// ---------------------------------------------------------------------------
// 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;
}
static jint Character_toTitleCaseImpl(JNIEnv*, jclass, jint codePoint) {
return u_totitle(codePoint);
}
//static jint Character_toTitleCaseImpl(JNIEnv*, jclass, jint codePoint) {
JNIEXPORT jint JNICALL
Java_java_lang_Character_toTitleCaseImpl(JNIEnv*, jclass, jint codePoint) {
return u_totitle(codePoint);
}
jint fastiva_vm_Character_C$__toTitleCaseImpl(jint codePoint) {
return u_totitle(codePoint);
}
uint32
BUnicodeChar::ToTitle(uint32 c)
{
BUnicodeChar();
return u_totitle(c);
}
static void demoCaseMapInC() {
/*
* input=
* "aB<capital sigma>"
* "iI<small dotless i><capital dotted I> "
* "<sharp s> <small lig. ffi>"
* "<small final sigma><small sigma><capital sigma>"
*/
static const UChar input[]={
0x61, 0x42, 0x3a3,
0x69, 0x49, 0x131, 0x130, 0x20,
0xdf, 0x20, 0xfb03,
0x3c2, 0x3c3, 0x3a3, 0
};
UChar buffer[32];
UErrorCode errorCode;
UChar32 c;
int32_t i, j, length;
UBool isError;
printf("\n* demoCaseMapInC() ----------------- ***\n\n");
/*
* First, use simple case mapping functions which provide
* 1:1 code point mappings without context/locale ID.
*
* Note that some mappings will not be "right" because some "real"
* case mappings require context, depend on the locale ID,
* and/or result in a change in the number of code points.
*/
printUString("input string: ", input, -1);
/* uppercase */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_toupper(c);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-uppercased: ", buffer, j);
/* lowercase */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_tolower(c);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-lowercased: ", buffer, j);
/* titlecase */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_totitle(c);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-titlecased: ", buffer, j);
/* case-fold/default */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_foldCase(c, U_FOLD_CASE_DEFAULT);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-case-folded/default: ", buffer, j);
/* case-fold/Turkic */
isError=FALSE;
for(i=j=0; j<UPRV_LENGTHOF(buffer) && !isError; /* U16_NEXT post-increments */) {
U16_NEXT(input, i, INT32_MAX, c); /* without length because NUL-terminated */
if(c==0) {
break; /* stop at terminating NUL, no need to terminate buffer */
}
c=u_foldCase(c, U_FOLD_CASE_EXCLUDE_SPECIAL_I);
U16_APPEND(buffer, j, UPRV_LENGTHOF(buffer), c, isError);
}
printUString("simple-case-folded/Turkic: ", buffer, j);
/*
* Second, use full case mapping functions which provide
* 1:n code point mappings (n can be 0!) and are sensitive to context and locale ID.
*
* Note that lower/upper/titlecasing take a locale ID while case-folding
* has bit flag options instead, by design of the Unicode SpecialCasing.txt UCD file.
*
* Also, string titlecasing requires a BreakIterator to find starts of words.
* The sample code here passes in a NULL pointer; u_strToTitle() will open and close a default
* titlecasing BreakIterator automatically.
* For production code where many strings are titlecased it would be more efficient
* to open a BreakIterator externally and pass it in.
*/
printUString("\ninput string: ", input, -1);
/* lowercase/English */
errorCode=U_ZERO_ERROR;
length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-lowercased/en: ", buffer, length);
} else {
printf("error in u_strToLower(en)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* lowercase/Turkish */
errorCode=U_ZERO_ERROR;
length=u_strToLower(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-lowercased/tr: ", buffer, length);
} else {
printf("error in u_strToLower(tr)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* uppercase/English */
errorCode=U_ZERO_ERROR;
length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "en", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-uppercased/en: ", buffer, length);
} else {
printf("error in u_strToUpper(en)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* uppercase/Turkish */
errorCode=U_ZERO_ERROR;
length=u_strToUpper(buffer, UPRV_LENGTHOF(buffer), input, -1, "tr", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-uppercased/tr: ", buffer, length);
} else {
printf("error in u_strToUpper(tr)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* titlecase/English */
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "en", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-titlecased/en: ", buffer, length);
} else {
printf("error in u_strToTitle(en)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* titlecase/Turkish */
errorCode=U_ZERO_ERROR;
length=u_strToTitle(buffer, UPRV_LENGTHOF(buffer), input, -1, NULL, "tr", &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-titlecased/tr: ", buffer, length);
} else {
printf("error in u_strToTitle(tr)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* case-fold/default */
errorCode=U_ZERO_ERROR;
length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_DEFAULT, &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-case-folded/default: ", buffer, length);
} else {
printf("error in u_strFoldCase(default)=%ld error=%s\n", length, u_errorName(errorCode));
}
/* case-fold/Turkic */
errorCode=U_ZERO_ERROR;
length=u_strFoldCase(buffer, UPRV_LENGTHOF(buffer), input, -1, U_FOLD_CASE_EXCLUDE_SPECIAL_I, &errorCode);
if(U_SUCCESS(errorCode)) {
printUString("full-case-folded/Turkic: ", buffer, length);
} else {
printf("error in u_strFoldCase(Turkic)=%ld error=%s\n", length, u_errorName(errorCode));
}
}
// ---------------------------------------------------------------------------
// 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;
}
