1 // Copyright 2006 Nemanja Trifunovic
4 Permission is hereby granted, free of charge, to any person or organization
5 obtaining a copy of the software and accompanying documentation covered by
6 this license (the "Software") to use, reproduce, display, distribute,
7 execute, and transmit the Software, and to prepare derivative works of the
8 Software, and to permit third-parties to whom the Software is furnished to
9 do so, all subject to the following:
11 The copyright notices in the Software and this entire statement, including
12 the above license grant, this restriction and the following disclaimer,
13 must be included in all copies of the Software, in whole or in part, and
14 all derivative works of the Software, unless such copies or derivative
15 works are solely in the form of machine-executable object code generated by
16 a source language processor.
18 THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
19 IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
20 FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT
21 SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE
22 FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
23 ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
24 DEALINGS IN THE SOFTWARE.
28 #ifndef UTF8_FOR_CPP_UNCHECKED_H_2675DCD0_9480_4c0c_B92A_CC14C027B731
29 #define UTF8_FOR_CPP_UNCHECKED_H_2675DCD0_9480_4c0c_B92A_CC14C027B731
37 template <typename octet_iterator>
38 octet_iterator append(uint32_t cp, octet_iterator result)
40 if (cp < 0x80) // one octet
41 *(result++) = static_cast<uint8_t>(cp);
42 else if (cp < 0x800) { // two octets
43 *(result++) = static_cast<uint8_t>((cp >> 6) | 0xc0);
44 *(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80);
46 else if (cp < 0x10000) { // three octets
47 *(result++) = static_cast<uint8_t>((cp >> 12) | 0xe0);
48 *(result++) = static_cast<uint8_t>(((cp >> 6) & 0x3f) | 0x80);
49 *(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80);
52 *(result++) = static_cast<uint8_t>((cp >> 18) | 0xf0);
53 *(result++) = static_cast<uint8_t>(((cp >> 12) & 0x3f)| 0x80);
54 *(result++) = static_cast<uint8_t>(((cp >> 6) & 0x3f) | 0x80);
55 *(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80);
60 template <typename octet_iterator, typename output_iterator>
61 output_iterator replace_invalid(octet_iterator start, octet_iterator end, output_iterator out, uint32_t replacement)
63 while (start != end) {
64 octet_iterator sequence_start = start;
65 internal::utf_error err_code = utf8::internal::validate_next(start, end);
67 case internal::UTF8_OK :
68 for (octet_iterator it = sequence_start; it != start; ++it)
71 case internal::NOT_ENOUGH_ROOM:
72 out = utf8::unchecked::append (replacement, out);
75 case internal::INVALID_LEAD:
76 out = utf8::unchecked::append (replacement, out);
79 case internal::INCOMPLETE_SEQUENCE:
80 case internal::OVERLONG_SEQUENCE:
81 case internal::INVALID_CODE_POINT:
82 out = utf8::unchecked::append (replacement, out);
84 // just one replacement mark for the sequence
85 while (start != end && utf8::internal::is_trail(*start))
93 template <typename octet_iterator, typename output_iterator>
94 inline output_iterator replace_invalid(octet_iterator start, octet_iterator end, output_iterator out)
96 static const uint32_t replacement_marker = utf8::internal::mask16(0xfffd);
97 return utf8::unchecked::replace_invalid(start, end, out, replacement_marker);
100 template <typename octet_iterator>
101 uint32_t next(octet_iterator& it)
103 uint32_t cp = utf8::internal::mask8(*it);
104 typename std::iterator_traits<octet_iterator>::difference_type length = utf8::internal::sequence_length(it);
110 cp = ((cp << 6) & 0x7ff) + ((*it) & 0x3f);
114 cp = ((cp << 12) & 0xffff) + ((utf8::internal::mask8(*it) << 6) & 0xfff);
120 cp = ((cp << 18) & 0x1fffff) + ((utf8::internal::mask8(*it) << 12) & 0x3ffff);
122 cp += (utf8::internal::mask8(*it) << 6) & 0xfff;
131 template <typename octet_iterator>
132 uint32_t peek_next(octet_iterator it)
134 return utf8::unchecked::next(it);
137 template <typename octet_iterator>
138 uint32_t prior(octet_iterator& it)
140 while (utf8::internal::is_trail(*(--it))) ;
141 octet_iterator temp = it;
142 return utf8::unchecked::next(temp);
145 template <typename octet_iterator, typename distance_type>
146 void advance (octet_iterator& it, distance_type n)
148 const distance_type zero(0);
151 for (distance_type i = n; i < zero; ++i)
152 utf8::unchecked::prior(it);
155 for (distance_type i = zero; i < n; ++i)
156 utf8::unchecked::next(it);
160 template <typename octet_iterator>
161 typename std::iterator_traits<octet_iterator>::difference_type
162 distance (octet_iterator first, octet_iterator last)
164 typename std::iterator_traits<octet_iterator>::difference_type dist;
165 for (dist = 0; first < last; ++dist)
166 utf8::unchecked::next(first);
170 template <typename u16bit_iterator, typename octet_iterator>
171 octet_iterator utf16to8 (u16bit_iterator start, u16bit_iterator end, octet_iterator result)
173 while (start != end) {
174 uint32_t cp = utf8::internal::mask16(*start++);
175 // Take care of surrogate pairs first
176 if (utf8::internal::is_lead_surrogate(cp)) {
177 uint32_t trail_surrogate = utf8::internal::mask16(*start++);
178 cp = (cp << 10) + trail_surrogate + internal::SURROGATE_OFFSET;
180 result = utf8::unchecked::append(cp, result);
185 template <typename u16bit_iterator, typename octet_iterator>
186 u16bit_iterator utf8to16 (octet_iterator start, octet_iterator end, u16bit_iterator result)
188 while (start < end) {
189 uint32_t cp = utf8::unchecked::next(start);
190 if (cp > 0xffff) { //make a surrogate pair
191 *result++ = static_cast<uint16_t>((cp >> 10) + internal::LEAD_OFFSET);
192 *result++ = static_cast<uint16_t>((cp & 0x3ff) + internal::TRAIL_SURROGATE_MIN);
195 *result++ = static_cast<uint16_t>(cp);
200 template <typename octet_iterator, typename u32bit_iterator>
201 octet_iterator utf32to8 (u32bit_iterator start, u32bit_iterator end, octet_iterator result)
204 result = utf8::unchecked::append(*(start++), result);
209 template <typename octet_iterator, typename u32bit_iterator>
210 u32bit_iterator utf8to32 (octet_iterator start, octet_iterator end, u32bit_iterator result)
213 (*result++) = utf8::unchecked::next(start);
218 // The iterator class
219 template <typename octet_iterator>
220 class iterator : public std::iterator <std::bidirectional_iterator_tag, uint32_t> {
224 explicit iterator (const octet_iterator& octet_it): it(octet_it) {}
225 // the default "big three" are OK
226 octet_iterator base () const { return it; }
227 uint32_t operator * () const
229 octet_iterator temp = it;
230 return utf8::unchecked::next(temp);
232 bool operator == (const iterator& rhs) const
234 return (it == rhs.it);
236 bool operator != (const iterator& rhs) const
238 return !(operator == (rhs));
240 iterator& operator ++ ()
242 ::std::advance(it, utf8::internal::sequence_length(it));
245 iterator operator ++ (int)
247 iterator temp = *this;
248 ::std::advance(it, utf8::internal::sequence_length(it));
251 iterator& operator -- ()
253 utf8::unchecked::prior(it);
256 iterator operator -- (int)
258 iterator temp = *this;
259 utf8::unchecked::prior(it);
264 } // namespace utf8::unchecked
268 #endif // header guard