Minor changes to simplify IsValidLocaleSyntax logic.

app/l10n_util.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 "app/l10n_util.h"
 
 #if defined(TOOLKIT_USES_GTK)
 #include <glib/gutils.h>
 #endif
 
@@ skipping 489 matching lines @@
 }
 
 bool IsValidLocaleSyntax(const std::string& locale) {
   // Check that the length is plausible.
   if (locale.size() < 2 || locale.size() >= ULOC_FULLNAME_CAPACITY)
     return false;
 
   // Strip off the part after an '@' sign, which might contain keywords,
   // as in en_IE@currency=IEP or fr@collation=phonebook;calendar=islamic-civil.
   // We don't validate that part much, just check that there's at least one
-  // equals sign in a plausible place.
-  std::string prefix = locale;
-  if (locale.find("@") != std::string::npos) {
-    size_t split_point = locale.find("@");
+  // equals sign in a plausible place. Normalize the prefix so that hyphens
+  // are changed to underscores.
+  std::string prefix = NormalizeLocale(locale);
+  size_t split_point = locale.find("@");
+  if (split_point != std::string::npos) {
     std::string keywords = locale.substr(split_point + 1);
     prefix = locale.substr(0, split_point);
 
     size_t equals_loc = keywords.find("=");
     if (equals_loc == std::string::npos ||
         equals_loc < 1 || equals_loc > keywords.size() - 2)
       return false;
   }
 
-  // Check that all characters before the at-sign are alphanumeric, hyphen,
-  // or underscore.
+  // Check that all characters before the at-sign are alphanumeric or
+  // underscore.
   for (size_t i = 0; i < prefix.size(); i++) {
     char ch = prefix[i];
-    if (!IsAsciiAlpha(ch) && !IsAsciiDigit(ch) && ch != '-' && ch != '_')
+    if (!IsAsciiAlpha(ch) && !IsAsciiDigit(ch) && ch != '_')
       return false;
   }
 
   // Check that the initial token (before the first hyphen/underscore)
   // is 1 - 3 alphabetical characters (a language tag).
   for (size_t i = 0; i < prefix.size(); i++) {
     char ch = prefix[i];
-    if (ch == '-' || ch == '_') {
+    if (ch == '_') {
       if (i < 1 || i > 3)
         return false;
       break;
     }
     if (!IsAsciiAlpha(ch))
       return false;
   }
 
   // Check that the all tokens after the initial token are 1 - 8 characters.
   // (Tokenize/StringTokenizer don't work here, they collapse multiple
   // delimiters into one.)
   int token_len = 0;
   int token_index = 0;
   for (size_t i = 0; i < prefix.size(); i++) {
     char ch = prefix[i];
-    if (ch == '-' || ch == '_') {
+    if (ch == '_') {

[Nebojša Ćirić, 2010/12/13 19:43:48]

You don't use ch anywhere else. Remove it and use prefix[i] directly.

You can also remove nested if (you can skip this):
for () {
  if (prefix[i] != '_') {
    token_len++;
    continue;
  }
  if (token_index == 0 && (token_len < 1)...) {
    return false;
  }

  token_index++;
  token_len = 0;
}

       if (token_index > 0 && (token_len < 1 || token_len > 8)) {
         return false;
       }
       token_index++;
       token_len = 0;
     } else {
       token_len++;
     }
   }
   if (token_index == 0 && (token_len < 1 || token_len > 3)) {
@@ skipping 397 matching lines @@
   for (size_t i = 0; i < arraysize(kAcceptLanguageList); ++i) {
     if (!IsLocaleNameTranslated(kAcceptLanguageList[i], display_locale))
       // TODO(jungshik) : Put them at the of the list with language codes
       // enclosed by brackets instead of skipping.
         continue;
     locale_codes->push_back(kAcceptLanguageList[i]);
   }
 }
 
 }  // namespace l10n_util