diff options
author | Roberto Ierusalimschy <roberto@inf.puc-rio.br> | 2018-07-12 15:56:44 -0300 |
---|---|---|
committer | Roberto Ierusalimschy <roberto@inf.puc-rio.br> | 2018-07-12 15:56:44 -0300 |
commit | fb18346dddcb0400d0396111c56a817a8d4bd8bd (patch) | |
tree | ba3c4afcbd9f300cec3b974471e37fd70f2a49dd /lutf8lib.c | |
parent | 96f9643f330a4bf0f8cd1973fecdd7161ffbbf68 (diff) | |
download | lua-fb18346dddcb0400d0396111c56a817a8d4bd8bd.tar.gz lua-fb18346dddcb0400d0396111c56a817a8d4bd8bd.tar.bz2 lua-fb18346dddcb0400d0396111c56a817a8d4bd8bd.zip |
Avoid using 'int' for UTF-8 values
An 'int' may have only 16 bits, so it may not be big enough for UTF-8
values. The new type 'utfint' (in the utf8 library) ensures at least
21 bits for those values.
Diffstat (limited to 'lutf8lib.c')
-rw-r--r-- | lutf8lib.c | 33 |
1 files changed, 22 insertions, 11 deletions
@@ -1,5 +1,5 @@ | |||
1 | /* | 1 | /* |
2 | ** $Id: lutf8lib.c,v 1.16 2016/12/22 13:08:50 roberto Exp roberto $ | 2 | ** $Id: lutf8lib.c $ |
3 | ** Standard library for UTF-8 manipulation | 3 | ** Standard library for UTF-8 manipulation |
4 | ** See Copyright Notice in lua.h | 4 | ** See Copyright Notice in lua.h |
5 | */ | 5 | */ |
@@ -20,8 +20,19 @@ | |||
20 | #include "lauxlib.h" | 20 | #include "lauxlib.h" |
21 | #include "lualib.h" | 21 | #include "lualib.h" |
22 | 22 | ||
23 | |||
23 | #define MAXUNICODE 0x10FFFF | 24 | #define MAXUNICODE 0x10FFFF |
24 | 25 | ||
26 | /* | ||
27 | ** Integer type for decoded UTF-8 values; MAXUNICODE needs 21 bits. | ||
28 | */ | ||
29 | #if LUAI_BITSINT >= 21 | ||
30 | typedef unsigned int utfint; | ||
31 | #else | ||
32 | typedef unsigned long utfint; | ||
33 | #endif | ||
34 | |||
35 | |||
25 | #define iscont(p) ((*(p) & 0xC0) == 0x80) | 36 | #define iscont(p) ((*(p) & 0xC0) == 0x80) |
26 | 37 | ||
27 | 38 | ||
@@ -37,11 +48,11 @@ static lua_Integer u_posrelat (lua_Integer pos, size_t len) { | |||
37 | /* | 48 | /* |
38 | ** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid. | 49 | ** Decode one UTF-8 sequence, returning NULL if byte sequence is invalid. |
39 | */ | 50 | */ |
40 | static const char *utf8_decode (const char *o, int *val) { | 51 | static const char *utf8_decode (const char *o, utfint *val) { |
41 | static const unsigned int limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFF}; | 52 | static const unsigned int limits[] = {0xFF, 0x7F, 0x7FF, 0xFFFF}; |
42 | const unsigned char *s = (const unsigned char *)o; | 53 | const unsigned char *s = (const unsigned char *)o; |
43 | unsigned int c = s[0]; | 54 | unsigned int c = s[0]; |
44 | unsigned int res = 0; /* final result */ | 55 | utfint res = 0; /* final result */ |
45 | if (c < 0x80) /* ascii? */ | 56 | if (c < 0x80) /* ascii? */ |
46 | res = c; | 57 | res = c; |
47 | else { | 58 | else { |
@@ -53,7 +64,7 @@ static const char *utf8_decode (const char *o, int *val) { | |||
53 | res = (res << 6) | (cc & 0x3F); /* add lower 6 bits from cont. byte */ | 64 | res = (res << 6) | (cc & 0x3F); /* add lower 6 bits from cont. byte */ |
54 | c <<= 1; /* to test next bit */ | 65 | c <<= 1; /* to test next bit */ |
55 | } | 66 | } |
56 | res |= ((c & 0x7F) << (count * 5)); /* add first byte */ | 67 | res |= ((utfint)(c & 0x7F) << (count * 5)); /* add first byte */ |
57 | if (count > 3 || res > MAXUNICODE || res <= limits[count]) | 68 | if (count > 3 || res > MAXUNICODE || res <= limits[count]) |
58 | return NULL; /* invalid byte sequence */ | 69 | return NULL; /* invalid byte sequence */ |
59 | s += count; /* skip continuation bytes read */ | 70 | s += count; /* skip continuation bytes read */ |
@@ -69,8 +80,8 @@ static const char *utf8_decode (const char *o, int *val) { | |||
69 | ** that interval | 80 | ** that interval |
70 | */ | 81 | */ |
71 | static int utflen (lua_State *L) { | 82 | static int utflen (lua_State *L) { |
72 | int n = 0; | 83 | lua_Integer n = 0; /* counter for the number of characters */ |
73 | size_t len; | 84 | size_t len; /* string length in bytes */ |
74 | const char *s = luaL_checklstring(L, 1, &len); | 85 | const char *s = luaL_checklstring(L, 1, &len); |
75 | lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len); | 86 | lua_Integer posi = u_posrelat(luaL_optinteger(L, 2, 1), len); |
76 | lua_Integer posj = u_posrelat(luaL_optinteger(L, 3, -1), len); | 87 | lua_Integer posj = u_posrelat(luaL_optinteger(L, 3, -1), len); |
@@ -109,12 +120,12 @@ static int codepoint (lua_State *L) { | |||
109 | if (posi > pose) return 0; /* empty interval; return no values */ | 120 | if (posi > pose) return 0; /* empty interval; return no values */ |
110 | if (pose - posi >= INT_MAX) /* (lua_Integer -> int) overflow? */ | 121 | if (pose - posi >= INT_MAX) /* (lua_Integer -> int) overflow? */ |
111 | return luaL_error(L, "string slice too long"); | 122 | return luaL_error(L, "string slice too long"); |
112 | n = (int)(pose - posi) + 1; | 123 | n = (int)(pose - posi) + 1; /* upper bound for number of returns */ |
113 | luaL_checkstack(L, n, "string slice too long"); | 124 | luaL_checkstack(L, n, "string slice too long"); |
114 | n = 0; | 125 | n = 0; /* count the number of returns */ |
115 | se = s + pose; | 126 | se = s + pose; /* string end */ |
116 | for (s += posi - 1; s < se;) { | 127 | for (s += posi - 1; s < se;) { |
117 | int code; | 128 | utfint code; |
118 | s = utf8_decode(s, &code); | 129 | s = utf8_decode(s, &code); |
119 | if (s == NULL) | 130 | if (s == NULL) |
120 | return luaL_error(L, "invalid UTF-8 code"); | 131 | return luaL_error(L, "invalid UTF-8 code"); |
@@ -211,7 +222,7 @@ static int iter_aux (lua_State *L) { | |||
211 | if (n >= (lua_Integer)len) | 222 | if (n >= (lua_Integer)len) |
212 | return 0; /* no more codepoints */ | 223 | return 0; /* no more codepoints */ |
213 | else { | 224 | else { |
214 | int code; | 225 | utfint code; |
215 | const char *next = utf8_decode(s + n, &code); | 226 | const char *next = utf8_decode(s + n, &code); |
216 | if (next == NULL || iscont(next)) | 227 | if (next == NULL || iscont(next)) |
217 | return luaL_error(L, "invalid UTF-8 code"); | 228 | return luaL_error(L, "invalid UTF-8 code"); |