void IntlTestSpoof::testSpoofAPI() {
TEST_SETUP
UnicodeString s("xyz"); // Many latin ranges are whole-script confusable with other scripts.
// If this test starts failing, consult confusablesWholeScript.txt
int32_t position = 666;
int32_t checkResults = uspoof_checkUnicodeString(sc, s, &position, &status);
TEST_ASSERT_SUCCESS(status);
TEST_ASSERT_EQ(0, checkResults);
TEST_ASSERT_EQ(666, position);
TEST_TEARDOWN;
TEST_SETUP
UnicodeString s1("cxs");
UnicodeString s2 = UnicodeString("\\u0441\\u0445\\u0455").unescape(); // Cyrillic "cxs"
int32_t checkResults = uspoof_areConfusableUnicodeString(sc, s1, s2, &status);
TEST_ASSERT_EQ(USPOOF_MIXED_SCRIPT_CONFUSABLE | USPOOF_WHOLE_SCRIPT_CONFUSABLE, checkResults);
TEST_TEARDOWN;
TEST_SETUP
UnicodeString s("I1l0O");
UnicodeString dest;
UnicodeString &retStr = uspoof_getSkeletonUnicodeString(sc, USPOOF_ANY_CASE, s, dest, &status);
TEST_ASSERT_SUCCESS(status);
TEST_ASSERT(UnicodeString("lllOO") == dest);
TEST_ASSERT(&dest == &retStr);
TEST_TEARDOWN;
}
U_CAPI int32_t U_EXPORT2
uspoof_getSkeletonUTF8(const USpoofChecker *sc,
uint32_t type,
const char *id, int32_t length,
char *dest, int32_t destCapacity,
UErrorCode *status) {
SpoofImpl::validateThis(sc, *status);
if (U_FAILURE(*status)) {
return 0;
}
if (length<-1 || destCapacity<0 || (destCapacity==0 && dest!=NULL)) {
*status = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
UnicodeString srcStr = UnicodeString::fromUTF8(StringPiece(id, length>=0 ? length : uprv_strlen(id)));
UnicodeString destStr;
uspoof_getSkeletonUnicodeString(sc, type, srcStr, destStr, status);
if (U_FAILURE(*status)) {
return 0;
}
int32_t lengthInUTF8 = 0;
u_strToUTF8(dest, destCapacity, &lengthInUTF8,
destStr.getBuffer(), destStr.length(), status);
return lengthInUTF8;
}
void IntlTestSpoof::testSpoofAPI() {
TEST_SETUP
UnicodeString s("uvw");
int32_t position = 666;
int32_t checkResults = uspoof_checkUnicodeString(sc, s, &position, &status);
TEST_ASSERT_SUCCESS(status);
TEST_ASSERT_EQ(0, checkResults);
TEST_ASSERT_EQ(666, position);
TEST_TEARDOWN;
TEST_SETUP
UnicodeString s1("cxs");
UnicodeString s2 = UnicodeString("\\u0441\\u0445\\u0455").unescape(); // Cyrillic "cxs"
int32_t checkResults = uspoof_areConfusableUnicodeString(sc, s1, s2, &status);
TEST_ASSERT_EQ(USPOOF_MIXED_SCRIPT_CONFUSABLE | USPOOF_WHOLE_SCRIPT_CONFUSABLE, checkResults);
TEST_TEARDOWN;
TEST_SETUP
UnicodeString s("I1l0O");
UnicodeString dest;
UnicodeString &retStr = uspoof_getSkeletonUnicodeString(sc, USPOOF_ANY_CASE, s, dest, &status);
TEST_ASSERT_SUCCESS(status);
TEST_ASSERT(UnicodeString("11100") == dest);
TEST_ASSERT(&dest == &retStr);
TEST_TEARDOWN;
}
//
// Run a single confusable skeleton transformation test case.
//
void IntlTestSpoof::checkSkeleton(const USpoofChecker *sc, uint32_t type,
const char *input, const char *expected, int32_t lineNum) {
UnicodeString uInput = UnicodeString(input).unescape();
UnicodeString uExpected = UnicodeString(expected).unescape();
UErrorCode status = U_ZERO_ERROR;
UnicodeString actual;
uspoof_getSkeletonUnicodeString(sc, type, uInput, actual, &status);
if (U_FAILURE(status)) {
errln("File %s, Line %d, Test case from line %d, status is %s", __FILE__, __LINE__, lineNum,
u_errorName(status));
return;
}
if (uExpected != actual) {
errln("File %s, Line %d, Test case from line %d, Actual and Expected skeletons differ.",
__FILE__, __LINE__, lineNum);
errln(UnicodeString(" Actual Skeleton: \"") + actual + UnicodeString("\"\n") +
UnicodeString(" Expected Skeleton: \"") + uExpected + UnicodeString("\""));
}
}
U_CAPI int32_t U_EXPORT2
uspoof_getSkeleton(const USpoofChecker *sc,
uint32_t type,
const UChar *id, int32_t length,
UChar *dest, int32_t destCapacity,
UErrorCode *status) {
SpoofImpl::validateThis(sc, *status);
if (U_FAILURE(*status)) {
return 0;
}
if (length<-1 || destCapacity<0 || (destCapacity==0 && dest!=NULL)) {
*status = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
UnicodeString idStr((length==-1), id, length); // Aliasing constructor
UnicodeString destStr;
uspoof_getSkeletonUnicodeString(sc, type, idStr, destStr, status);
destStr.extract(dest, destCapacity, *status);
return destStr.length();
}
// testConfData - Check each data item from the Unicode confusables.txt file,
// verify that it transforms correctly in a skeleton.
//
void IntlTestSpoof::testConfData() {
UErrorCode status = U_ZERO_ERROR;
const char *testDataDir = IntlTest::getSourceTestData(status);
TEST_ASSERT_SUCCESS(status);
char buffer[2000];
uprv_strcpy(buffer, testDataDir);
uprv_strcat(buffer, "confusables.txt");
LocalStdioFilePointer f(fopen(buffer, "rb"));
if (f.isNull()) {
errln("Skipping test spoof/testConfData. File confusables.txt not accessible.");
return;
}
fseek(f.getAlias(), 0, SEEK_END);
int32_t fileSize = ftell(f.getAlias());
LocalArray<char> fileBuf(new char[fileSize]);
fseek(f.getAlias(), 0, SEEK_SET);
int32_t amt_read = fread(fileBuf.getAlias(), 1, fileSize, f.getAlias());
TEST_ASSERT_EQ(amt_read, fileSize);
TEST_ASSERT(fileSize>0);
if (amt_read != fileSize || fileSize <=0) {
return;
}
UnicodeString confusablesTxt = UnicodeString::fromUTF8(StringPiece(fileBuf.getAlias(), fileSize));
LocalUSpoofCheckerPointer sc(uspoof_open(&status));
TEST_ASSERT_SUCCESS(status);
// Parse lines from the confusables.txt file. Example Line:
// FF44 ; 0064 ; SL # ( d -> d ) FULLWIDTH ....
// Three fields. The hex fields can contain more than one character,
// and each character may be more than 4 digits (for supplemntals)
// This regular expression matches lines and splits the fields into capture groups.
RegexMatcher parseLine("(?m)^([0-9A-F]{4}[^#;]*?);([^#;]*?);([^#]*)", confusablesTxt, 0, status);
TEST_ASSERT_SUCCESS(status);
while (parseLine.find()) {
UnicodeString from = parseHex(parseLine.group(1, status));
if (!Normalizer::isNormalized(from, UNORM_NFD, status)) {
// The source character was not NFD.
// Skip this case; the first step in obtaining a skeleton is to NFD the input,
// so the mapping in this line of confusables.txt will never be applied.
continue;
}
UnicodeString rawExpected = parseHex(parseLine.group(2, status));
UnicodeString expected;
Normalizer::decompose(rawExpected, FALSE /*NFD*/, 0, expected, status);
TEST_ASSERT_SUCCESS(status);
int32_t skeletonType = 0;
UnicodeString tableType = parseLine.group(3, status);
TEST_ASSERT_SUCCESS(status);
if (tableType.indexOf("SL") >= 0) {
skeletonType = USPOOF_SINGLE_SCRIPT_CONFUSABLE;
} else if (tableType.indexOf("SA") >= 0) {
skeletonType = USPOOF_SINGLE_SCRIPT_CONFUSABLE | USPOOF_ANY_CASE;
} else if (tableType.indexOf("ML") >= 0) {
skeletonType = 0;
} else if (tableType.indexOf("MA") >= 0) {
skeletonType = USPOOF_ANY_CASE;
}
UnicodeString actual;
uspoof_getSkeletonUnicodeString(sc.getAlias(), skeletonType, from, actual, &status);
TEST_ASSERT_SUCCESS(status);
TEST_ASSERT(actual == expected);
if (actual != expected) {
errln(parseLine.group(0, status));
UnicodeString line = "Actual: ";
int i = 0;
while (i < actual.length()) {
appendHexUChar(line, actual.char32At(i));
i = actual.moveIndex32(i, 1);
}
errln(line);
}
if (U_FAILURE(status)) {
break;
}
}
}
U_CAPI int32_t U_EXPORT2
uspoof_areConfusableUnicodeString(const USpoofChecker *sc,
const icu::UnicodeString &id1,
const icu::UnicodeString &id2,
UErrorCode *status) {
const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
if (U_FAILURE(*status)) {
return 0;
}
//
// See section 4 of UAX 39 for the algorithm for checking whether two strings are confusable,
// and for definitions of the types (single, whole, mixed-script) of confusables.
// We only care about a few of the check flags. Ignore the others.
// If no tests relavant to this function have been specified, return an error.
// TODO: is this really the right thing to do? It's probably an error on the caller's part,
// but logically we would just return 0 (no error).
if ((This->fChecks & (USPOOF_SINGLE_SCRIPT_CONFUSABLE | USPOOF_MIXED_SCRIPT_CONFUSABLE |
USPOOF_WHOLE_SCRIPT_CONFUSABLE)) == 0) {
*status = U_INVALID_STATE_ERROR;
return 0;
}
int32_t flagsForSkeleton = This->fChecks & USPOOF_ANY_CASE;
int32_t result = 0;
IdentifierInfo *identifierInfo = This->getIdentifierInfo(*status);
if (U_FAILURE(*status)) {
return 0;
}
identifierInfo->setIdentifier(id1, *status);
int32_t id1ScriptCount = identifierInfo->getScriptCount();
int32_t id1FirstScript = identifierInfo->getScripts()->nextSetBit(0);
identifierInfo->setIdentifier(id2, *status);
int32_t id2ScriptCount = identifierInfo->getScriptCount();
int32_t id2FirstScript = identifierInfo->getScripts()->nextSetBit(0);
This->releaseIdentifierInfo(identifierInfo);
identifierInfo = NULL;
if (This->fChecks & USPOOF_SINGLE_SCRIPT_CONFUSABLE) {
UnicodeString id1Skeleton;
UnicodeString id2Skeleton;
if (id1ScriptCount <= 1 && id2ScriptCount <= 1 && id1FirstScript == id2FirstScript) {
flagsForSkeleton |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id1, id1Skeleton, status);
uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id2, id2Skeleton, status);
if (id1Skeleton == id2Skeleton) {
result |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
}
}
}
if (result & USPOOF_SINGLE_SCRIPT_CONFUSABLE) {
// If the two inputs are single script confusable they cannot also be
// mixed or whole script confusable, according to the UAX39 definitions.
// So we can skip those tests.
return result;
}
// Two identifiers are whole script confusable if each is of a single script
// and they are mixed script confusable.
UBool possiblyWholeScriptConfusables =
id1ScriptCount <= 1 && id2ScriptCount <= 1 && (This->fChecks & USPOOF_WHOLE_SCRIPT_CONFUSABLE);
//
// Mixed Script Check
//
if ((This->fChecks & USPOOF_MIXED_SCRIPT_CONFUSABLE) || possiblyWholeScriptConfusables ) {
// For getSkeleton(), resetting the USPOOF_SINGLE_SCRIPT_CONFUSABLE flag will get us
// the mixed script table skeleton, which is what we want.
// The Any Case / Lower Case bit in the skelton flags was set at the top of the function.
UnicodeString id1Skeleton;
UnicodeString id2Skeleton;
flagsForSkeleton &= ~USPOOF_SINGLE_SCRIPT_CONFUSABLE;
uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id1, id1Skeleton, status);
uspoof_getSkeletonUnicodeString(sc, flagsForSkeleton, id2, id2Skeleton, status);
if (id1Skeleton == id2Skeleton) {
result |= USPOOF_MIXED_SCRIPT_CONFUSABLE;
if (possiblyWholeScriptConfusables) {
result |= USPOOF_WHOLE_SCRIPT_CONFUSABLE;
}
}
}
return result;
}
U_CAPI int32_t U_EXPORT2
uspoof_areConfusableUnicodeString(const USpoofChecker *sc,
const icu::UnicodeString &id1,
const icu::UnicodeString &id2,
UErrorCode *status) {
const SpoofImpl *This = SpoofImpl::validateThis(sc, *status);
if (U_FAILURE(*status)) {
return 0;
}
//
// See section 4 of UAX 39 for the algorithm for checking whether two strings are confusable,
// and for definitions of the types (single, whole, mixed-script) of confusables.
// We only care about a few of the check flags. Ignore the others.
// If no tests relavant to this function have been specified, return an error.
// TODO: is this really the right thing to do? It's probably an error on the caller's part,
// but logically we would just return 0 (no error).
if ((This->fChecks & USPOOF_CONFUSABLE) == 0) {
*status = U_INVALID_STATE_ERROR;
return 0;
}
// Compute the skeletons and check for confusability.
UnicodeString id1Skeleton;
uspoof_getSkeletonUnicodeString(sc, 0 /* deprecated */, id1, id1Skeleton, status);
UnicodeString id2Skeleton;
uspoof_getSkeletonUnicodeString(sc, 0 /* deprecated */, id2, id2Skeleton, status);
if (U_FAILURE(*status)) { return 0; }
if (id1Skeleton != id2Skeleton) {
return 0;
}
// If we get here, the strings are confusable. Now we just need to set the flags for the appropriate classes
// of confusables according to UTS 39 section 4.
// Start by computing the resolved script sets of id1 and id2.
ScriptSet id1RSS;
This->getResolvedScriptSet(id1, id1RSS, *status);
ScriptSet id2RSS;
This->getResolvedScriptSet(id2, id2RSS, *status);
// Turn on all applicable flags
int32_t result = 0;
if (id1RSS.intersects(id2RSS)) {
result |= USPOOF_SINGLE_SCRIPT_CONFUSABLE;
} else {
result |= USPOOF_MIXED_SCRIPT_CONFUSABLE;
if (!id1RSS.isEmpty() && !id2RSS.isEmpty()) {
result |= USPOOF_WHOLE_SCRIPT_CONFUSABLE;
}
}
// Turn off flags that the user doesn't want
if ((This->fChecks & USPOOF_SINGLE_SCRIPT_CONFUSABLE) == 0) {
result &= ~USPOOF_SINGLE_SCRIPT_CONFUSABLE;
}
if ((This->fChecks & USPOOF_MIXED_SCRIPT_CONFUSABLE) == 0) {
result &= ~USPOOF_MIXED_SCRIPT_CONFUSABLE;
}
if ((This->fChecks & USPOOF_WHOLE_SCRIPT_CONFUSABLE) == 0) {
result &= ~USPOOF_WHOLE_SCRIPT_CONFUSABLE;
}
return result;
}
