void XMLTreeBuilder::processStartTag(const AtomicXMLToken& token)
{
exitText();
bool isFirstElement = !m_sawFirstElement;
m_sawFirstElement = true;
NodeStackItem top = m_currentNodeStack.last();
processNamespaces(token, top);
QualifiedName qName(token.prefix(), token.name(), top.namespaceForPrefix(token.prefix(), top.namespaceURI()));
RefPtr<Element> newElement = m_document->createElement(qName, true);
processAttributes(token, top, newElement);
newElement->beginParsingChildren();
m_currentNodeStack.last().node()->parserAddChild(newElement.get());
top.setNode(newElement);
pushCurrentNode(top);
if (!newElement->attached())
newElement->attach();
if (isFirstElement && m_document->frame())
m_document->frame()->loader()->dispatchDocumentElementAvailable();
if (token.selfClosing())
closeElement(newElement);
}
void XMLTreeBuilder::processXMLEntity(const AtomicXMLToken& token)
{
DEFINE_STATIC_LOCAL(AtomicString, amp, ("amp"));
DEFINE_STATIC_LOCAL(AtomicString, apos, ("apos"));
DEFINE_STATIC_LOCAL(AtomicString, gt, ("gt"));
DEFINE_STATIC_LOCAL(AtomicString, lt, ("lt"));
DEFINE_STATIC_LOCAL(AtomicString, quot, ("quot"));
DEFINE_STATIC_LOCAL(String, ampS, ("&"));
DEFINE_STATIC_LOCAL(String, aposS, ("'"));
DEFINE_STATIC_LOCAL(String, gtS, (">"));
DEFINE_STATIC_LOCAL(String, ltS, ("<"));
DEFINE_STATIC_LOCAL(String, quotS, ("\""));
if (token.name() == amp)
appendToText(ampS.characters(), 1);
else if (token.name() == apos)
appendToText(aposS.characters(), 1);
else if (token.name() == gt)
appendToText(gtS.characters(), 1);
else if (token.name() == lt)
appendToText(ltS.characters(), 1);
else if (token.name() == quot)
appendToText(quotS.characters(), 1);
else
m_parser->stopParsing();
}
void XMLTreeBuilder::processDOCTYPE(const AtomicXMLToken& token)
{
DEFINE_STATIC_LOCAL(AtomicString, xhtmlTransitional, ("-//W3C//DTD XHTML 1.0 Transitional//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtml11, ("-//W3C//DTD XHTML 1.1//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtmlStrict, ("-//W3C//DTD XHTML 1.0 Strict//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtmlFrameset, ("-//W3C//DTD XHTML 1.0 Frameset//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtmlBasic, ("-//W3C//DTD XHTML Basic 1.0//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtmlMathML, ("-//W3C//DTD XHTML 1.1 plus MathML 2.0//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtmlMathMLSVG, ("-//W3C//DTD XHTML 1.1 plus MathML 2.0 plus SVG 1.1//EN"));
DEFINE_STATIC_LOCAL(AtomicString, xhtmlMobile, ("-//WAPFORUM//DTD XHTML Mobile 1.0//EN"));
if (!failOnText())
return;
AtomicString publicIdentifier(token.publicIdentifier().data(), token.publicIdentifier().size());
AtomicString systemIdentifier(token.systemIdentifier().data(), token.systemIdentifier().size());
RefPtr<DocumentType> doctype = DocumentType::create(m_document, token.name(), publicIdentifier, systemIdentifier);
m_document->setDocType(doctype);
m_document->parserAddChild(doctype);
if ((publicIdentifier == xhtmlTransitional)
|| (publicIdentifier == xhtml11)
|| (publicIdentifier == xhtmlStrict)
|| (publicIdentifier == xhtmlFrameset)
|| (publicIdentifier == xhtmlBasic)
|| (publicIdentifier == xhtmlMathML)
|| (publicIdentifier == xhtmlMathMLSVG)
|| (publicIdentifier == xhtmlMobile))
m_isXHTML = true;
}
void XMLTreeBuilder::processProcessingInstruction(const AtomicXMLToken& token)
{
if (!failOnText())
return;
// FIXME: fall back if we can't handle the PI ourself.
add(ProcessingInstruction::create(m_document, token.target(), token.data()));
}
void XMLTreeBuilder::processNamespaces(const AtomicXMLToken& token, NodeStackItem& stackItem)
{
for (unsigned i = 0; i < token.attributes().size(); ++i) {
const Attribute& tokenAttribute = token.attributes().at(i);
if (tokenAttribute.name().prefix() == xmlnsAtom)
stackItem.setNamespaceURI(tokenAttribute.name().localName(), tokenAttribute.value());
else if (tokenAttribute.name() == xmlnsAtom)
stackItem.setNamespaceURI(tokenAttribute.value());
}
}
void XMLTreeBuilder::processEndTag(const AtomicXMLToken& token)
{
exitText();
RefPtr<ContainerNode> node = m_currentNodeStack.last().node();
if (!node->hasTagName(QualifiedName(token.prefix(), token.name(), m_currentNodeStack.last().namespaceForPrefix(token.prefix(), m_currentNodeStack.last().namespaceURI()))))
m_parser->stopParsing();
closeElement(toElement(node.get()));
}
void XMLTreeBuilder::processNamespaces(const AtomicXMLToken& token, NodeStackItem& stackItem)
{
if (!token.attributes())
return;
for (size_t i = 0; i < token.attributes()->size(); ++i) {
Attribute* attribute = token.attributes()->attributeItem(i);
if (attribute->name().prefix() == xmlnsAtom)
stackItem.setNamespaceURI(attribute->name().localName(), attribute->value());
else if (attribute->name() == xmlnsAtom)
stackItem.setNamespaceURI(attribute->value());
}
}
void XMLTreeBuilder::processHTMLEntity(const AtomicXMLToken& token)
{
HTMLEntitySearch search;
const AtomicString& name = token.name();
for (size_t i = 0; i < name.length(); ++i) {
search.advance(name[i]);
if (!search.isEntityPrefix()) {
m_parser->stopParsing();
return;
}
}
search.advance(';');
if (!search.isEntityPrefix()) {
m_parser->stopParsing();
return;
}
UChar32 entityValue = search.mostRecentMatch()->firstValue;
// FIXME: We need to account for secondValue if any XML entities are longer
// than one unicode character.
ASSERT_NOT_REACHED();
// Darin Adler writes:
// You can see given the code above that this else is dead code. This code is in a strange state.
// And the reinterpret_cast to UChar* makes the code little-endian-specific. That is not good!
if (entityValue <= 0xFFFF)
appendToText(reinterpret_cast<UChar*>(&entityValue), 1);
else {
UChar utf16Pair[2] = { U16_LEAD(entityValue), U16_TRAIL(entityValue) };
appendToText(utf16Pair, 2);
}
}
void XMLTreeBuilder::processXMLDeclaration(const AtomicXMLToken& token)
{
if (!failOnText())
return;
ExceptionCode ec = 0;
m_document->setXMLVersion(String(token.xmlVersion()), ec);
if (ec)
m_parser->stopParsing();
m_document->setXMLStandalone(token.xmlStandalone(), ec);
if (ec)
m_parser->stopParsing();
// FIXME: how should this behave if standalone is not specified?
// FIXME: set encoding.
}
void XMLTreeBuilder::processAttributes(const AtomicXMLToken& token, NodeStackItem& stackItem, PassRefPtr<Element> newElement)
{
for (unsigned i = 0; i < token.attributes().size(); ++i) {
const Attribute& tokenAttribute = token.attributes().at(i);
ExceptionCode ec = 0;
if (tokenAttribute.name().prefix() == xmlnsAtom)
newElement->setAttributeNS(XMLNSNames::xmlnsNamespaceURI, "xmlns:" + tokenAttribute.name().localName(), tokenAttribute.value(), ec);
else if (tokenAttribute.name() == xmlnsAtom)
newElement->setAttributeNS(XMLNSNames::xmlnsNamespaceURI, xmlnsAtom, tokenAttribute.value(), ec);
else {
QualifiedName qName(tokenAttribute.prefix(), tokenAttribute.localName(), stackItem.namespaceForPrefix(tokenAttribute.prefix(), nullAtom));
newElement->setAttribute(qName, tokenAttribute.value());
}
if (ec) {
m_parser->stopParsing();
return;
}
}
}
void XMLTreeBuilder::processAttributes(const AtomicXMLToken& token, NodeStackItem& stackItem, PassRefPtr<Element> newElement)
{
if (!token.attributes())
return;
for (size_t i = 0; i < token.attributes()->size(); ++i) {
Attribute* attribute = token.attributes()->attributeItem(i);
ExceptionCode ec = 0;
if (attribute->name().prefix() == xmlnsAtom)
newElement->setAttributeNS(XMLNSNames::xmlnsNamespaceURI, "xmlns:" + attribute->name().localName(), attribute->value(), ec);
else if (attribute->name() == xmlnsAtom)
newElement->setAttributeNS(XMLNSNames::xmlnsNamespaceURI, xmlnsAtom, attribute->value(), ec);
else {
QualifiedName qName(attribute->prefix(), attribute->localName(), stackItem.namespaceForPrefix(attribute->prefix(), nullAtom));
newElement->setAttribute(qName, attribute->value());
}
if (ec) {
m_parser->stopParsing();
return;
}
}
}
void XMLTreeBuilder::processHTMLEntity(const AtomicXMLToken& token)
{
HTMLEntitySearch search;
const AtomicString& name = token.name();
for (size_t i = 0; i < name.length(); ++i) {
search.advance(name[i]);
if (!search.isEntityPrefix()) {
m_parser->stopParsing();
return;
}
}
search.advance(';');
UChar32 entityValue = search.currentValue();
if (entityValue <= 0xFFFF)
appendToText(reinterpret_cast<UChar*>(&entityValue), 1);
else {
UChar utf16Pair[2] = { U16_LEAD(entityValue), U16_TRAIL(entityValue) };
appendToText(utf16Pair, 2);
}
}
void XMLTreeBuilder::processToken(const AtomicXMLToken& token)
{
switch (token.type()) {
case XMLTokenTypes::Uninitialized:
ASSERT_NOT_REACHED();
break;
case XMLTokenTypes::ProcessingInstruction:
processProcessingInstruction(token);
break;
case XMLTokenTypes::XMLDeclaration:
processXMLDeclaration(token);
break;
case XMLTokenTypes::DOCTYPE:
processDOCTYPE(token);
break;
case XMLTokenTypes::StartTag:
processStartTag(token);
break;
case XMLTokenTypes::EndTag:
processEndTag(token);
break;
case XMLTokenTypes::CDATA:
processCDATA(token);
break;
case XMLTokenTypes::Character:
processCharacter(token);
break;
case XMLTokenTypes::Comment:
processComment(token);
break;
case XMLTokenTypes::Entity:
processEntity(token);
break;
case XMLTokenTypes::EndOfFile:
exitText();
return;
}
}
void XMLTreeBuilder::processComment(const AtomicXMLToken& token)
{
exitText();
add(Comment::create(m_document, token.comment()));
}
void XMLTreeBuilder::processCDATA(const AtomicXMLToken& token)
{
exitText();
add(CDATASection::create(m_document, token.data()));
}
void XMLTreeBuilder::processCharacter(const AtomicXMLToken& token)
{
appendToText(token.characters().data(), token.characters().size());
}