src/coding.h

   1 /* Header for coding system handler.
   2    Ver.1.0
   3    Copyright (C) 1995 Free Software Foundation, Inc.
   4    Copyright (C) 1995 Electrotechnical Laboratory, JAPAN.
   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., 59 Temple Place - Suite 330,
  21 Boston, MA 02111-1307, USA.  */
  22
  23 #ifndef _CODING_H
  24 #define _CODING_H
  25
  26 #ifndef _CCL_H
  27 #include "../src/ccl.h"
  28 #endif
  29
  30 /*** EMACS' INTERNAL FORMAT (emacs-mule) section ***/
  31
  32 /* All code (1-byte) of Emacs' internal format is classified into one
  33    of the followings.  See also `charset.h'.  */
  34 enum emacs_code_class_type
  35   {
  36     EMACS_control_code,         /* Control codes in the range
  37                                    0x00..0x1F and 0x7F except for the
  38                                    following two codes.  */
  39     EMACS_linefeed_code,        /* 0x0A (linefeed) to denote
  40                                    end-of-line.  */
  41     EMACS_carriage_return_code, /* 0x0D (carriage-return) to be used
  42                                    in selective display mode.  */
  43     EMACS_ascii_code,           /* ASCII characters.  */
  44     EMACS_leading_code_composition, /* Leading code of a composite
  45                                        character.  */
  46     EMACS_leading_code_2,       /* Base leading code of official
  47                                    TYPE9N character.  */
  48     EMACS_leading_code_3,       /* Base leading code of private TYPE9N
  49                                    or official TYPE9Nx9N character.  */
  50     EMACS_leading_code_4,       /* Base leading code of private
  51                                    TYPE9Nx9N character.  */
  52     EMACS_invalid_code          /* Invalid code, i.e. a base leading
  53                                    code not yet assigned to any
  54                                    charset, or a code of the range
  55                                    0xA0..0xFF.  */
  56   };
  57
  58 extern enum emacs_code_class_type emacs_code_class[256];
  59
  60 /*** ISO2022 section ***/
  61
  62 /* Macros to define code of control characters for ISO2022's functions.  */
  63                         /* code */      /* function */
  64 #define ISO_CODE_LF     0x0A            /* line-feed */
  65 #define ISO_CODE_CR     0x0D            /* carriage-return */
  66 #define ISO_CODE_SO     0x0E            /* shift-out */
  67 #define ISO_CODE_SI     0x0F            /* shift-in */
  68 #define ISO_CODE_SS2_7  0x19            /* single-shift-2 for 7-bit code */
  69 #define ISO_CODE_ESC    0x1B            /* escape */
  70 #define ISO_CODE_SS2    0x8E            /* single-shift-2 */
  71 #define ISO_CODE_SS3    0x8F            /* single-shift-3 */
  72 #define ISO_CODE_CSI    0x9B            /* control-sequence-introduce */
  73
  74 /* All code (1-byte) of ISO2022 is classified into one of the
  75    followings.  */
  76 enum iso_code_class_type
  77   {
  78     ISO_control_code,           /* Control codes in the range
  79                                    0x00..0x1F, 0x7F, and 0x80..0x9F,
  80                                    except for the following seven
  81                                    codes.  */
  82     ISO_carriage_return,        /* ISO_CODE_CR (0x0D) */
  83     ISO_shift_out,              /* ISO_CODE_SO (0x0E) */
  84     ISO_shift_in,               /* ISO_CODE_SI (0x0F) */
  85     ISO_single_shift_2_7,       /* ISO_CODE_SS2_7 (0x19) */
  86     ISO_escape,                 /* ISO_CODE_SO (0x1B) */
  87     ISO_single_shift_2,         /* ISO_CODE_SS2 (0x8E) */
  88     ISO_single_shift_3,         /* ISO_CODE_SS3 (0x8F) */
  89     ISO_control_sequence_introducer, /* ISO_CODE_CSI (0x9B) */
  90     ISO_0x20_or_0x7F,           /* Codes of the values 0x20 or 0x7F.  */
  91     ISO_graphic_plane_0,        /* Graphic codes in the range 0x21..0x7E.  */
  92     ISO_0xA0_or_0xFF,           /* Codes of the values 0xA0 or 0xFF.  */
  93     ISO_graphic_plane_1         /* Graphic codes in the range 0xA1..0xFE.  */
  94   };
  95
  96 /** The macros CODING_FLAG_ISO_XXX defines a flag bit of the `flags'
  97   element in the structure `coding_system'.  This information is used
  98   while encoding a text to ISO2022.  **/
  99
 100 /* If set, produce short-form designation sequence (e.g. ESC $ A)
 101    instead of long-form sequence (e.g. ESC $ ( A).  */
 102 #define CODING_FLAG_ISO_SHORT_FORM      0x0001
 103
 104 /* If set, reset graphic planes and registers at end-of-line to the
 105    initial state.  */
 106 #define CODING_FLAG_ISO_RESET_AT_EOL    0x0002
 107
 108 /* If set, reset graphic planes and registers before any control
 109    characters to the initial state.  */
 110 #define CODING_FLAG_ISO_RESET_AT_CNTL   0x0004
 111
 112 /* If set, encode by 7-bit environment.  */
 113 #define CODING_FLAG_ISO_SEVEN_BITS      0x0008
 114
 115 /* If set, use locking-shift function.  */
 116 #define CODING_FLAG_ISO_LOCKING_SHIFT   0x0010
 117
 118 /* If set, use single-shift function.  Overwrite
 119    CODING_FLAG_ISO_LOCKING_SHIFT.  */
 120 #define CODING_FLAG_ISO_SINGLE_SHIFT    0x0020
 121
 122 /* If set, designate JISX0201-Roman instead of ASCII.  */
 123 #define CODING_FLAG_ISO_USE_ROMAN       0x0040
 124
 125 /* If set, designate JISX0208-1978 instead of JISX0208-1983.  */
 126 #define CODING_FLAG_ISO_USE_OLDJIS      0x0080
 127
 128 /* If set, do not produce ISO6429's direction specifying sequence.  */
 129 #define CODING_FLAG_ISO_NO_DIRECTION    0x0100
 130
 131 /* If set, assume designation states are reset at beginning of line on
 132    output.  */
 133 #define CODING_FLAG_ISO_INIT_AT_BOL     0x0200
 134
 135 /* If set, designation sequence should be placed at beginning of line
 136    on output.  */
 137 #define CODING_FLAG_ISO_DESIGNATE_AT_BOL 0x0400
 138
 139 /* Structure of the field `spec.iso2022' in the structure `coding_system'.  */
 140 struct iso2022_spec
 141 {
 142   /* The current graphic register invoked to each graphic plane.  */
 143   int current_invocation[2];
 144
 145   /* The current charset designated to each graphic register.  */
 146   int current_designation[4];
 147
 148   /* A charset initially designated to each graphic register.  */
 149   int initial_designation[4];
 150
 151   /* A graphic register to which each charset should be designated.  */
 152   char requested_designation[MAX_CHARSET + 1];
 153
 154   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 155      by single-shift while encoding.  */
 156   int single_shifting;
 157
 158   /* Set to 1 temporarily only when processing at beginning of line.  */
 159   int bol;
 160 };
 161
 162 /* Macros to access each field in the structure `spec.iso2022'.  */
 163 #define CODING_SPEC_ISO_INVOCATION(coding, plane) \
 164   coding->spec.iso2022.current_invocation[plane]
 165 #define CODING_SPEC_ISO_DESIGNATION(coding, reg) \
 166   coding->spec.iso2022.current_designation[reg]
 167 #define CODING_SPEC_ISO_INITIAL_DESIGNATION(coding, reg) \
 168   coding->spec.iso2022.initial_designation[reg]
 169 #define CODING_SPEC_ISO_REQUESTED_DESIGNATION(coding, charset) \
 170   coding->spec.iso2022.requested_designation[charset]
 171 #define CODING_SPEC_ISO_SINGLE_SHIFTING(coding) \
 172   coding->spec.iso2022.single_shifting
 173 #define CODING_SPEC_ISO_BOL(coding) \
 174   coding->spec.iso2022.bol
 175
 176 /* Return a charset which is currently designated to the graphic plane
 177    PLANE in the coding-system CODING.  */
 178 #define CODING_SPEC_ISO_PLANE_CHARSET(coding, plane)    \
 179   ((CODING_SPEC_ISO_INVOCATION (coding, plane) < 0)     \
 180    ? -1                                                 \
 181    : CODING_SPEC_ISO_DESIGNATION (coding,               \
 182                                   CODING_SPEC_ISO_INVOCATION (coding, plane)))
 183
 184 /*** BIG5 section ***/
 185
 186 /* Macros to denote each type of BIG5 coding system.  */
 187 #define CODING_FLAG_BIG5_HKU    0x00 /* BIG5-HKU is one of variants of
 188                                         BIG5 developed by Hong Kong
 189                                         University.  */
 190 #define CODING_FLAG_BIG5_ETEN   0x01 /* BIG5_ETen is one of variants
 191                                         of BIG5 developed by the
 192                                         company ETen in Taiwan.  */
 193
 194 /*** GENERAL section ***/
 195
 196 /* Types of coding system.  */
 197 enum coding_type
 198   {
 199     coding_type_no_conversion,  /* A coding system which requires no
 200                                    conversion for reading and writing
 201                                    including end-of-line format.  */
 202     coding_type_emacs_mule,     /* A coding system used in Emacs'
 203                                    buffer and string.  Requires no
 204                                    conversion for reading and writing
 205                                    except for end-of-line format.  */
 206     coding_type_undecided,      /* A coding system which requires
 207                                    automatic detection of a real
 208                                    coding system.  */
 209     coding_type_sjis,           /* SJIS coding system for Japanese.  */
 210     coding_type_iso2022,        /* Any coding system of ISO2022
 211                                    variants.  */
 212     coding_type_big5,           /* BIG5 coding system for Chinese.  */
 213     coding_type_ccl             /* The coding system of which decoder
 214                                    and encoder are written in CCL.  */
 215   };
 216
 217 /* Formats of end-of-line.  */
 218 #define CODING_EOL_LF   0       /* Line-feed only, same as Emacs'
 219                                    internal format.  */
 220 #define CODING_EOL_CRLF 1       /* Sequence of carriage-return and
 221                                    line-feed.  */
 222 #define CODING_EOL_CR   2       /* Carriage-return only.  */
 223 #define CODING_EOL_UNDECIDED 3  /* This value is used to denote the
 224                                    eol-type is not yet decided.  */
 225
 226 /* Character composition status while encoding/decoding.  */
 227 #define COMPOSING_NO             0 /* not composing */
 228 #define COMPOSING_WITH_RULE_HEAD 1 /* 1st char of with-rule composing follow */
 229 #define COMPOSING_NO_RULE_HEAD   2 /* 1st char of no-rule composing follow */
 230 #define COMPOSING_WITH_RULE_TAIL 3 /* Nth char of with-rule composing follow */
 231 #define COMPOSING_NO_RULE_TAIL   4 /* Nth char of no-rule composing follow */
 232 #define COMPOSING_WITH_RULE_RULE 5 /* composition rule follow */
 233
 234 /* 1 iff composing.  */
 235 #define COMPOSING_P(composing) (composing)
 236 /* 1 iff 1st char of composing element follows.  */
 237 #define COMPOSING_HEAD_P(composing) \
 238   ((composing) && (composing) <= COMPOSING_NO_RULE_HEAD)
 239 /* 1 iff composing with embeded composition rule.  */
 240 #define COMPOSING_WITH_RULE_P(composing) ((composing) & 1)
 241
 242 struct coding_system
 243 {
 244   /* Type of the coding system.  */
 245   enum coding_type type;
 246
 247   /* If the coding system requires specific code to be attached at the
 248      tail of converted text, this value should be set to `1'.  */
 249   int require_flushing;
 250
 251   /* Flag bits of the coding system.  The meaning of each bit depends
 252      on the type of the coding system.  */
 253   unsigned int flags;
 254
 255   /* Type of end-of-line format (LF, CRLF, or CR) of the coding system.  */
 256   int eol_type;
 257
 258   /* Non-zero means that the current source text is the last block of the
 259      whole text to be converted.  */
 260   int last_block;
 261
 262   /* Non-zero means that characters are being composed currently while
 263      decoding or encoding.  See macros COMPOSING_XXXX above for the
 264      meaing of each non-zero value.  */
 265   int composing;
 266
 267   /* 0 (left-to-right) or 1 (right-to-left): the direction of the text
 268      being processed currently.  */
 269   int direction;
 270
 271   /* Non-zero means that the current source text is in a buffer which
 272      enables selective display.  */
 273   int selective;
 274
 275   /* Detailed information specific to each type of coding system.  */
 276   union spec
 277     {
 278       struct iso2022_spec iso2022;
 279       struct ccl_spec ccl;      /* Defined in ccl.h.  */
 280     } spec;
 281
 282   /* Backward pointer to the Lisp symbol of the coding system.  */
 283   Lisp_Object symbol;
 284
 285   /* Lisp function (symbol) to be called after decoding to do
 286      additional conversion. */
 287   Lisp_Object post_read_conversion;
 288
 289   /* Lisp function (symbol) to be called before encoding to do
 290      additional conversion. */
 291   Lisp_Object pre_write_conversion;
 292
 293   /* Character unification table to look up, or nil.  */
 294   Lisp_Object character_unification_table;
 295
 296   /* Carryover yielded by decoding/encoding incomplete source.  No
 297      coding-system yields more than 7-byte of carryover.  This does
 298      not include a text which is not processed because of short of
 299      output buffer.  */
 300   char carryover[8];
 301
 302   /* Actual data length in the above array.  */
 303   int carryover_size;
 304 };
 305
 306 /* Return 1 if the coding-system CODING requires conversion of
 307    representation of a visible character (text).  */
 308 #define CODING_REQUIRE_TEXT_CONVERSION(coding)  \
 309   ((coding)->type != coding_type_no_conversion          \
 310    && (coding)->type != coding_type_emacs_mule)
 311
 312 /* Return 1 if the coding-system CODING requires conversion of the
 313    format of end-of-line.  */
 314 #define CODING_REQUIRE_EOL_CONVERSION(coding)   \
 315   ((coding)->eol_type != CODING_EOL_UNDECIDED   \
 316    && (coding)->eol_type != CODING_EOL_LF)
 317
 318 /* Return 1 if the coding-system CODING requires some conversion.  */
 319 #define CODING_REQUIRE_CONVERSION(coding)       \
 320   (CODING_REQUIRE_TEXT_CONVERSION (coding)      \
 321    || CODING_REQUIRE_EOL_CONVERSION (coding))
 322
 323 /* Index for each coding category in `coding_category_table' */
 324 #define CODING_CATEGORY_IDX_EMACS_MULE  0
 325 #define CODING_CATEGORY_IDX_SJIS        1
 326 #define CODING_CATEGORY_IDX_ISO_7       2
 327 #define CODING_CATEGORY_IDX_ISO_8_1     3
 328 #define CODING_CATEGORY_IDX_ISO_8_2     4
 329 #define CODING_CATEGORY_IDX_ISO_ELSE    5
 330 #define CODING_CATEGORY_IDX_BIG5        6
 331 #define CODING_CATEGORY_IDX_BINARY      7
 332 #define CODING_CATEGORY_IDX_MAX         8
 333
 334 /* Definitions of flag bits returned by the function
 335    detect_coding_mask ().  */
 336 #define CODING_CATEGORY_MASK_EMACS_MULE (1 << CODING_CATEGORY_IDX_EMACS_MULE)
 337 #define CODING_CATEGORY_MASK_SJIS       (1 << CODING_CATEGORY_IDX_SJIS)
 338 #define CODING_CATEGORY_MASK_ISO_7      (1 << CODING_CATEGORY_IDX_ISO_7)
 339 #define CODING_CATEGORY_MASK_ISO_8_1    (1 << CODING_CATEGORY_IDX_ISO_8_1)
 340 #define CODING_CATEGORY_MASK_ISO_8_2    (1 << CODING_CATEGORY_IDX_ISO_8_2)
 341 #define CODING_CATEGORY_MASK_ISO_ELSE   (1 << CODING_CATEGORY_IDX_ISO_ELSE)
 342 #define CODING_CATEGORY_MASK_BIG5       (1 << CODING_CATEGORY_IDX_BIG5)
 343
 344 /* This value is returned if detect_coding_mask () find nothing other
 345    than ASCII characters.  */
 346 #define CODING_CATEGORY_MASK_ANY        \
 347   (  CODING_CATEGORY_MASK_EMACS_MULE    \
 348    | CODING_CATEGORY_MASK_SJIS          \
 349    | CODING_CATEGORY_MASK_ISO_7         \
 350    | CODING_CATEGORY_MASK_ISO_8_1       \
 351    | CODING_CATEGORY_MASK_ISO_8_2       \
 352    | CODING_CATEGORY_MASK_ISO_ELSE      \
 353    | CODING_CATEGORY_MASK_BIG5)
 354
 355 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 356    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 357    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 358    internal format.  */
 359
 360 #define DECODE_SJIS(s1, s2, c1, c2)                     \
 361   do {                                                  \
 362     if (s2 >= 0x9F)                                     \
 363       c1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),        \
 364       c2 = s2 - 0x7E;                                   \
 365     else                                                \
 366       c1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),      \
 367       c2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F);           \
 368   } while (0)
 369
 370 #define ENCODE_SJIS(c1, c2, s1, s2)                     \
 371   do {                                                  \
 372     if (c1 & 1)                                         \
 373       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x71 : 0xB1),        \
 374       s2 = c2 + ((c2 >= 0x60) ? 0x20 : 0x1F);           \
 375     else                                                \
 376       s1 = c1 / 2 + ((c1 < 0x5F) ? 0x70 : 0xB0),        \
 377       s2 = c2 + 0x7E;                                   \
 378   } while (0)
 379
 380 /* Extern declarations.  */
 381 extern int decode_coding (), encode_coding ();
 382 extern int decoding_buffer_size (), encoding_buffer_size ();
 383 extern int conversion_buffer_size;
 384 extern char *conversion_buffer, *get_conversion_buffer ();
 385 extern Lisp_Object Fcheck_coding_system ();
 386 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 387 extern Lisp_Object Qbuffer_file_coding_system;
 388 extern Lisp_Object Vcoding_category_list;
 389
 390 /* Mnemonic character to indicate each type of end-of-line.  */
 391 extern int eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 392 /* Mnemonic character to indicate type of end-of-line is not yet decided.  */
 393 extern int eol_mnemonic_undecided;
 394
 395 /* Table of coding-systems currently assigned to each coding-category.  */
 396 extern Lisp_Object coding_category_table[CODING_CATEGORY_IDX_MAX];
 397 /* Table of names of symbol for each coding-category.  */
 398 extern char *coding_category_name[CODING_CATEGORY_IDX_MAX];
 399
 400 #ifdef emacs
 401 extern Lisp_Object Qfile_coding_system;
 402 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 403 extern Lisp_Object Qstart_process, Qopen_network_stream;
 404
 405 /* Coding-system for reading files and receiving data from process.  */
 406 extern Lisp_Object Vcoding_system_for_read;
 407 /* Coding-system for writing files and sending data to process.  */
 408 extern Lisp_Object Vcoding_system_for_write;
 409 /* Coding-system actually used in the latest I/O.  */
 410 extern Lisp_Object Vlast_coding_system_used;
 411
 412 /* Coding-system to be used for encoding terminal output.  This
 413    structure contains information of a coding-system specified by the
 414    function `set-terminal-coding-system'.  */
 415 extern struct coding_system terminal_coding;
 416
 417 /* Coding-system of what is sent from terminal keyboard.  This
 418    structure contains information of a coding-system specified by the
 419    function `set-keyboard-coding-system'.  */
 420 extern struct coding_system keyboard_coding;
 421
 422 extern Lisp_Object Vcoding_system_alist;
 423
 424 #endif
 425
 426 #endif /* _CODING_H */