static int Character_forNameImpl(JNIEnv* env, jclass, jstring javaBlockName) {
ScopedUtfChars blockName(env, javaBlockName);
if (blockName.c_str() == NULL) {
return 0;
}
return u_getPropertyValueEnum(UCHAR_BLOCK, blockName.c_str());
}
ScriptSet &ScriptSet::parseScripts(const UnicodeString &scriptString, UErrorCode &status) {
resetAll();
if (U_FAILURE(status)) {
return *this;
}
UnicodeString oneScriptName;
for (int32_t i=0; i<scriptString.length();) {
UChar32 c = scriptString.char32At(i);
i = scriptString.moveIndex32(i, 1);
if (!u_isUWhiteSpace(c)) {
oneScriptName.append(c);
if (i < scriptString.length()) {
continue;
}
}
if (oneScriptName.length() > 0) {
char buf[40];
oneScriptName.extract(0, oneScriptName.length(), buf, sizeof(buf)-1, US_INV);
buf[sizeof(buf)-1] = 0;
int32_t sc = u_getPropertyValueEnum(UCHAR_SCRIPT, buf);
if (sc == UCHAR_INVALID_CODE) {
status = U_ILLEGAL_ARGUMENT_ERROR;
} else {
this->set((UScriptCode)sc, status);
}
if (U_FAILURE(status)) {
return *this;
}
oneScriptName.remove();
}
}
return *this;
}
static void U_CALLCONV
bidiClassLineFn(void *context,
char *fields[][2], int32_t fieldCount,
UErrorCode *pErrorCode) {
char *s;
uint32_t start, end, value;
/* get the code point range */
u_parseCodePointRange(fields[0][0], &start, &end, pErrorCode);
if(U_FAILURE(*pErrorCode)) {
fprintf(stderr, "genbidi: syntax error in DerivedBidiClass.txt field 0 at %s\n", fields[0][0]);
exit(*pErrorCode);
}
/* parse bidi class */
s=trimTerminateField(fields[1][0], fields[1][1]);
value=u_getPropertyValueEnum(UCHAR_BIDI_CLASS, s);
if((int32_t)value<0) {
fprintf(stderr, "genbidi error: unknown bidi class in DerivedBidiClass.txt field 1 at %s\n", s);
exit(U_PARSE_ERROR);
}
upvec_setValue(pv, start, end, 0, value, UBIDI_CLASS_MASK, pErrorCode);
if(U_FAILURE(*pErrorCode)) {
fprintf(stderr, "genbidi error: unable to set derived bidi class for U+%04x..U+%04x - %s\n",
(int)start, (int)end, u_errorName(*pErrorCode));
exit(*pErrorCode);
}
}
void TestBinaryValues() {
/*
* Unicode 5.1 explicitly defines binary property value aliases.
* Verify that they are all recognized.
*/
static const char *const falseValues[]={ "N", "No", "F", "False" };
static const char *const trueValues[]={ "Y", "Yes", "T", "True" };
int32_t i;
for(i=0; i<LENGTHOF(falseValues); ++i) {
if(FALSE!=u_getPropertyValueEnum(UCHAR_ALPHABETIC, falseValues[i])) {
log_data_err("u_getPropertyValueEnum(UCHAR_ALPHABETIC, \"%s\")!=FALSE (Are you missing data?)\n", falseValues[i]);
}
}
for(i=0; i<LENGTHOF(trueValues); ++i) {
if(TRUE!=u_getPropertyValueEnum(UCHAR_ALPHABETIC, trueValues[i])) {
log_data_err("u_getPropertyValueEnum(UCHAR_ALPHABETIC, \"%s\")!=TRUE (Are you missing data?)\n", trueValues[i]);
}
}
}
static void U_CALLCONV
singleEnumLineFn(void *context,
char *fields[][2], int32_t fieldCount,
UErrorCode *pErrorCode) {
const SingleEnum *sen;
char *s;
uint32_t start, end, uv;
int32_t value;
sen=(const SingleEnum *)context;
u_parseCodePointRange(fields[0][0], &start, &end, pErrorCode);
if(U_FAILURE(*pErrorCode)) {
fprintf(stderr, "genprops: syntax error in %s.txt field 0 at %s\n", sen->ucdFile, fields[0][0]);
exit(*pErrorCode);
}
/* parse property alias */
s=trimTerminateField(fields[1][0], fields[1][1]);
value=u_getPropertyValueEnum(sen->prop, s);
if(value<0) {
if(sen->prop==UCHAR_BLOCK) {
if(isToken("Greek", s)) {
value=UBLOCK_GREEK; /* Unicode 3.2 renames this to "Greek and Coptic" */
} else if(isToken("Combining Marks for Symbols", s)) {
value=UBLOCK_COMBINING_MARKS_FOR_SYMBOLS; /* Unicode 3.2 renames this to "Combining Diacritical Marks for Symbols" */
} else if(isToken("Private Use", s)) {
value=UBLOCK_PRIVATE_USE; /* Unicode 3.2 renames this to "Private Use Area" */
}
}
}
if(value<0) {
fprintf(stderr, "genprops error: unknown %s name in %s.txt field 1 at %s\n",
sen->propName, sen->ucdFile, s);
exit(U_PARSE_ERROR);
}
uv=(uint32_t)(value<<sen->vecShift);
if((uv&sen->vecMask)!=uv) {
fprintf(stderr, "genprops error: %s value overflow (0x%x) at %s\n",
sen->propName, (int)uv, s);
exit(U_INTERNAL_PROGRAM_ERROR);
}
if(start==0 && end==0x10ffff) {
/* Also set bits for initialValue and errorValue. */
end=UPVEC_MAX_CP;
}
upvec_setValue(pv, start, end, sen->vecWord, uv, sen->vecMask, pErrorCode);
if(U_FAILURE(*pErrorCode)) {
fprintf(stderr, "genprops error: unable to set %s code: %s\n",
sen->propName, u_errorName(*pErrorCode));
exit(*pErrorCode);
}
}
int fastiva_vm_Character_C$__forNameImpl(jstring javaBlockName) {
#if 1
FASTIVA_DBREAK();
return 0;
#else
ScopedUtfChars blockName(env, javaBlockName);
if (blockName.c_str() == NULL) {
return 0;
}
return u_getPropertyValueEnum(UCHAR_BLOCK, blockName.c_str());
#endif
}
// Implemented here because this calls uloc_addLikelySubtags().
U_CAPI UBool U_EXPORT2
uloc_isRightToLeft(const char *locale) {
UErrorCode errorCode = U_ZERO_ERROR;
char script[8];
int32_t scriptLength = uloc_getScript(locale, script, UPRV_LENGTHOF(script), &errorCode);
if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING ||
scriptLength == 0) {
// Fastpath: We know the likely scripts and their writing direction
// for some common languages.
errorCode = U_ZERO_ERROR;
char lang[8];
int32_t langLength = uloc_getLanguage(locale, lang, UPRV_LENGTHOF(lang), &errorCode);
if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING ||
langLength == 0) {
return FALSE;
}
const char* langPtr = uprv_strstr(LANG_DIR_STRING, lang);
if (langPtr != NULL) {
switch (langPtr[langLength]) {
case '-': return FALSE;
case '+': return TRUE;
default: break; // partial match of a longer code
}
}
// Otherwise, find the likely script.
errorCode = U_ZERO_ERROR;
char likely[ULOC_FULLNAME_CAPACITY];
(void)uloc_addLikelySubtags(locale, likely, UPRV_LENGTHOF(likely), &errorCode);
if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING) {
return FALSE;
}
scriptLength = uloc_getScript(likely, script, UPRV_LENGTHOF(script), &errorCode);
if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING ||
scriptLength == 0) {
return FALSE;
}
}
UScriptCode scriptCode = (UScriptCode)u_getPropertyValueEnum(UCHAR_SCRIPT, script);
return uscript_isRightToLeft(scriptCode);
}
int64_t HHVM_STATIC_METHOD(IntlChar, getPropertyValueEnum,
int64_t prop, const String& name) {
return u_getPropertyValueEnum((UProperty)prop, name.c_str());
}
UnicodeSet&
UnicodeSet::applyPropertyAlias(const UnicodeString& prop,
const UnicodeString& value,
UErrorCode& ec) {
if (U_FAILURE(ec) || isFrozen()) return *this;
// prop and value used to be converted to char * using the default
// converter instead of the invariant conversion.
// This should not be necessary because all Unicode property and value
// names use only invariant characters.
// If there are any variant characters, then we won't find them anyway.
// Checking first avoids assertion failures in the conversion.
if( !uprv_isInvariantUString(prop.getBuffer(), prop.length()) ||
!uprv_isInvariantUString(value.getBuffer(), value.length())
) {
FAIL(ec);
}
CharString pname, vname;
pname.appendInvariantChars(prop, ec);
vname.appendInvariantChars(value, ec);
if (U_FAILURE(ec)) return *this;
UProperty p;
int32_t v;
UBool mustNotBeEmpty = FALSE, invert = FALSE;
if (value.length() > 0) {
p = u_getPropertyEnum(pname.data());
if (p == UCHAR_INVALID_CODE) FAIL(ec);
// Treat gc as gcm
if (p == UCHAR_GENERAL_CATEGORY) {
p = UCHAR_GENERAL_CATEGORY_MASK;
}
if ((p >= UCHAR_BINARY_START && p < UCHAR_BINARY_LIMIT) ||
(p >= UCHAR_INT_START && p < UCHAR_INT_LIMIT) ||
(p >= UCHAR_MASK_START && p < UCHAR_MASK_LIMIT)) {
v = u_getPropertyValueEnum(p, vname.data());
if (v == UCHAR_INVALID_CODE) {
// Handle numeric CCC
if (p == UCHAR_CANONICAL_COMBINING_CLASS ||
p == UCHAR_TRAIL_CANONICAL_COMBINING_CLASS ||
p == UCHAR_LEAD_CANONICAL_COMBINING_CLASS) {
char* end;
double value = uprv_strtod(vname.data(), &end);
v = (int32_t) value;
if (v != value || v < 0 || *end != 0) {
// non-integral or negative value, or trailing junk
FAIL(ec);
}
// If the resultant set is empty then the numeric value
// was invalid.
mustNotBeEmpty = TRUE;
} else {
FAIL(ec);
}
}
}
else {
switch (p) {
case UCHAR_NUMERIC_VALUE:
{
char* end;
double value = uprv_strtod(vname.data(), &end);
if (*end != 0) {
FAIL(ec);
}
applyFilter(numericValueFilter, &value, UPROPS_SRC_CHAR, ec);
return *this;
}
case UCHAR_NAME:
{
// Must munge name, since u_charFromName() does not do
// 'loose' matching.
char buf[128]; // it suffices that this be > uprv_getMaxCharNameLength
if (!mungeCharName(buf, vname.data(), sizeof(buf))) FAIL(ec);
UChar32 ch = u_charFromName(U_EXTENDED_CHAR_NAME, buf, &ec);
if (U_SUCCESS(ec)) {
clear();
add(ch);
return *this;
} else {
FAIL(ec);
}
}
case UCHAR_UNICODE_1_NAME:
// ICU 49 deprecates the Unicode_1_Name property APIs.
FAIL(ec);
case UCHAR_AGE:
{
// Must munge name, since u_versionFromString() does not do
// 'loose' matching.
char buf[128];
if (!mungeCharName(buf, vname.data(), sizeof(buf))) FAIL(ec);
UVersionInfo version;
u_versionFromString(version, buf);
applyFilter(versionFilter, &version, UPROPS_SRC_PROPSVEC, ec);
return *this;
}
case UCHAR_SCRIPT_EXTENSIONS:
v = u_getPropertyValueEnum(UCHAR_SCRIPT, vname.data());
if (v == UCHAR_INVALID_CODE) {
FAIL(ec);
}
// fall through to calling applyIntPropertyValue()
break;
default:
// p is a non-binary, non-enumerated property that we
// don't support (yet).
FAIL(ec);
}
}
}
else {
// value is empty. Interpret as General Category, Script, or
// Binary property.
p = UCHAR_GENERAL_CATEGORY_MASK;
v = u_getPropertyValueEnum(p, pname.data());
if (v == UCHAR_INVALID_CODE) {
p = UCHAR_SCRIPT;
v = u_getPropertyValueEnum(p, pname.data());
if (v == UCHAR_INVALID_CODE) {
p = u_getPropertyEnum(pname.data());
if (p >= UCHAR_BINARY_START && p < UCHAR_BINARY_LIMIT) {
v = 1;
} else if (0 == uprv_comparePropertyNames(ANY, pname.data())) {
set(MIN_VALUE, MAX_VALUE);
return *this;
} else if (0 == uprv_comparePropertyNames(ASCII, pname.data())) {
set(0, 0x7F);
return *this;
} else if (0 == uprv_comparePropertyNames(ASSIGNED, pname.data())) {
// [:Assigned:]=[:^Cn:]
p = UCHAR_GENERAL_CATEGORY_MASK;
v = U_GC_CN_MASK;
invert = TRUE;
} else {
FAIL(ec);
}
}
}
}
applyIntPropertyValue(p, v, ec);
if(invert) {
complement();
}
if (U_SUCCESS(ec) && (mustNotBeEmpty && isEmpty())) {
// mustNotBeEmpty is set to true if an empty set indicates
// invalid input.
ec = U_ILLEGAL_ARGUMENT_ERROR;
}
if (isBogus() && U_SUCCESS(ec)) {
// We likely ran out of memory. AHHH!
ec = U_MEMORY_ALLOCATION_ERROR;
}
return *this;
}
/*
* imp: common/propname.cpp
* hdr: common/unicode/uchar.h
* @stable ICU 2.4
*/
U_CAPI int32_t U_EXPORT2
u_getPropertyValueEnum_4_0(UProperty property,
const char* alias)
{
return u_getPropertyValueEnum(property, alias);
}
int32_t
PropertyNames::getPropertyValueEnum(int32_t property, const char *name) const {
return u_getPropertyValueEnum((UProperty)property, name);
}
// Build the Whole Script Confusable data
//
// TODO: Reorganize. Either get rid of the WSConfusableDataBuilder class,
// because everything is local to this one build function anyhow,
// OR
// break this function into more reasonably sized pieces, with
// state in WSConfusableDataBuilder.
//
void buildWSConfusableData(SpoofImpl *spImpl, const char * confusablesWS,
int32_t confusablesWSLen, UParseError *pe, UErrorCode &status)
{
if (U_FAILURE(status)) {
return;
}
URegularExpression *parseRegexp = NULL;
int32_t inputLen = 0;
UChar *input = NULL;
int32_t lineNum = 0;
UVector *scriptSets = NULL;
uint32_t rtScriptSetsCount = 2;
UTrie2 *anyCaseTrie = NULL;
UTrie2 *lowerCaseTrie = NULL;
anyCaseTrie = utrie2_open(0, 0, &status);
lowerCaseTrie = utrie2_open(0, 0, &status);
// The scriptSets vector provides a mapping from TRIE values to the set of scripts.
//
// Reserved TRIE values:
// 0: Code point has no whole script confusables.
// 1: Code point is of script Common or Inherited.
// These code points do not participate in whole script confusable detection.
// (This is logically equivalent to saying that they contain confusables in
// all scripts)
//
// Because Trie values are indexes into the ScriptSets vector, pre-fill
// vector positions 0 and 1 to avoid conflicts with the reserved values.
scriptSets = new UVector(status);
if (scriptSets == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
goto cleanup;
}
scriptSets->addElement((void *)NULL, status);
scriptSets->addElement((void *)NULL, status);
// Convert the user input data from UTF-8 to UChar (UTF-16)
u_strFromUTF8(NULL, 0, &inputLen, confusablesWS, confusablesWSLen, &status);
if (status != U_BUFFER_OVERFLOW_ERROR) {
goto cleanup;
}
status = U_ZERO_ERROR;
input = static_cast<UChar *>(uprv_malloc((inputLen+1) * sizeof(UChar)));
if (input == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
goto cleanup;
}
u_strFromUTF8(input, inputLen+1, NULL, confusablesWS, confusablesWSLen, &status);
parseRegexp = uregex_openC(parseExp, 0, NULL, &status);
// Zap any Byte Order Mark at the start of input. Changing it to a space is benign
// given the syntax of the input.
if (*input == 0xfeff) {
*input = 0x20;
}
// Parse the input, one line per iteration of this loop.
uregex_setText(parseRegexp, input, inputLen, &status);
while (uregex_findNext(parseRegexp, &status)) {
lineNum++;
UChar line[200];
uregex_group(parseRegexp, 0, line, 200, &status);
if (uregex_start(parseRegexp, 1, &status) >= 0) {
// this was a blank or comment line.
continue;
}
if (uregex_start(parseRegexp, 8, &status) >= 0) {
// input file syntax error.
status = U_PARSE_ERROR;
goto cleanup;
}
if (U_FAILURE(status)) {
goto cleanup;
}
// Pick up the start and optional range end code points from the parsed line.
UChar32 startCodePoint = SpoofImpl::ScanHex(
input, uregex_start(parseRegexp, 2, &status), uregex_end(parseRegexp, 2, &status), status);
UChar32 endCodePoint = startCodePoint;
if (uregex_start(parseRegexp, 3, &status) >=0) {
endCodePoint = SpoofImpl::ScanHex(
input, uregex_start(parseRegexp, 3, &status), uregex_end(parseRegexp, 3, &status), status);
}
// Extract the two script names from the source line. We need these in an 8 bit
// default encoding (will be EBCDIC on IBM mainframes) in order to pass them on
// to the ICU u_getPropertyValueEnum() function. Ugh.
char srcScriptName[20];
char targScriptName[20];
extractGroup(parseRegexp, 4, srcScriptName, sizeof(srcScriptName), status);
extractGroup(parseRegexp, 5, targScriptName, sizeof(targScriptName), status);
UScriptCode srcScript =
static_cast<UScriptCode>(u_getPropertyValueEnum(UCHAR_SCRIPT, srcScriptName));
UScriptCode targScript =
static_cast<UScriptCode>(u_getPropertyValueEnum(UCHAR_SCRIPT, targScriptName));
if (U_FAILURE(status)) {
goto cleanup;
}
if (srcScript == USCRIPT_INVALID_CODE || targScript == USCRIPT_INVALID_CODE) {
status = U_INVALID_FORMAT_ERROR;
goto cleanup;
}
// select the table - (A) any case or (L) lower case only
UTrie2 *table = anyCaseTrie;
if (uregex_start(parseRegexp, 7, &status) >= 0) {
table = lowerCaseTrie;
}
// Build the set of scripts containing confusable characters for
// the code point(s) specified in this input line.
// Sanity check that the script of the source code point is the same
// as the source script indicated in the input file. Failure of this check is
// an error in the input file.
// Include the source script in the set (needed for Mixed Script Confusable detection).
//
UChar32 cp;
for (cp=startCodePoint; cp<=endCodePoint; cp++) {
int32_t setIndex = utrie2_get32(table, cp);
BuilderScriptSet *bsset = NULL;
if (setIndex > 0) {
U_ASSERT(setIndex < scriptSets->size());
bsset = static_cast<BuilderScriptSet *>(scriptSets->elementAt(setIndex));
} else {
bsset = new BuilderScriptSet();
if (bsset == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
goto cleanup;
}
bsset->codePoint = cp;
bsset->trie = table;
bsset->sset = new ScriptSet();
setIndex = scriptSets->size();
bsset->index = setIndex;
bsset->rindex = 0;
if (bsset->sset == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
goto cleanup;
}
scriptSets->addElement(bsset, status);
utrie2_set32(table, cp, setIndex, &status);
}
bsset->sset->Union(targScript);
bsset->sset->Union(srcScript);
if (U_FAILURE(status)) {
goto cleanup;
}
UScriptCode cpScript = uscript_getScript(cp, &status);
if (cpScript != srcScript) {
status = U_INVALID_FORMAT_ERROR;
goto cleanup;
}
}
}
// Eliminate duplicate script sets. At this point we have a separate
// script set for every code point that had data in the input file.
//
// We eliminate underlying ScriptSet objects, not the BuildScriptSets that wrap them
//
// printf("Number of scriptSets: %d\n", scriptSets->size());
{
int32_t duplicateCount = 0;
rtScriptSetsCount = 2;
for (int32_t outeri=2; outeri<scriptSets->size(); outeri++) {
BuilderScriptSet *outerSet = static_cast<BuilderScriptSet *>(scriptSets->elementAt(outeri));
if (outerSet->index != static_cast<uint32_t>(outeri)) {
// This set was already identified as a duplicate.
// It will not be allocated a position in the runtime array of ScriptSets.
continue;
}
outerSet->rindex = rtScriptSetsCount++;
for (int32_t inneri=outeri+1; inneri<scriptSets->size(); inneri++) {
BuilderScriptSet *innerSet = static_cast<BuilderScriptSet *>(scriptSets->elementAt(inneri));
if (*(outerSet->sset) == *(innerSet->sset) && outerSet->sset != innerSet->sset) {
delete innerSet->sset;
innerSet->scriptSetOwned = FALSE;
innerSet->sset = outerSet->sset;
innerSet->index = outeri;
innerSet->rindex = outerSet->rindex;
duplicateCount++;
}
// But this doesn't get all. We need to fix the TRIE.
}
}
// printf("Number of distinct script sets: %d\n", rtScriptSetsCount);
}
// Update the Trie values to be reflect the run time script indexes (after duplicate merging).
// (Trie Values 0 and 1 are reserved, and the corresponding slots in scriptSets
// are unused, which is why the loop index starts at 2.)
{
for (int32_t i=2; i<scriptSets->size(); i++) {
BuilderScriptSet *bSet = static_cast<BuilderScriptSet *>(scriptSets->elementAt(i));
if (bSet->rindex != (uint32_t)i) {
utrie2_set32(bSet->trie, bSet->codePoint, bSet->rindex, &status);
}
}
}
// For code points with script==Common or script==Inherited,
// Set the reserved value of 1 into both Tries. These characters do not participate
// in Whole Script Confusable detection; this reserved value is the means
// by which they are detected.
{
UnicodeSet ignoreSet;
ignoreSet.applyIntPropertyValue(UCHAR_SCRIPT, USCRIPT_COMMON, status);
UnicodeSet inheritedSet;
inheritedSet.applyIntPropertyValue(UCHAR_SCRIPT, USCRIPT_INHERITED, status);
ignoreSet.addAll(inheritedSet);
for (int32_t rn=0; rn<ignoreSet.getRangeCount(); rn++) {
UChar32 rangeStart = ignoreSet.getRangeStart(rn);
UChar32 rangeEnd = ignoreSet.getRangeEnd(rn);
utrie2_setRange32(anyCaseTrie, rangeStart, rangeEnd, 1, TRUE, &status);
utrie2_setRange32(lowerCaseTrie, rangeStart, rangeEnd, 1, TRUE, &status);
}
}
// Serialize the data to the Spoof Detector
{
utrie2_freeze(anyCaseTrie, UTRIE2_16_VALUE_BITS, &status);
int32_t size = utrie2_serialize(anyCaseTrie, NULL, 0, &status);
// printf("Any case Trie size: %d\n", size);
if (status != U_BUFFER_OVERFLOW_ERROR) {
goto cleanup;
}
status = U_ZERO_ERROR;
spImpl->fSpoofData->fRawData->fAnyCaseTrie = spImpl->fSpoofData->fMemLimit;
spImpl->fSpoofData->fRawData->fAnyCaseTrieLength = size;
spImpl->fSpoofData->fAnyCaseTrie = anyCaseTrie;
void *where = spImpl->fSpoofData->reserveSpace(size, status);
utrie2_serialize(anyCaseTrie, where, size, &status);
utrie2_freeze(lowerCaseTrie, UTRIE2_16_VALUE_BITS, &status);
size = utrie2_serialize(lowerCaseTrie, NULL, 0, &status);
// printf("Lower case Trie size: %d\n", size);
if (status != U_BUFFER_OVERFLOW_ERROR) {
goto cleanup;
}
status = U_ZERO_ERROR;
spImpl->fSpoofData->fRawData->fLowerCaseTrie = spImpl->fSpoofData->fMemLimit;
spImpl->fSpoofData->fRawData->fLowerCaseTrieLength = size;
spImpl->fSpoofData->fLowerCaseTrie = lowerCaseTrie;
where = spImpl->fSpoofData->reserveSpace(size, status);
utrie2_serialize(lowerCaseTrie, where, size, &status);
spImpl->fSpoofData->fRawData->fScriptSets = spImpl->fSpoofData->fMemLimit;
spImpl->fSpoofData->fRawData->fScriptSetsLength = rtScriptSetsCount;
ScriptSet *rtScriptSets = static_cast<ScriptSet *>
(spImpl->fSpoofData->reserveSpace(rtScriptSetsCount * sizeof(ScriptSet), status));
uint32_t rindex = 2;
for (int32_t i=2; i<scriptSets->size(); i++) {
BuilderScriptSet *bSet = static_cast<BuilderScriptSet *>(scriptSets->elementAt(i));
if (bSet->rindex < rindex) {
// We have already copied this script set to the serialized data.
continue;
}
U_ASSERT(rindex == bSet->rindex);
rtScriptSets[rindex] = *bSet->sset; // Assignment of a ScriptSet just copies the bits.
rindex++;
}
}
// Open new utrie2s from the serialized data. We don't want to keep the ones
// we just built because we would then have two copies of the data, one internal to
// the utries that we have already constructed, and one in the serialized data area.
// An alternative would be to not pre-serialize the Trie data, but that makes the
// spoof detector data different, depending on how the detector was constructed.
// It's simpler to keep the data always the same.
spImpl->fSpoofData->fAnyCaseTrie = utrie2_openFromSerialized(
UTRIE2_16_VALUE_BITS,
(const char *)spImpl->fSpoofData->fRawData + spImpl->fSpoofData->fRawData->fAnyCaseTrie,
spImpl->fSpoofData->fRawData->fAnyCaseTrieLength,
NULL,
&status);
spImpl->fSpoofData->fLowerCaseTrie = utrie2_openFromSerialized(
UTRIE2_16_VALUE_BITS,
(const char *)spImpl->fSpoofData->fRawData + spImpl->fSpoofData->fRawData->fLowerCaseTrie,
spImpl->fSpoofData->fRawData->fAnyCaseTrieLength,
NULL,
&status);
cleanup:
if (U_FAILURE(status)) {
pe->line = lineNum;
}
uregex_close(parseRegexp);
uprv_free(input);
int32_t i;
for (i=0; i<scriptSets->size(); i++) {
BuilderScriptSet *bsset = static_cast<BuilderScriptSet *>(scriptSets->elementAt(i));
delete bsset;
}
delete scriptSets;
utrie2_close(anyCaseTrie);
utrie2_close(lowerCaseTrie);
return;
}
