMsg_CSSInsertRequest, OnCSSInsertRequest)
@@ skipping 131 matching lines @@
 
   // Don't index/capture pages that failed to load.  This only checks the top
   // level frame so the thumbnail may contain a frame that failed to load.
   WebDataSource* ds = main_frame->dataSource();
   if (ds && ds->hasUnreachableURL())
     return;
 
   if (!preliminary_capture)
     last_indexed_page_id_ = load_id;
 
-  // Get the URL for this page.
+  // get the URL for this page
   GURL url(main_frame->url());
   if (url.is_empty())
     return;
 
-  // Retrieve the frame's full text.
+  // full text
   std::wstring contents;
   CaptureText(main_frame, &contents);
   if (contents.size()) {
-    // Send the text to the browser for indexing (the browser might decide not
-    // to index, if the URL is HTTPS for instance) and language discovery.
-    Send(new ViewHostMsg_PageContents(routing_id_, url, load_id, contents));
+    // Send the text to the browser for indexing.
+    Send(new ViewHostMsg_PageContents(url, load_id, contents));
+  }
+
+  // Send over text content of this page to the browser.
+  if (determine_page_text_after_loading_stops_) {
+    determine_page_text_after_loading_stops_ = false;
+    Send(new ViewMsg_DeterminePageText_Reply(routing_id_, contents));
   }
 
   // thumbnail
   SendThumbnail();
 }
 
 void RenderView::CaptureText(WebFrame* frame, std::wstring* contents) {
   contents->clear();
   if (!frame)
     return;
 
+  // Don't index any https pages. People generally don't want their bank
+  // accounts, etc. indexed on their computer, especially since some of these
+  // things are not marked cachable.
+  // TODO(brettw) we may want to consider more elaborate heuristics such as
+  // the cachability of the page. We may also want to consider subframes (this
+  // test will still index subframes if the subframe is SSL).
+  if (GURL(frame->url()).SchemeIsSecure())
+    return;
+
 #ifdef TIME_TEXT_RETRIEVAL
   double begin = time_util::GetHighResolutionTimeNow();
 #endif
 
   // get the contents of the frame
   *contents = UTF16ToWideHack(frame->contentAsText(kMaxIndexChars));
 
 #ifdef TIME_TEXT_RETRIEVAL
   double end = time_util::GetHighResolutionTimeNow();
   char buf[128];
@@ skipping 2327 matching lines @@
                                          true);  // reset the tickmarks
       }
 
       // Iterate to the next frame. The frame will not necessarily scope, for
       // example if it is not visible.
       search_frame = search_frame->traverseNext(true);
     } while (search_frame != main_frame);
   }
 }
 
+void RenderView::OnDeterminePageText() {
+  if (!is_loading_) {
+    if (!webview())
+      return;
+    WebFrame* main_frame = webview()->mainFrame();
+    std::wstring contents;
+    CaptureText(main_frame, &contents);
+    Send(new ViewMsg_DeterminePageText_Reply(routing_id_, contents));
+    determine_page_text_after_loading_stops_ = false;
+    return;
+  }
+
+  // We set |determine_page_text_after_loading_stops_| true here so that,
+  // after page has been loaded completely, the text in the page is captured.
+  determine_page_text_after_loading_stops_ = true;
+}
+
 void RenderView::DnsPrefetch(const std::vector<std::string>& host_names) {
   Send(new ViewHostMsg_DnsPrefetch(host_names));
 }
 
 void RenderView::OnZoom(PageZoom::Function function) {
   if (!webview())  // Not sure if this can happen, but no harm in being safe.
     return;
 
   int zoom_level = webview()->zoomLevel();
   int new_zoom_level = webview()->setZoomLevel(false,
@@ skipping 917 matching lines @@
       new PluginMsg_SignalModalDialogEvent(host_window_));
 
   message->EnableMessagePumping();  // Runs a nested message loop.
   bool rv = Send(message);
 
   PluginChannelHost::Broadcast(
       new PluginMsg_ResetModalDialogEvent(host_window_));
 
   return rv;
 }