Get the document language directly from WebKit

chrome/renderer/translate_helper.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/translate_helper.h"
 
 #include "base/bind.h"
 #include "base/compiler_specific.h"
 #include "base/logging.h"
 #include "base/message_loop.h"
 #include "base/metrics/histogram.h"
+#include "base/string16.h"
 #include "base/utf_string_conversions.h"
 #include "chrome/common/chrome_constants.h"
 #include "chrome/common/render_messages.h"
 #include "content/public/renderer/render_view.h"
 #include "third_party/WebKit/Source/WebKit/chromium/public/WebDocument.h"
 #include "third_party/WebKit/Source/WebKit/chromium/public/WebElement.h"
 #include "third_party/WebKit/Source/WebKit/chromium/public/WebFrame.h"
 #include "third_party/WebKit/Source/WebKit/chromium/public/WebScriptSource.h"
 #include "third_party/WebKit/Source/WebKit/chromium/public/WebView.h"
 #include "third_party/cld/encodings/compact_lang_det/win/cld_unicodetext.h"
@@ skipping 31 matching lines @@
       page_id_(-1),
       ALLOW_THIS_IN_INITIALIZER_LIST(weak_method_factory_(this)) {
 }
 
 TranslateHelper::~TranslateHelper() {
   CancelPendingTranslation();
 }
 
 void TranslateHelper::PageCaptured(const string16& contents) {
   WebDocument document = render_view()->GetWebView()->mainFrame()->document();
-  // If the page explicitly specifies a language, use it, otherwise we'll
-  // determine it based on the text content using the CLD.
-  std::string language = GetPageLanguageFromMetaTag(&document);
+
+  // Get the document language as set by WebKit from the http-equiv
+  // meta tag for "content-language".  This may or may not also
+  // have a value derived from the actual Content-Language HTTP
+  // header.  The two actually have different meanings (despite the
+  // original intent of http-equiv to be an equivalent) with the former
+  // being the language of the document and the latter being the
+  // language of the intended audience (a distinction really only
+  // relevant for things like langauge textbooks).  This distinction
+  // shouldn't affect translation.
+  std::string language = document.contentLanguage().utf8();
+  size_t coma_index = language.find(',');
+  if (coma_index != std::string::npos) {
+    // There are more than 1 language specified, just keep the first one.
+    language = language.substr(0, coma_index);
+  }
+  TrimWhitespaceASCII(language, TRIM_ALL, &language);
+  language = StringToLowerASCII(language);
+
   if (language.empty()) {
     base::TimeTicks begin_time = base::TimeTicks::Now();
     language = DetermineTextLanguage(contents);
     UMA_HISTOGRAM_MEDIUM_TIMES("Renderer4.LanguageDetection",
                                base::TimeTicks::Now() - begin_time);
   } else {
     VLOG(9) << "PageLanguageFromMetaTag: " << language;
   }
 
   Send(new ChromeViewHostMsg_TranslateLanguageDetermined(
@@ skipping 22 matching lines @@
       attribute = iter->getAttribute("content");
     if (attribute.isNull())
       continue;
     if (LowerCaseEqualsASCII(attribute, "notranslate"))
       return false;
   }
   return true;
 }
 
 // static
-std::string TranslateHelper::GetPageLanguageFromMetaTag(WebDocument* document) {
-  // The META language tag looks like:
-  // <meta http-equiv="content-language" content="en">
-  // It can contain more than one language:
-  // <meta http-equiv="content-language" content="en, fr">
-  std::vector<WebElement> meta_elements;
-  webkit_glue::GetMetaElementsWithAttribute(document,
-                                            ASCIIToUTF16("http-equiv"),
-                                            ASCIIToUTF16("content-language"),
-                                            &meta_elements);
-  if (meta_elements.empty())
-    return std::string();
-
-  // We don't expect more than one such tag. If there are several, just use the
-  // first one.
-  WebString attribute = meta_elements[0].getAttribute("content");
-  if (attribute.isEmpty())
-    return std::string();
-
-  // The value is supposed to be ASCII.
-  if (!IsStringASCII(attribute))
-    return std::string();
-
-  std::string language = StringToLowerASCII(UTF16ToASCII(attribute));
-  size_t coma_index = language.find(',');
-  if (coma_index != std::string::npos) {
-    // There are more than 1 language specified, just keep the first one.
-    language = language.substr(0, coma_index);
-  }
-  TrimWhitespaceASCII(language, TRIM_ALL, &language);
-  return language;
-}
-
-// static
 std::string TranslateHelper::DetermineTextLanguage(const string16& text) {
   std::string language = chrome::kUnknownLanguageCode;
   int num_languages = 0;
   int text_bytes = 0;
   bool is_reliable = false;
   Language cld_language =
       DetectLanguageOfUnicodeText(NULL, text.c_str(), true, &is_reliable,
                                   &num_languages, NULL, &text_bytes);
   // We don't trust the result if the CLD reports that the detection is not
   // reliable, or if the actual text used to detect the language was less than
@@ skipping 291 matching lines @@
   WebView* web_view = render_view()->GetWebView();
   if (!web_view) {
     // When the WebView is going away, the render view should have called
     // CancelPendingTranslation() which should have stopped any pending work, so
     // that case should not happen.
     NOTREACHED();
     return NULL;
   }
   return web_view->mainFrame();
 }