StringSearch::StringSearch(const UnicodeString &pattern,
CharacterIterator &text,
const Locale &locale,
BreakIterator *breakiter,
UErrorCode &status) :
SearchIterator(text, breakiter),
m_collator_(),
m_pattern_(pattern)
{
if (U_FAILURE(status)) {
m_strsrch_ = NULL;
return;
}
m_strsrch_ = usearch_open(m_pattern_.getBuffer(), m_pattern_.length(),
m_text_.getBuffer(), m_text_.length(),
locale.getName(), (UBreakIterator *)breakiter,
&status);
uprv_free(m_search_);
m_search_ = NULL;
if (U_SUCCESS(status)) {
int32_t length;
const UChar *rules = ucol_getRules(m_strsrch_->collator, &length);
m_collation_rules_.setTo(rules, length);
m_collator_.setUCollator((UCollator *)m_strsrch_->collator,
&m_collation_rules_);
// m_search_ has been created by the base SearchIterator class
m_search_ = m_strsrch_->search;
}
}
StringSearch::StringSearch(const UnicodeString &pattern,
const UnicodeString &text,
const Locale &locale,
BreakIterator *breakiter,
UErrorCode &status) :
SearchIterator(text, breakiter),
m_pattern_(pattern)
{
if (U_FAILURE(status)) {
m_strsrch_ = NULL;
return;
}
m_strsrch_ = usearch_open(m_pattern_.getBuffer(), m_pattern_.length(),
m_text_.getBuffer(), m_text_.length(),
locale.getName(), (UBreakIterator *)breakiter,
&status);
uprv_free(m_search_);
m_search_ = NULL;
if (U_SUCCESS(status)) {
// m_search_ has been created by the base SearchIterator class
m_search_ = m_strsrch_->search;
}
}
static UStringSearch* createSearcher()
{
// Provide a non-empty pattern and non-empty text so usearch_open will not fail,
// but it doesn't matter exactly what it is, since we don't perform any searches
// without setting both the pattern and the text.
UErrorCode status = U_ZERO_ERROR;
String searchCollatorName = currentSearchLocaleID() + String("@collation=search");
UStringSearch* searcher = usearch_open(&newlineCharacter, 1, &newlineCharacter, 1, searchCollatorName.utf8().data(), 0, &status);
ASSERT(status == U_ZERO_ERROR || status == U_USING_FALLBACK_WARNING || status == U_USING_DEFAULT_WARNING);
return searcher;
}
static void *engine_fixed_compile(error_t **error, UString *ustr, uint32_t flags)
{
UErrorCode status;
fixed_pattern_t *p;
p = mem_new(*p);
p->pattern = ustr; // not needed with usearch ?
p->flags = flags;
p->ubrk = NULL;
p->usearch = NULL;
status = U_ZERO_ERROR;
if (ustring_empty(ustr)) {
if (IS_WORD_BOUNDED(flags)) {
p->ubrk = ubrk_open(UBRK_WORD, NULL, NULL, 0, &status);
}
} else {
if (!IS_WHOLE_LINE(flags)) {
if (IS_WORD_BOUNDED(flags)) {
p->ubrk = ubrk_open(UBRK_WORD, NULL, NULL, 0, &status);
} else if (WITH_GRAPHEME()) {
p->ubrk = ubrk_open(UBRK_CHARACTER, NULL, NULL, 0, &status);
}
if (U_FAILURE(status)) {
fixed_pattern_destroy(p);
icu_error_set(error, FATAL, status, "ubrk_open");
return NULL;
}
}
if (IS_WORD_BOUNDED(flags) || (IS_CASE_INSENSITIVE(flags) && !IS_WHOLE_LINE(flags))) {
p->usearch = usearch_open(ustr->ptr, ustr->len, USEARCH_FAKE_USTR, uloc_getDefault(), p->ubrk, &status);
if (U_FAILURE(status)) {
if (NULL != p->ubrk) {
ubrk_close(p->ubrk);
}
fixed_pattern_destroy(p);
icu_error_set(error, FATAL, status, "usearch_open");
return NULL;
}
if (IS_CASE_INSENSITIVE(flags)) {
UCollator *ucol;
ucol = usearch_getCollator(p->usearch);
ucol_setStrength(ucol, (flags & ~OPT_MASK) > 1 ? UCOL_SECONDARY : UCOL_PRIMARY);
}
}
}
return p;
}
StringSearch::StringSearch(const UnicodeString &pattern,
const UnicodeString &text,
const Locale &locale,
BreakIterator *breakiter,
UErrorCode &status) :
SearchIterator(text, breakiter),
m_collator_(),
m_pattern_(pattern)
{
if (U_FAILURE(status)) {
m_strsrch_ = NULL;
return;
}
m_strsrch_ = usearch_open(m_pattern_.getBuffer(), m_pattern_.length(),
m_text_.getBuffer(), m_text_.length(),
locale.getName(), (UBreakIterator *)breakiter,
&status);
uprv_free(m_search_);
m_search_ = NULL;
// !!! dlf m_collator_ is an odd beast. basically it is an aliasing
// wrapper around the internal collator and rules, which (here) are
// owned by this stringsearch object. this means 1) it's destructor
// _should not_ delete the ucollator or rules, and 2) changes made
// to the exposed collator (setStrength etc) _should_ modify the
// ucollator. thus the collator is not a copy-on-write alias, and it
// needs to distinguish itself not merely from 'stand alone' colators
// but also from copy-on-write ones. it needs additional state, which
// setUCollator should set.
if (U_SUCCESS(status)) {
int32_t length;
const UChar *rules = ucol_getRules(m_strsrch_->collator, &length);
m_collation_rules_.setTo(rules, length);
m_collator_.setUCollator((UCollator *)m_strsrch_->collator,
&m_collation_rules_);
// m_search_ has been created by the base SearchIterator class
m_search_ = m_strsrch_->search;
}
}
StringSearchPerformanceTest::StringSearchPerformanceTest(int32_t argc, const char *argv[], UErrorCode &status)
:UPerfTest(argc,argv,status){
int32_t start, end;
#ifdef TEST_BOYER_MOORE_SEARCH
bms = NULL;
#else
srch = NULL;
#endif
pttrn = NULL;
if(status== U_ILLEGAL_ARGUMENT_ERROR || line_mode){
fprintf(stderr,gUsageString, "strsrchperf");
return;
}
/* Get the Text */
src = getBuffer(srcLen, status);
#if 0
/* Get a word to find. Do this by selecting a random word with a word breakiterator. */
UBreakIterator* brk = ubrk_open(UBRK_WORD, locale, src, srcLen, &status);
if(U_FAILURE(status)){
fprintf(stderr, "FAILED to create pattern for searching. Error: %s\n", u_errorName(status));
return;
}
start = ubrk_preceding(brk, 1000);
end = ubrk_following(brk, start);
pttrnLen = end - start;
UChar* temp = (UChar*)malloc(sizeof(UChar)*(pttrnLen));
for (int i = 0; i < pttrnLen; i++) {
temp[i] = src[start++];
}
pttrn = temp; /* store word in pttrn */
ubrk_close(brk);
#else
/* The first line of the file contains the pattern */
start = 0;
for(end = start; ; end += 1) {
UChar ch = src[end];
if (ch == 0x000A || ch == 0x000D || ch == 0x2028) {
break;
}
}
pttrnLen = end - start;
UChar* temp = (UChar*)malloc(sizeof(UChar)*(pttrnLen));
for (int i = 0; i < pttrnLen; i++) {
temp[i] = src[start++];
}
pttrn = temp; /* store word in pttrn */
#endif
#ifdef TEST_BOYER_MOORE_SEARCH
UnicodeString patternString(pttrn, pttrnLen);
UCollator *coll = ucol_open(locale, &status);
CollData *data = CollData::open(coll, status);
targetString = new UnicodeString(src, srcLen);
bms = new BoyerMooreSearch(data, patternString, targetString, status);
#else
/* Create the StringSearch object to be use in performance test. */
srch = usearch_open(pttrn, pttrnLen, src, srcLen, locale, NULL, &status);
#endif
if(U_FAILURE(status)){
fprintf(stderr, "FAILED to create UPerfTest object. Error: %s\n", u_errorName(status));
return;
}
}
/* {{{ grapheme_strpos_utf16 - strrpos using utf16*/
int32_t grapheme_strpos_utf16(char *haystack, size_t haystack_len, char *needle, size_t needle_len, int32_t offset, int32_t *puchar_pos, int f_ignore_case, int last)
{
UChar *uhaystack = NULL, *uneedle = NULL;
int32_t uhaystack_len = 0, uneedle_len = 0, char_pos, ret_pos, offset_pos = 0;
unsigned char u_break_iterator_buffer[U_BRK_SAFECLONE_BUFFERSIZE];
UBreakIterator* bi = NULL;
UErrorCode status;
UStringSearch* src = NULL;
UCollator *coll;
if(puchar_pos) {
*puchar_pos = -1;
}
/* convert the strings to UTF-16. */
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&uhaystack, &uhaystack_len, haystack, haystack_len, &status );
STRPOS_CHECK_STATUS(status, "Error converting input string to UTF-16");
status = U_ZERO_ERROR;
intl_convert_utf8_to_utf16(&uneedle, &uneedle_len, needle, needle_len, &status );
STRPOS_CHECK_STATUS(status, "Error converting needle string to UTF-16");
/* get a pointer to the haystack taking into account the offset */
status = U_ZERO_ERROR;
bi = grapheme_get_break_iterator(u_break_iterator_buffer, &status );
STRPOS_CHECK_STATUS(status, "Failed to get iterator");
status = U_ZERO_ERROR;
ubrk_setText(bi, uhaystack, uhaystack_len, &status);
STRPOS_CHECK_STATUS(status, "Failed to set up iterator");
status = U_ZERO_ERROR;
src = usearch_open(uneedle, uneedle_len, uhaystack, uhaystack_len, "", bi, &status);
STRPOS_CHECK_STATUS(status, "Error creating search object");
if(f_ignore_case) {
coll = usearch_getCollator(src);
status = U_ZERO_ERROR;
ucol_setAttribute(coll, UCOL_STRENGTH, UCOL_SECONDARY, &status);
STRPOS_CHECK_STATUS(status, "Error setting collation strength");
usearch_reset(src);
}
if(offset != 0) {
offset_pos = grapheme_get_haystack_offset(bi, offset);
if(offset_pos == -1) {
status = U_ILLEGAL_ARGUMENT_ERROR;
STRPOS_CHECK_STATUS(status, "Invalid search offset");
}
status = U_ZERO_ERROR;
usearch_setOffset(src, offset_pos, &status);
STRPOS_CHECK_STATUS(status, "Invalid search offset");
}
if(last) {
char_pos = usearch_last(src, &status);
if(char_pos < offset_pos) {
/* last one is beyound our start offset */
char_pos = USEARCH_DONE;
}
} else {
char_pos = usearch_next(src, &status);
}
STRPOS_CHECK_STATUS(status, "Error looking up string");
if(char_pos != USEARCH_DONE && ubrk_isBoundary(bi, char_pos)) {
ret_pos = grapheme_count_graphemes(bi, uhaystack,char_pos);
if(puchar_pos) {
*puchar_pos = char_pos;
}
} else {
ret_pos = -1;
}
if (uhaystack) {
efree( uhaystack );
}
if (uneedle) {
efree( uneedle );
}
ubrk_close (bi);
usearch_close (src);
return ret_pos;
}
