lua/lutf8lib.c

254 lines
6.8 KiB
C
Raw Normal View History

/*
** $Id: lutf8lib.c,v 1.9 2014/05/14 18:33:37 roberto Exp roberto $
** Standard library for UTF-8 manipulation
** See Copyright Notice in lua.h
*/
2014-02-06 21:32:33 +04:00
#include <assert.h>
#include <stdlib.h>
#include <string.h>
#define lutf8lib_c
#define LUA_LIB
2014-02-06 21:32:33 +04:00
#include "lua.h"
2014-02-06 21:32:33 +04:00
#include "lauxlib.h"
#include "lualib.h"
2014-02-06 21:32:33 +04:00
#define MAXUNICODE 0x10FFFF
#define iscont(p) ((*(p) & 0xC0) == 0x80)
/* from strlib */
/* translate a relative string position: negative means back from end */
static lua_Integer u_posrelat (lua_Integer pos, size_t len) {
2014-02-06 21:32:33 +04:00
if (pos >= 0) return pos;
else if (0u - (size_t)pos > len) return 0;
else return (lua_Integer)len + pos + 1;
}
/*
2014-04-11 22:19:07 +04:00
** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid.
2014-02-06 21:32:33 +04:00
*/
static const char *utf8_decode (const char *o, int *val) {
static unsigned int limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFF};
const unsigned char *s = (const unsigned char *)o;
unsigned int c = s[0];
unsigned int res = 0; /* final result */
if (c < 0x80) /* ascii? */
res = c;
else {
int count = 0; /* to count number of continuation bytes */
while (c & 0x40) { /* still have continuation bytes? */
int cc = s[++count]; /* read next byte */
if ((cc & 0xC0) != 0x80) /* not a continuation byte? */
return NULL; /* invalid byte sequence */
res = (res << 6) | (cc & 0x3F); /* add lower 6 bits from cont. byte */
c <<= 1; /* to test next bit */
}
res |= ((c & 0x7F) << (count * 5)); /* add first byte */
if (count > 3 || res > MAXUNICODE || res <= limits[count])
return NULL; /* invalid byte sequence */
s += count; /* skip continuation bytes read */
}
if (val) *val = res;
return (const char *)s + 1; /* +1 to include first byte */
}
/*
2014-04-11 22:19:07 +04:00
** utf8len(s [, i [, j]]) --> number of characters that start in the
** range [i,j], or nil + current position if 's' is not well formed in
** that interval
2014-02-06 21:32:33 +04:00
*/
static int utflen (lua_State *L) {
int n = 0;
size_t len;
const char *s = luaL_checklstring(L, 1, &len);
2014-04-02 21:01:22 +04:00
lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
lua_Integer posj = u_posrelat(luaL_optinteger(L, 3, -1), len);
luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 2,
2014-02-06 21:32:33 +04:00
"initial position out of string");
2014-04-02 21:01:22 +04:00
luaL_argcheck(L, --posj < (lua_Integer)len, 3,
"final position out of string");
while (posi <= posj) {
const char *s1 = utf8_decode(s + posi, NULL);
if (s1 == NULL) { /* conversion error? */
lua_pushnil(L); /* return nil ... */
lua_pushinteger(L, posi + 1); /* ... and current position */
return 2;
}
posi = s1 - s;
2014-02-06 21:32:33 +04:00
n++;
2014-04-02 21:01:22 +04:00
}
lua_pushinteger(L, n);
2014-02-06 21:32:33 +04:00
return 1;
}
/*
** codepoint(s, [i, [j]]) -> returns codepoints for all characters
2014-04-11 22:19:07 +04:00
** that start in the range [i,j]
2014-02-06 21:32:33 +04:00
*/
static int codepoint (lua_State *L) {
size_t len;
const char *s = luaL_checklstring(L, 1, &len);
lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
lua_Integer pose = u_posrelat(luaL_optinteger(L, 3, posi), len);
2014-02-06 21:32:33 +04:00
int n;
const char *se;
luaL_argcheck(L, posi >= 1, 2, "out of range");
luaL_argcheck(L, pose <= (lua_Integer)len, 3, "out of range");
if (posi > pose) return 0; /* empty interval; return no values */
n = (int)(pose - posi + 1);
if (posi + n <= pose) /* (lua_Integer -> int) overflow? */
return luaL_error(L, "string slice too long");
luaL_checkstack(L, n, "string slice too long");
n = 0;
se = s + pose;
for (s += posi - 1; s < se;) {
int code;
s = utf8_decode(s, &code);
if (s == NULL)
2014-03-20 23:36:02 +04:00
return luaL_error(L, "invalid UTF-8 code");
2014-02-06 21:32:33 +04:00
lua_pushinteger(L, code);
n++;
}
return n;
}
static void pushutfchar (lua_State *L, int arg) {
int code = luaL_checkint(L, arg);
luaL_argcheck(L, 0 <= code && code <= MAXUNICODE, arg, "value out of range");
lua_pushfstring(L, "%U", code);
}
/*
** utfchar(n1, n2, ...) -> char(n1)..char(n2)...
*/
static int utfchar (lua_State *L) {
int n = lua_gettop(L); /* number of arguments */
if (n == 1) /* optimize common case of single char */
pushutfchar(L, 1);
else {
int i;
luaL_Buffer b;
luaL_buffinit(L, &b);
for (i = 1; i <= n; i++) {
pushutfchar(L, i);
luaL_addvalue(&b);
}
luaL_pushresult(&b);
}
return 1;
}
/*
2014-04-03 17:45:09 +04:00
** offset(s, n, [i]) -> index where n-th character counting from
2014-02-06 21:32:33 +04:00
** position 'i' starts; 0 means character at 'i'.
*/
static int byteoffset (lua_State *L) {
size_t len;
const char *s = luaL_checklstring(L, 1, &len);
int n = luaL_checkint(L, 2);
2014-04-03 17:45:09 +04:00
lua_Integer posi = (n >= 0) ? 1 : len + 1;
posi = u_posrelat(luaL_optinteger(L, 3, posi), len);
luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 3,
2014-02-06 21:32:33 +04:00
"position out of range");
if (n == 0) {
/* find beginning of current byte sequence */
while (posi > 0 && iscont(s + posi)) posi--;
}
else {
2014-04-03 17:45:09 +04:00
if (iscont(s + posi))
luaL_error(L, "initial position is a continuation byte");
if (n < 0) {
while (n < 0 && posi > 0) { /* move back */
do { /* find beginning of previous character */
posi--;
} while (posi > 0 && iscont(s + posi));
n++;
}
}
else {
n--; /* do not move for 1st character */
while (n > 0 && posi < (lua_Integer)len) {
do { /* find beginning of next character */
posi++;
} while (iscont(s + posi)); /* (cannot pass final '\0') */
n--;
}
}
2014-02-06 21:32:33 +04:00
}
2014-04-03 17:45:09 +04:00
if (n == 0) /* did it find given character? */
2014-02-06 21:32:33 +04:00
lua_pushinteger(L, posi + 1);
2014-04-03 17:45:09 +04:00
else /* no such character */
lua_pushnil(L);
2014-02-06 21:32:33 +04:00
return 1;
}
static int iter_aux (lua_State *L) {
size_t len;
const char *s = luaL_checklstring(L, 1, &len);
2014-04-01 18:39:55 +04:00
lua_Integer n = lua_tointeger(L, 2) - 1;
2014-02-06 21:32:33 +04:00
if (n < 0) /* first iteration? */
n = 0; /* start from here */
else if (n < (lua_Integer)len) {
n++; /* skip current byte */
while (iscont(s + n)) n++; /* and its continuations */
}
if (n >= (lua_Integer)len)
return 0; /* no more codepoints */
else {
int code;
const char *next = utf8_decode(s + n, &code);
if (next == NULL || iscont(next))
2014-03-20 23:36:02 +04:00
return luaL_error(L, "invalid UTF-8 code");
2014-02-06 21:32:33 +04:00
lua_pushinteger(L, n + 1);
lua_pushinteger(L, code);
return 2;
}
}
static int iter_codes (lua_State *L) {
luaL_checkstring(L, 1);
lua_pushcfunction(L, iter_aux);
lua_pushvalue(L, 1);
lua_pushinteger(L, 0);
return 3;
}
/* pattern to match a single UTF-8 character */
#define UTF8PATT "[\0-\x7F\xC2-\xF4][\x80-\xBF]*"
static struct luaL_Reg funcs[] = {
{"offset", byteoffset},
{"codepoint", codepoint},
{"char", utfchar},
{"len", utflen},
{"codes", iter_codes},
/* placeholders */
{"charpatt", NULL},
2014-02-06 21:32:33 +04:00
{NULL, NULL}
};
LUAMOD_API int luaopen_utf8 (lua_State *L) {
2014-02-06 21:32:33 +04:00
luaL_newlib(L, funcs);
lua_pushliteral(L, UTF8PATT);
lua_setfield(L, -2, "charpatt");
return 1;
}