diff options
Diffstat (limited to 'src')
| -rw-r--r-- | src/corelib/text/qchar.cpp | 23 | ||||
| -rw-r--r-- | src/corelib/text/qstring.cpp | 34 | ||||
| -rw-r--r-- | src/corelib/text/qunicodetables.cpp | 2 |
3 files changed, 40 insertions, 19 deletions
diff --git a/src/corelib/text/qchar.cpp b/src/corelib/text/qchar.cpp index 6b1bb33949..1b8cfb642e 100644 --- a/src/corelib/text/qchar.cpp +++ b/src/corelib/text/qchar.cpp @@ -1550,6 +1550,29 @@ QChar::UnicodeVersion QChar::currentUnicodeVersion() noexcept return UNICODE_DATA_VERSION; } +using FullConvertCaseResult = std::array<char16_t, MaxSpecialCaseLength + 1>; +static FullConvertCaseResult fullConvertCase(char32_t uc, QUnicodeTables::Case which) noexcept +{ + FullConvertCaseResult result = {}; + auto pp = result.begin(); + + const auto fold = qGetProp(uc)->cases[which]; + const auto caseDiff = fold.diff; + + if (Q_UNLIKELY(fold.special)) { + const auto *specialCase = specialCaseMap + caseDiff; + auto length = *specialCase++; + while (length--) + *pp++ = *specialCase++; + } else if (Q_UNLIKELY(QChar::requiresSurrogates(uc))) { + // so far, case convertion never changes planes (guaranteed by the qunicodetables generator) + *pp++ = QChar::highSurrogate(uc); + *pp++ = QChar::lowSurrogate(uc + caseDiff); + } else { + *pp++ = uc + caseDiff; + } + return result; +} template <typename T> Q_DECL_CONST_FUNCTION static inline T convertCase_helper(T uc, QUnicodeTables::Case which) noexcept diff --git a/src/corelib/text/qstring.cpp b/src/corelib/text/qstring.cpp index 49ddb9279b..d8c74aac62 100644 --- a/src/corelib/text/qstring.cpp +++ b/src/corelib/text/qstring.cpp @@ -6533,36 +6533,32 @@ static QString detachAndConvertCase(T &str, QStringIterator it, QUnicodeTables:: QChar *pp = s.begin() + it.index(); // will detach if necessary do { - auto uc = it.nextUnchecked(); - - const auto fold = qGetProp(uc)->cases[which]; - signed short caseDiff = fold.diff; - - if (Q_UNLIKELY(fold.special)) { - const ushort *specialCase = specialCaseMap + caseDiff; - ushort length = *specialCase++; - - if (Q_LIKELY(length == 1)) { - *pp++ = QChar(*specialCase); + const auto folded = fullConvertCase(it.nextUnchecked(), which); + if (Q_UNLIKELY(folded[1])) { + if (folded[0] == *pp && !folded[2]) { + // special case: only second actually changed (e.g. surrogate pairs), + // avoid slow case + ++pp; + *pp++ = folded[1]; } else { // slow path: the string is growing int inpos = it.index() - 1; int outpos = pp - s.constBegin(); - s.replace(outpos, 1, reinterpret_cast<const QChar *>(specialCase), length); - pp = const_cast<QChar *>(s.constBegin()) + outpos + length; + int foldedSize = 2; // must be at least 2, b/c folded[1] != NUL + while (folded[foldedSize]) + ++foldedSize; + + s.replace(outpos, 1, reinterpret_cast<const QChar *>(folded.data()), foldedSize); + pp = const_cast<QChar *>(s.constBegin()) + outpos + foldedSize; // do we need to adjust the input iterator too? // if it is pointing to s's data, str is empty if (str.isEmpty()) - it = QStringIterator(s.constBegin(), inpos + length, s.constEnd()); + it = QStringIterator(s.constBegin(), inpos + foldedSize, s.constEnd()); } - } else if (Q_UNLIKELY(QChar::requiresSurrogates(uc))) { - // so far, case convertion never changes planes (guaranteed by the qunicodetables generator) - pp++; - *pp++ = QChar(QChar::lowSurrogate(uc + caseDiff)); } else { - *pp++ = QChar(uc + caseDiff); + *pp++ = folded[0]; } } while (it.hasNext()); diff --git a/src/corelib/text/qunicodetables.cpp b/src/corelib/text/qunicodetables.cpp index e6f6487126..865f93ba52 100644 --- a/src/corelib/text/qunicodetables.cpp +++ b/src/corelib/text/qunicodetables.cpp @@ -9946,6 +9946,8 @@ static const unsigned short specialCaseMap[] = { 0x1, 0xa64b }; +const unsigned int MaxSpecialCaseLength = 3; + static const unsigned short uc_decomposition_trie[] = { // 0 - 0x3400 |