add utf8 library from Lua 5.3; use struct for string packing

author: Philipp Janda <siffiejoe@gmx.net> 2015-01-14 03:52:56 +0100
committer: Philipp Janda <siffiejoe@gmx.net> 2015-01-14 03:52:56 +0100
commit: cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04 (patch)
tree: be89d436fec1054bf1b5cdcb14ffd13579f99d34 /lutf8lib.c
parent: 8ec87cc5b2afb05fd94fe11f35e997f26804cfef (diff)
download: lua-compat-5.3-cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04.tar.gz
lua-compat-5.3-cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04.tar.bz2
lua-compat-5.3-cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04.zip
1 files changed, 255 insertions, 0 deletions
diff --git a/lutf8lib.c b/lutf8lib.c
new file mode 100644
index 0000000..be4f087
--- /dev/null
+++ b/lutf8lib.c
@@ -0,0 +1,255 @@
+/*
+** $Id: lutf8lib.c,v 1.13 2014/11/02 19:19:04 roberto Exp $
+** Standard library for UTF-8 manipulation
+** See Copyright Notice in lua.h
+*/
+#define lutf8lib_c
+#define LUA_LIB
+#include "lprefix.h"
+#include <assert.h>
+#include <stdlib.h>
+#include <string.h>
+#include "lua.h"
+#include "lauxlib.h"
+#include "lualib.h"
+#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) {
+  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;
+}
+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");
+  lua_pushfstring(L, "%U", (long)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;
+}
+/*
+** 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 */
+#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 */
+  {"charpattern", NULL},
+  {NULL, NULL}
+};
+LUAMOD_API int luaopen_utf8 (lua_State *L) {
+  luaL_newlib(L, funcs);
+  lua_pushliteral(L, UTF8PATT);
+  lua_setfield(L, -2, "charpattern");
+  return 1;
+}
author	Philipp Janda <siffiejoe@gmx.net>	2015-01-14 03:52:56 +0100
committer	Philipp Janda <siffiejoe@gmx.net>	2015-01-14 03:52:56 +0100
commit	cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04 (patch)
tree	be89d436fec1054bf1b5cdcb14ffd13579f99d34 /lutf8lib.c
parent	8ec87cc5b2afb05fd94fe11f35e997f26804cfef (diff)
download	lua-compat-5.3-cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04.tar.gz lua-compat-5.3-cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04.tar.bz2 lua-compat-5.3-cdd787b876efe2e662ba7d6e8b9e38bf9f44bc04.zip

diff --git a/lutf8lib.c b/lutf8lib.c new file mode 100644 index 0000000..be4f087 --- /dev/null +++ b/lutf8lib.c
@@ -0,0 +1,255 @@
	1	/*
	2	** $Id: lutf8lib.c,v 1.13 2014/11/02 19:19:04 roberto Exp $
	3	** Standard library for UTF-8 manipulation
	4	** See Copyright Notice in lua.h
	5	*/
	6
	7	#define lutf8lib_c
	8	#define LUA_LIB
	9
	10	#include "lprefix.h"
	11
	12
	13	#include <assert.h>
	14	#include <stdlib.h>
	15	#include <string.h>
	16
	17	#include "lua.h"
	18
	19	#include "lauxlib.h"
	20	#include "lualib.h"
	21
	22	#define MAXUNICODE 0x10FFFF
	23
	24	#define iscont(p) ((*(p) & 0xC0) == 0x80)
	25
	26
	27	/* from strlib */
	28	/* translate a relative string position: negative means back from end */
	29	static lua_Integer u_posrelat (lua_Integer pos, size_t len) {
	30	if (pos >= 0) return pos;
	31	else if (0u - (size_t)pos > len) return 0;
	32	else return (lua_Integer)len + pos + 1;
	33	}
	34
	35
	36	/*
	37	** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid.
	38	*/
	39	static const char utf8_decode (const char o, int *val) {
	40	static unsigned int limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFF};
	41	const unsigned char s = (const unsigned char )o;
	42	unsigned int c = s[0];
	43	unsigned int res = 0; /* final result */
	44	if (c < 0x80) /* ascii? */
	45	res = c;
	46	else {
	47	int count = 0; /* to count number of continuation bytes */
	48	while (c & 0x40) { /* still have continuation bytes? */
	49	int cc = s[++count]; /* read next byte */
	50	if ((cc & 0xC0) != 0x80) /* not a continuation byte? */
	51	return NULL; /* invalid byte sequence */
	52	res = (res << 6) \| (cc & 0x3F); /* add lower 6 bits from cont. byte */
	53	c <<= 1; /* to test next bit */
	54	}
	55	res \|= ((c & 0x7F) << (count * 5)); /* add first byte */
	56	if (count > 3 \|\| res > MAXUNICODE \|\| res <= limits[count])
	57	return NULL; /* invalid byte sequence */
	58	s += count; /* skip continuation bytes read */
	59	}
	60	if (val) *val = res;
	61	return (const char )s + 1; / +1 to include first byte */
	62	}
	63
	64
	65	/*
	66	** utf8len(s [, i [, j]]) --> number of characters that start in the
	67	** range [i,j], or nil + current position if 's' is not well formed in
	68	** that interval
	69	*/
	70	static int utflen (lua_State *L) {
	71	int n = 0;
	72	size_t len;
	73	const char *s = luaL_checklstring(L, 1, &len);
	74	lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
	75	lua_Integer posj = u_posrelat(luaL_optinteger(L, 3, -1), len);
	76	luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 2,
	77	"initial position out of string");
	78	luaL_argcheck(L, --posj < (lua_Integer)len, 3,
	79	"final position out of string");
	80	while (posi <= posj) {
	81	const char *s1 = utf8_decode(s + posi, NULL);
	82	if (s1 == NULL) { /* conversion error? */
	83	lua_pushnil(L); /* return nil ... */
	84	lua_pushinteger(L, posi + 1); /* ... and current position */
	85	return 2;
	86	}
	87	posi = s1 - s;
	88	n++;
	89	}
	90	lua_pushinteger(L, n);
	91	return 1;
	92	}
	93
	94
	95	/*
	96	** codepoint(s, [i, [j]]) -> returns codepoints for all characters
	97	** that start in the range [i,j]
	98	*/
	99	static int codepoint (lua_State *L) {
	100	size_t len;
	101	const char *s = luaL_checklstring(L, 1, &len);
	102	lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len);
	103	lua_Integer pose = u_posrelat(luaL_optinteger(L, 3, posi), len);
	104	int n;
	105	const char *se;
	106	luaL_argcheck(L, posi >= 1, 2, "out of range");
	107	luaL_argcheck(L, pose <= (lua_Integer)len, 3, "out of range");
	108	if (posi > pose) return 0; /* empty interval; return no values */
	109	n = (int)(pose - posi + 1);
	110	if (posi + n <= pose) /* (lua_Integer -> int) overflow? */
	111	return luaL_error(L, "string slice too long");
	112	luaL_checkstack(L, n, "string slice too long");
	113	n = 0;
	114	se = s + pose;
	115	for (s += posi - 1; s < se;) {
	116	int code;
	117	s = utf8_decode(s, &code);
	118	if (s == NULL)
	119	return luaL_error(L, "invalid UTF-8 code");
	120	lua_pushinteger(L, code);
	121	n++;
	122	}
	123	return n;
	124	}
	125
	126
	127	static void pushutfchar (lua_State *L, int arg) {
	128	lua_Integer code = luaL_checkinteger(L, arg);
	129	luaL_argcheck(L, 0 <= code && code <= MAXUNICODE, arg, "value out of range");
	130	lua_pushfstring(L, "%U", (long)code);
	131	}
	132
	133
	134	/*
	135	** utfchar(n1, n2, ...) -> char(n1)..char(n2)...
	136	*/
	137	static int utfchar (lua_State *L) {
	138	int n = lua_gettop(L); /* number of arguments */
	139	if (n == 1) /* optimize common case of single char */
	140	pushutfchar(L, 1);
	141	else {
	142	int i;
	143	luaL_Buffer b;
	144	luaL_buffinit(L, &b);
	145	for (i = 1; i <= n; i++) {
	146	pushutfchar(L, i);
	147	luaL_addvalue(&b);
	148	}
	149	luaL_pushresult(&b);
	150	}
	151	return 1;
	152	}
	153
	154
	155	/*
	156	** offset(s, n, [i]) -> index where n-th character counting from
	157	** position 'i' starts; 0 means character at 'i'.
	158	*/
	159	static int byteoffset (lua_State *L) {
	160	size_t len;
	161	const char *s = luaL_checklstring(L, 1, &len);
	162	lua_Integer n = luaL_checkinteger(L, 2);
	163	lua_Integer posi = (n >= 0) ? 1 : len + 1;
	164	posi = u_posrelat(luaL_optinteger(L, 3, posi), len);
	165	luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 3,
	166	"position out of range");
	167	if (n == 0) {
	168	/* find beginning of current byte sequence */
	169	while (posi > 0 && iscont(s + posi)) posi--;
	170	}
	171	else {
	172	if (iscont(s + posi))
	173	luaL_error(L, "initial position is a continuation byte");
	174	if (n < 0) {
	175	while (n < 0 && posi > 0) { /* move back */
	176	do { /* find beginning of previous character */
	177	posi--;
	178	} while (posi > 0 && iscont(s + posi));
	179	n++;
	180	}
	181	}
	182	else {
	183	n--; /* do not move for 1st character */
	184	while (n > 0 && posi < (lua_Integer)len) {
	185	do { /* find beginning of next character */
	186	posi++;
	187	} while (iscont(s + posi)); /* (cannot pass final '\0') */
	188	n--;
	189	}
	190	}
	191	}
	192	if (n == 0) /* did it find given character? */
	193	lua_pushinteger(L, posi + 1);
	194	else /* no such character */
	195	lua_pushnil(L);
	196	return 1;
	197	}
	198
	199
	200	static int iter_aux (lua_State *L) {
	201	size_t len;
	202	const char *s = luaL_checklstring(L, 1, &len);
	203	lua_Integer n = lua_tointeger(L, 2) - 1;
	204	if (n < 0) /* first iteration? */
	205	n = 0; /* start from here */
	206	else if (n < (lua_Integer)len) {
	207	n++; /* skip current byte */
	208	while (iscont(s + n)) n++; /* and its continuations */
	209	}
	210	if (n >= (lua_Integer)len)
	211	return 0; /* no more codepoints */
	212	else {
	213	int code;
	214	const char *next = utf8_decode(s + n, &code);
	215	if (next == NULL \|\| iscont(next))
	216	return luaL_error(L, "invalid UTF-8 code");
	217	lua_pushinteger(L, n + 1);
	218	lua_pushinteger(L, code);
	219	return 2;
	220	}
	221	}
	222
	223
	224	static int iter_codes (lua_State *L) {
	225	luaL_checkstring(L, 1);
	226	lua_pushcfunction(L, iter_aux);
	227	lua_pushvalue(L, 1);
	228	lua_pushinteger(L, 0);
	229	return 3;
	230	}
	231
	232
	233	/* pattern to match a single UTF-8 character */
	234	#define UTF8PATT "[\0-\x7F\xC2-\xF4][\x80-\xBF]*"
	235
	236
	237	static struct luaL_Reg funcs[] = {
	238	{"offset", byteoffset},
	239	{"codepoint", codepoint},
	240	{"char", utfchar},
	241	{"len", utflen},
	242	{"codes", iter_codes},
	243	/* placeholders */
	244	{"charpattern", NULL},
	245	{NULL, NULL}
	246	};
	247
	248
	249	LUAMOD_API int luaopen_utf8 (lua_State *L) {
	250	luaL_newlib(L, funcs);
	251	lua_pushliteral(L, UTF8PATT);
	252	lua_setfield(L, -2, "charpattern");
	253	return 1;
	254	}
	255