iconvdata/utf-16.c

   1 /* Conversion module for UTF-16.
   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 <byteswap.h>
  22 #include <dlfcn.h>
  23 #include <gconv.h>
  24 #include <stddef.h>
  25 #include <stdint.h>
  26 #include <stdlib.h>
  27 #include <string.h>
  28
  29 /* This is the Byte Order Mark character (BOM).  */
  30 #define BOM     0xfeff
  31 /* And in the other byte order.  */
  32 #define BOM_OE  0xfffe
  33
  34
  35 /* Definitions used in the body of the `gconv' function.  */
  36 #define FROM_LOOP               from_utf16_loop
  37 #define TO_LOOP                 to_utf16_loop
  38 #define DEFINE_INIT             0
  39 #define DEFINE_FINI             0
  40 #define MIN_NEEDED_FROM         2
  41 #define MAX_NEEDED_FROM         4
  42 #define MIN_NEEDED_TO           4
  43 #define FROM_DIRECTION          (dir == from_utf16)
  44 #define PREPARE_LOOP \
  45   enum direction dir = ((struct utf16_data *) step->__data)->dir;             \
  46   enum variant var = ((struct utf16_data *) step->__data)->var;               \
  47   int swap = ((struct utf16_data *) step->__data)->swap;                      \
  48   if (FROM_DIRECTION || var == UTF_16)                                        \
  49     {                                                                         \
  50       if (data->__invocation_counter == 0)                                    \
  51         {                                                                     \
  52           /* We have to find out which byte order the file is encoded in.  */ \
  53           if (inptr + 2 > inend)                                              \
  54             return __GCONV_EMPTY_INPUT;                                       \
  55                                                                               \
  56           if (get16u (inptr) == BOM)                                          \
  57             /* Simply ignore the BOM character.  */                           \
  58             inptr += 2;                                                       \
  59           else if (get16u (inptr) == BOM_OE)                                  \
  60             {                                                                 \
  61               ((struct utf16_data *) step->__data)->swap = 1;                 \
  62               inptr += 2;                                                     \
  63             }                                                                 \
  64         }                                                                     \
  65     }                                                                         \
  66   else if (!FROM_DIRECTION && var == UTF_16 && !data->__internal_use          \
  67            && data->__invocation_counter == 0)                                \
  68     {                                                                         \
  69       /* Emit the Byte Order Mark.  */                                        \
  70       if (__builtin_expect (outbuf + 2 > outend, 0))                          \
  71         return __GCONV_FULL_OUTPUT;                                           \
  72                                                                               \
  73       put16u (outbuf, BOM);                                                   \
  74       outbuf += 2;                                                            \
  75     }
  76 #define EXTRA_LOOP_ARGS         , var, swap
  77
  78
  79 /* Direction of the transformation.  */
  80 enum direction
  81 {
  82   illegal_dir,
  83   to_utf16,
  84   from_utf16
  85 };
  86
  87 enum variant
  88 {
  89   illegal_var,
  90   UTF_16,
  91   UTF_16LE,
  92   UTF_16BE
  93 };
  94
  95 struct utf16_data
  96 {
  97   enum direction dir;
  98   enum variant var;
  99   int swap;
 100 };
 101
 102
 103 int
 104 gconv_init (struct __gconv_step *step)
 105 {
 106   /* Determine which direction.  */
 107   struct utf16_data *new_data;
 108   enum direction dir = illegal_dir;
 109   enum variant var = illegal_var;
 110   int result;
 111
 112   if (__strcasecmp (step->__from_name, "UTF-16") == 0)
 113     {
 114       dir = from_utf16;
 115       var = UTF_16;
 116     }
 117   else if (__strcasecmp (step->__to_name, "UTF-16") == 0)
 118     {
 119       dir = to_utf16;
 120       var = UTF_16;
 121     }
 122   else if (__strcasecmp (step->__from_name, "UTF-16BE") == 0)
 123     {
 124       dir = from_utf16;
 125       var = UTF_16BE;
 126     }
 127   else if (__strcasecmp (step->__to_name, "UTF-16BE") == 0)
 128     {
 129       dir = to_utf16;
 130       var = UTF_16BE;
 131     }
 132   else if (__strcasecmp (step->__from_name, "UTF-16LE") == 0)
 133     {
 134       dir = from_utf16;
 135       var = UTF_16LE;
 136     }
 137   else if (__strcasecmp (step->__to_name, "UTF-16LE") == 0)
 138     {
 139       dir = to_utf16;
 140       var = UTF_16LE;
 141     }
 142
 143   result = __GCONV_NOCONV;
 144   if (__builtin_expect (dir, to_utf16) != illegal_dir)
 145     {
 146       new_data = (struct utf16_data *) malloc (sizeof (struct utf16_data));
 147
 148       result = __GCONV_NOMEM;
 149       if (new_data != NULL)
 150         {
 151           new_data->dir = dir;
 152           new_data->var = var;
 153           new_data->swap = ((var == UTF_16LE && BYTE_ORDER == BIG_ENDIAN)
 154                             || (var == UTF_16BE
 155                                 && BYTE_ORDER == LITTLE_ENDIAN));
 156           step->__data = new_data;
 157
 158           if (dir == from_utf16)
 159             {
 160               step->__min_needed_from = MIN_NEEDED_FROM;
 161               step->__max_needed_from = MIN_NEEDED_FROM;
 162               step->__min_needed_to = MIN_NEEDED_TO;
 163               step->__max_needed_to = MIN_NEEDED_TO;
 164             }
 165           else
 166             {
 167               step->__min_needed_from = MIN_NEEDED_TO;
 168               step->__max_needed_from = MIN_NEEDED_TO;
 169               step->__min_needed_to = MIN_NEEDED_FROM;
 170               step->__max_needed_to = MIN_NEEDED_FROM;
 171             }
 172
 173           step->__stateful = 0;
 174
 175           result = __GCONV_OK;
 176         }
 177     }
 178
 179   return result;
 180 }
 181
 182
 183 void
 184 gconv_end (struct __gconv_step *data)
 185 {
 186   free (data->__data);
 187 }
 188
 189
 190 /* Convert from the internal (UCS4-like) format to UTF-16.  */
 191 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 192 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 193 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 194 #define LOOPFCT                 TO_LOOP
 195 #define BODY \
 196   {                                                                           \
 197     uint32_t c = get32 (inptr);                                               \
 198                                                                               \
 199     if (swap)                                                                 \
 200       {                                                                       \
 201         if (__builtin_expect (c, 0) >= 0x10000)                               \
 202           {                                                                   \
 203             if (__builtin_expect (c, 0) >= 0x110000)                          \
 204               {                                                               \
 205                 STANDARD_ERR_HANDLER (4);                                     \
 206               }                                                               \
 207                                                                               \
 208             /* Generate a surrogate character.  */                            \
 209             if (__builtin_expect (outptr + 4 > outend, 0))                    \
 210               {                                                               \
 211                 /* Overflow in the output buffer.  */                         \
 212                 result = __GCONV_FULL_OUTPUT;                                 \
 213                 break;                                                        \
 214               }                                                               \
 215                                                                               \
 216             put16 (outptr, bswap_16 (0xd7c0 + (c >> 10)));                    \
 217             outptr += 2;                                                      \
 218             put16 (outptr, bswap_16 (0xdc00 + (c & 0x3ff)));                  \
 219           }                                                                   \
 220         else                                                                  \
 221           put16 (outptr, bswap_16 (c));                                       \
 222       }                                                                       \
 223     else                                                                      \
 224       {                                                                       \
 225         if (__builtin_expect (c, 0) >= 0x10000)                               \
 226           {                                                                   \
 227             if (__builtin_expect (c, 0) >= 0x110000)                          \
 228               {                                                               \
 229                 STANDARD_ERR_HANDLER (4);                                     \
 230               }                                                               \
 231                                                                               \
 232             /* Generate a surrogate character.  */                            \
 233             if (__builtin_expect (outptr + 4 > outend, 0))                    \
 234               {                                                               \
 235                 /* Overflow in the output buffer.  */                         \
 236                 result = __GCONV_FULL_OUTPUT;                                 \
 237                 break;                                                        \
 238               }                                                               \
 239                                                                               \
 240             put16 (outptr, 0xd7c0 + (c >> 10));                               \
 241             outptr += 2;                                                      \
 242             put16 (outptr, 0xdc00 + (c & 0x3ff));                             \
 243           }                                                                   \
 244         else                                                                  \
 245           put16 (outptr, c);                                                  \
 246       }                                                                       \
 247     outptr += 2;                                                              \
 248     inptr += 4;                                                               \
 249   }
 250 #define LOOP_NEED_FLAGS
 251 #define EXTRA_LOOP_DECLS \
 252         , enum variant var, int swap
 253 #include <iconv/loop.c>
 254
 255
 256 /* Convert from UTF-16 to the internal (UCS4-like) format.  */
 257 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 258 #define MAX_NEEDED_INPUT        MAX_NEEDED_FROM
 259 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 260 #define LOOPFCT                 FROM_LOOP
 261 #define BODY \
 262   {                                                                           \
 263     uint16_t u1 = get16 (inptr);                                              \
 264                                                                               \
 265     if (swap)                                                                 \
 266       {                                                                       \
 267         u1 = bswap_16 (u1);                                                   \
 268                                                                               \
 269         if (__builtin_expect (u1, 0) < 0xd800 || u1 > 0xdfff)                 \
 270           {                                                                   \
 271             /* No surrogate.  */                                              \
 272             put32 (outptr, u1);                                               \
 273             inptr += 2;                                                       \
 274           }                                                                   \
 275         else                                                                  \
 276           {                                                                   \
 277             uint16_t u2;                                                      \
 278                                                                               \
 279             /* It's a surrogate character.  At least the first word says      \
 280                it is.  */                                                     \
 281             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 282               {                                                               \
 283                 /* We don't have enough input for another complete input      \
 284                    character.  */                                             \
 285                 result = __GCONV_INCOMPLETE_INPUT;                            \
 286                 break;                                                        \
 287               }                                                               \
 288                                                                               \
 289             inptr += 2;                                                       \
 290             u2 = bswap_16 (get16 (inptr));                                    \
 291             if (__builtin_expect (u2, 0xdc00) < 0xdc00                        \
 292                 || __builtin_expect (u2, 0xdc00) >= 0xdfff)                   \
 293               {                                                               \
 294                 /* This is no valid second word for a surrogate.  */          \
 295                 if (! ignore_errors_p ())                                     \
 296                   {                                                           \
 297                     inptr -= 2;                                               \
 298                     result = __GCONV_ILLEGAL_INPUT;                           \
 299                     break;                                                    \
 300                   }                                                           \
 301                                                                               \
 302                 ++*irreversible;                                              \
 303                 continue;                                                     \
 304               }                                                               \
 305                                                                               \
 306             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 307             inptr += 2;                                                       \
 308           }                                                                   \
 309       }                                                                       \
 310     else                                                                      \
 311       {                                                                       \
 312         if (__builtin_expect (u1, 0) < 0xd800 || u1 > 0xdfff)                 \
 313           {                                                                   \
 314             /* No surrogate.  */                                              \
 315             put32 (outptr, u1);                                               \
 316             inptr += 2;                                                       \
 317           }                                                                   \
 318         else                                                                  \
 319           {                                                                   \
 320             uint16_t u2;                                                      \
 321                                                                               \
 322             /* It's a surrogate character.  At least the first word says      \
 323                it is.  */                                                     \
 324             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 325               {                                                               \
 326                 /* We don't have enough input for another complete input      \
 327                    character.  */                                             \
 328                 result = __GCONV_INCOMPLETE_INPUT;                            \
 329                 break;                                                        \
 330               }                                                               \
 331                                                                               \
 332             inptr += 2;                                                       \
 333             u2 = get16 (inptr);                                               \
 334             if (__builtin_expect (u2, 0xdc00) < 0xdc00                        \
 335                 || __builtin_expect (u2, 0xdc00) >= 0xdfff)                   \
 336               {                                                               \
 337                 /* This is no valid second word for a surrogate.  */          \
 338                 if (! ignore_errors_p ())                                     \
 339                   {                                                           \
 340                     inptr -= 2;                                               \
 341                     result = __GCONV_ILLEGAL_INPUT;                           \
 342                     break;                                                    \
 343                   }                                                           \
 344                                                                               \
 345                 ++*irreversible;                                              \
 346                 continue;                                                     \
 347               }                                                               \
 348                                                                               \
 349             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 350             inptr += 2;                                                       \
 351           }                                                                   \
 352       }                                                                       \
 353     outptr += 4;                                                              \
 354   }
 355 #define LOOP_NEED_FLAGS
 356 #define EXTRA_LOOP_DECLS \
 357         , enum variant var, int swap
 358 #include <iconv/loop.c>
 359
 360
 361 /* Now define the toplevel functions.  */
 362 #include <iconv/skeleton.c>