LXR perl/utf8.h

   1/*    utf8.h
   2 *
   3 *    Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 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
  31START_EXTERN_C
  32
  33#ifdef DOINIT
  34EXTCONST unsigned char PL_utf8skip[] = {
  351,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
  361,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
  371,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
  381,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
  391,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */
  401,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */
  412,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* scripts */
  423,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,6,6,     /* cjk etc. */
  437,13, /* Perl extended (not UTF-8).  Up to 72bit allowed (64-bit + reserved). */
  44};
  45#else
  46EXTCONST unsigned char PL_utf8skip[];
  47#endif
  48
  49END_EXTERN_C
  50#define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)]
  51
  52/* Native character to iso-8859-1 */
  53#define NATIVE_TO_ASCII(ch)      (ch)
  54#define ASCII_TO_NATIVE(ch)      (ch)
  55/* Transform after encoding */
  56#define NATIVE_TO_UTF(ch)        (ch)
  57#define UTF_TO_NATIVE(ch)        (ch)
  58/* Transforms in wide UV chars */
  59#define UNI_TO_NATIVE(ch)        (ch)
  60#define NATIVE_TO_UNI(ch)        (ch)
  61/* Transforms in invariant space */
  62#define NATIVE_TO_NEED(enc,ch)   (ch)
  63#define ASCII_TO_NEED(enc,ch)    (ch)
  64
  65/* As there are no translations avoid the function wrapper */
  66#define utf8n_to_uvchr utf8n_to_uvuni
  67#define uvchr_to_utf8  uvuni_to_utf8
  68
  69/*
  70
  71 The following table is from Unicode 3.2.
  72
  73 Code Points            1st Byte  2nd Byte  3rd Byte  4th Byte
  74
  75   U+0000..U+007F       00..7F
  76   U+0080..U+07FF       C2..DF    80..BF
  77   U+0800..U+0FFF       E0        A0..BF    80..BF
  78   U+1000..U+CFFF       E1..EC    80..BF    80..BF
  79   U+D000..U+D7FF       ED        80..9F    80..BF
  80   U+D800..U+DFFF       ******* ill-formed *******
  81   U+E000..U+FFFF       EE..EF    80..BF    80..BF
  82  U+10000..U+3FFFF      F0        90..BF    80..BF    80..BF
  83  U+40000..U+FFFFF      F1..F3    80..BF    80..BF    80..BF
  84 U+100000..U+10FFFF     F4        80..8F    80..BF    80..BF
  85
  86Note the A0..BF in U+0800..U+0FFF, the 80..9F in U+D000...U+D7FF,
  87the 90..BF in U+10000..U+3FFFF, and the 80...8F in U+100000..U+10FFFF.
  88The "gaps" are caused by legal UTF-8 avoiding non-shortest encodings:
  89it is technically possible to UTF-8-encode a single code point in different
  90ways, but that is explicitly forbidden, and the shortest possible encoding
  91should always be used (and that is what Perl does).
  92
  93 */
  94
  95/*
  96 Another way to look at it, as bits:
  97
  98 Code Points                    1st Byte   2nd Byte  3rd Byte  4th Byte
  99
 100                    0aaaaaaa     0aaaaaaa
 101            00000bbbbbaaaaaa     110bbbbb  10aaaaaa
 102            ccccbbbbbbaaaaaa     1110cccc  10bbbbbb  10aaaaaa
 103  00000dddccccccbbbbbbaaaaaa     11110ddd  10cccccc  10bbbbbb  10aaaaaa
 104
 105As you can see, the continuation bytes all begin with C<10>, and the
 106leading bits of the start byte tell how many bytes the are in the
 107encoded character.
 108
 109*/
 110
 111
 112#define UNI_IS_INVARIANT(c)             (((UV)c) <  0x80)
 113#define UTF8_IS_INVARIANT(c)            UNI_IS_INVARIANT(NATIVE_TO_UTF(c))
 114#define NATIVE_IS_INVARIANT(c)          UNI_IS_INVARIANT(NATIVE_TO_ASCII(c))
 115#define UTF8_IS_START(c)                (((U8)c) >= 0xc0 && (((U8)c) <= 0xfd))
 116#define UTF8_IS_CONTINUATION(c)         (((U8)c) >= 0x80 && (((U8)c) <= 0xbf))
 117#define UTF8_IS_CONTINUED(c)            (((U8)c) &  0x80)
 118#define UTF8_IS_DOWNGRADEABLE_START(c)  (((U8)c & 0xfc) == 0xc0)
 119
 120#define UTF_START_MARK(len) ((len >  7) ? 0xFF : (0xFE << (7-len)))
 121#define UTF_START_MASK(len) ((len >= 7) ? 0x00 : (0x1F >> (len-2)))
 122
 123#define UTF_CONTINUATION_MARK           0x80
 124#define UTF_ACCUMULATION_SHIFT          6
 125#define UTF_CONTINUATION_MASK           ((U8)0x3f)
 126#define UTF8_ACCUMULATE(old, new)       (((old) << UTF_ACCUMULATION_SHIFT) | (((U8)new) & UTF_CONTINUATION_MASK))
 127
 128#define UTF8_EIGHT_BIT_HI(c)    ((((U8)(c))>>UTF_ACCUMULATION_SHIFT)|UTF_START_MARK(2))
 129#define UTF8_EIGHT_BIT_LO(c)    (((((U8)(c)))&UTF_CONTINUATION_MASK)|UTF_CONTINUATION_MARK)
 130
 131#ifdef HAS_QUAD
 132#define UNISKIP(uv) ( (uv) < 0x80           ? 1 : \
 133                      (uv) < 0x800          ? 2 : \
 134                      (uv) < 0x10000        ? 3 : \
 135                      (uv) < 0x200000       ? 4 : \
 136                      (uv) < 0x4000000      ? 5 : \
 137                      (uv) < 0x80000000     ? 6 : \
 138                      (uv) < UTF8_QUAD_MAX ? 7 : 13 )
 139#else
 140/* No, I'm not even going to *TRY* putting #ifdef inside a #define */
 141#define UNISKIP(uv) ( (uv) < 0x80           ? 1 : \
 142                      (uv) < 0x800          ? 2 : \
 143                      (uv) < 0x10000        ? 3 : \
 144                      (uv) < 0x200000       ? 4 : \
 145                      (uv) < 0x4000000      ? 5 : \
 146                      (uv) < 0x80000000     ? 6 : 7 )
 147#endif
 148
 149/*
 150 * Note: we try to be careful never to call the isXXX_utf8() functions
 151 * unless we're pretty sure we've seen the beginning of a UTF-8 character
 152 * (that is, the two high bits are set).  Otherwise we risk loading in the
 153 * heavy-duty swash_init and swash_fetch routines unnecessarily.
 154 */
 155#define isIDFIRST_lazy_if(p,c) ((IN_BYTES || (!c || (*((const U8*)p) < 0xc0))) \
 156                                ? isIDFIRST(*(p)) \
 157                                : isIDFIRST_utf8((const U8*)p))
 158#define isALNUM_lazy_if(p,c)   ((IN_BYTES || (!c || (*((const U8*)p) < 0xc0))) \
 159                                ? isALNUM(*(p)) \
 160                                : isALNUM_utf8((const U8*)p))
 161
 162
 163#endif /* EBCDIC vs ASCII */
 164
 165/* Rest of these are attributes of Unicode and perl's internals rather than the encoding */
 166
 167#define isIDFIRST_lazy(p)       isIDFIRST_lazy_if(p,1)
 168#define isALNUM_lazy(p)         isALNUM_lazy_if(p,1)
 169
 170#define UTF8_MAXBYTES 13
 171/* How wide can a single UTF-8 encoded character become in bytes.
 172 * NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8
 173 * since UTF-8 is an encoding of Unicode and given Unicode's current
 174 * upper limit only four bytes is possible.  Perl thinks of UTF-8
 175 * as a way to encode non-negative integers in a binary format. */
 176#define UTF8_MAXLEN UTF8_MAXBYTES
 177
 178#define UTF8_MAXLEN_UCLC 3              /* Obsolete, do not use. */
 179#define UTF8_MAXLEN_UCLC_MULT 39        /* Obsolete, do not use. */
 180#define UTF8_MAXLEN_FOLD 3              /* Obsolete, do not use. */
 181#define UTF8_MAXLEN_FOLD_MULT 39        /* Obsolete, do not use. */
 182
 183/* The maximum number of UTF-8 bytes a single Unicode character can
 184 * uppercase/lowercase/fold into; this number depends on the Unicode
 185 * version.  An example of maximal expansion is the U+03B0 which
 186 * uppercases to U+03C5 U+0308 U+0301.  The Unicode databases that
 187 * tell these things are UnicodeDatabase.txt, CaseFolding.txt, and
 188 * SpecialCasing.txt. */
 189#define UTF8_MAXBYTES_CASE      6
 190
 191#define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES)
 192#define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES)
 193
 194#define UTF8_ALLOW_EMPTY                0x0001
 195#define UTF8_ALLOW_CONTINUATION         0x0002
 196#define UTF8_ALLOW_NON_CONTINUATION     0x0004
 197#define UTF8_ALLOW_FE_FF                0x0008 /* Allow above 0x7fffFFFF */
 198#define UTF8_ALLOW_SHORT                0x0010
 199#define UTF8_ALLOW_SURROGATE            0x0020
 200#define UTF8_ALLOW_FFFF                 0x0040 /* Allow UNICODE_ILLEGAL */
 201#define UTF8_ALLOW_LONG                 0x0080
 202#define UTF8_ALLOW_ANYUV                (UTF8_ALLOW_EMPTY|UTF8_ALLOW_FE_FF|\
 203                                         UTF8_ALLOW_SURROGATE|UTF8_ALLOW_FFFF)
 204#define UTF8_ALLOW_ANY                  0x00FF
 205#define UTF8_CHECK_ONLY                 0x0200
 206#define UTF8_ALLOW_DEFAULT              (ckWARN(WARN_UTF8) ? 0 : \
 207                                         UTF8_ALLOW_ANYUV)
 208
 209#define UNICODE_SURROGATE_FIRST         0xD800
 210#define UNICODE_SURROGATE_LAST          0xDFFF
 211#define UNICODE_REPLACEMENT             0xFFFD
 212#define UNICODE_BYTE_ORDER_MARK         0xFEFF
 213#define UNICODE_ILLEGAL                 0xFFFF
 214
 215/* Though our UTF-8 encoding can go beyond this,
 216 * let's be conservative and do as Unicode 3.2 says. */
 217#define PERL_UNICODE_MAX        0x10FFFF
 218
 219#define UNICODE_ALLOW_SURROGATE 0x0001  /* Allow UTF-16 surrogates (EVIL) */
 220#define UNICODE_ALLOW_FDD0      0x0002  /* Allow the U+FDD0...U+FDEF */
 221#define UNICODE_ALLOW_FFFF      0x0004  /* Allow U+FFF[EF], U+1FFF[EF], ... */
 222#define UNICODE_ALLOW_SUPER     0x0008  /* Allow past 0x10FFFF */
 223#define UNICODE_ALLOW_ANY       0x000F
 224
 225#define UNICODE_IS_SURROGATE(c)         ((c) >= UNICODE_SURROGATE_FIRST && \
 226                                         (c) <= UNICODE_SURROGATE_LAST)
 227#define UNICODE_IS_REPLACEMENT(c)       ((c) == UNICODE_REPLACEMENT)
 228#define UNICODE_IS_BYTE_ORDER_MARK(c)   ((c) == UNICODE_BYTE_ORDER_MARK)
 229#define UNICODE_IS_ILLEGAL(c)           ((c) == UNICODE_ILLEGAL)
 230
 231#ifdef HAS_QUAD
 232#    define UTF8_QUAD_MAX       UINT64_C(0x1000000000)
 233#endif
 234
 235#define UTF8_IS_ASCII(c) UTF8_IS_INVARIANT(c)
 236
 237#define UNICODE_LATIN_SMALL_LETTER_SHARP_S      0x00DF
 238#define UNICODE_GREEK_CAPITAL_LETTER_SIGMA      0x03A3
 239#define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA  0x03C2
 240#define UNICODE_GREEK_SMALL_LETTER_SIGMA        0x03C3
 241
 242#define EBCDIC_LATIN_SMALL_LETTER_SHARP_S       0x0059
 243
 244#define UNI_DISPLAY_ISPRINT     0x0001
 245#define UNI_DISPLAY_BACKSLASH   0x0002
 246#define UNI_DISPLAY_QQ          (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
 247#define UNI_DISPLAY_REGEX       (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
 248
 249#ifdef EBCDIC
 250#   define ANYOF_FOLD_SHARP_S(node, input, end) \
 251        (ANYOF_BITMAP_TEST(node, EBCDIC_LATIN_SMALL_LETTER_SHARP_S) && \
 252         (ANYOF_FLAGS(node) & ANYOF_UNICODE) && \
 253         (ANYOF_FLAGS(node) & ANYOF_FOLD) && \
 254         ((end) > (input) + 1) && \
 255         toLOWER((input)[0]) == 's' && \
 256         toLOWER((input)[1]) == 's')
 257#else
 258#   define ANYOF_FOLD_SHARP_S(node, input, end) \
 259        (ANYOF_BITMAP_TEST(node, UNICODE_LATIN_SMALL_LETTER_SHARP_S) && \
 260         (ANYOF_FLAGS(node) & ANYOF_UNICODE) && \
 261         (ANYOF_FLAGS(node) & ANYOF_FOLD) && \
 262         ((end) > (input) + 1) && \
 263         toLOWER((input)[0]) == 's' && \
 264         toLOWER((input)[1]) == 's')
 265#endif
 266#define SHARP_S_SKIP 2
 267
 268#ifdef EBCDIC
 269/* IS_UTF8_CHAR() is not ported to EBCDIC */
 270#else
 271#define IS_UTF8_CHAR_1(p)       \
 272        ((p)[0] <= 0x7F)
 273#define IS_UTF8_CHAR_2(p)       \
 274        ((p)[0] >= 0xC2 && (p)[0] <= 0xDF && \
 275         (p)[1] >= 0x80 && (p)[1] <= 0xBF)
 276#define IS_UTF8_CHAR_3a(p)      \
 277        ((p)[0] == 0xE0 && \
 278         (p)[1] >= 0xA0 && (p)[1] <= 0xBF && \
 279         (p)[2] >= 0x80 && (p)[2] <= 0xBF)
 280#define IS_UTF8_CHAR_3b(p)      \
 281        ((p)[0] >= 0xE1 && (p)[0] <= 0xEC && \
 282         (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
 283         (p)[2] >= 0x80 && (p)[2] <= 0xBF)
 284#define IS_UTF8_CHAR_3c(p)      \
 285        ((p)[0] == 0xED && \
 286         (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
 287         (p)[2] >= 0x80 && (p)[2] <= 0xBF)
 288/* In IS_UTF8_CHAR_3c(p) one could use
 289 * (p)[1] >= 0x80 && (p)[1] <= 0x9F
 290 * if one wanted to exclude surrogates. */
 291#define IS_UTF8_CHAR_3d(p)      \
 292        ((p)[0] >= 0xEE && (p)[0] <= 0xEF && \
 293         (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
 294         (p)[2] >= 0x80 && (p)[2] <= 0xBF)
 295#define IS_UTF8_CHAR_4a(p)      \
 296        ((p)[0] == 0xF0 && \
 297         (p)[1] >= 0x90 && (p)[1] <= 0xBF && \
 298         (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
 299         (p)[3] >= 0x80 && (p)[3] <= 0xBF)
 300#define IS_UTF8_CHAR_4b(p)      \
 301        ((p)[0] >= 0xF1 && (p)[0] <= 0xF3 && \
 302         (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
 303         (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
 304         (p)[3] >= 0x80 && (p)[3] <= 0xBF)
 305/* In IS_UTF8_CHAR_4c(p) one could use
 306 * (p)[0] == 0xF4
 307 * if one wanted to stop at the Unicode limit U+10FFFF.
 308 * The 0xF7 allows us to go to 0x1fffff (0x200000 would
 309 * require five bytes).  Not doing any further code points
 310 * since that is not needed (and that would not be strict
 311 * UTF-8, anyway).  The "slow path" in Perl_is_utf8_char()
 312 * will take care of the "extended UTF-8". */
 313#define IS_UTF8_CHAR_4c(p)      \
 314        ((p)[0] == 0xF4 && (p)[0] <= 0xF7 && \
 315         (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
 316         (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
 317         (p)[3] >= 0x80 && (p)[3] <= 0xBF)
 318
 319#define IS_UTF8_CHAR_3(p)       \
 320        (IS_UTF8_CHAR_3a(p) || \
 321         IS_UTF8_CHAR_3b(p) || \
 322         IS_UTF8_CHAR_3c(p) || \
 323         IS_UTF8_CHAR_3d(p))
 324#define IS_UTF8_CHAR_4(p)       \
 325        (IS_UTF8_CHAR_4a(p) || \
 326         IS_UTF8_CHAR_4b(p) || \
 327         IS_UTF8_CHAR_4c(p))
 328
 329/* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it
 330 * (1) allows UTF-8 encoded UTF-16 surrogates
 331 * (2) it allows code points past U+10FFFF.
 332 * The Perl_is_utf8_char() full "slow" code will handle the Perl
 333 * "extended UTF-8". */
 334#define IS_UTF8_CHAR(p, n)      \
 335        ((n) == 1 ? IS_UTF8_CHAR_1(p) : \
 336         (n) == 2 ? IS_UTF8_CHAR_2(p) : \
 337         (n) == 3 ? IS_UTF8_CHAR_3(p) : \
 338         (n) == 4 ? IS_UTF8_CHAR_4(p) : 0)
 339
 340#define IS_UTF8_CHAR_FAST(n) ((n) <= 4)
 341
 342#endif /* IS_UTF8_CHAR() for UTF-8 */
 343