// ---------------------------------------------------------------------------
// RangeToken: Getter methods
// ---------------------------------------------------------------------------
RangeToken* RangeToken::getCaseInsensitiveToken(TokenFactory* const tokFactory) {
if (fCaseIToken == 0 && tokFactory) {
bool isNRange = (getTokenType() == T_NRANGE) ? true : false;
RangeToken* lwrToken = tokFactory->createRange(isNRange);
for (unsigned int i = 0; i < fElemCount - 1; i += 2) {
for (XMLInt32 ch = fRanges[i]; ch <= fRanges[i + 1]; ++ch) {
#if defined(XML_USE_ICU_TRANSCODER) || defined (XML_USE_UNICONV390_TRANSCODER)
const XMLInt32 upperCh = u_toupper(ch);
if (upperCh != ch)
{
lwrToken->addRange(upperCh, upperCh);
}
const XMLInt32 lowerCh = u_tolower(ch);
if (lowerCh != ch)
{
lwrToken->addRange(lowerCh, lowerCh);
}
const XMLInt32 titleCh = u_totitle(ch);
if (titleCh != ch && titleCh != upperCh)
{
lwrToken->addRange(titleCh, titleCh);
}
#else
if (ch >= chLatin_A && ch <= chLatin_Z)
{
ch += chLatin_a - chLatin_A;
lwrToken->addRange(ch, ch);
}
else if (ch >= chLatin_a && ch <= chLatin_z)
{
ch -= chLatin_a - chLatin_A;
lwrToken->addRange(ch, ch);
}
#endif
}
}
lwrToken->mergeRanges(this);
lwrToken->compactRanges();
lwrToken->createMap();
fCaseIToken = lwrToken;
}
return fCaseIToken;
}
// ---------------------------------------------------------------------------
// RangeToken: Getter methods
// ---------------------------------------------------------------------------
RangeToken* RangeToken::getCaseInsensitiveToken(TokenFactory* const tokFactory) {
// REVIST
// We will not build a token with case insenstive ranges
// For now we will return a copy of ourselves.
if (fCaseIToken == 0 && tokFactory) {
bool isNRange = (getTokenType() == T_NRANGE) ? true : false;
RangeToken* lwrToken = tokFactory->createRange(isNRange);
lwrToken->mergeRanges(this);
fCaseIToken = lwrToken;
}
return fCaseIToken;
}
RangeToken* RegxParser::parseCharacterClass(const bool useNRange) {
setParseContext(regexParserStateInBrackets);
processNext();
RangeToken* tok = 0;
bool isNRange = false;
if (getState() == REGX_T_CHAR && getCharData() == chCaret) {
isNRange = true;
processNext();
}
tok = fTokenFactory->createRange();
parserState type;
bool firstLoop = true;
bool wasDecoded;
while ( (type = getState()) != REGX_T_EOF) {
wasDecoded = false;
// single range | from-to-range | subtraction
if (type == REGX_T_CHAR && getCharData() == chCloseSquare && !firstLoop)
break;
XMLInt32 ch = getCharData();
bool end = false;
if (type == REGX_T_BACKSOLIDUS) {
switch(ch) {
case chLatin_d:
case chLatin_D:
case chLatin_w:
case chLatin_W:
case chLatin_s:
case chLatin_S:
case chLatin_i:
case chLatin_I:
case chLatin_c:
case chLatin_C:
{
tok->mergeRanges(getTokenForShorthand(ch));
end = true;
}
break;
case chLatin_p:
case chLatin_P:
{
RangeToken* tok2 = processBacksolidus_pP(ch);
if (tok2 == 0) {
ThrowXMLwithMemMgr(ParseException,XMLExcepts::Parser_Atom5, getMemoryManager());
}
tok->mergeRanges(tok2);
end = true;
}
break;
case chDash:
wasDecoded = true;
// fall thru to default.
default:
ch = decodeEscaped();
}
} // end if REGX_T_BACKSOLIDUS
else if (type == REGX_T_XMLSCHEMA_CC_SUBTRACTION && !firstLoop) {
if (isNRange)
{
tok = RangeToken::complementRanges(tok, fTokenFactory, fMemoryManager);
isNRange=false;
}
RangeToken* rangeTok = parseCharacterClass(false);
tok->subtractRanges(rangeTok);
if (getState() != REGX_T_CHAR || getCharData() != chCloseSquare) {
ThrowXMLwithMemMgr(ParseException,XMLExcepts::Parser_CC5, getMemoryManager());
}
break;
} // end if REGX_T_XMLSCHEMA...
processNext();
if (!end) {
if (type == REGX_T_CHAR
&& (ch == chOpenSquare
|| ch == chCloseSquare
|| (ch == chDash && getCharData() == chCloseSquare && firstLoop))) {
// if regex = [-] then invalid...
// '[', ']', '-' not allowed and should be escaped
XMLCh chStr[] = { ch, chNull };
ThrowXMLwithMemMgr2(ParseException,XMLExcepts::Parser_CC6, chStr, chStr, getMemoryManager());
}
if (ch == chDash && getCharData() == chDash && getState() != REGX_T_BACKSOLIDUS && !wasDecoded) {
XMLCh chStr[] = { ch, chNull };
ThrowXMLwithMemMgr2(ParseException,XMLExcepts::Parser_CC6, chStr, chStr, getMemoryManager());
}
if (getState() != REGX_T_CHAR || getCharData() != chDash) {
tok->addRange(ch, ch);
}
else {
processNext();
if ((type = getState()) == REGX_T_EOF)
ThrowXMLwithMemMgr(ParseException,XMLExcepts::Parser_CC2, getMemoryManager());
if (type == REGX_T_CHAR && getCharData() == chCloseSquare) {
tok->addRange(ch, ch);
tok->addRange(chDash, chDash);
}
else if (type == REGX_T_XMLSCHEMA_CC_SUBTRACTION) {
static const XMLCh dashStr[] = { chDash, chNull};
ThrowXMLwithMemMgr2(ParseException, XMLExcepts::Parser_CC6, dashStr, dashStr, getMemoryManager());
}
else {
XMLInt32 rangeEnd = getCharData();
XMLCh rangeEndStr[] = { rangeEnd, chNull };
if (type == REGX_T_CHAR) {
if (rangeEnd == chOpenSquare
|| rangeEnd == chCloseSquare
|| rangeEnd == chDash)
// '[', ']', '-' not allowed and should be escaped
ThrowXMLwithMemMgr2(ParseException, XMLExcepts::Parser_CC6, rangeEndStr, rangeEndStr, getMemoryManager());
}
else if (type == REGX_T_BACKSOLIDUS) {
rangeEnd = decodeEscaped();
}
processNext();
if (ch > rangeEnd) {
XMLCh chStr[] = { ch, chNull };
ThrowXMLwithMemMgr2(ParseException,XMLExcepts::Parser_Ope3, rangeEndStr, chStr, getMemoryManager());
}
tok->addRange(ch, rangeEnd);
}
}
}
firstLoop = false;
}
if (getState() == REGX_T_EOF)
ThrowXMLwithMemMgr(ParseException,XMLExcepts::Parser_CC2, getMemoryManager());
if (isNRange)
{
if(useNRange)
tok->setTokenType(Token::T_NRANGE);
else
tok = RangeToken::complementRanges(tok, fTokenFactory, fMemoryManager);
}
tok->sortRanges();
tok->compactRanges();
// If the case-insensitive option is enabled, we need to
// have the new RangeToken instance build its internal
// case-insensitive RangeToken.
if (RegularExpression::isSet(fOptions, RegularExpression::IGNORE_CASE))
{
tok->getCaseInsensitiveToken(fTokenFactory);
}
setParseContext(regexParserStateNormal);
processNext();
return tok;
}
// ---------------------------------------------------------------------------
// UnicodeRangeFactory: Range creation methods
// ---------------------------------------------------------------------------
void UnicodeRangeFactory::buildRanges(RangeTokenMap *rangeTokMap) {
if (fRangesCreated)
return;
if (!fKeywordsInitialized) {
initializeKeywordMap(rangeTokMap);
}
TokenFactory* tokFactory = rangeTokMap->getTokenFactory();
RangeToken* ranges[UNICATEGSIZE];
RangeToken* tok;
for (int i=0; i < UNICATEGSIZE; i++) {
ranges[i] = tokFactory->createRange();
}
for (int j=0; j < 0x10000; j++) {
unsigned short charType = XMLUniCharacter::getType(j);
ranges[charType]->addRange(j, j);
charType = getUniCategory(charType);
ranges[charType]->addRange(j, j);
}
ranges[XMLUniCharacter::UNASSIGNED]->addRange(0x10000, Token::UTF16_MAX);
for (int k=0; k < UNICATEGSIZE; k++) {
tok = (RangeToken*) RangeToken::complementRanges(ranges[k], tokFactory);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(uniCategNames[k], ranges[k]);
rangeTokMap->setRangeToken(uniCategNames[k], tok , true);
}
// Create all range
tok = tokFactory->createRange();
tok->addRange(0, Token::UTF16_MAX);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniAll, tok);
// Create alpha range
tok = tokFactory->createRange();
tok->mergeRanges(ranges[XMLUniCharacter::UPPERCASE_LETTER]);
tok->mergeRanges(ranges[XMLUniCharacter::LOWERCASE_LETTER]);
tok->mergeRanges(ranges[XMLUniCharacter::OTHER_LETTER]);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniIsAlpha, tok);
// Create alpha-num range
RangeToken* alnumTok = tokFactory->createRange();
alnumTok->mergeRanges(tok);
alnumTok->mergeRanges(ranges[XMLUniCharacter::DECIMAL_DIGIT_NUMBER]);
// build the internal map.
alnumTok->createMap();
rangeTokMap->setRangeToken(fgUniIsAlnum, alnumTok);
// Create word range
tok = tokFactory->createRange();
tok->mergeRanges(alnumTok);
tok->addRange(chUnderscore, chUnderscore);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniIsWord, tok);
tok = (RangeToken*) RangeToken::complementRanges(tok, tokFactory);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniIsWord, tok , true);
// Create assigned range
tok = (RangeToken*)RangeToken::complementRanges(
ranges[XMLUniCharacter::UNASSIGNED],
tokFactory,
tokFactory->getMemoryManager());
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniAssigned,tok);
// Create space range
tok = tokFactory->createRange();
tok->mergeRanges(ranges[XMLUniCharacter::SPACE_SEPARATOR]);
tok->mergeRanges(ranges[XMLUniCharacter::LINE_SEPARATOR]);
//tok->mergeRanges(ranges[XMLUniCharacter::PARAGRAPH_SEPARATOR]);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniIsSpace, tok);
tok = (RangeToken*) RangeToken::complementRanges(tok, tokFactory);
// build the internal map.
tok->createMap();
rangeTokMap->setRangeToken(fgUniIsSpace, tok , true);
// build the internal maps.
for (int l=0; l < UNICATEGSIZE; l++) {
ranges[l]->createMap();
}
fRangesCreated = true;
}
// ---------------------------------------------------------------------------
// RangeToken: Getter methods
// ---------------------------------------------------------------------------
RangeToken* RangeToken::getCaseInsensitiveToken(TokenFactory* const tokFactory) {
if (fCaseIToken == 0 && tokFactory && fRanges) {
bool isNRange = (getTokenType() == T_NRANGE) ? true : false;
RangeToken* lwrToken = tokFactory->createRange(isNRange);
#if XERCES_USE_TRANSCODER_ICU && ((U_ICU_VERSION_MAJOR_NUM > 2) || (U_ICU_VERSION_MAJOR_NUM == 2 && U_ICU_VERSION_MINOR_NUM >=4))
UChar* rangeStr=(UChar*)fMemoryManager->allocate(40*fElemCount*sizeof(UChar));
ArrayJanitor<UChar> janRange(rangeStr, fMemoryManager);
int c=0;
rangeStr[c++] = chOpenSquare;
for (unsigned int i = 0; i < fElemCount - 1; i += 2) {
XMLCh buffer[10];
XMLSize_t len, j;
rangeStr[c++] = chBackSlash;
rangeStr[c++] = chLatin_U;
XMLString::binToText(fRanges[i], buffer, 10, 16, fMemoryManager);
len = XMLString::stringLen(buffer);
for(j=0;j<(8-len);j++)
rangeStr[c++] = chDigit_0;
XMLCh* p=buffer;
while(*p)
rangeStr[c++] = *p++;
if(fRanges[i+1]!=fRanges[i])
{
rangeStr[c++] = chDash;
rangeStr[c++] = chBackSlash;
rangeStr[c++] = chLatin_U;
XMLString::binToText(fRanges[i+1], buffer, 10, 16, fMemoryManager);
len = XMLString::stringLen(buffer);
for(j=0;j<(8-len);j++)
rangeStr[c++] = chDigit_0;
p=buffer;
while(*p)
rangeStr[c++] = *p++;
}
}
rangeStr[c++] = chCloseSquare;
rangeStr[c++] = chNull;
UErrorCode ec=U_ZERO_ERROR;
USet* range=uset_openPatternOptions(rangeStr, -1, USET_CASE_INSENSITIVE, &ec);
if(range)
{
ec = U_ZERO_ERROR;
uint32_t cbCount=uset_serialize(range, NULL, 0, &ec);
uint16_t* buffer=(uint16_t*)fMemoryManager->allocate(cbCount*sizeof(uint16_t));
ArrayJanitor<uint16_t> janSet(buffer, fMemoryManager);
ec = U_ZERO_ERROR;
uset_serialize(range, buffer, cbCount, &ec);
USerializedSet serializedSet;
uset_getSerializedSet(&serializedSet, buffer, cbCount);
int32_t nSets=uset_getSerializedRangeCount(&serializedSet);
for(int32_t i=0; i<nSets; i++)
{
UChar32 start, end;
uset_getSerializedRange(&serializedSet, i, &start, &end);
lwrToken->addRange(start, end);
}
// does this release the memory allocated by the set?
uset_setSerializedToOne(&serializedSet, 32);
uset_close(range);
}
#else
unsigned int exceptIndex = 0;
for (unsigned int i = 0; i < fElemCount - 1; i += 2) {
for (XMLInt32 ch = fRanges[i]; ch <= fRanges[i + 1]; ++ch) {
#if XERCES_USE_TRANSCODER_ICU
const XMLInt32 upperCh = u_toupper(ch);
if (upperCh != ch)
{
lwrToken->addRange(upperCh, upperCh);
}
const XMLInt32 lowerCh = u_tolower(ch);
if (lowerCh != ch)
{
lwrToken->addRange(lowerCh, lowerCh);
}
const XMLInt32 titleCh = u_totitle(ch);
if (titleCh != ch && titleCh != upperCh)
{
lwrToken->addRange(titleCh, titleCh);
}
#else
if (ch >= chLatin_A && ch <= chLatin_Z)
{
ch += chLatin_a - chLatin_A;
lwrToken->addRange(ch, ch);
}
else if (ch >= chLatin_a && ch <= chLatin_z)
{
ch -= chLatin_a - chLatin_A;
lwrToken->addRange(ch, ch);
}
#endif
const unsigned int exceptionsSize =
sizeof(s_exceptions) / sizeof(s_exceptions[0]);
// Add any exception chars. These are characters where the the
// case mapping is not symmetric. (Unicode case mappings are not isomorphic...)
while (exceptIndex < exceptionsSize)
{
if (s_exceptions[exceptIndex].baseChar < ch)
{
++exceptIndex;
}
else if (s_exceptions[exceptIndex].baseChar == ch)
{
const XMLInt32 matchingChar =
s_exceptions[exceptIndex].matchingChar;
lwrToken->addRange(
matchingChar,
matchingChar);
++exceptIndex;
}
else
{
break;
}
}
}
}
lwrToken->mergeRanges(this);
#endif
lwrToken->compactRanges();
lwrToken->createMap();
fCaseIToken = lwrToken;
// TODO(dbertoni) This is a temporary hack until we can change the ABI.
// See Jira issue XERCESC-1866 for more details.
// Overload the fCaseIToken data member to be the case-insensitive token
// that's caching the case-insensitive one. We need this because tokens
// have varying lifetimes.
fCaseIToken->setCaseInsensitiveToken(this);
}
return fCaseIToken;
}