UChar decodeNamedEntity(const char* name)
{
HTMLEntitySearch search;
while (*name) {
search.advance(*name++);
if (!search.isEntityPrefix())
return 0;
}
search.advance(';');
UChar32 entityValue = search.currentValue();
if (U16_LENGTH(entityValue) != 1) {
// Callers need to move off this API if the entity table has values
// which do no fit in a 16 bit UChar!
ASSERT_NOT_REACHED();
return 0;
}
return static_cast<UChar>(entityValue);
}
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::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);
}
}
size_t decodeNamedEntityToUCharArray(const char* name, UChar result[4])
{
HTMLEntitySearch search;
while (*name) {
search.advance(*name++);
if (!search.isEntityPrefix())
return 0;
}
search.advance(';');
if (!search.isEntityPrefix())
return 0;
size_t numberOfCodePoints = appendUChar32ToUCharArray(search.mostRecentMatch()->firstValue, result);
if (!search.mostRecentMatch()->secondValue)
return numberOfCodePoints;
return numberOfCodePoints + appendUChar32ToUCharArray(search.mostRecentMatch()->secondValue, result + numberOfCodePoints);
}
static bool consumeNamedEntity(SegmentedString& source, StringBuilder& decodedEntity, bool& notEnoughCharacters, UChar additionalAllowedCharacter, UChar& cc)
{
StringBuilder consumedCharacters;
HTMLEntitySearch entitySearch;
while (!source.isEmpty()) {
cc = source.currentChar();
entitySearch.advance(cc);
if (!entitySearch.isEntityPrefix())
break;
consumedCharacters.append(cc);
source.advance();
}
notEnoughCharacters = source.isEmpty();
if (notEnoughCharacters) {
// We can't an entity because there might be a longer entity
// that we could match if we had more data.
unconsumeCharacters(source, consumedCharacters);
return false;
}
if (!entitySearch.mostRecentMatch()) {
unconsumeCharacters(source, consumedCharacters);
return false;
}
if (entitySearch.mostRecentMatch()->length != entitySearch.currentLength()) {
// We've consumed too many characters. We need to walk the
// source back to the point at which we had consumed an
// actual entity.
unconsumeCharacters(source, consumedCharacters);
consumedCharacters.clear();
const int length = entitySearch.mostRecentMatch()->length;
const LChar* reference = entitySearch.mostRecentMatch()->entity;
for (int i = 0; i < length; ++i) {
cc = source.currentChar();
ASSERT_UNUSED(reference, cc == *reference++);
consumedCharacters.append(cc);
source.advance();
ASSERT(!source.isEmpty());
}
cc = source.currentChar();
}
if (entitySearch.mostRecentMatch()->lastCharacter() == ';'
|| !additionalAllowedCharacter
|| !(isASCIIAlphanumeric(cc) || cc == '=')) {
decodedEntity.append(entitySearch.mostRecentMatch()->firstValue);
if (entitySearch.mostRecentMatch()->secondValue)
decodedEntity.append(entitySearch.mostRecentMatch()->secondValue);
return true;
}
unconsumeCharacters(source, consumedCharacters);
return false;
}
bool consumeHTMLEntity(SegmentedString& source, Vector<UChar, 16>& decodedEntity, bool& notEnoughCharacters, UChar additionalAllowedCharacter)
{
ASSERT(!additionalAllowedCharacter || additionalAllowedCharacter == '"' || additionalAllowedCharacter == '\'' || additionalAllowedCharacter == '>');
ASSERT(!notEnoughCharacters);
ASSERT(decodedEntity.isEmpty());
enum EntityState {
Initial,
Number,
MaybeHexLowerCaseX,
MaybeHexUpperCaseX,
Hex,
Decimal,
Named
};
EntityState entityState = Initial;
UChar32 result = 0;
Vector<UChar, 10> consumedCharacters;
while (!source.isEmpty()) {
UChar cc = *source;
switch (entityState) {
case Initial: {
if (cc == '\x09' || cc == '\x0A' || cc == '\x0C' || cc == ' ' || cc == '<' || cc == '&')
return false;
if (additionalAllowedCharacter && cc == additionalAllowedCharacter)
return false;
if (cc == '#') {
entityState = Number;
break;
}
if ((cc >= 'a' && cc <= 'z') || (cc >= 'A' && cc <= 'Z')) {
entityState = Named;
continue;
}
return false;
}
case Number: {
if (cc == 'x') {
entityState = MaybeHexLowerCaseX;
break;
}
if (cc == 'X') {
entityState = MaybeHexUpperCaseX;
break;
}
if (cc >= '0' && cc <= '9') {
entityState = Decimal;
continue;
}
source.push('#');
return false;
}
case MaybeHexLowerCaseX: {
if (isHexDigit(cc)) {
entityState = Hex;
continue;
}
source.push('#');
source.push('x');
return false;
}
case MaybeHexUpperCaseX: {
if (isHexDigit(cc)) {
entityState = Hex;
continue;
}
source.push('#');
source.push('X');
return false;
}
case Hex: {
if (cc >= '0' && cc <= '9')
result = result * 16 + cc - '0';
else if (cc >= 'a' && cc <= 'f')
result = result * 16 + 10 + cc - 'a';
else if (cc >= 'A' && cc <= 'F')
result = result * 16 + 10 + cc - 'A';
else {
if (cc == ';')
source.advanceAndASSERT(cc);
return convertToUTF16(legalEntityFor(result), decodedEntity);
}
break;
}
case Decimal: {
if (cc >= '0' && cc <= '9')
result = result * 10 + cc - '0';
else {
if (cc == ';')
source.advanceAndASSERT(cc);
return convertToUTF16(legalEntityFor(result), decodedEntity);
}
break;
}
case Named: {
HTMLEntitySearch entitySearch;
while (!source.isEmpty()) {
cc = *source;
entitySearch.advance(cc);
if (!entitySearch.isEntityPrefix())
break;
consumedCharacters.append(cc);
source.advanceAndASSERT(cc);
}
notEnoughCharacters = source.isEmpty();
if (notEnoughCharacters) {
// We can't an entity because there might be a longer entity
// that we could match if we had more data.
unconsumeCharacters(source, consumedCharacters);
return false;
}
if (!entitySearch.mostRecentMatch()) {
ASSERT(!entitySearch.currentValue());
unconsumeCharacters(source, consumedCharacters);
return false;
}
if (entitySearch.mostRecentMatch()->length != entitySearch.currentLength()) {
// We've consumed too many characters. We need to walk the
// source back to the point at which we had consumed an
// actual entity.
unconsumeCharacters(source, consumedCharacters);
consumedCharacters.clear();
const int length = entitySearch.mostRecentMatch()->length;
const UChar* reference = entitySearch.mostRecentMatch()->entity;
for (int i = 0; i < length; ++i) {
cc = *source;
ASSERT_UNUSED(reference, cc == *reference++);
consumedCharacters.append(cc);
source.advanceAndASSERT(cc);
ASSERT(!source.isEmpty());
}
cc = *source;
}
if (entitySearch.mostRecentMatch()->lastCharacter() == ';'
|| !additionalAllowedCharacter
|| !(isAlphaNumeric(cc) || cc == '=')) {
return convertToUTF16(entitySearch.mostRecentMatch()->value, decodedEntity);
}
unconsumeCharacters(source, consumedCharacters);
return false;
}
}
consumedCharacters.append(cc);
source.advanceAndASSERT(cc);
}
ASSERT(source.isEmpty());
notEnoughCharacters = true;
unconsumeCharacters(source, consumedCharacters);
return false;
}
