void U_EXPORT2
Normalizer::normalize(const UnicodeString& source,
UNormalizationMode mode, int32_t options,
UnicodeString& result,
UErrorCode &status) {
if(source.isBogus() || U_FAILURE(status)) {
result.setToBogus();
if(U_SUCCESS(status)) {
status=U_ILLEGAL_ARGUMENT_ERROR;
}
} else {
UnicodeString localDest;
UnicodeString *dest;
if(&source!=&result) {
dest=&result;
} else {
// the source and result strings are the same object, use a temporary one
dest=&localDest;
}
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, status);
if(U_SUCCESS(status)) {
if(options&UNORM_UNICODE_3_2) {
FilteredNormalizer2(*n2, *uniset_getUnicode32Instance(status)).
normalize(source, *dest, status);
} else {
n2->normalize(source, *dest, status);
}
}
if(dest==&localDest && U_SUCCESS(status)) {
result=*dest;
}
}
}
U_CAPI int32_t U_EXPORT2
unorm_concatenate(const UChar * left, int32_t leftLength,
const UChar * right, int32_t rightLength,
UChar * dest, int32_t destCapacity,
UNormalizationMode mode, int32_t options,
UErrorCode * pErrorCode)
{
const Normalizer2 * n2 = Normalizer2Factory::getInstance(mode, *pErrorCode);
const UnicodeSet * uni32;
if (options & UNORM_UNICODE_3_2)
{
uni32 = uniset_getUnicode32Instance(*pErrorCode);
}
else
{
uni32 = NULL; // unused
}
FilteredNormalizer2 fn2(*n2, *uni32);
if (options & UNORM_UNICODE_3_2)
{
n2 = &fn2;
}
if (U_FAILURE(*pErrorCode))
{
return 0;
}
if (destCapacity < 0 || (dest == NULL && destCapacity > 0) ||
left == NULL || leftLength < -1 ||
right == NULL || rightLength < -1
)
{
*pErrorCode = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
/* check for overlapping right and destination */
if (dest != NULL &&
((right >= dest && right < (dest + destCapacity)) ||
(rightLength > 0 && dest >= right && dest < (right + rightLength)))
)
{
*pErrorCode = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
/* allow left==dest */
UnicodeString destString;
if (left == dest)
{
destString.setTo(dest, leftLength, destCapacity);
}
else
{
destString.setTo(dest, 0, destCapacity);
destString.append(left, leftLength);
}
return n2->append(destString, UnicodeString(rightLength < 0, right, rightLength), *pErrorCode).
extract(dest, destCapacity, *pErrorCode);
}
U_CAPI UBool U_EXPORT2
unorm_isNormalizedWithOptions(const UChar *src, int32_t srcLength,
UNormalizationMode mode, int32_t options,
UErrorCode *pErrorCode) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, *pErrorCode);
if(options&UNORM_UNICODE_3_2) {
FilteredNormalizer2 fn2(*n2, *uniset_getUnicode32Instance(*pErrorCode));
return unorm2_isNormalized((const UNormalizer2 *)&fn2, src, srcLength, pErrorCode);
} else {
return unorm2_isNormalized((const UNormalizer2 *)n2, src, srcLength, pErrorCode);
}
}
void
Normalizer::init() {
UErrorCode errorCode=U_ZERO_ERROR;
fNorm2=Normalizer2Factory::getInstance(fUMode, errorCode);
if(fOptions&UNORM_UNICODE_3_2) {
delete fFilteredNorm2;
fNorm2=fFilteredNorm2=
new FilteredNormalizer2(*fNorm2, *uniset_getUnicode32Instance(errorCode));
}
if(U_FAILURE(errorCode)) {
errorCode=U_ZERO_ERROR;
fNorm2=Normalizer2Factory::getNoopInstance(errorCode);
}
}
U_CAPI UNormalizationCheckResult U_EXPORT2
unorm_quickCheckWithOptions(const UChar *src, int32_t srcLength,
UNormalizationMode mode, int32_t options,
UErrorCode *pErrorCode) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, *pErrorCode);
if(options&UNORM_UNICODE_3_2) {
FilteredNormalizer2 fn2(*n2, *uniset_getUnicode32Instance(*pErrorCode));
return unorm2_quickCheck(
reinterpret_cast<const UNormalizer2 *>(static_cast<Normalizer2 *>(&fn2)),
src, srcLength, pErrorCode);
} else {
return unorm2_quickCheck((const UNormalizer2 *)n2, src, srcLength, pErrorCode);
}
}
UnicodeString & U_EXPORT2
Normalizer::concatenate(UnicodeString & left, UnicodeString & right,
UnicodeString & result,
UNormalizationMode mode, int32_t options,
UErrorCode & errorCode)
{
if (left.isBogus() || right.isBogus() || U_FAILURE(errorCode))
{
result.setToBogus();
if (U_SUCCESS(errorCode))
{
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
}
}
else
{
UnicodeString localDest;
UnicodeString * dest;
if (&right != &result)
{
dest = &result;
}
else
{
// the right and result strings are the same object, use a temporary one
dest = &localDest;
}
*dest = left;
const Normalizer2 * n2 = Normalizer2Factory::getInstance(mode, errorCode);
if (U_SUCCESS(errorCode))
{
if (options & UNORM_UNICODE_3_2)
{
FilteredNormalizer2(*n2, *uniset_getUnicode32Instance(errorCode)).
append(*dest, right, errorCode);
}
else
{
n2->append(*dest, right, errorCode);
}
}
if (dest == &localDest && U_SUCCESS(errorCode))
{
result = *dest;
}
}
return result;
}
/** Public API for normalizing. */
U_CAPI int32_t U_EXPORT2
unorm_normalize(const UChar *src, int32_t srcLength,
UNormalizationMode mode, int32_t options,
UChar *dest, int32_t destCapacity,
UErrorCode *pErrorCode) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, *pErrorCode);
if(options&UNORM_UNICODE_3_2) {
FilteredNormalizer2 fn2(*n2, *uniset_getUnicode32Instance(*pErrorCode));
return unorm2_normalize((const UNormalizer2 *)&fn2,
src, srcLength, dest, destCapacity, pErrorCode);
} else {
return unorm2_normalize((const UNormalizer2 *)n2,
src, srcLength, dest, destCapacity, pErrorCode);
}
}
UBool
Normalizer::isNormalized(const UnicodeString& source,
UNormalizationMode mode, int32_t options,
UErrorCode &status) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, status);
if(U_SUCCESS(status)) {
if(options&UNORM_UNICODE_3_2) {
return FilteredNormalizer2(*n2, *uniset_getUnicode32Instance(status)).
isNormalized(source, status);
} else {
return n2->isNormalized(source, status);
}
} else {
return FALSE;
}
}
UNormalizationCheckResult
Normalizer::quickCheck(const UnicodeString& source,
UNormalizationMode mode, int32_t options,
UErrorCode &status) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, status);
if(U_SUCCESS(status)) {
if(options&UNORM_UNICODE_3_2) {
return FilteredNormalizer2(*n2, *uniset_getUnicode32Instance(status)).
quickCheck(source, status);
} else {
return n2->quickCheck(source, status);
}
} else {
return UNORM_MAYBE;
}
}
U_CAPI int32_t U_EXPORT2
unorm_concatenate(const UChar *left, int32_t leftLength,
const UChar *right, int32_t rightLength,
UChar *dest, int32_t destCapacity,
UNormalizationMode mode, int32_t options,
UErrorCode *pErrorCode) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, *pErrorCode);
if(options&UNORM_UNICODE_3_2) {
const UnicodeSet *uni32 = uniset_getUnicode32Instance(*pErrorCode);
if(U_FAILURE(*pErrorCode)) {
return 0;
}
FilteredNormalizer2 fn2(*n2, *uni32);
return _concatenate(left, leftLength, right, rightLength,
dest, destCapacity, &fn2, pErrorCode);
}
return _concatenate(left, leftLength, right, rightLength,
dest, destCapacity, n2, pErrorCode);
}
static int32_t
unorm_iterate(UCharIterator *src, UBool forward,
UChar *dest, int32_t destCapacity,
UNormalizationMode mode, int32_t options,
UBool doNormalize, UBool *pNeededToNormalize,
UErrorCode *pErrorCode) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, *pErrorCode);
if(options&UNORM_UNICODE_3_2) {
const UnicodeSet *uni32 = uniset_getUnicode32Instance(*pErrorCode);
if(U_FAILURE(*pErrorCode)) {
return 0;
}
FilteredNormalizer2 fn2(*n2, *uni32);
return _iterate(src, forward, dest, destCapacity,
&fn2, doNormalize, pNeededToNormalize, pErrorCode);
}
return _iterate(src, forward, dest, destCapacity,
n2, doNormalize, pNeededToNormalize, pErrorCode);
}
U_CAPI int32_t U_EXPORT2
unorm_compare(const UChar *s1, int32_t length1,
const UChar *s2, int32_t length2,
uint32_t options,
UErrorCode *pErrorCode) {
/* argument checking */
if(U_FAILURE(*pErrorCode)) {
return 0;
}
if(s1==0 || length1<-1 || s2==0 || length2<-1) {
*pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
UnicodeString fcd1, fcd2;
int32_t normOptions=(int32_t)(options>>UNORM_COMPARE_NORM_OPTIONS_SHIFT);
options|=_COMPARE_EQUIV;
/*
* UAX #21 Case Mappings, as fixed for Unicode version 4
* (see Jitterbug 2021), defines a canonical caseless match as
*
* A string X is a canonical caseless match
* for a string Y if and only if
* NFD(toCasefold(NFD(X))) = NFD(toCasefold(NFD(Y)))
*
* For better performance, we check for FCD (or let the caller tell us that
* both strings are in FCD) for the inner normalization.
* BasicNormalizerTest::FindFoldFCDExceptions() makes sure that
* case-folding preserves the FCD-ness of a string.
* The outer normalization is then only performed by unorm_cmpEquivFold()
* when there is a difference.
*
* Exception: When using the Turkic case-folding option, we do perform
* full NFD first. This is because in the Turkic case precomposed characters
* with 0049 capital I or 0069 small i fold differently whether they
* are first decomposed or not, so an FCD check - a check only for
* canonical order - is not sufficient.
*/
if(!(options&UNORM_INPUT_IS_FCD) || (options&U_FOLD_CASE_EXCLUDE_SPECIAL_I)) {
const Normalizer2 *n2;
if(options&U_FOLD_CASE_EXCLUDE_SPECIAL_I) {
n2=Normalizer2Factory::getNFDInstance(*pErrorCode);
} else {
n2=Normalizer2Factory::getFCDInstance(*pErrorCode);
}
if (U_FAILURE(*pErrorCode)) {
return 0;
}
// check if s1 and/or s2 fulfill the FCD conditions
const UnicodeSet *uni32;
if(normOptions&UNORM_UNICODE_3_2) {
uni32=uniset_getUnicode32Instance(*pErrorCode);
} else {
uni32=NULL; // unused
}
FilteredNormalizer2 fn2(*n2, *uni32);
if(normOptions&UNORM_UNICODE_3_2) {
n2=&fn2;
}
UnicodeString str1(length1<0, s1, length1);
UnicodeString str2(length2<0, s2, length2);
int32_t spanQCYes1=n2->spanQuickCheckYes(str1, *pErrorCode);
int32_t spanQCYes2=n2->spanQuickCheckYes(str2, *pErrorCode);
if(U_FAILURE(*pErrorCode)) {
return 0;
}
/*
* ICU 2.4 had a further optimization:
* If both strings were not in FCD, then they were both NFD'ed,
* and the _COMPARE_EQUIV option was turned off.
* It is not entirely clear that this is valid with the current
* definition of the canonical caseless match.
* Therefore, ICU 2.6 removes that optimization.
*/
if(spanQCYes1<str1.length()) {
UnicodeString unnormalized=str1.tempSubString(spanQCYes1);
fcd1.setTo(FALSE, str1.getBuffer(), spanQCYes1);
n2->normalizeSecondAndAppend(fcd1, unnormalized, *pErrorCode);
s1=fcd1.getBuffer();
length1=fcd1.length();
}
if(spanQCYes2<str2.length()) {
UnicodeString unnormalized=str2.tempSubString(spanQCYes2);
fcd2.setTo(FALSE, str2.getBuffer(), spanQCYes2);
n2->normalizeSecondAndAppend(fcd2, unnormalized, *pErrorCode);
s2=fcd2.getBuffer();
length2=fcd2.length();
}
}
if(U_SUCCESS(*pErrorCode)) {
return unorm_cmpEquivFold(s1, length1, s2, length2, options, pErrorCode);
} else {
return 0;
}
}
static int32_t
unorm_iterate(UCharIterator *src, UBool forward,
UChar *dest, int32_t destCapacity,
UNormalizationMode mode, int32_t options,
UBool doNormalize, UBool *pNeededToNormalize,
UErrorCode *pErrorCode) {
const Normalizer2 *n2=Normalizer2Factory::getInstance(mode, *pErrorCode);
const UnicodeSet *uni32;
if(options&UNORM_UNICODE_3_2) {
uni32=uniset_getUnicode32Instance(*pErrorCode);
} else {
uni32=NULL; // unused
}
if(U_FAILURE(*pErrorCode)) {
return 0;
}
FilteredNormalizer2 fn2(*n2, *uni32);
if(options&UNORM_UNICODE_3_2) {
n2=&fn2;
}
if( destCapacity<0 || (dest==NULL && destCapacity>0) ||
src==NULL
) {
*pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
if(pNeededToNormalize!=NULL) {
*pNeededToNormalize=FALSE;
}
if(!(forward ? src->hasNext(src) : src->hasPrevious(src))) {
return u_terminateUChars(dest, destCapacity, 0, pErrorCode);
}
UnicodeString buffer;
UChar32 c;
if(forward) {
/* get one character and ignore its properties */
buffer.append(uiter_next32(src));
/* get all following characters until we see a boundary */
while((c=uiter_next32(src))>=0) {
if(n2->hasBoundaryBefore(c)) {
/* back out the latest movement to stop at the boundary */
src->move(src, -U16_LENGTH(c), UITER_CURRENT);
break;
} else {
buffer.append(c);
}
}
} else {
while((c=uiter_previous32(src))>=0) {
/* always write this character to the front of the buffer */
buffer.insert(0, c);
/* stop if this just-copied character is a boundary */
if(n2->hasBoundaryBefore(c)) {
break;
}
}
}
UnicodeString destString(dest, 0, destCapacity);
if(buffer.length()>0 && doNormalize) {
n2->normalize(buffer, destString, *pErrorCode).extract(dest, destCapacity, *pErrorCode);
if(pNeededToNormalize!=NULL && U_SUCCESS(*pErrorCode)) {
*pNeededToNormalize= destString!=buffer;
}
return destString.length();
} else {
/* just copy the source characters */
return buffer.extract(dest, destCapacity, *pErrorCode);
}
}