code.delx.au - gnu-emacs/blob - 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_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 charactes 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 prohibited by CODING_FLAG_ISO_SAFE.  */
 150 #define CODING_INHIBIT_CHARACTER_SUBSTITUTION  077 /* 077 == `?' */
 151
 152 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 153 struct iso2022_spec
 154 {
 155   /* The current graphic register invoked to each graphic plane.  */
 156   int current_invocation[2];
 157
 158   /* The current charset designated to each graphic register.  */
 159   int current_designation[4];
 160
 161   /* A charset initially designated to each graphic register.  */
 162   int initial_designation[4];
 163
 164   /* If not -1, it is a graphic register specified in an invalid
 165      designation sequence.  */
 166   int last_invalid_designation_register;
 167
 168   /* A graphic register to which each charset should be designated.  */
 169   unsigned char requested_designation[MAX_CHARSET + 1];
 170
 171   /* A revision number to be specified for each charset on encoding.
 172      The value 255 means no revision number for the corresponding
 173      charset.  */
 174   unsigned char charset_revision_number[MAX_CHARSET + 1];
 175
 176   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 177      by single-shift while encoding.  */
 178   int single_shifting;
 179
 180   /* Set to 1 temporarily only when processing at beginning of line.  */
 181   int bol;
 182 };
 183
 184 /* Macros to access each field in the structure `spec.iso2022'.  */
 185 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 186   (coding)->spec.iso2022.current_invocation[plane]
 187 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 188   (coding)->spec.iso2022.current_designation[reg]
 189 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 190   (coding)->spec.iso2022.initial_designation[reg]
 191 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 192   (coding)->spec.iso2022.requested_designation[charset]
 193 #define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
 194   (coding)->spec.iso2022.charset_revision_number[charset]
 195 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 196   (coding)->spec.iso2022.single_shifting
 197 #define CODING_SPEC_ISO_BOL(coding) \
 198   (coding)->spec.iso2022.bol
 199
 200 /* A value which may appear in
 201    coding->spec.iso2022.requested_designation indicating that the
 202    corresponding charset does not request any graphic register to be
 203    designated.  */
 204 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 205
 206 /* Return a charset which is currently designated to the graphic plane
 207    PLANE in the coding-system CODING.  */
 208 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 209   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 210    ? -1                                                 \
 211    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 212                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 213
 214 /*** BIG5 section ***/
 215
 216 /* Macros to denote each type of BIG5 coding system.  */
 217 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 218                                         BIG5 developed by Hong Kong
 219                                         University.  */
 220 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 221                                         of BIG5 developed by the
 222                                         company ETen in Taiwan.  */
 223
 224 /*** GENERAL section ***/
 225
 226 /* Types of coding system.  */
 227 enum coding_type
 228   {
 229     coding_type_no_conversion,  /* A coding system which requires no
 230                                    conversion for reading and writing
 231                                    including end-of-line format.  */
 232     coding_type_emacs_mule,     /* A coding system used in Emacs'
 233                                    buffer and string.  Requires no
 234                                    conversion for reading and writing
 235                                    except for end-of-line format.  */
 236     coding_type_undecided,      /* A coding system which requires
 237                                    automatic detection of a real
 238                                    coding system.  */
 239     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 240     coding_type_iso2022,        /* Any coding system of ISO2022
 241                                    variants.  */
 242     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 243     coding_type_ccl,            /* The coding system of which decoder
 244                                    and encoder are written in CCL.  */
 245     coding_type_raw_text        /* A coding system for a text
 246                                    containing ramdom 8-bit code which
 247                                    does not require code conversion
 248                                    except for end-of-line format. */
 249   };
 250
 251 /* Formats of end-of-line.  */
 252 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 253                                    internal format.  */
 254 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 255                                    line-feed.  */
 256 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 257 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 258                                    eol-type is not yet decided.  */
 259 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 260                                      eol-type is not consistent
 261                                      through the file.  */
 262
 263 /* 1 iff composing.  */
 264 #define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
 265
 266 #define COMPOSITION_DATA_SIZE 4080
 267 #define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
 268
 269 /* Data structure to hold information about compositions of text that
 270    is being decoded or encode.  ISO 2022 base code conversion routines
 271    handle special ESC sequences for composition specification.  But,
 272    they can't get/put such information directly from/to a buffer in
 273    the deepest place.  So, they store or retrieve the information
 274    through this structure.
 275
 276    The encoder stores the information in this structure when it meets
 277    ESC sequences for composition while encoding codes, then, after all
 278    text codes are encoded, puts `composition' properties on the text
 279    by refering the structure.
 280
 281    The decoder at first stores the information of a text to be
 282    decoded, then, while decoding codes, generates ESC sequences for
 283    composition at proper places by refering the structure.  */
 284
 285 struct composition_data
 286 {
 287   /* The character position of the first character to be encoded or
 288      decoded.  START and END (see below) are relative to this
 289      position.  */
 290   int char_offset;
 291
 292   /* The composition data.  These elements are repeated for each
 293      composition:
 294         LENGTH START END METHOD [ COMPONENT ... ]
 295      where,
 296         LENGTH is the number of elements for this composition.
 297
 298         START and END are starting and ending character positions of
 299         the composition relative to `char_offset'.
 300
 301         METHOD is one of `enum cmposing_status' specifying the way of
 302         composition.
 303
 304         COMPONENT is a character or an encoded composition rule.  */
 305   int data[COMPOSITION_DATA_SIZE];
 306
 307   /* The number of elements in `data' currently used.  */
 308   int used;
 309
 310   /* Pointers to the previous and next structures.  When `data' is
 311      filled up, another structure is allocated and linked in `next'.
 312      The new struture has backward link to this struture in `prev'.
 313      The number of chaind structures depends on how many compositions
 314      the text being encoded or decoded contains.  */
 315   struct composition_data *prev, *next;
 316 };
 317
 318 /* Macros used for the member finish_status of the struct
 319    coding_system.  */
 320 #define CODING_FINISH_NORMAL            0
 321 #define CODING_FINISH_INSUFFICIENT_SRC  1
 322 #define CODING_FINISH_INSUFFICIENT_DST  2
 323 #define CODING_FINISH_INCONSISTENT_EOL  3
 324 #define CODING_FINISH_INSUFFICIENT_CMP  4
 325 #define CODING_FINISH_INTERRUPT         5
 326
 327 /* Macros used for the member `mode' of the struct coding_system.  */
 328
 329 /* If set, recover the original CR or LF of the already decoded text
 330    when the decoding routine encounters an inconsistent eol format.  */
 331 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 332
 333 /* If set, the decoding/encoding routines treat the current data as
 334    the last block of the whole text to be converted, and do
 335    appropriate fisishing job.  */
 336 #define CODING_MODE_LAST_BLOCK                  0x02
 337
 338 /* If set, it means that the current source text is in a buffer which
 339    enables selective display.  */
 340 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 341
 342 /* This flag is used by the decoding/encoding routines on the fly.  If
 343    set, it means that right-to-left text is being processed.  */
 344 #define CODING_MODE_DIRECTION                   0x08
 345
 346 struct coding_system
 347 {
 348   /* Type of the coding system.  */
 349   enum coding_type type;
 350
 351   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 352   int eol_type;
 353
 354   /* Flag bits of the coding system.  The meaning of each bit is common
 355      to all types of coding systems.  */
 356   unsigned int common_flags;
 357
 358   /* Flag bits of the coding system.  The meaning of each bit depends
 359      on the type of the coding system.  */
 360   unsigned int flags;
 361
 362   /* Mode bits of the coding system.  See the comments of the macros
 363      CODING_MODE_XXX.  */
 364   unsigned int mode;
 365
 366   /* Table of safe character sets for this coding system.  If the Nth
 367      element is 0, the charset of ID N is not a safe character set.
 368      Such a character set is not encoded when CODING_ISO_FLAG_SAFE is
 369      set.  */
 370   unsigned char safe_charsets[MAX_CHARSET + 1];
 371
 372   /* The current status of composition handling.  */
 373   int composing;
 374
 375   /* 1 iff the next character is a composition rule.  */
 376   int composition_rule_follows;
 377
 378   /* Information of compositions are stored here on decoding and set
 379      in advance on encoding.  */
 380   struct composition_data *cmp_data;
 381
 382   /* Index to cmp_data->data for the first element for the current
 383      composition.  */
 384   int cmp_data_start;
 385
 386   /* Index to cmp_data->data for the current element for the current
 387      composition.  */
 388   int cmp_data_index;
 389
 390   /* Detailed information specific to each type of coding system.  */
 391   union spec
 392     {
 393       struct iso2022_spec iso2022;
 394       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 395     } spec;
 396
 397   /* Index number of coding category of the coding system.  */
 398   int category_idx;
 399
 400   unsigned src_multibyte : 1;
 401   unsigned dst_multibyte : 1;
 402
 403   /* How may heading bytes we can skip for decoding.  This is set to
 404      -1 in setup_coding_system, and updated by detect_coding.  So,
 405      when this is equal to the byte length of the text being
 406      converted, we can skip the actual conversion process.  */
 407   int heading_ascii;
 408
 409   /* The following members are set by encoding/decoding routine.  */
 410   int produced, produced_char, consumed, consumed_char;
 411
 412   /* Number of error source data found in a decoding routine.  */
 413   int errors;
 414
 415   /* Finish status of code conversion.  It should be one of macros
 416      CODING_FINISH_XXXX.  */
 417   int result;
 418
 419   /* The following members are all Lisp symbols.  We don't have to
 420      protect them from GC because the current garbage collection
 421      doesn't relocate Lisp symbols.  But, when it is changed, we must
 422      find a way to protect them.  */
 423
 424   /* Backward pointer to the Lisp symbol of the coding system.  */
 425   Lisp_Object symbol;
 426
 427   /* Lisp function (symbol) to be called after decoding to do
 428      additional conversion, or nil.  */
 429   Lisp_Object post_read_conversion;
 430
 431   /* Lisp function (symbol) to be called before encoding to do
 432      additional conversion, or nil.  */
 433   Lisp_Object pre_write_conversion;
 434
 435   /* Character translation tables to look up, or nil.  */
 436   Lisp_Object translation_table_for_decode;
 437   Lisp_Object translation_table_for_encode;
 438 };
 439
 440 #define CODING_REQUIRE_FLUSHING_MASK    1
 441 #define CODING_REQUIRE_DECODING_MASK    2
 442 #define CODING_REQUIRE_ENCODING_MASK    4
 443 #define CODING_REQUIRE_DETECTION_MASK   8
 444
 445 /* Return 1 if the coding system CODING requires specific code to be
 446    attached at the tail of converted text.  */
 447 #define CODING_REQUIRE_FLUSHING(coding) \
 448   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 449
 450 /* Return 1 if the coding system CODING requires code conversion on
 451    decoding.  */
 452 #define CODING_REQUIRE_DECODING(coding) \
 453   ((coding)->dst_multibyte              \
 454    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 455
 456 /* Return 1 if the coding system CODING requires code conversion on
 457    encoding.  */
 458 #define CODING_REQUIRE_ENCODING(coding) \
 459   ((coding)->src_multibyte              \
 460    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 461
 462 /* Return 1 if the coding system CODING requires some kind of code
 463    detection.  */
 464 #define CODING_REQUIRE_DETECTION(coding) \
 465   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 466
 467 /* Return 1 if the coding system CODING requires code conversion on
 468    decoding or some kind of code detection.  */
 469 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 470   (CODING_REQUIRE_DECODING (coding)             \
 471    || CODING_REQUIRE_DETECTION (coding))
 472
 473 /* Index for each coding category in `coding_category_table' */
 474 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 475 #define CODING_CATEGORY_IDX_SJIS        1
 476 #define CODING_CATEGORY_IDX_ISO_7       2
 477 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 478 #define CODING_CATEGORY_IDX_ISO_8_1     4
 479 #define CODING_CATEGORY_IDX_ISO_8_2     5
 480 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 481 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 482 #define CODING_CATEGORY_IDX_CCL         8
 483 #define CODING_CATEGORY_IDX_BIG5        9
 484 #define CODING_CATEGORY_IDX_UTF_8       10
 485 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 486 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 487 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 488 #define CODING_CATEGORY_IDX_BINARY      14
 489 #define CODING_CATEGORY_IDX_MAX         15
 490
 491 /* Definitions of flag bits returned by the function
 492    detect_coding_mask ().  */
 493 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 494 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 495 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 496 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 497 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 498 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 499 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 500 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 501 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 502 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 503 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 504 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 505 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 506 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 507 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 508
 509 /* This value is returned if detect_coding_mask () find nothing other
 510    than ASCII characters.  */
 511 #define CODING_CATEGORY_MASK_ANY        \
 512   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 513    | CODING_CATEGORY_MASK_SJIS          \
 514    | CODING_CATEGORY_MASK_ISO_7         \
 515    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 516    | CODING_CATEGORY_MASK_ISO_8_1       \
 517    | CODING_CATEGORY_MASK_ISO_8_2       \
 518    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 519    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 520    | CODING_CATEGORY_MASK_CCL           \
 521    | CODING_CATEGORY_MASK_BIG5          \
 522    | CODING_CATEGORY_MASK_UTF_8         \
 523    | CODING_CATEGORY_MASK_UTF_16_BE     \
 524    | CODING_CATEGORY_MASK_UTF_16_LE)
 525
 526 #define CODING_CATEGORY_MASK_ISO_7BIT \
 527   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 528
 529 #define CODING_CATEGORY_MASK_ISO_8BIT \
 530   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 531
 532 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 533   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 534
 535 #define CODING_CATEGORY_MASK_ISO        \
 536   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 537    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 538    | CODING_CATEGORY_MASK_ISO_8BIT)
 539
 540 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 541    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 542
 543 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 544    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 545    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 546    internal format.  */
 547
 548 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 549   do {                                                  \
 550     if (s2 >= 0x9F)                                     \
 551       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 552       c2 = s2 - 0x7E;                                   \
 553     else                                                \
 554       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 555       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 556   } while (0)
 557
 558 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 559   do {                                                  \
 560     if (c1 & 1)                                         \
 561       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 562       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 563     else                                                \
 564       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 565       s2 = c2 + 0x7E;                                   \
 566   } while (0)
 567
 568 /* Encode the file name NAME using the specified coding system
 569    for file names, if any.  */
 570 #define ENCODE_FILE(name)                                                  \
 571   (! NILP (Vfile_name_coding_system)                                       \
 572    && XFASTINT (Vfile_name_coding_system) != 0                             \
 573    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 574    : (! NILP (Vdefault_file_name_coding_system)                            \
 575       && XFASTINT (Vdefault_file_name_coding_system) != 0                  \
 576       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 577       : name))
 578
 579 /* Decode the file name NAME using the specified coding system
 580    for file names, if any.  */
 581 #define DECODE_FILE(name)                                                  \
 582   (! NILP (Vfile_name_coding_system)                                       \
 583    && XFASTINT (Vfile_name_coding_system) != 0                             \
 584    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 585    : (! NILP (Vdefault_file_name_coding_system)                            \
 586       && XFASTINT (Vdefault_file_name_coding_system) != 0                  \
 587       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 588       : name))
 589
 590 /* Extern declarations.  */
 591 extern int decode_coding P_ ((struct coding_system *, unsigned char *,
 592                               unsigned char *, int, int));
 593 extern int encode_coding P_ ((struct coding_system *, unsigned char *,
 594                               unsigned char *, int, int));
 595 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 596                                          Lisp_Object));
 597 extern void coding_free_composition_data P_ ((struct coding_system *));
 598 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 599                                                   int));
 600 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 601                                     int, int));
 602 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 603 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 604 extern void detect_coding P_ ((struct coding_system *, unsigned char *, int));
 605 extern void detect_eol P_ ((struct coding_system *, unsigned char *, int));
 606 extern int conversion_buffer_size;
 607 extern char *conversion_buffer;
 608 extern char *get_conversion_buffer P_ ((int));
 609 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 610 extern Lisp_Object code_convert_string P_ ((Lisp_Object,
 611                                             struct coding_system *, int, int));
 612 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 613 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 614 extern Lisp_Object Qraw_text, Qemacs_mule;
 615 extern Lisp_Object Qbuffer_file_coding_system;
 616 extern Lisp_Object Vcoding_category_list;
 617
 618 extern Lisp_Object Qtranslation_table;
 619 extern Lisp_Object Qtranslation_table_id;
 620
 621 /* Mnemonic strings to indicate each type of end-of-line.  */
 622 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 623 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 624 extern Lisp_Object eol_mnemonic_undecided;
 625
 626 #ifdef emacs
 627 extern Lisp_Object Qfile_coding_system;
 628 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 629 extern Lisp_Object Qstart_process, Qopen_network_stream;
 630
 631 extern char *emacs_strerror P_ ((int));
 632
 633 /* Coding-system for reading files and receiving data from process.  */
 634 extern Lisp_Object Vcoding_system_for_read;
 635 /* Coding-system for writing files and sending data to process.  */
 636 extern Lisp_Object Vcoding_system_for_write;
 637 /* Coding-system actually used in the latest I/O.  */
 638 extern Lisp_Object Vlast_coding_system_used;
 639 /* Coding-system to use with system messages (e.g. strerror).  */
 640 extern Lisp_Object Vlocale_coding_system;
 641
 642 /* If non-zero, process buffer inherits the coding system used to decode
 643    the subprocess output.  */
 644 extern int inherit_process_coding_system;
 645
 646 /* Coding-system to be used for encoding terminal output.  This
 647    structure contains information of a coding-system specified by the
 648    function `set-terminal-coding-system'.  */
 649 extern struct coding_system terminal_coding;
 650
 651 /* Coding system to be used to encode text for terminal display when
 652    terminal coding system is nil.  */
 653 extern struct coding_system safe_terminal_coding;
 654
 655 /* Coding-system of what is sent from terminal keyboard.  This
 656    structure contains information of a coding-system specified by the
 657    function `set-keyboard-coding-system'.  */
 658 extern struct coding_system keyboard_coding;
 659
 660 /* Default coding system to be used to write a file.  */
 661 extern struct coding_system default_buffer_file_coding;
 662
 663 /* Default coding systems used for process I/O.  */
 664 extern Lisp_Object Vdefault_process_coding_system;
 665
 666 /* Function to call to force a user to force select a propert coding
 667    system.  */
 668 extern Lisp_Object Vselect_safe_coding_system_function;
 669
 670 /* Coding system for file names, or nil if none.  */
 671 extern Lisp_Object Vfile_name_coding_system;
 672
 673 /* Coding system for file names used only when
 674    Vfile_name_coding_system is nil.  */
 675 extern Lisp_Object Vdefault_file_name_coding_system;
 676 #endif
 677
 678 #endif /* _CODING_H */