iconvdata/iso-2022-cn-ext.c

   1 /* Conversion module for ISO-2022-CN-EXT.
   2    Copyright (C) 2000 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@cygnus.com>, 2000.
   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 "iso-ir-165.h"
  27 #include "cns11643.h"
  28 #include "cns11643l1.h"
  29 #include "cns11643l2.h"
  30
  31 #include <assert.h>
  32
  33 /* This makes obvious what everybody knows: 0x1b is the Esc character.  */
  34 #define ESC     0x1b
  35
  36 /* We have single-byte shift-in and shift-out sequences, and the single
  37    shift sequences SS2 and SS3 which replaces the SS2/SS3 designation for
  38    the next two bytes.  */
  39 #define SI      0x0f
  40 #define SO      0x0e
  41 #define SS2_0   ESC
  42 #define SS2_1   0x4e
  43 #define SS3_0   ESC
  44 #define SS3_1   0x4f
  45
  46 /* Definitions used in the body of the `gconv' function.  */
  47 #define CHARSET_NAME            "ISO-2022-CN-EXT//"
  48 #define DEFINE_INIT             1
  49 #define DEFINE_FINI             1
  50 #define FROM_LOOP               from_iso2022cn_ext_loop
  51 #define TO_LOOP                 to_iso2022cn_ext_loop
  52 #define MIN_NEEDED_FROM         1
  53 #define MAX_NEEDED_FROM         4
  54 #define MIN_NEEDED_TO           4
  55 #define MAX_NEEDED_TO           4
  56 #define PREPARE_LOOP \
  57   int save_set;                                                               \
  58   int *setp = &data->__statep->__count;
  59 #define EXTRA_LOOP_ARGS         , setp
  60
  61
  62 /* The charsets GB/T 12345-90, GB 7589-87, GB/T 13131-9X, GB 7590-87,
  63    and GB/T 13132-9X are not registered to the best of my knowledge and
  64    therefore have no escape sequence assigned.  We cannot handle them
  65    for this reason.  Tell the implementation about this.  */
  66 #define X12345  '\0'
  67 #define X7589   '\0'
  68 #define X13131  '\0'
  69 #define X7590   '\0'
  70 #define X13132  '\0'
  71
  72
  73 /* The COUNT element of the state keeps track of the currently selected
  74    character set.  The possible values are:  */
  75 enum
  76 {
  77   ASCII_set = 0,
  78   GB2312_set,
  79   GB12345_set,
  80   CNS11643_1_set,
  81   ISO_IR_165_set,
  82   SO_mask = 7,
  83
  84   GB7589_set = 1 << 3,
  85   GB13131_set = 2 << 3,
  86   CNS11643_2_set = 3 << 3,
  87   SS2_mask = 3 << 3,
  88
  89   GB7590_set = 1 << 5,
  90   GB13132_set = 2 << 5,
  91   CNS11643_3_set = 3 << 5,
  92   CNS11643_4_set = 4 << 5,
  93   CNS11643_5_set = 5 << 5,
  94   CNS11643_6_set = 6 << 5,
  95   CNS11643_7_set = 7 << 5,
  96   SS3_mask = 7 << 5,
  97
  98 #define CURRENT_MASK (SO_mask | SS2_mask | SS3_mask)
  99
 100   GB2312_ann = 1 << 8,
 101   GB12345_ann = 2 << 8,
 102   CNS11643_1_ann = 3 << 8,
 103   ISO_IR_165_ann = 4 << 8,
 104   SO_ann = 7 << 8,
 105
 106   GB7589_ann = 1 << 11,
 107   GB13131_ann = 2 << 11,
 108   CNS11643_2_ann = 3 << 11,
 109   SS2_ann = 3 << 11,
 110
 111   GB7590_ann = 1 << 13,
 112   GB13132_ann = 2 << 13,
 113   CNS11643_3_ann = 3 << 13,
 114   CNS11643_4_ann = 4 << 13,
 115   CNS11643_5_ann = 5 << 13,
 116   CNS11643_6_ann = 6 << 13,
 117   CNS11643_7_ann = 7 << 13,
 118   SS3_ann = 7 << 13
 119 };
 120
 121
 122 /* Since this is a stateful encoding we have to provide code which resets
 123    the output state to the initial state.  This has to be done during the
 124    flushing.  */
 125 #define EMIT_SHIFT_TO_INIT \
 126   if (data->__statep->__count >> 3 != ASCII_set)                              \
 127     {                                                                         \
 128       if (FROM_DIRECTION)                                                     \
 129         /* It's easy, we don't have to emit anything, we just reset the       \
 130            state for the input.  */                                           \
 131         data->__statep->__count = ASCII_set << 3;                             \
 132       else                                                                    \
 133         {                                                                     \
 134           unsigned char *outbuf = data->__outbuf;                             \
 135                                                                               \
 136           /* We are not in the initial state.  To switch back we have         \
 137              to emit `SI'.  */                                                \
 138           if (outbuf == data->__outbufend)                                    \
 139             /* We don't have enough room in the output buffer.  */            \
 140             status = __GCONV_FULL_OUTPUT;                                     \
 141           else                                                                \
 142             {                                                                 \
 143               /* Write out the shift sequence.  */                            \
 144               *outbuf++ = SI;                                                 \
 145               if (data->__flags & __GCONV_IS_LAST)                            \
 146                 *irreversible += 1;                                           \
 147               data->__outbuf = outbuf;                                        \
 148               data->__statep->__count = ASCII_set << 3;                       \
 149             }                                                                 \
 150         }                                                                     \
 151     }
 152
 153
 154 /* Since we might have to reset input pointer we must be able to save
 155    and retore the state.  */
 156 #define SAVE_RESET_STATE(Save) \
 157   if (Save)                                                                   \
 158     save_set = *setp;                                                         \
 159   else                                                                        \
 160     *setp = save_set
 161
 162
 163 /* First define the conversion function from ISO-2022-CN to UCS4.  */
 164 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 165 #define MAX_NEEDED_INPUT        MAX_NEEDED_FROM
 166 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 167 #define LOOPFCT                 FROM_LOOP
 168 #define BODY \
 169   {                                                                           \
 170     uint32_t ch = *inptr;                                                     \
 171                                                                               \
 172     /* This is a 7bit character set, disallow all 8bit characters.  */        \
 173     if (ch > 0x7f)                                                            \
 174       {                                                                       \
 175         if (! ignore_errors_p ())                                             \
 176           {                                                                   \
 177             result = __GCONV_ILLEGAL_INPUT;                                   \
 178             break;                                                            \
 179           }                                                                   \
 180         ++inptr;                                                              \
 181         ++*irreversible;                                                      \
 182         continue;                                                             \
 183       }                                                                       \
 184                                                                               \
 185     /* Recognize escape sequences.  */                                        \
 186     if (ch == ESC)                                                            \
 187       {                                                                       \
 188         /* There are three kinds of escape sequences we have to handle:       \
 189            - those announcing the use of GB and CNS characters on the         \
 190              line; we can simply ignore them                                  \
 191            - the initial byte of the SS2 sequence.                            \
 192            - the initial byte of the SS3 sequence.                            \
 193         */                                                                    \
 194         if (inptr + 2 > inend                                                 \
 195             || (inptr[1] == '$'                                               \
 196                 && (inptr + 3 > inend                                         \
 197                     || (inptr[2] == ')' && inptr + 4 > inend)                 \
 198                     || (inptr[2] == '*' && inptr + 4 > inend)                 \
 199                     || (inptr[2] == '+' && inptr + 4 > inend)))               \
 200             || (inptr[1] == SS2_1 && inptr + 4 > inend)                       \
 201             || (inptr[1] == SS3_1 && inptr + 4 > inend))                      \
 202           {                                                                   \
 203             result = __GCONV_INCOMPLETE_INPUT;                                \
 204             break;                                                            \
 205           }                                                                   \
 206         if (inptr[1] == '$'                                                   \
 207             && ((inptr[2] == ')'                                              \
 208                  && (inptr[3] == 'A'                                          \
 209                      || (X12345 != '\0' && inptr[3] == X12345)                \
 210                      || inptr[3] == 'E' || inptr[3] == 'G'))                  \
 211                 || (inptr[2] == '*'                                           \
 212                     && ((X7589 != '\0' && inptr[3] == X7589)                  \
 213                         || (X13131 != '\0' && inptr[3] == X13131)             \
 214                         || inptr[3] == 'H'))                                  \
 215                 || (inptr[2] == '+'                                           \
 216                     && ((X7590 != '\0' && inptr[3] == X7590)                  \
 217                         || (X13132 != '\0' && inptr[3] == X13132)             \
 218                         || inptr[3] == 'I' || inptr[3] == 'J'                 \
 219                         || inptr[3] == 'K' || inptr[3] == 'L'                 \
 220                         || inptr[3] == 'M'))))                                \
 221           {                                                                   \
 222             /* OK, we accept those character sets.  */                        \
 223             if (inptr[3] == 'A')                                              \
 224               ann = (ann & ~SO_ann) | GB2312_ann;                             \
 225             else if (inptr[3] == 'G')                                         \
 226               ann = (ann & ~SO_ann) | CNS11643_1_ann;                         \
 227             else if (inptr[3] == 'E')                                         \
 228               ann = (ann & ~SO_ann) | ISO_IR_165_ann;                         \
 229             else if (X12345 != '\0' && inptr[3] == X12345)                    \
 230               ann = (ann & ~SO_ann) | GB12345_ann;                            \
 231             else if (inptr[3] == 'H')                                         \
 232               ann = (ann & ~SS2_ann) | CNS11643_2_ann;                        \
 233             else if (X7589 != '\0' && inptr[3] == X7589)                      \
 234               ann = (ann & ~SS2_ann) | GB7589_ann;                            \
 235             else if (X13131 != '\0' && inptr[3] == X13131)                    \
 236               ann = (ann & ~SS2_ann) | GB13131_ann;                           \
 237             else if (inptr[3] == 'I')                                         \
 238               ann = (ann & ~SS3_ann) | CNS11643_3_ann;                        \
 239             else if (inptr[3] == 'J')                                         \
 240               ann = (ann & ~SS3_ann) | CNS11643_4_ann;                        \
 241             else if (inptr[3] == 'K')                                         \
 242               ann = (ann & ~SS3_ann) | CNS11643_5_ann;                        \
 243             else if (inptr[3] == 'L')                                         \
 244               ann = (ann & ~SS3_ann) | CNS11643_6_ann;                        \
 245             else if (inptr[3] == 'M')                                         \
 246               ann = (ann & ~SS3_ann) | CNS11643_7_ann;                        \
 247             else if (X7590 != '\0' && inptr[3] == X7590)                      \
 248               ann = (ann & ~SS3_ann) | GB7590_ann;                            \
 249             else if (X13132 != '\0' && inptr[3] == X13132)                    \
 250               ann = (ann & ~SS3_ann) | GB13132_ann;                           \
 251             inptr += 4;                                                       \
 252             continue;                                                         \
 253           }                                                                   \
 254       }                                                                       \
 255     else if (ch == SO)                                                        \
 256       {                                                                       \
 257         /* Switch to use GB2312, GB12345, CNS 11643 plane 1, or ISO-IR-165,   \
 258            depending on which S0 designation came last.  The only problem     \
 259            is what to do with faulty input files where no designator came.    \
 260            XXX For now I'll default to use GB2312.  If this is not the        \
 261            best behavior (e.g., we should flag an error) let me know.  */     \
 262         ++inptr;                                                              \
 263         switch (ann & SO_ann)                                                 \
 264           {                                                                   \
 265           case GB2312_ann:                                                    \
 266             set = GB2312_set;                                                 \
 267             break;                                                            \
 268           case GB12345_ann:                                                   \
 269             set = GB12345_set;                                                \
 270             break;                                                            \
 271           case CNS11643_1_ann:                                                \
 272             set = CNS11643_1_set;                                             \
 273             break;                                                            \
 274           default:                                                            \
 275             assert ((ann & SO_ann) == ISO_IR_165_ann);                        \
 276             set = ISO_IR_165_set;                                             \
 277             break;                                                            \
 278           }                                                                   \
 279         continue;                                                             \
 280       }                                                                       \
 281     else if (ch == SI)                                                        \
 282       {                                                                       \
 283         /* Switch to use ASCII.  */                                           \
 284         ++inptr;                                                              \
 285         set = ASCII_set;                                                      \
 286         continue;                                                             \
 287       }                                                                       \
 288                                                                               \
 289     if (ch == ESC && inptr[1] == SS2_1)                                       \
 290       {                                                                       \
 291         /* This is a character from CNS 11643 plane 2.                        \
 292            XXX We could test here whether the use of this character           \
 293            set was announced.                                                 \
 294            XXX Current GB7589 and GB13131 are not supported.  */              \
 295         inptr += 2;                                                           \
 296         ch = cns11643l2_to_ucs4 (&inptr, 2, 0);                               \
 297         if (ch == __UNKNOWN_10646_CHAR)                                       \
 298           {                                                                   \
 299             if (! ignore_errors_p ())                                         \
 300               {                                                               \
 301                 inptr -= 2;                                                   \
 302                 result = __GCONV_ILLEGAL_INPUT;                               \
 303                 break;                                                        \
 304               }                                                               \
 305             inptr += 2;                                                       \
 306             ++*irreversible;                                                  \
 307             continue;                                                         \
 308           }                                                                   \
 309       }                                                                       \
 310     /* Note that we can assume here that at least 4 bytes are available if    \
 311        the first byte is ESC since otherwise the first if would have been     \
 312        true.  */                                                              \
 313     else if (ch == ESC && inptr[1] == SS3_1)                                  \
 314       {                                                                       \
 315         /* This is a character from CNS 11643 plane 3 or higher.              \
 316            XXX Currently GB7590 and GB13132 are not supported.  */            \
 317         char buf[3];                                                          \
 318         const char *tmp = buf;                                                \
 319                                                                               \
 320         buf[1] = inptr[2];                                                    \
 321         buf[2] = inptr[3];                                                    \
 322         switch (ann & SS3_ann)                                                \
 323           {                                                                   \
 324           case CNS11643_3_ann:                                                \
 325             /* CNS 11643 plane 3 is part of the old CNS 11643 plane 14.  */   \
 326             if (buf[1] < 0x62 || (buf[1] == 0x62 && buf[2] <= 0x45))          \
 327               {                                                               \
 328                 buf[0] = 0x2e;                                                \
 329                 ch = cns11643_to_ucs4 (&tmp, 3, 0);                           \
 330               }                                                               \
 331             else                                                              \
 332               ch = __UNKNOWN_10646_CHAR;                                      \
 333             break;                                                            \
 334           default:                                                            \
 335             /* XXX Currently planes 4 to 7 are not supported.  */             \
 336             ch = __UNKNOWN_10646_CHAR;                                        \
 337             break;                                                            \
 338           }                                                                   \
 339         if (ch == __UNKNOWN_10646_CHAR)                                       \
 340           {                                                                   \
 341             if (! ignore_errors_p ())                                         \
 342               {                                                               \
 343                 result = __GCONV_ILLEGAL_INPUT;                               \
 344                 break;                                                        \
 345               }                                                               \
 346             inptr += 4;                                                       \
 347             ++*irreversible;                                                  \
 348             continue;                                                         \
 349           }                                                                   \
 350         assert (tmp == buf + 3);                                              \
 351         inptr += 4;                                                           \
 352       }                                                                       \
 353     else if (set == ASCII_set)                                                \
 354       {                                                                       \
 355         /* Almost done, just advance the input pointer.  */                   \
 356         ++inptr;                                                              \
 357       }                                                                       \
 358     else                                                                      \
 359       {                                                                       \
 360         /* That's pretty easy, we have a dedicated functions for this.  */    \
 361         if (inend - inptr < 2)                                                \
 362           {                                                                   \
 363             result = __GCONV_INCOMPLETE_INPUT;                                \
 364             break;                                                            \
 365           }                                                                   \
 366         if (set == GB2312_set)                                                \
 367           ch = gb2312_to_ucs4 (&inptr, inend - inptr, 0);                     \
 368         else if (set == ISO_IR_165_set)                                       \
 369           ch = isoir165_to_ucs4 (&inptr, inend - inptr);                      \
 370         else                                                                  \
 371           {                                                                   \
 372             assert (set == CNS11643_1_set);                                   \
 373             ch = cns11643l1_to_ucs4 (&inptr, inend - inptr, 0);               \
 374           }                                                                   \
 375                                                                               \
 376         if (ch == 0)                                                          \
 377           {                                                                   \
 378             result = __GCONV_INCOMPLETE_INPUT;                                \
 379             break;                                                            \
 380           }                                                                   \
 381         else if (ch == __UNKNOWN_10646_CHAR)                                  \
 382           {                                                                   \
 383             if (! ignore_errors_p ())                                         \
 384               {                                                               \
 385                 result = __GCONV_ILLEGAL_INPUT;                               \
 386                 break;                                                        \
 387               }                                                               \
 388             inptr += 2;                                                       \
 389             ++*irreversible;                                                  \
 390             continue;                                                         \
 391           }                                                                   \
 392       }                                                                       \
 393                                                                               \
 394     *((uint32_t *) outptr)++ = ch;                                            \
 395   }
 396 #define EXTRA_LOOP_DECLS        , int *setp
 397 #define INIT_PARAMS             int set = (*setp >> 3) & CURRENT_MASK; \
 398                                 int ann = (*setp >> 3) & ~CURRENT_MASK
 399 #define UPDATE_PARAMS           *setp = (set | ann) << 3
 400 #define LOOP_NEED_FLAGS
 401 #include <iconv/loop.c>
 402
 403
 404 /* Next, define the other direction.  */
 405 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 406 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 407 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 408 #define LOOPFCT                 TO_LOOP
 409 #define BODY \
 410   {                                                                           \
 411     uint32_t ch;                                                              \
 412     size_t written = 0;                                                       \
 413                                                                               \
 414     ch = *((const uint32_t *) inptr);                                         \
 415                                                                               \
 416     /* First see whether we can write the character using the currently       \
 417        selected character set.  */                                            \
 418     if (ch < 0x80)                                                            \
 419       {                                                                       \
 420         if (set != ASCII_set)                                                 \
 421           {                                                                   \
 422             *outptr++ = SI;                                                   \
 423             set = ASCII_set;                                                  \
 424             if (outptr == outend)                                             \
 425               {                                                               \
 426                 result = __GCONV_FULL_OUTPUT;                                 \
 427                 break;                                                        \
 428               }                                                               \
 429           }                                                                   \
 430                                                                               \
 431         *outptr++ = ch;                                                       \
 432         written = 1;                                                          \
 433                                                                               \
 434         /* At the end of the line we have to clear the `ann' flags since      \
 435            every line must contain this information again.  */                \
 436         if (ch == L'\n')                                                      \
 437           ann = 0;                                                            \
 438       }                                                                       \
 439     else                                                                      \
 440       {                                                                       \
 441         char buf[2];                                                          \
 442         int used;                                                             \
 443                                                                               \
 444         if (set == GB2312_set || ((ann & SO_ann) != CNS11643_1_ann            \
 445                                   && (ann & SO_ann) != ISO_IR_165_ann))       \
 446           {                                                                   \
 447             written = ucs4_to_gb2312 (ch, buf, 2);                            \
 448             used = GB2312_set;                                                \
 449           }                                                                   \
 450         else if (set == ISO_IR_165_set || (ann & SO_ann) == ISO_IR_165_set)   \
 451           {                                                                   \
 452             written = ucs4_to_isoir165 (ch, buf, 2);                          \
 453             used = ISO_IR_165_set;                                            \
 454           }                                                                   \
 455         else                                                                  \
 456           {                                                                   \
 457             written = ucs4_to_cns11643l1 (ch, buf, 2);                        \
 458             used = CNS11643_1_set;                                            \
 459           }                                                                   \
 460                                                                               \
 461         if (written == __UNKNOWN_10646_CHAR)                                  \
 462           {                                                                   \
 463             /* Cannot convert it using the currently selected SO set.         \
 464                Next try the SS2 set.  */                                      \
 465             written = ucs4_to_cns11643l2 (ch, buf, 2);                        \
 466             if (written != __UNKNOWN_10646_CHAR)                              \
 467               /* Yep, that worked.  */                                        \
 468               used = CNS11643_2_set;                                          \
 469             else                                                              \
 470               {                                                               \
 471                 char tmpbuf[3];                                               \
 472                                                                               \
 473                 switch (0)                                                    \
 474                   {                                                           \
 475                   default:                                                    \
 476                     /* Well, see whether we have to change the SO set.  */    \
 477                                                                               \
 478                     if (used != GB2312_set)                                   \
 479                       {                                                       \
 480                         written = ucs4_to_gb2312 (ch, buf, 2);                \
 481                         if (written != __UNKNOWN_10646_CHAR)                  \
 482                           {                                                   \
 483                             used = GB2312_set;                                \
 484                             break;                                            \
 485                           }                                                   \
 486                       }                                                       \
 487                                                                               \
 488                     if (used != ISO_IR_165_set)                               \
 489                       {                                                       \
 490                         written = ucs4_to_isoir165 (ch, buf, 2);              \
 491                         if (written != __UNKNOWN_10646_CHAR)                  \
 492                           {                                                   \
 493                             used = ISO_IR_165_set;                            \
 494                             break;                                            \
 495                           }                                                   \
 496                       }                                                       \
 497                                                                               \
 498                     if (used != CNS11643_1_set)                               \
 499                       {                                                       \
 500                         written = ucs4_to_cns11643l1 (ch, buf, 2);            \
 501                         if (written != __UNKNOWN_10646_CHAR)                  \
 502                           {                                                   \
 503                             used = CNS11643_1_set;                            \
 504                             break;                                            \
 505                           }                                                   \
 506                       }                                                       \
 507                                                                               \
 508                     written = ucs4_to_cns11643 (ch, tmpbuf, 3);               \
 509                     if (written == 3 && tmpbuf[0] != 1 && tmpbuf[0] != 2)     \
 510                       {                                                       \
 511                         buf[0] = tmpbuf[1];                                   \
 512                         buf[1] = tmpbuf[2];                                   \
 513                         written = 2;                                          \
 514                         /* CNS 11643 plane 3 is part of the old CNS 11643     \
 515                            plane 14.                                          \
 516                            XXX Currently planes 4 to 7 are not supported.  */ \
 517                         if (tmpbuf[0] == 14                                   \
 518                             && (tmpbuf[1] < 0x62                              \
 519                                 || (tmpbuf[1] == 0x62 && tmpbuf[2] <= 0x45))) \
 520                           {                                                   \
 521                             used = CNS11643_3_set;                            \
 522                             break;                                            \
 523                           }                                                   \
 524                       }                                                       \
 525                                                                               \
 526                     /* Even this does not work.  Error.  */                   \
 527                     used = ASCII_set;                                         \
 528                   }                                                           \
 529                 if (used == ASCII_set)                                        \
 530                   {                                                           \
 531                     STANDARD_ERR_HANDLER (4);                                 \
 532                   }                                                           \
 533               }                                                               \
 534           }                                                                   \
 535         assert (written == 2);                                                \
 536                                                                               \
 537         /* See whether we have to emit an escape sequence.  */                \
 538         if (set != used)                                                      \
 539           {                                                                   \
 540             /* First see whether we announced that we use this                \
 541                character set.  */                                             \
 542             if ((used & SO_mask) != 0 && (ann & SO_ann) != (used << 8))       \
 543               {                                                               \
 544                 const char *escseq;                                           \
 545                                                                               \
 546                 if (outptr + 4 > outend)                                      \
 547                   {                                                           \
 548                     result = __GCONV_FULL_OUTPUT;                             \
 549                     break;                                                    \
 550                   }                                                           \
 551                                                                               \
 552                 assert (used >= 1 && used <= 4);                              \
 553                 escseq = ")A\0\0)G)E" + (used - 1) * 2;                       \
 554                 *outptr++ = ESC;                                              \
 555                 *outptr++ = '$';                                              \
 556                 *outptr++ = *escseq++;                                        \
 557                 *outptr++ = *escseq++;                                        \
 558                                                                               \
 559                 ann = (ann & ~SO_ann) | (used << 8);                          \
 560               }                                                               \
 561             else if ((used & SS2_mask) != 0 && (ann & SS2_ann) != (used << 8))\
 562               {                                                               \
 563                 const char *escseq;                                           \
 564                                                                               \
 565                 assert (used == CNS11643_2_set); /* XXX */                    \
 566                 escseq = "*H";                                                \
 567                 *outptr++ = ESC;                                              \
 568                 *outptr++ = '$';                                              \
 569                 *outptr++ = *escseq++;                                        \
 570                 *outptr++ = *escseq++;                                        \
 571                                                                               \
 572                 ann = (ann & ~SS2_ann) | (used << 8);                         \
 573               }                                                               \
 574             else if ((used & SS3_mask) != 0 && (ann & SS3_ann) != (used << 8))\
 575               {                                                               \
 576                 const char *escseq;                                           \
 577                                                                               \
 578                 assert ((used >> 5) >= 3 && (used >> 5) <= 7);                \
 579                 escseq = "+I+J+K+L+M" + ((used >> 5) - 3) * 2;                \
 580                 *outptr++ = ESC;                                              \
 581                 *outptr++ = '$';                                              \
 582                 *outptr++ = *escseq++;                                        \
 583                 *outptr++ = *escseq++;                                        \
 584                                                                               \
 585                 ann = (ann & ~SS3_ann) | (used << 8);                         \
 586               }                                                               \
 587                                                                               \
 588             if (used == CNS11643_2_set)                                       \
 589               {                                                               \
 590                 if (outptr + 2 > outend)                                      \
 591                   {                                                           \
 592                     result = __GCONV_FULL_OUTPUT;                             \
 593                     break;                                                    \
 594                   }                                                           \
 595                 *outptr++ = SS2_0;                                            \
 596                 *outptr++ = SS2_1;                                            \
 597               }                                                               \
 598             else if (used >= CNS11643_3_set && used <= CNS11643_7_set)        \
 599               {                                                               \
 600                 if (outptr + 2 > outend)                                      \
 601                   {                                                           \
 602                     result = __GCONV_FULL_OUTPUT;                             \
 603                     break;                                                    \
 604                   }                                                           \
 605                 *outptr++ = SS3_0;                                            \
 606                 *outptr++ = SS3_1;                                            \
 607               }                                                               \
 608             else                                                              \
 609               {                                                               \
 610                 /* We only have to emit something if currently ASCII is       \
 611                    selected.  Otherwise we are switching within the           \
 612                    SO charset.  */                                            \
 613                 if (set == ASCII_set)                                         \
 614                   {                                                           \
 615                     if (outptr + 1 > outend)                                  \
 616                       {                                                       \
 617                         result = __GCONV_FULL_OUTPUT;                         \
 618                         break;                                                \
 619                       }                                                       \
 620                     *outptr++ = SO;                                           \
 621                   }                                                           \
 622               }                                                               \
 623                                                                               \
 624             /* Always test the length here since we have used up all the      \
 625                guaranteed output buffer slots.  */                            \
 626             if (outptr + 2 > outend)                                          \
 627               {                                                               \
 628                 result = __GCONV_FULL_OUTPUT;                                 \
 629                 break;                                                        \
 630               }                                                               \
 631           }                                                                   \
 632         else if (outptr + 2 > outend)                                         \
 633           {                                                                   \
 634             result = __GCONV_FULL_OUTPUT;                                     \
 635             break;                                                            \
 636           }                                                                   \
 637                                                                               \
 638         *outptr++ = buf[0];                                                   \
 639         *outptr++ = buf[1];                                                   \
 640         set = used;                                                           \
 641       }                                                                       \
 642                                                                               \
 643     /* Now that we wrote the output increment the input pointer.  */          \
 644     inptr += 4;                                                               \
 645   }
 646 #define EXTRA_LOOP_DECLS        , int *setp
 647 #define INIT_PARAMS             int set = (*setp >> 3) & CURRENT_MASK; \
 648                                 int ann = (*setp >> 3) & ~CURRENT_MASK
 649 #define UPDATE_PARAMS           *setp = (set | ann) << 3
 650 #define LOOP_NEED_FLAGS
 651 #include <iconv/loop.c>
 652
 653
 654 /* Now define the toplevel functions.  */
 655 #include <iconv/skeleton.c>