static UBool *
getResultsManually(const char** encodings, int32_t num_encodings,
const char *utf8, int32_t length,
const USet* excludedCodePoints, const UConverterUnicodeSet whichSet) {
UBool* resultsManually;
int32_t i;
resultsManually = (UBool*) uprv_malloc(gCountAvailable);
uprv_memset(resultsManually, 0, gCountAvailable);
for(i = 0 ; i < num_encodings ; i++) {
UErrorCode status = U_ZERO_ERROR;
/* get unicode set for that converter */
USet* set;
UConverter* test_converter;
UChar32 cp;
int32_t encIndex, offset;
set = uset_openEmpty();
test_converter = ucnv_open(encodings[i], &status);
ucnv_getUnicodeSet(test_converter, set,
whichSet, &status);
if (excludedCodePoints != NULL) {
uset_addAll(set, excludedCodePoints);
}
uset_freeze(set);
offset = 0;
cp = 0;
encIndex = findIndex(encodings[i]);
/*
* The following is almost, but not entirely, the same as
* resultsManually[encIndex] =
* (UBool)(uset_spanUTF8(set, utf8, length, USET_SPAN_SIMPLE) == length);
* They might be different if the set contains strings,
* or if the utf8 string contains an illegal sequence.
*
* The UConverterSelector does not currently handle strings that can be
* converted, and it treats an illegal sequence as convertible
* while uset_spanUTF8() treats it like U+FFFD which may not be convertible.
*/
resultsManually[encIndex] = TRUE;
while(offset<length) {
U8_NEXT(utf8, offset, length, cp);
if (cp >= 0 && !uset_contains(set, cp)) {
resultsManually[encIndex] = FALSE;
break;
}
}
uset_close(set);
ucnv_close(test_converter);
}
return resultsManually;
}
// pattern (ustr) is not expected to be a class here (no square brackets here)
USet *create_set_from_ustring(UString *ustr, UBool negate, error_t **UNUSED(error))
{
USet *uset;
uset = uset_openEmpty();
uset_addAllCodePoints(uset, ustr->ptr, ustr->len);
if (negate) {
uset_complement(uset);
}
uset_freeze(uset);
return uset;
}
U_CAPI USet* U_EXPORT2
utrans_getSourceSet(const UTransliterator* trans,
UBool ignoreFilter,
USet* fillIn,
UErrorCode* status) {
utrans_ENTRY(status) fillIn;
if (fillIn == NULL) {
fillIn = uset_openEmpty();
}
if (ignoreFilter) {
((Transliterator*) trans)->handleGetSourceSet(*((UnicodeSet*)fillIn));
} else {
((Transliterator*) trans)->getSourceSet(*((UnicodeSet*)fillIn));
}
return fillIn;
}
void SSearchTest::monkeyTest(char *params)
{
// ook!
UErrorCode status = U_ZERO_ERROR;
//UCollator *coll = ucol_open(NULL, &status);
UCollator *coll = ucol_openFromShortString("S1", FALSE, NULL, &status);
if (U_FAILURE(status)) {
errcheckln(status, "Failed to create collator in MonkeyTest! - %s", u_errorName(status));
return;
}
CollData *monkeyData = new CollData(coll, status);
USet *expansions = uset_openEmpty();
USet *contractions = uset_openEmpty();
ucol_getContractionsAndExpansions(coll, contractions, expansions, FALSE, &status);
U_STRING_DECL(letter_pattern, "[[:letter:]-[:ideographic:]-[:hangul:]]", 39);
U_STRING_INIT(letter_pattern, "[[:letter:]-[:ideographic:]-[:hangul:]]", 39);
USet *letters = uset_openPattern(letter_pattern, 39, &status);
SetMonkey letterMonkey(letters);
StringSetMonkey contractionMonkey(contractions, coll, monkeyData);
StringSetMonkey expansionMonkey(expansions, coll, monkeyData);
UnicodeString testCase;
UnicodeString alternate;
UnicodeString pattern, altPattern;
UnicodeString prefix, altPrefix;
UnicodeString suffix, altSuffix;
Monkey *monkeys[] = {
&letterMonkey,
&contractionMonkey,
&expansionMonkey,
&contractionMonkey,
&expansionMonkey,
&contractionMonkey,
&expansionMonkey,
&contractionMonkey,
&expansionMonkey};
int32_t monkeyCount = sizeof(monkeys) / sizeof(monkeys[0]);
// int32_t nonMatchCount = 0;
UCollationStrength strengths[] = {UCOL_PRIMARY, UCOL_SECONDARY, UCOL_TERTIARY};
const char *strengthNames[] = {"primary", "secondary", "tertiary"};
int32_t strengthCount = sizeof(strengths) / sizeof(strengths[0]);
int32_t loopCount = quick? 1000 : 10000;
int32_t firstStrength = 0;
int32_t lastStrength = strengthCount - 1; //*/ 0;
if (params != NULL) {
#if !UCONFIG_NO_REGULAR_EXPRESSIONS
UnicodeString p(params);
loopCount = getIntParam("loop", p, loopCount);
m_seed = getIntParam("seed", p, m_seed);
RegexMatcher m(" *strength *= *(primary|secondary|tertiary) *", p, 0, status);
if (m.find()) {
UnicodeString breakType = m.group(1, status);
for (int32_t s = 0; s < strengthCount; s += 1) {
if (breakType == strengthNames[s]) {
firstStrength = lastStrength = s;
break;
}
}
m.reset();
p = m.replaceFirst("", status);
}
if (RegexMatcher("\\S", p, 0, status).find()) {
// Each option is stripped out of the option string as it is processed.
// All options have been checked. The option string should have been completely emptied..
char buf[100];
p.extract(buf, sizeof(buf), NULL, status);
buf[sizeof(buf)-1] = 0;
errln("Unrecognized or extra parameter: %s\n", buf);
return;
}
#else
infoln("SSearchTest built with UCONFIG_NO_REGULAR_EXPRESSIONS: ignoring parameters.");
#endif
}
for(int32_t s = firstStrength; s <= lastStrength; s += 1) {
int32_t notFoundCount = 0;
logln("Setting strength to %s.", strengthNames[s]);
ucol_setStrength(coll, strengths[s]);
// TODO: try alternate prefix and suffix too?
// TODO: alterntaes are only equal at primary strength. Is this OK?
for(int32_t t = 0; t < loopCount; t += 1) {
uint32_t seed = m_seed;
// int32_t nmc = 0;
generateTestCase(coll, monkeys, monkeyCount, pattern, altPattern);
generateTestCase(coll, monkeys, monkeyCount, prefix, altPrefix);
generateTestCase(coll, monkeys, monkeyCount, suffix, altSuffix);
// pattern
notFoundCount += monkeyTestCase(coll, pattern, pattern, altPattern, "pattern", strengthNames[s], seed);
testCase.remove();
testCase.append(prefix);
testCase.append(/*alt*/pattern);
// prefix + pattern
notFoundCount += monkeyTestCase(coll, testCase, pattern, altPattern, "prefix + pattern", strengthNames[s], seed);
testCase.append(suffix);
// prefix + pattern + suffix
notFoundCount += monkeyTestCase(coll, testCase, pattern, altPattern, "prefix + pattern + suffix", strengthNames[s], seed);
testCase.remove();
testCase.append(pattern);
testCase.append(suffix);
// pattern + suffix
notFoundCount += monkeyTestCase(coll, testCase, pattern, altPattern, "pattern + suffix", strengthNames[s], seed);
}
logln("For strength %s the not found count is %d.", strengthNames[s], notFoundCount);
}
uset_close(contractions);
uset_close(expansions);
uset_close(letters);
delete monkeyData;
ucol_close(coll);
}
CollData::CollData(UCollator *collator, UErrorCode &status)
: coll(NULL), ceToCharsStartingWith(NULL)
{
// [:c:] == [[:cn:][:cc:][:co:][:cf:][:cs:]]
// i.e. other, control, private use, format, surrogate
U_STRING_DECL(test_pattern, "[[:assigned:]-[:c:]]", 20);
U_STRING_INIT(test_pattern, "[[:assigned:]-[:c:]]", 20);
USet *charsToTest = uset_openPattern(test_pattern, 20, &status);
// Han ext. A, Han, Jamo, Hangul, Han Ext. B
// i.e. all the characers we handle implicitly
U_STRING_DECL(remove_pattern, "[[\\u3400-\\u9FFF][\\u1100-\\u11F9][\\uAC00-\\uD7AF][\\U00020000-\\U0002A6DF]]", 70);
U_STRING_INIT(remove_pattern, "[[\\u3400-\\u9FFF][\\u1100-\\u11F9][\\uAC00-\\uD7AF][\\U00020000-\\U0002A6DF]]", 70);
USet *charsToRemove = uset_openPattern(remove_pattern, 70, &status);
if (U_FAILURE(status)) {
return;
}
USet *expansions = uset_openEmpty();
USet *contractions = uset_openEmpty();
int32_t itemCount;
ceToCharsStartingWith = new CEToStringsMap(status);
if (U_FAILURE(status)) {
goto bail;
}
#ifdef CLONE_COLLATOR
coll = ucol_safeClone(collator, NULL, NULL, &status);
if (U_FAILURE(status)) {
goto bail;
}
#else
coll = collator;
#endif
ucol_getContractionsAndExpansions(coll, contractions, expansions, FALSE, &status);
uset_addAll(charsToTest, contractions);
uset_addAll(charsToTest, expansions);
uset_removeAll(charsToTest, charsToRemove);
itemCount = uset_getItemCount(charsToTest);
for(int32_t item = 0; item < itemCount; item += 1) {
UChar32 start = 0, end = 0;
UChar buffer[16];
int32_t len = uset_getItem(charsToTest, item, &start, &end,
buffer, 16, &status);
if (len == 0) {
for (UChar32 ch = start; ch <= end; ch += 1) {
UnicodeString *st = new UnicodeString(ch);
if (st == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
break;
}
CEList *ceList = new CEList(coll, *st, status);
ceToCharsStartingWith->put(ceList->get(0), st, status);
delete ceList;
delete st;
}
} else if (len > 0) {
UnicodeString *st = new UnicodeString(buffer, len);
if (st == NULL) {
status = U_MEMORY_ALLOCATION_ERROR;
break;
}
CEList *ceList = new CEList(coll, *st, status);
ceToCharsStartingWith->put(ceList->get(0), st, status);
delete ceList;
delete st;
} else {
// shouldn't happen...
}
if (U_FAILURE(status)) {
break;
}
}
bail:
uset_close(contractions);
uset_close(expansions);
uset_close(charsToRemove);
uset_close(charsToTest);
if (U_FAILURE(status)) {
return;
}
UnicodeSet hanRanges(UNICODE_STRING_SIMPLE("[:Unified_Ideograph:]"), status);
if (U_FAILURE(status)) {
return;
}
UnicodeSetIterator hanIter(hanRanges);
UnicodeString hanString;
while(hanIter.nextRange()) {
hanString.append(hanIter.getCodepoint());
hanString.append(hanIter.getCodepointEnd());
}
// TODO: Why U+11FF? The old code had an outdated UCOL_LAST_T_JAMO=0x11F9,
// but as of Unicode 6.3 the 11xx block is filled,
// and there are also more Jamo T at U+D7CB..U+D7FB.
// Maybe use [:HST=T:] and look for the end of the last range?
// Maybe use script boundary mappings instead of this code??
UChar jamoRanges[] = {Hangul::JAMO_L_BASE, Hangul::JAMO_V_BASE, Hangul::JAMO_T_BASE + 1, 0x11FF};
UnicodeString jamoString(FALSE, jamoRanges, UPRV_LENGTHOF(jamoRanges));
CEList hanList(coll, hanString, status);
CEList jamoList(coll, jamoString, status);
int32_t j = 0;
if (U_FAILURE(status)) {
return;
}
for (int32_t c = 0; c < jamoList.size(); c += 1) {
uint32_t jce = jamoList[c];
if (! isContinuation(jce)) {
jamoLimits[j++] = jce;
}
}
jamoLimits[3] += (1 << UCOL_PRIMARYORDERSHIFT);
minHan = 0xFFFFFFFF;
maxHan = 0;
for(int32_t h = 0; h < hanList.size(); h += 2) {
uint32_t han = (uint32_t) hanList[h];
if (han < minHan) {
minHan = han;
}
if (han > maxHan) {
maxHan = han;
}
}
maxHan += (1 << UCOL_PRIMARYORDERSHIFT);
}
static void TestSelector()
{
TestText text;
USet* excluded_sets[3] = { NULL };
int32_t i, testCaseIdx;
if (!getAvailableNames()) {
return;
}
if (!text_open(&text)) {
releaseAvailableNames();;
}
excluded_sets[0] = uset_openEmpty();
for(i = 1 ; i < 3 ; i++) {
excluded_sets[i] = uset_open(i*30, i*30+500);
}
for(testCaseIdx = 0; testCaseIdx < UPRV_LENGTHOF(getEncodingsFns); testCaseIdx++)
{
int32_t excluded_set_id;
int32_t num_encodings;
const char **encodings = getEncodingsFns[testCaseIdx](&num_encodings);
if (getTestOption(QUICK_OPTION) && num_encodings > 25) {
uprv_free((void *)encodings);
continue;
}
/*
* for(excluded_set_id = 0 ; excluded_set_id < 3 ; excluded_set_id++)
*
* This loop was replaced by the following statement because
* the loop made the test run longer without adding to the code coverage.
* The handling of the exclusion set is independent of the
* set of encodings, so there is no need to test every combination.
*/
excluded_set_id = testCaseIdx % UPRV_LENGTHOF(excluded_sets);
{
UConverterSelector *sel_rt, *sel_fb;
char *buffer_fb = NULL;
UErrorCode status = U_ZERO_ERROR;
sel_rt = ucnvsel_open(encodings, num_encodings,
excluded_sets[excluded_set_id],
UCNV_ROUNDTRIP_SET, &status);
if (num_encodings == gCountAvailable) {
/* test the special "all converters" parameter values */
sel_fb = ucnvsel_open(NULL, 0,
excluded_sets[excluded_set_id],
UCNV_ROUNDTRIP_AND_FALLBACK_SET, &status);
} else if (uset_isEmpty(excluded_sets[excluded_set_id])) {
/* test that a NULL set gives the same results as an empty set */
sel_fb = ucnvsel_open(encodings, num_encodings,
NULL,
UCNV_ROUNDTRIP_AND_FALLBACK_SET, &status);
} else {
sel_fb = ucnvsel_open(encodings, num_encodings,
excluded_sets[excluded_set_id],
UCNV_ROUNDTRIP_AND_FALLBACK_SET, &status);
}
if (U_FAILURE(status)) {
log_err("ucnv_sel_open(encodings %ld) failed - %s\n", testCaseIdx, u_errorName(status));
ucnvsel_close(sel_rt);
uprv_free((void *)encodings);
continue;
}
text_reset(&text);
for (;;) {
UBool *manual_rt, *manual_fb;
static UChar utf16[10000];
char *s;
int32_t length8, length16;
s = text_nextString(&text, &length8);
if (s == NULL || (getTestOption(QUICK_OPTION) && text.number > 3)) {
break;
}
manual_rt = getResultsManually(encodings, num_encodings,
s, length8,
excluded_sets[excluded_set_id],
UCNV_ROUNDTRIP_SET);
manual_fb = getResultsManually(encodings, num_encodings,
s, length8,
excluded_sets[excluded_set_id],
UCNV_ROUNDTRIP_AND_FALLBACK_SET);
/* UTF-8 with length */
status = U_ZERO_ERROR;
verifyResult(ucnvsel_selectForUTF8(sel_rt, s, length8, &status), manual_rt);
verifyResult(ucnvsel_selectForUTF8(sel_fb, s, length8, &status), manual_fb);
/* UTF-8 NUL-terminated */
verifyResult(ucnvsel_selectForUTF8(sel_rt, s, -1, &status), manual_rt);
verifyResult(ucnvsel_selectForUTF8(sel_fb, s, -1, &status), manual_fb);
u_strFromUTF8(utf16, UPRV_LENGTHOF(utf16), &length16, s, length8, &status);
if (U_FAILURE(status)) {
log_err("error converting the test text (string %ld) to UTF-16 - %s\n",
(long)text.number, u_errorName(status));
} else {
if (text.number == 0) {
sel_fb = serializeAndUnserialize(sel_fb, &buffer_fb, &status);
}
if (U_SUCCESS(status)) {
/* UTF-16 with length */
verifyResult(ucnvsel_selectForString(sel_rt, utf16, length16, &status), manual_rt);
verifyResult(ucnvsel_selectForString(sel_fb, utf16, length16, &status), manual_fb);
/* UTF-16 NUL-terminated */
verifyResult(ucnvsel_selectForString(sel_rt, utf16, -1, &status), manual_rt);
verifyResult(ucnvsel_selectForString(sel_fb, utf16, -1, &status), manual_fb);
}
}
uprv_free(manual_rt);
uprv_free(manual_fb);
}
ucnvsel_close(sel_rt);
ucnvsel_close(sel_fb);
uprv_free(buffer_fb);
}
uprv_free((void *)encodings);
}
releaseAvailableNames();
text_close(&text);
for(i = 0 ; i < 3 ; i++) {
uset_close(excluded_sets[i]);
}
}
/**
* Basic API test for uset.x
*/
static void TestAPI() {
USet* set;
USet* set2;
UErrorCode ec;
/* [] */
set = uset_openEmpty();
expect(set, "", "abc{ab}", NULL);
uset_close(set);
set = uset_open(1, 0);
expect(set, "", "abc{ab}", NULL);
uset_close(set);
set = uset_open(1, 1);
uset_clear(set);
expect(set, "", "abc{ab}", NULL);
uset_close(set);
/* [ABC] */
set = uset_open(0x0041, 0x0043);
expect(set, "ABC", "DEF{ab}", NULL);
uset_close(set);
/* [a-c{ab}] */
ec = U_ZERO_ERROR;
set = uset_openPattern(PAT, PAT_LEN, &ec);
if(U_FAILURE(ec)) {
log_err("uset_openPattern([a-c{ab}]) failed - %s\n", u_errorName(ec));
return;
}
if(!uset_resemblesPattern(PAT, PAT_LEN, 0)) {
log_err("uset_resemblesPattern of PAT failed\n");
}
expect(set, "abc{ab}", "def{bc}", &ec);
/* [a-d{ab}] */
uset_add(set, 0x64);
expect(set, "abcd{ab}", "ef{bc}", NULL);
/* [acd{ab}{bc}] */
uset_remove(set, 0x62);
uset_addString(set, STR_bc, STR_bc_LEN);
expect(set, "acd{ab}{bc}", "bef{cd}", NULL);
/* [acd{bc}] */
uset_removeString(set, STR_ab, STR_ab_LEN);
expect(set, "acd{bc}", "bfg{ab}", NULL);
/* [^acd{bc}] */
uset_complement(set);
expect(set, "bef{bc}", "acd{ac}", NULL);
/* [a-e{bc}] */
uset_complement(set);
uset_addRange(set, 0x0062, 0x0065);
expect(set, "abcde{bc}", "fg{ab}", NULL);
/* [de{bc}] */
uset_removeRange(set, 0x0050, 0x0063);
expect(set, "de{bc}", "bcfg{ab}", NULL);
/* [g-l] */
uset_set(set, 0x0067, 0x006C);
expect(set, "ghijkl", "de{bc}", NULL);
if (uset_indexOf(set, 0x0067) != 0) {
log_err("uset_indexOf failed finding correct index of 'g'\n");
}
if (uset_charAt(set, 0) != 0x0067) {
log_err("uset_charAt failed finding correct char 'g' at index 0\n");
}
/* How to test this one...? */
uset_compact(set);
/* [g-i] */
uset_retain(set, 0x0067, 0x0069);
expect(set, "ghi", "dejkl{bc}", NULL);
/* UCHAR_ASCII_HEX_DIGIT */
uset_applyIntPropertyValue(set, UCHAR_ASCII_HEX_DIGIT, 1, &ec);
if(U_FAILURE(ec)) {
log_err("uset_applyIntPropertyValue([UCHAR_ASCII_HEX_DIGIT]) failed - %s\n", u_errorName(ec));
return;
}
expect(set, "0123456789ABCDEFabcdef", "GHIjkl{bc}", NULL);
/* [ab] */
uset_clear(set);
uset_addAllCodePoints(set, STR_ab, STR_ab_LEN);
expect(set, "ab", "def{ab}", NULL);
if (uset_containsAllCodePoints(set, STR_bc, STR_bc_LEN)){
log_err("set should not conatin all characters of \"bc\" \n");
}
/* [] */
set2 = uset_open(1, 1);
uset_clear(set2);
/* space */
uset_applyPropertyAlias(set2, PAT_lb, PAT_lb_LEN, VAL_SP, VAL_SP_LEN, &ec);
expect(set2, " ", "abcdefghi{bc}", NULL);
/* [a-c] */
uset_set(set2, 0x0061, 0x0063);
/* [g-i] */
uset_set(set, 0x0067, 0x0069);
/* [a-c g-i] */
if (uset_containsSome(set, set2)) {
log_err("set should not contain some of set2 yet\n");
}
uset_complementAll(set, set2);
if (!uset_containsSome(set, set2)) {
log_err("set should contain some of set2\n");
}
expect(set, "abcghi", "def{bc}", NULL);
/* [g-i] */
uset_removeAll(set, set2);
expect(set, "ghi", "abcdef{bc}", NULL);
/* [a-c g-i] */
uset_addAll(set2, set);
expect(set2, "abcghi", "def{bc}", NULL);
/* [g-i] */
uset_retainAll(set2, set);
expect(set2, "ghi", "abcdef{bc}", NULL);
uset_close(set);
uset_close(set2);
}
