Convert MediaQueryTokenizer's codepoints array to be static data

Source/core/css/parser/MediaQueryTokenizer.cpp

 // Copyright 2014 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 "config.h"
 #include "core/css/parser/MediaQueryTokenizer.h"
 
+namespace WebCore {
+#include "MediaQueryTokenizerCodepoints.cpp"
+}
+
 #include "core/css/parser/MediaQueryInputStream.h"
 #include "core/html/parser/HTMLParserIdioms.h"
 #include "wtf/unicode/CharacterNames.h"
 
 namespace WebCore {
 
-const unsigned codePointsNumber = SCHAR_MAX + 1;
-
-class MediaQueryTokenizer::CodePoints {
-public:
-    MediaQueryTokenizer::CodePoint codePoints[codePointsNumber];
-
-    // FIXME: Move the codePoint array to be a static one, generated by build scripts
-    CodePoints()
-    {
-        memset(codePoints, 0, codePointsNumber);
-        codePoints['\n'] = &MediaQueryTokenizer::whiteSpace;
-        codePoints['\r'] = &MediaQueryTokenizer::whiteSpace;
-        codePoints['\t'] = &MediaQueryTokenizer::whiteSpace;
-        codePoints[' '] = &MediaQueryTokenizer::whiteSpace;
-        codePoints['\f'] = &MediaQueryTokenizer::whiteSpace;
-        codePoints['('] = &MediaQueryTokenizer::leftParenthesis;
-        codePoints[')'] = &MediaQueryTokenizer::rightParenthesis;
-        codePoints['+'] = &MediaQueryTokenizer::plusOrFullStop;
-        codePoints['.'] = &MediaQueryTokenizer::plusOrFullStop;
-        codePoints[','] = &MediaQueryTokenizer::comma;
-        codePoints['-'] = &MediaQueryTokenizer::hyphenMinus;
-        codePoints['/'] = &MediaQueryTokenizer::solidus;
-        codePoints[':'] = &MediaQueryTokenizer::colon;
-        codePoints[';'] = &MediaQueryTokenizer::semiColon;
-        codePoints['\\'] = &MediaQueryTokenizer::reverseSolidus;
-        for (unsigned char digit = '0'; digit <= '9'; ++digit)
-            codePoints[digit] = &MediaQueryTokenizer::asciiDigit;
-        for (unsigned char alpha = 'a'; alpha <= 'z'; ++alpha)
-            codePoints[alpha] = &MediaQueryTokenizer::nameStart;
-        for (unsigned char alpha = 'A'; alpha <= 'Z'; ++alpha)
-            codePoints[alpha] = &MediaQueryTokenizer::nameStart;
-        codePoints['_'] = &MediaQueryTokenizer::nameStart;
-        codePoints[kEndOfFileMarker] = &MediaQueryTokenizer::endOfFile;
-    }
-};
-
-MediaQueryTokenizer::CodePoints* MediaQueryTokenizer::codePoints()
-{
-    static CodePoints codePoints;
-    return &codePoints;
-}
-
 // http://dev.w3.org/csswg/css-syntax/#name-start-code-point
 static bool isNameStart(UChar c)
 {
     if (isASCIIAlpha(c))
         return true;
     if (c == '_')
         return true;
     return !isASCII(c);
 }
 
@@ skipping 143 matching lines @@
     // We could move to the stateful model and instead create
     // states for all the "next 3 codepoints are X" cases.
     // State-machine tokenizers are easier to write to handle
     // incremental tokenization of partial sources.
     // However, for now we follow the spec exactly.
     UChar cc = consume();
     CodePoint codePointFunc = 0;
 
     if (isASCII(cc)) {
         ASSERT_WITH_SECURITY_IMPLICATION(cc < codePointsNumber);
-        codePointFunc = codePoints()->codePoints[cc];
+        codePointFunc = codePoints[cc];
     } else {
         codePointFunc = &MediaQueryTokenizer::nameStart;
     }
 
     if (codePointFunc)
         return ((this)->*(codePointFunc))(cc);
 
     return MediaQueryToken(DelimiterToken, cc);
 }
 
@@ skipping 192 matching lines @@
     if (firstChar == '-') {
         if (isNameStart(m_input.peek(1)))
             return true;
         return nextTwoCharsAreValidEscape();
     }
 
     return false;
 }
 
 } // namespace WebCore