1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
|
#ifndef UNICODE_H_
#define UNICODE_H_
#define UNICODE_BOX_CHARACTER 0x2610
#define UNICODE_CODE_POINTS 0x110000 // number of Unicode code points
static bool unicode_is_start_of_code_point(u8 byte) {
// see https://en.wikipedia.org/wiki/UTF-8#Encoding
// continuation bytes are of the form 10xxxxxx
return (byte & 0xC0) != 0x80;
}
static bool unicode_is_continuation_byte(u8 byte) {
return (byte & 0xC0) == 0x80;
}
// A lot like mbrtoc32. Doesn't depend on the locale though, for one thing.
// *c will be filled with the next UTF-8 code point in `str`. `bytes` refers to the maximum
// number of bytes that can be read from `str` (note: this function will never read past a null
// byte, even if `bytes` indicates that it could).
// Returns:
// 0 - if a null character was encountered or if `bytes == 0`
// (size_t)-1 - on invalid UTF-8
// (size_t)-2 - on incomplete code point (str should be longer)
// other - the number of bytes read from `str`.
static size_t unicode_utf8_to_utf32(char32_t *c, const char *str, size_t bytes) {
*c = 0;
if (bytes == 0) {
return 0;
}
// it's easier to do things with unsigned integers
const u8 *p = (const u8 *)str;
u8 first_byte = *p;
if (first_byte & 0x80) {
if ((first_byte & 0xE0) == 0xC0) {
// two-byte code point
if (bytes >= 2) {
++p;
u32 second_byte = *p;
if ((second_byte & 0xC0) != 0x80) return (size_t)-1;
u32 value = ((u32)first_byte & 0x1F) << 6
| (second_byte & 0x3F);
*c = (char32_t)value;
return 2;
} else {
// incomplete code point
return (size_t)-2;
}
}
if ((first_byte & 0xF0) == 0xE0) {
// three-byte code point
if (bytes >= 3) {
++p;
u32 second_byte = *p;
if ((second_byte & 0xC0) != 0x80) return (size_t)-1;
++p;
u32 third_byte = *p;
if ((third_byte & 0xC0) != 0x80) return (size_t)-1;
u32 value = ((u32)first_byte & 0x0F) << 12
| (second_byte & 0x3F) << 6
| (third_byte & 0x3F);
if (value < 0xD800 || value > 0xDFFF) {
*c = (char32_t)value;
return 3;
} else {
// reserved for UTF-16 surrogate halves
return (size_t)-1;
}
} else {
// incomplete
return (size_t)-2;
}
}
if ((first_byte & 0xF8) == 0xF0) {
// four-byte code point
if (bytes >= 4) {
++p;
u32 second_byte = *p;
if ((second_byte & 0xC0) != 0x80) return (size_t)-1;
++p;
u32 third_byte = *p;
if ((third_byte & 0xC0) != 0x80) return (size_t)-1;
++p;
u32 fourth_byte = *p;
if ((fourth_byte & 0xC0) != 0x80) return (size_t)-1;
u32 value = ((u32)first_byte & 0x07) << 18
| (second_byte & 0x3F) << 12
| (third_byte & 0x3F) << 6
| (fourth_byte & 0x3F);
if (value >= 0xD800 && value <= 0xDFFF) {
// reserved for UTF-16 surrogate halves
return (size_t)-1;
} else if (value <= 0x10FFFF) {
*c = (char32_t)value;
return 4;
} else {
// Code points this big can't be encoded by UTF-16 so are invalid UTF-8.
return (size_t)-1;
}
} else {
// incomplete
return (size_t)-2;
}
}
// invalid UTF-8
return (size_t)-1;
} else {
// ASCII character
if (first_byte == 0) {
return 0;
}
*c = first_byte;
return 1;
}
}
// A lot like c32rtomb
// Converts a UTF-32 codepoint to a UTF-8 string. Writes at most 4 bytes to s.
// NOTE: It is YOUR JOB to null-terminate your string if the UTF-32 isn't null-terminated!
// Returns the number of bytes written to s, or (size_t)-1 on invalid UTF-32.
static size_t unicode_utf32_to_utf8(char *s, char32_t c32) {
u8 *p = (u8 *)s;
if (c32 <= 0x7F) {
// ASCII
*p = (u8)c32;
return 1;
} else if (c32 <= 0x7FF) {
// two bytes needed
*p++ = (u8)(0xC0 | (c32 >> 6));
*p = (u8)(0x80 | (c32 & 0x3F));
return 2;
} else if (c32 <= 0x7FFF) {
if (c32 < 0xD800 || c32 > 0xDFFF) {
*p++ = (u8)(0xE0 | ( c32 >> 12));
*p++ = (u8)(0x80 | ((c32 >> 6) & 0x3F));
*p = (u8)(0x80 | ( c32 & 0x3F));
return 3;
} else {
// UTF-16 surrogate halves
*p = 0;
return (size_t)-1;
}
} else if (c32 <= 0x10FFFF) {
*p++ = (u8)(0xF0 | ( c32 >> 18));
*p++ = (u8)(0x80 | ((c32 >> 12) & 0x3F));
*p++ = (u8)(0x80 | ((c32 >> 6) & 0x3F));
*p = (u8)(0x80 | ( c32 & 0x3F));
return 4;
} else {
// code point too big
*p = 0;
return (size_t)-1;
}
}
// get the number of UTF-16 codepoints needed to encode `str`.
// returns (size_t)-1 on bad UTF-8
static size_t unicode_utf16_len(const char *str) {
size_t len = 0;
char32_t c = 0;
while (*str) {
size_t n = unicode_utf8_to_utf32(&c, str, 4);
if (n >= (size_t)-2)
return (size_t)-1;
if (c >= 0x10000)
len += 2;
else
len += 1;
str += n;
}
return len;
}
// returns the UTF-8 offset from `str` which corresponds to a UTF-16 offset of utf16_offset (rounds down if utf16_offset is in the middle of a codepoint).
// returns strlen(str) if utf16_offset == unicode_utf16_len(str)
// returns (size_t)-1 on bad UTF-8, or if utf16_offset > unicode_utf16_len(str)
static size_t unicode_utf16_to_utf8_offset(const char *str, size_t utf16_offset) {
size_t offset = 0;
char32_t c = 0;
while (*str) {
size_t n = unicode_utf8_to_utf32(&c, str, 4);
if (n >= (size_t)-2)
return (size_t)-1;
size_t u = c >= 0x10000 ? 2 : 1;
if (utf16_offset < u)
return offset;
utf16_offset -= u;
offset += n;
str += n;
}
if (utf16_offset == 0)
return offset;
return SIZE_MAX;
}
#endif // UNICODE_H_
|
