1 /* utf8.h 2 * 3 * Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009 by Larry Wall and others 4 * 5 * You may distribute under the terms of either the GNU General Public 6 * License or the Artistic License, as specified in the README file. 7 * 8 */ 9 10 /* Use UTF-8 as the default script encoding? 11 * Turning this on will break scripts having non-UTF-8 binary 12 * data (such as Latin-1) in string literals. */ 13 #ifdef USE_UTF8_SCRIPTS 14 # define USE_UTF8_IN_NAMES (!IN_BYTES) 15 #else 16 # define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8) 17 #endif 18 19 /* Source backward compatibility. */ 20 #define uvuni_to_utf8(d, uv) uvuni_to_utf8_flags(d, uv, 0) 21 #define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0) 22 23 #ifdef EBCDIC 24 /* The equivalent of these macros but implementing UTF-EBCDIC 25 are in the following header file: 26 */ 27 28 #include "utfebcdic.h" 29 30 #else /* ! EBCDIC */ 31 START_EXTERN_C 32 33 #ifdef DOINIT 34 EXTCONST unsigned char PL_utf8skip[] = { 35 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ 36 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ 37 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ 38 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */ 39 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */ 40 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */ 41 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* scripts */ 42 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,6,6, /* cjk etc. */ 43 7,13, /* Perl extended (not UTF-8). Up to 72bit allowed (64-bit + reserved). */ 44 }; 45 #else 46 EXTCONST unsigned char PL_utf8skip[]; 47 #endif 48 49 END_EXTERN_C 50 51 /* Native character to iso-8859-1 */ 52 #define NATIVE_TO_ASCII(ch) (ch) 53 #define ASCII_TO_NATIVE(ch) (ch) 54 /* Transform after encoding */ 55 #define NATIVE_TO_UTF(ch) (ch) 56 #define UTF_TO_NATIVE(ch) (ch) 57 /* Transforms in wide UV chars */ 58 #define UNI_TO_NATIVE(ch) (ch) 59 #define NATIVE_TO_UNI(ch) (ch) 60 /* Transforms in invariant space */ 61 #define NATIVE_TO_NEED(enc,ch) (ch) 62 #define ASCII_TO_NEED(enc,ch) (ch) 63 64 /* As there are no translations, avoid the function wrapper */ 65 #define utf8n_to_uvchr utf8n_to_uvuni 66 #define uvchr_to_utf8 uvuni_to_utf8 67 68 /* 69 70 The following table is from Unicode 3.2. 71 72 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte 73 74 U+0000..U+007F 00..7F 75 U+0080..U+07FF * C2..DF 80..BF 76 U+0800..U+0FFF E0 * A0..BF 80..BF 77 U+1000..U+CFFF E1..EC 80..BF 80..BF 78 U+D000..U+D7FF ED 80..9F 80..BF 79 U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++ 80 U+E000..U+FFFF EE..EF 80..BF 80..BF 81 U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF 82 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF 83 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF 84 85 Note the gaps before several of the byte entries above marked by '*'. These are 86 caused by legal UTF-8 avoiding non-shortest encodings: it is technically 87 possible to UTF-8-encode a single code point in different ways, but that is 88 explicitly forbidden, and the shortest possible encoding should always be used 89 (and that is what Perl does). 90 91 */ 92 93 /* 94 Another way to look at it, as bits: 95 96 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte 97 98 0aaaaaaa 0aaaaaaa 99 00000bbbbbaaaaaa 110bbbbb 10aaaaaa 100 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa 101 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa 102 103 As you can see, the continuation bytes all begin with C<10>, and the 104 leading bits of the start byte tell how many bytes there are in the 105 encoded character. 106 107 */ 108 109 110 #define UNI_IS_INVARIANT(c) (((UV)c) < 0x80) 111 /* Note that C0 and C1 are invalid in legal UTF8, so the lower bound of the 112 * below might ought to be C2 */ 113 #define UTF8_IS_START(c) (((U8)c) >= 0xc0 && (((U8)c) <= 0xfd)) 114 #define UTF8_IS_CONTINUATION(c) (((U8)c) >= 0x80 && (((U8)c) <= 0xbf)) 115 #define UTF8_IS_CONTINUED(c) (((U8)c) & 0x80) 116 #define UTF8_IS_DOWNGRADEABLE_START(c) (((U8)c & 0xfc) == 0xc0) 117 118 #define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFE << (7-(len)))) 119 #define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2))) 120 121 #define UTF_CONTINUATION_MARK 0x80 122 #define UTF_ACCUMULATION_SHIFT 6 123 #define UTF_CONTINUATION_MASK ((U8)0x3f) 124 125 #ifdef HAS_QUAD 126 #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \ 127 (uv) < 0x800 ? 2 : \ 128 (uv) < 0x10000 ? 3 : \ 129 (uv) < 0x200000 ? 4 : \ 130 (uv) < 0x4000000 ? 5 : \ 131 (uv) < 0x80000000 ? 6 : \ 132 (uv) < UTF8_QUAD_MAX ? 7 : 13 ) 133 #else 134 /* No, I'm not even going to *TRY* putting #ifdef inside a #define */ 135 #define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \ 136 (uv) < 0x800 ? 2 : \ 137 (uv) < 0x10000 ? 3 : \ 138 (uv) < 0x200000 ? 4 : \ 139 (uv) < 0x4000000 ? 5 : \ 140 (uv) < 0x80000000 ? 6 : 7 ) 141 #endif 142 143 #endif /* EBCDIC vs ASCII */ 144 145 /* Rest of these are attributes of Unicode and perl's internals rather than the 146 * encoding, or happen to be the same in both ASCII and EBCDIC (at least at 147 * this level; the macros that some of these call may have different 148 * definitions in the two encodings */ 149 150 #define NATIVE8_TO_UNI(ch) NATIVE_TO_ASCII(ch) /* a clearer synonym */ 151 152 #define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT) | (((U8)new) & UTF_CONTINUATION_MASK)) 153 154 #define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)] 155 156 #define UTF8_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE_TO_UTF(c)) 157 #define NATIVE_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE8_TO_UNI(c)) 158 159 #define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF /* constrained by EBCDIC */ 160 161 /* The macros in the next sets are used to generate the two utf8 or utfebcdic 162 * bytes from an ordinal that is known to fit into two bytes; it must be less 163 * than 0x3FF to work across both encodings. */ 164 /* Nocast allows these to be used in the case label of a switch statement */ 165 #define UTF8_TWO_BYTE_HI_nocast(c) UTF_TO_NATIVE(((c)>>UTF_ACCUMULATION_SHIFT)|UTF_START_MARK(2)) 166 #define UTF8_TWO_BYTE_LO_nocast(c) UTF_TO_NATIVE(((c)&UTF_CONTINUATION_MASK)|UTF_CONTINUATION_MARK) 167 168 #define UTF8_TWO_BYTE_HI(c) ((U8) (UTF8_TWO_BYTE_HI_nocast(c))) 169 #define UTF8_TWO_BYTE_LO(c) ((U8) (UTF8_TWO_BYTE_LO_nocast(c))) 170 171 /* This name is used when the source is a single byte */ 172 #define UTF8_EIGHT_BIT_HI(c) UTF8_TWO_BYTE_HI((U8)(c)) 173 #define UTF8_EIGHT_BIT_LO(c) UTF8_TWO_BYTE_LO((U8)(c)) 174 175 /* 176 * Note: we try to be careful never to call the isXXX_utf8() functions 177 * unless we're pretty sure we've seen the beginning of a UTF-8 or UTFEBCDIC 178 * character. Otherwise we risk loading in the heavy-duty swash_init and 179 * swash_fetch routines unnecessarily. 180 */ 181 #define isIDFIRST_lazy_if(p,c) ((IN_BYTES || (!c || ! UTF8_IS_START(*((const U8*)p)))) \ 182 ? isIDFIRST(*(p)) \ 183 : isIDFIRST_utf8((const U8*)p)) 184 #define isALNUM_lazy_if(p,c) ((IN_BYTES || (!c || ! UTF8_IS_START(*((const U8*)p)))) \ 185 ? isALNUM(*(p)) \ 186 : isALNUM_utf8((const U8*)p)) 187 188 #define isIDFIRST_lazy(p) isIDFIRST_lazy_if(p,1) 189 #define isALNUM_lazy(p) isALNUM_lazy_if(p,1) 190 191 #define UTF8_MAXBYTES 13 192 /* How wide can a single UTF-8 encoded character become in bytes. 193 * NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8 194 * since UTF-8 is an encoding of Unicode and given Unicode's current 195 * upper limit only four bytes is possible. Perl thinks of UTF-8 196 * as a way to encode non-negative integers in a binary format. */ 197 #define UTF8_MAXLEN UTF8_MAXBYTES 198 199 /* The maximum number of UTF-8 bytes a single Unicode character can 200 * uppercase/lowercase/fold into; this number depends on the Unicode 201 * version. An example of maximal expansion is the U+03B0 which 202 * uppercases to U+03C5 U+0308 U+0301. The Unicode databases that 203 * tell these things are UnicodeData.txt, CaseFolding.txt, and 204 * SpecialCasing.txt. */ 205 #define UTF8_MAXBYTES_CASE 6 206 207 #define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES) 208 #define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES) 209 #define IN_UNI_8_BIT ( (CopHINTS_get(PL_curcop) & HINT_UNI_8_BIT) \ 210 && ! IN_LOCALE_RUNTIME && ! IN_BYTES) 211 212 #define UTF8_ALLOW_EMPTY 0x0001 213 #define UTF8_ALLOW_CONTINUATION 0x0002 214 #define UTF8_ALLOW_NON_CONTINUATION 0x0004 215 #define UTF8_ALLOW_FE_FF 0x0008 /* Allow FE or FF start bytes, \ 216 yields above 0x7fffFFFF */ 217 #define UTF8_ALLOW_SHORT 0x0010 /* expecting more bytes */ 218 #define UTF8_ALLOW_SURROGATE 0x0020 219 #define UTF8_ALLOW_FFFF 0x0040 /* Allow UNICODE_ILLEGAL */ 220 #define UTF8_ALLOW_LONG 0x0080 /* expecting fewer bytes */ 221 #define UTF8_ALLOW_ANYUV (UTF8_ALLOW_EMPTY|UTF8_ALLOW_FE_FF|\ 222 UTF8_ALLOW_SURROGATE|UTF8_ALLOW_FFFF) 223 #define UTF8_ALLOW_ANY 0x00FF 224 #define UTF8_CHECK_ONLY 0x0200 225 #define UTF8_ALLOW_DEFAULT (ckWARN(WARN_UTF8) ? 0 : \ 226 UTF8_ALLOW_ANYUV) 227 228 #define UNICODE_SURROGATE_FIRST 0xD800 229 #define UNICODE_SURROGATE_LAST 0xDFFF 230 #define UNICODE_REPLACEMENT 0xFFFD 231 #define UNICODE_BYTE_ORDER_MARK 0xFEFF 232 #define UNICODE_ILLEGAL 0xFFFF 233 234 /* Though our UTF-8 encoding can go beyond this, 235 * let's be conservative and do as Unicode 5.1 says. */ 236 #define PERL_UNICODE_MAX 0x10FFFF 237 238 #define UNICODE_ALLOW_SURROGATE 0x0001 /* Allow UTF-16 surrogates (EVIL) */ 239 #define UNICODE_ALLOW_FDD0 0x0002 /* Allow the U+FDD0...U+FDEF */ 240 #define UNICODE_ALLOW_FFFF 0x0004 /* Allow U+FFF[EF], U+1FFF[EF], ... */ 241 #define UNICODE_ALLOW_SUPER 0x0008 /* Allow past 0x10FFFF */ 242 #define UNICODE_ALLOW_ANY 0x000F 243 244 #define UNICODE_IS_SURROGATE(c) ((c) >= UNICODE_SURROGATE_FIRST && \ 245 (c) <= UNICODE_SURROGATE_LAST) 246 #define UNICODE_IS_REPLACEMENT(c) ((c) == UNICODE_REPLACEMENT) 247 #define UNICODE_IS_BYTE_ORDER_MARK(c) ((c) == UNICODE_BYTE_ORDER_MARK) 248 #define UNICODE_IS_ILLEGAL(c) ((c) == UNICODE_ILLEGAL) 249 250 #ifdef HAS_QUAD 251 # define UTF8_QUAD_MAX UINT64_C(0x1000000000) 252 #endif 253 254 #define UTF8_IS_ASCII(c) UTF8_IS_INVARIANT(c) 255 256 #define UNICODE_GREEK_CAPITAL_LETTER_SIGMA 0x03A3 257 #define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA 0x03C2 258 #define UNICODE_GREEK_SMALL_LETTER_SIGMA 0x03C3 259 260 #define UNI_DISPLAY_ISPRINT 0x0001 261 #define UNI_DISPLAY_BACKSLASH 0x0002 262 #define UNI_DISPLAY_QQ (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH) 263 #define UNI_DISPLAY_REGEX (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH) 264 265 #ifndef EBCDIC 266 # define LATIN_SMALL_LETTER_SHARP_S 0x00DF 267 # define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS 0x00FF 268 # define MICRO_SIGN 0x00B5 269 #endif 270 271 #define ANYOF_FOLD_SHARP_S(node, input, end) \ 272 (ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \ 273 (ANYOF_FLAGS(node) & ANYOF_UNICODE) && \ 274 (ANYOF_FLAGS(node) & ANYOF_FOLD) && \ 275 ((end) > (input) + 1) && \ 276 toLOWER((input)[0]) == 's' && \ 277 toLOWER((input)[1]) == 's') 278 #define SHARP_S_SKIP 2 279 280 #ifdef EBCDIC 281 /* IS_UTF8_CHAR() is not ported to EBCDIC */ 282 #else 283 #define IS_UTF8_CHAR_1(p) \ 284 ((p)[0] <= 0x7F) 285 #define IS_UTF8_CHAR_2(p) \ 286 ((p)[0] >= 0xC2 && (p)[0] <= 0xDF && \ 287 (p)[1] >= 0x80 && (p)[1] <= 0xBF) 288 #define IS_UTF8_CHAR_3a(p) \ 289 ((p)[0] == 0xE0 && \ 290 (p)[1] >= 0xA0 && (p)[1] <= 0xBF && \ 291 (p)[2] >= 0x80 && (p)[2] <= 0xBF) 292 #define IS_UTF8_CHAR_3b(p) \ 293 ((p)[0] >= 0xE1 && (p)[0] <= 0xEC && \ 294 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ 295 (p)[2] >= 0x80 && (p)[2] <= 0xBF) 296 #define IS_UTF8_CHAR_3c(p) \ 297 ((p)[0] == 0xED && \ 298 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ 299 (p)[2] >= 0x80 && (p)[2] <= 0xBF) 300 /* In IS_UTF8_CHAR_3c(p) one could use 301 * (p)[1] >= 0x80 && (p)[1] <= 0x9F 302 * if one wanted to exclude surrogates. */ 303 #define IS_UTF8_CHAR_3d(p) \ 304 ((p)[0] >= 0xEE && (p)[0] <= 0xEF && \ 305 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ 306 (p)[2] >= 0x80 && (p)[2] <= 0xBF) 307 #define IS_UTF8_CHAR_4a(p) \ 308 ((p)[0] == 0xF0 && \ 309 (p)[1] >= 0x90 && (p)[1] <= 0xBF && \ 310 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ 311 (p)[3] >= 0x80 && (p)[3] <= 0xBF) 312 #define IS_UTF8_CHAR_4b(p) \ 313 ((p)[0] >= 0xF1 && (p)[0] <= 0xF3 && \ 314 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ 315 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ 316 (p)[3] >= 0x80 && (p)[3] <= 0xBF) 317 /* In IS_UTF8_CHAR_4c(p) one could use 318 * (p)[0] == 0xF4 319 * if one wanted to stop at the Unicode limit U+10FFFF. 320 * The 0xF7 allows us to go to 0x1fffff (0x200000 would 321 * require five bytes). Not doing any further code points 322 * since that is not needed (and that would not be strict 323 * UTF-8, anyway). The "slow path" in Perl_is_utf8_char() 324 * will take care of the "extended UTF-8". */ 325 #define IS_UTF8_CHAR_4c(p) \ 326 ((p)[0] == 0xF4 && (p)[0] <= 0xF7 && \ 327 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \ 328 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \ 329 (p)[3] >= 0x80 && (p)[3] <= 0xBF) 330 331 #define IS_UTF8_CHAR_3(p) \ 332 (IS_UTF8_CHAR_3a(p) || \ 333 IS_UTF8_CHAR_3b(p) || \ 334 IS_UTF8_CHAR_3c(p) || \ 335 IS_UTF8_CHAR_3d(p)) 336 #define IS_UTF8_CHAR_4(p) \ 337 (IS_UTF8_CHAR_4a(p) || \ 338 IS_UTF8_CHAR_4b(p) || \ 339 IS_UTF8_CHAR_4c(p)) 340 341 /* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it 342 * (1) allows UTF-8 encoded UTF-16 surrogates 343 * (2) it allows code points past U+10FFFF. 344 * The Perl_is_utf8_char() full "slow" code will handle the Perl 345 * "extended UTF-8". */ 346 #define IS_UTF8_CHAR(p, n) \ 347 ((n) == 1 ? IS_UTF8_CHAR_1(p) : \ 348 (n) == 2 ? IS_UTF8_CHAR_2(p) : \ 349 (n) == 3 ? IS_UTF8_CHAR_3(p) : \ 350 (n) == 4 ? IS_UTF8_CHAR_4(p) : 0) 351 352 #define IS_UTF8_CHAR_FAST(n) ((n) <= 4) 353 354 #endif /* IS_UTF8_CHAR() for UTF-8 */ 355 356 /* 357 * Local variables: 358 * c-indentation-style: bsd 359 * c-basic-offset: 4 360 * indent-tabs-mode: t 361 * End: 362 * 363 * ex: set ts=8 sts=4 sw=4 noet: 364 */ 365