diff options
Diffstat (limited to 'intl/icu/source/i18n/utf8collationiterator.cpp')
-rw-r--r-- | intl/icu/source/i18n/utf8collationiterator.cpp | 529 |
1 files changed, 529 insertions, 0 deletions
diff --git a/intl/icu/source/i18n/utf8collationiterator.cpp b/intl/icu/source/i18n/utf8collationiterator.cpp new file mode 100644 index 0000000000..ea7278afc6 --- /dev/null +++ b/intl/icu/source/i18n/utf8collationiterator.cpp @@ -0,0 +1,529 @@ +// © 2016 and later: Unicode, Inc. and others. +// License & terms of use: http://www.unicode.org/copyright.html +/* +******************************************************************************* +* Copyright (C) 2012-2014, International Business Machines +* Corporation and others. All Rights Reserved. +******************************************************************************* +* utf8collationiterator.cpp +* +* created on: 2012nov12 (from utf16collationiterator.cpp & uitercollationiterator.cpp) +* created by: Markus W. Scherer +*/ + +#include "unicode/utypes.h" + +#if !UCONFIG_NO_COLLATION + +#include "unicode/utf8.h" +#include "charstr.h" +#include "cmemory.h" +#include "collation.h" +#include "collationdata.h" +#include "collationfcd.h" +#include "collationiterator.h" +#include "normalizer2impl.h" +#include "uassert.h" +#include "utf8collationiterator.h" + +U_NAMESPACE_BEGIN + +UTF8CollationIterator::~UTF8CollationIterator() {} + +void +UTF8CollationIterator::resetToOffset(int32_t newOffset) { + reset(); + pos = newOffset; +} + +int32_t +UTF8CollationIterator::getOffset() const { + return pos; +} + +uint32_t +UTF8CollationIterator::handleNextCE32(UChar32 &c, UErrorCode & /*errorCode*/) { + if(pos == length) { + c = U_SENTINEL; + return Collation::FALLBACK_CE32; + } + // Optimized combination of U8_NEXT_OR_FFFD() and UTRIE2_U8_NEXT32(). + c = u8[pos++]; + if(U8_IS_SINGLE(c)) { + // ASCII 00..7F + return trie->data32[c]; + } + uint8_t t1, t2; + if(0xe0 <= c && c < 0xf0 && + ((pos + 1) < length || length < 0) && + U8_IS_VALID_LEAD3_AND_T1(c, t1 = u8[pos]) && + (t2 = (u8[pos + 1] - 0x80)) <= 0x3f) { + // U+0800..U+FFFF except surrogates + c = (((c & 0xf) << 12) | ((t1 & 0x3f) << 6) | t2); + pos += 2; + return UTRIE2_GET32_FROM_U16_SINGLE_LEAD(trie, c); + } else if(c < 0xe0 && c >= 0xc2 && pos != length && (t1 = (u8[pos] - 0x80)) <= 0x3f) { + // U+0080..U+07FF + uint32_t ce32 = trie->data32[trie->index[(UTRIE2_UTF8_2B_INDEX_2_OFFSET - 0xc0) + c] + t1]; + c = ((c & 0x1f) << 6) | t1; + ++pos; + return ce32; + } else { + // Function call for supplementary code points and error cases. + // Illegal byte sequences yield U+FFFD. + c = utf8_nextCharSafeBody(u8, &pos, length, c, -3); + return data->getCE32(c); + } +} + +UBool +UTF8CollationIterator::foundNULTerminator() { + if(length < 0) { + length = --pos; + return true; + } else { + return false; + } +} + +UBool +UTF8CollationIterator::forbidSurrogateCodePoints() const { + return true; +} + +UChar32 +UTF8CollationIterator::nextCodePoint(UErrorCode & /*errorCode*/) { + if(pos == length) { + return U_SENTINEL; + } + if(u8[pos] == 0 && length < 0) { + length = pos; + return U_SENTINEL; + } + UChar32 c; + U8_NEXT_OR_FFFD(u8, pos, length, c); + return c; +} + +UChar32 +UTF8CollationIterator::previousCodePoint(UErrorCode & /*errorCode*/) { + if(pos == 0) { + return U_SENTINEL; + } + UChar32 c; + U8_PREV_OR_FFFD(u8, 0, pos, c); + return c; +} + +void +UTF8CollationIterator::forwardNumCodePoints(int32_t num, UErrorCode & /*errorCode*/) { + U8_FWD_N(u8, pos, length, num); +} + +void +UTF8CollationIterator::backwardNumCodePoints(int32_t num, UErrorCode & /*errorCode*/) { + U8_BACK_N(u8, 0, pos, num); +} + +// FCDUTF8CollationIterator ------------------------------------------------ *** + +FCDUTF8CollationIterator::~FCDUTF8CollationIterator() {} + +void +FCDUTF8CollationIterator::resetToOffset(int32_t newOffset) { + reset(); + start = pos = newOffset; + state = CHECK_FWD; +} + +int32_t +FCDUTF8CollationIterator::getOffset() const { + if(state != IN_NORMALIZED) { + return pos; + } else if(pos == 0) { + return start; + } else { + return limit; + } +} + +uint32_t +FCDUTF8CollationIterator::handleNextCE32(UChar32 &c, UErrorCode &errorCode) { + for(;;) { + if(state == CHECK_FWD) { + // Combination of UTF8CollationIterator::handleNextCE32() with FCD check fastpath. + if(pos == length) { + c = U_SENTINEL; + return Collation::FALLBACK_CE32; + } + c = u8[pos++]; + if(U8_IS_SINGLE(c)) { + // ASCII 00..7F + return trie->data32[c]; + } + uint8_t t1, t2; + if(0xe0 <= c && c < 0xf0 && + ((pos + 1) < length || length < 0) && + U8_IS_VALID_LEAD3_AND_T1(c, t1 = u8[pos]) && + (t2 = (u8[pos + 1] - 0x80)) <= 0x3f) { + // U+0800..U+FFFF except surrogates + c = (((c & 0xf) << 12) | ((t1 & 0x3f) << 6) | t2); + pos += 2; + if(CollationFCD::hasTccc(c) && + (CollationFCD::maybeTibetanCompositeVowel(c) || + (pos != length && nextHasLccc()))) { + pos -= 3; + } else { + break; // return CE32(BMP) + } + } else if(c < 0xe0 && c >= 0xc2 && pos != length && (t1 = (u8[pos] - 0x80)) <= 0x3f) { + // U+0080..U+07FF + uint32_t ce32 = trie->data32[trie->index[(UTRIE2_UTF8_2B_INDEX_2_OFFSET - 0xc0) + c] + t1]; + c = ((c & 0x1f) << 6) | t1; + ++pos; + if(CollationFCD::hasTccc(c) && pos != length && nextHasLccc()) { + pos -= 2; + } else { + return ce32; + } + } else { + // Function call for supplementary code points and error cases. + // Illegal byte sequences yield U+FFFD. + c = utf8_nextCharSafeBody(u8, &pos, length, c, -3); + if(c == 0xfffd) { + return Collation::FFFD_CE32; + } else { + U_ASSERT(c > 0xffff); + if(CollationFCD::hasTccc(U16_LEAD(c)) && pos != length && nextHasLccc()) { + pos -= 4; + } else { + return data->getCE32FromSupplementary(c); + } + } + } + if(!nextSegment(errorCode)) { + c = U_SENTINEL; + return Collation::FALLBACK_CE32; + } + continue; + } else if(state == IN_FCD_SEGMENT && pos != limit) { + return UTF8CollationIterator::handleNextCE32(c, errorCode); + } else if(state == IN_NORMALIZED && pos != normalized.length()) { + c = normalized[pos++]; + break; + } else { + switchToForward(); + } + } + return UTRIE2_GET32_FROM_U16_SINGLE_LEAD(trie, c); +} + +UBool +FCDUTF8CollationIterator::nextHasLccc() const { + U_ASSERT(state == CHECK_FWD && pos != length); + // The lowest code point with ccc!=0 is U+0300 which is CC 80 in UTF-8. + // CJK U+4000..U+DFFF except U+Axxx are also FCD-inert. (Lead bytes E4..ED except EA.) + UChar32 c = u8[pos]; + if(c < 0xcc || (0xe4 <= c && c <= 0xed && c != 0xea)) { return false; } + int32_t i = pos; + U8_NEXT_OR_FFFD(u8, i, length, c); + if(c > 0xffff) { c = U16_LEAD(c); } + return CollationFCD::hasLccc(c); +} + +UBool +FCDUTF8CollationIterator::previousHasTccc() const { + U_ASSERT(state == CHECK_BWD && pos != 0); + UChar32 c = u8[pos - 1]; + if(U8_IS_SINGLE(c)) { return false; } + int32_t i = pos; + U8_PREV_OR_FFFD(u8, 0, i, c); + if(c > 0xffff) { c = U16_LEAD(c); } + return CollationFCD::hasTccc(c); +} + +char16_t +FCDUTF8CollationIterator::handleGetTrailSurrogate() { + if(state != IN_NORMALIZED) { return 0; } + U_ASSERT(pos < normalized.length()); + char16_t trail; + if(U16_IS_TRAIL(trail = normalized[pos])) { ++pos; } + return trail; +} + +UBool +FCDUTF8CollationIterator::foundNULTerminator() { + if(state == CHECK_FWD && length < 0) { + length = --pos; + return true; + } else { + return false; + } +} + +UChar32 +FCDUTF8CollationIterator::nextCodePoint(UErrorCode &errorCode) { + UChar32 c; + for(;;) { + if(state == CHECK_FWD) { + if(pos == length || ((c = u8[pos]) == 0 && length < 0)) { + return U_SENTINEL; + } + if(U8_IS_SINGLE(c)) { + ++pos; + return c; + } + U8_NEXT_OR_FFFD(u8, pos, length, c); + if(CollationFCD::hasTccc(c <= 0xffff ? c : U16_LEAD(c)) && + (CollationFCD::maybeTibetanCompositeVowel(c) || + (pos != length && nextHasLccc()))) { + // c is not FCD-inert, therefore it is not U+FFFD and it has a valid byte sequence + // and we can use U8_LENGTH() rather than a previous-position variable. + pos -= U8_LENGTH(c); + if(!nextSegment(errorCode)) { + return U_SENTINEL; + } + continue; + } + return c; + } else if(state == IN_FCD_SEGMENT && pos != limit) { + U8_NEXT_OR_FFFD(u8, pos, length, c); + return c; + } else if(state == IN_NORMALIZED && pos != normalized.length()) { + c = normalized.char32At(pos); + pos += U16_LENGTH(c); + return c; + } else { + switchToForward(); + } + } +} + +UChar32 +FCDUTF8CollationIterator::previousCodePoint(UErrorCode &errorCode) { + UChar32 c; + for(;;) { + if(state == CHECK_BWD) { + if(pos == 0) { + return U_SENTINEL; + } + if(U8_IS_SINGLE(c = u8[pos - 1])) { + --pos; + return c; + } + U8_PREV_OR_FFFD(u8, 0, pos, c); + if(CollationFCD::hasLccc(c <= 0xffff ? c : U16_LEAD(c)) && + (CollationFCD::maybeTibetanCompositeVowel(c) || + (pos != 0 && previousHasTccc()))) { + // c is not FCD-inert, therefore it is not U+FFFD and it has a valid byte sequence + // and we can use U8_LENGTH() rather than a previous-position variable. + pos += U8_LENGTH(c); + if(!previousSegment(errorCode)) { + return U_SENTINEL; + } + continue; + } + return c; + } else if(state == IN_FCD_SEGMENT && pos != start) { + U8_PREV_OR_FFFD(u8, 0, pos, c); + return c; + } else if(state >= IN_NORMALIZED && pos != 0) { + c = normalized.char32At(pos - 1); + pos -= U16_LENGTH(c); + return c; + } else { + switchToBackward(); + } + } +} + +void +FCDUTF8CollationIterator::forwardNumCodePoints(int32_t num, UErrorCode &errorCode) { + // Specify the class to avoid a virtual-function indirection. + // In Java, we would declare this class final. + while(num > 0 && FCDUTF8CollationIterator::nextCodePoint(errorCode) >= 0) { + --num; + } +} + +void +FCDUTF8CollationIterator::backwardNumCodePoints(int32_t num, UErrorCode &errorCode) { + // Specify the class to avoid a virtual-function indirection. + // In Java, we would declare this class final. + while(num > 0 && FCDUTF8CollationIterator::previousCodePoint(errorCode) >= 0) { + --num; + } +} + +void +FCDUTF8CollationIterator::switchToForward() { + U_ASSERT(state == CHECK_BWD || + (state == IN_FCD_SEGMENT && pos == limit) || + (state == IN_NORMALIZED && pos == normalized.length())); + if(state == CHECK_BWD) { + // Turn around from backward checking. + start = pos; + if(pos == limit) { + state = CHECK_FWD; // Check forward. + } else { // pos < limit + state = IN_FCD_SEGMENT; // Stay in FCD segment. + } + } else { + // Reached the end of the FCD segment. + if(state == IN_FCD_SEGMENT) { + // The input text segment is FCD, extend it forward. + } else { + // The input text segment needed to be normalized. + // Switch to checking forward from it. + start = pos = limit; + } + state = CHECK_FWD; + } +} + +UBool +FCDUTF8CollationIterator::nextSegment(UErrorCode &errorCode) { + if(U_FAILURE(errorCode)) { return false; } + U_ASSERT(state == CHECK_FWD && pos != length); + // The input text [start..pos[ passes the FCD check. + int32_t segmentStart = pos; + // Collect the characters being checked, in case they need to be normalized. + UnicodeString s; + uint8_t prevCC = 0; + for(;;) { + // Fetch the next character and its fcd16 value. + int32_t cpStart = pos; + UChar32 c; + U8_NEXT_OR_FFFD(u8, pos, length, c); + uint16_t fcd16 = nfcImpl.getFCD16(c); + uint8_t leadCC = (uint8_t)(fcd16 >> 8); + if(leadCC == 0 && cpStart != segmentStart) { + // FCD boundary before this character. + pos = cpStart; + break; + } + s.append(c); + if(leadCC != 0 && (prevCC > leadCC || CollationFCD::isFCD16OfTibetanCompositeVowel(fcd16))) { + // Fails FCD check. Find the next FCD boundary and normalize. + while(pos != length) { + cpStart = pos; + U8_NEXT_OR_FFFD(u8, pos, length, c); + if(nfcImpl.getFCD16(c) <= 0xff) { + pos = cpStart; + break; + } + s.append(c); + } + if(!normalize(s, errorCode)) { return false; } + start = segmentStart; + limit = pos; + state = IN_NORMALIZED; + pos = 0; + return true; + } + prevCC = (uint8_t)fcd16; + if(pos == length || prevCC == 0) { + // FCD boundary after the last character. + break; + } + } + limit = pos; + pos = segmentStart; + U_ASSERT(pos != limit); + state = IN_FCD_SEGMENT; + return true; +} + +void +FCDUTF8CollationIterator::switchToBackward() { + U_ASSERT(state == CHECK_FWD || + (state == IN_FCD_SEGMENT && pos == start) || + (state >= IN_NORMALIZED && pos == 0)); + if(state == CHECK_FWD) { + // Turn around from forward checking. + limit = pos; + if(pos == start) { + state = CHECK_BWD; // Check backward. + } else { // pos > start + state = IN_FCD_SEGMENT; // Stay in FCD segment. + } + } else { + // Reached the start of the FCD segment. + if(state == IN_FCD_SEGMENT) { + // The input text segment is FCD, extend it backward. + } else { + // The input text segment needed to be normalized. + // Switch to checking backward from it. + limit = pos = start; + } + state = CHECK_BWD; + } +} + +UBool +FCDUTF8CollationIterator::previousSegment(UErrorCode &errorCode) { + if(U_FAILURE(errorCode)) { return false; } + U_ASSERT(state == CHECK_BWD && pos != 0); + // The input text [pos..limit[ passes the FCD check. + int32_t segmentLimit = pos; + // Collect the characters being checked, in case they need to be normalized. + UnicodeString s; + uint8_t nextCC = 0; + for(;;) { + // Fetch the previous character and its fcd16 value. + int32_t cpLimit = pos; + UChar32 c; + U8_PREV_OR_FFFD(u8, 0, pos, c); + uint16_t fcd16 = nfcImpl.getFCD16(c); + uint8_t trailCC = (uint8_t)fcd16; + if(trailCC == 0 && cpLimit != segmentLimit) { + // FCD boundary after this character. + pos = cpLimit; + break; + } + s.append(c); + if(trailCC != 0 && ((nextCC != 0 && trailCC > nextCC) || + CollationFCD::isFCD16OfTibetanCompositeVowel(fcd16))) { + // Fails FCD check. Find the previous FCD boundary and normalize. + while(fcd16 > 0xff && pos != 0) { + cpLimit = pos; + U8_PREV_OR_FFFD(u8, 0, pos, c); + fcd16 = nfcImpl.getFCD16(c); + if(fcd16 == 0) { + pos = cpLimit; + break; + } + s.append(c); + } + s.reverse(); + if(!normalize(s, errorCode)) { return false; } + limit = segmentLimit; + start = pos; + state = IN_NORMALIZED; + pos = normalized.length(); + return true; + } + nextCC = (uint8_t)(fcd16 >> 8); + if(pos == 0 || nextCC == 0) { + // FCD boundary before the following character. + break; + } + } + start = pos; + pos = segmentLimit; + U_ASSERT(pos != start); + state = IN_FCD_SEGMENT; + return true; +} + +UBool +FCDUTF8CollationIterator::normalize(const UnicodeString &s, UErrorCode &errorCode) { + // NFD without argument checking. + U_ASSERT(U_SUCCESS(errorCode)); + nfcImpl.decompose(s, normalized, errorCode); + return U_SUCCESS(errorCode); +} + +U_NAMESPACE_END + +#endif // !UCONFIG_NO_COLLATION |