src/coding.h

   1 /* Header for coding system handler.
   2    Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
   3    Copyright (C) 1995, 1997, 1998, 2000
   4      National Institute of Advanced Industrial Science and Technology (AIST)
   5      Registration Number H14PRO021
   6
   7 This file is part of GNU Emacs.
   8
   9 GNU Emacs is free software; you can redistribute it and/or modify
  10 it under the terms of the GNU General Public License as published by
  11 the Free Software Foundation; either version 2, or (at your option)
  12 any later version.
  13
  14 GNU Emacs is distributed in the hope that it will be useful,
  15 but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17 GNU General Public License for more details.
  18
  19 You should have received a copy of the GNU General Public License
  20 along with GNU Emacs; see the file COPYING.  If not, write to
  21 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  22 Boston, MA 02110-1301, USA.  */
  23
  24 #ifndef EMACS_CODING_H
  25 #define EMACS_CODING_H
  26
  27 #include "ccl.h"
  28
  29 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  30
  31 /* All code (1-byte) of Emacs' internal format is classified into one
  32    of the followings.  See also `charset.h'.  */
  33 enum emacs_code_class_type
  34   {
  35     EMACS_control_code,         /* Control codes in the range
  36                                    0x00..0x1F and 0x7F except for the
  37                                    following two codes.  */
  38     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  39                                    end-of-line.  */
  40     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  41                                    in selective display mode.  */
  42     EMACS_ascii_code,           /* ASCII characters.  */
  43     EMACS_leading_code_2,       /* Base leading code of official
  44                                    TYPE9N character.  */
  45     EMACS_leading_code_3,       /* Base leading code of private TYPE9N
  46                                    or official TYPE9Nx9N character.  */
  47     EMACS_leading_code_4,       /* Base leading code of private
  48                                    TYPE9Nx9N character.  */
  49     EMACS_invalid_code          /* Invalid code, i.e. a base leading
  50                                    code not yet assigned to any
  51                                    charset, or a code of the range
  52                                    0xA0..0xFF.  */
  53   };
  54
  55 extern enum emacs_code_class_type emacs_code_class[256];
  56
  57 /*** ISO2022 section ***/
  58
  59 /* Macros to define code of control characters for ISO2022's functions.  */
  60                         /* code */      /* function */
  61 #define ISO_CODE_LF     0x0A            /* line-feed */
  62 #define ISO_CODE_CR     0x0D            /* carriage-return */
  63 #define ISO_CODE_SO     0x0E            /* shift-out */
  64 #define ISO_CODE_SI     0x0F            /* shift-in */
  65 #define ISO_CODE_SS2_7  0x19            /* single-shift-2 for 7-bit code */
  66 #define ISO_CODE_ESC    0x1B            /* escape */
  67 #define ISO_CODE_SS2    0x8E            /* single-shift-2 */
  68 #define ISO_CODE_SS3    0x8F            /* single-shift-3 */
  69 #define ISO_CODE_CSI    0x9B            /* control-sequence-introduce */
  70
  71 /* All code (1-byte) of ISO2022 is classified into one of the
  72    followings.  */
  73 enum iso_code_class_type
  74   {
  75     ISO_control_0,              /* Control codes in the range
  76                                    0x00..0x1F and 0x7F, except for the
  77                                    following 5 codes.  */
  78     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  79     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  80     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  81     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  82     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  83     ISO_control_1,              /* Control codes in the range
  84                                    0x80..0x9F, except for the
  85                                    following 3 codes.  */
  86     ISO_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  87     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  88     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  89     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  90     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  91     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  92     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  93   };
  94
  95 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  96   element in the structure `coding_system'.  This information is used
  97   while encoding a text to ISO2022.  **/
  98
  99 /* If set, produce short-form designation sequence (e.g. ESC $ A)
 100    instead of long-form sequence (e.g. ESC $ ( A).  */
 101 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 102
 103 /* If set, reset graphic planes and registers at end-of-line to the
 104    initial state.  */
 105 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 106
 107 /* If set, reset graphic planes and registers before any control
 108    characters to the initial state.  */
 109 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 110
 111 /* If set, encode by 7-bit environment.  */
 112 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 113
 114 /* If set, use locking-shift function.  */
 115 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 116
 117 /* If set, use single-shift function.  Overwrite
 118    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 119 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 120
 121 /* If set, designate JISX0201-Roman instead of ASCII.  */
 122 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 123
 124 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 125 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 126
 127 /* If set, do not produce ISO6429's direction specifying sequence.  */
 128 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 129
 130 /* If set, assume designation states are reset at beginning of line on
 131    output.  */
 132 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 133
 134 /* If set, designation sequence should be placed at beginning of line
 135    on output.  */
 136 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 137
 138 /* If set, do not encode unsafe characters on output.  */
 139 #define CODING_FLAG_ISO_SAFE            0x0800
 140
 141 /* If set, extra latin codes (128..159) are accepted as a valid code
 142    on input.  */
 143 #define CODING_FLAG_ISO_LATIN_EXTRA     0x1000
 144
 145 /* If set, use designation escape sequence.  */
 146 #define CODING_FLAG_ISO_DESIGNATION     0x10000
 147
 148 /* A character to be produced on output if encoding of the original
 149    character is inhibitted by CODING_MODE_INHIBIT_UNENCODABLE_CHAR.
 150    It must be an ASCII character.  */
 151 #define CODING_REPLACEMENT_CHARACTER '?'
 152
 153 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 154 struct iso2022_spec
 155 {
 156   /* The current graphic register invoked to each graphic plane.  */
 157   int current_invocation[2];
 158
 159   /* The current charset designated to each graphic register.  */
 160   int current_designation[4];
 161
 162   /* A charset initially designated to each graphic register.  */
 163   int initial_designation[4];
 164
 165   /* If not -1, it is a graphic register specified in an invalid
 166      designation sequence.  */
 167   int last_invalid_designation_register;
 168
 169   /* A graphic register to which each charset should be designated.  */
 170   unsigned char requested_designation[MAX_CHARSET + 1];
 171
 172   /* A revision number to be specified for each charset on encoding.
 173      The value 255 means no revision number for the corresponding
 174      charset.  */
 175   unsigned char charset_revision_number[MAX_CHARSET + 1];
 176
 177   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 178      by single-shift while encoding.  */
 179   int single_shifting;
 180
 181   /* Set to 1 temporarily only when processing at beginning of line.  */
 182   int bol;
 183 };
 184
 185 /* Macros to access each field in the structure `spec.iso2022'.  */
 186 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 187   (coding)->spec.iso2022.current_invocation[plane]
 188 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 189   (coding)->spec.iso2022.current_designation[reg]
 190 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 191   (coding)->spec.iso2022.initial_designation[reg]
 192 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 193   (coding)->spec.iso2022.requested_designation[charset]
 194 #define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
 195   (coding)->spec.iso2022.charset_revision_number[charset]
 196 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 197   (coding)->spec.iso2022.single_shifting
 198 #define CODING_SPEC_ISO_BOL(coding) \
 199   (coding)->spec.iso2022.bol
 200
 201 /* A value which may appear in
 202    coding->spec.iso2022.requested_designation indicating that the
 203    corresponding charset does not request any graphic register to be
 204    designated.  */
 205 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 206
 207 /* Return a charset which is currently designated to the graphic plane
 208    PLANE in the coding-system CODING.  */
 209 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 210   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 211    ? -1                                                 \
 212    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 213                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 214
 215 /*** BIG5 section ***/
 216
 217 /* Macros to denote each type of BIG5 coding system.  */
 218 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 219                                         BIG5 developed by Hong Kong
 220                                         University.  */
 221 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 222                                         of BIG5 developed by the
 223                                         company ETen in Taiwan.  */
 224
 225 /*** GENERAL section ***/
 226
 227 /* Types of coding system.  */
 228 enum coding_type
 229   {
 230     coding_type_no_conversion,  /* A coding system which requires no
 231                                    conversion for reading and writing
 232                                    including end-of-line format.  */
 233     coding_type_emacs_mule,     /* A coding system used in Emacs'
 234                                    buffer and string.  Requires no
 235                                    conversion for reading and writing
 236                                    except for end-of-line format.  */
 237     coding_type_undecided,      /* A coding system which requires
 238                                    automatic detection of a real
 239                                    coding system.  */
 240     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 241     coding_type_iso2022,        /* Any coding system of ISO2022
 242                                    variants.  */
 243     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 244     coding_type_ccl,            /* The coding system of which decoder
 245                                    and encoder are written in CCL.  */
 246     coding_type_raw_text        /* A coding system for a text
 247                                    containing random 8-bit code which
 248                                    does not require code conversion
 249                                    except for end-of-line format. */
 250   };
 251
 252 /* Formats of end-of-line.  */
 253 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 254                                    internal format.  */
 255 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 256                                    line-feed.  */
 257 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 258 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 259                                    eol-type is not yet decided.  */
 260 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 261                                      eol-type is not consistent
 262                                      through the file.  */
 263
 264 /* 1 iff composing.  */
 265 #define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
 266
 267 #define COMPOSITION_DATA_SIZE 4080
 268 #define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
 269
 270 /* Data structure to hold information about compositions of text that
 271    is being decoded or encode.  ISO 2022 base code conversion routines
 272    handle special ESC sequences for composition specification.  But,
 273    they can't get/put such information directly from/to a buffer in
 274    the deepest place.  So, they store or retrieve the information
 275    through this structure.
 276
 277    The encoder stores the information in this structure when it meets
 278    ESC sequences for composition while encoding codes, then, after all
 279    text codes are encoded, puts `composition' properties on the text
 280    by referring to the structure.
 281
 282    The decoder at first stores the information of a text to be
 283    decoded, then, while decoding codes, generates ESC sequences for
 284    composition at proper places by referring to the structure.  */
 285
 286 struct composition_data
 287 {
 288   /* The character position of the first character to be encoded or
 289      decoded.  START and END (see below) are relative to this
 290      position.  */
 291   int char_offset;
 292
 293   /* The composition data.  These elements are repeated for each
 294      composition:
 295         LENGTH START END METHOD [ COMPONENT ... ]
 296      where,
 297         LENGTH is the number of elements for this composition.
 298
 299         START and END are starting and ending character positions of
 300         the composition relative to `char_offset'.
 301
 302         METHOD is one of `enum composing_status' specifying the way of
 303         composition.
 304
 305         COMPONENT is a character or an encoded composition rule.  */
 306   int data[COMPOSITION_DATA_SIZE];
 307
 308   /* The number of elements in `data' currently used.  */
 309   int used;
 310
 311   /* Pointers to the previous and next structures.  When `data' is
 312      filled up, another structure is allocated and linked in `next'.
 313      The new structure has backward link to this structure in `prev'.
 314      The number of chained structures depends on how many compositions
 315      the text being encoded or decoded contains.  */
 316   struct composition_data *prev, *next;
 317 };
 318
 319 /* Macros used for the member `result' of the struct
 320    coding_system.  */
 321 #define CODING_FINISH_NORMAL            0
 322 #define CODING_FINISH_INSUFFICIENT_SRC  1
 323 #define CODING_FINISH_INSUFFICIENT_DST  2
 324 #define CODING_FINISH_INCONSISTENT_EOL  3
 325 #define CODING_FINISH_INSUFFICIENT_CMP  4
 326 #define CODING_FINISH_INTERRUPT         5
 327
 328 /* Macros used for the member `mode' of the struct coding_system.  */
 329
 330 /* If set, recover the original CR or LF of the already decoded text
 331    when the decoding routine encounters an inconsistent eol format.  */
 332 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 333
 334 /* If set, the decoding/encoding routines treat the current data as
 335    the last block of the whole text to be converted, and do
 336    appropriate finishing job.  */
 337 #define CODING_MODE_LAST_BLOCK                  0x02
 338
 339 /* If set, it means that the current source text is in a buffer which
 340    enables selective display.  */
 341 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 342
 343 /* If set, replace unencodabae characters by `?' on encoding.  */
 344 #define CODING_MODE_INHIBIT_UNENCODABLE_CHAR    0x08
 345
 346 /* This flag is used by the decoding/encoding routines on the fly.  If
 347    set, it means that right-to-left text is being processed.  */
 348 #define CODING_MODE_DIRECTION                   0x10
 349
 350 struct coding_system
 351 {
 352   /* Type of the coding system.  */
 353   enum coding_type type;
 354
 355   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 356   int eol_type;
 357
 358   /* Flag bits of the coding system.  The meaning of each bit is common
 359      to all types of coding systems.  */
 360   unsigned int common_flags;
 361
 362   /* Flag bits of the coding system.  The meaning of each bit depends
 363      on the type of the coding system.  */
 364   unsigned int flags;
 365
 366   /* Mode bits of the coding system.  See the comments of the macros
 367      CODING_MODE_XXX.  */
 368   unsigned int mode;
 369
 370   /* The current status of composition handling.  */
 371   int composing;
 372
 373   /* 1 iff the next character is a composition rule.  */
 374   int composition_rule_follows;
 375
 376   /* Information of compositions are stored here on decoding and set
 377      in advance on encoding.  */
 378   struct composition_data *cmp_data;
 379
 380   /* Index to cmp_data->data for the first element for the current
 381      composition.  */
 382   int cmp_data_start;
 383
 384   /* Index to cmp_data->data for the current element for the current
 385      composition.  */
 386   int cmp_data_index;
 387
 388   /* Detailed information specific to each type of coding system.  */
 389   union spec
 390     {
 391       struct iso2022_spec iso2022;
 392       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 393     } spec;
 394
 395   /* Index number of coding category of the coding system.  */
 396   int category_idx;
 397
 398   /* The following two members specify how characters 128..159 are
 399      represented in source and destination text respectively.  1 means
 400      they are represented by 2-byte sequence, 0 means they are
 401      represented by 1-byte as is (see the comment in charset.h).  */
 402   unsigned src_multibyte : 1;
 403   unsigned dst_multibyte : 1;
 404
 405   /* How may heading bytes we can skip for decoding.  This is set to
 406      -1 in setup_coding_system, and updated by detect_coding.  So,
 407      when this is equal to the byte length of the text being
 408      converted, we can skip the actual conversion process.  */
 409   int heading_ascii;
 410
 411   /* The following members are set by encoding/decoding routine.  */
 412   int produced, produced_char, consumed, consumed_char;
 413
 414   /* Number of error source data found in a decoding routine.  */
 415   int errors;
 416
 417   /* Finish status of code conversion.  It should be one of macros
 418      CODING_FINISH_XXXX.  */
 419   int result;
 420
 421   /* If nonzero, suppress error notification.  */
 422   int suppress_error;
 423
 424   /* The following members are all Lisp symbols.  We don't have to
 425      protect them from GC because the current garbage collection
 426      doesn't relocate Lisp symbols.  But, when it is changed, we must
 427      find a way to protect them.  */
 428
 429   /* Backward pointer to the Lisp symbol of the coding system.  */
 430   Lisp_Object symbol;
 431
 432   /* Lisp function (symbol) to be called after decoding to do
 433      additional conversion, or nil.  */
 434   Lisp_Object post_read_conversion;
 435
 436   /* Lisp function (symbol) to be called before encoding to do
 437      additional conversion, or nil.  */
 438   Lisp_Object pre_write_conversion;
 439
 440   /* Character translation tables to look up, or nil.  */
 441   Lisp_Object translation_table_for_decode;
 442   Lisp_Object translation_table_for_encode;
 443 };
 444
 445 #define CODING_REQUIRE_FLUSHING_MASK    1
 446 #define CODING_REQUIRE_DECODING_MASK    2
 447 #define CODING_REQUIRE_ENCODING_MASK    4
 448 #define CODING_REQUIRE_DETECTION_MASK   8
 449
 450 /* Return 1 if the coding system CODING requires specific code to be
 451    attached at the tail of converted text.  */
 452 #define CODING_REQUIRE_FLUSHING(coding) \
 453   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 454
 455 /* Return 1 if the coding system CODING requires code conversion on
 456    decoding.  */
 457 #define CODING_REQUIRE_DECODING(coding) \
 458   ((coding)->dst_multibyte              \
 459    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 460
 461 /* Return 1 if the coding system CODING requires code conversion on
 462    encoding.  */
 463 #define CODING_REQUIRE_ENCODING(coding) \
 464   ((coding)->src_multibyte              \
 465    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 466
 467 /* Return 1 if the coding system CODING requires some kind of code
 468    detection.  */
 469 #define CODING_REQUIRE_DETECTION(coding) \
 470   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 471
 472 /* Return 1 if the coding system CODING requires code conversion on
 473    decoding or some kind of code detection.  */
 474 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 475   (CODING_REQUIRE_DECODING (coding)             \
 476    || CODING_REQUIRE_DETECTION (coding))
 477
 478 /* Index for each coding category in `coding_category_table' */
 479 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 480 #define CODING_CATEGORY_IDX_SJIS        1
 481 #define CODING_CATEGORY_IDX_ISO_7       2
 482 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 483 #define CODING_CATEGORY_IDX_ISO_8_1     4
 484 #define CODING_CATEGORY_IDX_ISO_8_2     5
 485 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 486 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 487 #define CODING_CATEGORY_IDX_CCL         8
 488 #define CODING_CATEGORY_IDX_BIG5        9
 489 #define CODING_CATEGORY_IDX_UTF_8       10
 490 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 491 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 492 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 493 #define CODING_CATEGORY_IDX_BINARY      14
 494 #define CODING_CATEGORY_IDX_MAX         15
 495
 496 /* Definitions of flag bits returned by the function
 497    detect_coding_mask ().  */
 498 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 499 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 500 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 501 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 502 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 503 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 504 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 505 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 506 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 507 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 508 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 509 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 510 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 511 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 512 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 513
 514 /* This value is returned if detect_coding_mask () find nothing other
 515    than ASCII characters.  */
 516 #define CODING_CATEGORY_MASK_ANY        \
 517   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 518    | CODING_CATEGORY_MASK_SJIS          \
 519    | CODING_CATEGORY_MASK_ISO_7         \
 520    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 521    | CODING_CATEGORY_MASK_ISO_8_1       \
 522    | CODING_CATEGORY_MASK_ISO_8_2       \
 523    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 524    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 525    | CODING_CATEGORY_MASK_CCL           \
 526    | CODING_CATEGORY_MASK_BIG5          \
 527    | CODING_CATEGORY_MASK_UTF_8         \
 528    | CODING_CATEGORY_MASK_UTF_16_BE     \
 529    | CODING_CATEGORY_MASK_UTF_16_LE)
 530
 531 #define CODING_CATEGORY_MASK_ISO_7BIT \
 532   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 533
 534 #define CODING_CATEGORY_MASK_ISO_8BIT \
 535   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 536
 537 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 538   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 539
 540 #define CODING_CATEGORY_MASK_ISO        \
 541   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 542    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 543    | CODING_CATEGORY_MASK_ISO_8BIT)
 544
 545 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 546    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 547
 548 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 549    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 550    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 551    internal format.  */
 552
 553 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 554   do {                                                  \
 555     if (s2 >= 0x9F)                                     \
 556       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 557       c2 = s2 - 0x7E;                                   \
 558     else                                                \
 559       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 560       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 561   } while (0)
 562
 563 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 564   do {                                                  \
 565     if (c1 & 1)                                         \
 566       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 567       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 568     else                                                \
 569       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 570       s2 = c2 + 0x7E;                                   \
 571   } while (0)
 572
 573 /* Encode the file name NAME using the specified coding system
 574    for file names, if any.  */
 575 #define ENCODE_FILE(name)                                                  \
 576   (! NILP (Vfile_name_coding_system)                                       \
 577    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 578    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 579    : (! NILP (Vdefault_file_name_coding_system)                            \
 580       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 581       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 582       : name))
 583
 584 /* Decode the file name NAME using the specified coding system
 585    for file names, if any.  */
 586 #define DECODE_FILE(name)                                                  \
 587   (! NILP (Vfile_name_coding_system)                                       \
 588    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 589    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 590    : (! NILP (Vdefault_file_name_coding_system)                            \
 591       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 592       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 593       : name))
 594
 595 /* Encode the string STR using the specified coding system
 596    for system functions, if any.  */
 597 #define ENCODE_SYSTEM(str)                                                 \
 598   (! NILP (Vlocale_coding_system)                                          \
 599    && !EQ (Vlocale_coding_system, make_number (0))                         \
 600    ? code_convert_string_norecord (str, Vlocale_coding_system, 1)          \
 601    : str)
 602
 603 /* Decode the string STR using the specified coding system
 604    for system functions, if any.  */
 605 #define DECODE_SYSTEM(str)                                                 \
 606   (! NILP (Vlocale_coding_system)                                          \
 607    && !EQ (Vlocale_coding_system, make_number (0))                         \
 608    ? code_convert_string_norecord (str, Vlocale_coding_system, 0)          \
 609    : str)
 610
 611 #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1)
 612
 613 /* Extern declarations.  */
 614 extern int decode_coding P_ ((struct coding_system *, const unsigned char *,
 615                               unsigned char *, int, int));
 616 extern int encode_coding P_ ((struct coding_system *, const unsigned char *,
 617                               unsigned char *, int, int));
 618 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 619                                          Lisp_Object));
 620 extern void coding_free_composition_data P_ ((struct coding_system *));
 621 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 622                                                   int));
 623 extern void coding_allocate_composition_data P_ ((struct coding_system *,
 624                                                   int));
 625 extern void coding_restore_composition P_ ((struct coding_system *,
 626                                             Lisp_Object));
 627 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 628                                     int, int));
 629 extern Lisp_Object run_pre_post_conversion_on_str P_ ((Lisp_Object,
 630                                                        struct coding_system *,
 631                                                        int));
 632 extern void run_pre_write_conversin_on_c_str P_ ((unsigned char **, int *,
 633                                                   int, int,
 634                                                   struct coding_system *));
 635
 636 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 637 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 638 extern void detect_coding P_ ((struct coding_system *, const unsigned char *,
 639                                int));
 640 extern void detect_eol P_ ((struct coding_system *, const unsigned char *,
 641                             int));
 642 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 643 extern Lisp_Object code_convert_string P_ ((Lisp_Object,
 644                                             struct coding_system *, int, int));
 645 extern Lisp_Object code_convert_string1 P_ ((Lisp_Object, Lisp_Object,
 646                                              Lisp_Object, int));
 647 extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object,
 648                                                      int));
 649 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 650 extern Lisp_Object encode_coding_string P_ ((Lisp_Object,
 651                                              struct coding_system *, int));
 652 extern Lisp_Object decode_coding_string P_ ((Lisp_Object,
 653                                              struct coding_system *, int));
 654 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 655 extern Lisp_Object Qraw_text, Qemacs_mule;
 656 extern Lisp_Object Qbuffer_file_coding_system;
 657 extern Lisp_Object Vcoding_category_list;
 658 extern Lisp_Object Qutf_8;
 659
 660 extern Lisp_Object Qtranslation_table;
 661 extern Lisp_Object Qtranslation_table_id;
 662
 663 /* Mnemonic strings to indicate each type of end-of-line.  */
 664 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 665 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 666 extern Lisp_Object eol_mnemonic_undecided;
 667
 668 #ifdef emacs
 669 extern Lisp_Object Qfile_coding_system;
 670 extern Lisp_Object Qcall_process, Qcall_process_region;
 671 extern Lisp_Object Qstart_process, Qopen_network_stream;
 672 extern Lisp_Object Qwrite_region;
 673
 674 extern char *emacs_strerror P_ ((int));
 675
 676 /* Coding-system for reading files and receiving data from process.  */
 677 extern Lisp_Object Vcoding_system_for_read;
 678 /* Coding-system for writing files and sending data to process.  */
 679 extern Lisp_Object Vcoding_system_for_write;
 680 /* Coding-system actually used in the latest I/O.  */
 681 extern Lisp_Object Vlast_coding_system_used;
 682 /* Coding-system to use with system messages (e.g. strerror).  */
 683 extern Lisp_Object Vlocale_coding_system;
 684
 685 /* If non-zero, process buffer inherits the coding system used to decode
 686    the subprocess output.  */
 687 extern int inherit_process_coding_system;
 688
 689 /* Coding-system to be used for encoding terminal output.  This
 690    structure contains information of a coding-system specified by the
 691    function `set-terminal-coding-system'.  */
 692 extern struct coding_system terminal_coding;
 693
 694 /* Coding system to be used to encode text for terminal display when
 695    terminal coding system is nil.  */
 696 extern struct coding_system safe_terminal_coding;
 697
 698 /* Coding-system of what is sent from terminal keyboard.  This
 699    structure contains information of a coding-system specified by the
 700    function `set-keyboard-coding-system'.  */
 701 extern struct coding_system keyboard_coding;
 702
 703 /* Default coding system to be used to write a file.  */
 704 extern struct coding_system default_buffer_file_coding;
 705
 706 /* Default coding systems used for process I/O.  */
 707 extern Lisp_Object Vdefault_process_coding_system;
 708
 709 /* Function to call to force a user to force select a proper coding
 710    system.  */
 711 extern Lisp_Object Vselect_safe_coding_system_function;
 712
 713 /* If nonzero, on writing a file, Vselect_safe_coding_system_function
 714    is called even if Vcoding_system_for_write is non-nil.  */
 715 extern int coding_system_require_warning;
 716
 717 /* Coding system for file names, or nil if none.  */
 718 extern Lisp_Object Vfile_name_coding_system;
 719
 720 /* Coding system for file names used only when
 721    Vfile_name_coding_system is nil.  */
 722 extern Lisp_Object Vdefault_file_name_coding_system;
 723
 724 #endif
 725
 726 /* Error signaled when there's a problem with detecting coding system */
 727 extern Lisp_Object Qcoding_system_error;
 728
 729 #endif /* EMACS_CODING_H */
 730
 731 /* arch-tag: 2bc3b4fa-6870-4f64-8135-b962b2d290e4
 732    (do not change this comment) */