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

   1 /* Header for coding system handler.
   2    Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
   3    Licensed to the Free Software Foundation.
   4
   5 This file is part of GNU Emacs.
   6
   7 GNU Emacs is free software; you can redistribute it and/or modify
   8 it under the terms of the GNU General Public License as published by
   9 the Free Software Foundation; either version 2, or (at your option)
  10 any later version.
  11
  12 GNU Emacs is distributed in the hope that it will be useful,
  13 but WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15 GNU General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with GNU Emacs; see the file COPYING.  If not, write to
  19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  20 Boston, MA 02111-1307, USA.  */
  21
  22 #ifndef _CODING_H
  23 #define _CODING_H
  24
  25 #ifndef _CCL_H
  26 #include "../src/ccl.h"
  27 #endif
  28
  29 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  30
  31 /* All code (1-byte) of Emacs' internal format is classified into one
  32    of the followings.  See also `charset.h'.  */
  33 enum emacs_code_class_type
  34   {
  35     EMACS_control_code,         /* Control codes in the range
  36                                    0x00..0x1F and 0x7F except for the
  37                                    following two codes.  */
  38     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  39                                    end-of-line.  */
  40     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  41                                    in selective display mode.  */
  42     EMACS_ascii_code,           /* ASCII characters.  */
  43     EMACS_leading_code_composition, /* Leading code of a composite
  44                                        character.  */
  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_code,           /* Control codes in the range
  78                                    0x00..0x1F, 0x7F, and 0x80..0x9F,
  79                                    except for the following seven
  80                                    codes.  */
  81     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  82     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  83     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  84     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  85     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  86     ISO_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  87     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  88     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  89     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  90     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  91     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  92     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  93   };
  94
  95 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  96   element in the structure `coding_system'.  This information is used
  97   while encoding a text to ISO2022.  **/
  98
  99 /* If set, produce short-form designation sequence (e.g. ESC $ A)
 100    instead of long-form sequence (e.g. ESC $ ( A).  */
 101 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 102
 103 /* If set, reset graphic planes and registers at end-of-line to the
 104    initial state.  */
 105 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 106
 107 /* If set, reset graphic planes and registers before any control
 108    characters to the initial state.  */
 109 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 110
 111 /* If set, encode by 7-bit environment.  */
 112 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 113
 114 /* If set, use locking-shift function.  */
 115 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 116
 117 /* If set, use single-shift function.  Overwrite
 118    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 119 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 120
 121 /* If set, designate JISX0201-Roman instead of ASCII.  */
 122 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 123
 124 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 125 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 126
 127 /* If set, do not produce ISO6429's direction specifying sequence.  */
 128 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 129
 130 /* If set, assume designation states are reset at beginning of line on
 131    output.  */
 132 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 133
 134 /* If set, designation sequence should be placed at beginning of line
 135    on output.  */
 136 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 137
 138 /* If set, do not encode unexpected charactes on output.  */
 139 #define CODING_FLAG_ISO_SAFE            0x0800
 140
 141 /* If set, extra latin codes (128..159) are accepted as a valid code
 142    on input.  */
 143 #define CODING_FLAG_ISO_LATIN_EXTRA     0x1000
 144
 145 /* A character to be produced on output if encoding of the original
 146    character is prohibited by CODING_FLAG_ISO_SAFE.  */
 147 #define CODING_INHIBIT_CHARACTER_SUBSTITUTION  077 /* 077 == `?' */
 148
 149 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 150 struct iso2022_spec
 151 {
 152   /* The current graphic register invoked to each graphic plane.  */
 153   int current_invocation[2];
 154
 155   /* The current charset designated to each graphic register.  */
 156   int current_designation[4];
 157
 158   /* A charset initially designated to each graphic register.  */
 159   int initial_designation[4];
 160
 161   /* A graphic register to which each charset should be designated.  */
 162   unsigned char requested_designation[MAX_CHARSET + 1];
 163
 164   /* Table of expected character sets for this coding system.  If the
 165      Nth element is 0, the charset of ID N is not an expected
 166      character set.  Such a character set is not encoded when
 167      CODING_ISO_FLAG_SAFE is set.  */
 168   unsigned char expected_charsets[MAX_CHARSET + 1];
 169
 170   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 171      by single-shift while encoding.  */
 172   int single_shifting;
 173
 174   /* Set to 1 temporarily only when processing at beginning of line.  */
 175   int bol;
 176 };
 177
 178 /* Macros to access each field in the structure `spec.iso2022'.  */
 179 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 180   coding->spec.iso2022.current_invocation[plane]
 181 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 182   coding->spec.iso2022.current_designation[reg]
 183 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 184   coding->spec.iso2022.initial_designation[reg]
 185 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 186   coding->spec.iso2022.requested_designation[charset]
 187 #define CODING_SPEC_ISO_EXPECTED_CHARSETS(coding) \
 188   coding->spec.iso2022.expected_charsets
 189 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 190   coding->spec.iso2022.single_shifting
 191 #define CODING_SPEC_ISO_BOL(coding) \
 192   coding->spec.iso2022.bol
 193
 194 /* A value which may appear in
 195    coding->spec.iso2022.requested_designation indicating that the
 196    corresponding charset does not request any graphic register to be
 197    designated.  */
 198 #define CODING_SPEC_ISO_NO_REQUESTED_DESIGNATION 4
 199
 200 /* Return a charset which is currently designated to the graphic plane
 201    PLANE in the coding-system CODING.  */
 202 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 203   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 204    ? -1                                                 \
 205    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 206                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 207
 208 /*** BIG5 section ***/
 209
 210 /* Macros to denote each type of BIG5 coding system.  */
 211 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 212                                         BIG5 developed by Hong Kong
 213                                         University.  */
 214 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 215                                         of BIG5 developed by the
 216                                         company ETen in Taiwan.  */
 217
 218 /*** GENERAL section ***/
 219
 220 /* Types of coding system.  */
 221 enum coding_type
 222   {
 223     coding_type_no_conversion,  /* A coding system which requires no
 224                                    conversion for reading and writing
 225                                    including end-of-line format.  */
 226     coding_type_emacs_mule,     /* A coding system used in Emacs'
 227                                    buffer and string.  Requires no
 228                                    conversion for reading and writing
 229                                    except for end-of-line format.  */
 230     coding_type_undecided,      /* A coding system which requires
 231                                    automatic detection of a real
 232                                    coding system.  */
 233     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 234     coding_type_iso2022,        /* Any coding system of ISO2022
 235                                    variants.  */
 236     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 237     coding_type_ccl,            /* The coding system of which decoder
 238                                    and encoder are written in CCL.  */
 239     coding_type_raw_text        /* A coding system for a text
 240                                    containing ramdom 8-bit code which
 241                                    does not require code conversion
 242                                    except for end-of-line format. */
 243   };
 244
 245 /* Formats of end-of-line.  */
 246 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 247                                    internal format.  */
 248 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 249                                    line-feed.  */
 250 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 251 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 252                                    eol-type is not yet decided.  */
 253 #define CODING_EOL_INCONSISTENT 4 /* This value is used to denote the
 254                                      eol-type is not consistent
 255                                      through the file.  */
 256
 257 /* Character composition status while encoding/decoding.  */
 258 #define COMPOSING_NO             0 /* not composing */
 259 #define COMPOSING_WITH_RULE_HEAD 1 /* 1st char of with-rule composing follow */
 260 #define COMPOSING_NO_RULE_HEAD   2 /* 1st char of no-rule composing follow */
 261 #define COMPOSING_WITH_RULE_TAIL 3 /* Nth char of with-rule composing follow */
 262 #define COMPOSING_NO_RULE_TAIL   4 /* Nth char of no-rule composing follow */
 263 #define COMPOSING_WITH_RULE_RULE 5 /* composition rule follow */
 264
 265 /* 1 iff composing.  */
 266 #define COMPOSING_P(composing) (composing)
 267 /* 1 iff 1st char of composing element follows.  */
 268 #define COMPOSING_HEAD_P(composing) \
 269   ((composing) && (composing) <= COMPOSING_NO_RULE_HEAD)
 270 /* 1 iff composing with embeded composition rule.  */
 271 #define COMPOSING_WITH_RULE_P(composing) ((composing) & 1)
 272
 273 struct coding_system
 274 {
 275   /* Type of the coding system.  */
 276   enum coding_type type;
 277
 278   /* If the coding system requires specific code to be attached at the
 279      tail of converted text, this value should be set to `1'.  */
 280   int require_flushing;
 281
 282   /* Flag bits of the coding system.  The meaning of each bit depends
 283      on the type of the coding system.  */
 284   unsigned int flags;
 285
 286   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 287   int eol_type;
 288
 289   /* Non-zero means that the current source text is the last block of the
 290      whole text to be converted.  */
 291   int last_block;
 292
 293   /* Non-zero means that characters are being composed currently while
 294      decoding or encoding.  See macros COMPOSING_XXXX above for the
 295      meaing of each non-zero value.  */
 296   int composing;
 297
 298   /* 0 (left-to-right) or 1 (right-to-left): the direction of the text
 299      being processed currently.  */
 300   int direction;
 301
 302   /* Non-zero means that the current source text is in a buffer which
 303      enables selective display.  */
 304   int selective;
 305
 306   /* Detailed information specific to each type of coding system.  */
 307   union spec
 308     {
 309       struct iso2022_spec iso2022;
 310       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 311     } spec;
 312
 313   /* Backward pointer to the Lisp symbol of the coding system.  */
 314   Lisp_Object symbol;
 315
 316   /* Lisp function (symbol) to be called after decoding to do
 317      additional conversion. */
 318   Lisp_Object post_read_conversion;
 319
 320   /* Lisp function (symbol) to be called before encoding to do
 321      additional conversion. */
 322   Lisp_Object pre_write_conversion;
 323
 324   /* Character unification tables to look up, or nil.  */
 325   Lisp_Object character_unification_table_for_decode;
 326   Lisp_Object character_unification_table_for_encode;
 327
 328   /* Carryover yielded by decoding/encoding incomplete source.  No
 329      coding-system yields more than 7-byte of carryover.  This does
 330      not include a text which is not processed because of short of
 331      output buffer.  */
 332   char carryover[8];
 333
 334   /* Actual data length in the above array.  */
 335   int carryover_size;
 336 };
 337
 338 /* Return 1 if coding system CODING never requires any code conversion.  */
 339 #define CODING_REQUIRE_NO_CONVERSION(coding)            \
 340   ((coding)->type == coding_type_no_conversion          \
 341    || (((coding)->type == coding_type_emacs_mule        \
 342         || (coding)->type == coding_type_raw_text)      \
 343        && (coding)->eol_type == CODING_EOL_LF))
 344
 345 /* Return 1 if coding system CODING may not require code conversion.  */
 346 #define CODING_MAY_REQUIRE_NO_CONVERSION(coding)        \
 347   ((coding)->type == coding_type_no_conversion          \
 348    || (((coding)->type == coding_type_emacs_mule        \
 349         || (coding)->type == coding_type_undecided      \
 350         || (coding)->type == coding_type_raw_text)      \
 351        && ((coding)->eol_type == CODING_EOL_LF          \
 352            || (coding)->eol_type == CODING_EOL_UNDECIDED)))
 353
 354 /* Index for each coding category in `coding_category_table' */
 355 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 356 #define CODING_CATEGORY_IDX_SJIS        1
 357 #define CODING_CATEGORY_IDX_ISO_7       2
 358 #define CODING_CATEGORY_IDX_ISO_8_1     3
 359 #define CODING_CATEGORY_IDX_ISO_8_2     4
 360 #define CODING_CATEGORY_IDX_ISO_7_ELSE  5
 361 #define CODING_CATEGORY_IDX_ISO_8_ELSE  6
 362 #define CODING_CATEGORY_IDX_BIG5        7
 363 #define CODING_CATEGORY_IDX_RAW_TEXT    8
 364 #define CODING_CATEGORY_IDX_BINARY      9
 365 #define CODING_CATEGORY_IDX_MAX         10
 366
 367 /* Definitions of flag bits returned by the function
 368    detect_coding_mask ().  */
 369 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 370 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 371 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 372 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 373 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 374 #define CODING_CATEGORY_MASK_ISO_7_ELSE (1 << CODING_CATEGORY_IDX_ISO_7_ELSE)
 375 #define CODING_CATEGORY_MASK_ISO_8_ELSE (1 << CODING_CATEGORY_IDX_ISO_8_ELSE)
 376 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 377 #define CODING_CATEGORY_MASK_RAW_TEXT   (1 << CODING_CATEGORY_IDX_RAW_TEXT)
 378 #define CODING_CATEGORY_MASK_BINARY     (1 << CODING_CATEGORY_IDX_BINARY)
 379
 380 /* This value is returned if detect_coding_mask () find nothing other
 381    than ASCII characters.  */
 382 #define CODING_CATEGORY_MASK_ANY        \
 383   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 384    | CODING_CATEGORY_MASK_SJIS          \
 385    | CODING_CATEGORY_MASK_ISO_7         \
 386    | CODING_CATEGORY_MASK_ISO_8_1       \
 387    | CODING_CATEGORY_MASK_ISO_8_2       \
 388    | CODING_CATEGORY_MASK_ISO_7_ELSE    \
 389    | CODING_CATEGORY_MASK_ISO_8_ELSE    \
 390    | CODING_CATEGORY_MASK_BIG5)
 391
 392 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 393    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 394    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 395    internal format.  */
 396
 397 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 398   do {                                                  \
 399     if (s2 >= 0x9F)                                     \
 400       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 401       c2 = s2 - 0x7E;                                   \
 402     else                                                \
 403       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 404       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 405   } while (0)
 406
 407 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 408   do {                                                  \
 409     if (c1 & 1)                                         \
 410       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 411       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 412     else                                                \
 413       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 414       s2 = c2 + 0x7E;                                   \
 415   } while (0)
 416
 417 /* Extern declarations.  */
 418 extern int decode_coding (), encode_coding ();
 419 extern int decoding_buffer_size (), encoding_buffer_size ();
 420 extern int conversion_buffer_size;
 421 extern char *conversion_buffer, *get_conversion_buffer ();
 422 extern Lisp_Object Fcheck_coding_system ();
 423 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 424 extern Lisp_Object Qemacs_mule;
 425 extern Lisp_Object Qbuffer_file_coding_system;
 426 extern Lisp_Object Vcoding_category_list;
 427
 428 /* Mnemonic character to indicate each type of end-of-line.  */
 429 extern int eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 430 /* Mnemonic character to indicate type of end-of-line is not yet decided.  */
 431 extern int eol_mnemonic_undecided;
 432
 433 /* Table of coding-systems currently assigned to each coding-category.  */
 434 extern Lisp_Object coding_category_table[CODING_CATEGORY_IDX_MAX];
 435 /* Table of names of symbol for each coding-category.  */
 436 extern char *coding_category_name[CODING_CATEGORY_IDX_MAX];
 437
 438 #ifdef emacs
 439 extern Lisp_Object Qfile_coding_system;
 440 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 441 extern Lisp_Object Qstart_process, Qopen_network_stream;
 442
 443 /* Coding-system for reading files and receiving data from process.  */
 444 extern Lisp_Object Vcoding_system_for_read;
 445 /* Coding-system for writing files and sending data to process.  */
 446 extern Lisp_Object Vcoding_system_for_write;
 447 /* Coding-system actually used in the latest I/O.  */
 448 extern Lisp_Object Vlast_coding_system_used;
 449
 450 /* Coding-system to be used for encoding terminal output.  This
 451    structure contains information of a coding-system specified by the
 452    function `set-terminal-coding-system'.  */
 453 extern struct coding_system terminal_coding;
 454
 455 /* Coding system to be used to encode text for terminal display when
 456    terminal coding system is nil.  */
 457 extern struct coding_system safe_terminal_coding;
 458
 459 /* Coding-system of what is sent from terminal keyboard.  This
 460    structure contains information of a coding-system specified by the
 461    function `set-keyboard-coding-system'.  */
 462 extern struct coding_system keyboard_coding;
 463
 464 /* Default coding systems used for process I/O.  */
 465 extern Lisp_Object Vdefault_process_coding_system;
 466
 467 #endif
 468
 469 #endif /* _CODING_H */