static void TestBreakIteratorRefresh(void) {
/*
* RefreshInput changes out the input of a Break Iterator without
* changing anything else in the iterator's state. Used with Java JNI,
* when Java moves the underlying string storage. This test
* runs a ubrk_next() repeatedly, moving the text in the middle of the sequence.
* The right set of boundaries should still be found.
*/
UChar testStr[] = {0x20, 0x41, 0x20, 0x42, 0x20, 0x43, 0x20, 0x44, 0x0}; /* = " A B C D" */
UChar movedStr[] = {0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0};
UErrorCode status = U_ZERO_ERROR;
UBreakIterator *bi;
UText ut1 = UTEXT_INITIALIZER;
UText ut2 = UTEXT_INITIALIZER;
bi = ubrk_open(UBRK_LINE, "en_US", NULL, 0, &status);
TEST_ASSERT_SUCCESS(status);
if (U_FAILURE(status)) {
return;
}
utext_openUChars(&ut1, testStr, -1, &status);
TEST_ASSERT_SUCCESS(status);
ubrk_setUText(bi, &ut1, &status);
TEST_ASSERT_SUCCESS(status);
if (U_SUCCESS(status)) {
/* Line boundaries will occur before each letter in the original string */
TEST_ASSERT(1 == ubrk_next(bi));
TEST_ASSERT(3 == ubrk_next(bi));
/* Move the string, kill the original string. */
u_strcpy(movedStr, testStr);
u_memset(testStr, 0x20, u_strlen(testStr));
utext_openUChars(&ut2, movedStr, -1, &status);
TEST_ASSERT_SUCCESS(status);
ubrk_refreshUText(bi, &ut2, &status);
TEST_ASSERT_SUCCESS(status);
/* Find the following matches, now working in the moved string. */
TEST_ASSERT(5 == ubrk_next(bi));
TEST_ASSERT(7 == ubrk_next(bi));
TEST_ASSERT(8 == ubrk_next(bi));
TEST_ASSERT(UBRK_DONE == ubrk_next(bi));
TEST_ASSERT_SUCCESS(status);
utext_close(&ut1);
utext_close(&ut2);
}
ubrk_close(bi);
}
void RBBIAPITest::TestRefreshInputText() {
/*
* RefreshInput changes out the input of a Break Iterator without
* changing anything else in the iterator's state. Used with Java JNI,
* when Java moves the underlying string storage. This test
* runs BreakIterator::next() repeatedly, moving the text in the middle of the sequence.
* The right set of boundaries should still be found.
*/
UChar testStr[] = {0x20, 0x41, 0x20, 0x42, 0x20, 0x43, 0x20, 0x44, 0x0}; /* = " A B C D" */
UChar movedStr[] = {0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0x20, 0};
UErrorCode status = U_ZERO_ERROR;
UText ut1 = UTEXT_INITIALIZER;
UText ut2 = UTEXT_INITIALIZER;
RuleBasedBreakIterator *bi = (RuleBasedBreakIterator *)BreakIterator::createLineInstance(Locale::getEnglish(), status);
TEST_ASSERT_SUCCESS(status);
utext_openUChars(&ut1, testStr, -1, &status);
TEST_ASSERT_SUCCESS(status);
if (U_SUCCESS(status)) {
bi->setText(&ut1, status);
TEST_ASSERT_SUCCESS(status);
/* Line boundaries will occur before each letter in the original string */
TEST_ASSERT(1 == bi->next());
TEST_ASSERT(3 == bi->next());
/* Move the string, kill the original string. */
u_strcpy(movedStr, testStr);
u_memset(testStr, 0x20, u_strlen(testStr));
utext_openUChars(&ut2, movedStr, -1, &status);
TEST_ASSERT_SUCCESS(status);
RuleBasedBreakIterator *returnedBI = &bi->refreshInputText(&ut2, status);
TEST_ASSERT_SUCCESS(status);
TEST_ASSERT(bi == returnedBI);
/* Find the following matches, now working in the moved string. */
TEST_ASSERT(5 == bi->next());
TEST_ASSERT(7 == bi->next());
TEST_ASSERT(8 == bi->next());
TEST_ASSERT(UBRK_DONE == bi->next());
utext_close(&ut1);
utext_close(&ut2);
}
delete bi;
}
ScopedBreakIterator(JNIEnv* env, BreakIterator* breakIterator, jcharArray inputText,
jint length) : mBreakIterator(breakIterator), mChars(env, inputText) {
UErrorCode status = U_ZERO_ERROR;
mUText = utext_openUChars(NULL, mChars.get(), length, &status);
if (mUText == NULL) {
return;
}
mBreakIterator->setText(mUText, status);
}
U_CAPI void U_EXPORT2
ubrk_setText(UBreakIterator* bi,
const UChar* text,
int32_t textLength,
UErrorCode* status)
{
UText ut = UTEXT_INITIALIZER;
utext_openUChars(&ut, text, textLength, status);
((BreakIterator*)bi)->setText(&ut, *status);
// A stack allocated UText wrapping a UChar * string
// can be dumped without explicitly closing it.
}
virtual void call(UErrorCode *pErrorCode) {
UText text=UTEXT_INITIALIZER;
int32_t lengths[20];
const ULine *lines=perf.getCachedLines();
int32_t numLines=perf.getNumLines();
for(int32_t i=0; i<numLines; ++i) {
// Skip comment lines (start with a character below 'A').
if(lines[i].name[0]<0x41) {
continue;
}
utext_openUChars(&text, lines[i].name, lines[i].len, pErrorCode);
int32_t count=0;
ucharsTrieMatches(*trie, &text, lines[i].len,
lengths, count, UPRV_LENGTHOF(lengths));
if(count==0 || lengths[count-1]!=lines[i].len) {
fprintf(stderr, "word %ld (0-based) not found\n", (long)i);
}
}
}
MatcherAccessor(JNIEnv* env, jint addr, jstring javaInput, bool reset) {
init(env, addr);
mJavaInput = javaInput;
mChars = env->GetStringChars(mJavaInput, NULL);
if (mChars == NULL) {
return;
}
mUText = utext_openUChars(NULL, mChars, env->GetStringLength(mJavaInput), &mStatus);
if (mUText == NULL) {
return;
}
if (reset) {
mMatcher->reset(mUText);
} else {
mMatcher->refreshInputText(mUText, mStatus);
}
}
virtual void SetString(const char *s)
{
const char *string_base = s;
/* Unfortunately current ICU versions only provide rudimentary support
* for word break iterators (especially for CJK languages) in combination
* with UTF-8 input. As a work around we have to convert the input to
* UTF-16 and create a mapping back to UTF-8 character indices. */
this->utf16_str.Clear();
this->utf16_to_utf8.Clear();
while (*s != '\0') {
size_t idx = s - string_base;
WChar c = Utf8Consume(&s);
if (c < 0x10000) {
*this->utf16_str.Append() = (UChar)c;
} else {
/* Make a surrogate pair. */
*this->utf16_str.Append() = (UChar)(0xD800 + ((c - 0x10000) >> 10));
*this->utf16_str.Append() = (UChar)(0xDC00 + ((c - 0x10000) & 0x3FF));
*this->utf16_to_utf8.Append() = idx;
}
*this->utf16_to_utf8.Append() = idx;
}
*this->utf16_str.Append() = '\0';
*this->utf16_to_utf8.Append() = s - string_base;
UText text = UTEXT_INITIALIZER;
UErrorCode status = U_ZERO_ERROR;
utext_openUChars(&text, this->utf16_str.Begin(), this->utf16_str.Length() - 1, &status);
this->char_itr->setText(&text, status);
this->word_itr->setText(&text, status);
this->char_itr->first();
this->word_itr->first();
}
BreakIteratorAccessor(JNIEnv* env, jlong address, jstring javaInput, bool reset) {
init(env, address);
mJavaInput = javaInput;
if (mJavaInput == NULL) {
return;
}
mChars = env->GetStringChars(mJavaInput, NULL);
if (mChars == NULL) {
return;
}
mUText = utext_openUChars(NULL, mChars, env->GetStringLength(mJavaInput), &mStatus);
if (mUText == NULL) {
return;
}
if (reset) {
mBreakIterator->setText(mUText, mStatus);
} else {
mBreakIterator->refreshInputText(mUText, mStatus);
}
}
void BreakIterator::setText(const UChar* text, int32_t textLength, UErrorCode &status)
{
utext_openUChars(fText, text, textLength, &status);
this->first();
}
static void TestBreakIteratorCAPI()
{
UErrorCode status = U_ZERO_ERROR;
UBreakIterator *word, *sentence, *line, *character, *b, *bogus;
int32_t start,pos,end,to;
int32_t i;
int32_t count = 0;
UChar text[50];
/* Note: the adjacent "" are concatenating strings, not adding a \" to the
string, which is probably what whoever wrote this intended. Don't fix,
because it would throw off the hard coded break positions in the following
tests. */
u_uastrcpy(text, "He's from Africa. ""Mr. Livingston, I presume?"" Yeah");
/*test ubrk_open()*/
log_verbose("\nTesting BreakIterator open functions\n");
/* Use french for fun */
word = ubrk_open(UBRK_WORD, "en_US", text, u_strlen(text), &status);
if(status == U_FILE_ACCESS_ERROR) {
log_data_err("Check your data - it doesn't seem to be around\n");
return;
} else if(U_FAILURE(status)){
log_err_status(status, "FAIL: Error in ubrk_open() for word breakiterator: %s\n", myErrorName(status));
}
else{
log_verbose("PASS: Successfully opened word breakiterator\n");
}
sentence = ubrk_open(UBRK_SENTENCE, "en_US", text, u_strlen(text), &status);
if(U_FAILURE(status)){
log_err_status(status, "FAIL: Error in ubrk_open() for sentence breakiterator: %s\n", myErrorName(status));
return;
}
else{
log_verbose("PASS: Successfully opened sentence breakiterator\n");
}
line = ubrk_open(UBRK_LINE, "en_US", text, u_strlen(text), &status);
if(U_FAILURE(status)){
log_err("FAIL: Error in ubrk_open() for line breakiterator: %s\n", myErrorName(status));
return;
}
else{
log_verbose("PASS: Successfully opened line breakiterator\n");
}
character = ubrk_open(UBRK_CHARACTER, "en_US", text, u_strlen(text), &status);
if(U_FAILURE(status)){
log_err("FAIL: Error in ubrk_open() for character breakiterator: %s\n", myErrorName(status));
return;
}
else{
log_verbose("PASS: Successfully opened character breakiterator\n");
}
/*trying to open an illegal iterator*/
bogus = ubrk_open((UBreakIteratorType)5, "en_US", text, u_strlen(text), &status);
if(U_SUCCESS(status)){
log_err("FAIL: Error in ubrk_open() for BOGUS breakiterator. Expected U_ILLEGAL_ARGUMENT_ERROR\n");
}
if(U_FAILURE(status)){
if(status != U_ILLEGAL_ARGUMENT_ERROR){
log_err("FAIL: Error in ubrk_open() for BOGUS breakiterator. Expected U_ILLEGAL_ARGUMENT_ERROR\n Got %s\n", myErrorName(status));
}
}
status=U_ZERO_ERROR;
/* ======= Test ubrk_countAvialable() and ubrk_getAvialable() */
log_verbose("\nTesting ubrk_countAvailable() and ubrk_getAvailable()\n");
count=ubrk_countAvailable();
/* use something sensible w/o hardcoding the count */
if(count < 0){
log_err("FAIL: Error in ubrk_countAvialable() returned %d\n", count);
}
else{
log_verbose("PASS: ubrk_countAvialable() successful returned %d\n", count);
}
for(i=0;i<count;i++)
{
log_verbose("%s\n", ubrk_getAvailable(i));
if (ubrk_getAvailable(i) == 0)
log_err("No locale for which breakiterator is applicable\n");
else
log_verbose("A locale %s for which breakiterator is applicable\n",ubrk_getAvailable(i));
}
/*========Test ubrk_first(), ubrk_last()...... and other functions*/
log_verbose("\nTesting the functions for word\n");
start = ubrk_first(word);
if(start!=0)
log_err("error ubrk_start(word) did not return 0\n");
log_verbose("first (word = %d\n", (int32_t)start);
pos=ubrk_next(word);
if(pos!=4)
log_err("error ubrk_next(word) did not return 4\n");
log_verbose("next (word = %d\n", (int32_t)pos);
pos=ubrk_following(word, 4);
if(pos!=5)
log_err("error ubrl_following(word,4) did not return 6\n");
log_verbose("next (word = %d\n", (int32_t)pos);
end=ubrk_last(word);
if(end!=49)
log_err("error ubrk_last(word) did not return 49\n");
log_verbose("last (word = %d\n", (int32_t)end);
pos=ubrk_previous(word);
log_verbose("%d %d\n", end, pos);
pos=ubrk_previous(word);
log_verbose("%d \n", pos);
if (ubrk_isBoundary(word, 2) != FALSE) {
log_err("error ubrk_isBoundary(word, 2) did not return FALSE\n");
}
pos=ubrk_current(word);
if (pos != 4) {
log_err("error ubrk_current() != 4 after ubrk_isBoundary(word, 2)\n");
}
if (ubrk_isBoundary(word, 4) != TRUE) {
log_err("error ubrk_isBoundary(word, 4) did not return TRUE\n");
}
log_verbose("\nTesting the functions for character\n");
ubrk_first(character);
pos = ubrk_following(character, 5);
if(pos!=6)
log_err("error ubrk_following(character,5) did not return 6\n");
log_verbose("Following (character,5) = %d\n", (int32_t)pos);
pos=ubrk_following(character, 18);
if(pos!=19)
log_err("error ubrk_following(character,18) did not return 19\n");
log_verbose("Followingcharacter,18) = %d\n", (int32_t)pos);
pos=ubrk_preceding(character, 22);
if(pos!=21)
log_err("error ubrk_preceding(character,22) did not return 21\n");
log_verbose("preceding(character,22) = %d\n", (int32_t)pos);
log_verbose("\nTesting the functions for line\n");
pos=ubrk_first(line);
if(pos != 0)
log_err("error ubrk_first(line) returned %d, expected 0\n", (int32_t)pos);
pos = ubrk_next(line);
pos=ubrk_following(line, 18);
if(pos!=22)
log_err("error ubrk_following(line) did not return 22\n");
log_verbose("following (line) = %d\n", (int32_t)pos);
log_verbose("\nTesting the functions for sentence\n");
ubrk_first(sentence);
pos = ubrk_current(sentence);
log_verbose("Current(sentence) = %d\n", (int32_t)pos);
pos = ubrk_last(sentence);
if(pos!=49)
log_err("error ubrk_last for sentence did not return 49\n");
log_verbose("Last (sentence) = %d\n", (int32_t)pos);
ubrk_first(sentence);
to = ubrk_following( sentence, 0 );
if (to == 0) log_err("ubrk_following returned 0\n");
to = ubrk_preceding( sentence, to );
if (to != 0) log_err("ubrk_preceding didn't return 0\n");
if (ubrk_first(sentence)!=ubrk_current(sentence)) {
log_err("error in ubrk_first() or ubrk_current()\n");
}
/*---- */
/*Testing ubrk_open and ubrk_close()*/
log_verbose("\nTesting open and close for us locale\n");
b = ubrk_open(UBRK_WORD, "fr_FR", text, u_strlen(text), &status);
if (U_FAILURE(status)) {
log_err("ubrk_open for word returned NULL: %s\n", myErrorName(status));
}
ubrk_close(b);
/* Test setText and setUText */
{
UChar s1[] = {0x41, 0x42, 0x20, 0};
UChar s2[] = {0x41, 0x42, 0x43, 0x44, 0x45, 0};
UText *ut = NULL;
UBreakIterator *bb;
int j;
log_verbose("\nTesting ubrk_setText() and ubrk_setUText()\n");
status = U_ZERO_ERROR;
bb = ubrk_open(UBRK_WORD, "en_US", NULL, 0, &status);
TEST_ASSERT_SUCCESS(status);
ubrk_setText(bb, s1, -1, &status);
TEST_ASSERT_SUCCESS(status);
ubrk_first(bb);
j = ubrk_next(bb);
TEST_ASSERT(j == 2);
ut = utext_openUChars(ut, s2, -1, &status);
ubrk_setUText(bb, ut, &status);
TEST_ASSERT_SUCCESS(status);
j = ubrk_next(bb);
TEST_ASSERT(j == 5);
ubrk_close(bb);
utext_close(ut);
}
ubrk_close(word);
ubrk_close(sentence);
ubrk_close(line);
ubrk_close(character);
}
static void TestAPI(void) {
UErrorCode status = U_ZERO_ERROR;
UBool gFailed = FALSE;
(void)gFailed; /* Suppress set but not used warning. */
/* Open */
{
UText utLoc = UTEXT_INITIALIZER;
const char * cString = "\x61\x62\x63\x64";
UChar uString[] = {0x41, 0x42, 0x43, 0};
UText *uta;
UText *utb;
UChar c;
uta = utext_openUChars(NULL, uString, -1, &status);
TEST_SUCCESS(status);
c = utext_next32(uta);
TEST_ASSERT(c == 0x41);
utb = utext_close(uta);
TEST_ASSERT(utb == NULL);
uta = utext_openUTF8(&utLoc, cString, -1, &status);
TEST_SUCCESS(status);
TEST_ASSERT(uta == &utLoc);
uta = utext_close(&utLoc);
TEST_ASSERT(uta == &utLoc);
}
/* utext_clone() */
{
UChar uString[] = {0x41, 0x42, 0x43, 0};
int64_t len;
UText *uta;
UText *utb;
status = U_ZERO_ERROR;
uta = utext_openUChars(NULL, uString, -1, &status);
TEST_SUCCESS(status);
utb = utext_clone(NULL, uta, FALSE, FALSE, &status);
TEST_SUCCESS(status);
TEST_ASSERT(utb != NULL);
TEST_ASSERT(utb != uta);
len = utext_nativeLength(uta);
TEST_ASSERT(len == u_strlen(uString));
utext_close(uta);
utext_close(utb);
}
/* basic access functions */
{
UChar uString[] = {0x41, 0x42, 0x43, 0};
UText *uta;
UChar32 c;
int64_t len;
UBool b;
int64_t i;
status = U_ZERO_ERROR;
uta = utext_openUChars(NULL, uString, -1, &status);
TEST_ASSERT(uta!=NULL);
TEST_SUCCESS(status);
b = utext_isLengthExpensive(uta);
TEST_ASSERT(b==TRUE);
len = utext_nativeLength(uta);
TEST_ASSERT(len == u_strlen(uString));
b = utext_isLengthExpensive(uta);
TEST_ASSERT(b==FALSE);
c = utext_char32At(uta, 0);
TEST_ASSERT(c==uString[0]);
c = utext_current32(uta);
TEST_ASSERT(c==uString[0]);
c = utext_next32(uta);
TEST_ASSERT(c==uString[0]);
c = utext_current32(uta);
TEST_ASSERT(c==uString[1]);
c = utext_previous32(uta);
TEST_ASSERT(c==uString[0]);
c = utext_current32(uta);
TEST_ASSERT(c==uString[0]);
c = utext_next32From(uta, 1);
TEST_ASSERT(c==uString[1]);
c = utext_next32From(uta, u_strlen(uString));
TEST_ASSERT(c==U_SENTINEL);
c = utext_previous32From(uta, 2);
TEST_ASSERT(c==uString[1]);
i = utext_getNativeIndex(uta);
TEST_ASSERT(i == 1);
utext_setNativeIndex(uta, 0);
b = utext_moveIndex32(uta, 1);
TEST_ASSERT(b==TRUE);
i = utext_getNativeIndex(uta);
TEST_ASSERT(i==1);
b = utext_moveIndex32(uta, u_strlen(uString)-1);
TEST_ASSERT(b==TRUE);
i = utext_getNativeIndex(uta);
TEST_ASSERT(i==u_strlen(uString));
b = utext_moveIndex32(uta, 1);
TEST_ASSERT(b==FALSE);
i = utext_getNativeIndex(uta);
TEST_ASSERT(i==u_strlen(uString));
utext_setNativeIndex(uta, 0);
c = UTEXT_NEXT32(uta);
TEST_ASSERT(c==uString[0]);
c = utext_current32(uta);
TEST_ASSERT(c==uString[1]);
c = UTEXT_PREVIOUS32(uta);
TEST_ASSERT(c==uString[0]);
c = UTEXT_PREVIOUS32(uta);
TEST_ASSERT(c==U_SENTINEL);
utext_close(uta);
}
{
/*
* UText opened on a NULL string with zero length
*/
UText *uta;
UChar32 c;
status = U_ZERO_ERROR;
uta = utext_openUChars(NULL, NULL, 0, &status);
TEST_SUCCESS(status);
c = UTEXT_NEXT32(uta);
TEST_ASSERT(c == U_SENTINEL);
utext_close(uta);
uta = utext_openUTF8(NULL, NULL, 0, &status);
TEST_SUCCESS(status);
c = UTEXT_NEXT32(uta);
TEST_ASSERT(c == U_SENTINEL);
utext_close(uta);
}
{
/*
* extract
*/
UText *uta;
UChar uString[] = {0x41, 0x42, 0x43, 0};
UChar buf[100];
int32_t i;
/* Test pinning of input bounds */
UChar uString2[] = {0x41, 0x42, 0x43, 0x44, 0x45,
0x46, 0x47, 0x48, 0x49, 0x4A, 0};
UChar * uString2Ptr = uString2 + 5;
status = U_ZERO_ERROR;
uta = utext_openUChars(NULL, uString, -1, &status);
TEST_SUCCESS(status);
status = U_ZERO_ERROR;
i = utext_extract(uta, 0, 100, NULL, 0, &status);
TEST_ASSERT(status==U_BUFFER_OVERFLOW_ERROR);
TEST_ASSERT(i == u_strlen(uString));
status = U_ZERO_ERROR;
memset(buf, 0, sizeof(buf));
i = utext_extract(uta, 0, 100, buf, 100, &status);
TEST_SUCCESS(status);
TEST_ASSERT(i == u_strlen(uString));
i = u_strcmp(uString, buf);
TEST_ASSERT(i == 0);
utext_close(uta);
/* Test pinning of input bounds */
status = U_ZERO_ERROR;
uta = utext_openUChars(NULL, uString2Ptr, -1, &status);
TEST_SUCCESS(status);
status = U_ZERO_ERROR;
memset(buf, 0, sizeof(buf));
i = utext_extract(uta, -3, 20, buf, 100, &status);
TEST_SUCCESS(status);
TEST_ASSERT(i == u_strlen(uString2Ptr));
i = u_strcmp(uString2Ptr, buf);
TEST_ASSERT(i == 0);
utext_close(uta);
}
{
/*
* Copy, Replace, isWritable
* Can't create an editable UText from plain C, so all we
* can easily do is check that errors returned.
*/
UText *uta;
UChar uString[] = {0x41, 0x42, 0x43, 0};
UBool b;
status = U_ZERO_ERROR;
uta = utext_openUChars(NULL, uString, -1, &status);
TEST_SUCCESS(status);
b = utext_isWritable(uta);
TEST_ASSERT(b == FALSE);
b = utext_hasMetaData(uta);
TEST_ASSERT(b == FALSE);
utext_replace(uta,
0, 1, /* start, limit */
uString, -1, /* replacement, replacement length */
&status);
TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
utext_copy(uta,
0, 1, /* start, limit */
2, /* destination index */
FALSE, /* move flag */
&status);
TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
utext_close(uta);
}
}
RegexStaticSets::RegexStaticSets(UErrorCode *status)
:
fUnescapeCharSet(UnicodeString(TRUE, gUnescapeCharPattern, -1), *status),
fRuleDigitsAlias(NULL),
fEmptyText(NULL)
{
// First zero out everything
int i;
for (i=0; i<URX_LAST_SET; i++) {
fPropSets[i] = NULL;
}
// Then init the sets to their correct values.
fPropSets[URX_ISWORD_SET] = new UnicodeSet(UnicodeString(TRUE, gIsWordPattern, -1), *status);
fPropSets[URX_ISSPACE_SET] = new UnicodeSet(UnicodeString(TRUE, gIsSpacePattern, -1), *status);
fPropSets[URX_GC_EXTEND] = new UnicodeSet(UnicodeString(TRUE, gGC_ExtendPattern, -1), *status);
fPropSets[URX_GC_CONTROL] = new UnicodeSet(UnicodeString(TRUE, gGC_ControlPattern, -1), *status);
fPropSets[URX_GC_L] = new UnicodeSet(UnicodeString(TRUE, gGC_LPattern, -1), *status);
fPropSets[URX_GC_V] = new UnicodeSet(UnicodeString(TRUE, gGC_VPattern, -1), *status);
fPropSets[URX_GC_T] = new UnicodeSet(UnicodeString(TRUE, gGC_TPattern, -1), *status);
fPropSets[URX_GC_LV] = new UnicodeSet(UnicodeString(TRUE, gGC_LVPattern, -1), *status);
fPropSets[URX_GC_LVT] = new UnicodeSet(UnicodeString(TRUE, gGC_LVTPattern, -1), *status);
// Check for null pointers
if (fPropSets[URX_ISWORD_SET] == NULL || fPropSets[URX_ISSPACE_SET] == NULL || fPropSets[URX_GC_EXTEND] == NULL ||
fPropSets[URX_GC_CONTROL] == NULL || fPropSets[URX_GC_L] == NULL || fPropSets[URX_GC_V] == NULL ||
fPropSets[URX_GC_T] == NULL || fPropSets[URX_GC_LV] == NULL || fPropSets[URX_GC_LVT] == NULL) {
goto ExitConstrDeleteAll;
}
if (U_FAILURE(*status)) {
// Bail out if we were unable to create the above sets.
// The rest of the initialization needs them, so we cannot proceed.
return;
}
//
// The following sets are dynamically constructed, because their
// initialization strings would be unreasonable.
//
//
// "Normal" is the set of characters that don't need special handling
// when finding grapheme cluster boundaries.
//
fPropSets[URX_GC_NORMAL] = new UnicodeSet(0, UnicodeSet::MAX_VALUE);
// Null pointer check
if (fPropSets[URX_GC_NORMAL] == NULL) {
goto ExitConstrDeleteAll;
}
fPropSets[URX_GC_NORMAL]->remove(0xac00, 0xd7a4);
fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_CONTROL]);
fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_L]);
fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_V]);
fPropSets[URX_GC_NORMAL]->removeAll(*fPropSets[URX_GC_T]);
// Initialize the 8-bit fast bit sets from the parallel full
// UnicodeSets.
for (i=0; i<URX_LAST_SET; i++) {
if (fPropSets[i]) {
fPropSets[i]->compact();
fPropSets8[i].init(fPropSets[i]);
}
}
// Sets used while parsing rules, but not referenced from the parse state table
fRuleSets[kRuleSet_rule_char-128] = UnicodeSet(UnicodeString(TRUE, gRuleSet_rule_char_pattern, -1), *status);
fRuleSets[kRuleSet_digit_char-128].add((UChar)0x30, (UChar)0x39); // [0-9]
fRuleSets[kRuleSet_ascii_letter-128].add((UChar)0x41, (UChar)0x5A); // [A-Z]
fRuleSets[kRuleSet_ascii_letter-128].add((UChar)0x61, (UChar)0x7A); // [a-z]
fRuleDigitsAlias = &fRuleSets[kRuleSet_digit_char-128];
for (i=0; i<UPRV_LENGTHOF(fRuleSets); i++) {
fRuleSets[i].compact();
}
// Finally, initialize an empty string for utility purposes
fEmptyText = utext_openUChars(NULL, NULL, 0, status);
if (U_SUCCESS(*status)) {
return;
}
ExitConstrDeleteAll: // Remove fPropSets and fRuleSets and return error
for (i=0; i<URX_LAST_SET; i++) {
delete fPropSets[i];
fPropSets[i] = NULL;
}
if (U_SUCCESS(*status)) {
*status = U_MEMORY_ALLOCATION_ERROR;
}
}