void CollationThaiTest::TestReordering(void) {
const char *tests[] = {
"\\u0E41c\\u0301", "=", "\\u0E41\\u0107", // composition
"\\u0E41\\uD835\\uDFCE", "<", "\\u0E41\\uD835\\uDFCF", // supplementaries
"\\u0E41\\uD834\\uDD5F", "=", "\\u0E41\\uD834\\uDD58\\uD834\\uDD65", // supplementary composition decomps to supplementary
"\\u0E41\\uD87E\\uDC02", "=", "\\u0E41\\u4E41", // supplementary composition decomps to BMP
"\\u0E41\\u0301", "=", "\\u0E41\\u0301", // unsafe (just checking backwards iteration)
"\\u0E41\\u0301\\u0316", "=", "\\u0E41\\u0316\\u0301",
// after UCA 4.1, the two lines below are not equal anymore do not have equal sign
"\\u0e24\\u0e41", "<", "\\u0e41\\u0e24", // exiting contraction bug
"\\u0e3f\\u0e3f\\u0e24\\u0e41", "<", "\\u0e3f\\u0e3f\\u0e41\\u0e24",
"abc\\u0E41c\\u0301", "=", "abc\\u0E41\\u0107", // composition
"abc\\u0E41\\uD834\\uDC00", "<", "abc\\u0E41\\uD834\\uDC01", // supplementaries
"abc\\u0E41\\uD834\\uDD5F", "=", "abc\\u0E41\\uD834\\uDD58\\uD834\\uDD65", // supplementary composition decomps to supplementary
"abc\\u0E41\\uD87E\\uDC02", "=", "abc\\u0E41\\u4E41", // supplementary composition decomps to BMP
"abc\\u0E41\\u0301", "=", "abc\\u0E41\\u0301", // unsafe (just checking backwards iteration)
"abc\\u0E41\\u0301\\u0316", "=", "abc\\u0E41\\u0316\\u0301",
"\\u0E41c\\u0301abc", "=", "\\u0E41\\u0107abc", // composition
"\\u0E41\\uD834\\uDC00abc", "<", "\\u0E41\\uD834\\uDC01abc", // supplementaries
"\\u0E41\\uD834\\uDD5Fabc", "=", "\\u0E41\\uD834\\uDD58\\uD834\\uDD65abc", // supplementary composition decomps to supplementary
"\\u0E41\\uD87E\\uDC02abc", "=", "\\u0E41\\u4E41abc", // supplementary composition decomps to BMP
"\\u0E41\\u0301abc", "=", "\\u0E41\\u0301abc", // unsafe (just checking backwards iteration)
"\\u0E41\\u0301\\u0316abc", "=", "\\u0E41\\u0316\\u0301abc",
"abc\\u0E41c\\u0301abc", "=", "abc\\u0E41\\u0107abc", // composition
"abc\\u0E41\\uD834\\uDC00abc", "<", "abc\\u0E41\\uD834\\uDC01abc", // supplementaries
"abc\\u0E41\\uD834\\uDD5Fabc", "=", "abc\\u0E41\\uD834\\uDD58\\uD834\\uDD65abc", // supplementary composition decomps to supplementary
"abc\\u0E41\\uD87E\\uDC02abc", "=", "abc\\u0E41\\u4E41abc", // supplementary composition decomps to BMP
"abc\\u0E41\\u0301abc", "=", "abc\\u0E41\\u0301abc", // unsafe (just checking backwards iteration)
"abc\\u0E41\\u0301\\u0316abc", "=", "abc\\u0E41\\u0316\\u0301abc",
};
compareArray(*coll, tests, sizeof(tests)/sizeof(tests[0]));
const char *rule = "& c < ab";
const char *testcontraction[] = { "\\u0E41ab", ">", "\\u0E41c"}; // After UCA 4.1 Thai are normal so won't break a contraction
UnicodeString rules;
UErrorCode status = U_ZERO_ERROR;
parseChars(rules, rule);
RuleBasedCollator *rcoll = new RuleBasedCollator(rules, status);
if(U_SUCCESS(status)) {
compareArray(*rcoll, testcontraction, 3);
delete rcoll;
} else {
errln("Couldn't instantiate collator from rules");
}
}
void CollationThaiTest::TestReordering(void) {
// Until UCA 4.1, the collation code swapped Thai/Lao prevowels with the following consonants,
// resulting in consonant+prevowel == prevowel+consonant.
// From UCA 5.0 on, there are order-reversing contractions for prevowel+consonant.
// From UCA 5.0 until UCA 6.1, there was a tertiary difference between
// consonant+prevowel and prevowel+consonant.
// In UCA 6.2, they compare equal again.
// The test was modified to using a collator with strength=secondary,
// ignoring possible tertiary differences.
const char *tests[] = {
"\\u0E41c\\u0301", "=", "\\u0E41\\u0107", // composition
"\\u0E41\\U0001D7CE", "<", "\\u0E41\\U0001D7CF", // supplementaries
"\\u0E41\\U0001D15F", "=", "\\u0E41\\U0001D158\\U0001D165", // supplementary composition decomps to supplementary
"\\u0E41\\U0002F802", "=", "\\u0E41\\u4E41", // supplementary composition decomps to BMP
"\\u0E41\\u0301", "=", "\\u0E41\\u0301", // unsafe (just checking backwards iteration)
"\\u0E41\\u0301\\u0316", "=", "\\u0E41\\u0316\\u0301",
"\\u0e24\\u0e41", "=", "\\u0e41\\u0e24", // exiting contraction bug
"\\u0e3f\\u0e3f\\u0e24\\u0e41", "=", "\\u0e3f\\u0e3f\\u0e41\\u0e24",
"abc\\u0E41c\\u0301", "=", "abc\\u0E41\\u0107", // composition
"abc\\u0E41\\U0001D000", "<", "abc\\u0E41\\U0001D001", // supplementaries
"abc\\u0E41\\U0001D15F", "=", "abc\\u0E41\\U0001D158\\U0001D165", // supplementary composition decomps to supplementary
"abc\\u0E41\\U0002F802", "=", "abc\\u0E41\\u4E41", // supplementary composition decomps to BMP
"abc\\u0E41\\u0301", "=", "abc\\u0E41\\u0301", // unsafe (just checking backwards iteration)
"abc\\u0E41\\u0301\\u0316", "=", "abc\\u0E41\\u0316\\u0301",
"\\u0E41c\\u0301abc", "=", "\\u0E41\\u0107abc", // composition
"\\u0E41\\U0001D000abc", "<", "\\u0E41\\U0001D001abc", // supplementaries
"\\u0E41\\U0001D15Fabc", "=", "\\u0E41\\U0001D158\\U0001D165abc", // supplementary composition decomps to supplementary
"\\u0E41\\U0002F802abc", "=", "\\u0E41\\u4E41abc", // supplementary composition decomps to BMP
"\\u0E41\\u0301abc", "=", "\\u0E41\\u0301abc", // unsafe (just checking backwards iteration)
"\\u0E41\\u0301\\u0316abc", "=", "\\u0E41\\u0316\\u0301abc",
"abc\\u0E41c\\u0301abc", "=", "abc\\u0E41\\u0107abc", // composition
"abc\\u0E41\\U0001D000abc", "<", "abc\\u0E41\\U0001D001abc", // supplementaries
"abc\\u0E41\\U0001D15Fabc", "=", "abc\\u0E41\\U0001D158\\U0001D165abc", // supplementary composition decomps to supplementary
"abc\\u0E41\\U0002F802abc", "=", "abc\\u0E41\\u4E41abc", // supplementary composition decomps to BMP
"abc\\u0E41\\u0301abc", "=", "abc\\u0E41\\u0301abc", // unsafe (just checking backwards iteration)
"abc\\u0E41\\u0301\\u0316abc", "=", "abc\\u0E41\\u0316\\u0301abc",
};
LocalPointer<Collator> coll2(coll->clone());
UErrorCode status = U_ZERO_ERROR;
coll2->setAttribute(UCOL_STRENGTH, UCOL_SECONDARY, status);
if(U_FAILURE(status)) {
errln("Unable to set the Thai collator clone to secondary strength");
return;
}
compareArray(*coll2, tests, sizeof(tests)/sizeof(tests[0]));
const char *rule = "& c < ab";
const char *testcontraction[] = { "\\u0E41ab", ">", "\\u0E41c"}; // After UCA 4.1 Thai are normal so won't break a contraction
UnicodeString rules;
parseChars(rules, rule);
LocalPointer<RuleBasedCollator> rcoll(new RuleBasedCollator(rules, status));
if(U_SUCCESS(status)) {
compareArray(*rcoll, testcontraction, 3);
} else {
errln("Couldn't instantiate collator from rules");
}
}
void CollationThaiTest::compareArray(Collator& c, const char* tests[],
int32_t testsLength) {
for (int32_t i = 0; i < testsLength; i += 3) {
Collator::EComparisonResult expect;
if (tests[i+1][0] == '<') {
expect = Collator::LESS;
} else if (tests[i+1][0] == '>') {
expect = Collator::GREATER;
} else if (tests[i+1][0] == '=') {
expect = Collator::EQUAL;
} else {
// expect = Integer.decode(tests[i+1]).intValue();
errln((UnicodeString)"Error: unknown operator " + tests[i+1]);
return;
}
UnicodeString s1, s2;
parseChars(s1, tests[i]);
parseChars(s2, tests[i+2]);
doTest(&c, s1, s2, expect);
#if 0
UErrorCode status = U_ZERO_ERROR;
int32_t result = c.compare(s1, s2);
if (sign(result) != sign(expect))
{
UnicodeString t1, t2;
errln(UnicodeString("") +
i/3 + ": compare(" + IntlTest::prettify(s1, t1)
+ " , " + IntlTest::prettify(s2, t2)
+ ") got " + result + "; expected " + expect);
CollationKey k1, k2;
c.getCollationKey(s1, k1, status);
c.getCollationKey(s2, k2, status);
if (U_FAILURE(status)) {
errln((UnicodeString)"Fail: getCollationKey returned " + u_errorName(status));
return;
}
errln((UnicodeString)" key1: " + prettify(k1, t1) );
errln((UnicodeString)" key2: " + prettify(k2, t2) );
}
else
{
// Collator.compare worked OK; now try the collation keys
CollationKey k1, k2;
c.getCollationKey(s1, k1, status);
c.getCollationKey(s2, k2, status);
if (U_FAILURE(status)) {
errln((UnicodeString)"Fail: getCollationKey returned " + u_errorName(status));
return;
}
result = k1.compareTo(k2);
if (sign(result) != sign(expect)) {
UnicodeString t1, t2;
errln(UnicodeString("") +
i/3 + ": key(" + IntlTest::prettify(s1, t1)
+ ").compareTo(key(" + IntlTest::prettify(s2, t2)
+ ")) got " + result + "; expected " + expect);
errln((UnicodeString)" " + prettify(k1, t1) + " vs. " + prettify(k2, t2));
}
}
#endif
}
}
