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 EMACS_CODING_H
  23 #define EMACS_CODING_H
  24
  25 #include "ccl.h"
  26
  27 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  28
  29 /* All code (1-byte) of Emacs' internal format is classified into one
  30    of the followings.  See also `charset.h'.  */
  31 enum emacs_code_class_type
  32   {
  33     EMACS_control_code,         /* Control codes in the range
  34                                    0x00..0x1F and 0x7F except for the
  35                                    following two codes.  */
  36     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  37                                    end-of-line.  */
  38     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  39                                    in selective display mode.  */
  40     EMACS_ascii_code,           /* ASCII characters.  */
  41     EMACS_leading_code_2,       /* Base leading code of official
  42                                    TYPE9N character.  */
  43     EMACS_leading_code_3,       /* Base leading code of private TYPE9N
  44                                    or official TYPE9Nx9N character.  */
  45     EMACS_leading_code_4,       /* Base leading code of private
  46                                    TYPE9Nx9N character.  */
  47     EMACS_invalid_code          /* Invalid code, i.e. a base leading
  48                                    code not yet assigned to any
  49                                    charset, or a code of the range
  50                                    0xA0..0xFF.  */
  51   };
  52
  53 extern enum emacs_code_class_type emacs_code_class[256];
  54
  55 /*** ISO2022 section ***/
  56
  57 /* Macros to define code of control characters for ISO2022's functions.  */
  58                         /* code */      /* function */
  59 #define ISO_CODE_LF     0x0A            /* line-feed */
  60 #define ISO_CODE_CR     0x0D            /* carriage-return */
  61 #define ISO_CODE_SO     0x0E            /* shift-out */
  62 #define ISO_CODE_SI     0x0F            /* shift-in */
  63 #define ISO_CODE_SS2_7  0x19            /* single-shift-2 for 7-bit code */
  64 #define ISO_CODE_ESC    0x1B            /* escape */
  65 #define ISO_CODE_SS2    0x8E            /* single-shift-2 */
  66 #define ISO_CODE_SS3    0x8F            /* single-shift-3 */
  67 #define ISO_CODE_CSI    0x9B            /* control-sequence-introduce */
  68
  69 /* All code (1-byte) of ISO2022 is classified into one of the
  70    followings.  */
  71 enum iso_code_class_type
  72   {
  73     ISO_control_0,              /* Control codes in the range
  74                                    0x00..0x1F and 0x7F, except for the
  75                                    following 5 codes.  */
  76     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  77     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  78     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  79     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  80     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  81     ISO_control_1,              /* Control codes in the range
  82                                    0x80..0x9F, except for the
  83                                    following 3 codes.  */
  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 characters 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 inhibitted by CODING_MODE_INHIBIT_UNENCODABLE_CHAR.
 148    It must be an ASCII character.  */
 149 #define CODING_REPLACEMENT_CHARACTER '?'
 150
 151 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 152 struct iso2022_spec
 153 {
 154   /* The current graphic register invoked to each graphic plane.  */
 155   int current_invocation[2];
 156
 157   /* The current charset designated to each graphic register.  */
 158   int current_designation[4];
 159
 160   /* A charset initially designated to each graphic register.  */
 161   int initial_designation[4];
 162
 163   /* If not -1, it is a graphic register specified in an invalid
 164      designation sequence.  */
 165   int last_invalid_designation_register;
 166
 167   /* A graphic register to which each charset should be designated.  */
 168   unsigned char requested_designation[MAX_CHARSET + 1];
 169
 170   /* A revision number to be specified for each charset on encoding.
 171      The value 255 means no revision number for the corresponding
 172      charset.  */
 173   unsigned char charset_revision_number[MAX_CHARSET + 1];
 174
 175   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 176      by single-shift while encoding.  */
 177   int single_shifting;
 178
 179   /* Set to 1 temporarily only when processing at beginning of line.  */
 180   int bol;
 181 };
 182
 183 /* Macros to access each field in the structure `spec.iso2022'.  */
 184 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 185   (coding)->spec.iso2022.current_invocation[plane]
 186 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 187   (coding)->spec.iso2022.current_designation[reg]
 188 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 189   (coding)->spec.iso2022.initial_designation[reg]
 190 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 191   (coding)->spec.iso2022.requested_designation[charset]
 192 #define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
 193   (coding)->spec.iso2022.charset_revision_number[charset]
 194 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 195   (coding)->spec.iso2022.single_shifting
 196 #define CODING_SPEC_ISO_BOL(coding) \
 197   (coding)->spec.iso2022.bol
 198
 199 /* A value which may appear in
 200    coding->spec.iso2022.requested_designation indicating that the
 201    corresponding charset does not request any graphic register to be
 202    designated.  */
 203 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 204
 205 /* Return a charset which is currently designated to the graphic plane
 206    PLANE in the coding-system CODING.  */
 207 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 208   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 209    ? -1                                                 \
 210    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 211                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 212
 213 /*** BIG5 section ***/
 214
 215 /* Macros to denote each type of BIG5 coding system.  */
 216 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 217                                         BIG5 developed by Hong Kong
 218                                         University.  */
 219 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 220                                         of BIG5 developed by the
 221                                         company ETen in Taiwan.  */
 222
 223 /*** GENERAL section ***/
 224
 225 /* Types of coding system.  */
 226 enum coding_type
 227   {
 228     coding_type_no_conversion,  /* A coding system which requires no
 229                                    conversion for reading and writing
 230                                    including end-of-line format.  */
 231     coding_type_emacs_mule,     /* A coding system used in Emacs'
 232                                    buffer and string.  Requires no
 233                                    conversion for reading and writing
 234                                    except for end-of-line format.  */
 235     coding_type_undecided,      /* A coding system which requires
 236                                    automatic detection of a real
 237                                    coding system.  */
 238     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 239     coding_type_iso2022,        /* Any coding system of ISO2022
 240                                    variants.  */
 241     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 242     coding_type_ccl,            /* The coding system of which decoder
 243                                    and encoder are written in CCL.  */
 244     coding_type_raw_text        /* A coding system for a text
 245                                    containing random 8-bit code which
 246                                    does not require code conversion
 247                                    except for end-of-line format. */
 248   };
 249
 250 /* Formats of end-of-line.  */
 251 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 252                                    internal format.  */
 253 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 254                                    line-feed.  */
 255 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 256 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 257                                    eol-type is not yet decided.  */
 258 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 259                                      eol-type is not consistent
 260                                      through the file.  */
 261
 262 /* 1 iff composing.  */
 263 #define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
 264
 265 #define COMPOSITION_DATA_SIZE 4080
 266 #define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
 267
 268 /* Data structure to hold information about compositions of text that
 269    is being decoded or encode.  ISO 2022 base code conversion routines
 270    handle special ESC sequences for composition specification.  But,
 271    they can't get/put such information directly from/to a buffer in
 272    the deepest place.  So, they store or retrieve the information
 273    through this structure.
 274
 275    The encoder stores the information in this structure when it meets
 276    ESC sequences for composition while encoding codes, then, after all
 277    text codes are encoded, puts `composition' properties on the text
 278    by referring to the structure.
 279
 280    The decoder at first stores the information of a text to be
 281    decoded, then, while decoding codes, generates ESC sequences for
 282    composition at proper places by referring to the structure.  */
 283
 284 struct composition_data
 285 {
 286   /* The character position of the first character to be encoded or
 287      decoded.  START and END (see below) are relative to this
 288      position.  */
 289   int char_offset;
 290
 291   /* The composition data.  These elements are repeated for each
 292      composition:
 293         LENGTH START END METHOD [ COMPONENT ... ]
 294      where,
 295         LENGTH is the number of elements for this composition.
 296
 297         START and END are starting and ending character positions of
 298         the composition relative to `char_offset'.
 299
 300         METHOD is one of `enum composing_status' specifying the way of
 301         composition.
 302
 303         COMPONENT is a character or an encoded composition rule.  */
 304   int data[COMPOSITION_DATA_SIZE];
 305
 306   /* The number of elements in `data' currently used.  */
 307   int used;
 308
 309   /* Pointers to the previous and next structures.  When `data' is
 310      filled up, another structure is allocated and linked in `next'.
 311      The new structure has backward link to this structure in `prev'.
 312      The number of chained structures depends on how many compositions
 313      the text being encoded or decoded contains.  */
 314   struct composition_data *prev, *next;
 315 };
 316
 317 /* Macros used for the member `result' of the struct
 318    coding_system.  */
 319 #define CODING_FINISH_NORMAL            0
 320 #define CODING_FINISH_INSUFFICIENT_SRC  1
 321 #define CODING_FINISH_INSUFFICIENT_DST  2
 322 #define CODING_FINISH_INCONSISTENT_EOL  3
 323 #define CODING_FINISH_INSUFFICIENT_CMP  4
 324 #define CODING_FINISH_INTERRUPT         5
 325
 326 /* Macros used for the member `mode' of the struct coding_system.  */
 327
 328 /* If set, recover the original CR or LF of the already decoded text
 329    when the decoding routine encounters an inconsistent eol format.  */
 330 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 331
 332 /* If set, the decoding/encoding routines treat the current data as
 333    the last block of the whole text to be converted, and do
 334    appropriate finishing job.  */
 335 #define CODING_MODE_LAST_BLOCK                  0x02
 336
 337 /* If set, it means that the current source text is in a buffer which
 338    enables selective display.  */
 339 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 340
 341 /* If set, replace unencodabae characters by `?' on encoding.  */
 342 #define CODING_MODE_INHIBIT_UNENCODABLE_CHAR    0x08
 343
 344 /* This flag is used by the decoding/encoding routines on the fly.  If
 345    set, it means that right-to-left text is being processed.  */
 346 #define CODING_MODE_DIRECTION                   0x10
 347
 348 struct coding_system
 349 {
 350   /* Type of the coding system.  */
 351   enum coding_type type;
 352
 353   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 354   int eol_type;
 355
 356   /* Flag bits of the coding system.  The meaning of each bit is common
 357      to all types of coding systems.  */
 358   unsigned int common_flags;
 359
 360   /* Flag bits of the coding system.  The meaning of each bit depends
 361      on the type of the coding system.  */
 362   unsigned int flags;
 363
 364   /* Mode bits of the coding system.  See the comments of the macros
 365      CODING_MODE_XXX.  */
 366   unsigned int mode;
 367
 368   /* The current status of composition handling.  */
 369   int composing;
 370
 371   /* 1 iff the next character is a composition rule.  */
 372   int composition_rule_follows;
 373
 374   /* Information of compositions are stored here on decoding and set
 375      in advance on encoding.  */
 376   struct composition_data *cmp_data;
 377
 378   /* Index to cmp_data->data for the first element for the current
 379      composition.  */
 380   int cmp_data_start;
 381
 382   /* Index to cmp_data->data for the current element for the current
 383      composition.  */
 384   int cmp_data_index;
 385
 386   /* Detailed information specific to each type of coding system.  */
 387   union spec
 388     {
 389       struct iso2022_spec iso2022;
 390       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 391     } spec;
 392
 393   /* Index number of coding category of the coding system.  */
 394   int category_idx;
 395
 396   /* The following two members specify how characters 128..159 are
 397      represented in source and destination text respectively.  1 means
 398      they are represented by 2-byte sequence, 0 means they are
 399      represented by 1-byte as is (see the comment in charset.h).  */
 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   /* If nonzero, suppress error notification.  */
 420   int suppress_error;
 421
 422   /* The following members are all Lisp symbols.  We don't have to
 423      protect them from GC because the current garbage collection
 424      doesn't relocate Lisp symbols.  But, when it is changed, we must
 425      find a way to protect them.  */
 426
 427   /* Backward pointer to the Lisp symbol of the coding system.  */
 428   Lisp_Object symbol;
 429
 430   /* Lisp function (symbol) to be called after decoding to do
 431      additional conversion, or nil.  */
 432   Lisp_Object post_read_conversion;
 433
 434   /* Lisp function (symbol) to be called before encoding to do
 435      additional conversion, or nil.  */
 436   Lisp_Object pre_write_conversion;
 437
 438   /* Character translation tables to look up, or nil.  */
 439   Lisp_Object translation_table_for_decode;
 440   Lisp_Object translation_table_for_encode;
 441 };
 442
 443 #define CODING_REQUIRE_FLUSHING_MASK    1
 444 #define CODING_REQUIRE_DECODING_MASK    2
 445 #define CODING_REQUIRE_ENCODING_MASK    4
 446 #define CODING_REQUIRE_DETECTION_MASK   8
 447
 448 /* Return 1 if the coding system CODING requires specific code to be
 449    attached at the tail of converted text.  */
 450 #define CODING_REQUIRE_FLUSHING(coding) \
 451   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 452
 453 /* Return 1 if the coding system CODING requires code conversion on
 454    decoding.  */
 455 #define CODING_REQUIRE_DECODING(coding) \
 456   ((coding)->dst_multibyte              \
 457    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 458
 459 /* Return 1 if the coding system CODING requires code conversion on
 460    encoding.  */
 461 #define CODING_REQUIRE_ENCODING(coding) \
 462   ((coding)->src_multibyte              \
 463    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 464
 465 /* Return 1 if the coding system CODING requires some kind of code
 466    detection.  */
 467 #define CODING_REQUIRE_DETECTION(coding) \
 468   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 469
 470 /* Return 1 if the coding system CODING requires code conversion on
 471    decoding or some kind of code detection.  */
 472 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 473   (CODING_REQUIRE_DECODING (coding)             \
 474    || CODING_REQUIRE_DETECTION (coding))
 475
 476 /* Index for each coding category in `coding_category_table' */
 477 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 478 #define CODING_CATEGORY_IDX_SJIS        1
 479 #define CODING_CATEGORY_IDX_ISO_7       2
 480 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 481 #define CODING_CATEGORY_IDX_ISO_8_1     4
 482 #define CODING_CATEGORY_IDX_ISO_8_2     5
 483 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 484 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 485 #define CODING_CATEGORY_IDX_CCL         8
 486 #define CODING_CATEGORY_IDX_BIG5        9
 487 #define CODING_CATEGORY_IDX_UTF_8       10
 488 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 489 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 490 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 491 #define CODING_CATEGORY_IDX_BINARY      14
 492 #define CODING_CATEGORY_IDX_MAX         15
 493
 494 /* Definitions of flag bits returned by the function
 495    detect_coding_mask ().  */
 496 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 497 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 498 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 499 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 500 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 501 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 502 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 503 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 504 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 505 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 506 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 507 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 508 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 509 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 510 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 511
 512 /* This value is returned if detect_coding_mask () find nothing other
 513    than ASCII characters.  */
 514 #define CODING_CATEGORY_MASK_ANY        \
 515   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 516    | CODING_CATEGORY_MASK_SJIS          \
 517    | CODING_CATEGORY_MASK_ISO_7         \
 518    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 519    | CODING_CATEGORY_MASK_ISO_8_1       \
 520    | CODING_CATEGORY_MASK_ISO_8_2       \
 521    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 522    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 523    | CODING_CATEGORY_MASK_CCL           \
 524    | CODING_CATEGORY_MASK_BIG5          \
 525    | CODING_CATEGORY_MASK_UTF_8         \
 526    | CODING_CATEGORY_MASK_UTF_16_BE     \
 527    | CODING_CATEGORY_MASK_UTF_16_LE)
 528
 529 #define CODING_CATEGORY_MASK_ISO_7BIT \
 530   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 531
 532 #define CODING_CATEGORY_MASK_ISO_8BIT \
 533   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 534
 535 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 536   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 537
 538 #define CODING_CATEGORY_MASK_ISO        \
 539   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 540    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 541    | CODING_CATEGORY_MASK_ISO_8BIT)
 542
 543 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 544    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 545
 546 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 547    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 548    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 549    internal format.  */
 550
 551 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 552   do {                                                  \
 553     if (s2 >= 0x9F)                                     \
 554       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 555       c2 = s2 - 0x7E;                                   \
 556     else                                                \
 557       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 558       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 559   } while (0)
 560
 561 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 562   do {                                                  \
 563     if (c1 & 1)                                         \
 564       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 565       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 566     else                                                \
 567       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 568       s2 = c2 + 0x7E;                                   \
 569   } while (0)
 570
 571 /* Encode the file name NAME using the specified coding system
 572    for file names, if any.  */
 573 #define ENCODE_FILE(name)                                                  \
 574   (! NILP (Vfile_name_coding_system)                                       \
 575    && XFASTINT (Vfile_name_coding_system) != 0                             \
 576    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 577    : (! NILP (Vdefault_file_name_coding_system)                            \
 578       && XFASTINT (Vdefault_file_name_coding_system) != 0                  \
 579       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 580       : name))
 581
 582 /* Decode the file name NAME using the specified coding system
 583    for file names, if any.  */
 584 #define DECODE_FILE(name)                                                  \
 585   (! NILP (Vfile_name_coding_system)                                       \
 586    && XFASTINT (Vfile_name_coding_system) != 0                             \
 587    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 588    : (! NILP (Vdefault_file_name_coding_system)                            \
 589       && XFASTINT (Vdefault_file_name_coding_system) != 0                  \
 590       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 591       : name))
 592
 593 #ifdef WINDOWSNT
 594 /* Encode the string STR using the specified coding system
 595    for w32 system functions, if any.  */
 596 #define ENCODE_SYSTEM(str)                                                 \
 597   (! NILP (Vlocale_coding_system)                                          \
 598    && XFASTINT (Vlocale_coding_system) != 0                                \
 599    ? code_convert_string_norecord (str, Vlocale_coding_system, 1)          \
 600    : str)
 601
 602 /* Decode the string STR using the specified coding system
 603    for w32 system functions, if any.  */
 604 #define DECODE_SYSTEM(name)                                                \
 605   (! NILP (Vlocale_coding_system)                                          \
 606    && XFASTINT (Vlocale_coding_system) != 0                                \
 607    ? code_convert_string_norecord (str, Vlocale_coding_system, 0)          \
 608    : str)
 609
 610 #else /* WINDOWSNT */
 611
 612 #define ENCODE_SYSTEM(str) string_make_unibyte(str)
 613 #define DECODE_SYSTEM(name) name
 614
 615 #endif /* !WINDOWSNT */
 616
 617 #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1)
 618
 619 /* Extern declarations.  */
 620 extern int decode_coding P_ ((struct coding_system *, const unsigned char *,
 621                               unsigned char *, int, int));
 622 extern int encode_coding P_ ((struct coding_system *, const unsigned char *,
 623                               unsigned char *, int, int));
 624 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 625                                          Lisp_Object));
 626 extern void coding_free_composition_data P_ ((struct coding_system *));
 627 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 628                                                   int));
 629 extern void coding_allocate_composition_data P_ ((struct coding_system *,
 630                                                   int));
 631 extern void coding_restore_composition P_ ((struct coding_system *,
 632                                             Lisp_Object));
 633 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 634                                     int, int));
 635 extern Lisp_Object run_pre_post_conversion_on_str P_ ((Lisp_Object,
 636                                                        struct coding_system *,
 637                                                        int));
 638 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 639 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 640 extern void detect_coding P_ ((struct coding_system *, const unsigned char *,
 641                                int));
 642 extern void detect_eol P_ ((struct coding_system *, const unsigned char *,
 643                             int));
 644 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 645 extern Lisp_Object code_convert_string P_ ((Lisp_Object,
 646                                             struct coding_system *, int, int));
 647 extern Lisp_Object code_convert_string1 P_ ((Lisp_Object, Lisp_Object,
 648                                              Lisp_Object, int));
 649 extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object,
 650                                                      int));
 651 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 652 extern Lisp_Object encode_coding_string P_ ((Lisp_Object,
 653                                              struct coding_system *, int));
 654 extern Lisp_Object decode_coding_string P_ ((Lisp_Object,
 655                                              struct coding_system *, int));
 656 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 657 extern Lisp_Object Qraw_text, Qemacs_mule;
 658 extern Lisp_Object Qbuffer_file_coding_system;
 659 extern Lisp_Object Vcoding_category_list;
 660 extern Lisp_Object Qutf_8;
 661
 662 extern Lisp_Object Qtranslation_table;
 663 extern Lisp_Object Qtranslation_table_id;
 664
 665 /* Mnemonic strings to indicate each type of end-of-line.  */
 666 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 667 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 668 extern Lisp_Object eol_mnemonic_undecided;
 669
 670 #ifdef emacs
 671 extern Lisp_Object Qfile_coding_system;
 672 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 673 extern Lisp_Object Qstart_process, Qopen_network_stream;
 674 extern Lisp_Object Qwrite_region;
 675
 676 extern char *emacs_strerror P_ ((int));
 677
 678 /* Coding-system for reading files and receiving data from process.  */
 679 extern Lisp_Object Vcoding_system_for_read;
 680 /* Coding-system for writing files and sending data to process.  */
 681 extern Lisp_Object Vcoding_system_for_write;
 682 /* Coding-system actually used in the latest I/O.  */
 683 extern Lisp_Object Vlast_coding_system_used;
 684 /* Coding-system to use with system messages (e.g. strerror).  */
 685 extern Lisp_Object Vlocale_coding_system;
 686
 687 /* If non-zero, process buffer inherits the coding system used to decode
 688    the subprocess output.  */
 689 extern int inherit_process_coding_system;
 690
 691 /* Coding-system to be used for encoding terminal output.  This
 692    structure contains information of a coding-system specified by the
 693    function `set-terminal-coding-system'.  */
 694 extern struct coding_system terminal_coding;
 695
 696 /* Coding system to be used to encode text for terminal display when
 697    terminal coding system is nil.  */
 698 extern struct coding_system safe_terminal_coding;
 699
 700 /* Coding-system of what is sent from terminal keyboard.  This
 701    structure contains information of a coding-system specified by the
 702    function `set-keyboard-coding-system'.  */
 703 extern struct coding_system keyboard_coding;
 704
 705 /* Default coding system to be used to write a file.  */
 706 extern struct coding_system default_buffer_file_coding;
 707
 708 /* Default coding systems used for process I/O.  */
 709 extern Lisp_Object Vdefault_process_coding_system;
 710
 711 /* Function to call to force a user to force select a proper coding
 712    system.  */
 713 extern Lisp_Object Vselect_safe_coding_system_function;
 714
 715 /* If nonzero, on writing a file, Vselect_safe_coding_system_function
 716    is called even if Vcoding_system_for_write is non-nil.  */
 717 extern int coding_system_require_warning;
 718
 719 /* Coding system for file names, or nil if none.  */
 720 extern Lisp_Object Vfile_name_coding_system;
 721
 722 /* Coding system for file names used only when
 723    Vfile_name_coding_system is nil.  */
 724 extern Lisp_Object Vdefault_file_name_coding_system;
 725
 726 #endif
 727
 728 /* Error signaled when there's a problem with detecting coding system */
 729 extern Lisp_Object Qcoding_system_error;
 730
 731 #endif /* EMACS_CODING_H */