src/coding.h

   1 /* Header for coding system handler.
   2    Copyright (C) 2002, 2003, 2004, 2005,
   3                  2006 Free Software Foundation, Inc.
   4    Copyright (C) 1995, 1997, 1998, 2000
   5      National Institute of Advanced Industrial Science and Technology (AIST)
   6      Registration Number H14PRO021
   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 2, 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
  22 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  23 Boston, MA 02110-1301, USA.  */
  24
  25 #ifndef EMACS_CODING_H
  26 #define EMACS_CODING_H
  27
  28 #include "ccl.h"
  29
  30 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  31
  32 /* All code (1-byte) of Emacs' internal format is classified into one
  33    of the followings.  See also `charset.h'.  */
  34 enum emacs_code_class_type
  35   {
  36     EMACS_control_code,         /* Control codes in the range
  37                                    0x00..0x1F and 0x7F except for the
  38                                    following two codes.  */
  39     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  40                                    end-of-line.  */
  41     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  42                                    in selective display mode.  */
  43     EMACS_ascii_code,           /* ASCII characters.  */
  44     EMACS_leading_code_2,       /* Base leading code of official
  45                                    TYPE9N character.  */
  46     EMACS_leading_code_3,       /* Base leading code of private TYPE9N
  47                                    or official TYPE9Nx9N character.  */
  48     EMACS_leading_code_4,       /* Base leading code of private
  49                                    TYPE9Nx9N character.  */
  50     EMACS_invalid_code          /* Invalid code, i.e. a base leading
  51                                    code not yet assigned to any
  52                                    charset, or a code of the range
  53                                    0xA0..0xFF.  */
  54   };
  55
  56 extern enum emacs_code_class_type emacs_code_class[256];
  57
  58 /*** ISO2022 section ***/
  59
  60 /* Macros to define code of control characters for ISO2022's functions.  */
  61                         /* code */      /* function */
  62 #define ISO_CODE_LF     0x0A            /* line-feed */
  63 #define ISO_CODE_CR     0x0D            /* carriage-return */
  64 #define ISO_CODE_SO     0x0E            /* shift-out */
  65 #define ISO_CODE_SI     0x0F            /* shift-in */
  66 #define ISO_CODE_SS2_7  0x19            /* single-shift-2 for 7-bit code */
  67 #define ISO_CODE_ESC    0x1B            /* escape */
  68 #define ISO_CODE_SS2    0x8E            /* single-shift-2 */
  69 #define ISO_CODE_SS3    0x8F            /* single-shift-3 */
  70 #define ISO_CODE_CSI    0x9B            /* control-sequence-introduce */
  71
  72 /* All code (1-byte) of ISO2022 is classified into one of the
  73    followings.  */
  74 enum iso_code_class_type
  75   {
  76     ISO_control_0,              /* Control codes in the range
  77                                    0x00..0x1F and 0x7F, except for the
  78                                    following 5 codes.  */
  79     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  80     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  81     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  82     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  83     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  84     ISO_control_1,              /* Control codes in the range
  85                                    0x80..0x9F, except for the
  86                                    following 3 codes.  */
  87     ISO_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  88     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  89     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  90     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  91     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  92     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  93     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  94   };
  95
  96 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  97   element in the structure `coding_system'.  This information is used
  98   while encoding a text to ISO2022.  **/
  99
 100 /* If set, produce short-form designation sequence (e.g. ESC $ A)
 101    instead of long-form sequence (e.g. ESC $ ( A).  */
 102 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 103
 104 /* If set, reset graphic planes and registers at end-of-line to the
 105    initial state.  */
 106 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 107
 108 /* If set, reset graphic planes and registers before any control
 109    characters to the initial state.  */
 110 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 111
 112 /* If set, encode by 7-bit environment.  */
 113 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 114
 115 /* If set, use locking-shift function.  */
 116 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 117
 118 /* If set, use single-shift function.  Overwrite
 119    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 120 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 121
 122 /* If set, designate JISX0201-Roman instead of ASCII.  */
 123 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 124
 125 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 126 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 127
 128 /* If set, do not produce ISO6429's direction specifying sequence.  */
 129 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 130
 131 /* If set, assume designation states are reset at beginning of line on
 132    output.  */
 133 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 134
 135 /* If set, designation sequence should be placed at beginning of line
 136    on output.  */
 137 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 138
 139 /* If set, do not encode unsafe characters on output.  */
 140 #define CODING_FLAG_ISO_SAFE            0x0800
 141
 142 /* If set, extra latin codes (128..159) are accepted as a valid code
 143    on input.  */
 144 #define CODING_FLAG_ISO_LATIN_EXTRA     0x1000
 145
 146 /* If set, use designation escape sequence.  */
 147 #define CODING_FLAG_ISO_DESIGNATION     0x10000
 148
 149 /* A character to be produced on output if encoding of the original
 150    character is inhibitted by CODING_MODE_INHIBIT_UNENCODABLE_CHAR.
 151    It must be an ASCII character.  */
 152 #define CODING_REPLACEMENT_CHARACTER '?'
 153
 154 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 155 struct iso2022_spec
 156 {
 157   /* The current graphic register invoked to each graphic plane.  */
 158   int current_invocation[2];
 159
 160   /* The current charset designated to each graphic register.  */
 161   int current_designation[4];
 162
 163   /* A charset initially designated to each graphic register.  */
 164   int initial_designation[4];
 165
 166   /* If not -1, it is a graphic register specified in an invalid
 167      designation sequence.  */
 168   int last_invalid_designation_register;
 169
 170   /* A graphic register to which each charset should be designated.  */
 171   unsigned char requested_designation[MAX_CHARSET + 1];
 172
 173   /* A revision number to be specified for each charset on encoding.
 174      The value 255 means no revision number for the corresponding
 175      charset.  */
 176   unsigned char charset_revision_number[MAX_CHARSET + 1];
 177
 178   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 179      by single-shift while encoding.  */
 180   int single_shifting;
 181
 182   /* Set to 1 temporarily only when processing at beginning of line.  */
 183   int bol;
 184 };
 185
 186 /* Macros to access each field in the structure `spec.iso2022'.  */
 187 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 188   (coding)->spec.iso2022.current_invocation[plane]
 189 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 190   (coding)->spec.iso2022.current_designation[reg]
 191 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 192   (coding)->spec.iso2022.initial_designation[reg]
 193 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 194   (coding)->spec.iso2022.requested_designation[charset]
 195 #define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
 196   (coding)->spec.iso2022.charset_revision_number[charset]
 197 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 198   (coding)->spec.iso2022.single_shifting
 199 #define CODING_SPEC_ISO_BOL(coding) \
 200   (coding)->spec.iso2022.bol
 201
 202 /* A value which may appear in
 203    coding->spec.iso2022.requested_designation indicating that the
 204    corresponding charset does not request any graphic register to be
 205    designated.  */
 206 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 207
 208 /* Return a charset which is currently designated to the graphic plane
 209    PLANE in the coding-system CODING.  */
 210 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 211   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 212    ? -1                                                 \
 213    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 214                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 215
 216 /*** BIG5 section ***/
 217
 218 /* Macros to denote each type of BIG5 coding system.  */
 219 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 220                                         BIG5 developed by Hong Kong
 221                                         University.  */
 222 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 223                                         of BIG5 developed by the
 224                                         company ETen in Taiwan.  */
 225
 226 /*** GENERAL section ***/
 227
 228 /* Types of coding system.  */
 229 enum coding_type
 230   {
 231     coding_type_no_conversion,  /* A coding system which requires no
 232                                    conversion for reading and writing
 233                                    including end-of-line format.  */
 234     coding_type_emacs_mule,     /* A coding system used in Emacs'
 235                                    buffer and string.  Requires no
 236                                    conversion for reading and writing
 237                                    except for end-of-line format.  */
 238     coding_type_undecided,      /* A coding system which requires
 239                                    automatic detection of a real
 240                                    coding system.  */
 241     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 242     coding_type_iso2022,        /* Any coding system of ISO2022
 243                                    variants.  */
 244     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 245     coding_type_ccl,            /* The coding system of which decoder
 246                                    and encoder are written in CCL.  */
 247     coding_type_raw_text        /* A coding system for a text
 248                                    containing random 8-bit code which
 249                                    does not require code conversion
 250                                    except for end-of-line format. */
 251   };
 252
 253 /* Formats of end-of-line.  */
 254 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 255                                    internal format.  */
 256 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 257                                    line-feed.  */
 258 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 259 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 260                                    eol-type is not yet decided.  */
 261 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 262                                      eol-type is not consistent
 263                                      through the file.  */
 264
 265 /* 1 iff composing.  */
 266 #define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
 267
 268 #define COMPOSITION_DATA_SIZE 4080
 269 #define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
 270
 271 /* Data structure to hold information about compositions of text that
 272    is being decoded or encode.  ISO 2022 base code conversion routines
 273    handle special ESC sequences for composition specification.  But,
 274    they can't get/put such information directly from/to a buffer in
 275    the deepest place.  So, they store or retrieve the information
 276    through this structure.
 277
 278    The encoder stores the information in this structure when it meets
 279    ESC sequences for composition while encoding codes, then, after all
 280    text codes are encoded, puts `composition' properties on the text
 281    by referring to the structure.
 282
 283    The decoder at first stores the information of a text to be
 284    decoded, then, while decoding codes, generates ESC sequences for
 285    composition at proper places by referring to the structure.  */
 286
 287 struct composition_data
 288 {
 289   /* The character position of the first character to be encoded or
 290      decoded.  START and END (see below) are relative to this
 291      position.  */
 292   int char_offset;
 293
 294   /* The composition data.  These elements are repeated for each
 295      composition:
 296         LENGTH START END METHOD [ COMPONENT ... ]
 297      where,
 298         LENGTH is the number of elements for this composition.
 299
 300         START and END are starting and ending character positions of
 301         the composition relative to `char_offset'.
 302
 303         METHOD is one of `enum composing_status' specifying the way of
 304         composition.
 305
 306         COMPONENT is a character or an encoded composition rule.  */
 307   int data[COMPOSITION_DATA_SIZE];
 308
 309   /* The number of elements in `data' currently used.  */
 310   int used;
 311
 312   /* Pointers to the previous and next structures.  When `data' is
 313      filled up, another structure is allocated and linked in `next'.
 314      The new structure has backward link to this structure in `prev'.
 315      The number of chained structures depends on how many compositions
 316      the text being encoded or decoded contains.  */
 317   struct composition_data *prev, *next;
 318 };
 319
 320 /* Macros used for the member `result' of the struct
 321    coding_system.  */
 322 #define CODING_FINISH_NORMAL            0
 323 #define CODING_FINISH_INSUFFICIENT_SRC  1
 324 #define CODING_FINISH_INSUFFICIENT_DST  2
 325 #define CODING_FINISH_INCONSISTENT_EOL  3
 326 #define CODING_FINISH_INSUFFICIENT_CMP  4
 327 #define CODING_FINISH_INTERRUPT         5
 328
 329 /* Macros used for the member `mode' of the struct coding_system.  */
 330
 331 /* If set, recover the original CR or LF of the already decoded text
 332    when the decoding routine encounters an inconsistent eol format.  */
 333 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 334
 335 /* If set, the decoding/encoding routines treat the current data as
 336    the last block of the whole text to be converted, and do
 337    appropriate finishing job.  */
 338 #define CODING_MODE_LAST_BLOCK                  0x02
 339
 340 /* If set, it means that the current source text is in a buffer which
 341    enables selective display.  */
 342 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 343
 344 /* If set, replace unencodabae characters by `?' on encoding.  */
 345 #define CODING_MODE_INHIBIT_UNENCODABLE_CHAR    0x08
 346
 347 /* This flag is used by the decoding/encoding routines on the fly.  If
 348    set, it means that right-to-left text is being processed.  */
 349 #define CODING_MODE_DIRECTION                   0x10
 350
 351 struct coding_system
 352 {
 353   /* Type of the coding system.  */
 354   enum coding_type type;
 355
 356   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 357   int eol_type;
 358
 359   /* Flag bits of the coding system.  The meaning of each bit is common
 360      to all types of coding systems.  */
 361   unsigned int common_flags;
 362
 363   /* Flag bits of the coding system.  The meaning of each bit depends
 364      on the type of the coding system.  */
 365   unsigned int flags;
 366
 367   /* Mode bits of the coding system.  See the comments of the macros
 368      CODING_MODE_XXX.  */
 369   unsigned int mode;
 370
 371   /* The current status of composition handling.  */
 372   int composing;
 373
 374   /* 1 iff the next character is a composition rule.  */
 375   int composition_rule_follows;
 376
 377   /* Information of compositions are stored here on decoding and set
 378      in advance on encoding.  */
 379   struct composition_data *cmp_data;
 380
 381   /* Index to cmp_data->data for the first element for the current
 382      composition.  */
 383   int cmp_data_start;
 384
 385   /* Index to cmp_data->data for the current element for the current
 386      composition.  */
 387   int cmp_data_index;
 388
 389   /* Detailed information specific to each type of coding system.  */
 390   union spec
 391     {
 392       struct iso2022_spec iso2022;
 393       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 394     } spec;
 395
 396   /* Index number of coding category of the coding system.  */
 397   int category_idx;
 398
 399   /* The following two members specify how characters 128..159 are
 400      represented in source and destination text respectively.  1 means
 401      they are represented by 2-byte sequence, 0 means they are
 402      represented by 1-byte as is (see the comment in charset.h).  */
 403   unsigned src_multibyte : 1;
 404   unsigned dst_multibyte : 1;
 405
 406   /* How may heading bytes we can skip for decoding.  This is set to
 407      -1 in setup_coding_system, and updated by detect_coding.  So,
 408      when this is equal to the byte length of the text being
 409      converted, we can skip the actual conversion process.  */
 410   int heading_ascii;
 411
 412   /* The following members are set by encoding/decoding routine.  */
 413   int produced, produced_char, consumed, consumed_char;
 414
 415   /* Number of error source data found in a decoding routine.  */
 416   int errors;
 417
 418   /* Finish status of code conversion.  It should be one of macros
 419      CODING_FINISH_XXXX.  */
 420   int result;
 421
 422   /* If nonzero, suppress error notification.  */
 423   int suppress_error;
 424
 425   /* The following members are all Lisp symbols.  We don't have to
 426      protect them from GC because the current garbage collection
 427      doesn't relocate Lisp symbols.  But, when it is changed, we must
 428      find a way to protect them.  */
 429
 430   /* Backward pointer to the Lisp symbol of the coding system.  */
 431   Lisp_Object symbol;
 432
 433   /* Lisp function (symbol) to be called after decoding to do
 434      additional conversion, or nil.  */
 435   Lisp_Object post_read_conversion;
 436
 437   /* Lisp function (symbol) to be called before encoding to do
 438      additional conversion, or nil.  */
 439   Lisp_Object pre_write_conversion;
 440
 441   /* Character translation tables to look up, or nil.  */
 442   Lisp_Object translation_table_for_decode;
 443   Lisp_Object translation_table_for_encode;
 444 };
 445
 446 #define CODING_REQUIRE_FLUSHING_MASK    1
 447 #define CODING_REQUIRE_DECODING_MASK    2
 448 #define CODING_REQUIRE_ENCODING_MASK    4
 449 #define CODING_REQUIRE_DETECTION_MASK   8
 450
 451 /* Return 1 if the coding system CODING requires specific code to be
 452    attached at the tail of converted text.  */
 453 #define CODING_REQUIRE_FLUSHING(coding) \
 454   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 455
 456 /* Return 1 if the coding system CODING requires code conversion on
 457    decoding.  */
 458 #define CODING_REQUIRE_DECODING(coding) \
 459   ((coding)->dst_multibyte              \
 460    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 461
 462 /* Return 1 if the coding system CODING requires code conversion on
 463    encoding.  */
 464 #define CODING_REQUIRE_ENCODING(coding) \
 465   ((coding)->src_multibyte              \
 466    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 467
 468 /* Return 1 if the coding system CODING requires some kind of code
 469    detection.  */
 470 #define CODING_REQUIRE_DETECTION(coding) \
 471   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 472
 473 /* Return 1 if the coding system CODING requires code conversion on
 474    decoding or some kind of code detection.  */
 475 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 476   (CODING_REQUIRE_DECODING (coding)             \
 477    || CODING_REQUIRE_DETECTION (coding))
 478
 479 /* Index for each coding category in `coding_category_table' */
 480 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 481 #define CODING_CATEGORY_IDX_SJIS        1
 482 #define CODING_CATEGORY_IDX_ISO_7       2
 483 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 484 #define CODING_CATEGORY_IDX_ISO_8_1     4
 485 #define CODING_CATEGORY_IDX_ISO_8_2     5
 486 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 487 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 488 #define CODING_CATEGORY_IDX_CCL         8
 489 #define CODING_CATEGORY_IDX_BIG5        9
 490 #define CODING_CATEGORY_IDX_UTF_8       10
 491 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 492 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 493 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 494 #define CODING_CATEGORY_IDX_BINARY      14
 495 #define CODING_CATEGORY_IDX_MAX         15
 496
 497 /* Definitions of flag bits returned by the function
 498    detect_coding_mask ().  */
 499 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 500 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 501 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 502 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 503 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 504 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 505 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 506 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 507 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 508 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 509 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 510 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 511 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 512 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 513 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 514
 515 /* This value is returned if detect_coding_mask () find nothing other
 516    than ASCII characters.  */
 517 #define CODING_CATEGORY_MASK_ANY        \
 518   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 519    | CODING_CATEGORY_MASK_SJIS          \
 520    | CODING_CATEGORY_MASK_ISO_7         \
 521    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 522    | CODING_CATEGORY_MASK_ISO_8_1       \
 523    | CODING_CATEGORY_MASK_ISO_8_2       \
 524    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 525    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 526    | CODING_CATEGORY_MASK_CCL           \
 527    | CODING_CATEGORY_MASK_BIG5          \
 528    | CODING_CATEGORY_MASK_UTF_8         \
 529    | CODING_CATEGORY_MASK_UTF_16_BE     \
 530    | CODING_CATEGORY_MASK_UTF_16_LE)
 531
 532 #define CODING_CATEGORY_MASK_ISO_7BIT \
 533   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 534
 535 #define CODING_CATEGORY_MASK_ISO_8BIT \
 536   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 537
 538 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 539   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 540
 541 #define CODING_CATEGORY_MASK_ISO        \
 542   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 543    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 544    | CODING_CATEGORY_MASK_ISO_8BIT)
 545
 546 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 547    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 548
 549 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 550    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 551    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 552    internal format.  */
 553
 554 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 555   do {                                                  \
 556     if (s2 >= 0x9F)                                     \
 557       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 558       c2 = s2 - 0x7E;                                   \
 559     else                                                \
 560       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 561       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 562   } while (0)
 563
 564 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 565   do {                                                  \
 566     if (c1 & 1)                                         \
 567       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 568       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 569     else                                                \
 570       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 571       s2 = c2 + 0x7E;                                   \
 572   } while (0)
 573
 574 /* Encode the file name NAME using the specified coding system
 575    for file names, if any.  */
 576 #define ENCODE_FILE(name)                                                  \
 577   (! NILP (Vfile_name_coding_system)                                       \
 578    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 579    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 580    : (! NILP (Vdefault_file_name_coding_system)                            \
 581       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 582       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 583       : name))
 584
 585 /* Decode the file name NAME using the specified coding system
 586    for file names, if any.  */
 587 #define DECODE_FILE(name)                                                  \
 588   (! NILP (Vfile_name_coding_system)                                       \
 589    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 590    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 591    : (! NILP (Vdefault_file_name_coding_system)                            \
 592       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 593       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 594       : name))
 595
 596 /* Encode the string STR using the specified coding system
 597    for system functions, if any.  */
 598 #define ENCODE_SYSTEM(str)                                                 \
 599   (! NILP (Vlocale_coding_system)                                          \
 600    && !EQ (Vlocale_coding_system, make_number (0))                         \
 601    ? code_convert_string_norecord (str, Vlocale_coding_system, 1)          \
 602    : str)
 603
 604 /* Decode the string STR using the specified coding system
 605    for system functions, if any.  */
 606 #define DECODE_SYSTEM(str)                                                 \
 607   (! NILP (Vlocale_coding_system)                                          \
 608    && !EQ (Vlocale_coding_system, make_number (0))                         \
 609    ? code_convert_string_norecord (str, Vlocale_coding_system, 0)          \
 610    : str)
 611
 612 #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1)
 613
 614 /* Extern declarations.  */
 615 extern int decode_coding P_ ((struct coding_system *, const unsigned char *,
 616                               unsigned char *, int, int));
 617 extern int encode_coding P_ ((struct coding_system *, const unsigned char *,
 618                               unsigned char *, int, int));
 619 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 620                                          Lisp_Object));
 621 extern void coding_free_composition_data P_ ((struct coding_system *));
 622 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 623                                                   int));
 624 extern void coding_allocate_composition_data P_ ((struct coding_system *,
 625                                                   int));
 626 extern void coding_restore_composition P_ ((struct coding_system *,
 627                                             Lisp_Object));
 628 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 629                                     int, int));
 630 extern Lisp_Object run_pre_post_conversion_on_str P_ ((Lisp_Object,
 631                                                        struct coding_system *,
 632                                                        int));
 633 extern void run_pre_write_conversin_on_c_str P_ ((unsigned char **, int *,
 634                                                   int, int,
 635                                                   struct coding_system *));
 636
 637 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 638 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 639 extern void detect_coding P_ ((struct coding_system *, const unsigned char *,
 640                                int));
 641 extern void detect_eol P_ ((struct coding_system *, const unsigned char *,
 642                             int));
 643 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 644 extern Lisp_Object code_convert_string P_ ((Lisp_Object,
 645                                             struct coding_system *, int, int));
 646 extern Lisp_Object code_convert_string1 P_ ((Lisp_Object, Lisp_Object,
 647                                              Lisp_Object, int));
 648 extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object,
 649                                                      int));
 650 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 651 extern Lisp_Object encode_coding_string P_ ((Lisp_Object,
 652                                              struct coding_system *, int));
 653 extern Lisp_Object decode_coding_string P_ ((Lisp_Object,
 654                                              struct coding_system *, int));
 655 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 656 extern Lisp_Object Qraw_text, Qemacs_mule;
 657 extern Lisp_Object Qbuffer_file_coding_system;
 658 extern Lisp_Object Vcoding_category_list;
 659 extern Lisp_Object Qutf_8;
 660
 661 extern Lisp_Object Qtranslation_table;
 662 extern Lisp_Object Qtranslation_table_id;
 663
 664 /* Mnemonic strings to indicate each type of end-of-line.  */
 665 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 666 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 667 extern Lisp_Object eol_mnemonic_undecided;
 668
 669 #ifdef emacs
 670 extern Lisp_Object Qfile_coding_system;
 671 extern Lisp_Object Qcall_process, Qcall_process_region;
 672 extern Lisp_Object Qstart_process, Qopen_network_stream;
 673 extern Lisp_Object Qwrite_region;
 674
 675 extern char *emacs_strerror P_ ((int));
 676
 677 /* Coding-system for reading files and receiving data from process.  */
 678 extern Lisp_Object Vcoding_system_for_read;
 679 /* Coding-system for writing files and sending data to process.  */
 680 extern Lisp_Object Vcoding_system_for_write;
 681 /* Coding-system actually used in the latest I/O.  */
 682 extern Lisp_Object Vlast_coding_system_used;
 683 /* Coding-system to use with system messages (e.g. strerror).  */
 684 extern Lisp_Object Vlocale_coding_system;
 685
 686 /* If non-zero, process buffer inherits the coding system used to decode
 687    the subprocess output.  */
 688 extern int inherit_process_coding_system;
 689
 690 /* Coding-system to be used for encoding terminal output.  This
 691    structure contains information of a coding-system specified by the
 692    function `set-terminal-coding-system'.  */
 693 extern struct coding_system terminal_coding;
 694
 695 /* Coding system to be used to encode text for terminal display when
 696    terminal coding system is nil.  */
 697 extern struct coding_system safe_terminal_coding;
 698
 699 /* Coding-system of what is sent from terminal keyboard.  This
 700    structure contains information of a coding-system specified by the
 701    function `set-keyboard-coding-system'.  */
 702 extern struct coding_system keyboard_coding;
 703
 704 /* Default coding system to be used to write a file.  */
 705 extern struct coding_system default_buffer_file_coding;
 706
 707 /* Default coding systems used for process I/O.  */
 708 extern Lisp_Object Vdefault_process_coding_system;
 709
 710 /* Function to call to force a user to force select a proper coding
 711    system.  */
 712 extern Lisp_Object Vselect_safe_coding_system_function;
 713
 714 /* If nonzero, on writing a file, Vselect_safe_coding_system_function
 715    is called even if Vcoding_system_for_write is non-nil.  */
 716 extern int coding_system_require_warning;
 717
 718 /* Coding system for file names, or nil if none.  */
 719 extern Lisp_Object Vfile_name_coding_system;
 720
 721 /* Coding system for file names used only when
 722    Vfile_name_coding_system is nil.  */
 723 extern Lisp_Object Vdefault_file_name_coding_system;
 724
 725 #endif
 726
 727 /* Error signaled when there's a problem with detecting coding system */
 728 extern Lisp_Object Qcoding_system_error;
 729
 730 #endif /* EMACS_CODING_H */
 731
 732 /* arch-tag: 2bc3b4fa-6870-4f64-8135-b962b2d290e4
 733    (do not change this comment) */