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    Copyright (C) 2003
   6      National Institute of Advanced Industrial Science and Technology (AIST)
   7      Registration Number H13PRO009
   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 2, 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., 59 Temple Place - Suite 330,
  24 Boston, MA 02111-1307, USA.  */
  25
  26 #ifndef EMACS_CODING_H
  27 #define EMACS_CODING_H
  28
  29 /* Index to arguments of Fdefine_coding_system_internal.  */
  30
  31 enum define_coding_system_arg_index
  32   {
  33     coding_arg_name,
  34     coding_arg_mnemonic,
  35     coding_arg_coding_type,
  36     coding_arg_charset_list,
  37     coding_arg_ascii_compatible_p,
  38     coding_arg_decode_translation_table,
  39     coding_arg_encode_translation_table,
  40     coding_arg_post_read_conversion,
  41     coding_arg_pre_write_conversion,
  42     coding_arg_default_char,
  43     coding_arg_for_unibyte,
  44     coding_arg_plist,
  45     coding_arg_eol_type,
  46     coding_arg_max
  47   };
  48
  49 enum define_coding_iso2022_arg_index
  50   {
  51     coding_arg_iso2022_initial = coding_arg_max,
  52     coding_arg_iso2022_reg_usage,
  53     coding_arg_iso2022_request,
  54     coding_arg_iso2022_flags,
  55     coding_arg_iso2022_max
  56   };
  57
  58 enum define_coding_utf16_arg_index
  59   {
  60     coding_arg_utf16_bom = coding_arg_max,
  61     coding_arg_utf16_endian,
  62     coding_arg_utf16_max
  63   };
  64
  65 enum define_coding_ccl_arg_index
  66   {
  67     coding_arg_ccl_decoder = coding_arg_max,
  68     coding_arg_ccl_encoder,
  69     coding_arg_ccl_valids,
  70     coding_arg_ccl_max
  71   };
  72
  73 /* Hash table for all coding systems.  Keys are coding system symbols
  74    and values are spec vectors of the corresponding coding system.  A
  75    spec vector has the form [ ATTRS ALIASES EOL-TYPE ].  ATTRS is a
  76    vector of attribute of the coding system.  ALIASES is a list of
  77    aliases (symbols) of the coding system.  EOL-TYPE is `unix', `dos',
  78    `mac' or a vector of coding systems (symbols).  */
  79
  80 extern Lisp_Object Vcoding_system_hash_table;
  81
  82
  83 /* Enumeration of coding system type.  */
  84
  85 enum coding_system_type
  86   {
  87     coding_type_charset,
  88     coding_type_utf_8,
  89     coding_type_utf_16,
  90     coding_type_iso_2022,
  91     coding_type_emacs_mule,
  92     coding_type_sjis,
  93     coding_type_ccl,
  94     coding_type_raw_text,
  95     coding_type_undecided,
  96     coding_type_max
  97   };
  98
  99
 100 /* Enumeration of end-of-line format type.  */
 101
 102 enum end_of_line_type
 103   {
 104     eol_lf,             /* Line-feed only, same as Emacs' internal
 105                            format.  */
 106     eol_crlf,           /* Sequence of carriage-return and
 107                            line-feed.  */
 108     eol_cr,             /* Carriage-return only.  */
 109     eol_any,            /* Accept any of above.  Produce line-feed
 110                            only.  */
 111     eol_undecided,      /* This value is used to denote that the
 112                            eol-type is not yet undecided.  */
 113     eol_type_max
 114   };
 115
 116 /* Enumeration of index to an attribute vector of a coding system.  */
 117
 118 enum coding_attr_index
 119   {
 120     coding_attr_base_name,
 121     coding_attr_docstring,
 122     coding_attr_mnemonic,
 123     coding_attr_type,
 124     coding_attr_charset_list,
 125     coding_attr_ascii_compat,
 126     coding_attr_decode_tbl,
 127     coding_attr_encode_tbl,
 128     coding_attr_trans_tbl,
 129     coding_attr_post_read,
 130     coding_attr_pre_write,
 131     coding_attr_default_char,
 132     coding_attr_for_unibyte,
 133     coding_attr_plist,
 134
 135     coding_attr_category,
 136     coding_attr_safe_charsets,
 137
 138     /* The followings are extra attributes for each type.  */
 139     coding_attr_charset_valids,
 140
 141     coding_attr_ccl_decoder,
 142     coding_attr_ccl_encoder,
 143     coding_attr_ccl_valids,
 144
 145     coding_attr_iso_initial,
 146     coding_attr_iso_usage,
 147     coding_attr_iso_request,
 148     coding_attr_iso_flags,
 149
 150     coding_attr_utf_16_bom,
 151     coding_attr_utf_16_endian,
 152
 153     coding_attr_emacs_mule_full,
 154
 155     coding_attr_last_index
 156   };
 157
 158
 159 /* Macros to access an element of an attribute vector.  */
 160
 161 #define CODING_ATTR_BASE_NAME(attrs)    AREF (attrs, coding_attr_base_name)
 162 #define CODING_ATTR_TYPE(attrs)         AREF (attrs, coding_attr_type)
 163 #define CODING_ATTR_CHARSET_LIST(attrs) AREF (attrs, coding_attr_charset_list)
 164 #define CODING_ATTR_MNEMONIC(attrs)     AREF (attrs, coding_attr_mnemonic)
 165 #define CODING_ATTR_DOCSTRING(attrs)    AREF (attrs, coding_attr_docstring)
 166 #define CODING_ATTR_ASCII_COMPAT(attrs) AREF (attrs, coding_attr_ascii_compat)
 167 #define CODING_ATTR_DECODE_TBL(attrs)   AREF (attrs, coding_attr_decode_tbl)
 168 #define CODING_ATTR_ENCODE_TBL(attrs)   AREF (attrs, coding_attr_encode_tbl)
 169 #define CODING_ATTR_TRANS_TBL(attrs)    AREF (attrs, coding_attr_trans_tbl)
 170 #define CODING_ATTR_POST_READ(attrs)    AREF (attrs, coding_attr_post_read)
 171 #define CODING_ATTR_PRE_WRITE(attrs)    AREF (attrs, coding_attr_pre_write)
 172 #define CODING_ATTR_DEFAULT_CHAR(attrs) AREF (attrs, coding_attr_default_char)
 173 #define CODING_ATTR_FOR_UNIBYTE(attrs)  AREF (attrs, coding_attr_for_unibyte)
 174 #define CODING_ATTR_FLUSHING(attrs)     AREF (attrs, coding_attr_flushing)
 175 #define CODING_ATTR_PLIST(attrs)        AREF (attrs, coding_attr_plist)
 176 #define CODING_ATTR_CATEGORY(attrs)     AREF (attrs, coding_attr_category)
 177 #define CODING_ATTR_SAFE_CHARSETS(attrs)AREF (attrs, coding_attr_safe_charsets)
 178
 179
 180 /* Return the name of a coding system specified by ID.  */
 181 #define CODING_ID_NAME(id) \
 182   (HASH_KEY (XHASH_TABLE (Vcoding_system_hash_table), id))
 183
 184 /* Return the attribute vector of a coding system specified by ID.  */
 185
 186 #define CODING_ID_ATTRS(id)     \
 187   (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 0))
 188
 189 /* Return the list of aliases of a coding system specified by ID.  */
 190
 191 #define CODING_ID_ALIASES(id)   \
 192   (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 1))
 193
 194 /* Return the eol-type of a coding system specified by ID.  */
 195
 196 #define CODING_ID_EOL_TYPE(id)  \
 197   (AREF (HASH_VALUE (XHASH_TABLE (Vcoding_system_hash_table), id), 2))
 198
 199
 200 /* Return the spec vector of CODING_SYSTEM_SYMBOL.  */
 201
 202 #define CODING_SYSTEM_SPEC(coding_system_symbol)        \
 203   (Fgethash (coding_system_symbol, Vcoding_system_hash_table, Qnil))
 204
 205
 206 /* Return the ID of CODING_SYSTEM_SYMBOL.  */
 207
 208 #define CODING_SYSTEM_ID(coding_system_symbol)                  \
 209   hash_lookup (XHASH_TABLE (Vcoding_system_hash_table),         \
 210                coding_system_symbol, NULL)
 211
 212 /* Return 1 iff CODING_SYSTEM_SYMBOL is a coding system.  */
 213
 214 #define CODING_SYSTEM_P(coding_system_symbol)   \
 215   (! NILP (CODING_SYSTEM_SPEC (coding_system_symbol)))
 216
 217 /* Check if X is a coding system or not.  */
 218
 219 #define CHECK_CODING_SYSTEM(x)                          \
 220   do {                                                  \
 221     if (!CODING_SYSTEM_P (x))                           \
 222       wrong_type_argument (Qcoding_system_p, (x));      \
 223   } while (0)
 224
 225
 226 /* Check if X is a coding system or not.  If it is, set SEPC to the
 227    spec vector of the coding system.  */
 228
 229 #define CHECK_CODING_SYSTEM_GET_SPEC(x, spec)           \
 230   do {                                                  \
 231     spec = CODING_SYSTEM_SPEC (x);                      \
 232     if (NILP (spec))                                    \
 233       x = wrong_type_argument (Qcoding_system_p, (x));  \
 234   } while (0)
 235
 236
 237 /* Check if X is a coding system or not.  If it is, set ID to the
 238    ID of the coding system.  */
 239
 240 #define CHECK_CODING_SYSTEM_GET_ID(x, id)                       \
 241   do                                                            \
 242     {                                                           \
 243       id = CODING_SYSTEM_ID (x);                                \
 244       if (id < 0)                                               \
 245         x = wrong_type_argument (Qcoding_system_p, (x));        \
 246     } while (0)
 247
 248
 249 /*** GENERAL section ***/
 250
 251 /* Enumeration of result code of code conversion.  */
 252 enum coding_result_code
 253   {
 254     CODING_RESULT_SUCCESS,
 255     CODING_RESULT_INSUFFICIENT_SRC,
 256     CODING_RESULT_INSUFFICIENT_DST,
 257     CODING_RESULT_INCONSISTENT_EOL,
 258     CODING_RESULT_INVALID_SRC,
 259     CODING_RESULT_INTERRUPT,
 260     CODING_RESULT_INSUFFICIENT_MEM
 261   };
 262
 263
 264 /* Macros used for the member `mode' of the struct coding_system.  */
 265
 266 /* If set, recover the original CR or LF of the already decoded text
 267    when the decoding routine encounters an inconsistent eol format.  */
 268 #define CODING_MODE_INHIBIT_INCONSISTENT_EOL    0x01
 269
 270 /* If set, the decoding/encoding routines treat the current data as
 271    the last block of the whole text to be converted, and do the
 272    appropriate finishing job.  */
 273 #define CODING_MODE_LAST_BLOCK                  0x02
 274
 275 /* If set, it means that the current source text is in a buffer which
 276    enables selective display.  */
 277 #define CODING_MODE_SELECTIVE_DISPLAY           0x04
 278
 279 /* This flag is used by the decoding/encoding routines on the fly.  If
 280    set, it means that right-to-left text is being processed.  */
 281 #define CODING_MODE_DIRECTION                   0x08
 282
 283 #define CODING_MODE_FIXED_DESTINATION           0x10
 284
 285 /* If set, it means that the encoding routines produces some safe
 286    ASCII characters (usually '?') for unsupported characters.  */
 287 #define CODING_MODE_SAFE_ENCODING               0x20
 288
 289 /* Structure of the field `spec.iso_2022' in the structure
 290    `coding_system'.  */
 291 struct iso_2022_spec
 292 {
 293   /* Bit-wise-or of CODING_ISO_FLAG_XXX.  */
 294   unsigned flags;
 295
 296   /* The current graphic register invoked to each graphic plane.  */
 297   int current_invocation[2];
 298
 299   /* The current charset designated to each graphic register.  The
 300      value -1 means that not charset is designated, -2 means that
 301      there was an invalid designation previously.  */
 302   int current_designation[4];
 303
 304   /* Set to 1 temporarily only when graphic register 2 or 3 is invoked
 305      by single-shift while encoding.  */
 306   int single_shifting;
 307
 308   /* Set to 1 temporarily only when processing at beginning of line.  */
 309   int bol;
 310 };
 311
 312 struct ccl_spec;
 313
 314 enum utf_16_bom_type
 315   {
 316     utf_16_detect_bom,
 317     utf_16_without_bom,
 318     utf_16_with_bom
 319   };
 320
 321 enum utf_16_endian_type
 322   {
 323     utf_16_big_endian,
 324     utf_16_little_endian
 325   };
 326
 327 struct utf_16_spec
 328 {
 329   enum utf_16_bom_type bom;
 330   enum utf_16_endian_type endian;
 331   int surrogate;
 332 };
 333
 334 struct coding_detection_info
 335 {
 336   /* Values of these members are bitwise-OR of CATEGORY_MASK_XXXs.  */
 337   /* Which categories are already checked.  */
 338   int checked;
 339   /* Which categories are strongly found.  */
 340   int found;
 341   /* Which categories are rejected.  */
 342   int rejected;
 343 };
 344
 345
 346 struct coding_system
 347 {
 348   /* ID number of the coding system.  This is an index to
 349      Vcoding_system_hash_table.  This value is set by
 350      setup_coding_system.  At the early stage of building time, this
 351      value is -1 in the array coding_categories to indicate that no
 352      coding-system of that category is yet defined.  */
 353   int id;
 354
 355   /* Flag bits of the coding system.  The meaning of each bit is common
 356      to all types of coding systems.  */
 357   int common_flags;
 358
 359   /* Mode bits of the coding system.  See the comments of the macros
 360      CODING_MODE_XXX.  */
 361   unsigned int mode;
 362
 363   /* Detailed information specific to each type of coding system.  */
 364   union
 365     {
 366       struct iso_2022_spec iso_2022;
 367       struct ccl_spec *ccl;     /* Defined in ccl.h.  */
 368       struct utf_16_spec utf_16;
 369       int emacs_mule_full_support;
 370     } spec;
 371
 372   int max_charset_id;
 373   char *safe_charsets;
 374
 375   /* The following two members specify how binary 8-bit code 128..255
 376      are represented in source and destination text respectively.  1
 377      means they are represented by 2-byte sequence, 0 means they are
 378      represented by 1-byte as is (see the comment in character.h).  */
 379   unsigned src_multibyte : 1;
 380   unsigned dst_multibyte : 1;
 381
 382   /* How may heading bytes we can skip for decoding.  This is set to
 383      -1 in setup_coding_system, and updated by detect_coding.  So,
 384      when this is equal to the byte length of the text being
 385      converted, we can skip the actual conversion process.  */
 386   int head_ascii;
 387
 388   /* The following members are set by encoding/decoding routine.  */
 389   EMACS_INT produced, produced_char, consumed, consumed_char;
 390
 391   /* Number of error source data found in a decoding routine.  */
 392   int errors;
 393
 394   /* Store the positions of error source data. */
 395   EMACS_INT *error_positions;
 396
 397   /* Finish status of code conversion.  */
 398   enum coding_result_code result;
 399
 400   EMACS_INT src_pos, src_pos_byte, src_chars, src_bytes;
 401   Lisp_Object src_object;
 402   const unsigned char *source;
 403
 404   EMACS_INT dst_pos, dst_pos_byte, dst_bytes;
 405   Lisp_Object dst_object;
 406   unsigned char *destination;
 407
 408   /* Set to 1 iff the source of conversion is not in the member
 409      `charbuf', but at `src_object'.  */
 410   int chars_at_source;
 411
 412   /* If an element is non-negative, it is a character code.
 413
 414      If it is in the range -128..-1, it is a 8-bit character code
 415      minus 256.
 416
 417      If it is less than -128, it specifies the start of an annotation
 418      chunk.  The length of the chunk is -128 minus the value of the
 419      element.  The following elements are OFFSET, ANNOTATION-TYPE, and
 420      a sequence of actual data for the annotation.  OFFSET is a
 421      character position offset from dst_pos or src_pos,
 422      ANNOTATION-TYPE specfies the meaning of the annotation and how to
 423      handle the following data..  */
 424   int *charbuf;
 425   int charbuf_size, charbuf_used;
 426
 427   /* Set to 1 if charbuf contains an annotation.  */
 428   int annotated;
 429
 430   unsigned char carryover[64];
 431   int carryover_bytes;
 432
 433   int default_char;
 434
 435   int (*detector) P_ ((struct coding_system *,
 436                        struct coding_detection_info *));
 437   void (*decoder) P_ ((struct coding_system *));
 438   int (*encoder) P_ ((struct coding_system *));
 439 };
 440
 441 /* Meanings of bits in the member `common_flags' of the structure
 442    coding_system.  The lowest 8 bits are reserved for various kind of
 443    annotations (currently two of them are used).  */
 444 #define CODING_ANNOTATION_MASK                  0x00FF
 445 #define CODING_ANNOTATE_COMPOSITION_MASK        0x0001
 446 #define CODING_ANNOTATE_DIRECTION_MASK          0x0002
 447 #define CODING_ANNOTATE_CHARSET_MASK            0x0003
 448 #define CODING_FOR_UNIBYTE_MASK                 0x0100
 449 #define CODING_REQUIRE_FLUSHING_MASK            0x0200
 450 #define CODING_REQUIRE_DECODING_MASK            0x0400
 451 #define CODING_REQUIRE_ENCODING_MASK            0x0800
 452 #define CODING_REQUIRE_DETECTION_MASK           0x1000
 453 #define CODING_RESET_AT_BOL_MASK                0x2000
 454
 455 /* Return 1 if the coding context CODING requires annotaion
 456    handling.  */
 457 #define CODING_REQUIRE_ANNOTATION(coding) \
 458   ((coding)->common_flags & CODING_ANNOTATION_MASK)
 459
 460 /* Return 1 if the coding context CODING prefers decoding into unibyte.  */
 461 #define CODING_FOR_UNIBYTE(coding) \
 462   ((coding)->common_flags & CODING_FOR_UNIBYTE_MASK)
 463
 464 /* Return 1 if the coding context CODING requires specific code to be
 465    attached at the tail of converted text.  */
 466 #define CODING_REQUIRE_FLUSHING(coding) \
 467   ((coding)->common_flags & CODING_REQUIRE_FLUSHING_MASK)
 468
 469 /* Return 1 if the coding context CODING requires code conversion on
 470    decoding.  */
 471 #define CODING_REQUIRE_DECODING(coding) \
 472   ((coding)->dst_multibyte              \
 473    || (coding)->common_flags & CODING_REQUIRE_DECODING_MASK)
 474
 475
 476 /* Return 1 if the coding context CODING requires code conversion on
 477    encoding.  */
 478 #define CODING_REQUIRE_ENCODING(coding)                         \
 479   ((coding)->src_multibyte                                      \
 480    || (coding)->common_flags & CODING_REQUIRE_ENCODING_MASK     \
 481    || (coding)->mode & CODING_MODE_SELECTIVE_DISPLAY)
 482
 483
 484 /* Return 1 if the coding context CODING requires some kind of code
 485    detection.  */
 486 #define CODING_REQUIRE_DETECTION(coding) \
 487   ((coding)->common_flags & CODING_REQUIRE_DETECTION_MASK)
 488
 489 /* Return 1 if the coding context CODING requires code conversion on
 490    decoding or some kind of code detection.  */
 491 #define CODING_MAY_REQUIRE_DECODING(coding)     \
 492   (CODING_REQUIRE_DECODING (coding)             \
 493    || CODING_REQUIRE_DETECTION (coding))
 494
 495 /* Macros to decode or encode a character of JISX0208 in SJIS.  S1 and
 496    S2 are the 1st and 2nd position-codes of JISX0208 in SJIS coding
 497    system.  C1 and C2 are the 1st and 2nd position codes of Emacs'
 498    internal format.  */
 499
 500 #define SJIS_TO_JIS(code)                               \
 501   do {                                                  \
 502     int s1, s2, j1, j2;                                 \
 503                                                         \
 504     s1 = (code) >> 8, s2 = (code) & 0xFF;               \
 505                                                         \
 506     if (s2 >= 0x9F)                                     \
 507       (j1 = s1 * 2 - (s1 >= 0xE0 ? 0x160 : 0xE0),       \
 508        j2 = s2 - 0x7E);                                 \
 509     else                                                \
 510       (j1 = s1 * 2 - ((s1 >= 0xE0) ? 0x161 : 0xE1),     \
 511        j2 = s2 - ((s2 >= 0x7F) ? 0x20 : 0x1F));         \
 512     (code) = (j1 << 8) | j2;                            \
 513   } while (0)
 514
 515 #define SJIS_TO_JIS2(code)                              \
 516   do {                                                  \
 517     int s1, s2, j1, j2;                                 \
 518                                                         \
 519     s1 = (code) >> 8, s2 = (code) & 0xFF;               \
 520                                                         \
 521     if (s2 >= 0x9F)                                     \
 522       {                                                 \
 523         j1 = (s1 == 0xF0 ? 0x28                         \
 524               : s1 == 0xF1 ? 0x24                       \
 525               : s1 == 0xF2 ? 0x2C                       \
 526               : s1 == 0xF3 ? 0x2E                       \
 527               : 0x6E + (s1 - 0xF4) * 2);                \
 528         j2 = s2 - 0x7E;                                 \
 529       }                                                 \
 530     else                                                \
 531       {                                                 \
 532         j1 = (s1 <= 0xF2 ? 0x21 + (s1 - 0xF0) * 2       \
 533               : s1 <= 0xF4 ? 0x2D + (s1 - 0xF3) * 2     \
 534               : 0x6F + (s1 - 0xF5) * 2);                \
 535         j2 = s2 - ((s2 >= 0x7F ? 0x20 : 0x1F));         \
 536       }                                                 \
 537     (code) = (j1 << 8) | j2;                            \
 538   } while (0)
 539
 540
 541 #define JIS_TO_SJIS(code)                               \
 542   do {                                                  \
 543     int s1, s2, j1, j2;                                 \
 544                                                         \
 545     j1 = (code) >> 8, j2 = (code) & 0xFF;               \
 546     if (j1 & 1)                                         \
 547       (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x71 : 0xB1),       \
 548        s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F));         \
 549     else                                                \
 550       (s1 = j1 / 2 + ((j1 < 0x5F) ? 0x70 : 0xB0),       \
 551        s2 = j2 + 0x7E);                                 \
 552     (code) = (s1 << 8) | s2;                            \
 553   } while (0)
 554
 555 #define JIS_TO_SJIS2(code)                              \
 556   do {                                                  \
 557     int s1, s2, j1, j2;                                 \
 558                                                         \
 559     j1 = (code) >> 8, j2 = (code) & 0xFF;               \
 560     if (j1 & 1)                                         \
 561       {                                                 \
 562         s1 = (j1 <= 0x25 ? 0xF0 + (j1 - 0x21) / 2       \
 563               : j1 <= 0x27 ? 0xF3 + (j1 - 0x2D) / 2     \
 564               : 0xF5 + (j1 - 0x6F) / 2);                \
 565         s2 = j2 + ((j2 >= 0x60) ? 0x20 : 0x1F);         \
 566       }                                                 \
 567     else                                                \
 568       {                                                 \
 569         s1 = (j1 == 0x28 ? 0xF0                         \
 570               : j1 == 0x24 ? 0xF1                       \
 571               : j1 == 0x2C ? 0xF2                       \
 572               : j1 == 0x2E ? 0xF3                       \
 573               : 0xF4 + (j1 - 0x6E) / 2);                \
 574         s2 = j2 + 0x7E;                                 \
 575       }                                                 \
 576     (code) = (s1 << 8) | s2;                            \
 577   } while (0)
 578
 579 /* Encode the file name NAME using the specified coding system
 580    for file names, if any.  */
 581 #define ENCODE_FILE(name)                                                  \
 582   (! NILP (Vfile_name_coding_system)                                       \
 583    && !EQ (Vfile_name_coding_system, make_number (0))                      \
 584    ? code_convert_string_norecord (name, Vfile_name_coding_system, 1)      \
 585    : (! NILP (Vdefault_file_name_coding_system)                            \
 586       && !EQ (Vdefault_file_name_coding_system, make_number (0))           \
 587       ? code_convert_string_norecord (name, Vdefault_file_name_coding_system, 1) \
 588       : name))
 589
 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
 603 /* Encode the string STR using the specified coding system
 604    for system functions, if any.  */
 605 #define ENCODE_SYSTEM(str)                                                 \
 606   (! NILP (Vlocale_coding_system)                                          \
 607    && !EQ (Vlocale_coding_system, make_number (0))                         \
 608    ? code_convert_string_norecord (str, Vlocale_coding_system, 1)          \
 609    : str)
 610
 611 /* Decode the string STR using the specified coding system
 612    for system functions, if any.  */
 613 #define DECODE_SYSTEM(name)                                                \
 614   (! NILP (Vlocale_coding_system)                                          \
 615    && !EQ (Vlocale_coding_system, make_number (0))                         \
 616    ? code_convert_string_norecord (str, Vlocale_coding_system, 0)          \
 617    : str)
 618
 619 /* Used by the gtk menu code.  Note that this encodes utf-8, not
 620    utf-8-emacs, so it's not a no-op.  */
 621 #define ENCODE_UTF_8(str) code_convert_string_norecord (str, Qutf_8, 1)
 622
 623 /* Extern declarations.  */
 624 extern Lisp_Object code_conversion_save P_ ((int, int));
 625 extern int decoding_buffer_size P_ ((struct coding_system *, int));
 626 extern int encoding_buffer_size P_ ((struct coding_system *, int));
 627 extern void setup_coding_system P_ ((Lisp_Object, struct coding_system *));
 628 extern Lisp_Object coding_charset_list P_ ((struct coding_system *));
 629 extern void detect_coding P_ ((struct coding_system *));
 630 extern Lisp_Object code_convert_region P_ ((Lisp_Object, Lisp_Object,
 631                                             Lisp_Object, Lisp_Object,
 632                                             int, int));
 633 extern Lisp_Object code_convert_string P_ ((Lisp_Object, Lisp_Object,
 634                                             Lisp_Object, int, int, int));
 635 extern Lisp_Object code_convert_string_norecord P_ ((Lisp_Object, Lisp_Object,
 636                                                      int));
 637 extern Lisp_Object raw_text_coding_system P_ ((Lisp_Object));
 638 extern Lisp_Object coding_inherit_eol_type P_ ((Lisp_Object, Lisp_Object));
 639
 640 extern int decode_coding_gap P_ ((struct coding_system *,
 641                                   EMACS_INT, EMACS_INT));
 642 extern int encode_coding_gap P_ ((struct coding_system *,
 643                                   EMACS_INT, EMACS_INT));
 644 extern void decode_coding_object P_ ((struct coding_system *,
 645                                       Lisp_Object, EMACS_INT, EMACS_INT,
 646                                       EMACS_INT, EMACS_INT, Lisp_Object));
 647 extern void encode_coding_object P_ ((struct coding_system *,
 648                                       Lisp_Object, EMACS_INT, EMACS_INT,
 649                                       EMACS_INT, EMACS_INT, Lisp_Object));
 650
 651 /* Macros for backward compatibility.  */
 652
 653 #define decode_coding_region(coding, from, to)          \
 654   decode_coding_object (coding, Fcurrent_buffer (),     \
 655                         from, CHAR_TO_BYTE (from),      \
 656                         to, CHAR_TO_BYTE (to), Fcurrent_buffer ())
 657
 658
 659 #define encode_coding_region(coding, from, to)          \
 660   encode_coding_object (coding, Fcurrent_buffer (),     \
 661                         from, CHAR_TO_BYTE (from),      \
 662                         to, CHAR_TO_BYTE (to), Fcurrent_buffer ())
 663
 664
 665 #define decode_coding_string(coding, string, nocopy)                    \
 666   decode_coding_object (coding, string, 0, 0, XSTRING (string)->size,   \
 667                         STRING_BYTES (XSTRING (string)), Qt)
 668
 669 #define encode_coding_string(coding, string, nocopy)                    \
 670   (encode_coding_object (coding, string, 0, 0, XSTRING (string)->size,  \
 671                          STRING_BYTES (XSTRING (string)), Qt),          \
 672    (coding)->dst_object)
 673
 674
 675 #define decode_coding_c_string(coding, src, bytes, dst_object)          \
 676   do {                                                                  \
 677     (coding)->source = (src);                                           \
 678     (coding)->src_chars = (coding)->src_bytes = (bytes);                \
 679     decode_coding_object ((coding), Qnil, 0, 0, (bytes), (bytes),       \
 680                           (dst_object));                                \
 681   } while (0)
 682
 683
 684 extern Lisp_Object preferred_coding_system P_ (());
 685
 686
 687 extern Lisp_Object Qutf_8, Qutf_8_emacs;
 688
 689 extern Lisp_Object Qcoding_system, Qeol_type, Qcoding_category_index;
 690 extern Lisp_Object Qcoding_system_p;
 691 extern Lisp_Object Qraw_text, Qemacs_mule, Qno_conversion, Qundecided;
 692 extern Lisp_Object Qiso_2022;
 693 extern Lisp_Object Qbuffer_file_coding_system;
 694
 695 extern Lisp_Object Qunix, Qdos, Qmac;
 696
 697 extern Lisp_Object Qtranslation_table;
 698 extern Lisp_Object Qtranslation_table_id;
 699
 700 /* Mnemonic strings to indicate each type of end-of-line.  */
 701 extern Lisp_Object eol_mnemonic_unix, eol_mnemonic_dos, eol_mnemonic_mac;
 702 /* Mnemonic string to indicate type of end-of-line is not yet decided.  */
 703 extern Lisp_Object eol_mnemonic_undecided;
 704
 705 #ifdef emacs
 706 extern Lisp_Object Qfile_coding_system;
 707 extern Lisp_Object Qcall_process, Qcall_process_region, Qprocess_argument;
 708 extern Lisp_Object Qstart_process, Qopen_network_stream;
 709 extern Lisp_Object Qwrite_region;
 710
 711 extern char *emacs_strerror P_ ((int));
 712
 713 /* Coding-system for reading files and receiving data from process.  */
 714 extern Lisp_Object Vcoding_system_for_read;
 715 /* Coding-system for writing files and sending data to process.  */
 716 extern Lisp_Object Vcoding_system_for_write;
 717 /* Coding-system actually used in the latest I/O.  */
 718 extern Lisp_Object Vlast_coding_system_used;
 719 /* Coding-system to use with system messages (e.g. strerror).  */
 720 extern Lisp_Object Vlocale_coding_system;
 721
 722 /* If non-zero, process buffer inherits the coding system used to decode
 723    the subprocess output.  */
 724 extern int inherit_process_coding_system;
 725
 726 /* Coding-system to be used for encoding terminal output.  This
 727    structure contains information of a coding-system specified by the
 728    function `set-terminal-coding-system'.  */
 729 extern struct coding_system terminal_coding;
 730
 731 /* Coding system to be used to encode text for terminal display when
 732    terminal coding system is nil.  */
 733 extern struct coding_system safe_terminal_coding;
 734
 735 /* Coding-system of what is sent from terminal keyboard.  This
 736    structure contains information of a coding-system specified by the
 737    function `set-keyboard-coding-system'.  */
 738 extern struct coding_system keyboard_coding;
 739
 740 /* Default coding systems used for process I/O.  */
 741 extern Lisp_Object Vdefault_process_coding_system;
 742
 743 /* Function to call to force a user to force select a propert coding
 744    system.  */
 745 extern Lisp_Object Vselect_safe_coding_system_function;
 746
 747 /* If nonzero, on writing a file, Vselect_safe_coding_system_function
 748    is called even if Vcoding_system_for_write is non-nil.  */
 749 extern int coding_system_require_warning;
 750
 751 /* Coding system for file names, or nil if none.  */
 752 extern Lisp_Object Vfile_name_coding_system;
 753
 754 /* Coding system for file names used only when
 755    Vfile_name_coding_system is nil.  */
 756 extern Lisp_Object Vdefault_file_name_coding_system;
 757
 758 #endif
 759
 760 /* Error signaled when there's a problem with detecting coding system */
 761 extern Lisp_Object Qcoding_system_error;
 762
 763 extern char emacs_mule_bytes[256];
 764 extern int emacs_mule_string_char P_ ((unsigned char *));
 765
 766 #endif /* EMACS_CODING_H */
 767
 768 /* arch-tag: 2bc3b4fa-6870-4f64-8135-b962b2d290e4
 769    (do not change this comment) */