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 extern int gconv_init (struct __gconv_step *step);
 104 int
 105 gconv_init (struct __gconv_step *step)
 106 {
 107   /* Determine which direction.  */
 108   struct utf16_data *new_data;
 109   enum direction dir = illegal_dir;
 110   enum variant var = illegal_var;
 111   int result;
 112
 113   if (__strcasecmp (step->__from_name, "UTF-16//") == 0)
 114     {
 115       dir = from_utf16;
 116       var = UTF_16;
 117     }
 118   else if (__strcasecmp (step->__to_name, "UTF-16//") == 0)
 119     {
 120       dir = to_utf16;
 121       var = UTF_16;
 122     }
 123   else if (__strcasecmp (step->__from_name, "UTF-16BE//") == 0)
 124     {
 125       dir = from_utf16;
 126       var = UTF_16BE;
 127     }
 128   else if (__strcasecmp (step->__to_name, "UTF-16BE//") == 0)
 129     {
 130       dir = to_utf16;
 131       var = UTF_16BE;
 132     }
 133   else if (__strcasecmp (step->__from_name, "UTF-16LE//") == 0)
 134     {
 135       dir = from_utf16;
 136       var = UTF_16LE;
 137     }
 138   else if (__strcasecmp (step->__to_name, "UTF-16LE//") == 0)
 139     {
 140       dir = to_utf16;
 141       var = UTF_16LE;
 142     }
 143
 144   result = __GCONV_NOCONV;
 145   if (__builtin_expect (dir, to_utf16) != illegal_dir)
 146     {
 147       new_data = (struct utf16_data *) malloc (sizeof (struct utf16_data));
 148
 149       result = __GCONV_NOMEM;
 150       if (new_data != NULL)
 151         {
 152           new_data->dir = dir;
 153           new_data->var = var;
 154           new_data->swap = ((var == UTF_16LE && BYTE_ORDER == BIG_ENDIAN)
 155                             || (var == UTF_16BE
 156                                 && BYTE_ORDER == LITTLE_ENDIAN));
 157           step->__data = new_data;
 158
 159           if (dir == from_utf16)
 160             {
 161               step->__min_needed_from = MIN_NEEDED_FROM;
 162               step->__max_needed_from = MIN_NEEDED_FROM;
 163               step->__min_needed_to = MIN_NEEDED_TO;
 164               step->__max_needed_to = MIN_NEEDED_TO;
 165             }
 166           else
 167             {
 168               step->__min_needed_from = MIN_NEEDED_TO;
 169               step->__max_needed_from = MIN_NEEDED_TO;
 170               step->__min_needed_to = MIN_NEEDED_FROM;
 171               step->__max_needed_to = MIN_NEEDED_FROM;
 172             }
 173
 174           step->__stateful = 0;
 175
 176           result = __GCONV_OK;
 177         }
 178     }
 179
 180   return result;
 181 }
 182
 183
 184 extern void gconv_end (struct __gconv_step *data);
 185 void
 186 gconv_end (struct __gconv_step *data)
 187 {
 188   free (data->__data);
 189 }
 190
 191
 192 /* Convert from the internal (UCS4-like) format to UTF-16.  */
 193 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 194 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 195 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 196 #define LOOPFCT                 TO_LOOP
 197 #define BODY \
 198   {                                                                           \
 199     uint32_t c = get32 (inptr);                                               \
 200                                                                               \
 201     if (__builtin_expect (c >= 0xd800 && c < 0xe000, 0))                      \
 202       {                                                                       \
 203         /* Surrogate characters in UCS-4 input are not valid.                 \
 204            We must catch this.  If we let surrogates pass through,            \
 205            attackers could make a security hole exploit by                    \
 206            synthesizing any desired plane 1-16 character.  */                 \
 207         if (! ignore_errors_p ())                                             \
 208           {                                                                   \
 209             result = __GCONV_ILLEGAL_INPUT;                                   \
 210             break;                                                            \
 211           }                                                                   \
 212         inptr += 4;                                                           \
 213         ++*irreversible;                                                      \
 214         continue;                                                             \
 215       }                                                                       \
 216                                                                               \
 217     if (swap)                                                                 \
 218       {                                                                       \
 219         if (__builtin_expect (c, 0) >= 0x10000)                               \
 220           {                                                                   \
 221             if (__builtin_expect (c, 0) >= 0x110000)                          \
 222               {                                                               \
 223                 STANDARD_ERR_HANDLER (4);                                     \
 224               }                                                               \
 225                                                                               \
 226             /* Generate a surrogate character.  */                            \
 227             if (__builtin_expect (outptr + 4 > outend, 0))                    \
 228               {                                                               \
 229                 /* Overflow in the output buffer.  */                         \
 230                 result = __GCONV_FULL_OUTPUT;                                 \
 231                 break;                                                        \
 232               }                                                               \
 233                                                                               \
 234             put16 (outptr, bswap_16 (0xd7c0 + (c >> 10)));                    \
 235             outptr += 2;                                                      \
 236             put16 (outptr, bswap_16 (0xdc00 + (c & 0x3ff)));                  \
 237           }                                                                   \
 238         else                                                                  \
 239           put16 (outptr, bswap_16 (c));                                       \
 240       }                                                                       \
 241     else                                                                      \
 242       {                                                                       \
 243         if (__builtin_expect (c, 0) >= 0x10000)                               \
 244           {                                                                   \
 245             if (__builtin_expect (c, 0) >= 0x110000)                          \
 246               {                                                               \
 247                 STANDARD_ERR_HANDLER (4);                                     \
 248               }                                                               \
 249                                                                               \
 250             /* Generate a surrogate character.  */                            \
 251             if (__builtin_expect (outptr + 4 > outend, 0))                    \
 252               {                                                               \
 253                 /* Overflow in the output buffer.  */                         \
 254                 result = __GCONV_FULL_OUTPUT;                                 \
 255                 break;                                                        \
 256               }                                                               \
 257                                                                               \
 258             put16 (outptr, 0xd7c0 + (c >> 10));                               \
 259             outptr += 2;                                                      \
 260             put16 (outptr, 0xdc00 + (c & 0x3ff));                             \
 261           }                                                                   \
 262         else                                                                  \
 263           put16 (outptr, c);                                                  \
 264       }                                                                       \
 265     outptr += 2;                                                              \
 266     inptr += 4;                                                               \
 267   }
 268 #define LOOP_NEED_FLAGS
 269 #define EXTRA_LOOP_DECLS \
 270         , enum variant var, int swap
 271 #include <iconv/loop.c>
 272
 273
 274 /* Convert from UTF-16 to the internal (UCS4-like) format.  */
 275 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 276 #define MAX_NEEDED_INPUT        MAX_NEEDED_FROM
 277 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 278 #define LOOPFCT                 FROM_LOOP
 279 #define BODY \
 280   {                                                                           \
 281     uint16_t u1 = get16 (inptr);                                              \
 282                                                                               \
 283     if (swap)                                                                 \
 284       {                                                                       \
 285         u1 = bswap_16 (u1);                                                   \
 286                                                                               \
 287         if (__builtin_expect (u1, 0) < 0xd800 || u1 > 0xdfff)                 \
 288           {                                                                   \
 289             /* No surrogate.  */                                              \
 290             put32 (outptr, u1);                                               \
 291             inptr += 2;                                                       \
 292           }                                                                   \
 293         else                                                                  \
 294           {                                                                   \
 295             uint16_t u2;                                                      \
 296                                                                               \
 297             /* It's a surrogate character.  At least the first word says      \
 298                it is.  */                                                     \
 299             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 300               {                                                               \
 301                 /* We don't have enough input for another complete input      \
 302                    character.  */                                             \
 303                 result = __GCONV_INCOMPLETE_INPUT;                            \
 304                 break;                                                        \
 305               }                                                               \
 306                                                                               \
 307             inptr += 2;                                                       \
 308             u2 = bswap_16 (get16 (inptr));                                    \
 309             if (__builtin_expect (u2, 0xdc00) < 0xdc00                        \
 310                 || __builtin_expect (u2, 0xdc00) >= 0xdfff)                   \
 311               {                                                               \
 312                 /* This is no valid second word for a surrogate.  */          \
 313                 if (! ignore_errors_p ())                                     \
 314                   {                                                           \
 315                     inptr -= 2;                                               \
 316                     result = __GCONV_ILLEGAL_INPUT;                           \
 317                     break;                                                    \
 318                   }                                                           \
 319                                                                               \
 320                 ++*irreversible;                                              \
 321                 continue;                                                     \
 322               }                                                               \
 323                                                                               \
 324             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 325             inptr += 2;                                                       \
 326           }                                                                   \
 327       }                                                                       \
 328     else                                                                      \
 329       {                                                                       \
 330         if (__builtin_expect (u1, 0) < 0xd800 || u1 > 0xdfff)                 \
 331           {                                                                   \
 332             /* No surrogate.  */                                              \
 333             put32 (outptr, u1);                                               \
 334             inptr += 2;                                                       \
 335           }                                                                   \
 336         else                                                                  \
 337           {                                                                   \
 338             uint16_t u2;                                                      \
 339                                                                               \
 340             /* It's a surrogate character.  At least the first word says      \
 341                it is.  */                                                     \
 342             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 343               {                                                               \
 344                 /* We don't have enough input for another complete input      \
 345                    character.  */                                             \
 346                 result = __GCONV_INCOMPLETE_INPUT;                            \
 347                 break;                                                        \
 348               }                                                               \
 349                                                                               \
 350             inptr += 2;                                                       \
 351             u2 = get16 (inptr);                                               \
 352             if (__builtin_expect (u2, 0xdc00) < 0xdc00                        \
 353                 || __builtin_expect (u2, 0xdc00) >= 0xdfff)                   \
 354               {                                                               \
 355                 /* This is no valid second word for a surrogate.  */          \
 356                 if (! ignore_errors_p ())                                     \
 357                   {                                                           \
 358                     inptr -= 2;                                               \
 359                     result = __GCONV_ILLEGAL_INPUT;                           \
 360                     break;                                                    \
 361                   }                                                           \
 362                                                                               \
 363                 ++*irreversible;                                              \
 364                 continue;                                                     \
 365               }                                                               \
 366                                                                               \
 367             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 368             inptr += 2;                                                       \
 369           }                                                                   \
 370       }                                                                       \
 371     outptr += 4;                                                              \
 372   }
 373 #define LOOP_NEED_FLAGS
 374 #define EXTRA_LOOP_DECLS \
 375         , enum variant var, int swap
 376 #include <iconv/loop.c>
 377
 378
 379 /* Now define the toplevel functions.  */
 380 #include <iconv/skeleton.c>