src/coding.h

   1 /* Header for coding system handler.
   2    Copyright (C) 2001, 2002, 2003, 2004, 2005,
   3                  2006, 2007  Free Software Foundation, Inc.
   4    Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   5      2005, 2006, 2007
   6      National Institute of Advanced Industrial Science and Technology (AIST)
   7      Registration Number H14PRO021
   8
   9 This file is part of GNU Emacs.
  10
  11 GNU Emacs is free software; you can redistribute it and/or modify
  12 it under the terms of the GNU General Public License as published by
  13 the Free Software Foundation; either version 3, or (at your option)
  14 any later version.
  15
  16 GNU Emacs is distributed in the hope that it will be useful,
  17 but WITHOUT ANY WARRANTY; without even the implied warranty of
  18 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 GNU General Public License for more details.
  20
  21 You should have received a copy of the GNU General Public License
  22 along with GNU Emacs; see the file COPYING.  If not, write to
  23 the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  24 Boston, MA 02110-1301, USA.  */
  25
  26 #ifndef EMACS_CODING_H
  27 #define EMACS_CODING_H
  28
  29 #include "ccl.h"
  30
  31 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  32
  33 /* All code (1-byte) of Emacs' internal format is classified into one
  34    of the followings.  See also `charset.h'.  */
  35 enum emacs_code_class_type
  36   {
  37     EMACS_control_code,         /* Control codes in the range
  38                                    0x00..0x1F and 0x7F except for the
  39                                    following two codes.  */
  40     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  41                                    end-of-line.  */
  42     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  43                                    in selective display mode.  */
  44     EMACS_ascii_code,           /* ASCII characters.  */
  45     EMACS_leading_code_2,       /* Base leading code of official
  46                                    TYPE9N character.  */
  47     EMACS_leading_code_3,       /* Base leading code of private TYPE9N
  48                                    or official TYPE9Nx9N character.  */
  49     EMACS_leading_code_4,       /* Base leading code of private
  50                                    TYPE9Nx9N character.  */
  51     EMACS_invalid_code          /* Invalid code, i.e. a base leading
  52                                    code not yet assigned to any
  53                                    charset, or a code of the range
  54                                    0xA0..0xFF.  */
  55   };
  56
  57 extern enum emacs_code_class_type emacs_code_class[256];
  58
  59 /*** ISO2022 section ***/
  60
  61 /* Macros to define code of control characters for ISO2022's functions.  */
  62                         /* code */      /* function */
  63 #define ISO_CODE_LF     0x0A            /* line-feed */
  64 #define ISO_CODE_CR     0x0D            /* carriage-return */
  65 #define ISO_CODE_SO     0x0E            /* shift-out */
  66 #define ISO_CODE_SI     0x0F            /* shift-in */
  67 #define ISO_CODE_SS2_7  0x19            /* single-shift-2 for 7-bit code */
  68 #define ISO_CODE_ESC    0x1B            /* escape */
  69 #define ISO_CODE_SS2    0x8E            /* single-shift-2 */
  70 #define ISO_CODE_SS3    0x8F            /* single-shift-3 */
  71 #define ISO_CODE_CSI    0x9B            /* control-sequence-introduce */
  72
  73 /* All code (1-byte) of ISO2022 is classified into one of the
  74    followings.  */
  75 enum iso_code_class_type
  76   {
  77     ISO_control_0,              /* Control codes in the range
  78                                    0x00..0x1F and 0x7F, except for the
  79                                    following 5 codes.  */
  80     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  81     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  82     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  83     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  84     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  85     ISO_control_1,              /* Control codes in the range
  86                                    0x80..0x9F, except for the
  87                                    following 3 codes.  */
  88     ISO_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  89     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  90     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  91     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  92     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  93     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  94     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  95   };
  96
  97 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  98   element in the structure `coding_system'.  This information is used
  99   while encoding a text to ISO2022.  **/
 100
 101 /* If set, produce short-form designation sequence (e.g. ESC $ A)
 102    instead of long-form sequence (e.g. ESC $ ( A).  */
 103 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 104
 105 /* If set, reset graphic planes and registers at end-of-line to the
 106    initial state.  */
 107 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 108
 109 /* If set, reset graphic planes and registers before any control
 110    characters to the initial state.  */
 111 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 112
 113 /* If set, encode by 7-bit environment.  */
 114 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 115
 116 /* If set, use locking-shift function.  */
 117 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 118
 119 /* If set, use single-shift function.  Overwrite
 120    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 121 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 122
 123 /* If set, designate JISX0201-Roman instead of ASCII.  */
 124 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 125
 126 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 127 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 128
 129 /* If set, do not produce ISO6429's direction specifying sequence.  */
 130 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 131
 132 /* If set, assume designation states are reset at beginning of line on
 133    output.  */
 134 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 135
 136 /* If set, designation sequence should be placed at beginning of line
 137    on output.  */
 138 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 139
 140 /* If set, do not encode unsafe characters on output.  */
 141 #define CODING_FLAG_ISO_SAFE            0x0800
 142
 143 /* If set, extra latin codes (128..159) are accepted as a valid code
 144    on input.  */
 145 #define CODING_FLAG_ISO_LATIN_EXTRA     0x1000
 146
 147 /* If set, use designation escape sequence.  */
 148 #define CODING_FLAG_ISO_DESIGNATION     0x10000
 149
 150 /* A character to be produced on output if encoding of the original
 151    character is inhibitted by CODING_MODE_INHIBIT_UNENCODABLE_CHAR.
 152    It must be an ASCII character.  */
 153 #define CODING_REPLACEMENT_CHARACTER '?'
 154
 155 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 156 struct iso2022_spec
 157 {
 158   /* The current graphic register invoked to each graphic plane.  */
 159   int current_invocation[2];
 160
 161   /* The current charset designated to each graphic register.  */
 162   int current_designation[4];
 163
 164   /* A charset initially designated to each graphic register.  */
 165   int initial_designation[4];
 166
 167   /* If not -1, it is a graphic register specified in an invalid
 168      designation sequence.  */
 169   int last_invalid_designation_register;
 170
 171   /* A graphic register to which each charset should be designated.  */
 172   unsigned char requested_designation[MAX_CHARSET + 1];
 173
 174   /* A revision number to be specified for each charset on encoding.
 175      The value 255 means no revision number for the corresponding
 176      charset.  */
 177   unsigned char charset_revision_number[MAX_CHARSET + 1];
 178
 179   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 180      by single-shift while encoding.  */
 181   int single_shifting;
 182
 183   /* Set to 1 temporarily only when processing at beginning of line.  */
 184   int bol;
 185 };
 186
 187 /* Macros to access each field in the structure `spec.iso2022'.  */
 188 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 189   (coding)->spec.iso2022.current_invocation[plane]
 190 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 191   (coding)->spec.iso2022.current_designation[reg]
 192 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 193   (coding)->spec.iso2022.initial_designation[reg]
 194 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 195   (coding)->spec.iso2022.requested_designation[charset]
 196 #define CODING_SPEC_ISO_REVISION_NUMBER(coding, charset) \
 197   (coding)->spec.iso2022.charset_revision_number[charset]
 198 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 199   (coding)->spec.iso2022.single_shifting
 200 #define CODING_SPEC_ISO_BOL(coding) \
 201   (coding)->spec.iso2022.bol
 202
 203 /* A value which may appear in
 204    coding->spec.iso2022.requested_designation indicating that the
 205    corresponding charset does not request any graphic register to be
 206    designated.  */
 207 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 208
 209 /* Return a charset which is currently designated to the graphic plane
 210    PLANE in the coding-system CODING.  */
 211 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 212   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 213    ? -1                                                 \
 214    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 215                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 216
 217 /*** BIG5 section ***/
 218
 219 /* Macros to denote each type of BIG5 coding system.  */
 220 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 221                                         BIG5 developed by Hong Kong
 222                                         University.  */
 223 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 224                                         of BIG5 developed by the
 225                                         company ETen in Taiwan.  */
 226
 227 /*** GENERAL section ***/
 228
 229 /* Types of coding system.  */
 230 enum coding_type
 231   {
 232     coding_type_no_conversion,  /* A coding system which requires no
 233                                    conversion for reading and writing
 234                                    including end-of-line format.  */
 235     coding_type_emacs_mule,     /* A coding system used in Emacs'
 236                                    buffer and string.  Requires no
 237                                    conversion for reading and writing
 238                                    except for end-of-line format.  */
 239     coding_type_undecided,      /* A coding system which requires
 240                                    automatic detection of a real
 241                                    coding system.  */
 242     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 243     coding_type_iso2022,        /* Any coding system of ISO2022
 244                                    variants.  */
 245     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 246     coding_type_ccl,            /* The coding system of which decoder
 247                                    and encoder are written in CCL.  */
 248     coding_type_raw_text        /* A coding system for a text
 249                                    containing random 8-bit code which
 250                                    does not require code conversion
 251                                    except for end-of-line format. */
 252   };
 253
 254 /* Formats of end-of-line.  */
 255 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 256                                    internal format.  */
 257 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 258                                    line-feed.  */
 259 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 260 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 261                                    eol-type is not yet decided.  */
 262 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 263                                      eol-type is not consistent
 264                                      through the file.  */
 265
 266 /* 1 if composing.  */
 267 #define COMPOSING_P(coding) ((int) coding->composing > (int) COMPOSITION_NO)
 268
 269 #define COMPOSITION_DATA_SIZE 4080
 270 #define COMPOSITION_DATA_MAX_BUNCH_LENGTH (4 + MAX_COMPOSITION_COMPONENTS*2)
 271
 272 /* Data structure to hold information about compositions of text that
 273    is being decoded or encode.  ISO 2022 base code conversion routines
 274    handle special ESC sequences for composition specification.  But,
 275    they can't get/put such information directly from/to a buffer in
 276    the deepest place.  So, they store or retrieve the information
 277    through this structure.
 278
 279    The encoder stores the information in this structure when it meets
 280    ESC sequences for composition while encoding codes, then, after all
 281    text codes are encoded, puts `composition' properties on the text
 282    by referring to the structure.
 283
 284    The decoder at first stores the information of a text to be
 285    decoded, then, while decoding codes, generates ESC sequences for
 286    composition at proper places by referring to the structure.  */
 287
 288 struct composition_data
 289 {
 290   /* The character position of the first character to be encoded or
 291      decoded.  START and END (see below) are relative to this
 292      position.  */
 293   int char_offset;
 294
 295   /* The composition data.  These elements are repeated for each
 296      composition:
 297         LENGTH START END METHOD [ COMPONENT ... ]
 298      where,
 299         LENGTH is the number of elements for this composition.
 300
 301         START and END are starting and ending character positions of
 302         the composition relative to `char_offset'.
 303
 304         METHOD is one of `enum composing_status' specifying the way of
 305         composition.
 306
 307         COMPONENT is a character or an encoded composition rule.  */
 308   int data[COMPOSITION_DATA_SIZE];
 309
 310   /* The number of elements in `data' currently used.  */
 311   int used;
 312
 313   /* Pointers to the previous and next structures.  When `data' is
 314      filled up, another structure is allocated and linked in `next'.
 315      The new structure has backward link to this structure in `prev'.
 316      The number of chained structures depends on how many compositions
 317      the text being encoded or decoded contains.  */
 318   struct composition_data *prev, *next;
 319 };
 320
 321 /* Macros used for the member `result' of the struct
 322    coding_system.  */
 323 #define CODING_FINISH_NORMAL            0
 324 #define CODING_FINISH_INSUFFICIENT_SRC  1
 325 #define CODING_FINISH_INSUFFICIENT_DST  2
 326 #define CODING_FINISH_INCONSISTENT_EOL  3
 327 #define CODING_FINISH_INSUFFICIENT_CMP  4
 328 #define CODING_FINISH_INTERRUPT         5
 329
 330 /* Macros used for the member `mode' of the struct coding_system.  */
 331
 332 /* If set, recover the original CR or LF of the already decoded text
 333    when the decoding routine encounters an inconsistent eol format.  */
 334 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 335
 336 /* If set, the decoding/encoding routines treat the current data as
 337    the last block of the whole text to be converted, and do
 338    appropriate finishing job.  */
 339 #define CODING_MODE_LAST_BLOCK                  0x02
 340
 341 /* If set, it means that the current source text is in a buffer which
 342    enables selective display.  */
 343 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 344
 345 /* If set, replace unencodabae characters by `?' on encoding.  */
 346 #define CODING_MODE_INHIBIT_UNENCODABLE_CHAR    0x08
 347
 348 /* This flag is used by the decoding/encoding routines on the fly.  If
 349    set, it means that right-to-left text is being processed.  */
 350 #define CODING_MODE_DIRECTION                   0x10
 351
 352 struct coding_system
 353 {
 354   /* Type of the coding system.  */
 355   enum coding_type type;
 356
 357   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 358   int eol_type;
 359
 360   /* Flag bits of the coding system.  The meaning of each bit is common
 361      to all types of coding systems.  */
 362   unsigned int common_flags;
 363
 364   /* Flag bits of the coding system.  The meaning of each bit depends
 365      on the type of the coding system.  */
 366   unsigned int flags;
 367
 368   /* Mode bits of the coding system.  See the comments of the macros
 369      CODING_MODE_XXX.  */
 370   unsigned int mode;
 371
 372   /* The current status of composition handling.  */
 373   int composing;
 374
 375   /* 1 if 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   /* The following two members specify how characters 128..159 are
 401      represented in source and destination text respectively.  1 means
 402      they are represented by 2-byte sequence, 0 means they are
 403      represented by 1-byte as is (see the comment in charset.h).  */
 404   unsigned src_multibyte : 1;
 405   unsigned dst_multibyte : 1;
 406
 407   /* How may heading bytes we can skip for decoding.  This is set to
 408      -1 in setup_coding_system, and updated by detect_coding.  So,
 409      when this is equal to the byte length of the text being
 410      converted, we can skip the actual conversion process.  */
 411   int heading_ascii;
 412
 413   /* The following members are set by encoding/decoding routine.  */
 414   int produced, produced_char, consumed, consumed_char;
 415
 416   /* Number of error source data found in a decoding routine.  */
 417   int errors;
 418
 419   /* Finish status of code conversion.  It should be one of macros
 420      CODING_FINISH_XXXX.  */
 421   int result;
 422
 423   /* If nonzero, suppress error notification.  */
 424   int suppress_error;
 425
 426   /* The following members are all Lisp symbols.  We don't have to
 427      protect them from GC because the current garbage collection
 428      doesn't relocate Lisp symbols.  But, when it is changed, we must
 429      find a way to protect them.  */
 430
 431   /* Backward pointer to the Lisp symbol of the coding system.  */
 432   Lisp_Object symbol;
 433
 434   /* Lisp function (symbol) to be called after decoding to do
 435      additional conversion, or nil.  */
 436   Lisp_Object post_read_conversion;
 437
 438   /* Lisp function (symbol) to be called before encoding to do
 439      additional conversion, or nil.  */
 440   Lisp_Object pre_write_conversion;
 441
 442   /* Character translation tables to look up, or nil.  */
 443   Lisp_Object translation_table_for_decode;
 444   Lisp_Object translation_table_for_encode;
 445 };
 446
 447 /* Mask bits for (struct coding_system *)->common_flags.  */
 448 #define CODING_REQUIRE_FLUSHING_MASK    0x01
 449 #define CODING_REQUIRE_DECODING_MASK    0x02
 450 #define CODING_REQUIRE_ENCODING_MASK    0x04
 451 #define CODING_REQUIRE_DETECTION_MASK   0x08
 452 #define CODING_ASCII_INCOMPATIBLE_MASK  0x10
 453
 454 /* Return 1 if the coding system CODING requires specific code to be
 455    attached at the tail of converted text.  */
 456 #define CODING_REQUIRE_FLUSHING(coding) \
 457   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 458
 459 /* Return 1 if the coding system CODING requires code conversion on
 460    decoding.  */
 461 #define CODING_REQUIRE_DECODING(coding) \
 462   ((coding)->dst_multibyte              \
 463    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 464
 465 /* Return 1 if the coding system CODING requires code conversion on
 466    encoding.
 467    The non-multibyte part of the condition is to support encoding of
 468    unibyte strings/buffers generated by string-as-unibyte or
 469    (set-buffer-multibyte nil) from multibyte strings/buffers.  */
 470 #define CODING_REQUIRE_ENCODING(coding) \
 471   ((coding)->src_multibyte              \
 472    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK)
 473
 474 /* Return 1 if the coding system CODING requires some kind of code
 475    detection.  */
 476 #define CODING_REQUIRE_DETECTION(coding) \
 477   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 478
 479 /* Return 1 if the coding system CODING requires code conversion on
 480    decoding or some kind of code detection.  */
 481 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 482   (CODING_REQUIRE_DECODING (coding)             \
 483    || CODING_REQUIRE_DETECTION (coding))
 484
 485 /* Index for each coding category in `coding_category_table' */
 486 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 487 #define CODING_CATEGORY_IDX_SJIS        1
 488 #define CODING_CATEGORY_IDX_ISO_7       2
 489 #define CODING_CATEGORY_IDX_ISO_7_TIGHT 3
 490 #define CODING_CATEGORY_IDX_ISO_8_1     4
 491 #define CODING_CATEGORY_IDX_ISO_8_2     5
 492 #define CODING_CATEGORY_IDX_ISO_7_ELSE  6
 493 #define CODING_CATEGORY_IDX_ISO_8_ELSE  7
 494 #define CODING_CATEGORY_IDX_CCL         8
 495 #define CODING_CATEGORY_IDX_BIG5        9
 496 #define CODING_CATEGORY_IDX_UTF_8       10
 497 #define CODING_CATEGORY_IDX_UTF_16_BE   11
 498 #define CODING_CATEGORY_IDX_UTF_16_LE   12
 499 #define CODING_CATEGORY_IDX_RAW_TEXT    13
 500 #define CODING_CATEGORY_IDX_BINARY      14
 501 #define CODING_CATEGORY_IDX_MAX         15
 502
 503 /* Definitions of flag bits returned by the function
 504    detect_coding_mask ().  */
 505 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 506 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 507 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 508 #define CODING_CATEGORY_MASK_ISO_7_TIGHT (1 << CODING_CATEGORY_IDX_ISO_7_TIGHT)
 509 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 510 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 511 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 512 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 513 #define CODING_CATEGORY_MASK_CCL        (1 << CODING_CATEGORY_IDX_CCL)
 514 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 515 #define CODING_CATEGORY_MASK_UTF_8      (1 << CODING_CATEGORY_IDX_UTF_8)
 516 #define CODING_CATEGORY_MASK_UTF_16_BE  (1 << CODING_CATEGORY_IDX_UTF_16_BE)
 517 #define CODING_CATEGORY_MASK_UTF_16_LE  (1 << CODING_CATEGORY_IDX_UTF_16_LE)
 518 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 519 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 520
 521 /* This value is returned if detect_coding_mask () find nothing other
 522    than ASCII characters.  */
 523 #define CODING_CATEGORY_MASK_ANY        \
 524   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 525    | CODING_CATEGORY_MASK_SJIS          \
 526    | CODING_CATEGORY_MASK_ISO_7         \
 527    | CODING_CATEGORY_MASK_ISO_7_TIGHT   \
 528    | CODING_CATEGORY_MASK_ISO_8_1       \
 529    | CODING_CATEGORY_MASK_ISO_8_2       \
 530    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 531    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 532    | CODING_CATEGORY_MASK_CCL           \
 533    | CODING_CATEGORY_MASK_BIG5          \
 534    | CODING_CATEGORY_MASK_UTF_8         \
 535    | CODING_CATEGORY_MASK_UTF_16_BE     \
 536    | CODING_CATEGORY_MASK_UTF_16_LE)
 537
 538 #define CODING_CATEGORY_MASK_ISO_7BIT \
 539   (CODING_CATEGORY_MASK_ISO_7 | CODING_CATEGORY_MASK_ISO_7_TIGHT)
 540
 541 #define CODING_CATEGORY_MASK_ISO_8BIT \
 542   (CODING_CATEGORY_MASK_ISO_8_1 | CODING_CATEGORY_MASK_ISO_8_2)
 543
 544 #define CODING_CATEGORY_MASK_ISO_SHIFT \
 545   (CODING_CATEGORY_MASK_ISO_7_ELSE | CODING_CATEGORY_MASK_ISO_8_ELSE)
 546
 547 #define CODING_CATEGORY_MASK_ISO        \
 548   (  CODING_CATEGORY_MASK_ISO_7BIT      \
 549    | CODING_CATEGORY_MASK_ISO_SHIFT     \
 550    | CODING_CATEGORY_MASK_ISO_8BIT)
 551
 552 #define CODING_CATEGORY_MASK_UTF_16_BE_LE \
 553    (CODING_CATEGORY_MASK_UTF_16_BE | CODING_CATEGORY_MASK_UTF_16_LE)
 554
 555 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 556    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 557    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 558    internal format.  */
 559
 560 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 561   do {                                                  \
 562     if (s2 >= 0x9F)                                     \
 563       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 564       c2 = s2 - 0x7E;                                   \
 565     else                                                \
 566       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 567       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 568   } while (0)
 569
 570 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 571   do {                                                  \
 572     if (c1 & 1)                                         \
 573       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 574       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 575     else                                                \
 576       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 577       s2 = c2 + 0x7E;                                   \
 578   } while (0)
 579
 580 /* Encode the file name NAME using the specified coding system
 581    for file names, if any.  */
 582 #define ENCODE_FILE(name)                                                  \
 583   (! NILP (Vfile_name_coding_system)                                       \
 584    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 585    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 586    : (! NILP (Vdefault_file_name_coding_system)                            \
 587       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 588       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 589       : name))
 590
 591 /* Decode the file name NAME using the specified coding system
 592    for file names, if any.  */
 593 #define DECODE_FILE(name)                                                  \
 594   (! NILP (Vfile_name_coding_system)                                       \
 595    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 596    ? code_convert_string_norecord (name, Vfile_name_coding_system, 0)      \
 597    : (! NILP (Vdefault_file_name_coding_system)                            \
 598       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 599       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 0) \
 600       : name))
 601
 602 /* Encode the string STR using the specified coding system
 603    for system functions, if any.  */
 604 #define ENCODE_SYSTEM(str)                                                 \
 605   (! NILP (Vlocale_coding_system)                                          \
 606    && !EQ (Vlocale_coding_system, make_number (0))                         \
 607    ? code_convert_string_norecord (str, Vlocale_coding_system, 1)          \
 608    : str)
 609
 610 /* Decode the string STR using the specified coding system
 611    for system functions, if any.  */
 612 #define DECODE_SYSTEM(str)                                                 \
 613   (! NILP (Vlocale_coding_system)                                          \
 614    && !EQ (Vlocale_coding_system, make_number (0))                         \
 615    ? code_convert_string_norecord (str, Vlocale_coding_system, 0)          \
 616    : str)
 617
 618 #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1)
 619
 620 /* Extern declarations.  */
 621 extern int decode_coding P_ ((struct coding_system *, const unsigned char *,
 622                               unsigned char *, int, int));
 623 extern int encode_coding P_ ((struct coding_system *, const unsigned char *,
 624                               unsigned char *, int, int));
 625 extern void coding_save_composition P_ ((struct coding_system *, int, int,
 626                                          Lisp_Object));
 627 extern void coding_free_composition_data P_ ((struct coding_system *));
 628 extern void coding_adjust_composition_offset P_ ((struct coding_system *,
 629                                                   int));
 630 extern void coding_allocate_composition_data P_ ((struct coding_system *,
 631                                                   int));
 632 extern void coding_restore_composition P_ ((struct coding_system *,
 633                                             Lisp_Object));
 634 extern int code_convert_region P_ ((int, int, int, int, struct coding_system *,
 635                                     int, int));
 636 extern Lisp_Object run_pre_post_conversion_on_str P_ ((Lisp_Object,
 637                                                        struct coding_system *,
 638                                                        int));
 639 extern void run_pre_write_conversin_on_c_str P_ ((unsigned char **, int *,
 640                                                   int, int,
 641                                                   struct coding_system *));
 642
 643 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 644 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 645 extern void detect_coding P_ ((struct coding_system *, const unsigned char *,
 646                                int));
 647 extern void detect_eol P_ ((struct coding_system *, const unsigned char *,
 648                             int));
 649 extern int setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 650 extern Lisp_Object code_convert_string P_ ((Lisp_Object,
 651                                             struct coding_system *, int, int));
 652 extern Lisp_Object code_convert_string1 P_ ((Lisp_Object, Lisp_Object,
 653                                              Lisp_Object, int));
 654 extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object,
 655                                                      int));
 656 extern void setup_raw_text_coding_system P_ ((struct coding_system *));
 657 extern Lisp_Object encode_coding_string P_ ((Lisp_Object,
 658                                              struct coding_system *, int));
 659 extern Lisp_Object decode_coding_string P_ ((Lisp_Object,
 660                                              struct coding_system *, int));
 661 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 662 extern Lisp_Object Qraw_text, Qemacs_mule;
 663 extern Lisp_Object Qbuffer_file_coding_system;
 664 extern Lisp_Object Vcoding_category_list;
 665 extern Lisp_Object Qutf_8;
 666
 667 extern Lisp_Object Qtranslation_table;
 668 extern Lisp_Object Qtranslation_table_id;
 669
 670 /* Mnemonic strings to indicate each type of end-of-line.  */
 671 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 672 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 673 extern Lisp_Object eol_mnemonic_undecided;
 674
 675 /* Format of end-of-line decided by system.  */
 676 extern int system_eol_type;
 677
 678 #ifdef emacs
 679 extern Lisp_Object Qfile_coding_system;
 680 extern Lisp_Object Qcall_process, Qcall_process_region;
 681 extern Lisp_Object Qstart_process, Qopen_network_stream;
 682 extern Lisp_Object Qwrite_region;
 683
 684 extern char *emacs_strerror P_ ((int));
 685
 686 /* Coding-system for reading files and receiving data from process.  */
 687 extern Lisp_Object Vcoding_system_for_read;
 688 /* Coding-system for writing files and sending data to process.  */
 689 extern Lisp_Object Vcoding_system_for_write;
 690 /* Coding-system actually used in the latest I/O.  */
 691 extern Lisp_Object Vlast_coding_system_used;
 692 /* Coding-system to use with system messages (e.g. strerror).  */
 693 extern Lisp_Object Vlocale_coding_system;
 694
 695 /* If non-zero, process buffer inherits the coding system used to decode
 696    the subprocess output.  */
 697 extern int inherit_process_coding_system;
 698
 699 /* Coding-system to be used for encoding terminal output.  This
 700    structure contains information of a coding-system specified by the
 701    function `set-terminal-coding-system'.  */
 702 extern struct coding_system terminal_coding;
 703
 704 /* Coding system to be used to encode text for terminal display when
 705    terminal coding system is nil.  */
 706 extern struct coding_system safe_terminal_coding;
 707
 708 /* Coding-system of what is sent from terminal keyboard.  This
 709    structure contains information of a coding-system specified by the
 710    function `set-keyboard-coding-system'.  */
 711 extern struct coding_system keyboard_coding;
 712
 713 /* Default coding system to be used to write a file.  */
 714 extern struct coding_system default_buffer_file_coding;
 715
 716 /* Default coding systems used for process I/O.  */
 717 extern Lisp_Object Vdefault_process_coding_system;
 718
 719 /* Function to call to force a user to force select a proper coding
 720    system.  */
 721 extern Lisp_Object Vselect_safe_coding_system_function;
 722
 723 /* If nonzero, on writing a file, Vselect_safe_coding_system_function
 724    is called even if Vcoding_system_for_write is non-nil.  */
 725 extern int coding_system_require_warning;
 726
 727 /* Coding system for file names, or nil if none.  */
 728 extern Lisp_Object Vfile_name_coding_system;
 729
 730 /* Coding system for file names used only when
 731    Vfile_name_coding_system is nil.  */
 732 extern Lisp_Object Vdefault_file_name_coding_system;
 733
 734 #endif
 735
 736 /* Error signaled when there's a problem with detecting coding system */
 737 extern Lisp_Object Qcoding_system_error;
 738
 739 #endif /* EMACS_CODING_H */
 740
 741 /* arch-tag: 2bc3b4fa-6870-4f64-8135-b962b2d290e4
 742    (do not change this comment) */