void StringSearch::setPattern(const UnicodeString &pattern,
UErrorCode &status)
{
if (U_SUCCESS(status)) {
m_pattern_ = pattern;
usearch_setPattern(m_strsrch_, m_pattern_.getBuffer(), m_pattern_.length(),
&status);
}
}
inline SearchBuffer::~SearchBuffer()
{
// Leave the static object pointing to valid strings (pattern=targer,
// text=buffer). Otheriwse, usearch_reset() will results in 'use-after-free'
// error.
UErrorCode status = U_ZERO_ERROR;
usearch_setPattern(blink::searcher(), &newlineCharacter, 1, &status);
usearch_setText(blink::searcher(), &newlineCharacter, 1, &status);
ASSERT(status == U_ZERO_ERROR);
unlockSearcher();
}
inline SearchBuffer::SearchBuffer(const String& target, FindOptions options)
: m_options(options)
, m_prefixLength(0)
, m_numberOfCharactersJustAppended(0)
, m_atBreak(true)
, m_needsMoreContext(options & AtWordStarts)
, m_targetRequiresKanaWorkaround(containsKanaLetters(target))
{
ASSERT(!target.isEmpty());
target.appendTo(m_target);
// FIXME: We'd like to tailor the searcher to fold quote marks for us instead
// of doing it in a separate replacement pass here, but ICU doesn't offer a way
// to add tailoring on top of the locale-specific tailoring as of this writing.
foldQuoteMarksAndSoftHyphens(m_target.data(), m_target.size());
size_t targetLength = m_target.size();
m_buffer.reserveInitialCapacity(std::max(targetLength * 8, minimumSearchBufferSize));
m_overlap = m_buffer.capacity() / 4;
if ((m_options & AtWordStarts) && targetLength) {
UChar32 targetFirstCharacter;
U16_GET(m_target.data(), 0, 0, targetLength, targetFirstCharacter);
// Characters in the separator category never really occur at the beginning of a word,
// so if the target begins with such a character, we just ignore the AtWordStart option.
if (isSeparator(targetFirstCharacter)) {
m_options &= ~AtWordStarts;
m_needsMoreContext = false;
}
}
// Grab the single global searcher.
// If we ever have a reason to do more than once search buffer at once, we'll have
// to move to multiple searchers.
lockSearcher();
UStringSearch* searcher = blink::searcher();
UCollator* collator = usearch_getCollator(searcher);
UCollationStrength strength = m_options & CaseInsensitive ? UCOL_PRIMARY : UCOL_TERTIARY;
if (ucol_getStrength(collator) != strength) {
ucol_setStrength(collator, strength);
usearch_reset(searcher);
}
UErrorCode status = U_ZERO_ERROR;
usearch_setPattern(searcher, m_target.data(), targetLength, &status);
ASSERT(status == U_ZERO_ERROR);
// The kana workaround requires a normalized copy of the target string.
if (m_targetRequiresKanaWorkaround)
normalizeCharactersIntoNFCForm(m_target.data(), m_target.size(), m_normalizedTarget);
}
int32_t SSearchTest::monkeyTestCase(UCollator *coll, const UnicodeString &testCase, const UnicodeString &pattern, const UnicodeString &altPattern,
const char *name, const char *strength, uint32_t seed)
{
UErrorCode status = U_ZERO_ERROR;
int32_t actualStart = -1, actualEnd = -1;
//int32_t expectedStart = prefix.length(), expectedEnd = prefix.length() + altPattern.length();
int32_t expectedStart = -1, expectedEnd = -1;
int32_t notFoundCount = 0;
LocalUStringSearchPointer uss(usearch_openFromCollator(pattern.getBuffer(), pattern.length(),
testCase.getBuffer(), testCase.length(),
coll,
NULL, // the break iterator
&status));
// **** TODO: find *all* matches, not just first one ****
simpleSearch(coll, testCase, 0, pattern, expectedStart, expectedEnd);
usearch_search(uss.getAlias(), 0, &actualStart, &actualEnd, &status);
if (expectedStart >= 0 && (actualStart != expectedStart || actualEnd != expectedEnd)) {
errln("Search for <pattern> in <%s> failed: expected [%d, %d], got [%d, %d]\n"
" strength=%s seed=%d",
name, expectedStart, expectedEnd, actualStart, actualEnd, strength, seed);
}
if (expectedStart == -1 && actualStart == -1) {
notFoundCount += 1;
}
// **** TODO: find *all* matches, not just first one ****
simpleSearch(coll, testCase, 0, altPattern, expectedStart, expectedEnd);
usearch_setPattern(uss.getAlias(), altPattern.getBuffer(), altPattern.length(), &status);
usearch_search(uss.getAlias(), 0, &actualStart, &actualEnd, &status);
if (expectedStart >= 0 && (actualStart != expectedStart || actualEnd != expectedEnd)) {
errln("Search for <alt_pattern> in <%s> failed: expected [%d, %d], got [%d, %d]\n"
" strength=%s seed=%d",
name, expectedStart, expectedEnd, actualStart, actualEnd, strength, seed);
}
if (expectedStart == -1 && actualStart == -1) {
notFoundCount += 1;
}
return notFoundCount;
}
void TextSearcherICU::setPattern(const UChar* pattern, size_t length) {
UErrorCode status = U_ZERO_ERROR;
usearch_setPattern(m_searcher, pattern, length, &status);
DCHECK_EQ(status, U_ZERO_ERROR);
}
