Get the document language directly from WebKit rather than trying to figure

it out ourselves.  Since WebKit is already case-insensitive, this fixes
test-case #3.  Test-Case #1 requires WebKit support for extracting the
language from the HTTP headers (https://bugs.webkit.org/show_bug.cgi?id=97929).

Note that this patch will fail completely until a WebKit patch to export
this information is accepted (https://bugs.webkit.org/show_bug.cgi?id=98066).

BUG=145689


Review URL: https://chromiumcodereview.appspot.com/11052002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@162212 0039d316-1c4b-4281-b951-d872f2087c98

chrome/renderer/translate_helper.cc

@@ -9,6 +9,7 @@
 #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"
@@ -60,9 +61,25 @@ TranslateHelper::~TranslateHelper() {
 
 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);
@@ -104,40 +121,6 @@ bool TranslateHelper::IsPageTranslatable(WebDocument* document) {
   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;

chrome/renderer/translate_helper.h

@@ -70,13 +70,6 @@ class TranslateHelper : public content::RenderViewObserver {
   // should not be translated.
   static bool IsPageTranslatable(WebKit::WebDocument* document);
 
-  // Returns the language specified in the language meta tag of |document|, or
-  // an empty string if no such tag was found.
-  // The tag may specify several languages, the first one is returned.
-  // Example of such meta-tag:
-  // <meta http-equiv="content-language" content="en, fr">
-  static std::string GetPageLanguageFromMetaTag(WebKit::WebDocument* document);
-
   // Returns the ISO 639_1 language code of the specified |text|, or 'unknown'
   // if it failed.
   static std::string DetermineTextLanguage(const string16& text);

chrome/renderer/translate_helper_browsertest.cc

@@ -367,6 +367,35 @@ TEST_F(ChromeRenderViewTest, LanguageMetaTag) {
   EXPECT_EQ("fr", params.a);
 }
 
+// Tests that the language meta tag works even with non-all-lower-case.
+// http://code.google.com/p/chromium/issues/detail?id=145689
+TEST_F(ChromeRenderViewTest, LanguageMetaTagCase) {
+  // Suppress the normal delay that occurs when the page is loaded before which
+  // the renderer sends the page contents to the browser.
+  SendContentStateImmediately();
+
+  LoadHTML("<html><head><meta http-equiv=\"Content-Language\" content=\"es\">"
+           "</head><body>A random page with random content.</body></html>");
+  const IPC::Message* message = render_thread_->sink().GetUniqueMessageMatching(
+      ChromeViewHostMsg_TranslateLanguageDetermined::ID);
+  ASSERT_NE(static_cast<IPC::Message*>(NULL), message);
+  ChromeViewHostMsg_TranslateLanguageDetermined::Param params;
+  ChromeViewHostMsg_TranslateLanguageDetermined::Read(message, &params);
+  EXPECT_EQ("es", params.a);
+  render_thread_->sink().ClearMessages();
+
+  // Makes sure we support multiple languages specified.
+  LoadHTML("<html><head><meta http-equiv=\"Content-Language\" "
+           "content=\" fr , es,en \">"
+           "</head><body>A random page with random content.</body></html>");
+  message = render_thread_->sink().GetUniqueMessageMatching(
+      ChromeViewHostMsg_TranslateLanguageDetermined::ID);
+  ASSERT_NE(static_cast<IPC::Message*>(NULL), message);
+  ChromeViewHostMsg_TranslateLanguageDetermined::Read(message, &params);
+  EXPECT_EQ("fr", params.a);
+}
+
+
 // Tests that a back navigation gets a translate language message.
 TEST_F(ChromeRenderViewTest, BackToTranslatablePage) {
   SendContentStateImmediately();