code.delx.au - gnu-emacs/blob - src/coding.h

   1 /* Header for coding system handler.
   2    Copyright (C) 2004, 2005  Free Software Foundation, Inc.
   3    Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
   4    Licensed to the Free Software Foundation.
   5
   6 This file is part of GNU Emacs.
   7
   8 GNU Emacs is free software; you can redistribute it and/or modify
   9 it under the terms of the GNU General Public License as published by
  10 the Free Software Foundation; either version 2, or (at your option)
  11 any later version.
  12
  13 GNU Emacs is distributed in the hope that it will be useful,
  14 but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 GNU General Public License for more details.
  17
  18 You should have received a copy of the GNU General Public License
  19 along with GNU Emacs; see the file COPYING.  If not, write to
  20 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  21 Boston, MA 02110-1301, USA.  */
  22
  23 #ifndef EMACS_CODING_H
  24 #define EMACS_CODING_H
  25
  26 #include "ccl.h"
  27
  28 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  29
  30 /* All code (1-byte) of Emacs' internal format is classified into one
  31    of the followings.  See also `charset.h'.  */
  32 enum emacs_code_class_type
  33   {
  34     EMACS_control_code,         /* Control codes in the range
  35                                    0x00..0x1F and 0x7F except for the
  36                                    following two codes.  */
  37     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  38                                    end-of-line.  */
  39     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  40                                    in selective display mode.  */
  41     EMACS_ascii_code,           /* ASCII characters.  */
  42     EMACS_leading_code_2,       /* Base leading code of official
  43                                    TYPE9N character.  */
  44     EMACS_leading_code_3,       /* Base leading code of private TYPE9N
  45                                    or official TYPE9Nx9N character.  */
  46     EMACS_leading_code_4,       /* Base leading code of private
  47                                    TYPE9Nx9N character.  */
  48     EMACS_invalid_code          /* Invalid code, i.e. a base leading
  49                                    code not yet assigned to any
  50                                    charset, or a code of the range
  51                                    0xA0..0xFF.  */
  52   };
  53
  54 extern enum emacs_code_class_type emacs_code_class[256];
  55
  56 /*** ISO2022 section ***/
  57
  58 /* Macros to define code of control characters for ISO2022's functions.  */
  59                         /* code */      /* function */
  60 #define ISO_CODE_LF     0x0A            /* line-feed */
  61 #define ISO_CODE_CR     0x0D            /* carriage-return */
  62 #define ISO_CODE_SO     0x0E            /* shift-out */
  63 #define ISO_CODE_SI     0x0F            /* shift-in */
  64 #define ISO_CODE_SS2_7  0x19            /* single-shift-2 for 7-bit code */
  65 #define ISO_CODE_ESC    0x1B            /* escape */
  66 #define ISO_CODE_SS2    0x8E            /* single-shift-2 */
  67 #define ISO_CODE_SS3    0x8F            /* single-shift-3 */
  68 #define ISO_CODE_CSI    0x9B            /* control-sequence-introduce */
  69
  70 /* All code (1-byte) of ISO2022 is classified into one of the
  71    followings.  */
  72 enum iso_code_class_type
  73   {
  74     ISO_control_0,              /* Control codes in the range
  75                                    0x00..0x1F and 0x7F, except for the
  76                                    following 5 codes.  */
  77     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  78     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  79     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  80     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  81     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  82     ISO_control_1,              /* Control codes in the range
  83                                    0x80..0x9F, except for the
  84                                    following 3 codes.  */
  85     ISO_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  86     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  87     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  88     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  89     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  90     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  91     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  92   };
  93
  94 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  95   element in the structure `coding_system'.  This information is used
  96   while encoding a text to ISO2022.  **/
  97
  98 /* If set, produce short-form designation sequence (e.g. ESC $ A)
  99    instead of long-form sequence (e.g. ESC $ ( A).  */
 100 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 101
 102 /* If set, reset graphic planes and registers at end-of-line to the
 103    initial state.  */
 104 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 105
 106 /* If set, reset graphic planes and registers before any control
 107    characters to the initial state.  */
 108 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 109
 110 /* If set, encode by 7-bit environment.  */
 111 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 112
 113 /* If set, use locking-shift function.  */
 114 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 115
 116 /* If set, use single-shift function.  Overwrite
 117    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 118 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 119
 120 /* If set, designate JISX0201-Roman instead of ASCII.  */
 121 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 122
 123 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 124 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 125
 126 /* If set, do not produce ISO6429's direction specifying sequence.  */
 127 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 128
 129 /* If set, assume designation states are reset at beginning of line on
 130    output.  */
 131 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 132
 133 /* If set, designation sequence should be placed at beginning of line
 134    on output.  */
 135 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 136
 137 /* If set, do not encode unsafe characters on output.  */
 138 #define CODING_FLAG_ISO_SAFE            0x0800
 139
 140 /* If set, extra latin codes (128..159) are accepted as a valid code
 141    on input.  */
 142 #define CODING_FLAG_ISO_LATIN_EXTRA     0x1000
 143
 144 /* If set, use designation escape sequence.  */
 145 #define CODING_FLAG_ISO_DESIGNATION     0x10000
 146
 147 /* A character to be produced on output if encoding of the original
 148    character is inhibitted by CODING_MODE_INHIBIT_UNENCODABLE_CHAR.
 149    It must be an ASCII character.  */
 150 #define CODING_REPLACEMENT_CHARACTER '?'
 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 random 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 referring to 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 referring to 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 composing_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 structure has backward link to this structure in `prev'.
 313      The number of chained 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 `result' 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 finishing 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 /* If set, replace unencodabae characters by `?' on encoding.  */
 343 #define CODING_MODE_INHIBIT_UNENCODABLE_CHAR    0x08
 344
 345 /* This flag is used by the decoding/encoding routines on the fly.  If
 346    set, it means that right-to-left text is being processed.  */
 347 #define CODING_MODE_DIRECTION                   0x10
 348
 349 struct coding_system
 350 {
 351   /* Type of the coding system.  */
 352   enum coding_type type;
 353
 354   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 355   int eol_type;
 356
 357   /* Flag bits of the coding system.  The meaning of each bit is common
 358      to all types of coding systems.  */
 359   unsigned int common_flags;
 360
 361   /* Flag bits of the coding system.  The meaning of each bit depends
 362      on the type of the coding system.  */
 363   unsigned int flags;
 364
 365   /* Mode bits of the coding system.  See the comments of the macros
 366      CODING_MODE_XXX.  */
 367   unsigned int mode;
 368
 369   /* The current status of composition handling.  */
 370   int composing;
 371
 372   /* 1 iff the next character is a composition rule.  */
 373   int composition_rule_follows;
 374
 375   /* Information of compositions are stored here on decoding and set
 376      in advance on encoding.  */
 377   struct composition_data *cmp_data;
 378
 379   /* Index to cmp_data->data for the first element for the current
 380      composition.  */
 381   int cmp_data_start;
 382
 383   /* Index to cmp_data->data for the current element for the current
 384      composition.  */
 385   int cmp_data_index;
 386
 387   /* Detailed information specific to each type of coding system.  */
 388   union spec
 389     {
 390       struct iso2022_spec iso2022;
 391       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 392     } spec;
 393
 394   /* Index number of coding category of the coding system.  */
 395   int category_idx;
 396
 397   /* The following two members specify how characters 128..159 are
 398      represented in source and destination text respectively.  1 means
 399      they are represented by 2-byte sequence, 0 means they are
 400      represented by 1-byte as is (see the comment in charset.h).  */
 401   unsigned src_multibyte : 1;
 402   unsigned dst_multibyte : 1;
 403
 404   /* How may heading bytes we can skip for decoding.  This is set to
 405      -1 in setup_coding_system, and updated by detect_coding.  So,
 406      when this is equal to the byte length of the text being
 407      converted, we can skip the actual conversion process.  */
 408   int heading_ascii;
 409
 410   /* The following members are set by encoding/decoding routine.  */
 411   int produced, produced_char, consumed, consumed_char;
 412
 413   /* Number of error source data found in a decoding routine.  */
 414   int errors;
 415
 416   /* Finish status of code conversion.  It should be one of macros
 417      CODING_FINISH_XXXX.  */
 418   int result;
 419
 420   /* If nonzero, suppress error notification.  */
 421   int suppress_error;
 422
 423   /* The following members are all Lisp symbols.  We don't have to
 424      protect them from GC because the current garbage collection
 425      doesn't relocate Lisp symbols.  But, when it is changed, we must
 426      find a way to protect them.  */
 427
 428   /* Backward pointer to the Lisp symbol of the coding system.  */
 429   Lisp_Object symbol;
 430
 431   /* Lisp function (symbol) to be called after decoding to do
 432      additional conversion, or nil.  */
 433   Lisp_Object post_read_conversion;
 434
 435   /* Lisp function (symbol) to be called before encoding to do
 436      additional conversion, or nil.  */
 437   Lisp_Object pre_write_conversion;
 438
 439   /* Character translation tables to look up, or nil.  */
 440   Lisp_Object translation_table_for_decode;
 441   Lisp_Object translation_table_for_encode;
 442 };
 443
 444 #define CODING_REQUIRE_FLUSHING_MASK    1
 445 #define CODING_REQUIRE_DECODING_MASK    2
 446 #define CODING_REQUIRE_ENCODING_MASK    4
 447 #define CODING_REQUIRE_DETECTION_MASK   8
 448
 449 /* Return 1 if the coding system CODING requires specific code to be
 450    attached at the tail of converted text.  */
 451 #define CODING_REQUIRE_FLUSHING(coding) \
 452   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 453
 454 /* Return 1 if the coding system CODING requires code conversion on
 455    decoding.  */
 456 #define CODING_REQUIRE_DECODING(coding) \
 457   ((coding)->dst_multibyte              \
 458    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 459
 460 /* Return 1 if the coding system CODING requires code conversion on
 461    encoding.  */
 462 #define CODING_REQUIRE_ENCODING(coding) \
 463   ((coding)->src_multibyte              \
 464    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 465
 466 /* Return 1 if the coding system CODING requires some kind of code
 467    detection.  */
 468 #define CODING_REQUIRE_DETECTION(coding) \
 469   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 470
 471 /* Return 1 if the coding system CODING requires code conversion on
 472    decoding or some kind of code detection.  */
 473 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 474   (CODING_REQUIRE_DECODING (coding)             \
 475    || CODING_REQUIRE_DETECTION (coding))
 476
 477 /* Index for each coding category in `coding_category_table' */
 478 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 479 #define CODING_CATEGORY_IDX_SJIS        1
 480 #define CODING_CATEGORY_IDX_ISO_7       2
 481 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 482 #define CODING_CATEGORY_IDX_ISO_8_1     4
 483 #define CODING_CATEGORY_IDX_ISO_8_2     5
 484 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 485 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 486 #define CODING_CATEGORY_IDX_CCL         8
 487 #define CODING_CATEGORY_IDX_BIG5        9
 488 #define CODING_CATEGORY_IDX_UTF_8       10
 489 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 490 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 491 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 492 #define CODING_CATEGORY_IDX_BINARY      14
 493 #define CODING_CATEGORY_IDX_MAX         15
 494
 495 /* Definitions of flag bits returned by the function
 496    detect_coding_mask ().  */
 497 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 498 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 499 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 500 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 501 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 502 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 503 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 504 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 505 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 506 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 507 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 508 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 509 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 510 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 511 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 512
 513 /* This value is returned if detect_coding_mask () find nothing other
 514    than ASCII characters.  */
 515 #define CODING_CATEGORY_MASK_ANY        \
 516   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 517    | CODING_CATEGORY_MASK_SJIS          \
 518    | CODING_CATEGORY_MASK_ISO_7         \
 519    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 520    | CODING_CATEGORY_MASK_ISO_8_1       \
 521    | CODING_CATEGORY_MASK_ISO_8_2       \
 522    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 523    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 524    | CODING_CATEGORY_MASK_CCL           \
 525    | CODING_CATEGORY_MASK_BIG5          \
 526    | CODING_CATEGORY_MASK_UTF_8         \
 527    | CODING_CATEGORY_MASK_UTF_16_BE     \
 528    | CODING_CATEGORY_MASK_UTF_16_LE)
 529
 530 #define CODING_CATEGORY_MASK_ISO_7BIT \
 531   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 532
 533 #define CODING_CATEGORY_MASK_ISO_8BIT \
 534   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 535
 536 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 537   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 538
 539 #define CODING_CATEGORY_MASK_ISO        \
 540   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 541    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 542    | CODING_CATEGORY_MASK_ISO_8BIT)
 543
 544 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 545    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 546
 547 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 548    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 549    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 550    internal format.  */
 551
 552 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 553   do {                                                  \
 554     if (s2 >= 0x9F)                                     \
 555       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 556       c2 = s2 - 0x7E;                                   \
 557     else                                                \
 558       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 559       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 560   } while (0)
 561
 562 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 563   do {                                                  \
 564     if (c1 & 1)                                         \
 565       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 566       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 567     else                                                \
 568       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 569       s2 = c2 + 0x7E;                                   \
 570   } while (0)
 571
 572 /* Encode the file name NAME using the specified coding system
 573    for file names, if any.  */
 574 #define ENCODE_FILE(name)                                                  \
 575   (! NILP (Vfile_name_coding_system)                                       \
 576    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 577    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 578    : (! NILP (Vdefault_file_name_coding_system)                            \
 579       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 580       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 581       : name))
 582
 583 /* Decode the file name NAME using the specified coding system
 584    for file names, if any.  */
 585 #define DECODE_FILE(name)                                                  \
 586   (! NILP (Vfile_name_coding_system)                                       \
 587    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 588    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 589    : (! NILP (Vdefault_file_name_coding_system)                            \
 590       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 591       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 592       : name))
 593
 594 /* Encode the string STR using the specified coding system
 595    for system functions, if any.  */
 596 #define ENCODE_SYSTEM(str)                                                 \
 597   (! NILP (Vlocale_coding_system)                                          \
 598    && !EQ (Vlocale_coding_system, make_number (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 system functions, if any.  */
 604 #define DECODE_SYSTEM(name)                                                \
 605   (! NILP (Vlocale_coding_system)                                          \
 606    && !EQ (Vlocale_coding_system, make_number (0))                         \
 607    ? code_convert_string_norecord (str, Vlocale_coding_system, 0)          \
 608    : str)
 609
 610 #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1)
 611
 612 /* Extern declarations.  */
 613 extern int decode_coding P_ ((struct coding_system *, const unsigned char *,
 614                               unsigned char *, int, int));
 615 extern int encode_coding P_ ((struct coding_system *, const unsigned char *,
 616                               unsigned char *, int, int));
 617 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 618                                          Lisp_Object));
 619 extern void coding_free_composition_data P_ ((struct coding_system *));
 620 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 621                                                   int));
 622 extern void coding_allocate_composition_data P_ ((struct coding_system *,
 623                                                   int));
 624 extern void coding_restore_composition P_ ((struct coding_system *,
 625                                             Lisp_Object));
 626 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 627                                     int, int));
 628 extern Lisp_Object run_pre_post_conversion_on_str P_ ((Lisp_Object,
 629                                                        struct coding_system *,
 630                                                        int));
 631 extern void run_pre_write_conversin_on_c_str P_ ((unsigned char **, int *,
 632                                                   int, int,
 633                                                   struct coding_system *));
 634
 635 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 636 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 637 extern void detect_coding P_ ((struct coding_system *, const unsigned char *,
 638                                int));
 639 extern void detect_eol P_ ((struct coding_system *, const unsigned char *,
 640                             int));
 641 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 642 extern Lisp_Object code_convert_string P_ ((Lisp_Object,
 643                                             struct coding_system *, int, int));
 644 extern Lisp_Object code_convert_string1 P_ ((Lisp_Object, Lisp_Object,
 645                                              Lisp_Object, int));
 646 extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object,
 647                                                      int));
 648 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 649 extern Lisp_Object encode_coding_string P_ ((Lisp_Object,
 650                                              struct coding_system *, int));
 651 extern Lisp_Object decode_coding_string P_ ((Lisp_Object,
 652                                              struct coding_system *, int));
 653 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 654 extern Lisp_Object Qraw_text, Qemacs_mule;
 655 extern Lisp_Object Qbuffer_file_coding_system;
 656 extern Lisp_Object Vcoding_category_list;
 657 extern Lisp_Object Qutf_8;
 658
 659 extern Lisp_Object Qtranslation_table;
 660 extern Lisp_Object Qtranslation_table_id;
 661
 662 /* Mnemonic strings to indicate each type of end-of-line.  */
 663 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 664 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 665 extern Lisp_Object eol_mnemonic_undecided;
 666
 667 #ifdef emacs
 668 extern Lisp_Object Qfile_coding_system;
 669 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 670 extern Lisp_Object Qstart_process, Qopen_network_stream;
 671 extern Lisp_Object Qwrite_region;
 672
 673 extern char *emacs_strerror P_ ((int));
 674
 675 /* Coding-system for reading files and receiving data from process.  */
 676 extern Lisp_Object Vcoding_system_for_read;
 677 /* Coding-system for writing files and sending data to process.  */
 678 extern Lisp_Object Vcoding_system_for_write;
 679 /* Coding-system actually used in the latest I/O.  */
 680 extern Lisp_Object Vlast_coding_system_used;
 681 /* Coding-system to use with system messages (e.g. strerror).  */
 682 extern Lisp_Object Vlocale_coding_system;
 683
 684 /* If non-zero, process buffer inherits the coding system used to decode
 685    the subprocess output.  */
 686 extern int inherit_process_coding_system;
 687
 688 /* Coding system to be used to encode text for terminal display when
 689    terminal coding system is nil.  */
 690 extern struct coding_system safe_terminal_coding;
 691
 692 /* Default coding system to be used to write a file.  */
 693 extern struct coding_system default_buffer_file_coding;
 694
 695 /* Default coding systems used for process I/O.  */
 696 extern Lisp_Object Vdefault_process_coding_system;
 697
 698 /* Function to call to force a user to force select a proper coding
 699    system.  */
 700 extern Lisp_Object Vselect_safe_coding_system_function;
 701
 702 /* If nonzero, on writing a file, Vselect_safe_coding_system_function
 703    is called even if Vcoding_system_for_write is non-nil.  */
 704 extern int coding_system_require_warning;
 705
 706 /* Coding system for file names, or nil if none.  */
 707 extern Lisp_Object Vfile_name_coding_system;
 708
 709 /* Coding system for file names used only when
 710    Vfile_name_coding_system is nil.  */
 711 extern Lisp_Object Vdefault_file_name_coding_system;
 712
 713 #endif
 714
 715 /* Error signaled when there's a problem with detecting coding system */
 716 extern Lisp_Object Qcoding_system_error;
 717
 718 #endif /* EMACS_CODING_H */
 719
 720 /* arch-tag: 2bc3b4fa-6870-4f64-8135-b962b2d290e4
 721    (do not change this comment) */