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 <dlfcn.h>
  22 #include <gconv.h>
  23 #include <stdint.h>
  24 #include <string.h>
  25 #include "gb2312.h"
  26 #include "cns11643l1.h"
  27 #include "cns11643l2.h"
  28
  29 #include <assert.h>
  30
  31 /* This makes obvious what everybody knows: 0x1b is the Esc character.  */
  32 #define ESC     0x1b
  33
  34 /* We have single-byte shift-in and shift-out sequences, and the single
  35    shift sequence SS2 which replaces the SS2 designation for the next
  36    two bytes.  */
  37 #define SI      0x0f
  38 #define SO      0x0e
  39 #define SS2_0   ESC
  40 #define SS2_1   0x4e
  41
  42 /* Definitions used in the body of the `gconv' function.  */
  43 #define CHARSET_NAME            "ISO-2022-CN//"
  44 #define DEFINE_INIT             1
  45 #define DEFINE_FINI             1
  46 #define FROM_LOOP               from_iso2022cn_loop
  47 #define TO_LOOP                 to_iso2022cn_loop
  48 #define MIN_NEEDED_FROM         1
  49 #define MAX_NEEDED_FROM         4
  50 #define MIN_NEEDED_TO           4
  51 #define MAX_NEEDED_TO           4
  52 #define PREPARE_LOOP \
  53   int save_set;                                                               \
  54   int *setp = &data->__statep->__count;
  55 #define EXTRA_LOOP_ARGS         , setp
  56
  57
  58 /* The COUNT element of the state keeps track of the currently selected
  59    character set.  The possible values are:  */
  60 enum
  61 {
  62   ASCII_set = 0,
  63   GB2312_set = 8,
  64   CNS11643_1_set = 16,
  65   CNS11643_2_set = 24,
  66   CURRENT_SEL_MASK = 24,
  67   GB2312_ann = 32,
  68   CNS11643_1_ann = 64,
  69   CNS11643_2_ann = 128,
  70   CURRENT_ANN_MASK = 224
  71 };
  72
  73
  74 /* Since this is a stateful encoding we have to provide code which resets
  75    the output state to the initial state.  This has to be done during the
  76    flushing.  */
  77 #define EMIT_SHIFT_TO_INIT \
  78   if (data->__statep->__count != ASCII_set)                                   \
  79     {                                                                         \
  80       if (FROM_DIRECTION)                                                     \
  81         /* It's easy, we don't have to emit anything, we just reset the       \
  82            state for the input.  */                                           \
  83         data->__statep->__count = ASCII_set;                                  \
  84       else                                                                    \
  85         {                                                                     \
  86           unsigned char *outbuf = data->__outbuf;                             \
  87                                                                               \
  88           /* We are not in the initial state.  To switch back we have         \
  89              to emit `SI'.  */                                                \
  90           if (__builtin_expect (outbuf == data->__outbufend, 0))              \
  91             /* We don't have enough room in the output buffer.  */            \
  92             status = __GCONV_FULL_OUTPUT;                                     \
  93           else                                                                \
  94             {                                                                 \
  95               /* Write out the shift sequence.  */                            \
  96               *outbuf++ = SI;                                                 \
  97               data->__outbuf = outbuf;                                        \
  98               data->__statep->__count = ASCII_set;                            \
  99             }                                                                 \
 100         }                                                                     \
 101     }
 102
 103
 104 /* Since we might have to reset input pointer we must be able to save
 105    and retore the state.  */
 106 #define SAVE_RESET_STATE(Save) \
 107   if (Save)                                                                   \
 108     save_set = *setp;                                                         \
 109   else                                                                        \
 110     *setp = save_set
 111
 112
 113 /* First define the conversion function from ISO-2022-CN to UCS4.  */
 114 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 115 #define MAX_NEEDED_INPUT        MAX_NEEDED_FROM
 116 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 117 #define LOOPFCT                 FROM_LOOP
 118 #define BODY \
 119   {                                                                           \
 120     uint32_t ch = *inptr;                                                     \
 121                                                                               \
 122     /* This is a 7bit character set, disallow all 8bit characters.  */        \
 123     if (__builtin_expect (ch, 0) > 0x7f)                                      \
 124       {                                                                       \
 125         if (! ignore_errors_p ())                                             \
 126           {                                                                   \
 127             result = __GCONV_ILLEGAL_INPUT;                                   \
 128             break;                                                            \
 129           }                                                                   \
 130                                                                               \
 131         ++inptr;                                                              \
 132         ++*irreversible;                                                      \
 133         continue;                                                             \
 134       }                                                                       \
 135                                                                               \
 136     /* Recognize escape sequences.  */                                        \
 137     if (__builtin_expect (ch, 0) == ESC)                                      \
 138       {                                                                       \
 139         /* There are two kinds of escape sequences we have to handle:         \
 140            - those announcing the use of GB and CNS characters on the         \
 141              line; we can simply ignore them                                  \
 142            - the initial byte of the SS2 sequence.                            \
 143         */                                                                    \
 144         if (__builtin_expect (inptr + 2 > inend, 0)                           \
 145             || (inptr[1] == '$'                                               \
 146                 && (__builtin_expect (inptr + 3 > inend, 0)                   \
 147                     || (inptr[2] == ')'                                       \
 148                         && __builtin_expect (inptr + 4 > inend, 0))           \
 149                     || (inptr[2] == '*'                                       \
 150                         && __builtin_expect (inptr + 4 > inend, 0))))         \
 151             || (inptr[1] == SS2_1                                             \
 152                 && __builtin_expect (inptr + 4 > inend, 0)))                  \
 153           {                                                                   \
 154             result = __GCONV_INCOMPLETE_INPUT;                                \
 155             break;                                                            \
 156           }                                                                   \
 157         if (inptr[1] == '$'                                                   \
 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             ++*irreversible;                                                  \
 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, inend - inptr, 0);                     \
 220         else                                                                  \
 221           {                                                                   \
 222             assert (set == CNS11643_1_set);                                   \
 223             ch = cns11643l1_to_ucs4 (&inptr, inend - inptr, 0);               \
 224           }                                                                   \
 225                                                                               \
 226         if (__builtin_expect (ch, 1) == 0)                                    \
 227           {                                                                   \
 228             result = __GCONV_INCOMPLETE_INPUT;                                \
 229             break;                                                            \
 230           }                                                                   \
 231         else if (__builtin_expect (ch, 1) == __UNKNOWN_10646_CHAR)            \
 232           {                                                                   \
 233             if (! ignore_errors_p ())                                         \
 234               {                                                               \
 235                 /* This is an illegal character.  */                          \
 236                 result = __GCONV_ILLEGAL_INPUT;                               \
 237                 break;                                                        \
 238               }                                                               \
 239                                                                               \
 240             ++inptr;                                                          \
 241             ++*irreversible;                                                  \
 242             continue;                                                         \
 243           }                                                                   \
 244       }                                                                       \
 245                                                                               \
 246     put32 (outptr, ch);                                                       \
 247     outptr += 4;                                                              \
 248   }
 249 #define LOOP_NEED_FLAGS
 250 #define EXTRA_LOOP_DECLS        , int *setp
 251 #define INIT_PARAMS             int set = *setp & CURRENT_SEL_MASK; \
 252                                 int ann = *setp & CURRENT_ANN_MASK
 253 #define UPDATE_PARAMS           *setp = set | ann
 254 #include <iconv/loop.c>
 255
 256
 257 /* Next, define the other direction.  */
 258 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 259 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 260 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 261 #define LOOPFCT                 TO_LOOP
 262 #define BODY \
 263   {                                                                           \
 264     uint32_t ch = get32 (inptr);                                              \
 265                                                                               \
 266     /* First see whether we can write the character using the currently       \
 267        selected character set.  */                                            \
 268     if (ch < 0x80)                                                            \
 269       {                                                                       \
 270         if (set != ASCII_set)                                                 \
 271           {                                                                   \
 272             *outptr++ = SI;                                                   \
 273             set = ASCII_set;                                                  \
 274             if (__builtin_expect (outptr == outend, 0))                       \
 275               {                                                               \
 276                 result = __GCONV_FULL_OUTPUT;                                 \
 277                 break;                                                        \
 278               }                                                               \
 279           }                                                                   \
 280                                                                               \
 281         *outptr++ = ch;                                                       \
 282                                                                               \
 283         /* At the end of the line we have to clear the `ann' flags since      \
 284            every line must contain this information again.  */                \
 285         if (ch == L'\n')                                                      \
 286           ann = 0;                                                            \
 287       }                                                                       \
 288     else                                                                      \
 289       {                                                                       \
 290         char buf[2];                                                          \
 291         int used;                                                             \
 292         size_t written = 0;                                                   \
 293                                                                               \
 294         if (set == GB2312_set || (ann & CNS11643_1_ann) == 0)                 \
 295           {                                                                   \
 296             written = ucs4_to_gb2312 (ch, buf, 2);                            \
 297             used = GB2312_set;                                                \
 298           }                                                                   \
 299         else                                                                  \
 300           {                                                                   \
 301             written = ucs4_to_cns11643l1 (ch, buf, 2);                        \
 302             used = CNS11643_1_set;                                            \
 303           }                                                                   \
 304                                                                               \
 305         if (written == __UNKNOWN_10646_CHAR)                                  \
 306           {                                                                   \
 307             /* Cannot convert it using the currently selected SO set.         \
 308                Next try the SS2 set.  */                                      \
 309             written = ucs4_to_cns11643l2 (ch, buf, 2);                        \
 310             if (written != __UNKNOWN_10646_CHAR)                              \
 311               /* Yep, that worked.  */                                        \
 312               used = CNS11643_2_set;                                          \
 313             else                                                              \
 314               {                                                               \
 315                 /* Well, see whether we have to change the SO set.  */        \
 316                 if (used == GB2312_set)                                       \
 317                   written = ucs4_to_cns11643l1 (ch, buf, 2);                  \
 318                 else                                                          \
 319                   written = ucs4_to_gb2312 (ch, buf, 2);                      \
 320                                                                               \
 321                 if (__builtin_expect (written, 0) != __UNKNOWN_10646_CHAR)    \
 322                   /* Oh well, then switch SO.  */                             \
 323                   used = GB2312_set + CNS11643_1_set - used;                  \
 324                 else                                                          \
 325                   {                                                           \
 326                     /* Even this does not work.  Error.  */                   \
 327                     STANDARD_ERR_HANDLER (4);                                 \
 328                   }                                                           \
 329               }                                                               \
 330           }                                                                   \
 331         assert (written == 2);                                                \
 332                                                                               \
 333         /* See whether we have to emit an escape sequence.  */                \
 334         if (set != used)                                                      \
 335           {                                                                   \
 336             /* First see whether we announced that we use this                \
 337                character set.  */                                             \
 338             if ((ann & (16 << (used >> 3))) == 0)                             \
 339               {                                                               \
 340                 const char *escseq;                                           \
 341                                                                               \
 342                 if (__builtin_expect (outptr + 4 > outend, 0))                \
 343                   {                                                           \
 344                     result = __GCONV_FULL_OUTPUT;                             \
 345                     break;                                                    \
 346                   }                                                           \
 347                                                                               \
 348                 assert ((used >> 3) >= 1 && (used >> 3) <= 3);                \
 349                 escseq = ")A)G*H" + ((used >> 3) - 1) * 2;                    \
 350                 *outptr++ = ESC;                                              \
 351                 *outptr++ = '$';                                              \
 352                 *outptr++ = *escseq++;                                        \
 353                 *outptr++ = *escseq++;                                        \
 354                                                                               \
 355                 if (used == GB2312_set)                                       \
 356                   ann = (ann & CNS11643_2_ann) | GB2312_ann;                  \
 357                 else if (used == CNS11643_1_set)                              \
 358                   ann = (ann & CNS11643_2_ann) | CNS11643_1_ann;              \
 359                 else                                                          \
 360                   ann |= CNS11643_2_ann;                                      \
 361               }                                                               \
 362                                                                               \
 363             if (used == CNS11643_2_set)                                       \
 364               {                                                               \
 365                 if (__builtin_expect (outptr + 2 > outend, 0))                \
 366                   {                                                           \
 367                     result = __GCONV_FULL_OUTPUT;                             \
 368                     break;                                                    \
 369                   }                                                           \
 370                 *outptr++ = SS2_0;                                            \
 371                 *outptr++ = SS2_1;                                            \
 372               }                                                               \
 373             else                                                              \
 374               {                                                               \
 375                 /* We only have to emit something is currently ASCII is       \
 376                    selected.  Otherwise we are switching within the           \
 377                    SO charset.  */                                            \
 378                 if (set == ASCII_set)                                         \
 379                   {                                                           \
 380                     if (__builtin_expect (outptr + 1 > outend, 0))            \
 381                       {                                                       \
 382                         result = __GCONV_FULL_OUTPUT;                         \
 383                         break;                                                \
 384                       }                                                       \
 385                     *outptr++ = SO;                                           \
 386                   }                                                           \
 387               }                                                               \
 388                                                                               \
 389             /* Always test the length here since we have used up all the      \
 390                guaranteed output buffer slots.  */                            \
 391             if (__builtin_expect (outptr + 2 > outend, 0))                    \
 392               {                                                               \
 393                 result = __GCONV_FULL_OUTPUT;                                 \
 394                 break;                                                        \
 395               }                                                               \
 396           }                                                                   \
 397         else if (__builtin_expect (outptr + 2 > outend, 0))                   \
 398           {                                                                   \
 399             result = __GCONV_FULL_OUTPUT;                                     \
 400             break;                                                            \
 401           }                                                                   \
 402                                                                               \
 403         *outptr++ = buf[0];                                                   \
 404         *outptr++ = buf[1];                                                   \
 405         set = used;                                                           \
 406       }                                                                       \
 407                                                                               \
 408     /* Now that we wrote the output increment the input pointer.  */          \
 409     inptr += 4;                                                               \
 410   }
 411 #define LOOP_NEED_FLAGS
 412 #define EXTRA_LOOP_DECLS        , int *setp
 413 #define INIT_PARAMS             int set = *setp & CURRENT_SEL_MASK; \
 414                                 int ann = *setp & CURRENT_ANN_MASK
 415 #define UPDATE_PARAMS           *setp = set | ann
 416 #include <iconv/loop.c>
 417
 418
 419 /* Now define the toplevel functions.  */
 420 #include <iconv/skeleton.c>