mirror of
https://gitcode.com/JianFeeeee/LuaCangjia_api.git
synced 2026-09-20 17:08:12 +00:00
feat: 适配 Lua 5.5.1 最新版 (lua_5.5.1 分支)
- lib/lua 替换为 Lua 5.5.1 官方源码(2026-07 发布, 当前最新) - luaconf.h LUA_EXTRASPACE 扩为 272 字节(与 5.4.8 同方案) - C API 完全兼容, 桥接层零改动 - PackagePathConfig 测试沿用跨版本语义断言(5.4+ require 返回 loader data 路径) - 验证: GTest 44/44 通过
This commit is contained in:
JianFeeeee
99
lib/lua/lutf8lib.c
Executable file → Normal file
99
lib/lua/lutf8lib.c
Executable file → Normal file
@ -10,7 +10,6 @@
|
||||
#include "lprefix.h"
|
||||
|
||||
|
||||
#include <assert.h>
|
||||
#include <limits.h>
|
||||
#include <stdlib.h>
|
||||
#include <string.h>
|
||||
@ -19,6 +18,7 @@
|
||||
|
||||
#include "lauxlib.h"
|
||||
#include "lualib.h"
|
||||
#include "llimits.h"
|
||||
|
||||
|
||||
#define MAXUNICODE 0x10FFFFu
|
||||
@ -28,15 +28,6 @@
|
||||
|
||||
#define MSGInvalid "invalid UTF-8 code"
|
||||
|
||||
/*
|
||||
** Integer type for decoded UTF-8 values; MAXUTF needs 31 bits.
|
||||
*/
|
||||
#if (UINT_MAX >> 30) >= 1
|
||||
typedef unsigned int utfint;
|
||||
#else
|
||||
typedef unsigned long utfint;
|
||||
#endif
|
||||
|
||||
|
||||
#define iscont(c) (((c) & 0xC0) == 0x80)
|
||||
#define iscontp(p) iscont(*(p))
|
||||
@ -55,16 +46,18 @@ static lua_Integer u_posrelat (lua_Integer pos, size_t len) {
|
||||
** Decode one UTF-8 sequence, returning NULL if byte sequence is
|
||||
** invalid. The array 'limits' stores the minimum value for each
|
||||
** sequence length, to check for overlong representations. Its first
|
||||
** entry forces an error for non-ascii bytes with no continuation
|
||||
** entry forces an error for non-ASCII bytes with no continuation
|
||||
** bytes (count == 0).
|
||||
*/
|
||||
static const char *utf8_decode (const char *s, utfint *val, int strict) {
|
||||
static const utfint limits[] =
|
||||
{~(utfint)0, 0x80, 0x800, 0x10000u, 0x200000u, 0x4000000u};
|
||||
static const char *utf8_decode (const char *s, l_uint32 *val, int strict) {
|
||||
static const l_uint32 limits[] =
|
||||
{~(l_uint32)0, 0x80, 0x800, 0x10000u, 0x200000u, 0x4000000u};
|
||||
unsigned int c = (unsigned char)s[0];
|
||||
utfint res = 0; /* final result */
|
||||
if (c < 0x80) /* ascii? */
|
||||
l_uint32 res = 0; /* final result */
|
||||
if (c < 0x80) /* ASCII? */
|
||||
res = c;
|
||||
else if (c >= 0xfe) /* c >= 1111 1110b ? */
|
||||
return NULL; /* would need six or more continuation bytes */
|
||||
else {
|
||||
int count = 0; /* to count number of continuation bytes */
|
||||
for (; c & 0x40; c <<= 1) { /* while it needs continuation bytes... */
|
||||
@ -73,8 +66,9 @@ static const char *utf8_decode (const char *s, utfint *val, int strict) {
|
||||
return NULL; /* invalid byte sequence */
|
||||
res = (res << 6) | (cc & 0x3F); /* add lower 6 bits from cont. byte */
|
||||
}
|
||||
res |= ((utfint)(c & 0x7F) << (count * 5)); /* add first byte */
|
||||
if (count > 5 || res > MAXUTF || res < limits[count])
|
||||
lua_assert(count <= 5);
|
||||
res |= ((l_uint32)(c & 0x7F) << (count * 5)); /* add first byte */
|
||||
if (res > MAXUTF || res < limits[count])
|
||||
return NULL; /* invalid byte sequence */
|
||||
s += count; /* skip continuation bytes read */
|
||||
}
|
||||
@ -111,7 +105,7 @@ static int utflen (lua_State *L) {
|
||||
lua_pushinteger(L, posi + 1); /* ... and current position */
|
||||
return 2;
|
||||
}
|
||||
posi = s1 - s;
|
||||
posi = ct_diff2S(s1 - s);
|
||||
n++;
|
||||
}
|
||||
lua_pushinteger(L, n);
|
||||
@ -141,11 +135,11 @@ static int codepoint (lua_State *L) {
|
||||
n = 0; /* count the number of returns */
|
||||
se = s + pose; /* string end */
|
||||
for (s += posi - 1; s < se;) {
|
||||
utfint code;
|
||||
l_uint32 code;
|
||||
s = utf8_decode(s, &code, !lax);
|
||||
if (s == NULL)
|
||||
return luaL_error(L, MSGInvalid);
|
||||
lua_pushinteger(L, code);
|
||||
lua_pushinteger(L, l_castU2S(code));
|
||||
n++;
|
||||
}
|
||||
return n;
|
||||
@ -155,7 +149,7 @@ static int codepoint (lua_State *L) {
|
||||
static void pushutfchar (lua_State *L, int arg) {
|
||||
lua_Unsigned code = (lua_Unsigned)luaL_checkinteger(L, arg);
|
||||
luaL_argcheck(L, code <= MAXUTF, arg, "value out of range");
|
||||
lua_pushfstring(L, "%U", (long)code);
|
||||
lua_pushfstring(L, "%U", cast(unsigned long, code));
|
||||
}
|
||||
|
||||
|
||||
@ -181,14 +175,14 @@ static int utfchar (lua_State *L) {
|
||||
|
||||
|
||||
/*
|
||||
** offset(s, n, [i]) -> index where n-th character counting from
|
||||
** position 'i' starts; 0 means character at 'i'.
|
||||
** offset(s, n, [i]) -> indices where n-th character counting from
|
||||
** position 'i' starts and ends; 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;
|
||||
lua_Integer posi = (n >= 0) ? 1 : cast_st2S(len) + 1;
|
||||
posi = u_posrelat(luaL_optinteger(L, 3, posi), len);
|
||||
luaL_argcheck(L, 1 <= posi && --posi <= (lua_Integer)len, 3,
|
||||
"position out of bounds");
|
||||
@ -200,28 +194,37 @@ static int byteoffset (lua_State *L) {
|
||||
if (iscontp(s + posi))
|
||||
return 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 && iscontp(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 (iscontp(s + posi)); /* (cannot pass final '\0') */
|
||||
n--;
|
||||
}
|
||||
}
|
||||
while (n < 0 && posi > 0) { /* move back */
|
||||
do { /* find beginning of previous character */
|
||||
posi--;
|
||||
} while (posi > 0 && iscontp(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 (iscontp(s + posi)); /* (cannot pass final '\0') */
|
||||
n--;
|
||||
}
|
||||
}
|
||||
}
|
||||
if (n == 0) /* did it find given character? */
|
||||
lua_pushinteger(L, posi + 1);
|
||||
else /* no such character */
|
||||
if (n != 0) { /* did not find given character? */
|
||||
luaL_pushfail(L);
|
||||
return 1;
|
||||
return 1;
|
||||
}
|
||||
lua_pushinteger(L, posi + 1); /* initial position */
|
||||
if ((s[posi] & 0x80) != 0) { /* multi-byte character? */
|
||||
if (iscont(s[posi]))
|
||||
return luaL_error(L, "initial position is a continuation byte");
|
||||
while (iscontp(s + posi + 1))
|
||||
posi++; /* skip to last continuation byte */
|
||||
}
|
||||
/* else one-byte character: final position is the initial one */
|
||||
lua_pushinteger(L, posi + 1); /* 'posi' now is the final position */
|
||||
return 2;
|
||||
}
|
||||
|
||||
|
||||
@ -235,12 +238,12 @@ static int iter_aux (lua_State *L, int strict) {
|
||||
if (n >= len) /* (also handles original 'n' being negative) */
|
||||
return 0; /* no more codepoints */
|
||||
else {
|
||||
utfint code;
|
||||
l_uint32 code;
|
||||
const char *next = utf8_decode(s + n, &code, strict);
|
||||
if (next == NULL || iscontp(next))
|
||||
return luaL_error(L, MSGInvalid);
|
||||
lua_pushinteger(L, n + 1);
|
||||
lua_pushinteger(L, code);
|
||||
lua_pushinteger(L, l_castU2S(n + 1));
|
||||
lua_pushinteger(L, l_castU2S(code));
|
||||
return 2;
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user