Clean up third_party/cld.

1. Remove files we don't use that come from toolbar
2. Simplify the directory structure by removing the redundant
intermediary directories bar/toolbar/cld/i18n

This will bring CLD a step closer to where it can be open-sourced separately (or along with CED for encoding detection.). It's not there yet.


In addition to the clean-up, change |LanguageCode*| return 
'nb', 'he', 'fil' instead of 'no', 'iw', and 'tl'. 

Also, use LanguageCodeWithDialects instead of LanguageCode to get 'zh-CN' 
instead of 'zh'. This is to simplify the 3-way mapping between 
Chrome's UI locale code, CLD's language code and what's accepted by 
Google Translate. It's also another preparation for open-sourcing because
we'd better use the standard code rather than the obsolete code like 'iw'
when open-sourcing.

BUG=32759, 33613
TEST=third_party/cld builds on all platforms and pass the following tests:
- unit_tests: Extension*.DetectTabLanguage and CompactLangDetTest.* 
- browser_tests: ExtensionBrowserTest.Toolstrip


git-svn-id: svn://svn.chromium.org/chrome/trunk/src@38146 0039d316-1c4b-4281-b951-d872f2087c98

3 files changed, 6 insertions, 11 deletions

diff --git a/chrome/renderer/DEPS b/chrome/renderer/DEPS
index 35f9fb6..55936c2 100644
--- a/chrome/renderer/DEPS
+++ b/chrome/renderer/DEPS
@@ -12,7 +12,7 @@ include_rules = [
   "+webkit/glue",
   "+webkit/glue/plugins",
   "+v8/include",
-  "+third_party/cld/bar/toolbar/cld/i18n/encodings/compact_lang_det/win",
+  "+third_party/cld/encodings/compact_lang_det/win",
   "+third_party/npapi/bindings",
   "+third_party/sqlite/preprocessed",
   "+third_party/tcmalloc",
diff --git a/chrome/renderer/render_view.cc b/chrome/renderer/render_view.cc
index 872c518..cf74f81 100644
--- a/chrome/renderer/render_view.cc
+++ b/chrome/renderer/render_view.cc
@@ -67,7 +67,7 @@
 #include "net/base/net_errors.h"
 #include "skia/ext/bitmap_platform_device.h"
 #include "skia/ext/image_operations.h"
-#include "third_party/cld/bar/toolbar/cld/i18n/encodings/compact_lang_det/win/cld_unicodetext.h"
+#include "third_party/cld/encodings/compact_lang_det/win/cld_unicodetext.h"
 #include "third_party/WebKit/WebKit/chromium/public/WebAccessibilityCache.h"
 #include "third_party/WebKit/WebKit/chromium/public/WebAccessibilityObject.h"
 #include "third_party/WebKit/WebKit/chromium/public/WebCString.h"
diff --git a/chrome/renderer/render_view.h b/chrome/renderer/render_view.h
index c09bf32..7245df1 100644
--- a/chrome/renderer/render_view.h
+++ b/chrome/renderer/render_view.h
@@ -458,15 +458,10 @@ class RenderView : public RenderWidget,
 
   PageTranslator* page_translator() const { return page_translator_.get(); }
 
-  // Returns the ISO 639 language code of the current page (e.g. en, fr, zh).
-  // If ISO 639-1 code is not available for the language, ISO 639-2 3-letter code
-  // will be returned (e.g. kha for Khasi and und for undtermined). For traditional
-  // Chinse, 'zh-TW' will be returned while for simplified Chinse, 'zh' will be
-  // returned.
-  // TODO(jungshik): Make it return 'he' (the correct ISO 639 code for Hebrew)
-  // instead of the obsolete 'iw'. Perhaps, it's also better to return 'zh-Hans'
-  // (or 'zh-CN') for Simplified Chinese instead of 'zh' to be aligned with
-  // 'zh-TW' for Traditional Chinse.
+  // Returns the ISO 639 language code of the current page (e.g. en, fr).
+  // If ISO 639-1 code is not available for the language, ISO 639-2 3-letter
+  // code will be returned (e.g. fil for Filipino and und for undtermined). For
+  // traditional and simplified Chinse, 'zh-TW'  and 'zh-CN' will be returned.
   std::string DetectLanguage();
 
 #if defined(OS_MACOSX)