1 /* $NetBSD: lutf8lib.c,v 1.3 2015/10/08 13:21:00 mbalmer Exp $ */
2
3 /*
4 ** Id: lutf8lib.c,v 1.15 2015/03/28 19:16:55 roberto Exp
5 ** Standard library for UTF-8 manipulation
6 ** See Copyright Notice in lua.h
7 */
8
9 #define lutf8lib_c
10 #define LUA_LIB
11
12 #include "lprefix.h"
13
14
15 #ifndef _KERNEL
16 #include <assert.h>
17 #include <limits.h>
18 #include <stdlib.h>
19 #include <string.h>
20 #endif
21
22 #include "lua.h"
23
24 #include "lauxlib.h"
25 #include "lualib.h"
26
27 #define MAXUNICODE 0x10FFFF
28
29 #define iscont(p) ((*(p) & 0xC0) == 0x80)
30
31
32 /* from strlib */
33 /* translate a relative string position: negative means back from end */
u_posrelat(lua_Integer pos,size_t len)34 static lua_Integer u_posrelat (lua_Integer pos, size_t len) {
35 if (pos >= 0) return pos;
36 else if (0u - (size_t)pos > len) return 0;
37 else return (lua_Integer)len + pos + 1;
38 }
39
40
41 /*
42 ** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid.
43 */
utf8_decode(const char * o,int * val)44 static const char *utf8_decode (const char *o, int *val) {
45 static const unsigned int limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFF};
46 const unsigned char *s = (const unsigned char *)o;
47 unsigned int c = s[0];
48 unsigned int res = 0; /* final result */
49 if (c < 0x80) /* ascii? */
50 res = c;
51 else {
52 int count = 0; /* to count number of continuation bytes */
53 while (c & 0x40) { /* still have continuation bytes? */
54 int cc = s[++count]; /* read next byte */
55 if ((cc & 0xC0) != 0x80) /* not a continuation byte? */
56 return NULL; /* invalid byte sequence */
57 res = (res << 6) | (cc & 0x3F); /* add lower 6 bits from cont. byte */
58 c <<= 1; /* to test next bit */
59 }
60 res |= ((c & 0x7F) << (count * 5)); /* add first byte */
61 if (count > 3 || res > MAXUNICODE || res <= limits[count])
62 return NULL; /* invalid byte sequence */
63 s += count; /* skip continuation bytes read */
64 }
65 if (val) *val = res;
66 return (const char *)s + 1; /* +1 to include first byte */
67 }
68
69
70 /*
71 ** utf8len(s [, i [, j]]) --> number of characters that start in the
72 ** range [i,j], or nil + current position if 's' is not well formed in
73 ** that interval
74 */
utflen(lua_State * L)75 static int utflen (lua_State *L) {
76 int n = 0;
77 size_t len;
78 const char *s = luaL_checklstring(L, 1, &len);
79 lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
80 lua_Integer posj = u_posrelat(luaL_optinteger(L, 3, -1), len);
81 luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 2,
82 "initial position out of string");
83 luaL_argcheck(L, --posj < (lua_Integer)len, 3,
84 "final position out of string");
85 while (posi <= posj) {
86 const char *s1 = utf8_decode(s + posi, NULL);
87 if (s1 == NULL) { /* conversion error? */
88 lua_pushnil(L); /* return nil ... */
89 lua_pushinteger(L, posi + 1); /* ... and current position */
90 return 2;
91 }
92 posi = s1 - s;
93 n++;
94 }
95 lua_pushinteger(L, n);
96 return 1;
97 }
98
99
100 /*
101 ** codepoint(s, [i, [j]]) -> returns codepoints for all characters
102 ** that start in the range [i,j]
103 */
codepoint(lua_State * L)104 static int codepoint (lua_State *L) {
105 size_t len;
106 const char *s = luaL_checklstring(L, 1, &len);
107 lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
108 lua_Integer pose = u_posrelat(luaL_optinteger(L, 3, posi), len);
109 int n;
110 const char *se;
111 luaL_argcheck(L, posi >= 1, 2, "out of range");
112 luaL_argcheck(L, pose <= (lua_Integer)len, 3, "out of range");
113 if (posi > pose) return 0; /* empty interval; return no values */
114 if (pose - posi >= INT_MAX) /* (lua_Integer -> int) overflow? */
115 return luaL_error(L, "string slice too long");
116 n = (int)(pose - posi) + 1;
117 luaL_checkstack(L, n, "string slice too long");
118 n = 0;
119 se = s + pose;
120 for (s += posi - 1; s < se;) {
121 int code;
122 s = utf8_decode(s, &code);
123 if (s == NULL)
124 return luaL_error(L, "invalid UTF-8 code");
125 lua_pushinteger(L, code);
126 n++;
127 }
128 return n;
129 }
130
131
pushutfchar(lua_State * L,int arg)132 static void pushutfchar (lua_State *L, int arg) {
133 lua_Integer code = luaL_checkinteger(L, arg);
134 luaL_argcheck(L, 0 <= code && code <= MAXUNICODE, arg, "value out of range");
135 lua_pushfstring(L, "%U", (long)code);
136 }
137
138
139 /*
140 ** utfchar(n1, n2, ...) -> char(n1)..char(n2)...
141 */
utfchar(lua_State * L)142 static int utfchar (lua_State *L) {
143 int n = lua_gettop(L); /* number of arguments */
144 if (n == 1) /* optimize common case of single char */
145 pushutfchar(L, 1);
146 else {
147 int i;
148 luaL_Buffer b;
149 luaL_buffinit(L, &b);
150 for (i = 1; i <= n; i++) {
151 pushutfchar(L, i);
152 luaL_addvalue(&b);
153 }
154 luaL_pushresult(&b);
155 }
156 return 1;
157 }
158
159
160 /*
161 ** offset(s, n, [i]) -> index where n-th character counting from
162 ** position 'i' starts; 0 means character at 'i'.
163 */
byteoffset(lua_State * L)164 static int byteoffset (lua_State *L) {
165 size_t len;
166 const char *s = luaL_checklstring(L, 1, &len);
167 lua_Integer n = luaL_checkinteger(L, 2);
168 lua_Integer posi = (n >= 0) ? 1 : len + 1;
169 posi = u_posrelat(luaL_optinteger(L, 3, posi), len);
170 luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 3,
171 "position out of range");
172 if (n == 0) {
173 /* find beginning of current byte sequence */
174 while (posi > 0 && iscont(s + posi)) posi--;
175 }
176 else {
177 if (iscont(s + posi))
178 luaL_error(L, "initial position is a continuation byte");
179 if (n < 0) {
180 while (n < 0 && posi > 0) { /* move back */
181 do { /* find beginning of previous character */
182 posi--;
183 } while (posi > 0 && iscont(s + posi));
184 n++;
185 }
186 }
187 else {
188 n--; /* do not move for 1st character */
189 while (n > 0 && posi < (lua_Integer)len) {
190 do { /* find beginning of next character */
191 posi++;
192 } while (iscont(s + posi)); /* (cannot pass final '\0') */
193 n--;
194 }
195 }
196 }
197 if (n == 0) /* did it find given character? */
198 lua_pushinteger(L, posi + 1);
199 else /* no such character */
200 lua_pushnil(L);
201 return 1;
202 }
203
204
iter_aux(lua_State * L)205 static int iter_aux (lua_State *L) {
206 size_t len;
207 const char *s = luaL_checklstring(L, 1, &len);
208 lua_Integer n = lua_tointeger(L, 2) - 1;
209 if (n < 0) /* first iteration? */
210 n = 0; /* start from here */
211 else if (n < (lua_Integer)len) {
212 n++; /* skip current byte */
213 while (iscont(s + n)) n++; /* and its continuations */
214 }
215 if (n >= (lua_Integer)len)
216 return 0; /* no more codepoints */
217 else {
218 int code;
219 const char *next = utf8_decode(s + n, &code);
220 if (next == NULL || iscont(next))
221 return luaL_error(L, "invalid UTF-8 code");
222 lua_pushinteger(L, n + 1);
223 lua_pushinteger(L, code);
224 return 2;
225 }
226 }
227
228
iter_codes(lua_State * L)229 static int iter_codes (lua_State *L) {
230 luaL_checkstring(L, 1);
231 lua_pushcfunction(L, iter_aux);
232 lua_pushvalue(L, 1);
233 lua_pushinteger(L, 0);
234 return 3;
235 }
236
237
238 /* pattern to match a single UTF-8 character */
239 #define UTF8PATT "[\0-\x7F\xC2-\xF4][\x80-\xBF]*"
240
241
242 static const luaL_Reg funcs[] = {
243 {"offset", byteoffset},
244 {"codepoint", codepoint},
245 {"char", utfchar},
246 {"len", utflen},
247 {"codes", iter_codes},
248 /* placeholders */
249 {"charpattern", NULL},
250 {NULL, NULL}
251 };
252
253
luaopen_utf8(lua_State * L)254 LUAMOD_API int luaopen_utf8 (lua_State *L) {
255 luaL_newlib(L, funcs);
256 lua_pushlstring(L, UTF8PATT, sizeof(UTF8PATT)/sizeof(char) - 1);
257 lua_setfield(L, -2, "charpattern");
258 return 1;
259 }
260
261