Revert "Moved text decoding to the parser thread"

Source/core/html/parser/HTMLMetaCharsetParser.cpp

 /*
  * Copyright (C) 2010 Google Inc. All Rights Reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
  * 2. Redistributions in binary form must reproduce the above copyright
  *    notice, this list of conditions and the following disclaimer in the
@@ skipping 33 matching lines @@
     , m_assumedCodec(newTextCodec(Latin1Encoding()))
     , m_inHeadSection(true)
     , m_doneChecking(false)
 {
 }
 
 HTMLMetaCharsetParser::~HTMLMetaCharsetParser()
 {
 }
 
+static const char charsetString[] = "charset";
+static const size_t charsetLength = sizeof("charset") - 1;
+
+String HTMLMetaCharsetParser::extractCharset(const String& value)
+{
+    size_t pos = 0;
+    unsigned length = value.length();
+
+    while (pos < length) {
+        pos = value.find(charsetString, pos, false);
+        if (pos == kNotFound)
+            break;
+
+        pos += charsetLength;
+
+        // Skip whitespace.
+        while (pos < length && value[pos] <= ' ')
+            ++pos;
+
+        if (value[pos] != '=')
+            continue;
+
+        ++pos;
+
+        while (pos < length && value[pos] <= ' ')
+            ++pos;
+
+        char quoteMark = 0;
+        if (pos < length && (value[pos] == '"' || value[pos] == '\'')) {
+            quoteMark = static_cast<char>(value[pos++]);
+            ASSERT(!(quoteMark & 0x80));
+        }
+
+        if (pos == length)
+            break;
+
+        unsigned end = pos;
+        while (end < length && ((quoteMark && value[end] != quoteMark) || (!quoteMark && value[end] > ' ' && value[end] != '"' && value[end] != '\'' && value[end] != ';')))
+            ++end;
+
+        if (quoteMark && (end == length))
+            break; // Close quote not found.
+
+        return value.substring(pos, end - pos);
+    }
+
+    return "";
+}
+
 bool HTMLMetaCharsetParser::processMeta()
 {
     const HTMLToken::AttributeList& tokenAttributes = m_token.attributes();
-    HTMLAttributeList attributes;
+    AttributeList attributes;
     for (HTMLToken::AttributeList::const_iterator iter = tokenAttributes.begin(); iter != tokenAttributes.end(); ++iter) {
-        String attributeName = attemptStaticStringCreation(iter->name, Likely8Bit);
+        String attributeName = StringImpl::create8BitIfPossible(iter->name);
         String attributeValue = StringImpl::create8BitIfPossible(iter->value);
         attributes.append(std::make_pair(attributeName, attributeValue));
     }
 
     m_encoding = encodingFromMetaAttributes(attributes);
     return m_encoding.isValid();
 }
 
+WTF::TextEncoding HTMLMetaCharsetParser::encodingFromMetaAttributes(const AttributeList& attributes)
+{
+    bool gotPragma = false;
+    Mode mode = None;
+    String charset;
+
+    for (AttributeList::const_iterator iter = attributes.begin(); iter != attributes.end(); ++iter) {
+        const AtomicString& attributeName = AtomicString(iter->first);
+        const String& attributeValue = iter->second;
+
+        if (attributeName == http_equivAttr) {
+            if (equalIgnoringCase(attributeValue, "content-type"))
+                gotPragma = true;
+        } else if (charset.isEmpty()) {
+            if (attributeName == charsetAttr) {
+                charset = attributeValue;
+                mode = Charset;
+            } else if (attributeName == contentAttr) {
+                charset = extractCharset(attributeValue);
+                if (charset.length())
+                    mode = Pragma;
+            }
+        }
+    }
+
+    if (mode == Charset || (mode == Pragma && gotPragma))
+        return WTF::TextEncoding(stripLeadingAndTrailingHTMLSpaces(charset));
+
+    return WTF::TextEncoding();
+}
+
 static const int bytesToCheckUnconditionally = 1024; // That many input bytes will be checked for meta charset even if <head> section is over.
 
 bool HTMLMetaCharsetParser::checkForMetaCharset(const char* data, size_t length)
 {
     if (m_doneChecking)
         return true;
 
     ASSERT(!m_encoding.isValid());
 
     // We still don't have an encoding, and are in the head.
@@ skipping 12 matching lines @@
 
     // Since many sites have charset declarations after <body> or other tags
     // that are disallowed in <head>, we don't bail out until we've checked at
     // least bytesToCheckUnconditionally bytes of input.
 
     m_input.append(SegmentedString(m_assumedCodec->decode(data, length)));
 
     while (m_tokenizer->nextToken(m_input, m_token)) {
         bool end = m_token.type() == HTMLToken::EndTag;
         if (end || m_token.type() == HTMLToken::StartTag) {
-            String tagName = attemptStaticStringCreation(m_token.name(), Likely8Bit);
+            AtomicString tagName(m_token.name());
             if (!end) {
                 m_tokenizer->updateStateFor(tagName);
-                if (threadSafeMatch(tagName, metaTag) && processMeta()) {
+                if (tagName == metaTag && processMeta()) {
                     m_doneChecking = true;
                     return true;
                 }
             }
 
-            if (!threadSafeMatch(tagName, scriptTag) && !threadSafeMatch(tagName, noscriptTag)
-                && !threadSafeMatch(tagName, styleTag) && !threadSafeMatch(tagName, linkTag)
-                && !threadSafeMatch(tagName, metaTag) && !threadSafeMatch(tagName, objectTag)
-                && !threadSafeMatch(tagName, titleTag) && !threadSafeMatch(tagName, baseTag)
-                && (end || !threadSafeMatch(tagName, htmlTag)) && (end || !threadSafeMatch(tagName, headTag))) {
+            if (tagName != scriptTag && tagName != noscriptTag
+                && tagName != styleTag && tagName != linkTag
+                && tagName != metaTag && tagName != objectTag
+                && tagName != titleTag && tagName != baseTag
+                && (end || tagName != htmlTag) && (end || tagName != headTag)) {
                 m_inHeadSection = false;
             }
         }
 
         if (!m_inHeadSection && m_input.numberOfCharactersConsumed() >= bytesToCheckUnconditionally) {
             m_doneChecking = true;
             return true;
         }
 
         m_token.clear();
     }
 
     return false;
 }
 
 }