+/*
+**********************************************************************
+* Copyright (c) 2013-2014, International Business Machines
+* Corporation and others. All Rights Reserved.
+**********************************************************************
+*/
+
+#include <string.h>
+#include "unicode/localpointer.h"
+#include "unicode/uperf.h"
+#include "unicode/ucol.h"
+#include "unicode/coll.h"
+#include "unicode/uiter.h"
+#include "unicode/ustring.h"
+#include "unicode/sortkey.h"
+#include "uarrsort.h"
+#include "uoptions.h"
+#include "ustr_imp.h"
+
+#define COMPACT_ARRAY(CompactArrays, UNIT) \
+struct CompactArrays{\
+ CompactArrays(const CompactArrays & );\
+ CompactArrays & operator=(const CompactArrays & );\
+ int32_t count;/*total number of the strings*/ \
+ int32_t * index;/*relative offset in data*/ \
+ UNIT * data; /*the real space to hold strings*/ \
+ \
+ ~CompactArrays(){free(index);free(data);} \
+ CompactArrays() : count(0), index(NULL), data(NULL) { \
+ index = (int32_t *) realloc(index, sizeof(int32_t)); \
+ index[0] = 0; \
+ } \
+ void append_one(int32_t theLen){ /*include terminal NULL*/ \
+ count++; \
+ index = (int32_t *) realloc(index, sizeof(int32_t) * (count + 1)); \
+ index[count] = index[count - 1] + theLen; \
+ data = (UNIT *) realloc(data, sizeof(UNIT) * index[count]); \
+ } \
+ UNIT * last(){return data + index[count - 1];} \
+ const UNIT * dataOf(int32_t i) const {return data + index[i];} \
+ int32_t lengthOf(int i) const {return index[i+1] - index[i] - 1; } /*exclude terminating NULL*/ \
+};
+
+COMPACT_ARRAY(CA_uchar, UChar)
+COMPACT_ARRAY(CA_char, char)
+
+#define MAX_TEST_STRINGS_FOR_PERMUTING 1000
+
+// C API test cases
+
+//
+// Test case taking a single test data array, calling ucol_strcoll by permuting the test data
+//
+class Strcoll : public UPerfFunction
+{
+public:
+ Strcoll(const UCollator* coll, const CA_uchar* source, UBool useLen);
+ ~Strcoll();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_uchar *source;
+ UBool useLen;
+ int32_t maxTestStrings;
+};
+
+Strcoll::Strcoll(const UCollator* coll, const CA_uchar* source, UBool useLen)
+ : coll(coll),
+ source(source),
+ useLen(useLen)
+{
+ maxTestStrings = source->count > MAX_TEST_STRINGS_FOR_PERMUTING ? MAX_TEST_STRINGS_FOR_PERMUTING : source->count;
+}
+
+Strcoll::~Strcoll()
+{
+}
+
+void Strcoll::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ // call strcoll for permutation
+ int32_t divisor = source->count / maxTestStrings;
+ int32_t srcLen, tgtLen;
+ int32_t cmp = 0;
+ for (int32_t i = 0, numTestStringsI = 0; i < source->count && numTestStringsI < maxTestStrings; i++) {
+ if (i % divisor) continue;
+ numTestStringsI++;
+ srcLen = useLen ? source->lengthOf(i) : -1;
+ for (int32_t j = 0, numTestStringsJ = 0; j < source->count && numTestStringsJ < maxTestStrings; j++) {
+ if (j % divisor) continue;
+ numTestStringsJ++;
+ tgtLen = useLen ? source->lengthOf(j) : -1;
+ cmp += ucol_strcoll(coll, source->dataOf(i), srcLen, source->dataOf(j), tgtLen);
+ }
+ }
+ // At the end, cmp must be 0
+ if (cmp != 0) {
+ *status = U_INTERNAL_PROGRAM_ERROR;
+ }
+}
+
+long Strcoll::getOperationsPerIteration()
+{
+ return maxTestStrings * maxTestStrings;
+}
+
+//
+// Test case taking two test data arrays, calling ucol_strcoll for strings at a same index
+//
+class Strcoll_2 : public UPerfFunction
+{
+public:
+ Strcoll_2(const UCollator* coll, const CA_uchar* source, const CA_uchar* target, UBool useLen);
+ ~Strcoll_2();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_uchar *source;
+ const CA_uchar *target;
+ UBool useLen;
+};
+
+Strcoll_2::Strcoll_2(const UCollator* coll, const CA_uchar* source, const CA_uchar* target, UBool useLen)
+ : coll(coll),
+ source(source),
+ target(target),
+ useLen(useLen)
+{
+}
+
+Strcoll_2::~Strcoll_2()
+{
+}
+
+void Strcoll_2::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ // call strcoll for two strings at the same index
+ if (source->count < target->count) {
+ *status = U_ILLEGAL_ARGUMENT_ERROR;
+ } else {
+ for (int32_t i = 0; i < source->count; i++) {
+ int32_t srcLen = useLen ? source->lengthOf(i) : -1;
+ int32_t tgtLen = useLen ? target->lengthOf(i) : -1;
+ ucol_strcoll(coll, source->dataOf(i), srcLen, target->dataOf(i), tgtLen);
+ }
+ }
+}
+
+long Strcoll_2::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+
+//
+// Test case taking a single test data array, calling ucol_strcollUTF8 by permuting the test data
+//
+class StrcollUTF8 : public UPerfFunction
+{
+public:
+ StrcollUTF8(const UCollator* coll, const CA_char* source, UBool useLen);
+ ~StrcollUTF8();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_char *source;
+ UBool useLen;
+ int32_t maxTestStrings;
+};
+
+StrcollUTF8::StrcollUTF8(const UCollator* coll, const CA_char* source, UBool useLen)
+ : coll(coll),
+ source(source),
+ useLen(useLen)
+{
+ maxTestStrings = source->count > MAX_TEST_STRINGS_FOR_PERMUTING ? MAX_TEST_STRINGS_FOR_PERMUTING : source->count;
+}
+
+StrcollUTF8::~StrcollUTF8()
+{
+}
+
+void StrcollUTF8::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ // call strcollUTF8 for permutation
+ int32_t divisor = source->count / maxTestStrings;
+ int32_t srcLen, tgtLen;
+ int32_t cmp = 0;
+ for (int32_t i = 0, numTestStringsI = 0; U_SUCCESS(*status) && i < source->count && numTestStringsI < maxTestStrings; i++) {
+ if (i % divisor) continue;
+ numTestStringsI++;
+ srcLen = useLen ? source->lengthOf(i) : -1;
+ for (int32_t j = 0, numTestStringsJ = 0; U_SUCCESS(*status) && j < source->count && numTestStringsJ < maxTestStrings; j++) {
+ if (j % divisor) continue;
+ numTestStringsJ++;
+ tgtLen = useLen ? source->lengthOf(j) : -1;
+ cmp += ucol_strcollUTF8(coll, source->dataOf(i), srcLen, source->dataOf(j), tgtLen, status);
+ }
+ }
+ // At the end, cmp must be 0
+ if (cmp != 0) {
+ *status = U_INTERNAL_PROGRAM_ERROR;
+ }
+}
+
+long StrcollUTF8::getOperationsPerIteration()
+{
+ return maxTestStrings * maxTestStrings;
+}
+
+//
+// Test case taking two test data arrays, calling ucol_strcoll for strings at a same index
+//
+class StrcollUTF8_2 : public UPerfFunction
+{
+public:
+ StrcollUTF8_2(const UCollator* coll, const CA_char* source, const CA_char* target, UBool useLen);
+ ~StrcollUTF8_2();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_char *source;
+ const CA_char *target;
+ UBool useLen;
+};
+
+StrcollUTF8_2::StrcollUTF8_2(const UCollator* coll, const CA_char* source, const CA_char* target, UBool useLen)
+ : coll(coll),
+ source(source),
+ target(target),
+ useLen(useLen)
+{
+}
+
+StrcollUTF8_2::~StrcollUTF8_2()
+{
+}
+
+void StrcollUTF8_2::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ // call strcoll for two strings at the same index
+ if (source->count < target->count) {
+ *status = U_ILLEGAL_ARGUMENT_ERROR;
+ } else {
+ for (int32_t i = 0; U_SUCCESS(*status) && i < source->count; i++) {
+ int32_t srcLen = useLen ? source->lengthOf(i) : -1;
+ int32_t tgtLen = useLen ? target->lengthOf(i) : -1;
+ ucol_strcollUTF8(coll, source->dataOf(i), srcLen, target->dataOf(i), tgtLen, status);
+ }
+ }
+}
+
+long StrcollUTF8_2::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+//
+// Test case taking a single test data array, calling ucol_getSortKey for each
+//
+class GetSortKey : public UPerfFunction
+{
+public:
+ GetSortKey(const UCollator* coll, const CA_uchar* source, UBool useLen);
+ ~GetSortKey();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_uchar *source;
+ UBool useLen;
+};
+
+GetSortKey::GetSortKey(const UCollator* coll, const CA_uchar* source, UBool useLen)
+ : coll(coll),
+ source(source),
+ useLen(useLen)
+{
+}
+
+GetSortKey::~GetSortKey()
+{
+}
+
+#define KEY_BUF_SIZE 512
+
+void GetSortKey::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ uint8_t key[KEY_BUF_SIZE];
+ int32_t len;
+
+ if (useLen) {
+ for (int32_t i = 0; i < source->count; i++) {
+ len = ucol_getSortKey(coll, source->dataOf(i), source->lengthOf(i), key, KEY_BUF_SIZE);
+ }
+ } else {
+ for (int32_t i = 0; i < source->count; i++) {
+ len = ucol_getSortKey(coll, source->dataOf(i), -1, key, KEY_BUF_SIZE);
+ }
+ }
+}
+
+long GetSortKey::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+//
+// Test case taking a single test data array in UTF-16, calling ucol_nextSortKeyPart for each for the
+// given buffer size
+//
+class NextSortKeyPart : public UPerfFunction
+{
+public:
+ NextSortKeyPart(const UCollator* coll, const CA_uchar* source, int32_t bufSize, int32_t maxIteration = -1);
+ ~NextSortKeyPart();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+ virtual long getEventsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_uchar *source;
+ int32_t bufSize;
+ int32_t maxIteration;
+ long events;
+};
+
+// Note: maxIteration = -1 -> repeat until the end of collation key
+NextSortKeyPart::NextSortKeyPart(const UCollator* coll, const CA_uchar* source, int32_t bufSize, int32_t maxIteration /* = -1 */)
+ : coll(coll),
+ source(source),
+ bufSize(bufSize),
+ maxIteration(maxIteration),
+ events(0)
+{
+}
+
+NextSortKeyPart::~NextSortKeyPart()
+{
+}
+
+void NextSortKeyPart::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ uint8_t *part = (uint8_t *)malloc(bufSize);
+ uint32_t state[2];
+ UCharIterator iter;
+
+ events = 0;
+ for (int i = 0; i < source->count && U_SUCCESS(*status); i++) {
+ uiter_setString(&iter, source->dataOf(i), source->lengthOf(i));
+ state[0] = 0;
+ state[1] = 0;
+ int32_t partLen = bufSize;
+ for (int32_t n = 0; U_SUCCESS(*status) && partLen == bufSize && (maxIteration < 0 || n < maxIteration); n++) {
+ partLen = ucol_nextSortKeyPart(coll, &iter, state, part, bufSize, status);
+ events++;
+ }
+ }
+ free(part);
+}
+
+long NextSortKeyPart::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+long NextSortKeyPart::getEventsPerIteration()
+{
+ return events;
+}
+
+//
+// Test case taking a single test data array in UTF-8, calling ucol_nextSortKeyPart for each for the
+// given buffer size
+//
+class NextSortKeyPartUTF8 : public UPerfFunction
+{
+public:
+ NextSortKeyPartUTF8(const UCollator* coll, const CA_char* source, int32_t bufSize, int32_t maxIteration = -1);
+ ~NextSortKeyPartUTF8();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+ virtual long getEventsPerIteration();
+
+private:
+ const UCollator *coll;
+ const CA_char *source;
+ int32_t bufSize;
+ int32_t maxIteration;
+ long events;
+};
+
+// Note: maxIteration = -1 -> repeat until the end of collation key
+NextSortKeyPartUTF8::NextSortKeyPartUTF8(const UCollator* coll, const CA_char* source, int32_t bufSize, int32_t maxIteration /* = -1 */)
+ : coll(coll),
+ source(source),
+ bufSize(bufSize),
+ maxIteration(maxIteration),
+ events(0)
+{
+}
+
+NextSortKeyPartUTF8::~NextSortKeyPartUTF8()
+{
+}
+
+void NextSortKeyPartUTF8::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ uint8_t *part = (uint8_t *)malloc(bufSize);
+ uint32_t state[2];
+ UCharIterator iter;
+
+ events = 0;
+ for (int i = 0; i < source->count && U_SUCCESS(*status); i++) {
+ uiter_setUTF8(&iter, source->dataOf(i), source->lengthOf(i));
+ state[0] = 0;
+ state[1] = 0;
+ int32_t partLen = bufSize;
+ for (int32_t n = 0; U_SUCCESS(*status) && partLen == bufSize && (maxIteration < 0 || n < maxIteration); n++) {
+ partLen = ucol_nextSortKeyPart(coll, &iter, state, part, bufSize, status);
+ events++;
+ }
+ }
+ free(part);
+}
+
+long NextSortKeyPartUTF8::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+long NextSortKeyPartUTF8::getEventsPerIteration()
+{
+ return events;
+}
+
+// CPP API test cases
+
+//
+// Test case taking a single test data array, calling Collator::compare by permuting the test data
+//
+class CppCompare : public UPerfFunction
+{
+public:
+ CppCompare(const Collator* coll, const CA_uchar* source, UBool useLen);
+ ~CppCompare();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const Collator *coll;
+ const CA_uchar *source;
+ UBool useLen;
+ int32_t maxTestStrings;
+};
+
+CppCompare::CppCompare(const Collator* coll, const CA_uchar* source, UBool useLen)
+ : coll(coll),
+ source(source),
+ useLen(useLen)
+{
+ maxTestStrings = source->count > MAX_TEST_STRINGS_FOR_PERMUTING ? MAX_TEST_STRINGS_FOR_PERMUTING : source->count;
+}
+
+CppCompare::~CppCompare()
+{
+}
+
+void CppCompare::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ // call compare for permutation of test data
+ int32_t divisor = source->count / maxTestStrings;
+ int32_t srcLen, tgtLen;
+ int32_t cmp = 0;
+ for (int32_t i = 0, numTestStringsI = 0; i < source->count && numTestStringsI < maxTestStrings; i++) {
+ if (i % divisor) continue;
+ numTestStringsI++;
+ srcLen = useLen ? source->lengthOf(i) : -1;
+ for (int32_t j = 0, numTestStringsJ = 0; j < source->count && numTestStringsJ < maxTestStrings; j++) {
+ if (j % divisor) continue;
+ numTestStringsJ++;
+ tgtLen = useLen ? source->lengthOf(j) : -1;
+ cmp += coll->compare(source->dataOf(i), srcLen, source->dataOf(j), tgtLen);
+ }
+ }
+ // At the end, cmp must be 0
+ if (cmp != 0) {
+ *status = U_INTERNAL_PROGRAM_ERROR;
+ }
+}
+
+long CppCompare::getOperationsPerIteration()
+{
+ return maxTestStrings * maxTestStrings;
+}
+
+//
+// Test case taking two test data arrays, calling Collator::compare for strings at a same index
+//
+class CppCompare_2 : public UPerfFunction
+{
+public:
+ CppCompare_2(const Collator* coll, const CA_uchar* source, const CA_uchar* target, UBool useLen);
+ ~CppCompare_2();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const Collator *coll;
+ const CA_uchar *source;
+ const CA_uchar *target;
+ UBool useLen;
+};
+
+CppCompare_2::CppCompare_2(const Collator* coll, const CA_uchar* source, const CA_uchar* target, UBool useLen)
+ : coll(coll),
+ source(source),
+ target(target),
+ useLen(useLen)
+{
+}
+
+CppCompare_2::~CppCompare_2()
+{
+}
+
+void CppCompare_2::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ // call strcoll for two strings at the same index
+ if (source->count < target->count) {
+ *status = U_ILLEGAL_ARGUMENT_ERROR;
+ } else {
+ for (int32_t i = 0; i < source->count; i++) {
+ int32_t srcLen = useLen ? source->lengthOf(i) : -1;
+ int32_t tgtLen = useLen ? target->lengthOf(i) : -1;
+ coll->compare(source->dataOf(i), srcLen, target->dataOf(i), tgtLen);
+ }
+ }
+}
+
+long CppCompare_2::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+
+//
+// Test case taking a single test data array, calling Collator::compareUTF8 by permuting the test data
+//
+class CppCompareUTF8 : public UPerfFunction
+{
+public:
+ CppCompareUTF8(const Collator* coll, const CA_char* source, UBool useLen);
+ ~CppCompareUTF8();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const Collator *coll;
+ const CA_char *source;
+ UBool useLen;
+ int32_t maxTestStrings;
+};
+
+CppCompareUTF8::CppCompareUTF8(const Collator* coll, const CA_char* source, UBool useLen)
+ : coll(coll),
+ source(source),
+ useLen(useLen)
+{
+ maxTestStrings = source->count > MAX_TEST_STRINGS_FOR_PERMUTING ? MAX_TEST_STRINGS_FOR_PERMUTING : source->count;
+}
+
+CppCompareUTF8::~CppCompareUTF8()
+{
+}
+
+void CppCompareUTF8::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ // call compareUTF8 for all permutations
+ int32_t divisor = source->count / maxTestStrings;
+ StringPiece src, tgt;
+ int32_t cmp = 0;
+ for (int32_t i = 0, numTestStringsI = 0; U_SUCCESS(*status) && i < source->count && numTestStringsI < maxTestStrings; i++) {
+ if (i % divisor) continue;
+ numTestStringsI++;
+
+ if (useLen) {
+ src.set(source->dataOf(i), source->lengthOf(i));
+ } else {
+ src.set(source->dataOf(i));
+ }
+ for (int32_t j = 0, numTestStringsJ = 0; U_SUCCESS(*status) && j < source->count && numTestStringsJ < maxTestStrings; j++) {
+ if (j % divisor) continue;
+ numTestStringsJ++;
+
+ if (useLen) {
+ tgt.set(source->dataOf(i), source->lengthOf(i));
+ } else {
+ tgt.set(source->dataOf(i));
+ }
+ cmp += coll->compareUTF8(src, tgt, *status);
+ }
+ }
+ // At the end, cmp must be 0
+ if (cmp != 0) {
+ *status = U_INTERNAL_PROGRAM_ERROR;
+ }
+}
+
+long CppCompareUTF8::getOperationsPerIteration()
+{
+ return maxTestStrings * maxTestStrings;
+}
+
+
+//
+// Test case taking two test data arrays, calling Collator::compareUTF8 for strings at a same index
+//
+class CppCompareUTF8_2 : public UPerfFunction
+{
+public:
+ CppCompareUTF8_2(const Collator* coll, const CA_char* source, const CA_char* target, UBool useLen);
+ ~CppCompareUTF8_2();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const Collator *coll;
+ const CA_char *source;
+ const CA_char *target;
+ UBool useLen;
+};
+
+CppCompareUTF8_2::CppCompareUTF8_2(const Collator* coll, const CA_char* source, const CA_char* target, UBool useLen)
+ : coll(coll),
+ source(source),
+ target(target),
+ useLen(useLen)
+{
+}
+
+CppCompareUTF8_2::~CppCompareUTF8_2()
+{
+}
+
+void CppCompareUTF8_2::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ // call strcoll for two strings at the same index
+ StringPiece src, tgt;
+ if (source->count < target->count) {
+ *status = U_ILLEGAL_ARGUMENT_ERROR;
+ } else {
+ for (int32_t i = 0; U_SUCCESS(*status) && i < source->count; i++) {
+ if (useLen) {
+ src.set(source->dataOf(i), source->lengthOf(i));
+ tgt.set(target->dataOf(i), target->lengthOf(i));
+ } else {
+ src.set(source->dataOf(i));
+ tgt.set(target->dataOf(i));
+ }
+ coll->compareUTF8(src, tgt, *status);
+ }
+ }
+}
+
+long CppCompareUTF8_2::getOperationsPerIteration()
+{
+ return source->count;
+}
+
+
+//
+// Test case taking a single test data array, calling Collator::getCollationKey for each
+//
+class CppGetCollationKey : public UPerfFunction
+{
+public:
+ CppGetCollationKey(const Collator* coll, const CA_uchar* source, UBool useLen);
+ ~CppGetCollationKey();
+ virtual void call(UErrorCode* status);
+ virtual long getOperationsPerIteration();
+
+private:
+ const Collator *coll;
+ const CA_uchar *source;
+ UBool useLen;
+};
+
+CppGetCollationKey::CppGetCollationKey(const Collator* coll, const CA_uchar* source, UBool useLen)
+ : coll(coll),
+ source(source),
+ useLen(useLen)
+{
+}
+
+CppGetCollationKey::~CppGetCollationKey()
+{
+}
+
+void CppGetCollationKey::call(UErrorCode* status)
+{
+ if (U_FAILURE(*status)) return;
+
+ CollationKey key;
+ for (int32_t i = 0; U_SUCCESS(*status) && i < source->count; i++) {
+ coll->getCollationKey(source->dataOf(i), source->lengthOf(i), key, *status);
+ }
+}
+
+long CppGetCollationKey::getOperationsPerIteration() {
+ return source->count;
+}
+
+namespace {
+
+struct CollatorAndCounter {
+ CollatorAndCounter(const Collator& coll) : coll(coll), ucoll(NULL), counter(0) {}
+ CollatorAndCounter(const Collator& coll, const UCollator *ucoll)
+ : coll(coll), ucoll(ucoll), counter(0) {}
+ const Collator& coll;
+ const UCollator *ucoll;
+ int32_t counter;
+};
+
+int32_t U_CALLCONV
+UniStrCollatorComparator(const void* context, const void* left, const void* right) {
+ CollatorAndCounter& cc = *(CollatorAndCounter*)context;
+ const UnicodeString& leftString = **(const UnicodeString**)left;
+ const UnicodeString& rightString = **(const UnicodeString**)right;
+ UErrorCode errorCode = U_ZERO_ERROR;
+ ++cc.counter;
+ return cc.coll.compare(leftString, rightString, errorCode);
+}
+
+} // namespace
+
+class CollPerfFunction : public UPerfFunction {
+public:
+ CollPerfFunction(const Collator& coll, const UCollator *ucoll)
+ : coll(coll), ucoll(ucoll), ops(0) {}
+ virtual ~CollPerfFunction();
+ /** Calls call() to set the ops field, and returns that. */
+ virtual long getOperationsPerIteration();
+
+protected:
+ const Collator& coll;
+ const UCollator *ucoll;
+ int32_t ops;
+};
+
+CollPerfFunction::~CollPerfFunction() {}
+
+long CollPerfFunction::getOperationsPerIteration() {
+ UErrorCode errorCode = U_ZERO_ERROR;
+ call(&errorCode);
+ return U_SUCCESS(errorCode) ? ops : 0;
+}
+
+class UniStrCollPerfFunction : public CollPerfFunction {
+public:
+ UniStrCollPerfFunction(const Collator& coll, const UCollator *ucoll, const CA_uchar* data16)
+ : CollPerfFunction(coll, ucoll), d16(data16),
+ source(new UnicodeString*[d16->count]) {
+ for (int32_t i = 0; i < d16->count; ++i) {
+ source[i] = new UnicodeString(TRUE, d16->dataOf(i), d16->lengthOf(i));
+ }
+ }
+ virtual ~UniStrCollPerfFunction();
+
+protected:
+ const CA_uchar* d16;
+ UnicodeString** source;
+};
+
+UniStrCollPerfFunction::~UniStrCollPerfFunction() {
+ for (int32_t i = 0; i < d16->count; ++i) {
+ delete source[i];
+ }
+ delete[] source;
+}
+
+//
+// Test case sorting an array of UnicodeString pointers.
+//
+class UniStrSort : public UniStrCollPerfFunction {
+public:
+ UniStrSort(const Collator& coll, const UCollator *ucoll, const CA_uchar* data16)
+ : UniStrCollPerfFunction(coll, ucoll, data16),
+ dest(new UnicodeString*[d16->count]) {}
+ virtual ~UniStrSort();
+ virtual void call(UErrorCode* status);
+
+private:
+ UnicodeString** dest; // aliases only
+};
+
+UniStrSort::~UniStrSort() {
+ delete[] dest;
+}
+
+void UniStrSort::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ CollatorAndCounter cc(coll);
+ int32_t count = d16->count;
+ memcpy(dest, source, count * sizeof(UnicodeString *));
+ uprv_sortArray(dest, count, (int32_t)sizeof(UnicodeString *),
+ UniStrCollatorComparator, &cc, TRUE, status);
+ ops = cc.counter;
+}
+
+namespace {
+
+int32_t U_CALLCONV
+StringPieceCollatorComparator(const void* context, const void* left, const void* right) {
+ CollatorAndCounter& cc = *(CollatorAndCounter*)context;
+ const StringPiece& leftString = *(const StringPiece*)left;
+ const StringPiece& rightString = *(const StringPiece*)right;
+ UErrorCode errorCode = U_ZERO_ERROR;
+ ++cc.counter;
+ return cc.coll.compareUTF8(leftString, rightString, errorCode);
+}
+
+int32_t U_CALLCONV
+StringPieceUCollatorComparator(const void* context, const void* left, const void* right) {
+ CollatorAndCounter& cc = *(CollatorAndCounter*)context;
+ const StringPiece& leftString = *(const StringPiece*)left;
+ const StringPiece& rightString = *(const StringPiece*)right;
+ UErrorCode errorCode = U_ZERO_ERROR;
+ ++cc.counter;
+ return ucol_strcollUTF8(cc.ucoll,
+ leftString.data(), leftString.length(),
+ rightString.data(), rightString.length(), &errorCode);
+}
+
+} // namespace
+
+class StringPieceCollPerfFunction : public CollPerfFunction {
+public:
+ StringPieceCollPerfFunction(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : CollPerfFunction(coll, ucoll), d8(data8),
+ source(new StringPiece[d8->count]) {
+ for (int32_t i = 0; i < d8->count; ++i) {
+ source[i].set(d8->dataOf(i), d8->lengthOf(i));
+ }
+ }
+ virtual ~StringPieceCollPerfFunction();
+
+protected:
+ const CA_char* d8;
+ StringPiece* source;
+};
+
+StringPieceCollPerfFunction::~StringPieceCollPerfFunction() {
+ delete[] source;
+}
+
+class StringPieceSort : public StringPieceCollPerfFunction {
+public:
+ StringPieceSort(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : StringPieceCollPerfFunction(coll, ucoll, data8),
+ dest(new StringPiece[d8->count]) {}
+ virtual ~StringPieceSort();
+
+protected:
+ StringPiece* dest;
+};
+
+StringPieceSort::~StringPieceSort() {
+ delete[] dest;
+}
+
+//
+// Test case sorting an array of UTF-8 StringPiece's with Collator::compareUTF8().
+//
+class StringPieceSortCpp : public StringPieceSort {
+public:
+ StringPieceSortCpp(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : StringPieceSort(coll, ucoll, data8) {}
+ virtual ~StringPieceSortCpp();
+ virtual void call(UErrorCode* status);
+};
+
+StringPieceSortCpp::~StringPieceSortCpp() {}
+
+void StringPieceSortCpp::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ CollatorAndCounter cc(coll);
+ int32_t count = d8->count;
+ memcpy(dest, source, count * sizeof(StringPiece));
+ uprv_sortArray(dest, count, (int32_t)sizeof(StringPiece),
+ StringPieceCollatorComparator, &cc, TRUE, status);
+ ops = cc.counter;
+}
+
+//
+// Test case sorting an array of UTF-8 StringPiece's with ucol_strcollUTF8().
+//
+class StringPieceSortC : public StringPieceSort {
+public:
+ StringPieceSortC(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : StringPieceSort(coll, ucoll, data8) {}
+ virtual ~StringPieceSortC();
+ virtual void call(UErrorCode* status);
+};
+
+StringPieceSortC::~StringPieceSortC() {}
+
+void StringPieceSortC::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ CollatorAndCounter cc(coll, ucoll);
+ int32_t count = d8->count;
+ memcpy(dest, source, count * sizeof(StringPiece));
+ uprv_sortArray(dest, count, (int32_t)sizeof(StringPiece),
+ StringPieceUCollatorComparator, &cc, TRUE, status);
+ ops = cc.counter;
+}
+
+//
+// Test case performing binary searches in a sorted array of UnicodeString pointers.
+//
+class UniStrBinSearch : public UniStrCollPerfFunction {
+public:
+ UniStrBinSearch(const Collator& coll, const UCollator *ucoll, const CA_uchar* data16)
+ : UniStrCollPerfFunction(coll, ucoll, data16) {}
+ virtual ~UniStrBinSearch();
+ virtual void call(UErrorCode* status);
+};
+
+UniStrBinSearch::~UniStrBinSearch() {}
+
+void UniStrBinSearch::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ CollatorAndCounter cc(coll);
+ int32_t count = d16->count;
+ for (int32_t i = 0; i < count; ++i) {
+ (void)uprv_stableBinarySearch((char *)source, count,
+ source + i, (int32_t)sizeof(UnicodeString *),
+ UniStrCollatorComparator, &cc);
+ }
+ ops = cc.counter;
+}
+
+class StringPieceBinSearch : public StringPieceCollPerfFunction {
+public:
+ StringPieceBinSearch(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : StringPieceCollPerfFunction(coll, ucoll, data8) {}
+ virtual ~StringPieceBinSearch();
+};
+
+StringPieceBinSearch::~StringPieceBinSearch() {}
+
+//
+// Test case performing binary searches in a sorted array of UTF-8 StringPiece's
+// with Collator::compareUTF8().
+//
+class StringPieceBinSearchCpp : public StringPieceBinSearch {
+public:
+ StringPieceBinSearchCpp(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : StringPieceBinSearch(coll, ucoll, data8) {}
+ virtual ~StringPieceBinSearchCpp();
+ virtual void call(UErrorCode* status);
+};
+
+StringPieceBinSearchCpp::~StringPieceBinSearchCpp() {}
+
+void StringPieceBinSearchCpp::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ CollatorAndCounter cc(coll);
+ int32_t count = d8->count;
+ for (int32_t i = 0; i < count; ++i) {
+ (void)uprv_stableBinarySearch((char *)source, count,
+ source + i, (int32_t)sizeof(StringPiece),
+ StringPieceCollatorComparator, &cc);
+ }
+ ops = cc.counter;
+}
+
+//
+// Test case performing binary searches in a sorted array of UTF-8 StringPiece's
+// with ucol_strcollUTF8().
+//
+class StringPieceBinSearchC : public StringPieceBinSearch {
+public:
+ StringPieceBinSearchC(const Collator& coll, const UCollator *ucoll, const CA_char* data8)
+ : StringPieceBinSearch(coll, ucoll, data8) {}
+ virtual ~StringPieceBinSearchC();
+ virtual void call(UErrorCode* status);
+};
+
+StringPieceBinSearchC::~StringPieceBinSearchC() {}
+
+void StringPieceBinSearchC::call(UErrorCode* status) {
+ if (U_FAILURE(*status)) return;
+
+ CollatorAndCounter cc(coll, ucoll);
+ int32_t count = d8->count;
+ for (int32_t i = 0; i < count; ++i) {
+ (void)uprv_stableBinarySearch((char *)source, count,
+ source + i, (int32_t)sizeof(StringPiece),
+ StringPieceUCollatorComparator, &cc);
+ }
+ ops = cc.counter;
+}
+
+
+class CollPerf2Test : public UPerfTest
+{
+public:
+ CollPerf2Test(int32_t argc, const char *argv[], UErrorCode &status);
+ ~CollPerf2Test();
+ virtual UPerfFunction* runIndexedTest(
+ int32_t index, UBool exec, const char *&name, char *par = NULL);
+
+private:
+ UCollator* coll;
+ Collator* collObj;
+
+ int32_t count;
+ CA_uchar* data16;
+ CA_char* data8;
+
+ CA_uchar* modData16;
+ CA_char* modData8;
+
+ CA_uchar* sortedData16;
+ CA_char* sortedData8;
+
+ CA_uchar* randomData16;
+ CA_char* randomData8;
+
+ const CA_uchar* getData16(UErrorCode &status);
+ const CA_char* getData8(UErrorCode &status);
+
+ const CA_uchar* getModData16(UErrorCode &status);
+ const CA_char* getModData8(UErrorCode &status);
+
+ const CA_uchar* getSortedData16(UErrorCode &status);
+ const CA_char* getSortedData8(UErrorCode &status);
+
+ const CA_uchar* getRandomData16(UErrorCode &status);
+ const CA_char* getRandomData8(UErrorCode &status);
+
+ static CA_uchar* sortData16(
+ const CA_uchar* d16,
+ UComparator *cmp, const void *context,
+ UErrorCode &status);
+ static CA_char* getData8FromData16(const CA_uchar* d16, UErrorCode &status);
+
+ UPerfFunction* TestStrcoll();
+ UPerfFunction* TestStrcollNull();
+ UPerfFunction* TestStrcollSimilar();
+
+ UPerfFunction* TestStrcollUTF8();
+ UPerfFunction* TestStrcollUTF8Null();
+ UPerfFunction* TestStrcollUTF8Similar();
+
+ UPerfFunction* TestGetSortKey();
+ UPerfFunction* TestGetSortKeyNull();
+
+ UPerfFunction* TestNextSortKeyPart_4All();
+ UPerfFunction* TestNextSortKeyPart_4x2();
+ UPerfFunction* TestNextSortKeyPart_4x4();
+ UPerfFunction* TestNextSortKeyPart_4x8();
+ UPerfFunction* TestNextSortKeyPart_32All();
+ UPerfFunction* TestNextSortKeyPart_32x2();
+
+ UPerfFunction* TestNextSortKeyPartUTF8_4All();
+ UPerfFunction* TestNextSortKeyPartUTF8_4x2();
+ UPerfFunction* TestNextSortKeyPartUTF8_4x4();
+ UPerfFunction* TestNextSortKeyPartUTF8_4x8();
+ UPerfFunction* TestNextSortKeyPartUTF8_32All();
+ UPerfFunction* TestNextSortKeyPartUTF8_32x2();
+
+ UPerfFunction* TestCppCompare();
+ UPerfFunction* TestCppCompareNull();
+ UPerfFunction* TestCppCompareSimilar();
+
+ UPerfFunction* TestCppCompareUTF8();
+ UPerfFunction* TestCppCompareUTF8Null();
+ UPerfFunction* TestCppCompareUTF8Similar();
+
+ UPerfFunction* TestCppGetCollationKey();
+ UPerfFunction* TestCppGetCollationKeyNull();
+
+ UPerfFunction* TestUniStrSort();
+ UPerfFunction* TestStringPieceSortCpp();
+ UPerfFunction* TestStringPieceSortC();
+
+ UPerfFunction* TestUniStrBinSearch();
+ UPerfFunction* TestStringPieceBinSearchCpp();
+ UPerfFunction* TestStringPieceBinSearchC();
+};
+
+CollPerf2Test::CollPerf2Test(int32_t argc, const char *argv[], UErrorCode &status) :
+ UPerfTest(argc, argv, status),
+ coll(NULL),
+ collObj(NULL),
+ count(0),
+ data16(NULL),
+ data8(NULL),
+ modData16(NULL),
+ modData8(NULL),
+ sortedData16(NULL),
+ sortedData8(NULL),
+ randomData16(NULL),
+ randomData8(NULL)
+{
+ if (U_FAILURE(status)) {
+ return;
+ }
+
+ if (locale == NULL){
+ locale = "en_US"; // set default locale
+ }
+
+ // Set up an ICU collator
+ coll = ucol_open(locale, &status);
+ collObj = Collator::createInstance(locale, status);
+
+ // Keyword support should be actually a part of ICU collator, see ICU ticket #8260.
+ char keyBuffer[256];
+ UColAttributeValue val;
+ if (uloc_getKeywordValue(locale, "strength", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "primary") == 0) {
+ val = UCOL_PRIMARY;
+ } else if (strcmp(keyBuffer, "secondary") == 0) {
+ val = UCOL_SECONDARY;
+ } else if (strcmp(keyBuffer, "tertiary") == 0) {
+ val = UCOL_TERTIARY;
+ } else if (strcmp(keyBuffer, "quaternary") == 0) {
+ val = UCOL_QUATERNARY;
+ } else if (strcmp(keyBuffer, "identical") == 0) {
+ val = UCOL_IDENTICAL;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_STRENGTH, val, &status);
+ collObj->setAttribute(UCOL_STRENGTH, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "alternate", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "non-ignorable") == 0) {
+ val = UCOL_NON_IGNORABLE;
+ } else if (strcmp(keyBuffer, "shifted") == 0) {
+ val = UCOL_SHIFTED;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_ALTERNATE_HANDLING, val, &status);
+ collObj->setAttribute(UCOL_ALTERNATE_HANDLING, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "backwards", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "on") == 0) {
+ val = UCOL_ON;
+ } else if (strcmp(keyBuffer, "off") == 0) {
+ val = UCOL_OFF;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_FRENCH_COLLATION, val, &status);
+ collObj->setAttribute(UCOL_FRENCH_COLLATION, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "normalization", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "on") == 0) {
+ val = UCOL_ON;
+ } else if (strcmp(keyBuffer, "off") == 0) {
+ val = UCOL_OFF;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_NORMALIZATION_MODE, val, &status);
+ collObj->setAttribute(UCOL_NORMALIZATION_MODE, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "caseLevel", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "on") == 0) {
+ val = UCOL_ON;
+ } else if (strcmp(keyBuffer, "off") == 0) {
+ val = UCOL_OFF;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_CASE_LEVEL, val, &status);
+ collObj->setAttribute(UCOL_CASE_LEVEL, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "caseFirst", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "upper") == 0) {
+ val = UCOL_UPPER_FIRST;
+ } else if (strcmp(keyBuffer, "lower") == 0) {
+ val = UCOL_LOWER_FIRST;
+ } else if (strcmp(keyBuffer, "off") == 0) {
+ val = UCOL_OFF;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_CASE_FIRST, val, &status);
+ collObj->setAttribute(UCOL_CASE_FIRST, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "hiraganaQuaternary", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "on") == 0) {
+ val = UCOL_ON;
+ } else if (strcmp(keyBuffer, "off") == 0) {
+ val = UCOL_OFF;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_HIRAGANA_QUATERNARY_MODE, val, &status);
+ collObj->setAttribute(UCOL_HIRAGANA_QUATERNARY_MODE, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "numeric", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ if (strcmp(keyBuffer, "on") == 0) {
+ val = UCOL_ON;
+ } else if (strcmp(keyBuffer, "off") == 0) {
+ val = UCOL_OFF;
+ } else {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ }
+ if (U_SUCCESS(status)) {
+ ucol_setAttribute(coll, UCOL_NUMERIC_COLLATION, val, &status);
+ collObj->setAttribute(UCOL_NUMERIC_COLLATION, val, status);
+ }
+ }
+ if (uloc_getKeywordValue(locale, "variableTop", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ // no support for now
+ status = U_UNSUPPORTED_ERROR;
+ }
+ if (uloc_getKeywordValue(locale, "reorder", keyBuffer, sizeof(keyBuffer)/sizeof(keyBuffer[0]), &status)) {
+ // no support for now
+ status = U_UNSUPPORTED_ERROR;
+ }
+}
+
+CollPerf2Test::~CollPerf2Test()
+{
+ ucol_close(coll);
+ delete collObj;
+
+ delete data16;
+ delete data8;
+ delete modData16;
+ delete modData8;
+ delete sortedData16;
+ delete sortedData8;
+ delete randomData16;
+ delete randomData8;
+}
+
+#define MAX_NUM_DATA 10000
+
+const CA_uchar* CollPerf2Test::getData16(UErrorCode &status)
+{
+ if (U_FAILURE(status)) return NULL;
+ if (data16) return data16;
+
+ CA_uchar* d16 = new CA_uchar();
+ const UChar *line = NULL;
+ int32_t len = 0;
+ int32_t numData = 0;
+
+ for (;;) {
+ line = ucbuf_readline(ucharBuf, &len, &status);
+ if (line == NULL || U_FAILURE(status)) break;
+
+ // Refer to the source code of ucbuf_readline()
+ // 1. 'len' includes the line terminal symbols
+ // 2. The length of the line terminal symbols is only one character
+ // 3. The Windows CR LF line terminal symbols will be converted to CR
+
+ if (len == 1 || line[0] == 0x23 /* '#' */) {
+ continue; // skip empty/comment line
+ } else {
+ d16->append_one(len);
+ u_memcpy(d16->last(), line, len);
+
+ numData++;
+ if (numData >= MAX_NUM_DATA) break;
+ }
+ }
+
+ if (U_SUCCESS(status)) {
+ data16 = d16;
+ } else {
+ delete d16;
+ }
+
+ return data16;
+}
+
+const CA_char* CollPerf2Test::getData8(UErrorCode &status)
+{
+ if (U_FAILURE(status)) return NULL;
+ if (data8) return data8;
+ return data8 = getData8FromData16(getData16(status), status);
+}
+
+const CA_uchar* CollPerf2Test::getModData16(UErrorCode &status)
+{
+ if (U_FAILURE(status)) return NULL;
+ if (modData16) return modData16;
+
+ const CA_uchar* d16 = getData16(status);
+ if (U_FAILURE(status)) return NULL;
+
+ CA_uchar* modData16 = new CA_uchar();
+
+ for (int32_t i = 0; i < d16->count; i++) {
+ const UChar *s = d16->dataOf(i);
+ int32_t len = d16->lengthOf(i) + 1; // including NULL terminator
+
+ modData16->append_one(len);
+ u_memcpy(modData16->last(), s, len);
+
+ // replacing the last character with a different character
+ UChar *lastChar = &modData16->last()[len -2];
+ for (int32_t j = i + 1; j != i; j++) {
+ if (j >= d16->count) {
+ j = 0;
+ }
+ const UChar *s1 = d16->dataOf(j);
+ UChar lastChar1 = s1[d16->lengthOf(j) - 1];
+ if (*lastChar != lastChar1) {
+ *lastChar = lastChar1;
+ break;
+ }
+ }
+ }
+
+ return modData16;
+}
+
+const CA_char* CollPerf2Test::getModData8(UErrorCode &status)
+{
+ if (U_FAILURE(status)) return NULL;
+ if (modData8) return modData8;
+ return modData8 = getData8FromData16(getModData16(status), status);
+}
+
+namespace {
+
+struct ArrayAndColl {
+ ArrayAndColl(const CA_uchar* a, const Collator& c) : d16(a), coll(c) {}
+ const CA_uchar* d16;
+ const Collator& coll;
+};
+
+int32_t U_CALLCONV
+U16CollatorComparator(const void* context, const void* left, const void* right) {
+ const ArrayAndColl& ac = *(const ArrayAndColl*)context;
+ const CA_uchar* d16 = ac.d16;
+ int32_t leftIndex = *(const int32_t*)left;
+ int32_t rightIndex = *(const int32_t*)right;
+ UErrorCode errorCode = U_ZERO_ERROR;
+ return ac.coll.compare(d16->dataOf(leftIndex), d16->lengthOf(leftIndex),
+ d16->dataOf(rightIndex), d16->lengthOf(rightIndex),
+ errorCode);
+}
+
+int32_t U_CALLCONV
+U16HashComparator(const void* context, const void* left, const void* right) {
+ const CA_uchar* d16 = (const CA_uchar*)context;
+ int32_t leftIndex = *(const int32_t*)left;
+ int32_t rightIndex = *(const int32_t*)right;
+ int32_t leftHash = ustr_hashUCharsN(d16->dataOf(leftIndex), d16->lengthOf(leftIndex));
+ int32_t rightHash = ustr_hashUCharsN(d16->dataOf(rightIndex), d16->lengthOf(rightIndex));
+ return leftHash < rightHash ? -1 : leftHash == rightHash ? 0 : 1;
+}
+
+} // namespace
+
+const CA_uchar* CollPerf2Test::getSortedData16(UErrorCode &status) {
+ if (U_FAILURE(status)) return NULL;
+ if (sortedData16) return sortedData16;
+
+ ArrayAndColl ac(getData16(status), *collObj);
+ return sortedData16 = sortData16(ac.d16, U16CollatorComparator, &ac, status);
+}
+
+const CA_char* CollPerf2Test::getSortedData8(UErrorCode &status) {
+ if (U_FAILURE(status)) return NULL;
+ if (sortedData8) return sortedData8;
+ return sortedData8 = getData8FromData16(getSortedData16(status), status);
+}
+
+const CA_uchar* CollPerf2Test::getRandomData16(UErrorCode &status) {
+ if (U_FAILURE(status)) return NULL;
+ if (randomData16) return randomData16;
+
+ // Sort the strings by their hash codes, which should be a reasonably pseudo-random order.
+ const CA_uchar* d16 = getData16(status);
+ return randomData16 = sortData16(d16, U16HashComparator, d16, status);
+}
+
+const CA_char* CollPerf2Test::getRandomData8(UErrorCode &status) {
+ if (U_FAILURE(status)) return NULL;
+ if (randomData8) return randomData8;
+ return randomData8 = getData8FromData16(getRandomData16(status), status);
+}
+
+CA_uchar* CollPerf2Test::sortData16(const CA_uchar* d16,
+ UComparator *cmp, const void *context,
+ UErrorCode &status) {
+ if (U_FAILURE(status)) return NULL;
+
+ LocalArray<int32_t> indexes(new int32_t[d16->count]);
+ for (int32_t i = 0; i < d16->count; ++i) {
+ indexes[i] = i;
+ }
+ uprv_sortArray(indexes.getAlias(), d16->count, 4, cmp, context, TRUE, &status);
+ if (U_FAILURE(status)) return NULL;
+
+ // Copy the strings in sorted order into a new array.
+ LocalPointer<CA_uchar> newD16(new CA_uchar());
+ for (int32_t i = 0; i < d16->count; i++) {
+ const UChar* s = d16->dataOf(i);
+ int32_t len = d16->lengthOf(i);
+ int32_t capacity = len + 1; // including NULL terminator
+ newD16->append_one(capacity);
+ u_memcpy(newD16->last(), s, capacity);
+ }
+
+ if (U_SUCCESS(status)) {
+ return newD16.orphan();
+ } else {
+ return NULL;
+ }
+}
+
+CA_char* CollPerf2Test::getData8FromData16(const CA_uchar* d16, UErrorCode &status) {
+ if (U_FAILURE(status)) return NULL;
+
+ // UTF-16 -> UTF-8 conversion
+ LocalPointer<CA_char> d8(new CA_char());
+ for (int32_t i = 0; i < d16->count; i++) {
+ const UChar *s16 = d16->dataOf(i);
+ int32_t length16 = d16->lengthOf(i);
+
+ // get length in UTF-8
+ int32_t length8;
+ u_strToUTF8(NULL, 0, &length8, s16, length16, &status);
+ if (status == U_BUFFER_OVERFLOW_ERROR || status == U_ZERO_ERROR){
+ status = U_ZERO_ERROR;
+ } else {
+ break;
+ }
+ int32_t capacity8 = length8 + 1; // plus terminal NULL
+ d8->append_one(capacity8);
+
+ // convert to UTF-8
+ u_strToUTF8(d8->last(), capacity8, NULL, s16, length16, &status);
+ if (U_FAILURE(status)) break;
+ }
+
+ if (U_SUCCESS(status)) {
+ return d8.orphan();
+ } else {
+ return NULL;
+ }
+}
+
+UPerfFunction*
+CollPerf2Test::runIndexedTest(int32_t index, UBool exec, const char *&name, char *par /*= NULL*/)
+{
+ (void)par;
+ TESTCASE_AUTO_BEGIN;
+
+ TESTCASE_AUTO(TestStrcoll);
+ TESTCASE_AUTO(TestStrcollNull);
+ TESTCASE_AUTO(TestStrcollSimilar);
+
+ TESTCASE_AUTO(TestStrcollUTF8);
+ TESTCASE_AUTO(TestStrcollUTF8Null);
+ TESTCASE_AUTO(TestStrcollUTF8Similar);
+
+ TESTCASE_AUTO(TestGetSortKey);
+ TESTCASE_AUTO(TestGetSortKeyNull);
+
+ TESTCASE_AUTO(TestNextSortKeyPart_4All);
+ TESTCASE_AUTO(TestNextSortKeyPart_4x4);
+ TESTCASE_AUTO(TestNextSortKeyPart_4x8);
+ TESTCASE_AUTO(TestNextSortKeyPart_32All);
+ TESTCASE_AUTO(TestNextSortKeyPart_32x2);
+
+ TESTCASE_AUTO(TestNextSortKeyPartUTF8_4All);
+ TESTCASE_AUTO(TestNextSortKeyPartUTF8_4x4);
+ TESTCASE_AUTO(TestNextSortKeyPartUTF8_4x8);
+ TESTCASE_AUTO(TestNextSortKeyPartUTF8_32All);
+ TESTCASE_AUTO(TestNextSortKeyPartUTF8_32x2);
+
+ TESTCASE_AUTO(TestCppCompare);
+ TESTCASE_AUTO(TestCppCompareNull);
+ TESTCASE_AUTO(TestCppCompareSimilar);
+
+ TESTCASE_AUTO(TestCppCompareUTF8);
+ TESTCASE_AUTO(TestCppCompareUTF8Null);
+ TESTCASE_AUTO(TestCppCompareUTF8Similar);
+
+ TESTCASE_AUTO(TestCppGetCollationKey);
+ TESTCASE_AUTO(TestCppGetCollationKeyNull);
+
+ TESTCASE_AUTO(TestUniStrSort);
+ TESTCASE_AUTO(TestStringPieceSortCpp);
+ TESTCASE_AUTO(TestStringPieceSortC);
+
+ TESTCASE_AUTO(TestUniStrBinSearch);
+ TESTCASE_AUTO(TestStringPieceBinSearchCpp);
+ TESTCASE_AUTO(TestStringPieceBinSearchC);
+
+ TESTCASE_AUTO_END;
+ return NULL;
+}
+
+
+
+UPerfFunction* CollPerf2Test::TestStrcoll()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ Strcoll *testCase = new Strcoll(coll, getData16(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStrcollNull()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ Strcoll *testCase = new Strcoll(coll, getData16(status), FALSE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStrcollSimilar()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ Strcoll_2 *testCase = new Strcoll_2(coll, getData16(status), getModData16(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStrcollUTF8()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ StrcollUTF8 *testCase = new StrcollUTF8(coll, getData8(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStrcollUTF8Null()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ StrcollUTF8 *testCase = new StrcollUTF8(coll, getData8(status),FALSE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStrcollUTF8Similar()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ StrcollUTF8_2 *testCase = new StrcollUTF8_2(coll, getData8(status), getModData8(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestGetSortKey()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ GetSortKey *testCase = new GetSortKey(coll, getData16(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestGetSortKeyNull()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ GetSortKey *testCase = new GetSortKey(coll, getData16(status), FALSE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPart_4All()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPart *testCase = new NextSortKeyPart(coll, getData16(status), 4 /* bufSize */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPart_4x4()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPart *testCase = new NextSortKeyPart(coll, getData16(status), 4 /* bufSize */, 4 /* maxIteration */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPart_4x8()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPart *testCase = new NextSortKeyPart(coll, getData16(status), 4 /* bufSize */, 8 /* maxIteration */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPart_32All()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPart *testCase = new NextSortKeyPart(coll, getData16(status), 32 /* bufSize */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPart_32x2()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPart *testCase = new NextSortKeyPart(coll, getData16(status), 32 /* bufSize */, 2 /* maxIteration */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPartUTF8_4All()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPartUTF8 *testCase = new NextSortKeyPartUTF8(coll, getData8(status), 4 /* bufSize */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPartUTF8_4x4()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPartUTF8 *testCase = new NextSortKeyPartUTF8(coll, getData8(status), 4 /* bufSize */, 4 /* maxIteration */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPartUTF8_4x8()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPartUTF8 *testCase = new NextSortKeyPartUTF8(coll, getData8(status), 4 /* bufSize */, 8 /* maxIteration */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPartUTF8_32All()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPartUTF8 *testCase = new NextSortKeyPartUTF8(coll, getData8(status), 32 /* bufSize */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestNextSortKeyPartUTF8_32x2()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ NextSortKeyPartUTF8 *testCase = new NextSortKeyPartUTF8(coll, getData8(status), 32 /* bufSize */, 2 /* maxIteration */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppCompare()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppCompare *testCase = new CppCompare(collObj, getData16(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppCompareNull()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppCompare *testCase = new CppCompare(collObj, getData16(status), FALSE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppCompareSimilar()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppCompare_2 *testCase = new CppCompare_2(collObj, getData16(status), getModData16(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppCompareUTF8()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppCompareUTF8 *testCase = new CppCompareUTF8(collObj, getData8(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppCompareUTF8Null()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppCompareUTF8 *testCase = new CppCompareUTF8(collObj, getData8(status), FALSE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppCompareUTF8Similar()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppCompareUTF8_2 *testCase = new CppCompareUTF8_2(collObj, getData8(status), getModData8(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppGetCollationKey()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppGetCollationKey *testCase = new CppGetCollationKey(collObj, getData16(status), TRUE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestCppGetCollationKeyNull()
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CppGetCollationKey *testCase = new CppGetCollationKey(collObj, getData16(status), FALSE /* useLen */);
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestUniStrSort() {
+ UErrorCode status = U_ZERO_ERROR;
+ UPerfFunction *testCase = new UniStrSort(*collObj, coll, getRandomData16(status));
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStringPieceSortCpp() {
+ UErrorCode status = U_ZERO_ERROR;
+ UPerfFunction *testCase = new StringPieceSortCpp(*collObj, coll, getRandomData8(status));
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStringPieceSortC() {
+ UErrorCode status = U_ZERO_ERROR;
+ UPerfFunction *testCase = new StringPieceSortC(*collObj, coll, getRandomData8(status));
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestUniStrBinSearch() {
+ UErrorCode status = U_ZERO_ERROR;
+ UPerfFunction *testCase = new UniStrBinSearch(*collObj, coll, getSortedData16(status));
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStringPieceBinSearchCpp() {
+ UErrorCode status = U_ZERO_ERROR;
+ UPerfFunction *testCase = new StringPieceBinSearchCpp(*collObj, coll, getSortedData8(status));
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+UPerfFunction* CollPerf2Test::TestStringPieceBinSearchC() {
+ UErrorCode status = U_ZERO_ERROR;
+ UPerfFunction *testCase = new StringPieceBinSearchC(*collObj, coll, getSortedData8(status));
+ if (U_FAILURE(status)) {
+ delete testCase;
+ return NULL;
+ }
+ return testCase;
+}
+
+
+int main(int argc, const char *argv[])
+{
+ UErrorCode status = U_ZERO_ERROR;
+ CollPerf2Test test(argc, argv, status);
+
+ if (U_FAILURE(status)){
+ printf("The error is %s\n", u_errorName(status));
+ //TODO: print usage here
+ return status;
+ }
+
+ if (test.run() == FALSE){
+ fprintf(stderr, "FAILED: Tests could not be run please check the arguments.\n");
+ return -1;
+ }
+ return 0;
+}
projects / apple / icu.git / blobdiff