1 /* Copyright (c) 2008-2009 Bjoern Hoehrmann <bjoern@hoehrmann.de>
2  *
3  * Permission is hereby granted, free of charge, to any person
4  * obtaining a copy of this software and associated documentation
5  * files (the "Software"), to deal in the Software without
6  * restriction, including without limitation the rights to use, copy,
7  * modify, merge, publish, distribute, sublicense, and/or sell copies
8  * of the Software, and to permit persons to whom the Software is
9  * furnished to do so, subject to the following conditions:
10  *
11  * The above copyright notice and this permission notice shall be
12  * included in all copies or substantial portions of the Software.
13  *
14  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
15  * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
16  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
17  * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
18  * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
19  * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
20  * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21  * SOFTWARE.
22  */
23 
24 #include "compat.h"
25 
26 #include <assert.h>
27 #include <stddef.h>
28 #include <stdint.h>
29 #include <wchar.h>
30 
31 #include "utf8.h"
32 
33 #define UTF8_ACCEPT 0
34 #define UTF8_REJECT 1
35 
36 static const uint8_t utf8d[] = {
37 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 00..1f
38 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 20..3f
39 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 40..5f
40 	0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 60..7f
41 	1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, // 80..9f
42 	7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, // a0..bf
43 	8,8,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, // c0..df
44 	0xa,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x4,0x3,0x3, // e0..ef
45 	0xb,0x6,0x6,0x6,0x5,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8, // f0..ff
46 	0x0,0x1,0x2,0x3,0x5,0x8,0x7,0x1,0x1,0x1,0x4,0x6,0x1,0x1,0x1,0x1, // s0..s0
47 	1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,1,1,1,1,1,0,1,0,1,1,1,1,1,1, // s1..s2
48 	1,2,1,1,1,1,1,2,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,2,1,1,1,1,1,1,1,1, // s3..s4
49 	1,2,1,1,1,1,1,1,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,3,1,3,1,1,1,1,1,1, // s5..s6
50 	1,3,1,1,1,1,1,3,1,3,1,1,1,1,1,1,1,3,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // s7..s8
51 };
52 
53 static inline uint32_t
54 decode(uint32_t* restrict state, uint32_t* restrict codep, uint8_t byte)
55 {
56 	uint32_t type = utf8d[byte];
57 
58 	*codep = (*state != UTF8_ACCEPT) ?
59 		(byte & 0x3fu) | (*codep << 6) :
60 		(0xff >> type) & (byte);
61 
62 	*state = utf8d[256 + *state*16 + type];
63 	return *state;
64 }
65 
66 
67 /* end of the converter, utility functions ahead */
68 
69 #define ZERO_WIDTH_SPACE 0x200B
70 
71 /* public version of decode */
72 uint32_t
73 utf8_decode(uint32_t* restrict state, uint32_t* restrict codep, uint8_t byte)
74 {
75 	return decode(state, codep, byte);
76 }
77 
78 /* encode cp in s.  s must be at least 4 bytes wide */
79 size_t
80 utf8_encode(uint32_t cp, char *s)
81 {
82 	if (cp <= 0x7F) {
83 		*s = (uint8_t)cp;
84 		return 1;
85 	} else if (cp <= 0x7FF) {
86 		s[1] = (uint8_t)(( cp        & 0x3F ) + 0x80);
87 		s[0] = (uint8_t)(((cp >>  6) & 0x1F) + 0xC0);
88 		return 2;
89 	} else if (cp <= 0xFFFF) {
90 		s[2] = (uint8_t)(( cp        & 0x3F) + 0x80);
91 		s[1] = (uint8_t)(((cp >>  6) & 0x3F) + 0x80);
92 		s[0] = (uint8_t)(((cp >> 12) & 0x0F) + 0xE0);
93 		return 3;
94 	} else if (cp <= 0x10FFFF) {
95 		s[3] = (uint8_t)(( cp        & 0x3F) + 0x80);
96 		s[2] = (uint8_t)(((cp >>  6) & 0x3F) + 0x80);
97 		s[1] = (uint8_t)(((cp >> 12) & 0x3F) + 0x80);
98 		s[0] = (uint8_t)(((cp >> 18) & 0x07) + 0xF0);
99 		return 4;
100 	} else {
101 		s[0] = '\0';
102 		return 0;
103 	}
104 }
105 
106 char *
107 utf8_nth(char *s, size_t n)
108 {
109 	size_t i;
110 	uint32_t cp = 0, state = 0;
111 
112 	for (i = 0; *s && i < n; ++s)
113 		if (!decode(&state, &cp, *s))
114 			++i;
115 
116 	if (state != UTF8_ACCEPT)
117 		return NULL;
118 	if (i == n)
119 		return s;
120 	return NULL;
121 }
122 
123 size_t
124 utf8_cplen(char *s)
125 {
126 	uint32_t cp = 0, state = 0;
127 	size_t len;
128 
129 	len = 0;
130 	for (; *s; ++s)
131 		if (!decode(&state, &cp, *s))
132 			len++;
133 	return len;
134 }
135 
136 size_t
137 utf8_ncplen(const char *s, size_t slen)
138 {
139 	uint32_t cp = 0, state = 0;
140 	size_t len = 0;
141 
142 	for (; slen > 0 && *s; ++s, --slen)
143 		if (!decode(&state, &cp, *s))
144 			len++;
145 	return len;
146 }
147 
148 /* returns only 0, 1, 2 or 8.  assumes sizeof(wchar_t) is 4 */
149 size_t
150 utf8_chwidth(uint32_t cp)
151 {
152 	/* XXX: if we're running on a platform where sizeof(wchar_t)
153 	 * == 2 what to do?  The manpage for wcwidth and wcs isn't
154 	 * clear about the encoding, but if it's 16 bit wide I assume
155 	 * it must use UTF-16... right? */
156 	assert(sizeof(wchar_t) == 4);
157 
158 	/*
159 	 * quick and dirty fix for the tabs.  In the future we may
160 	 * want to expand tabs into N spaces, but for the time being
161 	 * this seems to be good enough (tm).
162 	 */
163 	if (cp == '\t')
164 		return 8;
165 
166 	return wcwidth((wchar_t)cp);
167 }
168 
169 /* NOTE: n is the number of codepoints, NOT the byte length.  In
170  * other words, s MUST be NUL-terminated. */
171 size_t
172 utf8_snwidth(const char *s, size_t n)
173 {
174 	size_t i, tot;
175 	uint32_t cp = 0, state = 0;
176 
177 	tot = 0;
178 	for (i = 0; *s && i < n; ++s)
179 		if (!decode(&state, &cp, *s)) {
180 			i++;
181 			tot += utf8_chwidth(cp);
182 		}
183 
184 	return tot;
185 }
186 
187 size_t
188 utf8_swidth(const char *s)
189 {
190 	size_t tot;
191 	uint32_t cp = 0, state = 0;
192 
193 	tot = 0;
194 	for (; *s; ++s)
195 		if (!decode(&state, &cp, *s))
196 			tot += utf8_chwidth(cp);
197 
198 	return tot;
199 }
200 
201 size_t
202 utf8_swidth_between(const char *str, const char *end)
203 {
204 	size_t tot;
205 	uint32_t cp = 0, state = 0;
206 
207 	tot = 0;
208 	for (; *str && str < end; ++str)
209 		if (!decode(&state, &cp, *str))
210 			tot += utf8_chwidth(cp);
211 	return tot;
212 }
213 
214 char *
215 utf8_next_cp(const char *s)
216 {
217 	uint32_t cp = 0, state = 0;
218 
219 	for (; *s; ++s)
220 		if (!decode(&state, &cp, *s))
221 			break;
222 	return (char*)s+1;
223 }
224 
225 char *
226 utf8_prev_cp(const char *start, const char *base)
227 {
228 	uint8_t c;
229 
230 	for (; start > base; start--) {
231 		c = *start;
232 		if ((c & 0xC0) != 0x80)
233 			return (char*)start;
234 	}
235 
236 	return (char*)base;
237 }
238 
239 /*
240  * XXX: This is not correct.  There are codepoints classified as
241  * "emoji", but these can be joined toghether to form more complex
242  * emoji.  There is an official list of what these valid combinations
243  * are, but it would require a costly lookup (a trie can be used to
244  * reduce the times, but...).  The following approach is conceptually
245  * simpler: if there is a sequence of "emoji codepoints" (or ZWS) and
246  * then a space, consider everything before the space a single emoji.
247  * It needs a special check for numbers (yes, 0..9 and # are
248  * technically speaking emojis) but otherwise seems to work well in
249  * practice.
250  */
251 int
252 emojied_line(const char *s, const char **space_ret)
253 {
254 	uint32_t cp = 0, state = 0;
255 	int only_numbers = 1;
256 
257 	for (; *s; ++s) {
258 		if (!decode(&state, &cp, *s)) {
259 			if (cp == ZERO_WIDTH_SPACE)
260 				continue;
261 			if (cp == ' ') {
262 				*space_ret = s;
263 				return !only_numbers;
264 			}
265 			if (!is_emoji(cp))
266 				return 0;
267 			if (cp < '0' || cp > '9')
268 				only_numbers = 0;
269 		}
270 	}
271 
272 	return 0;
273 }