U_CAPI int32_t U_EXPORT2
uspoof_check(const USpoofChecker *sc,
const UChar *text, int32_t length,
int32_t *position,
UErrorCode *status) {
const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
if (This == NULL) {
return 0;
}
if (length < -1) {
*status = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
if (length == -1) {
// It's not worth the bother to handle nul terminated strings everywhere.
// Just get the length and be done with it.
length = u_strlen(text);
}
int32_t result = 0;
int32_t failPos = 0x7fffffff; // TODO: do we have a #define for max int32?
// A count of the number of non-Common or inherited scripts.
// Needed for both the SINGLE_SCRIPT and the WHOLE/MIXED_SCIRPT_CONFUSABLE tests.
// Share the computation when possible. scriptCount == -1 means that we haven't
// done it yet.
int32_t scriptCount = -1;
if ((This->fChecks) & USPOOF_SINGLE_SCRIPT) {
scriptCount = This->scriptScan(text, length, failPos, *status);
// printf("scriptCount (clipped to 2) = %d\n", scriptCount);
if ( scriptCount >= 2) {
// Note: scriptCount == 2 covers all cases of the number of scripts >= 2
result |= USPOOF_SINGLE_SCRIPT;
}
}
if (This->fChecks & USPOOF_CHAR_LIMIT) {
int32_t i;
UChar32 c;
for (i=0; i<length ;) {
U16_NEXT(text, i, length, c);
if (!This->fAllowedCharsSet->contains(c)) {
result |= USPOOF_CHAR_LIMIT;
if (i < failPos) {
failPos = i;
}
break;
}
}
}
if (This->fChecks &
(USPOOF_WHOLE_SCRIPT_CONFUSABLE | USPOOF_MIXED_SCRIPT_CONFUSABLE | USPOOF_INVISIBLE)) {
// These are the checks that need to be done on NFD input
NFDBuffer normalizedInput(text, length, *status);
const UChar *nfdText = normalizedInput.getBuffer();
int32_t nfdLength = normalizedInput.getLength();
if (This->fChecks & USPOOF_INVISIBLE) {
// scan for more than one occurence of the same non-spacing mark
// in a sequence of non-spacing marks.
int32_t i;
UChar32 c;
UChar32 firstNonspacingMark = 0;
UBool haveMultipleMarks = FALSE;
UnicodeSet marksSeenSoFar; // Set of combining marks in a single combining sequence.
for (i=0; i<nfdLength ;) {
U16_NEXT(nfdText, i, nfdLength, c);
if (u_charType(c) != U_NON_SPACING_MARK) {
firstNonspacingMark = 0;
if (haveMultipleMarks) {
marksSeenSoFar.clear();
haveMultipleMarks = FALSE;
}
continue;
}
if (firstNonspacingMark == 0) {
firstNonspacingMark = c;
continue;
}
if (!haveMultipleMarks) {
marksSeenSoFar.add(firstNonspacingMark);
haveMultipleMarks = TRUE;
}
if (marksSeenSoFar.contains(c)) {
// report the error, and stop scanning.
// No need to find more than the first failure.
result |= USPOOF_INVISIBLE;
failPos = i;
// TODO: Bug 8655: failPos is the position in the NFD buffer, but what we want
// to give back to our caller is a position in the original input string.
if (failPos > length) {
failPos = length;
}
break;
}
marksSeenSoFar.add(c);
}
}
if (This->fChecks & (USPOOF_WHOLE_SCRIPT_CONFUSABLE | USPOOF_MIXED_SCRIPT_CONFUSABLE)) {
// The basic test is the same for both whole and mixed script confusables.
// Compute the set of scripts that every input character has a confusable in.
// For this computation an input character is always considered to be
// confusable with itself in its own script.
// If the number of such scripts is two or more, and the input consisted of
// characters all from a single script, we have a whole script confusable.
// (The two scripts will be the original script and the one that is confusable)
// If the number of such scripts >= one, and the original input contained characters from
// more than one script, we have a mixed script confusable. (We can transform
// some of the characters, and end up with a visually similar string all in
// one script.)
if (scriptCount == -1) {
int32_t t;
scriptCount = This->scriptScan(text, length, t, *status);
}
ScriptSet scripts;
This->wholeScriptCheck(nfdText, nfdLength, &scripts, *status);
int32_t confusableScriptCount = scripts.countMembers();
//printf("confusableScriptCount = %d\n", confusableScriptCount);
if ((This->fChecks & USPOOF_WHOLE_SCRIPT_CONFUSABLE) &&
confusableScriptCount >= 2 &&
scriptCount == 1) {
result |= USPOOF_WHOLE_SCRIPT_CONFUSABLE;
}
if ((This->fChecks & USPOOF_MIXED_SCRIPT_CONFUSABLE) &&
confusableScriptCount >= 1 &&
scriptCount > 1) {
result |= USPOOF_MIXED_SCRIPT_CONFUSABLE;
}
}
}
if (position != NULL && failPos != 0x7fffffff) {
*position = failPos;
}
return result;
}
IdentifierInfo &IdentifierInfo::setIdentifier(const UnicodeString &identifier, UErrorCode &status) {
if (U_FAILURE(status)) {
return *this;
}
*fIdentifier = identifier;
clear();
ScriptSet scriptsForCP;
UChar32 cp;
for (int32_t i = 0; i < identifier.length(); i += U16_LENGTH(cp)) {
cp = identifier.char32At(i);
// Store a representative character for each kind of decimal digit
if (u_charType(cp) == U_DECIMAL_DIGIT_NUMBER) {
// Just store the zero character as a representative for comparison. Unicode guarantees it is cp - value
fNumerics->add(cp - (UChar32)u_getNumericValue(cp));
}
UScriptCode extensions[500];
int32_t extensionsCount = uscript_getScriptExtensions(cp, extensions, UPRV_LENGTHOF(extensions), &status);
if (U_FAILURE(status)) {
return *this;
}
scriptsForCP.resetAll();
for (int32_t j=0; j<extensionsCount; j++) {
scriptsForCP.set(extensions[j], status);
}
scriptsForCP.reset(USCRIPT_COMMON, status);
scriptsForCP.reset(USCRIPT_INHERITED, status);
switch (scriptsForCP.countMembers()) {
case 0: break;
case 1:
// Single script, record it.
fRequiredScripts->Union(scriptsForCP);
break;
default:
if (!fRequiredScripts->intersects(scriptsForCP)
&& !uhash_geti(fScriptSetSet, &scriptsForCP)) {
// If the set hasn't been added already, add it
// (Add a copy, fScriptSetSet takes ownership of the copy.)
uhash_puti(fScriptSetSet, new ScriptSet(scriptsForCP), 1, &status);
}
break;
}
}
// Now make a final pass through ScriptSetSet to remove alternates that came before singles.
// [Kana], [Kana Hira] => [Kana]
// This is relatively infrequent, so doesn't have to be optimized.
// We also compute any commonalities among the alternates.
if (uhash_count(fScriptSetSet) > 0) {
fCommonAmongAlternates->setAll();
for (int32_t it = UHASH_FIRST;;) {
const UHashElement *nextHashEl = uhash_nextElement(fScriptSetSet, &it);
if (nextHashEl == NULL) {
break;
}
ScriptSet *next = static_cast<ScriptSet *>(nextHashEl->key.pointer);
// [Kana], [Kana Hira] => [Kana]
if (fRequiredScripts->intersects(*next)) {
uhash_removeElement(fScriptSetSet, nextHashEl);
} else {
fCommonAmongAlternates->intersect(*next);
// [[Arab Syrc Thaa]; [Arab Syrc]] => [[Arab Syrc]]
for (int32_t otherIt = UHASH_FIRST;;) {
const UHashElement *otherHashEl = uhash_nextElement(fScriptSetSet, &otherIt);
if (otherHashEl == NULL) {
break;
}
ScriptSet *other = static_cast<ScriptSet *>(otherHashEl->key.pointer);
if (next != other && next->contains(*other)) {
uhash_removeElement(fScriptSetSet, nextHashEl);
break;
}
}
}
}
}
if (uhash_count(fScriptSetSet) == 0) {
fCommonAmongAlternates->resetAll();
}
return *this;
}
U_CAPI int32_t U_EXPORT2
uspoof_checkUnicodeString(const USpoofChecker *sc,
const icu::UnicodeString &id,
int32_t *position,
UErrorCode *status) {
const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
if (This == NULL) {
return 0;
}
int32_t result = 0;
IdentifierInfo *identifierInfo = NULL;
if ((This->fChecks) & (USPOOF_RESTRICTION_LEVEL | USPOOF_MIXED_NUMBERS)) {
identifierInfo = This->getIdentifierInfo(*status);
if (U_FAILURE(*status)) {
goto cleanupAndReturn;
}
identifierInfo->setIdentifier(id, *status);
identifierInfo->setIdentifierProfile(*This->fAllowedCharsSet);
}
if ((This->fChecks) & USPOOF_RESTRICTION_LEVEL) {
URestrictionLevel idRestrictionLevel = identifierInfo->getRestrictionLevel(*status);
if (idRestrictionLevel > This->fRestrictionLevel) {
result |= USPOOF_RESTRICTION_LEVEL;
}
if (This->fChecks & USPOOF_AUX_INFO) {
result |= idRestrictionLevel;
}
}
if ((This->fChecks) & USPOOF_MIXED_NUMBERS) {
const UnicodeSet *numerics = identifierInfo->getNumerics();
if (numerics->size() > 1) {
result |= USPOOF_MIXED_NUMBERS;
}
// TODO: ICU4J returns the UnicodeSet of the numerics found in the identifier.
// We have no easy way to do the same in C.
// if (checkResult != null) {
// checkResult.numerics = numerics;
// }
}
if (This->fChecks & (USPOOF_CHAR_LIMIT)) {
int32_t i;
UChar32 c;
int32_t length = id.length();
for (i=0; i<length ;) {
c = id.char32At(i);
i += U16_LENGTH(c);
if (!This->fAllowedCharsSet->contains(c)) {
result |= USPOOF_CHAR_LIMIT;
break;
}
}
}
if (This->fChecks &
(USPOOF_WHOLE_SCRIPT_CONFUSABLE | USPOOF_MIXED_SCRIPT_CONFUSABLE | USPOOF_INVISIBLE)) {
// These are the checks that need to be done on NFD input
UnicodeString nfdText;
gNfdNormalizer->normalize(id, nfdText, *status);
int32_t nfdLength = nfdText.length();
if (This->fChecks & USPOOF_INVISIBLE) {
// scan for more than one occurence of the same non-spacing mark
// in a sequence of non-spacing marks.
int32_t i;
UChar32 c;
UChar32 firstNonspacingMark = 0;
UBool haveMultipleMarks = FALSE;
UnicodeSet marksSeenSoFar; // Set of combining marks in a single combining sequence.
for (i=0; i<nfdLength ;) {
c = nfdText.char32At(i);
i += U16_LENGTH(c);
if (u_charType(c) != U_NON_SPACING_MARK) {
firstNonspacingMark = 0;
if (haveMultipleMarks) {
marksSeenSoFar.clear();
haveMultipleMarks = FALSE;
}
continue;
}
if (firstNonspacingMark == 0) {
firstNonspacingMark = c;
continue;
}
if (!haveMultipleMarks) {
marksSeenSoFar.add(firstNonspacingMark);
haveMultipleMarks = TRUE;
}
if (marksSeenSoFar.contains(c)) {
// report the error, and stop scanning.
// No need to find more than the first failure.
result |= USPOOF_INVISIBLE;
break;
}
marksSeenSoFar.add(c);
}
}
if (This->fChecks & (USPOOF_WHOLE_SCRIPT_CONFUSABLE | USPOOF_MIXED_SCRIPT_CONFUSABLE)) {
// The basic test is the same for both whole and mixed script confusables.
// Compute the set of scripts that every input character has a confusable in.
// For this computation an input character is always considered to be
// confusable with itself in its own script.
//
// If the number of such scripts is two or more, and the input consisted of
// characters all from a single script, we have a whole script confusable.
// (The two scripts will be the original script and the one that is confusable)
//
// If the number of such scripts >= one, and the original input contained characters from
// more than one script, we have a mixed script confusable. (We can transform
// some of the characters, and end up with a visually similar string all in
// one script.)
if (identifierInfo == NULL) {
identifierInfo = This->getIdentifierInfo(*status);
if (U_FAILURE(*status)) {
goto cleanupAndReturn;
}
identifierInfo->setIdentifier(id, *status);
}
int32_t scriptCount = identifierInfo->getScriptCount();
ScriptSet scripts;
This->wholeScriptCheck(nfdText, &scripts, *status);
int32_t confusableScriptCount = scripts.countMembers();
//printf("confusableScriptCount = %d\n", confusableScriptCount);
if ((This->fChecks & USPOOF_WHOLE_SCRIPT_CONFUSABLE) &&
confusableScriptCount >= 2 &&
scriptCount == 1) {
result |= USPOOF_WHOLE_SCRIPT_CONFUSABLE;
}
if ((This->fChecks & USPOOF_MIXED_SCRIPT_CONFUSABLE) &&
confusableScriptCount >= 1 &&
scriptCount > 1) {
result |= USPOOF_MIXED_SCRIPT_CONFUSABLE;
}
}
}
cleanupAndReturn:
This->releaseIdentifierInfo(identifierInfo);
if (position != NULL) {
*position = 0;
}
return result;
}
