873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
1 |
// Copyright 2006 Nemanja Trifunovic
|
2 |
||
3 |
/*
|
|
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:
|
|
10 |
||
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.
|
|
17 |
||
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.
|
|
25 |
*/
|
|
26 |
||
27 |
||
1122.2.10
by Monty Taylor
Fixed all of the include guards. |
28 |
#ifndef DRIZZLED_UTF8_UNCHECKED_H
|
29 |
#define DRIZZLED_UTF8_UNCHECKED_H
|
|
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
30 |
|
31 |
#include "drizzled/utf8/core.h" |
|
32 |
||
1823.4.2
by Monty Taylor
Cleaned things up a little bit namespace wise |
33 |
namespace drizzled |
34 |
{
|
|
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
35 |
namespace utf8 |
36 |
{
|
|
37 |
namespace unchecked |
|
38 |
{
|
|
39 |
template <typename octet_iterator> |
|
40 |
octet_iterator append(uint32_t cp, octet_iterator result) |
|
41 |
{
|
|
42 |
if (cp < 0x80) // one octet |
|
43 |
*(result++) = static_cast<uint8_t>(cp); |
|
44 |
else if (cp < 0x800) { // two octets |
|
45 |
*(result++) = static_cast<uint8_t>((cp >> 6) | 0xc0); |
|
46 |
*(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80); |
|
47 |
}
|
|
48 |
else if (cp < 0x10000) { // three octets |
|
49 |
*(result++) = static_cast<uint8_t>((cp >> 12) | 0xe0); |
|
1823.4.3
by Monty Taylor
Merged in new lib version. |
50 |
*(result++) = static_cast<uint8_t>(((cp >> 6) & 0x3f) | 0x80); |
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
51 |
*(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80); |
52 |
}
|
|
53 |
else { // four octets |
|
54 |
*(result++) = static_cast<uint8_t>((cp >> 18) | 0xf0); |
|
1823.4.3
by Monty Taylor
Merged in new lib version. |
55 |
*(result++) = static_cast<uint8_t>(((cp >> 12) & 0x3f)| 0x80); |
56 |
*(result++) = static_cast<uint8_t>(((cp >> 6) & 0x3f) | 0x80); |
|
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
57 |
*(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80); |
58 |
}
|
|
59 |
return result; |
|
60 |
}
|
|
61 |
||
62 |
template <typename octet_iterator> |
|
63 |
uint32_t next(octet_iterator& it) |
|
64 |
{
|
|
65 |
uint32_t cp = internal::mask8(*it); |
|
66 |
typename std::iterator_traits<octet_iterator>::difference_type length = utf8::internal::sequence_length(it); |
|
67 |
switch (length) { |
|
68 |
case 1: |
|
69 |
break; |
|
70 |
case 2: |
|
71 |
it++; |
|
72 |
cp = ((cp << 6) & 0x7ff) + ((*it) & 0x3f); |
|
73 |
break; |
|
74 |
case 3: |
|
75 |
++it; |
|
76 |
cp = ((cp << 12) & 0xffff) + ((internal::mask8(*it) << 6) & 0xfff); |
|
77 |
++it; |
|
78 |
cp += (*it) & 0x3f; |
|
79 |
break; |
|
80 |
case 4: |
|
81 |
++it; |
|
82 |
cp = ((cp << 18) & 0x1fffff) + ((internal::mask8(*it) << 12) & 0x3ffff); |
|
83 |
++it; |
|
84 |
cp += (internal::mask8(*it) << 6) & 0xfff; |
|
85 |
++it; |
|
86 |
cp += (*it) & 0x3f; |
|
87 |
break; |
|
88 |
}
|
|
89 |
++it; |
|
90 |
return cp; |
|
91 |
}
|
|
92 |
||
93 |
template <typename octet_iterator> |
|
94 |
uint32_t peek_next(octet_iterator it) |
|
95 |
{
|
|
96 |
return next(it); |
|
97 |
}
|
|
98 |
||
99 |
template <typename octet_iterator> |
|
100 |
uint32_t prior(octet_iterator& it) |
|
101 |
{
|
|
102 |
while (internal::is_trail(*(--it))) ; |
|
103 |
octet_iterator temp = it; |
|
104 |
return next(temp); |
|
105 |
}
|
|
106 |
||
107 |
// Deprecated in versions that include prior, but only for the sake of consistency (see utf8::previous)
|
|
108 |
template <typename octet_iterator> |
|
109 |
inline uint32_t previous(octet_iterator& it) |
|
110 |
{
|
|
111 |
return prior(it); |
|
112 |
}
|
|
113 |
||
114 |
template <typename octet_iterator, typename distance_type> |
|
115 |
void advance (octet_iterator& it, distance_type n) |
|
116 |
{
|
|
117 |
for (distance_type i = 0; i < n; ++i) |
|
118 |
next(it); |
|
119 |
}
|
|
120 |
||
121 |
template <typename octet_iterator> |
|
122 |
typename std::iterator_traits<octet_iterator>::difference_type |
|
123 |
distance (octet_iterator first, octet_iterator last) |
|
124 |
{
|
|
125 |
typename std::iterator_traits<octet_iterator>::difference_type dist; |
|
126 |
for (dist = 0; first < last; ++dist) |
|
127 |
next(first); |
|
128 |
return dist; |
|
129 |
}
|
|
130 |
||
131 |
template <typename u16bit_iterator, typename octet_iterator> |
|
132 |
octet_iterator utf16to8 (u16bit_iterator start, u16bit_iterator end, octet_iterator result) |
|
133 |
{
|
|
134 |
while (start != end) { |
|
135 |
uint32_t cp = internal::mask16(*start++); |
|
136 |
// Take care of surrogate pairs first
|
|
1823.4.3
by Monty Taylor
Merged in new lib version. |
137 |
if (internal::is_lead_surrogate(cp)) { |
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
138 |
uint32_t trail_surrogate = internal::mask16(*start++); |
139 |
cp = (cp << 10) + trail_surrogate + internal::SURROGATE_OFFSET; |
|
140 |
}
|
|
141 |
result = append(cp, result); |
|
142 |
}
|
|
143 |
return result; |
|
144 |
}
|
|
145 |
||
146 |
template <typename u16bit_iterator, typename octet_iterator> |
|
147 |
u16bit_iterator utf8to16 (octet_iterator start, octet_iterator end, u16bit_iterator result) |
|
148 |
{
|
|
1823.4.3
by Monty Taylor
Merged in new lib version. |
149 |
while (start < end) { |
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
150 |
uint32_t cp = next(start); |
151 |
if (cp > 0xffff) { //make a surrogate pair |
|
152 |
*result++ = static_cast<uint16_t>((cp >> 10) + internal::LEAD_OFFSET); |
|
153 |
*result++ = static_cast<uint16_t>((cp & 0x3ff) + internal::TRAIL_SURROGATE_MIN); |
|
154 |
}
|
|
155 |
else
|
|
156 |
*result++ = static_cast<uint16_t>(cp); |
|
157 |
}
|
|
158 |
return result; |
|
159 |
}
|
|
160 |
||
161 |
template <typename octet_iterator, typename u32bit_iterator> |
|
162 |
octet_iterator utf32to8 (u32bit_iterator start, u32bit_iterator end, octet_iterator result) |
|
163 |
{
|
|
164 |
while (start != end) |
|
165 |
result = append(*(start++), result); |
|
166 |
||
167 |
return result; |
|
168 |
}
|
|
169 |
||
170 |
template <typename octet_iterator, typename u32bit_iterator> |
|
171 |
u32bit_iterator utf8to32 (octet_iterator start, octet_iterator end, u32bit_iterator result) |
|
172 |
{
|
|
173 |
while (start < end) |
|
174 |
(*result++) = next(start); |
|
175 |
||
176 |
return result; |
|
177 |
}
|
|
178 |
||
179 |
// The iterator class
|
|
180 |
template <typename octet_iterator> |
|
181 |
class iterator : public std::iterator <std::bidirectional_iterator_tag, uint32_t> { |
|
182 |
octet_iterator it; |
|
183 |
public: |
|
184 |
iterator () {}; |
|
185 |
explicit iterator (const octet_iterator& octet_it): it(octet_it) {} |
|
186 |
// the default "big three" are OK
|
|
187 |
octet_iterator base () const { return it; } |
|
188 |
uint32_t operator * () const |
|
189 |
{
|
|
190 |
octet_iterator temp = it; |
|
191 |
return next(temp); |
|
192 |
}
|
|
193 |
bool operator == (const iterator& rhs) const |
|
194 |
{
|
|
195 |
return (it == rhs.it); |
|
196 |
}
|
|
197 |
bool operator != (const iterator& rhs) const |
|
198 |
{
|
|
199 |
return !(operator == (rhs)); |
|
200 |
}
|
|
201 |
iterator& operator ++ () |
|
202 |
{
|
|
203 |
std::advance(it, internal::sequence_length(it)); |
|
204 |
return *this; |
|
205 |
}
|
|
206 |
iterator operator ++ (int) |
|
207 |
{
|
|
208 |
iterator temp = *this; |
|
209 |
std::advance(it, internal::sequence_length(it)); |
|
210 |
return temp; |
|
211 |
}
|
|
212 |
iterator& operator -- () |
|
213 |
{
|
|
214 |
prior(it); |
|
215 |
return *this; |
|
216 |
}
|
|
217 |
iterator operator -- (int) |
|
218 |
{
|
|
219 |
iterator temp = *this; |
|
220 |
prior(it); |
|
221 |
return temp; |
|
222 |
}
|
|
223 |
}; // class iterator |
|
224 |
||
225 |
} // namespace utf8::unchecked |
|
226 |
} // namespace utf8 |
|
1823.4.3
by Monty Taylor
Merged in new lib version. |
227 |
} // namespace drizzled |
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
228 |
|
229 |
||
1122.2.10
by Monty Taylor
Fixed all of the include guards. |
230 |
#endif /* DRIZZLED_UTF8_UNCHECKED_H */ |
873.2.27
by Monty Taylor
Added in utf8cpp. We want this to be an external lib, but it needs packaged. |
231 |