src/character.h

   1 /* Header for multibyte character handler.
   2    Copyright (C) 1995, 1997, 1998 Electrotechnical Laboratory, JAPAN.
   3      Licensed to the Free Software Foundation.
   4    Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008
   5      National Institute of Advanced Industrial Science and Technology (AIST)
   6      Registration Number H13PRO009
   7
   8 This file is part of GNU Emacs.
   9
  10 GNU Emacs is free software; you can redistribute it and/or modify
  11 it under the terms of the GNU General Public License as published by
  12 the Free Software Foundation; either version 3, or (at your option)
  13 any later version.
  14
  15 GNU Emacs is distributed in the hope that it will be useful,
  16 but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 GNU General Public License for more details.
  19
  20 You should have received a copy of the GNU General Public License
  21 along with GNU Emacs; see the file COPYING.  If not, write to the
  22 Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  23 Boston, MA 02110-1301, USA.  */
  24
  25 #ifndef EMACS_CHARACTER_H
  26 #define EMACS_CHARACTER_H
  27
  28 /* character code       1st byte   byte sequence
  29    --------------       --------   -------------
  30         0-7F            00..7F     0xxxxxxx
  31        80-7FF           C2..DF     110xxxxx 10xxxxxx
  32       800-FFFF          E0..EF     1110xxxx 10xxxxxx 10xxxxxx
  33     10000-1FFFFF        F0..F7     11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
  34    200000-3FFF7F        F8         11111000 1000xxxx 10xxxxxx 10xxxxxx 10xxxxxx
  35    3FFF80-3FFFFF        C0..C1     1100000x 10xxxxxx (for eight-bit-char)
  36    400000-...           invalid
  37
  38    invalid 1st byte     80..BF     10xxxxxx
  39                         F9..FF     11111xxx (xxx != 000)
  40 */
  41
  42 /* Maximum character code ((1 << CHARACTERBITS) - 1).  */
  43 #define MAX_CHAR  0x3FFFFF
  44
  45 /* Maximum Unicode character code.  */
  46 #define MAX_UNICODE_CHAR 0x10FFFF
  47
  48 /* Maximum N-byte character codes.  */
  49 #define MAX_1_BYTE_CHAR 0x7F
  50 #define MAX_2_BYTE_CHAR 0x7FF
  51 #define MAX_3_BYTE_CHAR 0xFFFF
  52 #define MAX_4_BYTE_CHAR 0x1FFFFF
  53 #define MAX_5_BYTE_CHAR 0x3FFF7F
  54
  55 /* Minimum leading code of multibyte characters.  */
  56 #define MIN_MULTIBYTE_LEADING_CODE 0xC0
  57 /* Maximum leading code of multibyte characters.  */
  58 #define MAX_MULTIBYTE_LEADING_CODE 0xF8
  59
  60 /* Nonzero iff C is a character that corresponds to a raw 8-bit
  61    byte.  */
  62 #define CHAR_BYTE8_P(c) ((c) > MAX_5_BYTE_CHAR)
  63
  64 /* Return the character code for raw 8-bit byte BYTE.  */
  65 #define BYTE8_TO_CHAR(byte) ((byte) + 0x3FFF00)
  66
  67 /* Return the raw 8-bit byte for character C.  */
  68 #define CHAR_TO_BYTE8(c)        \
  69   (CHAR_BYTE8_P (c)             \
  70    ? (c) - 0x3FFF00             \
  71    : multibyte_char_to_unibyte (c, Qnil))
  72
  73 /* Nonzero iff BYTE is the 1st byte of a multibyte form of a character
  74    that corresponds to a raw 8-bit byte.  */
  75 #define CHAR_BYTE8_HEAD_P(byte) ((byte) == 0xC0 || (byte) == 0xC1)
  76
  77 /* Mapping table from unibyte chars to multibyte chars.  */
  78 extern int unibyte_to_multibyte_table[256];
  79
  80 /* Convert the unibyte character C to the corresponding multibyte
  81    character.  If C can't be converted, return C.  */
  82 #define unibyte_char_to_multibyte(c)    \
  83   ((c) < 256 ? unibyte_to_multibyte_table[(c)] : (c))
  84
  85 /* Nth element is 1 iff unibyte char N can be mapped to a multibyte
  86    char.  */
  87 extern char unibyte_has_multibyte_table[256];
  88
  89 #define UNIBYTE_CHAR_HAS_MULTIBYTE_P(c) (unibyte_has_multibyte_table[(c)])
  90
  91 /* If C is not ASCII, make it unibyte. */
  92 #define MAKE_CHAR_UNIBYTE(c)    \
  93   do {                          \
  94     if (! ASCII_CHAR_P (c))     \
  95       c = CHAR_TO_BYTE8 (c);    \
  96   } while (0)
  97
  98
  99 /* If C is not ASCII, make it multibyte.  It assumes C < 256.  */
 100 #define MAKE_CHAR_MULTIBYTE(c) ((c) = unibyte_to_multibyte_table[(c)])
 101
 102 /* This is the maximum byte length of multibyte form.  */
 103 #define MAX_MULTIBYTE_LENGTH 5
 104
 105 /* Return a Lisp character whose character code is C.  It assumes C is
 106    a valid character code.  */
 107 #define make_char(c) make_number (c)
 108
 109 /* Nonzero iff C is an ASCII byte.  */
 110 #define ASCII_BYTE_P(c) ((unsigned) (c) < 0x80)
 111
 112 /* Nonzero iff X is a character.  */
 113 #define CHARACTERP(x) (NATNUMP (x) && XFASTINT (x) <= MAX_CHAR)
 114
 115 /* Nonzero iff C is valid as a character code.  GENERICP is not used
 116    now.  */
 117 #define CHAR_VALID_P(c, genericp) ((unsigned) (c) <= MAX_CHAR)
 118
 119 /* Check if Lisp object X is a character or not.  */
 120 #define CHECK_CHARACTER(x) \
 121   CHECK_TYPE (CHARACTERP (x), Qcharacterp, x)
 122
 123 #define CHECK_CHARACTER_CAR(x) \
 124   do {                                  \
 125     Lisp_Object tmp = XCAR (x);         \
 126     CHECK_CHARACTER (tmp);              \
 127     XSETCAR ((x), tmp);                 \
 128   } while (0)
 129
 130 #define CHECK_CHARACTER_CDR(x) \
 131   do {                                  \
 132     Lisp_Object tmp = XCDR (x);         \
 133     CHECK_CHARACTER (tmp);              \
 134     XSETCDR ((x), tmp);                 \
 135   } while (0)
 136
 137 /* Nonzero iff C is an ASCII character.  */
 138 #define ASCII_CHAR_P(c) ((unsigned) (c) < 0x80)
 139
 140 /* Nonzero iff C is a character of code less than 0x100.  */
 141 #define SINGLE_BYTE_CHAR_P(c) ((unsigned) (c) < 0x100)
 142
 143 /* Nonzero if character C has a printable glyph.  */
 144 #define CHAR_PRINTABLE_P(c)     \
 145   (((c) >= 32 && ((c) < 127)    \
 146     || ! NILP (CHAR_TABLE_REF (Vprintable_chars, (c)))))
 147
 148 /* Return byte length of multibyte form for character C.  */
 149 #define CHAR_BYTES(c)                   \
 150   ( (c) <= MAX_1_BYTE_CHAR ? 1          \
 151     : (c) <= MAX_2_BYTE_CHAR ? 2        \
 152     : (c) <= MAX_3_BYTE_CHAR ? 3        \
 153     : (c) <= MAX_4_BYTE_CHAR ? 4        \
 154     : (c) <= MAX_5_BYTE_CHAR ? 5        \
 155     : 2)
 156
 157
 158 /* Return the leading code of multibyte form of C.  */
 159 #define CHAR_LEADING_CODE(c)                            \
 160   ((c) <= MAX_1_BYTE_CHAR ? c                           \
 161    : (c) <= MAX_2_BYTE_CHAR ? (0xC0 | ((c) >> 6))       \
 162    : (c) <= MAX_3_BYTE_CHAR ? (0xE0 | ((c) >> 12))      \
 163    : (c) <= MAX_4_BYTE_CHAR ? (0xF0 | ((c) >> 18))      \
 164    : (c) <= MAX_5_BYTE_CHAR ? 0xF8                      \
 165    : (0xC0 | (((c) >> 6) & 0x01)))
 166
 167
 168 /* Store multibyte form of the character C in P.  The caller should
 169    allocate at least MAX_MULTIBYTE_LENGTH bytes area at P in advance.
 170    Returns the length of the multibyte form.  */
 171
 172 #define CHAR_STRING(c, p)                       \
 173   ((unsigned) (c) <= MAX_1_BYTE_CHAR            \
 174    ? ((p)[0] = (c),                             \
 175       1)                                        \
 176    : (unsigned) (c) <= MAX_2_BYTE_CHAR          \
 177    ? ((p)[0] = (0xC0 | ((c) >> 6)),             \
 178       (p)[1] = (0x80 | ((c) & 0x3F)),           \
 179       2)                                        \
 180    : (unsigned) (c) <= MAX_3_BYTE_CHAR          \
 181    ? ((p)[0] = (0xE0 | ((c) >> 12)),            \
 182       (p)[1] = (0x80 | (((c) >> 6) & 0x3F)),    \
 183       (p)[2] = (0x80 | ((c) & 0x3F)),           \
 184       3)                                        \
 185    : char_string ((unsigned) c, p))
 186
 187 /* Store multibyte form of byte B in P.  The caller should allocate at
 188    least MAX_MULTIBYTE_LENGTH bytes area at P in advance.  Returns the
 189    length of the multibyte form.  */
 190
 191 #define BYTE8_STRING(b, p)                      \
 192   ((p)[0] = (0xC0 | (((b) >> 6) & 0x01)),       \
 193    (p)[1] = (0x80 | ((b) & 0x3F)),              \
 194    2)
 195
 196
 197 /* Store multibyte form of the character C in P.  The caller should
 198    allocate at least MAX_MULTIBYTE_LENGTH bytes area at P in advance.
 199    And, advance P to the end of the multibyte form.  */
 200
 201 #define CHAR_STRING_ADVANCE(c, p)               \
 202   do {                                          \
 203     if ((c) <= MAX_1_BYTE_CHAR)                 \
 204       *(p)++ = (c);                             \
 205     else if ((c) <= MAX_2_BYTE_CHAR)            \
 206       *(p)++ = (0xC0 | ((c) >> 6)),             \
 207         *(p)++ = (0x80 | ((c) & 0x3F));         \
 208     else if ((c) <= MAX_3_BYTE_CHAR)            \
 209       *(p)++ = (0xE0 | ((c) >> 12)),            \
 210         *(p)++ = (0x80 | (((c) >> 6) & 0x3F)),  \
 211         *(p)++ = (0x80 | ((c) & 0x3F));         \
 212     else                                        \
 213       (p) += char_string ((c), (p));            \
 214   } while (0)
 215
 216
 217 /* Nonzero iff BYTE starts a non-ASCII character in a multibyte
 218    form.  */
 219 #define LEADING_CODE_P(byte) (((byte) & 0xC0) == 0xC0)
 220
 221 /* Nonzero iff BYTE is a trailing code of a non-ASCII character in a
 222    multibyte form.  */
 223 #define TRAILING_CODE_P(byte) (((byte) & 0xC0) == 0x80)
 224
 225 /* Nonzero iff BYTE starts a character in a multibyte form.
 226    This is equivalent to:
 227         (ASCII_BYTE_P (byte) || LEADING_CODE_P (byte))  */
 228 #define CHAR_HEAD_P(byte) (((byte) & 0xC0) != 0x80)
 229
 230 /* Just kept for backward compatibility.  This macro will be removed
 231    in the future.  */
 232 #define BASE_LEADING_CODE_P LEADING_CODE_P
 233
 234 /* How many bytes a character that starts with BYTE occupies in a
 235    multibyte form.  */
 236 #define BYTES_BY_CHAR_HEAD(byte)        \
 237   (!((byte) & 0x80) ? 1                 \
 238    : !((byte) & 0x20) ? 2               \
 239    : !((byte) & 0x10) ? 3               \
 240    : !((byte) & 0x08) ? 4               \
 241    : 5)
 242
 243
 244 /* Return the length of the multi-byte form at string STR of length
 245    LEN while assuming that STR points a valid multi-byte form.  As
 246    this macro isn't necessary anymore, all callers will be changed to
 247    use BYTES_BY_CHAR_HEAD directly in the future.  */
 248
 249 #define MULTIBYTE_FORM_LENGTH(str, len)         \
 250   BYTES_BY_CHAR_HEAD (*(str))
 251
 252 /* Parse multibyte string STR of length LENGTH and set BYTES to the
 253    byte length of a character at STR while assuming that STR points a
 254    valid multibyte form.  As this macro isn't necessary anymore, all
 255    callers will be changed to use BYTES_BY_CHAR_HEAD directly in the
 256    future.  */
 257
 258 #define PARSE_MULTIBYTE_SEQ(str, length, bytes) \
 259   (bytes) = BYTES_BY_CHAR_HEAD (*(str))
 260
 261 /* The byte length of multibyte form at unibyte string P ending at
 262    PEND.  If STR doesn't point a valid multibyte form, return 0.  */
 263
 264 #define MULTIBYTE_LENGTH(p, pend)                               \
 265   (p >= pend ? 0                                                \
 266    : !((p)[0] & 0x80) ? 1                                       \
 267    : ((p + 1 >= pend) || (((p)[1] & 0xC0) != 0x80)) ? 0         \
 268    : ((p)[0] & 0xE0) == 0xC0 ? 2                                \
 269    : ((p + 2 >= pend) || (((p)[2] & 0xC0) != 0x80)) ? 0         \
 270    : ((p)[0] & 0xF0) == 0xE0 ? 3                                \
 271    : ((p + 3 >= pend) || (((p)[3] & 0xC0) != 0x80)) ? 0         \
 272    : ((p)[0] & 0xF8) == 0xF0 ? 4                                \
 273    : ((p + 4 >= pend) || (((p)[4] & 0xC0) != 0x80)) ? 0         \
 274    : (p)[0] == 0xF8 && ((p)[1] & 0xF0) == 0x80 ? 5              \
 275    : 0)
 276
 277
 278 /* Like MULTIBYTE_LENGTH but don't check the ending address.  */
 279
 280 #define MULTIBYTE_LENGTH_NO_CHECK(p)                    \
 281   (!((p)[0] & 0x80) ? 1                                 \
 282    : ((p)[1] & 0xC0) != 0x80 ? 0                        \
 283    : ((p)[0] & 0xE0) == 0xC0 ? 2                        \
 284    : ((p)[2] & 0xC0) != 0x80 ? 0                        \
 285    : ((p)[0] & 0xF0) == 0xE0 ? 3                        \
 286    : ((p)[3] & 0xC0) != 0x80 ? 0                        \
 287    : ((p)[0] & 0xF8) == 0xF0 ? 4                        \
 288    : ((p)[4] & 0xC0) != 0x80 ? 0                        \
 289    : (p)[0] == 0xF8 && ((p)[1] & 0xF0) == 0x80 ? 5      \
 290    : 0)
 291
 292 /* If P is before LIMIT, advance P to the next character boundary.  It
 293    assumes that P is already at a character boundary of the sane
 294    mulitbyte form whose end address is LIMIT.  */
 295
 296 #define NEXT_CHAR_BOUNDARY(p, limit)    \
 297   do {                                  \
 298     if ((p) < (limit))                  \
 299       (p) += BYTES_BY_CHAR_HEAD (*(p)); \
 300   } while (0)
 301
 302
 303 /* If P is after LIMIT, advance P to the previous character boundary.
 304    It assumes that P is already at a character boundary of the sane
 305    mulitbyte form whose beginning address is LIMIT.  */
 306
 307 #define PREV_CHAR_BOUNDARY(p, limit)                                    \
 308   do {                                                                  \
 309     if ((p) > (limit))                                                  \
 310       {                                                                 \
 311         const unsigned char *p0 = (p);                                  \
 312         do {                                                            \
 313           p0--;                                                         \
 314         } while (p0 >= limit && ! CHAR_HEAD_P (*p0));                   \
 315         (p) = (BYTES_BY_CHAR_HEAD (*p0) == (p) - p0) ? p0 : (p) - 1;    \
 316       }                                                                 \
 317   } while (0)
 318
 319 /* Return the character code of character whose multibyte form is at
 320    P.  The argument LEN is ignored.  It will be removed in the
 321    future.  */
 322
 323 #define STRING_CHAR(p, len)                                     \
 324   (!((p)[0] & 0x80)                                             \
 325    ? (p)[0]                                                     \
 326    : ! ((p)[0] & 0x20)                                          \
 327    ? (((((p)[0] & 0x1F) << 6)                                   \
 328        | ((p)[1] & 0x3F))                                       \
 329       + (((unsigned char) (p)[0]) < 0xC2 ? 0x3FFF80 : 0))       \
 330    : ! ((p)[0] & 0x10)                                          \
 331    ? ((((p)[0] & 0x0F) << 12)                                   \
 332       | (((p)[1] & 0x3F) << 6)                                  \
 333       | ((p)[2] & 0x3F))                                        \
 334    : string_char ((p), NULL, NULL))
 335
 336
 337 /* Like STRING_CHAR but set ACTUAL_LEN to the length of multibyte
 338    form.  The argument LEN is ignored.  It will be removed in the
 339    future.  */
 340
 341 #define STRING_CHAR_AND_LENGTH(p, len, actual_len)              \
 342   (!((p)[0] & 0x80)                                             \
 343    ? ((actual_len) = 1, (p)[0])                                 \
 344    : ! ((p)[0] & 0x20)                                          \
 345    ? ((actual_len) = 2,                                         \
 346       (((((p)[0] & 0x1F) << 6)                                  \
 347         | ((p)[1] & 0x3F))                                      \
 348        + (((unsigned char) (p)[0]) < 0xC2 ? 0x3FFF80 : 0)))     \
 349    : ! ((p)[0] & 0x10)                                          \
 350    ? ((actual_len) = 3,                                         \
 351       ((((p)[0] & 0x0F) << 12)                                  \
 352        | (((p)[1] & 0x3F) << 6)                                 \
 353        | ((p)[2] & 0x3F)))                                      \
 354    : string_char ((p), NULL, &actual_len))
 355
 356
 357 /* Like STRING_CHAR but advance P to the end of multibyte form.  */
 358
 359 #define STRING_CHAR_ADVANCE(p)                                  \
 360   (!((p)[0] & 0x80)                                             \
 361    ? *(p)++                                                     \
 362    : ! ((p)[0] & 0x20)                                          \
 363    ? ((p) += 2,                                                 \
 364       ((((p)[-2] & 0x1F) << 6)                                  \
 365        | ((p)[-1] & 0x3F)                                       \
 366        | ((unsigned char) ((p)[-2]) < 0xC2 ? 0x3FFF80 : 0)))    \
 367    : ! ((p)[0] & 0x10)                                          \
 368    ? ((p) += 3,                                                 \
 369       ((((p)[-3] & 0x0F) << 12)                                 \
 370        | (((p)[-2] & 0x3F) << 6)                                \
 371        | ((p)[-1] & 0x3F)))                                     \
 372    : string_char ((p), &(p), NULL))
 373
 374
 375 /* Fetch the "next" character from Lisp string STRING at byte position
 376    BYTEIDX, character position CHARIDX.  Store it into OUTPUT.
 377
 378    All the args must be side-effect-free.
 379    BYTEIDX and CHARIDX must be lvalues;
 380    we increment them past the character fetched.  */
 381
 382 #define FETCH_STRING_CHAR_ADVANCE(OUTPUT, STRING, CHARIDX, BYTEIDX)     \
 383   do                                                                    \
 384     {                                                                   \
 385       CHARIDX++;                                                        \
 386       if (STRING_MULTIBYTE (STRING))                                    \
 387         {                                                               \
 388           unsigned char *ptr = &SDATA (STRING)[BYTEIDX];                \
 389           int len;                                                      \
 390                                                                         \
 391           OUTPUT = STRING_CHAR_AND_LENGTH (ptr, 0, len);                \
 392           BYTEIDX += len;                                               \
 393         }                                                               \
 394       else                                                              \
 395         {                                                               \
 396           OUTPUT = SREF (STRING, BYTEIDX);                              \
 397           BYTEIDX++;                                                    \
 398         }                                                               \
 399     }                                                                   \
 400   while (0)
 401
 402 /* Like FETCH_STRING_CHAR_ADVANCE but return a multibyte character eve
 403    if STRING is unibyte.  */
 404
 405 #define FETCH_STRING_CHAR_AS_MULTIBYTE_ADVANCE(OUTPUT, STRING, CHARIDX, BYTEIDX) \
 406   do                                                                          \
 407     {                                                                         \
 408       CHARIDX++;                                                              \
 409       if (STRING_MULTIBYTE (STRING))                                          \
 410         {                                                                     \
 411           unsigned char *ptr = &SDATA (STRING)[BYTEIDX];                      \
 412           int len;                                                            \
 413                                                                               \
 414           OUTPUT = STRING_CHAR_AND_LENGTH (ptr, 0, len);                      \
 415           BYTEIDX += len;                                                     \
 416         }                                                                     \
 417       else                                                                    \
 418         {                                                                     \
 419           OUTPUT = SREF (STRING, BYTEIDX);                                    \
 420           BYTEIDX++;                                                          \
 421           MAKE_CHAR_MULTIBYTE (OUTPUT);                                       \
 422         }                                                                     \
 423     }                                                                         \
 424   while (0)
 425
 426
 427 /* Like FETCH_STRING_CHAR_ADVANCE but assumes STRING is multibyte.  */
 428
 429 #define FETCH_STRING_CHAR_ADVANCE_NO_CHECK(OUTPUT, STRING, CHARIDX, BYTEIDX) \
 430   do                                                                         \
 431     {                                                                        \
 432       unsigned char *ptr = &SDATA (STRING)[BYTEIDX];                         \
 433       int len;                                                               \
 434                                                                              \
 435       OUTPUT = STRING_CHAR_AND_LENGTH (ptr, 0, len);                         \
 436       BYTEIDX += len;                                                        \
 437       CHARIDX++;                                                             \
 438     }                                                                        \
 439   while (0)
 440
 441
 442 /* Like FETCH_STRING_CHAR_ADVANCE but fetch character from the current
 443    buffer.  */
 444
 445 #define FETCH_CHAR_ADVANCE(OUTPUT, CHARIDX, BYTEIDX)            \
 446   do                                                            \
 447     {                                                           \
 448       CHARIDX++;                                                \
 449       if (!NILP (current_buffer->enable_multibyte_characters))  \
 450         {                                                       \
 451           unsigned char *ptr = BYTE_POS_ADDR (BYTEIDX);         \
 452           int len;                                              \
 453                                                                 \
 454           OUTPUT= STRING_CHAR_AND_LENGTH (ptr, 0, len);         \
 455           BYTEIDX += len;                                       \
 456         }                                                       \
 457       else                                                      \
 458         {                                                       \
 459           OUTPUT = *(BYTE_POS_ADDR (BYTEIDX));                  \
 460           BYTEIDX++;                                            \
 461         }                                                       \
 462     }                                                           \
 463   while (0)
 464
 465
 466 /* Like FETCH_CHAR_ADVANCE but assumes the current buffer is multibyte.  */
 467
 468 #define FETCH_CHAR_ADVANCE_NO_CHECK(OUTPUT, CHARIDX, BYTEIDX)   \
 469   do                                                            \
 470     {                                                           \
 471       unsigned char *ptr = BYTE_POS_ADDR (BYTEIDX);             \
 472       int len;                                                  \
 473                                                                 \
 474       OUTPUT= STRING_CHAR_AND_LENGTH (ptr, 0, len);             \
 475       BYTEIDX += len;                                           \
 476       CHARIDX++;                                                \
 477     }                                                           \
 478   while (0)
 479
 480
 481 /* Increase the buffer byte position POS_BYTE of the current buffer to
 482    the next character boundary.  No range checking of POS.  */
 483
 484 #define INC_POS(pos_byte)                               \
 485   do {                                                  \
 486     unsigned char *p = BYTE_POS_ADDR (pos_byte);        \
 487     pos_byte += BYTES_BY_CHAR_HEAD (*p);                \
 488   } while (0)
 489
 490
 491 /* Decrease the buffer byte position POS_BYTE of the current buffer to
 492    the previous character boundary.  No range checking of POS.  */
 493
 494 #define DEC_POS(pos_byte)                       \
 495   do {                                          \
 496     unsigned char *p;                           \
 497                                                 \
 498     pos_byte--;                                 \
 499     if (pos_byte < GPT_BYTE)                    \
 500       p = BEG_ADDR + pos_byte - BEG_BYTE;       \
 501     else                                        \
 502       p = BEG_ADDR + GAP_SIZE + pos_byte - BEG_BYTE;\
 503     while (!CHAR_HEAD_P (*p))                   \
 504       {                                         \
 505         p--;                                    \
 506         pos_byte--;                             \
 507       }                                         \
 508   } while (0)
 509
 510 /* Increment both CHARPOS and BYTEPOS, each in the appropriate way.  */
 511
 512 #define INC_BOTH(charpos, bytepos)                              \
 513   do                                                            \
 514     {                                                           \
 515       (charpos)++;                                              \
 516       if (NILP (current_buffer->enable_multibyte_characters))   \
 517         (bytepos)++;                                            \
 518       else                                                      \
 519         INC_POS ((bytepos));                                    \
 520     }                                                           \
 521   while (0)
 522
 523
 524 /* Decrement both CHARPOS and BYTEPOS, each in the appropriate way.  */
 525
 526 #define DEC_BOTH(charpos, bytepos)                              \
 527   do                                                            \
 528     {                                                           \
 529       (charpos)--;                                              \
 530       if (NILP (current_buffer->enable_multibyte_characters))   \
 531         (bytepos)--;                                            \
 532       else                                                      \
 533         DEC_POS ((bytepos));                                    \
 534     }                                                           \
 535   while (0)
 536
 537
 538 /* Increase the buffer byte position POS_BYTE of the current buffer to
 539    the next character boundary.  This macro relies on the fact that
 540    *GPT_ADDR and *Z_ADDR are always accessible and the values are
 541    '\0'.  No range checking of POS_BYTE.  */
 542
 543 #define BUF_INC_POS(buf, pos_byte)                              \
 544   do {                                                          \
 545     unsigned char *p = BUF_BYTE_ADDRESS (buf, pos_byte);        \
 546     pos_byte += BYTES_BY_CHAR_HEAD (*p);                        \
 547   } while (0)
 548
 549
 550 /* Decrease the buffer byte position POS_BYTE of the current buffer to
 551    the previous character boundary.  No range checking of POS_BYTE.  */
 552
 553 #define BUF_DEC_POS(buf, pos_byte)                                      \
 554   do {                                                                  \
 555     unsigned char *p;                                                   \
 556     pos_byte--;                                                         \
 557     if (pos_byte < BUF_GPT_BYTE (buf))                                  \
 558       p = BUF_BEG_ADDR (buf) + pos_byte - BEG_BYTE;                     \
 559     else                                                                \
 560       p = BUF_BEG_ADDR (buf) + BUF_GAP_SIZE (buf) + pos_byte - BEG_BYTE;\
 561     while (!CHAR_HEAD_P (*p))                                           \
 562       {                                                                 \
 563         p--;                                                            \
 564         pos_byte--;                                                     \
 565       }                                                                 \
 566   } while (0)
 567
 568
 569 /* If C is a character to be unified with a Unicode character, return
 570    the unified Unicode character.  */
 571
 572 #define MAYBE_UNIFY_CHAR(c)                                     \
 573   if (c > MAX_UNICODE_CHAR                                      \
 574       && CHAR_TABLE_P (Vchar_unify_table))                      \
 575     {                                                           \
 576       Lisp_Object val;                                          \
 577       int unified;                                              \
 578                                                                 \
 579       val = CHAR_TABLE_REF (Vchar_unify_table, c);              \
 580       if (! NILP (val))                                         \
 581         {                                                       \
 582           if (SYMBOLP (val))                                    \
 583             {                                                   \
 584               Funify_charset (val, Qnil, Qnil);                 \
 585               val = CHAR_TABLE_REF (Vchar_unify_table, c);      \
 586             }                                                   \
 587           if ((unified = XINT (val)) >= 0)                      \
 588             c = unified;                                        \
 589         }                                                       \
 590     }                                                           \
 591   else
 592
 593
 594 /* Return the width of ASCII character C.  The width is measured by
 595    how many columns occupied on the screen when displayed in the
 596    current buffer.  */
 597
 598 #define ASCII_CHAR_WIDTH(c)                                             \
 599   (c < 0x20                                                             \
 600    ? (c == '\t'                                                         \
 601       ? XFASTINT (current_buffer->tab_width)                            \
 602       : (c == '\n' ? 0 : (NILP (current_buffer->ctl_arrow) ? 4 : 2)))   \
 603    : (c < 0x7f                                                          \
 604       ? 1                                                               \
 605       : ((NILP (current_buffer->ctl_arrow) ? 4 : 2))))
 606
 607 /* Return the width of character C.  The width is measured by how many
 608    columns occupied on the screen when displayed in the current
 609    buffer.  */
 610
 611 #define CHAR_WIDTH(c)           \
 612   (ASCII_CHAR_P (c)             \
 613    ? ASCII_CHAR_WIDTH (c)       \
 614    : XINT (CHAR_TABLE_REF (Vchar_width_table, c)))
 615
 616 extern int char_resolve_modifier_mask P_ ((int));
 617 extern int char_string P_ ((unsigned, unsigned char *));
 618 extern int string_char P_ ((const unsigned char *,
 619                             const unsigned char **, int *));
 620
 621 extern int translate_char P_ ((Lisp_Object, int c));
 622 extern int char_printable_p P_ ((int c));
 623 extern void parse_str_as_multibyte P_ ((const unsigned char *, int, int *,
 624                                         int *));
 625 extern int parse_str_to_multibyte P_ ((unsigned char *, int));
 626 extern int str_as_multibyte P_ ((unsigned char *, int, int, int *));
 627 extern int str_to_multibyte P_ ((unsigned char *, int, int));
 628 extern int str_as_unibyte P_ ((unsigned char *, int));
 629 extern int strwidth P_ ((unsigned char *, int));
 630 extern int c_string_width P_ ((const unsigned char *, int, int, int *, int *));
 631 extern int lisp_string_width P_ ((Lisp_Object, int, int *, int *));
 632
 633 extern Lisp_Object Vprintable_chars;
 634
 635 extern Lisp_Object Qcharacterp, Qauto_fill_chars;
 636 extern Lisp_Object Vtranslation_table_vector;
 637 extern Lisp_Object Vchar_width_table;
 638 extern Lisp_Object Vchar_direction_table;
 639 extern Lisp_Object Vchar_unify_table;
 640
 641 extern Lisp_Object string_escape_byte8 P_ ((Lisp_Object));
 642
 643 /* Return a translation table of id number ID.  */
 644 #define GET_TRANSLATION_TABLE(id) \
 645   (XCDR(XVECTOR(Vtranslation_table_vector)->contents[(id)]))
 646
 647 /* A char-table for characters which may invoke auto-filling.  */
 648 extern Lisp_Object Vauto_fill_chars;
 649
 650 extern Lisp_Object Vchar_script_table;
 651 extern Lisp_Object Vscript_representative_chars;
 652
 653 /* Copy LEN bytes from FROM to TO.  This macro should be used only
 654    when a caller knows that LEN is short and the obvious copy loop is
 655    faster than calling bcopy which has some overhead.  Copying a
 656    multibyte sequence of a character is the typical case.  */
 657
 658 #define BCOPY_SHORT(from, to, len)              \
 659   do {                                          \
 660     int i = len;                                \
 661     unsigned char *from_p = from, *to_p = to;   \
 662     while (i--) *to_p++ = *from_p++;            \
 663   } while (0)
 664
 665 #define DEFSYM(sym, name)       \
 666   do { (sym) = intern ((name)); staticpro (&(sym)); } while (0)
 667
 668 #endif /* EMACS_CHARACTER_H */
 669
 670 /* arch-tag: 4ef86004-2eff-4073-8cea-cfcbcf7188ac
 671    (do not change this comment) */