Use JSON schema compiler for i18n api

chrome/browser/extensions/extension_i18n_api.cc

 // Copyright (c) 2010 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/browser/extensions/extension_i18n_api.h"
 
+#include <string>
+#include <vector>
+
 #include "base/string_piece.h"
 #include "base/utf_string_conversions.h"
 #include "chrome/browser/prefs/pref_service.h"
 #include "chrome/browser/profiles/profile.h"
 #include "chrome/common/pref_names.h"
+#include "chrome/common/extensions/api/i18n.h"
+
+namespace GetAcceptLanguages = extensions::api::i18n::GetAcceptLanguages;
 
 // Errors.
 static const char kEmptyAcceptLanguagesError[] = "accept-languages is empty.";
 
 bool GetAcceptLanguagesFunction::RunImpl() {
   string16 acceptLanguages =

[not at google - send to devlin, 2012/06/29 04:58:00]

Style should be accept_languages not acceptLanguages.

       UTF8ToUTF16(profile()->GetPrefs()->GetString(prefs::kAcceptLanguages));
   // Currently, there are 2 ways to set browser's accept-languages: through UI
   // or directly modify the preference file. The accept-languages set through
   // UI is guranteed to be valid, and the accept-languages string returned from
   // profile()->GetPrefs()->GetString(prefs::kAcceptLanguages) is guranteed to
   // be valid and well-formed, which means each accept-langauge is a valid
   // code, and accept-languages are seperatd by "," without surrrounding
   // spaces. But we do not do any validation (either the format or the validity
   // of the language code) on accept-languages set through editing preference
   // file directly. So, here, we're adding extra checks to be resistant to
   // crashes caused by data corruption.
-  result_.reset(new ListValue());
   if (acceptLanguages.empty()) {
     error_ = kEmptyAcceptLanguagesError;
     return false;
   }
   size_t begin = 0;
   size_t end;
+  std::vector<std::string> languages_vec;

[not at google - send to devlin, 2012/06/29 04:58:00]

just "languages"?

[mitchellwrosen, 2012/07/27 23:51:41]

Done.

   while (1) {
     end = acceptLanguages.find(',', begin);
     if (end > begin) {
       // Guard against a malformed value with multiple "," in a row.
-      string16 acceptLang = acceptLanguages.substr(begin, end - begin);
-      static_cast<ListValue*>(result_.get())->
-          Append(Value::CreateStringValue(acceptLang));
+      languages_vec.push_back(
+          UTF16ToUTF8(acceptLanguages.substr(begin, end - begin)));
     }
     begin = end + 1;
     // 'begin >= acceptLanguages.length()' to guard against a value
     // ending with ','.
     if (end == string16::npos || begin >= acceptLanguages.length())
       break;
   }

[not at google - send to devlin, 2012/06/29 04:58:00]

This code should be using base::SplitString (base/string_split.h) on
accept_langauges with ',', filtering out empty strings (e.g. using
std::remove_if or some other stl function I'm not aware of), then returning an
error if it's empty.

However I don't think the goal of this patch is to cleanup messy code :) so if
you don't feel like changing it, don't worry about it.

-  if (static_cast<ListValue*>(result_.get())->GetSize() == 0) {
+
+  if (languages_vec.empty()) {
     error_ = kEmptyAcceptLanguagesError;
     return false;
   }
+
+  result_.reset(GetAcceptLanguages::Result::Create(languages_vec));
   return true;
 }