The Tor Browser / file comparison
comparison: gfx/graphite2/src/inc/UtfCodec.h
gfx/graphite2/src/inc/UtfCodec.h
- branch
- TOR_BUG_9701
- changeset 15
- b8a032363ba2
equal
deleted
inserted
replaced
| -1:000000000000 | 0:8f303fea2069 |
|---|---|
| 1 /* GRAPHITE2 LICENSING | |
| 2 | |
| 3 Copyright 2011, SIL International | |
| 4 All rights reserved. | |
| 5 | |
| 6 This library is free software; you can redistribute it and/or modify | |
| 7 it under the terms of the GNU Lesser General Public License as published | |
| 8 by the Free Software Foundation; either version 2.1 of License, or | |
| 9 (at your option) any later version. | |
| 10 | |
| 11 This program is distributed in the hope that it will be useful, | |
| 12 but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
| 14 Lesser General Public License for more details. | |
| 15 | |
| 16 You should also have received a copy of the GNU Lesser General Public | |
| 17 License along with this library in the file named "LICENSE". | |
| 18 If not, write to the Free Software Foundation, 51 Franklin Street, | |
| 19 Suite 500, Boston, MA 02110-1335, USA or visit their web page on the | |
| 20 internet at http://www.fsf.org/licenses/lgpl.html. | |
| 21 | |
| 22 Alternatively, the contents of this file may be used under the terms of the | |
| 23 Mozilla Public License (http://mozilla.org/MPL) or the GNU General Public | |
| 24 License, as published by the Free Software Foundation, either version 2 | |
| 25 of the License or (at your option) any later version. | |
| 26 */ | |
| 27 #pragma once | |
| 28 | |
| 29 #include <cstdlib> | |
| 30 #include "inc/Main.h" | |
| 31 | |
| 32 namespace graphite2 { | |
| 33 | |
| 34 typedef uint32 uchar_t; | |
| 35 | |
| 36 template <int N> | |
| 37 struct _utf_codec | |
| 38 { | |
| 39 typedef uchar_t codeunit_t; | |
| 40 | |
| 41 static void put(codeunit_t * cp, const uchar_t , int8 & len) throw(); | |
| 42 static uchar_t get(const codeunit_t * cp, int8 & len) throw(); | |
| 43 }; | |
| 44 | |
| 45 | |
| 46 template <> | |
| 47 struct _utf_codec<32> | |
| 48 { | |
| 49 private: | |
| 50 static const uchar_t limit = 0x110000; | |
| 51 public: | |
| 52 typedef uint32 codeunit_t; | |
| 53 | |
| 54 inline | |
| 55 static void put(codeunit_t * cp, const uchar_t usv, int8 & l) throw() | |
| 56 { | |
| 57 *cp = usv; l = 1; | |
| 58 } | |
| 59 | |
| 60 inline | |
| 61 static uchar_t get(const codeunit_t * cp, int8 & l) throw() | |
| 62 { | |
| 63 if (cp[0] < limit) { l = 1; return cp[0]; } | |
| 64 else { l = -1; return 0xFFFD; } | |
| 65 } | |
| 66 }; | |
| 67 | |
| 68 | |
| 69 template <> | |
| 70 struct _utf_codec<16> | |
| 71 { | |
| 72 private: | |
| 73 static const int32 lead_offset = 0xD800 - (0x10000 >> 10); | |
| 74 static const int32 surrogate_offset = 0x10000 - (0xD800 << 10) - 0xDC00; | |
| 75 public: | |
| 76 typedef uint16 codeunit_t; | |
| 77 | |
| 78 inline | |
| 79 static void put(codeunit_t * cp, const uchar_t usv, int8 & l) throw() | |
| 80 { | |
| 81 if (usv < 0x10000) { l = 1; cp[0] = codeunit_t(usv); } | |
| 82 else | |
| 83 { | |
| 84 cp[0] = codeunit_t(lead_offset + (usv >> 10)); | |
| 85 cp[1] = codeunit_t(0xDC00 + (usv & 0x3FF)); | |
| 86 l = 2; | |
| 87 } | |
| 88 } | |
| 89 | |
| 90 inline | |
| 91 static uchar_t get(const codeunit_t * cp, int8 & l) throw() | |
| 92 { | |
| 93 const uint32 uh = cp[0]; | |
| 94 l = 1; | |
| 95 | |
| 96 if (0xD800 > uh || uh > 0xDFFF) { return uh; } | |
| 97 const uint32 ul = cp[1]; | |
| 98 if (uh > 0xDBFF || 0xDC00 > ul || ul > 0xDFFF) { l = -1; return 0xFFFD; } | |
| 99 ++l; | |
| 100 return (uh<<10) + ul + surrogate_offset; | |
| 101 } | |
| 102 }; | |
| 103 | |
| 104 | |
| 105 template <> | |
| 106 struct _utf_codec<8> | |
| 107 { | |
| 108 private: | |
| 109 static const int8 sz_lut[16]; | |
| 110 static const byte mask_lut[5]; | |
| 111 | |
| 112 | |
| 113 public: | |
| 114 typedef uint8 codeunit_t; | |
| 115 | |
| 116 inline | |
| 117 static void put(codeunit_t * cp, const uchar_t usv, int8 & l) throw() | |
| 118 { | |
| 119 if (usv < 0x80) {l = 1; cp[0] = usv; return; } | |
| 120 if (usv < 0x0800) {l = 2; cp[0] = 0xC0 + (usv >> 6); cp[1] = 0x80 + (usv & 0x3F); return; } | |
| 121 if (usv < 0x10000) {l = 3; cp[0] = 0xE0 + (usv >> 12); cp[1] = 0x80 + ((usv >> 6) & 0x3F); cp[2] = 0x80 + (usv & 0x3F); return; } | |
| 122 else {l = 4; cp[0] = 0xF0 + (usv >> 18); cp[1] = 0x80 + ((usv >> 12) & 0x3F); cp[2] = 0x80 + ((usv >> 6) & 0x3F); cp[3] = 0x80 + (usv & 0x3F); return; } | |
| 123 } | |
| 124 | |
| 125 inline | |
| 126 static uchar_t get(const codeunit_t * cp, int8 & l) throw() | |
| 127 { | |
| 128 const int8 seq_sz = sz_lut[*cp >> 4]; | |
| 129 uchar_t u = *cp & mask_lut[seq_sz]; | |
| 130 l = 1; | |
| 131 bool toolong = false; | |
| 132 | |
| 133 switch(seq_sz) { | |
| 134 case 4: u <<= 6; u |= *++cp & 0x3F; if (*cp >> 6 != 2) break; ++l; toolong = (u < 0x10); // no break | |
| 135 case 3: u <<= 6; u |= *++cp & 0x3F; if (*cp >> 6 != 2) break; ++l; toolong |= (u < 0x20); // no break | |
| 136 case 2: u <<= 6; u |= *++cp & 0x3F; if (*cp >> 6 != 2) break; ++l; toolong |= (u < 0x80); // no break | |
| 137 case 1: break; | |
| 138 case 0: l = -1; return 0xFFFD; | |
| 139 } | |
| 140 | |
| 141 if (l != seq_sz || toolong) | |
| 142 { | |
| 143 l = -l; | |
| 144 return 0xFFFD; | |
| 145 } | |
| 146 return u; | |
| 147 } | |
| 148 }; | |
| 149 | |
| 150 | |
| 151 template <typename C> | |
| 152 class _utf_iterator | |
| 153 { | |
| 154 typedef _utf_codec<sizeof(C)*8> codec; | |
| 155 | |
| 156 C * cp; | |
| 157 mutable int8 sl; | |
| 158 | |
| 159 public: | |
| 160 typedef C codeunit_type; | |
| 161 typedef uchar_t value_type; | |
| 162 typedef uchar_t * pointer; | |
| 163 | |
| 164 class reference | |
| 165 { | |
| 166 const _utf_iterator & _i; | |
| 167 | |
| 168 reference(const _utf_iterator & i): _i(i) {} | |
| 169 public: | |
| 170 operator value_type () const throw () { return codec::get(_i.cp, _i.sl); } | |
| 171 reference & operator = (const value_type usv) throw() { codec::put(_i.cp, usv, _i.sl); return *this; } | |
| 172 | |
| 173 friend class _utf_iterator; | |
| 174 }; | |
| 175 | |
| 176 | |
| 177 _utf_iterator(const void * us=0) : cp(reinterpret_cast<C *>(const_cast<void *>(us))), sl(1) { } | |
| 178 | |
| 179 _utf_iterator & operator ++ () { cp += abs(sl); return *this; } | |
| 180 _utf_iterator operator ++ (int) { _utf_iterator tmp(*this); operator++(); return tmp; } | |
| 181 | |
| 182 bool operator == (const _utf_iterator & rhs) const throw() { return cp >= rhs.cp; } | |
| 183 bool operator != (const _utf_iterator & rhs) const throw() { return !operator==(rhs); } | |
| 184 | |
| 185 reference operator * () const throw() { return *this; } | |
| 186 pointer operator ->() const throw() { return &operator *(); } | |
| 187 | |
| 188 operator codeunit_type * () const throw() { return cp; } | |
| 189 | |
| 190 bool error() const throw() { return sl < 1; } | |
| 191 }; | |
| 192 | |
| 193 template <typename C> | |
| 194 struct utf | |
| 195 { | |
| 196 typedef typename _utf_codec<sizeof(C)*8>::codeunit_t codeunit_t; | |
| 197 | |
| 198 typedef _utf_iterator<C> iterator; | |
| 199 typedef _utf_iterator<const C> const_iterator; | |
| 200 }; | |
| 201 | |
| 202 | |
| 203 typedef utf<uint32> utf32; | |
| 204 typedef utf<uint16> utf16; | |
| 205 typedef utf<uint8> utf8; | |
| 206 | |
| 207 } // namespace graphite2 |
