xref: /netbsd-src/external/mit/lua/dist/src/lutf8lib.c (revision f89f6560d453f5e37386cc7938c072d2f528b9fa)
1 /*	$NetBSD: lutf8lib.c,v 1.2 2015/02/02 14:03:05 lneto Exp $	*/
2 
3 /*
4 ** Id: lutf8lib.c,v 1.13 2014/11/02 19:19:04 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 <stdlib.h>
18 #include <string.h>
19 #endif
20 
21 #include "lua.h"
22 
23 #include "lauxlib.h"
24 #include "lualib.h"
25 
26 #define MAXUNICODE	0x10FFFF
27 
28 #define iscont(p)	((*(p) & 0xC0) == 0x80)
29 
30 
31 /* from strlib */
32 /* translate a relative string position: negative means back from end */
33 static lua_Integer u_posrelat (lua_Integer pos, size_t len) {
34   if (pos >= 0) return pos;
35   else if (0u - (size_t)pos > len) return 0;
36   else return (lua_Integer)len + pos + 1;
37 }
38 
39 
40 /*
41 ** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid.
42 */
43 static const char *utf8_decode (const char *o, int *val) {
44   static unsigned int limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFF};
45   const unsigned char *s = (const unsigned char *)o;
46   unsigned int c = s[0];
47   unsigned int res = 0;  /* final result */
48   if (c < 0x80)  /* ascii? */
49     res = c;
50   else {
51     int count = 0;  /* to count number of continuation bytes */
52     while (c & 0x40) {  /* still have continuation bytes? */
53       int cc = s[++count];  /* read next byte */
54       if ((cc & 0xC0) != 0x80)  /* not a continuation byte? */
55         return NULL;  /* invalid byte sequence */
56       res = (res << 6) | (cc & 0x3F);  /* add lower 6 bits from cont. byte */
57       c <<= 1;  /* to test next bit */
58     }
59     res |= ((c & 0x7F) << (count * 5));  /* add first byte */
60     if (count > 3 || res > MAXUNICODE || res <= limits[count])
61       return NULL;  /* invalid byte sequence */
62     s += count;  /* skip continuation bytes read */
63   }
64   if (val) *val = res;
65   return (const char *)s + 1;  /* +1 to include first byte */
66 }
67 
68 
69 /*
70 ** utf8len(s [, i [, j]]) --> number of characters that start in the
71 ** range [i,j], or nil + current position if 's' is not well formed in
72 ** that interval
73 */
74 static int utflen (lua_State *L) {
75   int n = 0;
76   size_t len;
77   const char *s = luaL_checklstring(L, 1, &len);
78   lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
79   lua_Integer posj = u_posrelat(luaL_optinteger(L, 3, -1), len);
80   luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 2,
81                    "initial position out of string");
82   luaL_argcheck(L, --posj < (lua_Integer)len, 3,
83                    "final position out of string");
84   while (posi <= posj) {
85     const char *s1 = utf8_decode(s + posi, NULL);
86     if (s1 == NULL) {  /* conversion error? */
87       lua_pushnil(L);  /* return nil ... */
88       lua_pushinteger(L, posi + 1);  /* ... and current position */
89       return 2;
90     }
91     posi = s1 - s;
92     n++;
93   }
94   lua_pushinteger(L, n);
95   return 1;
96 }
97 
98 
99 /*
100 ** codepoint(s, [i, [j]])  -> returns codepoints for all characters
101 ** that start in the range [i,j]
102 */
103 static int codepoint (lua_State *L) {
104   size_t len;
105   const char *s = luaL_checklstring(L, 1, &len);
106   lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
107   lua_Integer pose = u_posrelat(luaL_optinteger(L, 3, posi), len);
108   int n;
109   const char *se;
110   luaL_argcheck(L, posi >= 1, 2, "out of range");
111   luaL_argcheck(L, pose <= (lua_Integer)len, 3, "out of range");
112   if (posi > pose) return 0;  /* empty interval; return no values */
113   n = (int)(pose -  posi + 1);
114   if (posi + n <= pose)  /* (lua_Integer -> int) overflow? */
115     return luaL_error(L, "string slice too long");
116   luaL_checkstack(L, n, "string slice too long");
117   n = 0;
118   se = s + pose;
119   for (s += posi - 1; s < se;) {
120     int code;
121     s = utf8_decode(s, &code);
122     if (s == NULL)
123       return luaL_error(L, "invalid UTF-8 code");
124     lua_pushinteger(L, code);
125     n++;
126   }
127   return n;
128 }
129 
130 
131 static void pushutfchar (lua_State *L, int arg) {
132   lua_Integer code = luaL_checkinteger(L, arg);
133   luaL_argcheck(L, 0 <= code && code <= MAXUNICODE, arg, "value out of range");
134   lua_pushfstring(L, "%U", (long)code);
135 }
136 
137 
138 /*
139 ** utfchar(n1, n2, ...)  -> char(n1)..char(n2)...
140 */
141 static int utfchar (lua_State *L) {
142   int n = lua_gettop(L);  /* number of arguments */
143   if (n == 1)  /* optimize common case of single char */
144     pushutfchar(L, 1);
145   else {
146     int i;
147     luaL_Buffer b;
148     luaL_buffinit(L, &b);
149     for (i = 1; i <= n; i++) {
150       pushutfchar(L, i);
151       luaL_addvalue(&b);
152     }
153     luaL_pushresult(&b);
154   }
155   return 1;
156 }
157 
158 
159 /*
160 ** offset(s, n, [i])  -> index where n-th character counting from
161 **   position 'i' starts; 0 means character at 'i'.
162 */
163 static int byteoffset (lua_State *L) {
164   size_t len;
165   const char *s = luaL_checklstring(L, 1, &len);
166   lua_Integer n  = luaL_checkinteger(L, 2);
167   lua_Integer posi = (n >= 0) ? 1 : len + 1;
168   posi = u_posrelat(luaL_optinteger(L, 3, posi), len);
169   luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 3,
170                    "position out of range");
171   if (n == 0) {
172     /* find beginning of current byte sequence */
173     while (posi > 0 && iscont(s + posi)) posi--;
174   }
175   else {
176     if (iscont(s + posi))
177       luaL_error(L, "initial position is a continuation byte");
178     if (n < 0) {
179        while (n < 0 && posi > 0) {  /* move back */
180          do {  /* find beginning of previous character */
181            posi--;
182          } while (posi > 0 && iscont(s + posi));
183          n++;
184        }
185      }
186      else {
187        n--;  /* do not move for 1st character */
188        while (n > 0 && posi < (lua_Integer)len) {
189          do {  /* find beginning of next character */
190            posi++;
191          } while (iscont(s + posi));  /* (cannot pass final '\0') */
192          n--;
193        }
194      }
195   }
196   if (n == 0)  /* did it find given character? */
197     lua_pushinteger(L, posi + 1);
198   else  /* no such character */
199     lua_pushnil(L);
200   return 1;
201 }
202 
203 
204 static int iter_aux (lua_State *L) {
205   size_t len;
206   const char *s = luaL_checklstring(L, 1, &len);
207   lua_Integer n = lua_tointeger(L, 2) - 1;
208   if (n < 0)  /* first iteration? */
209     n = 0;  /* start from here */
210   else if (n < (lua_Integer)len) {
211     n++;  /* skip current byte */
212     while (iscont(s + n)) n++;  /* and its continuations */
213   }
214   if (n >= (lua_Integer)len)
215     return 0;  /* no more codepoints */
216   else {
217     int code;
218     const char *next = utf8_decode(s + n, &code);
219     if (next == NULL || iscont(next))
220       return luaL_error(L, "invalid UTF-8 code");
221     lua_pushinteger(L, n + 1);
222     lua_pushinteger(L, code);
223     return 2;
224   }
225 }
226 
227 
228 static int iter_codes (lua_State *L) {
229   luaL_checkstring(L, 1);
230   lua_pushcfunction(L, iter_aux);
231   lua_pushvalue(L, 1);
232   lua_pushinteger(L, 0);
233   return 3;
234 }
235 
236 
237 /* pattern to match a single UTF-8 character */
238 #define UTF8PATT	"[\0-\x7F\xC2-\xF4][\x80-\xBF]*"
239 
240 
241 static struct luaL_Reg funcs[] = {
242   {"offset", byteoffset},
243   {"codepoint", codepoint},
244   {"char", utfchar},
245   {"len", utflen},
246   {"codes", iter_codes},
247   /* placeholders */
248   {"charpattern", NULL},
249   {NULL, NULL}
250 };
251 
252 
253 LUAMOD_API int luaopen_utf8 (lua_State *L) {
254   luaL_newlib(L, funcs);
255   lua_pushliteral(L, UTF8PATT);
256   lua_setfield(L, -2, "charpattern");
257   return 1;
258 }
259 
260