iconvdata/iso-2022-cn.c

   1 /* Conversion module for ISO-2022-CN.
   2    Copyright (C) 1999, 2000 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1999.
   5
   6    The GNU C Library is free software; you can redistribute it and/or
   7    modify it under the terms of the GNU Library General Public License as
   8    published by the Free Software Foundation; either version 2 of the
   9    License, or (at your option) any later version.
  10
  11    The GNU C Library is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14    Library General Public License for more details.
  15
  16    You should have received a copy of the GNU Library General Public
  17    License along with the GNU C Library; see the file COPYING.LIB.  If not,
  18    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  19    Boston, MA 02111-1307, USA.  */
  20
  21 #include <gconv.h>
  22 #include <stdint.h>
  23 #include <string.h>
  24 #include "gb2312.h"
  25 #include "cns11643l1.h"
  26 #include "cns11643l2.h"
  27
  28 #include <assert.h>
  29
  30 /* This makes obvious what everybody knows: 0x1b is the Esc character.  */
  31 #define ESC     0x1b
  32
  33 /* We have single-byte shift-in and shift-out sequences, and the single
  34    shift sequence SS2 which replaces the SS2 designation for the next
  35    two bytes.  */
  36 #define SI      0x0f
  37 #define SO      0x0e
  38 #define SS2_0   ESC
  39 #define SS2_1   0x4e
  40
  41 /* Definitions used in the body of the `gconv' function.  */
  42 #define CHARSET_NAME            "ISO-2022-CN//"
  43 #define DEFINE_INIT             1
  44 #define DEFINE_FINI             1
  45 #define FROM_LOOP               from_iso2022cn_loop
  46 #define TO_LOOP                 to_iso2022cn_loop
  47 #define MIN_NEEDED_FROM         1
  48 #define MAX_NEEDED_FROM         4
  49 #define MIN_NEEDED_TO           4
  50 #define MAX_NEEDED_TO           4
  51 #define PREPARE_LOOP \
  52   int save_set;                                                               \
  53   int *setp = &data->__statep->__count;
  54 #define EXTRA_LOOP_ARGS         , setp
  55
  56
  57 /* The COUNT element of the state keeps track of the currently selected
  58    character set.  The possible values are:  */
  59 enum
  60 {
  61   ASCII_set = 0,
  62   GB2312_set = 8,
  63   CNS11643_1_set = 16,
  64   CNS11643_2_set = 24,
  65   CURRENT_SEL_MASK = 24,
  66   GB2312_ann = 32,
  67   CNS11643_1_ann = 64,
  68   CNS11643_2_ann = 128,
  69   CURRENT_ANN_MASK = 224
  70 };
  71
  72
  73 /* Since this is a stateful encoding we have to provide code which resets
  74    the output state to the initial state.  This has to be done during the
  75    flushing.  */
  76 #define EMIT_SHIFT_TO_INIT \
  77   if (data->__statep->__count != ASCII_set)                                   \
  78     {                                                                         \
  79       if (FROM_DIRECTION)                                                     \
  80         /* It's easy, we don't have to emit anything, we just reset the       \
  81            state for the input.  */                                           \
  82         data->__statep->__count = ASCII_set;                                  \
  83       else                                                                    \
  84         {                                                                     \
  85           unsigned char *outbuf = data->__outbuf;                             \
  86                                                                               \
  87           /* We are not in the initial state.  To switch back we have         \
  88              to emit `SI'.  */                                                \
  89           if (__builtin_expect (outbuf == data->__outbufend, 0))              \
  90             /* We don't have enough room in the output buffer.  */            \
  91             status = __GCONV_FULL_OUTPUT;                                     \
  92           else                                                                \
  93             {                                                                 \
  94               /* Write out the shift sequence.  */                            \
  95               *outbuf++ = SI;                                                 \
  96               data->__outbuf = outbuf;                                        \
  97               data->__statep->__count = ASCII_set;                            \
  98             }                                                                 \
  99         }                                                                     \
 100     }
 101
 102
 103 /* Since we might have to reset input pointer we must be able to save
 104    and retore the state.  */
 105 #define SAVE_RESET_STATE(Save) \
 106   if (Save)                                                                   \
 107     save_set = *setp;                                                         \
 108   else                                                                        \
 109     *setp = save_set
 110
 111
 112 /* First define the conversion function from ISO-2022-CN to UCS4.  */
 113 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 114 #define MAX_NEEDED_INPUT        MAX_NEEDED_FROM
 115 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 116 #define LOOPFCT                 FROM_LOOP
 117 #define BODY \
 118   {                                                                           \
 119     uint32_t ch = *inptr;                                                     \
 120                                                                               \
 121     /* This is a 7bit character set, disallow all 8bit characters.  */        \
 122     if (__builtin_expect (ch, 0) > 0x7f)                                      \
 123       {                                                                       \
 124         if (! ignore_errors_p ())                                             \
 125           {                                                                   \
 126             result = __GCONV_ILLEGAL_INPUT;                                   \
 127             break;                                                            \
 128           }                                                                   \
 129                                                                               \
 130         ++inptr;                                                              \
 131         ++*converted;                                                         \
 132         continue;                                                             \
 133       }                                                                       \
 134                                                                               \
 135     /* Recognize escape sequences.  */                                        \
 136     if (__builtin_expect (ch, 0) == ESC)                                      \
 137       {                                                                       \
 138         /* There are two kinds of escape sequences we have to handle:         \
 139            - those announcing the use of GB and CNS characters on the         \
 140              line; we can simply ignore them                                  \
 141            - the initial byte of the SS2 sequence.                            \
 142         */                                                                    \
 143         if (NEED_LENGTH_TEST                                                  \
 144             && (__builtin_expect (inptr + 1 > inend, 0)                       \
 145                 || (inptr[1] == '$'                                           \
 146                     && (__builtin_expect (inptr + 2 > inend, 0)               \
 147                         || (inptr[2] == ')'                                   \
 148                             && __builtin_expect (inptr + 3 > inend, 0))       \
 149                         || (inptr[2] == '*'                                   \
 150                             && __builtin_expect (inptr + 3 > inend, 0))))     \
 151                 || (inptr[1] == SS2_1                                         \
 152                     && __builtin_expect (inptr + 3 > inend, 0))))             \
 153           {                                                                   \
 154             result = __GCONV_EMPTY_INPUT;                                     \
 155             break;                                                            \
 156           }                                                                   \
 157         if (inptr[1] == '$'                                                   \
 158             && ((inptr[2] == ')' && (inptr[3] == 'A' || inptr[3] == 'G'))     \
 159                 || (inptr[2] == '*' && inptr[3] == 'H')))                     \
 160           {                                                                   \
 161             /* OK, we accept those character sets.  */                        \
 162             if (inptr[3] == 'A')                                              \
 163               ann = GB2312_ann;                                               \
 164             else if (inptr[3] == 'G')                                         \
 165               ann = CNS11643_1_ann;                                           \
 166             inptr += 4;                                                       \
 167             continue;                                                         \
 168           }                                                                   \
 169       }                                                                       \
 170     else if (__builtin_expect (ch, 0) == SO)                                  \
 171       {                                                                       \
 172         /* Switch to use GB2312 or CNS 11643 plane 1, depending on which      \
 173            S0 designation came last.  The only problem is what to do with     \
 174            faulty input files where no designator came.                       \
 175            XXX For now I'll default to use GB2312.  If this is not the        \
 176            best behaviour (e.g., we should flag an error) let me know.  */    \
 177         ++inptr;                                                              \
 178         set = ann == CNS11643_1_ann ? CNS11643_1_set : GB2312_set;            \
 179         continue;                                                             \
 180       }                                                                       \
 181     else if (__builtin_expect (ch, 0) == SI)                                  \
 182       {                                                                       \
 183         /* Switch to use ASCII.  */                                           \
 184         ++inptr;                                                              \
 185         set = ASCII_set;                                                      \
 186         continue;                                                             \
 187       }                                                                       \
 188                                                                               \
 189     if (__builtin_expect (ch, 0) == ESC && inptr[1] == SS2_1)                 \
 190       {                                                                       \
 191         /* This is a character from CNS 11643 plane 2.                        \
 192            XXX We could test here whether the use of this character           \
 193            set was announced.  */                                             \
 194         inptr += 2;                                                           \
 195         ch = cns11643l2_to_ucs4 (&inptr, 2, 0);                               \
 196         if (__builtin_expect (ch, 0) == __UNKNOWN_10646_CHAR)                 \
 197           {                                                                   \
 198             if (! ignore_errors_p ())                                         \
 199               {                                                               \
 200                 /* This is an illegal character.  */                          \
 201                 inptr -= 2;                                                   \
 202                 result = __GCONV_ILLEGAL_INPUT;                               \
 203                 break;                                                        \
 204               }                                                               \
 205                                                                               \
 206             ++*converted;                                                     \
 207             continue;                                                         \
 208           }                                                                   \
 209       }                                                                       \
 210     else if (set == ASCII_set)                                                \
 211       {                                                                       \
 212         /* Almost done, just advance the input pointer.  */                   \
 213         ++inptr;                                                              \
 214       }                                                                       \
 215     else                                                                      \
 216       {                                                                       \
 217         /* That's pretty easy, we have a dedicated functions for this.  */    \
 218         if (set == GB2312_set)                                                \
 219           ch = gb2312_to_ucs4 (&inptr,                                        \
 220                                NEED_LENGTH_TEST ? inend - inptr : 2, 0);      \
 221         else                                                                  \
 222           {                                                                   \
 223             assert (set == CNS11643_1_set);                                   \
 224             ch = cns11643l1_to_ucs4 (&inptr,                                  \
 225                                      NEED_LENGTH_TEST ? inend - inptr : 2, 0);\
 226           }                                                                   \
 227                                                                               \
 228         if (NEED_LENGTH_TEST && __builtin_expect (ch, 1) == 0)                \
 229           {                                                                   \
 230             result = __GCONV_EMPTY_INPUT;                                     \
 231             break;                                                            \
 232           }                                                                   \
 233         else if (__builtin_expect (ch, 1) == __UNKNOWN_10646_CHAR)            \
 234           {                                                                   \
 235             if (! ignore_errors_p ())                                         \
 236               {                                                               \
 237                 /* This is an illegal character.  */                          \
 238                 result = __GCONV_ILLEGAL_INPUT;                               \
 239                 break;                                                        \
 240               }                                                               \
 241                                                                               \
 242             ++inptr;                                                          \
 243             ++*converted;                                                     \
 244             continue;                                                         \
 245           }                                                                   \
 246       }                                                                       \
 247                                                                               \
 248     put32 (outptr, ch);                                                       \
 249     outptr += 4;                                                              \
 250   }
 251 #define EXTRA_LOOP_DECLS        , int *setp
 252 #define INIT_PARAMS             int set = *setp & CURRENT_SEL_MASK; \
 253                                 int ann = *setp & CURRENT_ANN_MASK
 254 #define UPDATE_PARAMS           *setp = set | ann
 255 #include <iconv/loop.c>
 256
 257
 258 /* Next, define the other direction.  */
 259 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 260 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 261 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 262 #define LOOPFCT                 TO_LOOP
 263 #define BODY \
 264   {                                                                           \
 265     uint32_t ch = get32 (inptr);                                              \
 266                                                                               \
 267     /* First see whether we can write the character using the currently       \
 268        selected character set.  */                                            \
 269     if (ch < 0x80)                                                            \
 270       {                                                                       \
 271         if (set != ASCII_set)                                                 \
 272           {                                                                   \
 273             *outptr++ = SI;                                                   \
 274             set = ASCII_set;                                                  \
 275             if (NEED_LENGTH_TEST && __builtin_expect (outptr == outend, 0))   \
 276               {                                                               \
 277                 result = __GCONV_FULL_OUTPUT;                                 \
 278                 break;                                                        \
 279               }                                                               \
 280           }                                                                   \
 281                                                                               \
 282         *outptr++ = ch;                                                       \
 283                                                                               \
 284         /* At the end of the line we have to clear the `ann' flags since      \
 285            every line must contain this information again.  */                \
 286         if (ch == L'\n')                                                      \
 287           ann = 0;                                                            \
 288       }                                                                       \
 289     else                                                                      \
 290       {                                                                       \
 291         char buf[2];                                                          \
 292         int used;                                                             \
 293         size_t written = 0;                                                   \
 294                                                                               \
 295         if (set == GB2312_set || (ann & CNS11643_1_ann) == 0)                 \
 296           {                                                                   \
 297             written = ucs4_to_gb2312 (ch, buf, 2);                            \
 298             used = GB2312_set;                                                \
 299           }                                                                   \
 300         else                                                                  \
 301           {                                                                   \
 302             written = ucs4_to_cns11643l1 (ch, buf, 2);                        \
 303             used = CNS11643_1_set;                                            \
 304           }                                                                   \
 305                                                                               \
 306         if (written == __UNKNOWN_10646_CHAR)                                  \
 307           {                                                                   \
 308             /* Cannot convert it using the currently selected SO set.         \
 309                Next try the SS2 set.  */                                      \
 310             written = ucs4_to_cns11643l2 (ch, buf, 2);                        \
 311             if (written != __UNKNOWN_10646_CHAR)                              \
 312               /* Yep, that worked.  */                                        \
 313               used = CNS11643_2_set;                                          \
 314             else                                                              \
 315               {                                                               \
 316                 /* Well, see whether we have to change the SO set.  */        \
 317                 if (set == GB2312_set)                                        \
 318                   written = ucs4_to_cns11643l1 (ch, buf, 2);                  \
 319                 else                                                          \
 320                   written = ucs4_to_gb2312 (ch, buf, 2);                      \
 321                                                                               \
 322                 if (__builtin_expect (written, 0) != __UNKNOWN_10646_CHAR)    \
 323                   /* Oh well, then switch SO.  */                             \
 324                   used = GB2312_set + CNS11643_1_set - set;                   \
 325                 else                                                          \
 326                   {                                                           \
 327                     /* Even this does not work.  Error.  */                   \
 328                     if (! ignore_errors_p ())                                 \
 329                       {                                                       \
 330                         result = __GCONV_ILLEGAL_INPUT;                       \
 331                         break;                                                \
 332                       }                                                       \
 333                                                                               \
 334                     inptr += 4;                                               \
 335                     ++*converted;                                             \
 336                     continue;                                                 \
 337                   }                                                           \
 338               }                                                               \
 339           }                                                                   \
 340         assert (written == 2);                                                \
 341                                                                               \
 342         /* See whether we have to emit an escape sequence.  */                \
 343         if (set != used)                                                      \
 344           {                                                                   \
 345             /* First see whether we announced that we use this                \
 346                character set.  */                                             \
 347             if ((ann & (2 << used)) == 0)                                     \
 348               {                                                               \
 349                 const char *escseq;                                           \
 350                                                                               \
 351                 if (NEED_LENGTH_TEST                                          \
 352                     && __builtin_expect (outptr + 4 > outend, 0))             \
 353                   {                                                           \
 354                     result = __GCONV_FULL_OUTPUT;                             \
 355                     break;                                                    \
 356                   }                                                           \
 357                                                                               \
 358                 assert (used >= 1 && used <= 3);                              \
 359                 escseq = "\e$)A\e$)G\e$*H" + (used - 1) * 4;                  \
 360                 *outptr++ = *escseq++;                                        \
 361                 *outptr++ = *escseq++;                                        \
 362                 *outptr++ = *escseq++;                                        \
 363                 *outptr++ = *escseq++;                                        \
 364                                                                               \
 365                 if (used == GB2312_set)                                       \
 366                   ann = (ann & CNS11643_2_ann) | GB2312_ann;                  \
 367                 else if (used == CNS11643_1_set)                              \
 368                   ann = (ann & CNS11643_2_ann) | CNS11643_1_ann;              \
 369                 else                                                          \
 370                   ann |= CNS11643_2_ann;                                      \
 371               }                                                               \
 372                                                                               \
 373             if (used == CNS11643_2_set)                                       \
 374               {                                                               \
 375                 if (__builtin_expect (outptr + 2 > outend, 0))                \
 376                   {                                                           \
 377                     result = __GCONV_FULL_OUTPUT;                             \
 378                     break;                                                    \
 379                   }                                                           \
 380                 *outptr++ = SS2_0;                                            \
 381                 *outptr++ = SS2_1;                                            \
 382               }                                                               \
 383             else                                                              \
 384               {                                                               \
 385                 /* We only have to emit something is currently ASCII is       \
 386                    selected.  Otherwise we are switching within the           \
 387                    SO charset.  */                                            \
 388                 if (set == ASCII_set)                                         \
 389                   {                                                           \
 390                     if (__builtin_expect (outptr + 1 > outend, 0))            \
 391                       {                                                       \
 392                         result = __GCONV_FULL_OUTPUT;                         \
 393                         break;                                                \
 394                       }                                                       \
 395                     *outptr++ = SO;                                           \
 396                   }                                                           \
 397               }                                                               \
 398                                                                               \
 399             /* Always test the length here since we have used up all the      \
 400                guaranteed output buffer slots.  */                            \
 401             if (__builtin_expect (outptr + 2 > outend, 0))                    \
 402               {                                                               \
 403                 result = __GCONV_FULL_OUTPUT;                                 \
 404                 break;                                                        \
 405               }                                                               \
 406           }                                                                   \
 407         else if (NEED_LENGTH_TEST                                             \
 408                  && __builtin_expect (outptr + 2 > outend, 0))                \
 409           {                                                                   \
 410             result = __GCONV_FULL_OUTPUT;                                     \
 411             break;                                                            \
 412           }                                                                   \
 413                                                                               \
 414         *outptr++ = buf[0];                                                   \
 415         *outptr++ = buf[1];                                                   \
 416       }                                                                       \
 417                                                                               \
 418     /* Now that we wrote the output increment the input pointer.  */          \
 419     inptr += 4;                                                               \
 420   }
 421 #define EXTRA_LOOP_DECLS        , int *setp
 422 #define INIT_PARAMS             int set = *setp & CURRENT_SEL_MASK; \
 423                                 int ann = *setp & CURRENT_ANN_MASK
 424 #define UPDATE_PARAMS           *setp = set | ann
 425 #include <iconv/loop.c>
 426
 427
 428 /* Now define the toplevel functions.  */
 429 #include <iconv/skeleton.c>