+#endif /* !UCONFIG_NO_NORMALIZATION */
+}
+
+void UnicodeTest::TestBinaryValues() {
+ /*
+ * Unicode 5.1 explicitly defines binary property value aliases.
+ * Verify that they are all recognized.
+ */
+ UErrorCode errorCode=U_ZERO_ERROR;
+ UnicodeSet alpha(UNICODE_STRING_SIMPLE("[:Alphabetic:]"), errorCode);
+ if(U_FAILURE(errorCode)) {
+ dataerrln("UnicodeSet([:Alphabetic:]) failed - %s", u_errorName(errorCode));
+ return;
+ }
+
+ 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) {
+ UnicodeString pattern=UNICODE_STRING_SIMPLE("[:Alphabetic=:]");
+ pattern.insert(pattern.length()-2, UnicodeString(falseValues[i], -1, US_INV));
+ errorCode=U_ZERO_ERROR;
+ UnicodeSet set(pattern, errorCode);
+ if(U_FAILURE(errorCode)) {
+ errln("UnicodeSet([:Alphabetic=%s:]) failed - %s\n", falseValues[i], u_errorName(errorCode));
+ continue;
+ }
+ set.complement();
+ if(set!=alpha) {
+ errln("UnicodeSet([:Alphabetic=%s:]).complement()!=UnicodeSet([:Alphabetic:])\n", falseValues[i]);
+ }
+ }
+ for(i=0; i<LENGTHOF(trueValues); ++i) {
+ UnicodeString pattern=UNICODE_STRING_SIMPLE("[:Alphabetic=:]");
+ pattern.insert(pattern.length()-2, UnicodeString(trueValues[i], -1, US_INV));
+ errorCode=U_ZERO_ERROR;
+ UnicodeSet set(pattern, errorCode);
+ if(U_FAILURE(errorCode)) {
+ errln("UnicodeSet([:Alphabetic=%s:]) failed - %s\n", trueValues[i], u_errorName(errorCode));
+ continue;
+ }
+ if(set!=alpha) {
+ errln("UnicodeSet([:Alphabetic=%s:])!=UnicodeSet([:Alphabetic:])\n", trueValues[i]);
+ }
+ }
+}
+
+void UnicodeTest::TestConsistency() {
+#if !UCONFIG_NO_NORMALIZATION
+ /*
+ * Test for an example that getCanonStartSet() delivers
+ * all characters that compose from the input one,
+ * even in multiple steps.
+ * For example, the set for "I" (0049) should contain both
+ * I-diaeresis (00CF) and I-diaeresis-acute (1E2E).
+ * In general, the set for the middle such character should be a subset
+ * of the set for the first.
+ */
+ IcuTestErrorCode errorCode(*this, "TestConsistency");
+ const Normalizer2 *nfd=Normalizer2::getNFDInstance(errorCode);
+ const Normalizer2Impl *nfcImpl=Normalizer2Factory::getNFCImpl(errorCode);
+ if(!nfcImpl->ensureCanonIterData(errorCode) || errorCode.isFailure()) {
+ dataerrln("Normalizer2::getInstance(NFD) or Normalizer2Factory::getNFCImpl() failed - %s\n",
+ errorCode.errorName());
+ errorCode.reset();
+ return;
+ }
+
+ UnicodeSet set1, set2;
+ if (nfcImpl->getCanonStartSet(0x49, set1)) {
+ /* enumerate all characters that are plausible to be latin letters */
+ for(UChar start=0xa0; start<0x2000; ++start) {
+ UnicodeString decomp=nfd->normalize(UnicodeString(start), errorCode);
+ if(decomp.length()>1 && decomp[0]==0x49) {
+ set2.add(start);
+ }
+ }
+
+ if (set1!=set2) {
+ errln("[canon start set of 0049] != [all c with canon decomp with 0049]");
+ }
+ // This was available in cucdtst.c but the test had to move to intltest
+ // because the new internal normalization functions are in C++.
+ //compareUSets(set1, set2,
+ // "[canon start set of 0049]", "[all c with canon decomp with 0049]",
+ // TRUE);
+ } else {
+ errln("NFC.getCanonStartSet() returned FALSE");
+ }
+#endif
+}
+
+/**
+ * Test various implementations of Pattern_Syntax & Pattern_White_Space.
+ */
+void UnicodeTest::TestPatternProperties() {
+ IcuTestErrorCode errorCode(*this, "TestPatternProperties()");
+ UnicodeSet syn_pp;
+ UnicodeSet syn_prop(UNICODE_STRING_SIMPLE("[:Pattern_Syntax:]"), errorCode);
+ UnicodeSet syn_list(
+ "[!-/\\:-@\\[-\\^`\\{-~"
+ "\\u00A1-\\u00A7\\u00A9\\u00AB\\u00AC\\u00AE\\u00B0\\u00B1\\u00B6\\u00BB\\u00BF\\u00D7\\u00F7"
+ "\\u2010-\\u2027\\u2030-\\u203E\\u2041-\\u2053\\u2055-\\u205E\\u2190-\\u245F\\u2500-\\u2775"
+ "\\u2794-\\u2BFF\\u2E00-\\u2E7F\\u3001-\\u3003\\u3008-\\u3020\\u3030\\uFD3E\\uFD3F\\uFE45\\uFE46]", errorCode);
+ UnicodeSet ws_pp;
+ UnicodeSet ws_prop(UNICODE_STRING_SIMPLE("[:Pattern_White_Space:]"), errorCode);
+ UnicodeSet ws_list(UNICODE_STRING_SIMPLE("[\\u0009-\\u000D\\ \\u0085\\u200E\\u200F\\u2028\\u2029]"), errorCode);
+ UnicodeSet syn_ws_pp;
+ UnicodeSet syn_ws_prop(syn_prop);
+ syn_ws_prop.addAll(ws_prop);
+ for(UChar32 c=0; c<=0xffff; ++c) {
+ if(PatternProps::isSyntax(c)) {
+ syn_pp.add(c);
+ }
+ if(PatternProps::isWhiteSpace(c)) {
+ ws_pp.add(c);
+ }
+ if(PatternProps::isSyntaxOrWhiteSpace(c)) {
+ syn_ws_pp.add(c);
+ }
+ }
+ compareUSets(syn_pp, syn_prop,
+ "PatternProps.isSyntax()", "[:Pattern_Syntax:]", TRUE);
+ compareUSets(syn_pp, syn_list,
+ "PatternProps.isSyntax()", "[Pattern_Syntax ranges]", TRUE);
+ compareUSets(ws_pp, ws_prop,
+ "PatternProps.isWhiteSpace()", "[:Pattern_White_Space:]", TRUE);
+ compareUSets(ws_pp, ws_list,
+ "PatternProps.isWhiteSpace()", "[Pattern_White_Space ranges]", TRUE);
+ compareUSets(syn_ws_pp, syn_ws_prop,
+ "PatternProps.isSyntaxOrWhiteSpace()",
+ "[[:Pattern_Syntax:][:Pattern_White_Space:]]", TRUE);
+}
+
+// So far only minimal port of Java & cucdtst.c compareUSets().
+UBool
+UnicodeTest::compareUSets(const UnicodeSet &a, const UnicodeSet &b,
+ const char *a_name, const char *b_name,
+ UBool diffIsError) {
+ UBool same= a==b;
+ if(!same && diffIsError) {
+ errln("Sets are different: %s vs. %s\n", a_name, b_name);
+ }
+ return same;
+}
+
+namespace {
+
+/**
+ * Maps a special script code to the most common script of its encoded characters.
+ */
+UScriptCode getCharScript(UScriptCode script) {
+ switch(script) {
+ case USCRIPT_SIMPLIFIED_HAN:
+ case USCRIPT_TRADITIONAL_HAN:
+ return USCRIPT_HAN;
+ case USCRIPT_JAPANESE:
+ return USCRIPT_HIRAGANA;
+ case USCRIPT_KOREAN:
+ return USCRIPT_HANGUL;
+ default:
+ return script;
+ }
+}
+
+} // namespace
+
+void UnicodeTest::TestScriptMetadata() {
+ IcuTestErrorCode errorCode(*this, "TestScriptMetadata()");
+ UnicodeSet rtl("[[:bc=R:][:bc=AL:]-[:Cn:]-[:sc=Common:]]", errorCode);
+ // So far, sample characters are uppercase.
+ // Georgian is special.
+ UnicodeSet cased("[[:Lu:]-[:sc=Common:]-[:sc=Geor:]]", errorCode);
+ for(int32_t sci = 0; sci < USCRIPT_CODE_LIMIT; ++sci) {
+ UScriptCode sc = (UScriptCode)sci;
+ // Run the test with -v to see which script has failures:
+ // .../intltest$ make && ./intltest utility/UnicodeTest/TestScriptMetadata -v | grep -C 3 FAIL
+ logln(uscript_getShortName(sc));
+ UScriptUsage usage = uscript_getUsage(sc);
+ UnicodeString sample = uscript_getSampleUnicodeString(sc);
+ UnicodeSet scriptSet;
+ scriptSet.applyIntPropertyValue(UCHAR_SCRIPT, sc, errorCode);
+ if(usage == USCRIPT_USAGE_NOT_ENCODED) {
+ assertTrue("not encoded, no sample", sample.isEmpty());
+ assertFalse("not encoded, not RTL", uscript_isRightToLeft(sc));
+ assertFalse("not encoded, not LB letters", uscript_breaksBetweenLetters(sc));
+ assertFalse("not encoded, not cased", uscript_isCased(sc));
+ assertTrue("not encoded, no characters", scriptSet.isEmpty());
+ } else {
+ assertFalse("encoded, has a sample character", sample.isEmpty());
+ UChar32 firstChar = sample.char32At(0);
+ UScriptCode charScript = getCharScript(sc);
+ assertEquals("script(sample(script))",
+ charScript, uscript_getScript(firstChar, errorCode));
+ assertEquals("RTL vs. set", rtl.contains(firstChar), uscript_isRightToLeft(sc));
+ assertEquals("cased vs. set", cased.contains(firstChar), uscript_isCased(sc));
+ assertEquals("encoded, has characters", sc == charScript, !scriptSet.isEmpty());
+ if(uscript_isRightToLeft(sc)) {
+ rtl.removeAll(scriptSet);
+ }
+ if(uscript_isCased(sc)) {
+ cased.removeAll(scriptSet);
+ }
+ }
+ }
+ UnicodeString pattern;
+ assertEquals("no remaining RTL characters",
+ UnicodeString("[]"), rtl.toPattern(pattern));
+ assertEquals("no remaining cased characters",
+ UnicodeString("[]"), cased.toPattern(pattern));
+
+ assertTrue("Hani breaks between letters", uscript_breaksBetweenLetters(USCRIPT_HAN));
+ assertTrue("Thai breaks between letters", uscript_breaksBetweenLetters(USCRIPT_THAI));
+ assertFalse("Latn does not break between letters", uscript_breaksBetweenLetters(USCRIPT_LATIN));
projects / apple / icu.git / blobdiff