love/src/libraries/lua53/lutf8lib.c

/*
** $Id: lutf8lib.c,v 1.13 2014/11/02 19:19:04 roberto Exp $
** Standard library for UTF-8 manipulation
** Modified by the LOVE Development Team to work with Lua 5.1's API
*/

/******************************************************************************
 * Copyright (C) 1994-2015 Lua.org, PUC-Rio, 2015 LOVE Development Team.
 *
 * Permission is hereby granted, free of charge, to any person obtaining
 * a copy of this software and associated documentation files (the
 * "Software"), to deal in the Software without restriction, including
 * without limitation the rights to use, copy, modify, merge, publish,
 * distribute, sublicense, and/or sell copies of the Software, and to
 * permit persons to whom the Software is furnished to do so, subject to
 * the following conditions:
 *
 * The above copyright notice and this permission notice shall be
 * included in all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
 * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
 * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 ******************************************************************************/

#define lutf8lib_c

#include "lprefix.h"


#include <assert.h>
#include <stdlib.h>
#include <string.h>

#include "lutf8lib.h"

#include "lauxlib.h"
#include "lualib.h"

#define MAXUNICODE	0x10FFFF

/* size of buffer for 'utf8esc' function (taken from lobject.h) */
#define UTF8BUFFSZ	8

#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) {
  if (pos >= 0) return pos;
  else if (0u - (size_t)pos > len) return 0;
  else return (lua_Integer)len + pos + 1;
}


/*
** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid.
*/
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 */
}


/*
** 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
*/
static int utflen (lua_State *L) {
  int n = 0;
  size_t len;
  const char *s = luaL_checklstring(L, 1, &len);
  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,
                   "initial position out of string");
  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;
    n++;
  }
  lua_pushinteger(L, n);
  return 1;
}


/*
** codepoint(s, [i, [j]])  -> returns codepoints for all characters
** that start in the range [i,j]
*/
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);
  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)
      return luaL_error(L, "invalid UTF-8 code");
    lua_pushinteger(L, code);
    n++;
  }
  return n;
}


/* taken from lobject.c */
static int utf8esc (char *buff, unsigned long x) {
	int n = 1;  /* number of bytes put in buffer (backwards) */
	lua_assert(x <= 0x10FFFF);
	if (x < 0x80)  /* ascii? */
		buff[UTF8BUFFSZ - 1] = (char) x;
	else {  /* need continuation bytes */
		unsigned int mfb = 0x3f;  /* maximum that fits in first byte */
		do {  /* add continuation bytes */
			buff[UTF8BUFFSZ - (n++)] = (char) (0x80 | (x & 0x3f));
			x >>= 6;  /* remove added bits */
			mfb >>= 1;  /* now there is one less bit available in first byte */
		} while (x > mfb);  /* still needs continuation byte? */
		buff[UTF8BUFFSZ - n] = (char) ((~mfb << 1) | x);  /* add first byte */
	}
	return n;
}

static void pushutfchar (lua_State *L, int arg) {
  lua_Integer code = luaL_checkinteger(L, arg);
  luaL_argcheck(L, 0 <= code && code <= MAXUNICODE, arg, "value out of range");

  /* the %U string format does not exist in lua 5.1 or 5.2, so we emulate it */
  /* (code from luaO_pushvfstring in lobject.c) */
  char buff[UTF8BUFFSZ];
  int l = utf8esc(buff, (long) code);
  lua_pushlstring(L, buff + UTF8BUFFSZ - l, l);
}


/*
** 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;
}


/*
** offset(s, n, [i])  -> index where n-th character counting from
**   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);
  lua_Integer n  = luaL_checkinteger(L, 2);
  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,
                   "position out of range");
  if (n == 0) {
    /* find beginning of current byte sequence */
    while (posi > 0 && iscont(s + posi)) posi--;
  }
  else {
    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--;
       }
     }
  }
  if (n == 0)  /* did it find given character? */
    lua_pushinteger(L, posi + 1);
  else  /* no such character */
    lua_pushnil(L);
  return 1;  
}


static int iter_aux (lua_State *L) {
  size_t len;
  const char *s = luaL_checklstring(L, 1, &len);
  lua_Integer n = lua_tointeger(L, 2) - 1;
  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))
      return luaL_error(L, "invalid UTF-8 code");
    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 */
#if LUA_VERSION_NUM >= 502
#define UTF8PATT	"[\0-\x7F\xC2-\xF4][\x80-\xBF]*"
#else
/* lua 5.1 doesn't support literal null bytes in patterns */
#define UTF8PATT	"[%z\x01-\x7F\xC2-\xF4][\x80-\xBF]*"
#endif


static struct luaL_Reg funcs[] = {
  {"offset", byteoffset},
  {"codepoint", codepoint},
  {"char", utfchar},
  {"len", utflen},
  {"codes", iter_codes},
  /* placeholders */
  {"charpattern", NULL},
  {NULL, NULL}
};


/* modified version of luaopen_utf8, designed to work with lua 5.1-5.3 */
int luaopen_luautf8 (lua_State *L) {
  luaL_Reg *l;
  lua_createtable(L, 0, (int) (sizeof(funcs) / sizeof(luaL_Reg)) - 1);
  for (l = funcs; l->name != NULL; l++) {
    if (l->func != NULL) {
      lua_pushcfunction(L, l->func);
      lua_setfield(L, -2, l->name);
    }
  }
  lua_pushliteral(L, UTF8PATT);
  lua_setfield(L, -2, "charpattern");
  return 1;
}