// Computes the augmented script set for a code point, according to UTS 39 section 5.1.
void SpoofImpl::getAugmentedScriptSet(UChar32 codePoint, ScriptSet& result, UErrorCode& status) {
result.resetAll();
result.setScriptExtensions(codePoint, status);
if (U_FAILURE(status)) { return; }
// Section 5.1 step 1
if (result.test(USCRIPT_HAN, status)) {
result.set(USCRIPT_HAN_WITH_BOPOMOFO, status);
result.set(USCRIPT_JAPANESE, status);
result.set(USCRIPT_KOREAN, status);
}
if (result.test(USCRIPT_HIRAGANA, status)) {
result.set(USCRIPT_JAPANESE, status);
}
if (result.test(USCRIPT_KATAKANA, status)) {
result.set(USCRIPT_JAPANESE, status);
}
if (result.test(USCRIPT_HANGUL, status)) {
result.set(USCRIPT_KOREAN, status);
}
if (result.test(USCRIPT_BOPOMOFO, status)) {
result.set(USCRIPT_HAN_WITH_BOPOMOFO, status);
}
// Section 5.1 step 2
if (result.test(USCRIPT_COMMON, status) || result.test(USCRIPT_INHERITED, status)) {
result.setAll();
}
}
// Computes the resolved script set for a string, omitting characters having the specified script.
// If USCRIPT_CODE_LIMIT is passed as the second argument, all characters are included.
void SpoofImpl::getResolvedScriptSetWithout(const UnicodeString& input, UScriptCode script, ScriptSet& result, UErrorCode& status) const {
result.setAll();
ScriptSet temp;
UChar32 codePoint;
for (int32_t i = 0; i < input.length(); i += U16_LENGTH(codePoint)) {
codePoint = input.char32At(i);
// Compute the augmented script set for the character
getAugmentedScriptSet(codePoint, temp, status);
if (U_FAILURE(status)) { return; }
// Intersect the augmented script set with the resolved script set, but only if the character doesn't
// have the script specified in the function call
if (script == USCRIPT_CODE_LIMIT || !temp.test(script, status)) {
result.intersect(temp);
}
}
}