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_CORE_H_2675DCD0_9480_4c0c_B92A_CC14C027B731
29 #define UTF8_FOR_CPP_CORE_H_2675DCD0_9480_4c0c_B92A_CC14C027B731
35 // The typedefs for 8-bit, 16-bit and 32-bit unsigned integers
36 // You may need to change them to match your system.
37 // These typedefs have the same names as ones from cstdint, or boost/cstdint
38 typedef unsigned char uint8_t;
39 typedef unsigned short uint16_t;
40 typedef unsigned int uint32_t;
42 // Helper code - not intended to be directly called by the library users. May be changed at any time
46 // Leading (high) surrogates: 0xd800 - 0xdbff
47 // Trailing (low) surrogates: 0xdc00 - 0xdfff
48 const uint16_t LEAD_SURROGATE_MIN
= 0xd800u
;
49 const uint16_t LEAD_SURROGATE_MAX
= 0xdbffu
;
50 const uint16_t TRAIL_SURROGATE_MIN
= 0xdc00u
;
51 const uint16_t TRAIL_SURROGATE_MAX
= 0xdfffu
;
52 const uint16_t LEAD_OFFSET
= LEAD_SURROGATE_MIN
- (0x10000 >> 10);
53 const uint32_t SURROGATE_OFFSET
= 0x10000u
- (LEAD_SURROGATE_MIN
<< 10) - TRAIL_SURROGATE_MIN
;
55 // Maximum valid value for a Unicode code point
56 const uint32_t CODE_POINT_MAX
= 0x0010ffffu
;
58 template<typename octet_type
>
59 inline uint8_t mask8(octet_type oc
)
61 return static_cast<uint8_t>(0xff & oc
);
63 template<typename u16_type
>
64 inline uint16_t mask16(u16_type oc
)
66 return static_cast<uint16_t>(0xffff & oc
);
68 template<typename octet_type
>
69 inline bool is_trail(octet_type oc
)
71 return ((mask8(oc
) >> 6) == 0x2);
74 template <typename u16
>
75 inline bool is_lead_surrogate(u16 cp
)
77 return (cp
>= LEAD_SURROGATE_MIN
&& cp
<= LEAD_SURROGATE_MAX
);
80 template <typename u16
>
81 inline bool is_trail_surrogate(u16 cp
)
83 return (cp
>= TRAIL_SURROGATE_MIN
&& cp
<= TRAIL_SURROGATE_MAX
);
86 template <typename u16
>
87 inline bool is_surrogate(u16 cp
)
89 return (cp
>= LEAD_SURROGATE_MIN
&& cp
<= TRAIL_SURROGATE_MAX
);
92 template <typename u32
>
93 inline bool is_code_point_valid(u32 cp
)
95 return (cp
<= CODE_POINT_MAX
&& !is_surrogate(cp
) && cp
!= 0xfffe && cp
!= 0xffff);
98 template <typename octet_iterator
>
99 inline typename
std::iterator_traits
<octet_iterator
>::difference_type
100 sequence_length(octet_iterator lead_it
)
102 uint8_t lead
= mask8(*lead_it
);
105 else if ((lead
>> 5) == 0x6)
107 else if ((lead
>> 4) == 0xe)
109 else if ((lead
>> 3) == 0x1e)
115 inline bool is_overlong_sequence(uint32_t cp
, int length
)
121 else if (cp
< 0x800) {
125 else if (cp
< 0x10000) {
133 enum utf_error
{UTF8_OK
, NOT_ENOUGH_ROOM
, INVALID_LEAD
, INCOMPLETE_SEQUENCE
, OVERLONG_SEQUENCE
, INVALID_CODE_POINT
};
135 /// get_sequence_x functions decode utf-8 sequences of the length x
137 template <typename octet_iterator
>
138 utf_error
get_sequence_1(octet_iterator
& it
, octet_iterator end
, uint32_t* code_point
)
142 *code_point
= mask8(*it
);
145 return NOT_ENOUGH_ROOM
;
148 template <typename octet_iterator
>
149 utf_error
get_sequence_2(octet_iterator
& it
, octet_iterator end
, uint32_t* code_point
)
151 utf_error ret_code
= NOT_ENOUGH_ROOM
;
154 uint32_t cp
= mask8(*it
);
157 cp
= ((cp
<< 6) & 0x7ff) + ((*it
) & 0x3f);
164 ret_code
= INCOMPLETE_SEQUENCE
;
167 ret_code
= NOT_ENOUGH_ROOM
;
173 template <typename octet_iterator
>
174 utf_error
get_sequence_3(octet_iterator
& it
, octet_iterator end
, uint32_t* code_point
)
176 utf_error ret_code
= NOT_ENOUGH_ROOM
;
179 uint32_t cp
= mask8(*it
);
182 cp
= ((cp
<< 12) & 0xffff) + ((mask8(*it
) << 6) & 0xfff);
192 ret_code
= INCOMPLETE_SEQUENCE
;
195 ret_code
= NOT_ENOUGH_ROOM
;
198 ret_code
= INCOMPLETE_SEQUENCE
;
201 ret_code
= NOT_ENOUGH_ROOM
;
207 template <typename octet_iterator
>
208 utf_error
get_sequence_4(octet_iterator
& it
, octet_iterator end
, uint32_t* code_point
)
210 utf_error ret_code
= NOT_ENOUGH_ROOM
;
213 uint32_t cp
= mask8(*it
);
216 cp
= ((cp
<< 18) & 0x1fffff) + ((mask8(*it
) << 12) & 0x3ffff);
219 cp
+= (mask8(*it
) << 6) & 0xfff;
229 ret_code
= INCOMPLETE_SEQUENCE
;
232 ret_code
= NOT_ENOUGH_ROOM
;
235 ret_code
= INCOMPLETE_SEQUENCE
;
238 ret_code
= NOT_ENOUGH_ROOM
;
241 ret_code
= INCOMPLETE_SEQUENCE
;
244 ret_code
= NOT_ENOUGH_ROOM
;
250 template <typename octet_iterator
>
251 utf_error
validate_next(octet_iterator
& it
, octet_iterator end
, uint32_t* code_point
)
253 // Save the original value of it so we can go back in case of failure
254 // Of course, it does not make much sense with i.e. stream iterators
255 octet_iterator original_it
= it
;
258 // Determine the sequence length based on the lead octet
259 typedef typename
std::iterator_traits
<octet_iterator
>::difference_type octet_difference_type
;
260 octet_difference_type length
= sequence_length(it
);
264 // Now that we have a valid sequence length, get trail octets and calculate the code point
265 utf_error err
= UTF8_OK
;
268 err
= get_sequence_1(it
, end
, &cp
);
271 err
= get_sequence_2(it
, end
, &cp
);
274 err
= get_sequence_3(it
, end
, &cp
);
277 err
= get_sequence_4(it
, end
, &cp
);
281 if (err
== UTF8_OK
) {
282 // Decoding succeeded. Now, security checks...
283 if (is_code_point_valid(cp
)) {
284 if (!is_overlong_sequence(cp
, length
)){
285 // Passed! Return here.
292 err
= OVERLONG_SEQUENCE
;
295 err
= INVALID_CODE_POINT
;
298 // Failure branch - restore the original value of the iterator
303 template <typename octet_iterator
>
304 inline utf_error
validate_next(octet_iterator
& it
, octet_iterator end
) {
305 return validate_next(it
, end
, 0);
308 } // namespace internal
310 /// The library API - functions intended to be called by the users
313 const uint8_t bom
[] = {0xef, 0xbb, 0xbf};
315 template <typename octet_iterator
>
316 octet_iterator
find_invalid(octet_iterator start
, octet_iterator end
)
318 octet_iterator result
= start
;
319 while (result
!= end
) {
320 internal::utf_error err_code
= internal::validate_next(result
, end
);
321 if (err_code
!= internal::UTF8_OK
)
327 template <typename octet_iterator
>
328 inline bool is_valid(octet_iterator start
, octet_iterator end
)
330 return (find_invalid(start
, end
) == end
);
333 template <typename octet_iterator
>
334 inline bool is_bom (octet_iterator it
)
337 (internal::mask8(*it
++)) == bom
[0] &&
338 (internal::mask8(*it
++)) == bom
[1] &&
339 (internal::mask8(*it
)) == bom
[2]
344 #endif // header guard