// @bug 4114077
//
// Collation with decomposition off doesn't work for Europe
//
void CollationRegressionTest::Test4114077(/* char* par */)
{
// Ensure that we get the same results with decomposition off
// as we do with it on....
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::TERTIARY);
static const UChar test1[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x00C0, 0}, {0x3d, 0}, {0x41, 0x0300, 0}, // Should be equivalent
{0x70, 0x00ea, 0x63, 0x68, 0x65, 0}, {0x3e, 0}, {0x70, 0x00e9, 0x63, 0x68, 0x00e9, 0},
{0x0204, 0}, {0x3d, 0}, {0x45, 0x030F, 0},
{0x01fa, 0}, {0x3d, 0}, {0x41, 0x030a, 0x0301, 0}, // a-ring-acute -> a-ring, acute
// -> a, ring, acute
{0x41, 0x0300, 0x0316, 0}, {0x3c, 0}, {0x41, 0x0316, 0x0300, 0} // No reordering --> unequal
};
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_OFF, status);
compareArray(*c, test1, UPRV_LENGTHOF(test1));
static const UChar test2[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x41, 0x0300, 0x0316, 0}, {0x3d, 0}, {0x41, 0x0316, 0x0300, 0} // Reordering --> equal
};
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status);
compareArray(*c, test2, UPRV_LENGTHOF(test2));
delete c;
}
// @bug 4087241
//
// string comparison errors in Scandinavian collators
//
void CollationRegressionTest::Test4087241(/* char* par */)
{
UErrorCode status = U_ZERO_ERROR;
Locale da_DK("da", "DK");
RuleBasedCollator *c = NULL;
c = (RuleBasedCollator *) Collator::createInstance(da_DK, status);
if (c == NULL || U_FAILURE(status))
{
errln("Failed to create collator for da_DK locale");
delete c;
return;
}
c->setStrength(Collator::SECONDARY);
static const UChar tests[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x7a, 0}, {0x3c, 0}, {0x00E6, 0}, // z < ae
{0x61, 0x0308, 0}, {0x3c, 0}, {0x61, 0x030A, 0}, // a-umlaut < a-ring
{0x59, 0}, {0x3c, 0}, {0x75, 0x0308, 0}, // Y < u-umlaut
};
compareArray(*c, tests, UPRV_LENGTHOF(tests));
delete c;
}
// @bug 4062418
//
// Secondary/Tertiary comparison incorrect in French Secondary
//
void CollationRegressionTest::Test4062418(/* char* par */)
{
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = NULL;
c = (RuleBasedCollator *) Collator::createInstance(Locale::getCanadaFrench(), status);
if (c == NULL || U_FAILURE(status))
{
errln("Failed to create collator for Locale::getCanadaFrench()");
delete c;
return;
}
c->setStrength(Collator::SECONDARY);
/*
String[] tests = {
"p\u00eache", "<", "p\u00e9ch\u00e9", // Comparing accents from end, p\u00e9ch\u00e9 is greater
};
*/
static const UChar tests[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x70, 0x00EA, 0x63, 0x68, 0x65, 0}, {0x3c, 0}, {0x70, 0x00E9, 0x63, 0x68, 0x00E9, 0}
};
compareArray(*c, tests, UPRV_LENGTHOF(tests));
delete c;
}
void CollationIteratorTest::TestStrengthOrder()
{
int order = 0x0123ABCD;
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *coll =
(RuleBasedCollator *)Collator::createInstance(status);
if (coll == NULL || U_FAILURE(status))
{
errln("Couldn't create a default collator.");
return;
}
coll->setStrength(Collator::PRIMARY);
CollationElementIterator *iter =
coll->createCollationElementIterator(test1);
if (iter == NULL) {
errln("Couldn't create a collation element iterator from default collator");
return;
}
if (iter->strengthOrder(order) != 0x01230000) {
errln("Strength order for a primary strength collator should be the first 2 bytes");
return;
}
coll->setStrength(Collator::SECONDARY);
if (iter->strengthOrder(order) != 0x0123AB00) {
errln("Strength order for a secondary strength collator should be the third byte");
return;
}
coll->setStrength(Collator::TERTIARY);
if (iter->strengthOrder(order) != order) {
errln("Strength order for a tertiary strength collator should be the third byte");
return;
}
delete iter;
delete coll;
}
// @bug 4087243
//
// CollationKey takes ignorable strings into account when it shouldn't
//
void CollationRegressionTest::Test4087243(/* char* par */)
{
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::TERTIARY);
static const UChar tests[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x31, 0x32, 0x33, 0}, {0x3d, 0}, {0x31, 0x32, 0x33, 0x0001, 0} // 1 2 3 = 1 2 3 ctrl-A
};
compareArray(*c, tests, ARRAY_LENGTH(tests));
delete c;
}
// @bug 4103436
//
// Collator::compare not handling spaces properly
//
void CollationRegressionTest::Test4103436(/* char* par */)
{
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::TERTIARY);
static const UChar tests[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x66, 0x69, 0x6c, 0x65, 0}, {0x3c, 0}, {0x66, 0x69, 0x6c, 0x65, 0x20, 0x61, 0x63, 0x63, 0x65, 0x73, 0x73, 0},
{0x66, 0x69, 0x6c, 0x65, 0}, {0x3c, 0}, {0x66, 0x69, 0x6c, 0x65, 0x61, 0x63, 0x63, 0x65, 0x73, 0x73, 0}
};
compareArray(*c, tests, ARRAY_LENGTH(tests));
delete c;
}
// @bug 4076676
//
// Bad canonicalization of same-class combining characters
//
void CollationRegressionTest::Test4076676(/* char* par */)
{
// These combining characters are all in the same class, so they should not
// be reordered, and they should compare as unequal.
static const UChar s1[] = {0x41, 0x0301, 0x0302, 0x0300, 0};
static const UChar s2[] = {0x41, 0x0302, 0x0300, 0x0301, 0};
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::TERTIARY);
if (c->compare(s1,s2) == 0)
{
errln("Same-class combining chars were reordered");
}
delete c;
}
// @bug 4054736
//
// Full Decomposition mode not implemented
//
void CollationRegressionTest::Test4054736(/* char* par */)
{
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::SECONDARY);
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status);
static const UChar tests[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0xFB4F, 0}, {0x3d, 0}, {0x05D0, 0x05DC} // Alef-Lamed vs. Alef, Lamed
};
compareArray(*c, tests, ARRAY_LENGTH(tests));
delete c;
}
// @bug 4114076
//
// Collation not Unicode conformant with Hangul syllables
//
void CollationRegressionTest::Test4114076(/* char* par */)
{
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::TERTIARY);
//
// With Canonical decomposition, Hangul syllables should get decomposed
// into Jamo, but Jamo characters should not be decomposed into
// conjoining Jamo
//
static const UChar test1[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0xd4db, 0}, {0x3d, 0}, {0x1111, 0x1171, 0x11b6, 0}
};
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status);
compareArray(*c, test1, ARRAY_LENGTH(test1));
// From UTR #15:
// *In earlier versions of Unicode, jamo characters like ksf
// had compatibility mappings to kf + sf. These mappings were
// removed in Unicode 2.1.9 to ensure that Hangul syllables are maintained.)
// That is, the following test is obsolete as of 2.1.9
//obsolete- // With Full decomposition, it should go all the way down to
//obsolete- // conjoining Jamo characters.
//obsolete- //
//obsolete- static const UChar test2[][CollationRegressionTest::MAX_TOKEN_LEN] =
//obsolete- {
//obsolete- {0xd4db, 0}, {0x3d, 0}, {0x1111, 0x116e, 0x1175, 0x11af, 0x11c2, 0}
//obsolete- };
//obsolete-
//obsolete- c->setDecomposition(Normalizer::DECOMP_COMPAT);
//obsolete- compareArray(*c, test2, ARRAY_LENGTH(test2));
delete c;
}
// @bug 4066696
//
// French secondary collation checking at the end of compare iteration fails
//
void CollationRegressionTest::Test4066696(/* char* par */)
{
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = NULL;
c = (RuleBasedCollator *)Collator::createInstance(Locale::getCanadaFrench(), status);
if (c == NULL || U_FAILURE(status))
{
errln("Failure creating collator for Locale::getCanadaFrench()");
delete c;
return;
}
c->setStrength(Collator::SECONDARY);
/*
String[] tests = {
"\u00e0", "<", "\u01fa", // a-grave < A-ring-acute
};
should be:
String[] tests = {
"\u00e0", ">", "\u01fa", // a-grave < A-ring-acute
};
*/
static const UChar tests[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x00E0, 0}, {0x3e, 0}, {0x01FA, 0}
};
compareArray(*c, tests, ARRAY_LENGTH(tests));
delete c;
}
// @bug 4081866
//
// Combining characters in different classes not reordered properly.
//
void CollationRegressionTest::Test4081866(/* char* par */)
{
// These combining characters are all in different classes,
// so they should be reordered and the strings should compare as equal.
static const UChar s1[] = {0x41, 0x0300, 0x0316, 0x0327, 0x0315, 0};
static const UChar s2[] = {0x41, 0x0327, 0x0316, 0x0315, 0x0300, 0};
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::TERTIARY);
// Now that the default collators are set to NO_DECOMPOSITION
// (as a result of fixing bug 4114077), we must set it explicitly
// when we're testing reordering behavior. -- lwerner, 5/5/98
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status);
if (c->compare(s1,s2) != 0)
{
errln("Combining chars were not reordered");
}
delete c;
}
// @bug 4054734
//
// Collator::IDENTICAL documented but not implemented
//
void CollationRegressionTest::Test4054734(/* char* par */)
{
/*
Here's the original Java:
String[] decomp = {
"\u0001", "<", "\u0002",
"\u0001", "=", "\u0001",
"A\u0001", ">", "~\u0002", // Ensure A and ~ are not compared bitwise
"\u00C0", "=", "A\u0300" // Decomp should make these equal
};
String[] nodecomp = {
"\u00C0", ">", "A\u0300" // A-grave vs. A combining-grave
};
*/
static const UChar decomp[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x0001, 0}, {0x3c, 0}, {0x0002, 0},
{0x0001, 0}, {0x3d, 0}, {0x0001, 0},
{0x41, 0x0001, 0}, {0x3e, 0}, {0x7e, 0x0002, 0},
{0x00c0, 0}, {0x3d, 0}, {0x41, 0x0300, 0}
};
UErrorCode status = U_ZERO_ERROR;
RuleBasedCollator *c = (RuleBasedCollator *) en_us->clone();
c->setStrength(Collator::IDENTICAL);
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status);
compareArray(*c, decomp, ARRAY_LENGTH(decomp));
delete c;
}
// @bug 4060154
//
// MergeCollation::fixEntry broken for "& H < \u0131, \u0130, i, I"
//
void CollationRegressionTest::Test4060154(/* char* par */)
{
UErrorCode status = U_ZERO_ERROR;
UnicodeString rules;
rules += "&f < g, G < h, H < i, I < j, J";
rules += " & H < ";
rules += (UChar)0x0131;
rules += ", ";
rules += (UChar)0x0130;
rules += ", i, I";
RuleBasedCollator *c = NULL;
c = new RuleBasedCollator(rules, status);
if (c == NULL || U_FAILURE(status))
{
errln("failure building collator.");
delete c;
return;
}
c->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status);
/*
String[] tertiary = {
"A", "<", "B",
"H", "<", "\u0131",
"H", "<", "I",
"\u0131", "<", "\u0130",
"\u0130", "<", "i",
"\u0130", ">", "H",
};
*/
static const UChar tertiary[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x41, 0}, {0x3c, 0}, {0x42, 0},
{0x48, 0}, {0x3c, 0}, {0x0131, 0},
{0x48, 0}, {0x3c, 0}, {0x49, 0},
{0x0131, 0}, {0x3c, 0}, {0x0130, 0},
{0x0130, 0}, {0x3c, 0}, {0x69, 0},
{0x0130, 0}, {0x3e, 0}, {0x48, 0}
};
c->setStrength(Collator::TERTIARY);
compareArray(*c, tertiary, ARRAY_LENGTH(tertiary));
/*
String[] secondary = {
"H", "<", "I",
"\u0131", "=", "\u0130",
};
*/
static const UChar secondary[][CollationRegressionTest::MAX_TOKEN_LEN] =
{
{0x48, 0}, {0x3c, 0}, {0x49, 0},
{0x0131, 0}, {0x3d, 0}, {0x0130, 0}
};
c->setStrength(Collator::PRIMARY);
compareArray(*c, secondary, ARRAY_LENGTH(secondary));
delete c;
}
