src/coding.h

   1 /* Header for coding system handler.
   2    Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
   3    Licensed to the Free Software Foundation.
   4
   5 This file is part of GNU Emacs.
   6
   7 GNU Emacs is free software; you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License as published by
   9 the Free Software Foundation; either version 2, or (at your option)
  10 any later version.
  11
  12 GNU Emacs is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with GNU Emacs; see the file COPYING.  If not, write to
  19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  20 Boston, MA 02111-1307, USA.  */
  21
  22 #ifndef _CODING_H
  23 #define _CODING_H
  24
  25 #ifndef _CCL_H
  26 #include "ccl.h"
  27 #endif
  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_code,           /* Control codes in the range
  76                                    0x00..0x1F, 0x7F, and 0x80..0x9F,
  77                                    except for the following seven
  78                                    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_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  85     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  86     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  87     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  88     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  89     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  90     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  91   };
  92
  93 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  94   element in the structure `coding_system'.  This information is used
  95   while encoding a text to ISO2022.  **/
  96
  97 /* If set, produce short-form designation sequence (e.g. ESC $ A)
  98    instead of long-form sequence (e.g. ESC $ ( A).  */
  99 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 100
 101 /* If set, reset graphic planes and registers at end-of-line to the
 102    initial state.  */
 103 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 104
 105 /* If set, reset graphic planes and registers before any control
 106    characters to the initial state.  */
 107 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 108
 109 /* If set, encode by 7-bit environment.  */
 110 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 111
 112 /* If set, use locking-shift function.  */
 113 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 114
 115 /* If set, use single-shift function.  Overwrite
 116    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 117 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 118
 119 /* If set, designate JISX0201-Roman instead of ASCII.  */
 120 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 121
 122 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 123 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 124
 125 /* If set, do not produce ISO6429's direction specifying sequence.  */
 126 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 127
 128 /* If set, assume designation states are reset at beginning of line on
 129    output.  */
 130 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 131
 132 /* If set, designation sequence should be placed at beginning of line
 133    on output.  */
 134 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 135
 136 /* If set, do not encode unsafe charactes on output.  */
 137 #define CODING_FLAG_ISO_SAFE            0x0800
 138
 139 /* If set, extra latin codes (128..159) are accepted as a valid code
 140    on input.  */
 141 #define CODING_FLAG_ISO_LATIN_EXTRA     0x1000
 142
 143 /* If set, use designation escape sequence.  */
 144 #define CODING_FLAG_ISO_DESIGNATION     0x10000
 145
 146 /* A character to be produced on output if encoding of the original
 147    character is prohibited by CODING_FLAG_ISO_SAFE.  */
 148 #define CODING_INHIBIT_CHARACTER_SUBSTITUTION  077 /* 077 == `?' */
 149
 150 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 151 struct iso2022_spec
 152 {
 153   /* The current graphic register invoked to each graphic plane.  */
 154   int current_invocation[2];
 155
 156   /* The current charset designated to each graphic register.  */
 157   int current_designation[4];
 158
 159   /* A charset initially designated to each graphic register.  */
 160   int initial_designation[4];
 161
 162   /* If not -1, it is a graphic register specified in an invalid
 163      designation sequence.  */
 164   int last_invalid_designation_register;
 165
 166   /* A graphic register to which each charset should be designated.  */
 167   unsigned char requested_designation[MAX_CHARSET + 1];
 168
 169   /* A revision number to be specified for each charset on encoding.
 170      The value 255 means no revision number for the corresponding
 171      charset.  */
 172   unsigned char charset_revision_number[MAX_CHARSET + 1];
 173
 174   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 175      by single-shift while encoding.  */
 176   int single_shifting;
 177
 178   /* Set to 1 temporarily only when processing at beginning of line.  */
 179   int bol;
 180 };
 181
 182 /* Macros to access each field in the structure `spec.iso2022'.  */
 183 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 184   (coding)->spec.iso2022.current_invocation[plane]
 185 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 186   (coding)->spec.iso2022.current_designation[reg]
 187 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 188   (coding)->spec.iso2022.initial_designation[reg]
 189 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 190   (coding)->spec.iso2022.requested_designation[charset]
 191 #define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
 192   (coding)->spec.iso2022.charset_revision_number[charset]
 193 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 194   (coding)->spec.iso2022.single_shifting
 195 #define CODING_SPEC_ISO_BOL(coding) \
 196   (coding)->spec.iso2022.bol
 197
 198 /* A value which may appear in
 199    coding->spec.iso2022.requested_designation indicating that the
 200    corresponding charset does not request any graphic register to be
 201    designated.  */
 202 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 203
 204 /* Return a charset which is currently designated to the graphic plane
 205    PLANE in the coding-system CODING.  */
 206 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 207   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 208    ? -1                                                 \
 209    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 210                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 211
 212 /*** BIG5 section ***/
 213
 214 /* Macros to denote each type of BIG5 coding system.  */
 215 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 216                                         BIG5 developed by Hong Kong
 217                                         University.  */
 218 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 219                                         of BIG5 developed by the
 220                                         company ETen in Taiwan.  */
 221
 222 /*** GENERAL section ***/
 223
 224 /* Types of coding system.  */
 225 enum coding_type
 226   {
 227     coding_type_no_conversion,  /* A coding system which requires no
 228                                    conversion for reading and writing
 229                                    including end-of-line format.  */
 230     coding_type_emacs_mule,     /* A coding system used in Emacs'
 231                                    buffer and string.  Requires no
 232                                    conversion for reading and writing
 233                                    except for end-of-line format.  */
 234     coding_type_undecided,      /* A coding system which requires
 235                                    automatic detection of a real
 236                                    coding system.  */
 237     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 238     coding_type_iso2022,        /* Any coding system of ISO2022
 239                                    variants.  */
 240     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 241     coding_type_ccl,            /* The coding system of which decoder
 242                                    and encoder are written in CCL.  */
 243     coding_type_raw_text        /* A coding system for a text
 244                                    containing ramdom 8-bit code which
 245                                    does not require code conversion
 246                                    except for end-of-line format. */
 247   };
 248
 249 /* Formats of end-of-line.  */
 250 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 251                                    internal format.  */
 252 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 253                                    line-feed.  */
 254 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 255 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 256                                    eol-type is not yet decided.  */
 257 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 258                                      eol-type is not consistent
 259                                      through the file.  */
 260
 261 /* 1 iff composing.  */
 262 #define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
 263
 264 #define COMPOSITION_DATA_SIZE 4080
 265 #define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
 266
 267 /* Data structure to hold information about compositions of text that
 268    is being decoded or encode.  ISO 2022 base code conversion routines
 269    handle special ESC sequences for composition specification.  But,
 270    they can't get/put such information directly from/to a buffer in
 271    the deepest place.  So, they store or retrieve the information
 272    through this structure.
 273
 274    The encoder stores the information in this structure when it meets
 275    ESC sequences for composition while encoding codes, then, after all
 276    text codes are encoded, puts `composition' properties on the text
 277    by refering the structure.
 278
 279    The decoder at first stores the information of a text to be
 280    decoded, then, while decoding codes, generates ESC sequences for
 281    composition at proper places by refering the structure.  */
 282
 283 struct composition_data
 284 {
 285   /* The character position of the first character to be encoded or
 286      decoded.  START and END (see below) are relative to this
 287      position.  */
 288   int char_offset;
 289
 290   /* The composition data.  These elements are repeated for each
 291      composition:
 292         LENGTH START END METHOD [ COMPONENT ... ]
 293      where,
 294         LENGTH is the number of elements for this composition.
 295
 296         START and END are starting and ending character positions of
 297         the composition relative to `char_offset'.
 298
 299         METHOD is one of `enum cmposing_status' specifying the way of
 300         composition.
 301
 302         COMPONENT is a character or an encoded composition rule.  */
 303   int data[COMPOSITION_DATA_SIZE];
 304
 305   /* The number of elements in `data' currently used.  */
 306   int used;
 307
 308   /* Pointers to the previous and next structures.  When `data' is
 309      filled up, another structure is allocated and linked in `next'.
 310      The new struture has backward link to this struture in `prev'.
 311      The number of chaind structures depends on how many compositions
 312      the text being encoded or decoded contains.  */
 313   struct composition_data *prev, *next;
 314 };
 315
 316 /* Macros used for the member finish_status of the struct
 317    coding_system.  */
 318 #define CODING_FINISH_NORMAL            0
 319 #define CODING_FINISH_INSUFFICIENT_SRC  1
 320 #define CODING_FINISH_INSUFFICIENT_DST  2
 321 #define CODING_FINISH_INCONSISTENT_EOL  3
 322 #define CODING_FINISH_INSUFFICIENT_CMP  4
 323 #define CODING_FINISH_INTERRUPT         5
 324
 325 /* Macros used for the member `mode' of the struct coding_system.  */
 326
 327 /* If set, recover the original CR or LF of the already decoded text
 328    when the decoding routine encounters an inconsistent eol format.  */
 329 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 330
 331 /* If set, the decoding/encoding routines treat the current data as
 332    the last block of the whole text to be converted, and do
 333    appropriate fisishing job.  */
 334 #define CODING_MODE_LAST_BLOCK                  0x02
 335
 336 /* If set, it means that the current source text is in a buffer which
 337    enables selective display.  */
 338 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 339
 340 /* This flag is used by the decoding/encoding routines on the fly.  If
 341    set, it means that right-to-left text is being processed.  */
 342 #define CODING_MODE_DIRECTION                   0x08
 343
 344 struct coding_system
 345 {
 346   /* Type of the coding system.  */
 347   enum coding_type type;
 348
 349   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 350   int eol_type;
 351
 352   /* Flag bits of the coding system.  The meaning of each bit is common
 353      to all types of coding systems.  */
 354   unsigned int common_flags;
 355
 356   /* Flag bits of the coding system.  The meaning of each bit depends
 357      on the type of the coding system.  */
 358   unsigned int flags;
 359
 360   /* Mode bits of the coding system.  See the comments of the macros
 361      CODING_MODE_XXX.  */
 362   unsigned int mode;
 363
 364   /* Table of safe character sets for this coding system.  If the Nth
 365      element is 0, the charset of ID N is not a safe character set.
 366      Such a character set is not encoded when CODING_ISO_FLAG_SAFE is
 367      set.  */
 368   unsigned char safe_charsets[MAX_CHARSET + 1];
 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   /* How may heading bytes we can skip for decoding.  This is set to
 399      -1 in setup_coding_system, and updated by detect_coding.  So,
 400      when this is equal to the byte length of the text being
 401      converted, we can skip the actual conversion process.  */
 402   int heading_ascii;
 403
 404   /* The following members are set by encoding/decoding routine.  */
 405   int produced, produced_char, consumed, consumed_char;
 406
 407   /* Encoding routines set this to 1 when they produce a byte sequence
 408      which can be parsed as a multibyte character.  Decoding routines
 409      set this to 1 when they encounter an invalid code and, as the
 410      result, produce an unexpected multibyte character.  */
 411   int fake_multibyte;
 412
 413   /* The following members are all Lisp symbols.  We don't have to
 414      protect them from GC because the current garbage collection
 415      doesn't relocate Lisp symbols.  But, when it is changed, we must
 416      find a way to protect them.  */
 417
 418   /* Backward pointer to the Lisp symbol of the coding system.  */
 419   Lisp_Object symbol;
 420
 421   /* Lisp function (symbol) to be called after decoding to do
 422      additional conversion, or nil.  */
 423   Lisp_Object post_read_conversion;
 424
 425   /* Lisp function (symbol) to be called before encoding to do
 426      additional conversion, or nil.  */
 427   Lisp_Object pre_write_conversion;
 428
 429   /* Character translation tables to look up, or nil.  */
 430   Lisp_Object translation_table_for_decode;
 431   Lisp_Object translation_table_for_encode;
 432 };
 433
 434 #define CODING_REQUIRE_FLUSHING_MASK    1
 435 #define CODING_REQUIRE_DECODING_MASK    2
 436 #define CODING_REQUIRE_ENCODING_MASK    4
 437 #define CODING_REQUIRE_DETECTION_MASK   8
 438
 439 /* Return 1 if the coding system CODING requires specific code to be
 440    attached at the tail of converted text.  */
 441 #define CODING_REQUIRE_FLUSHING(coding) \
 442   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 443
 444 /* Return 1 if the coding system CODING requires code conversion on
 445    decoding.  */
 446 #define CODING_REQUIRE_DECODING(coding) \
 447   ((coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 448
 449 /* Return 1 if the coding system CODING requires code conversion on
 450    encoding.  */
 451 #define CODING_REQUIRE_ENCODING(coding) \
 452   ((coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 453
 454 /* Return 1 if the coding system CODING requires some kind of code
 455    detection.  */
 456 #define CODING_REQUIRE_DETECTION(coding) \
 457   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 458
 459 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 460   ((coding)->common_flags                       \
 461    & (CODING_REQUIRE_DETECTION_MASK | CODING_REQUIRE_DECODING_MASK))
 462
 463 /* Index for each coding category in `coding_category_table' */
 464 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 465 #define CODING_CATEGORY_IDX_SJIS        1
 466 #define CODING_CATEGORY_IDX_ISO_7       2
 467 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 468 #define CODING_CATEGORY_IDX_ISO_8_1     4
 469 #define CODING_CATEGORY_IDX_ISO_8_2     5
 470 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 471 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 472 #define CODING_CATEGORY_IDX_CCL         8
 473 #define CODING_CATEGORY_IDX_BIG5        9
 474 #define CODING_CATEGORY_IDX_UTF_8       10
 475 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 476 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 477 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 478 #define CODING_CATEGORY_IDX_BINARY      14
 479 #define CODING_CATEGORY_IDX_MAX         15
 480
 481 /* Definitions of flag bits returned by the function
 482    detect_coding_mask ().  */
 483 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 484 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 485 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 486 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 487 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 488 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 489 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 490 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 491 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 492 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 493 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 494 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 495 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 496 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 497 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 498
 499 /* This value is returned if detect_coding_mask () find nothing other
 500    than ASCII characters.  */
 501 #define CODING_CATEGORY_MASK_ANY        \
 502   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 503    | CODING_CATEGORY_MASK_SJIS          \
 504    | CODING_CATEGORY_MASK_ISO_7         \
 505    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 506    | CODING_CATEGORY_MASK_ISO_8_1       \
 507    | CODING_CATEGORY_MASK_ISO_8_2       \
 508    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 509    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 510    | CODING_CATEGORY_MASK_CCL           \
 511    | CODING_CATEGORY_MASK_BIG5          \
 512    | CODING_CATEGORY_MASK_UTF_8         \
 513    | CODING_CATEGORY_MASK_UTF_16_BE     \
 514    | CODING_CATEGORY_MASK_UTF_16_LE)
 515
 516 #define CODING_CATEGORY_MASK_ISO_7BIT \
 517   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 518
 519 #define CODING_CATEGORY_MASK_ISO_8BIT \
 520   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 521
 522 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 523   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 524
 525 #define CODING_CATEGORY_MASK_ISO        \
 526   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 527    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 528    | CODING_CATEGORY_MASK_ISO_8BIT)
 529
 530 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 531    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 532
 533 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 534    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 535    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 536    internal format.  */
 537
 538 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 539   do {                                                  \
 540     if (s2 >= 0x9F)                                     \
 541       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 542       c2 = s2 - 0x7E;                                   \
 543     else                                                \
 544       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 545       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 546   } while (0)
 547
 548 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 549   do {                                                  \
 550     if (c1 & 1)                                         \
 551       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 552       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 553     else                                                \
 554       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 555       s2 = c2 + 0x7E;                                   \
 556   } while (0)
 557
 558 /* Encode the file name NAME using the specified coding system
 559    for file names, if any.  */
 560 #define ENCODE_FILE(name)                                                  \
 561   (! NILP (Vfile_name_coding_system)                                       \
 562    && XFASTINT (Vfile_name_coding_system) != 0                             \
 563    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 564    : (! NILP (Vdefault_file_name_coding_system)                            \
 565       && XFASTINT (Vdefault_file_name_coding_system) != 0                  \
 566       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 567       : name))
 568
 569 /* Decode the file name NAME using the specified coding system
 570    for file names, if any.  */
 571 #define DECODE_FILE(name)                                                  \
 572   (! NILP (Vfile_name_coding_system)                                       \
 573    && XFASTINT (Vfile_name_coding_system) != 0                             \
 574    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 575    : (! NILP (Vdefault_file_name_coding_system)                            \
 576       && XFASTINT (Vdefault_file_name_coding_system) != 0                  \
 577       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 578       : name))
 579
 580 /* Extern declarations.  */
 581 extern int decode_coding P_ ((struct coding_system *, unsigned char *,
 582                               unsigned char *, int, int));
 583 extern int encode_coding P_ ((struct coding_system *, unsigned char *,
 584                               unsigned char *, int, int));
 585 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 586                                          Lisp_Object));
 587 extern void coding_free_composition_data P_ ((struct coding_system *));
 588 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 589                                                   int));
 590 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 591                                     int, int));
 592 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 593 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 594 extern void detect_coding P_ ((struct coding_system *, unsigned char *, int));
 595 extern void detect_eol P_ ((struct coding_system *, unsigned char *, int));
 596 extern int conversion_buffer_size;
 597 extern char *conversion_buffer;
 598 extern char *get_conversion_buffer P_ ((int));
 599 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 600 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 601 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 602 extern Lisp_Object Qraw_text, Qemacs_mule;
 603 extern Lisp_Object Qbuffer_file_coding_system;
 604 extern Lisp_Object Vcoding_category_list;
 605
 606 extern Lisp_Object Qtranslation_table;
 607 extern Lisp_Object Qtranslation_table_id;
 608
 609 /* Mnemonic strings to indicate each type of end-of-line.  */
 610 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 611 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 612 extern Lisp_Object eol_mnemonic_undecided;
 613
 614 #ifdef emacs
 615 extern Lisp_Object Qfile_coding_system;
 616 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 617 extern Lisp_Object Qstart_process, Qopen_network_stream;
 618
 619 extern char *emacs_strerror P_ ((int));
 620
 621 /* Coding-system for reading files and receiving data from process.  */
 622 extern Lisp_Object Vcoding_system_for_read;
 623 /* Coding-system for writing files and sending data to process.  */
 624 extern Lisp_Object Vcoding_system_for_write;
 625 /* Coding-system actually used in the latest I/O.  */
 626 extern Lisp_Object Vlast_coding_system_used;
 627 /* Coding-system to use with system messages (e.g. strerror).  */
 628 extern Lisp_Object Vlocale_coding_system;
 629
 630 /* If non-zero, process buffer inherits the coding system used to decode
 631    the subprocess output.  */
 632 extern int inherit_process_coding_system;
 633
 634 /* Coding-system to be used for encoding terminal output.  This
 635    structure contains information of a coding-system specified by the
 636    function `set-terminal-coding-system'.  */
 637 extern struct coding_system terminal_coding;
 638
 639 /* Coding system to be used to encode text for terminal display when
 640    terminal coding system is nil.  */
 641 extern struct coding_system safe_terminal_coding;
 642
 643 /* Coding-system of what is sent from terminal keyboard.  This
 644    structure contains information of a coding-system specified by the
 645    function `set-keyboard-coding-system'.  */
 646 extern struct coding_system keyboard_coding;
 647
 648 /* Default coding system to be used to write a file.  */
 649 extern struct coding_system default_buffer_file_coding;
 650
 651 /* Default coding systems used for process I/O.  */
 652 extern Lisp_Object Vdefault_process_coding_system;
 653
 654 /* Function to call to force a user to force select a propert coding
 655    system.  */
 656 extern Lisp_Object Vselect_safe_coding_system_function;
 657
 658 /* Coding system for file names, or nil if none.  */
 659 extern Lisp_Object Vfile_name_coding_system;
 660
 661 /* Coding system for file names used only when
 662    Vfile_name_coding_system is nil.  */
 663 extern Lisp_Object Vdefault_file_name_coding_system;
 664 #endif
 665
 666 #endif /* _CODING_H */