| 1 | // Copyright (C) 2016 The Qt Company Ltd. |
| 2 | // Copyright (C) 2014 Klarälvdalens Datakonsult AB, a KDAB Group company, info@kdab.com, author Giuseppe D'Angelo <giuseppe.dangelo@kdab.com> |
| 3 | // SPDX-License-Identifier: LicenseRef-Qt-Commercial OR LGPL-3.0-only OR GPL-2.0-only OR GPL-3.0-only |
| 4 | |
| 5 | #ifndef QSTRINGITERATOR_H |
| 6 | #define QSTRINGITERATOR_H |
| 7 | |
| 8 | // |
| 9 | // W A R N I N G |
| 10 | // ------------- |
| 11 | // |
| 12 | // This file is not part of the Qt API. It exists purely as an |
| 13 | // implementation detail. This header file may change from version to |
| 14 | // version without notice, or even be removed. |
| 15 | // |
| 16 | // We mean it. |
| 17 | // |
| 18 | |
| 19 | #include <QtCore/private/qglobal_p.h> |
| 20 | #include <QtCore/qstring.h> |
| 21 | |
| 22 | QT_BEGIN_NAMESPACE |
| 23 | |
| 24 | class QStringIterator |
| 25 | { |
| 26 | QString::const_iterator i, pos, e; |
| 27 | static_assert((std::is_same<QString::const_iterator, const QChar *>::value)); |
| 28 | static bool less(const QChar *lhs, const QChar *rhs) noexcept |
| 29 | { return std::less{}(lhs, rhs); } |
| 30 | public: |
| 31 | explicit QStringIterator(QStringView string, qsizetype idx = 0) |
| 32 | : i(string.begin()), |
| 33 | pos(i + idx), |
| 34 | e(string.end()) |
| 35 | { |
| 36 | } |
| 37 | |
| 38 | inline explicit QStringIterator(const QChar *begin, const QChar *end) |
| 39 | : i(begin), |
| 40 | pos(begin), |
| 41 | e(end) |
| 42 | { |
| 43 | } |
| 44 | |
| 45 | explicit QStringIterator(const QChar *begin, qsizetype idx, const QChar *end) |
| 46 | : i(begin), |
| 47 | pos(begin + idx), |
| 48 | e(end) |
| 49 | { |
| 50 | } |
| 51 | |
| 52 | inline QString::const_iterator position() const |
| 53 | { |
| 54 | return pos; |
| 55 | } |
| 56 | |
| 57 | qsizetype index() const |
| 58 | { |
| 59 | return pos - i; |
| 60 | } |
| 61 | |
| 62 | inline void setPosition(QString::const_iterator position) |
| 63 | { |
| 64 | Q_ASSERT_X(!less(position, i) && !less(e, position), |
| 65 | Q_FUNC_INFO, "position out of bounds" ); |
| 66 | pos = position; |
| 67 | } |
| 68 | |
| 69 | // forward iteration |
| 70 | |
| 71 | inline bool hasNext() const |
| 72 | { |
| 73 | return less(lhs: pos, rhs: e); |
| 74 | } |
| 75 | |
| 76 | inline void advance() |
| 77 | { |
| 78 | Q_ASSERT_X(hasNext(), Q_FUNC_INFO, "iterator hasn't a next item" ); |
| 79 | |
| 80 | if (Q_UNLIKELY((pos++)->isHighSurrogate())) { |
| 81 | if (Q_LIKELY(pos != e && pos->isLowSurrogate())) |
| 82 | ++pos; |
| 83 | } |
| 84 | } |
| 85 | |
| 86 | inline void advanceUnchecked() |
| 87 | { |
| 88 | Q_ASSERT_X(hasNext(), Q_FUNC_INFO, "iterator hasn't a next item" ); |
| 89 | |
| 90 | if (Q_UNLIKELY((pos++)->isHighSurrogate())) { |
| 91 | Q_ASSERT(hasNext() && pos->isLowSurrogate()); |
| 92 | ++pos; |
| 93 | } |
| 94 | } |
| 95 | |
| 96 | inline char32_t peekNextUnchecked() const |
| 97 | { |
| 98 | Q_ASSERT_X(hasNext(), Q_FUNC_INFO, "iterator hasn't a next item" ); |
| 99 | |
| 100 | if (Q_UNLIKELY(pos->isHighSurrogate())) { |
| 101 | Q_ASSERT(less(pos + 1, e) && pos[1].isLowSurrogate()); |
| 102 | return QChar::surrogateToUcs4(high: pos[0], low: pos[1]); |
| 103 | } |
| 104 | |
| 105 | return pos->unicode(); |
| 106 | } |
| 107 | |
| 108 | inline char32_t peekNext(char32_t invalidAs = QChar::ReplacementCharacter) const |
| 109 | { |
| 110 | Q_ASSERT_X(hasNext(), Q_FUNC_INFO, "iterator hasn't a next item" ); |
| 111 | |
| 112 | if (Q_UNLIKELY(pos->isSurrogate())) { |
| 113 | if (Q_LIKELY(pos->isHighSurrogate())) { |
| 114 | const QChar *low = pos + 1; |
| 115 | if (Q_LIKELY(low != e && low->isLowSurrogate())) |
| 116 | return QChar::surrogateToUcs4(high: *pos, low: *low); |
| 117 | } |
| 118 | return invalidAs; |
| 119 | } |
| 120 | |
| 121 | return pos->unicode(); |
| 122 | } |
| 123 | |
| 124 | inline char32_t nextUnchecked() |
| 125 | { |
| 126 | Q_ASSERT_X(hasNext(), Q_FUNC_INFO, "iterator hasn't a next item" ); |
| 127 | |
| 128 | const QChar cur = *pos++; |
| 129 | if (Q_UNLIKELY(cur.isHighSurrogate())) { |
| 130 | Q_ASSERT(hasNext() && pos->isLowSurrogate()); |
| 131 | return QChar::surrogateToUcs4(high: cur, low: *pos++); |
| 132 | } |
| 133 | return cur.unicode(); |
| 134 | } |
| 135 | |
| 136 | inline char32_t next(char32_t invalidAs = QChar::ReplacementCharacter) |
| 137 | { |
| 138 | Q_ASSERT_X(hasNext(), Q_FUNC_INFO, "iterator hasn't a next item" ); |
| 139 | |
| 140 | const QChar uc = *pos++; |
| 141 | if (Q_UNLIKELY(uc.isSurrogate())) { |
| 142 | if (Q_LIKELY(uc.isHighSurrogate() && hasNext() && pos->isLowSurrogate())) |
| 143 | return QChar::surrogateToUcs4(high: uc, low: *pos++); |
| 144 | return invalidAs; |
| 145 | } |
| 146 | |
| 147 | return uc.unicode(); |
| 148 | } |
| 149 | |
| 150 | // backwards iteration |
| 151 | |
| 152 | inline bool hasPrevious() const |
| 153 | { |
| 154 | return less(lhs: i, rhs: pos); |
| 155 | } |
| 156 | |
| 157 | inline void recede() |
| 158 | { |
| 159 | Q_ASSERT_X(hasPrevious(), Q_FUNC_INFO, "iterator hasn't a previous item" ); |
| 160 | |
| 161 | if (Q_UNLIKELY((--pos)->isLowSurrogate())) { |
| 162 | const QChar *high = pos - 1; |
| 163 | if (Q_LIKELY(high != i - 1 && high->isHighSurrogate())) |
| 164 | --pos; |
| 165 | } |
| 166 | } |
| 167 | |
| 168 | inline void recedeUnchecked() |
| 169 | { |
| 170 | Q_ASSERT_X(hasPrevious(), Q_FUNC_INFO, "iterator hasn't a previous item" ); |
| 171 | |
| 172 | if (Q_UNLIKELY((--pos)->isLowSurrogate())) { |
| 173 | Q_ASSERT(hasPrevious() && pos[-1].isHighSurrogate()); |
| 174 | --pos; |
| 175 | } |
| 176 | } |
| 177 | |
| 178 | inline char32_t peekPreviousUnchecked() const |
| 179 | { |
| 180 | Q_ASSERT_X(hasPrevious(), Q_FUNC_INFO, "iterator hasn't a previous item" ); |
| 181 | |
| 182 | if (Q_UNLIKELY(pos[-1].isLowSurrogate())) { |
| 183 | Q_ASSERT(less(i + 1, pos) && pos[-2].isHighSurrogate()); |
| 184 | return QChar::surrogateToUcs4(high: pos[-2], low: pos[-1]); |
| 185 | } |
| 186 | return pos[-1].unicode(); |
| 187 | } |
| 188 | |
| 189 | inline char32_t peekPrevious(char32_t invalidAs = QChar::ReplacementCharacter) const |
| 190 | { |
| 191 | Q_ASSERT_X(hasPrevious(), Q_FUNC_INFO, "iterator hasn't a previous item" ); |
| 192 | |
| 193 | if (Q_UNLIKELY(pos[-1].isSurrogate())) { |
| 194 | if (Q_LIKELY(pos[-1].isLowSurrogate())) { |
| 195 | const QChar *high = pos - 2; |
| 196 | if (Q_LIKELY(high != i - 1 && high->isHighSurrogate())) |
| 197 | return QChar::surrogateToUcs4(high: *high, low: pos[-1]); |
| 198 | } |
| 199 | return invalidAs; |
| 200 | } |
| 201 | |
| 202 | return pos[-1].unicode(); |
| 203 | } |
| 204 | |
| 205 | inline char32_t previousUnchecked() |
| 206 | { |
| 207 | Q_ASSERT_X(hasPrevious(), Q_FUNC_INFO, "iterator hasn't a previous item" ); |
| 208 | |
| 209 | const QChar cur = *--pos; |
| 210 | if (Q_UNLIKELY(cur.isLowSurrogate())) { |
| 211 | Q_ASSERT(hasPrevious() && pos[-1].isHighSurrogate()); |
| 212 | return QChar::surrogateToUcs4(high: *--pos, low: cur); |
| 213 | } |
| 214 | return cur.unicode(); |
| 215 | } |
| 216 | |
| 217 | inline char32_t previous(char32_t invalidAs = QChar::ReplacementCharacter) |
| 218 | { |
| 219 | Q_ASSERT_X(hasPrevious(), Q_FUNC_INFO, "iterator hasn't a previous item" ); |
| 220 | |
| 221 | const QChar uc = *--pos; |
| 222 | if (Q_UNLIKELY(uc.isSurrogate())) { |
| 223 | if (Q_LIKELY(uc.isLowSurrogate() && hasPrevious() && pos[-1].isHighSurrogate())) |
| 224 | return QChar::surrogateToUcs4(high: *--pos, low: uc); |
| 225 | return invalidAs; |
| 226 | } |
| 227 | |
| 228 | return uc.unicode(); |
| 229 | } |
| 230 | }; |
| 231 | |
| 232 | QT_END_NAMESPACE |
| 233 | |
| 234 | #endif // QSTRINGITERATOR_H |
| 235 | |