UnicodeString &
FilteredNormalizer2::normalizeSecondAndAppend(UnicodeString &first,
const UnicodeString &second,
UBool doNormalize,
UErrorCode &errorCode) const {
uprv_checkCanGetBuffer(first, errorCode);
uprv_checkCanGetBuffer(second, errorCode);
if(U_FAILURE(errorCode)) {
return first;
}
if(&first==&second) {
errorCode=U_ILLEGAL_ARGUMENT_ERROR;
return first;
}
if(first.isEmpty()) {
if(doNormalize) {
return normalize(second, first, errorCode);
} else {
return first=second;
}
}
// merge the in-filter suffix of the first string with the in-filter prefix of the second
int32_t prefixLimit=set.span(second, 0, USET_SPAN_SIMPLE);
if(prefixLimit!=0) {
UnicodeString prefix(second.tempSubString(0, prefixLimit));
int32_t suffixStart=set.spanBack(first, INT32_MAX, USET_SPAN_SIMPLE);
if(suffixStart==0) {
if(doNormalize) {
norm2.normalizeSecondAndAppend(first, prefix, errorCode);
} else {
norm2.append(first, prefix, errorCode);
}
} else {
UnicodeString middle(first, suffixStart, INT32_MAX);
if(doNormalize) {
norm2.normalizeSecondAndAppend(middle, prefix, errorCode);
} else {
norm2.append(middle, prefix, errorCode);
}
first.replace(suffixStart, INT32_MAX, middle);
}
}
if(prefixLimit<second.length()) {
UnicodeString rest(second.tempSubString(prefixLimit, INT32_MAX));
if(doNormalize) {
normalize(rest, first, USET_SPAN_NOT_CONTAINED, errorCode);
} else {
first.append(rest);
}
}
return first;
}
UNormalizationCheckResult
FilteredNormalizer2::quickCheck(const UnicodeString &s, UErrorCode &errorCode) const {
uprv_checkCanGetBuffer(s, errorCode);
if(U_FAILURE(errorCode)) {
return UNORM_MAYBE;
}
UNormalizationCheckResult result=UNORM_YES;
USetSpanCondition spanCondition=USET_SPAN_SIMPLE;
for(int32_t prevSpanLimit=0; prevSpanLimit<s.length();) {
int32_t spanLimit=set.span(s, prevSpanLimit, spanCondition);
if(spanCondition==USET_SPAN_NOT_CONTAINED) {
spanCondition=USET_SPAN_SIMPLE;
} else {
UNormalizationCheckResult qcResult=
norm2.quickCheck(s.tempSubStringBetween(prevSpanLimit, spanLimit), errorCode);
if(U_FAILURE(errorCode) || qcResult==UNORM_NO) {
return qcResult;
} else if(qcResult==UNORM_MAYBE) {
result=qcResult;
}
spanCondition=USET_SPAN_NOT_CONTAINED;
}
prevSpanLimit=spanLimit;
}
return result;
}
int32_t
FilteredNormalizer2::spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const {
uprv_checkCanGetBuffer(s, errorCode);
if(U_FAILURE(errorCode)) {
return 0;
}
USetSpanCondition spanCondition=USET_SPAN_SIMPLE;
for(int32_t prevSpanLimit=0; prevSpanLimit<s.length();) {
int32_t spanLimit=set.span(s, prevSpanLimit, spanCondition);
if(spanCondition==USET_SPAN_NOT_CONTAINED) {
spanCondition=USET_SPAN_SIMPLE;
} else {
int32_t yesLimit=
prevSpanLimit+
norm2.spanQuickCheckYes(
s.tempSubStringBetween(prevSpanLimit, spanLimit), errorCode);
if(U_FAILURE(errorCode) || yesLimit<spanLimit) {
return yesLimit;
}
spanCondition=USET_SPAN_NOT_CONTAINED;
}
prevSpanLimit=spanLimit;
}
return s.length();
}
UBool
FilteredNormalizer2::isNormalized(const UnicodeString & s, UErrorCode & errorCode) const
{
uprv_checkCanGetBuffer(s, errorCode);
if (U_FAILURE(errorCode))
{
return FALSE;
}
USetSpanCondition spanCondition = USET_SPAN_SIMPLE;
for (int32_t prevSpanLimit = 0; prevSpanLimit < s.length();)
{
int32_t spanLimit = set.span(s, prevSpanLimit, spanCondition);
if (spanCondition == USET_SPAN_NOT_CONTAINED)
{
spanCondition = USET_SPAN_SIMPLE;
}
else
{
if (!norm2.isNormalized(s.tempSubStringBetween(prevSpanLimit, spanLimit), errorCode) ||
U_FAILURE(errorCode)
)
{
return FALSE;
}
spanCondition = USET_SPAN_NOT_CONTAINED;
}
prevSpanLimit = spanLimit;
}
return TRUE;
}
UnicodeString &
FilteredNormalizer2::normalize(const UnicodeString &src,
UnicodeString &dest,
UErrorCode &errorCode) const {
uprv_checkCanGetBuffer(src, errorCode);
if(U_FAILURE(errorCode)) {
dest.setToBogus();
return dest;
}
if(&dest==&src) {
errorCode=U_ILLEGAL_ARGUMENT_ERROR;
return dest;
}
dest.remove();
return normalize(src, dest, USET_SPAN_SIMPLE, errorCode);
}
