iconvdata/utf-16.c

   1 /* Conversion module for UTF-16.
   2    Copyright (C) 1999, 2000-2002 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 Lesser General Public
   8    License as published by the Free Software Foundation; either
   9    version 2.1 of the 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    Lesser General Public License for more details.
  15
  16    You should have received a copy of the GNU Lesser General Public
  17    License along with the GNU C Library; if not, write to the Free
  18    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  19    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;                                                                   \
  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             *inptrp = inptr += 2;                                             \
  59           else if (get16u (inptr) == BOM_OE)                                  \
  60             {                                                                 \
  61               ((struct utf16_data *) step->__data)->swap = 1;                 \
  62               *inptrp = 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   swap = ((struct utf16_data *) step->__data)->swap;
  77 #define EXTRA_LOOP_ARGS         , var, swap
  78
  79
  80 /* Direction of the transformation.  */
  81 enum direction
  82 {
  83   illegal_dir,
  84   to_utf16,
  85   from_utf16
  86 };
  87
  88 enum variant
  89 {
  90   illegal_var,
  91   UTF_16,
  92   UTF_16LE,
  93   UTF_16BE
  94 };
  95
  96 struct utf16_data
  97 {
  98   enum direction dir;
  99   enum variant var;
 100   int swap;
 101 };
 102
 103
 104 extern int gconv_init (struct __gconv_step *step);
 105 int
 106 gconv_init (struct __gconv_step *step)
 107 {
 108   /* Determine which direction.  */
 109   struct utf16_data *new_data;
 110   enum direction dir = illegal_dir;
 111   enum variant var = illegal_var;
 112   int result;
 113
 114   if (__strcasecmp (step->__from_name, "UTF-16//") == 0)
 115     {
 116       dir = from_utf16;
 117       var = UTF_16;
 118     }
 119   else if (__strcasecmp (step->__to_name, "UTF-16//") == 0)
 120     {
 121       dir = to_utf16;
 122       var = UTF_16;
 123     }
 124   else if (__strcasecmp (step->__from_name, "UTF-16BE//") == 0)
 125     {
 126       dir = from_utf16;
 127       var = UTF_16BE;
 128     }
 129   else if (__strcasecmp (step->__to_name, "UTF-16BE//") == 0)
 130     {
 131       dir = to_utf16;
 132       var = UTF_16BE;
 133     }
 134   else if (__strcasecmp (step->__from_name, "UTF-16LE//") == 0)
 135     {
 136       dir = from_utf16;
 137       var = UTF_16LE;
 138     }
 139   else if (__strcasecmp (step->__to_name, "UTF-16LE//") == 0)
 140     {
 141       dir = to_utf16;
 142       var = UTF_16LE;
 143     }
 144
 145   result = __GCONV_NOCONV;
 146   if (__builtin_expect (dir, to_utf16) != illegal_dir)
 147     {
 148       new_data = (struct utf16_data *) malloc (sizeof (struct utf16_data));
 149
 150       result = __GCONV_NOMEM;
 151       if (new_data != NULL)
 152         {
 153           new_data->dir = dir;
 154           new_data->var = var;
 155           new_data->swap = ((var == UTF_16LE && BYTE_ORDER == BIG_ENDIAN)
 156                             || (var == UTF_16BE
 157                                 && BYTE_ORDER == LITTLE_ENDIAN));
 158           step->__data = new_data;
 159
 160           if (dir == from_utf16)
 161             {
 162               step->__min_needed_from = MIN_NEEDED_FROM;
 163               step->__max_needed_from = MAX_NEEDED_FROM;
 164               step->__min_needed_to = MIN_NEEDED_TO;
 165               step->__max_needed_to = MIN_NEEDED_TO;
 166             }
 167           else
 168             {
 169               step->__min_needed_from = MIN_NEEDED_TO;
 170               step->__max_needed_from = MIN_NEEDED_TO;
 171               step->__min_needed_to = MIN_NEEDED_FROM;
 172               step->__max_needed_to = MAX_NEEDED_FROM;
 173             }
 174
 175           step->__stateful = 0;
 176
 177           result = __GCONV_OK;
 178         }
 179     }
 180
 181   return result;
 182 }
 183
 184
 185 extern void gconv_end (struct __gconv_step *data);
 186 void
 187 gconv_end (struct __gconv_step *data)
 188 {
 189   free (data->__data);
 190 }
 191
 192
 193 /* Convert from the internal (UCS4-like) format to UTF-16.  */
 194 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 195 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 196 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 197 #define LOOPFCT                 TO_LOOP
 198 #define BODY \
 199   {                                                                           \
 200     uint32_t c = get32 (inptr);                                               \
 201                                                                               \
 202     if (__builtin_expect (c >= 0xd800 && c < 0xe000, 0))                      \
 203       {                                                                       \
 204         /* Surrogate characters in UCS-4 input are not valid.                 \
 205            We must catch this.  If we let surrogates pass through,            \
 206            attackers could make a security hole exploit by                    \
 207            synthesizing any desired plane 1-16 character.  */                 \
 208         result = __GCONV_ILLEGAL_INPUT;                                       \
 209         if (! ignore_errors_p ())                                             \
 210           break;                                                              \
 211         inptr += 4;                                                           \
 212         ++*irreversible;                                                      \
 213         continue;                                                             \
 214       }                                                                       \
 215                                                                               \
 216     if (swap)                                                                 \
 217       {                                                                       \
 218         if (__builtin_expect (c >= 0x10000, 0))                               \
 219           {                                                                   \
 220             if (__builtin_expect (c >= 0x110000, 0))                          \
 221               {                                                               \
 222                 STANDARD_TO_LOOP_ERR_HANDLER (4);                             \
 223               }                                                               \
 224                                                                               \
 225             /* Generate a surrogate character.  */                            \
 226             if (__builtin_expect (outptr + 4 > outend, 0))                    \
 227               {                                                               \
 228                 /* Overflow in the output buffer.  */                         \
 229                 result = __GCONV_FULL_OUTPUT;                                 \
 230                 break;                                                        \
 231               }                                                               \
 232                                                                               \
 233             put16 (outptr, bswap_16 (0xd7c0 + (c >> 10)));                    \
 234             outptr += 2;                                                      \
 235             put16 (outptr, bswap_16 (0xdc00 + (c & 0x3ff)));                  \
 236           }                                                                   \
 237         else                                                                  \
 238           put16 (outptr, bswap_16 (c));                                       \
 239       }                                                                       \
 240     else                                                                      \
 241       {                                                                       \
 242         if (__builtin_expect (c >= 0x10000, 0))                               \
 243           {                                                                   \
 244             if (__builtin_expect (c >= 0x110000, 0))                          \
 245               {                                                               \
 246                 STANDARD_TO_LOOP_ERR_HANDLER (4);                             \
 247               }                                                               \
 248                                                                               \
 249             /* Generate a surrogate character.  */                            \
 250             if (__builtin_expect (outptr + 4 > outend, 0))                    \
 251               {                                                               \
 252                 /* Overflow in the output buffer.  */                         \
 253                 result = __GCONV_FULL_OUTPUT;                                 \
 254                 break;                                                        \
 255               }                                                               \
 256                                                                               \
 257             put16 (outptr, 0xd7c0 + (c >> 10));                               \
 258             outptr += 2;                                                      \
 259             put16 (outptr, 0xdc00 + (c & 0x3ff));                             \
 260           }                                                                   \
 261         else                                                                  \
 262           put16 (outptr, c);                                                  \
 263       }                                                                       \
 264     outptr += 2;                                                              \
 265     inptr += 4;                                                               \
 266   }
 267 #define LOOP_NEED_FLAGS
 268 #define EXTRA_LOOP_DECLS \
 269         , enum variant var, int swap
 270 #include <iconv/loop.c>
 271
 272
 273 /* Convert from UTF-16 to the internal (UCS4-like) format.  */
 274 #define MIN_NEEDED_INPUT        MIN_NEEDED_FROM
 275 #define MAX_NEEDED_INPUT        MAX_NEEDED_FROM
 276 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_TO
 277 #define LOOPFCT                 FROM_LOOP
 278 #define BODY \
 279   {                                                                           \
 280     uint16_t u1 = get16 (inptr);                                              \
 281                                                                               \
 282     if (swap)                                                                 \
 283       {                                                                       \
 284         u1 = bswap_16 (u1);                                                   \
 285                                                                               \
 286         if (__builtin_expect (u1 < 0xd800, 1) || u1 > 0xdfff)                 \
 287           {                                                                   \
 288             /* No surrogate.  */                                              \
 289             put32 (outptr, u1);                                               \
 290             inptr += 2;                                                       \
 291           }                                                                   \
 292         else                                                                  \
 293           {                                                                   \
 294             uint16_t u2;                                                      \
 295                                                                               \
 296             /* It's a surrogate character.  At least the first word says      \
 297                it is.  */                                                     \
 298             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 299               {                                                               \
 300                 /* We don't have enough input for another complete input      \
 301                    character.  */                                             \
 302                 result = __GCONV_INCOMPLETE_INPUT;                            \
 303                 break;                                                        \
 304               }                                                               \
 305                                                                               \
 306             inptr += 2;                                                       \
 307             u2 = bswap_16 (get16 (inptr));                                    \
 308             if (__builtin_expect (u2 < 0xdc00, 0)                             \
 309                 || __builtin_expect (u2 == 0xdfff, 0))                        \
 310               {                                                               \
 311                 /* This is no valid second word for a surrogate.  */          \
 312                 inptr -= 2;                                                   \
 313                 STANDARD_FROM_LOOP_ERR_HANDLER (2);                           \
 314               }                                                               \
 315                                                                               \
 316             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 317             inptr += 2;                                                       \
 318           }                                                                   \
 319       }                                                                       \
 320     else                                                                      \
 321       {                                                                       \
 322         if (__builtin_expect (u1 < 0xd800, 1) || u1 > 0xdfff)                 \
 323           {                                                                   \
 324             /* No surrogate.  */                                              \
 325             put32 (outptr, u1);                                               \
 326             inptr += 2;                                                       \
 327           }                                                                   \
 328         else                                                                  \
 329           {                                                                   \
 330             uint16_t u2;                                                      \
 331                                                                               \
 332             /* It's a surrogate character.  At least the first word says      \
 333                it is.  */                                                     \
 334             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 335               {                                                               \
 336                 /* We don't have enough input for another complete input      \
 337                    character.  */                                             \
 338                 result = __GCONV_INCOMPLETE_INPUT;                            \
 339                 break;                                                        \
 340               }                                                               \
 341                                                                               \
 342             inptr += 2;                                                       \
 343             u2 = get16 (inptr);                                               \
 344             if (__builtin_expect (u2 < 0xdc00, 0)                             \
 345                 || __builtin_expect (u2 >= 0xdfff, 0))                        \
 346               {                                                               \
 347                 /* This is no valid second word for a surrogate.  */          \
 348                 inptr -= 2;                                                   \
 349                 STANDARD_FROM_LOOP_ERR_HANDLER (2);                           \
 350               }                                                               \
 351                                                                               \
 352             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 353             inptr += 2;                                                       \
 354           }                                                                   \
 355       }                                                                       \
 356     outptr += 4;                                                              \
 357   }
 358 #define LOOP_NEED_FLAGS
 359 #define EXTRA_LOOP_DECLS \
 360         , enum variant var, int swap
 361 #include <iconv/loop.c>
 362
 363
 364 /* Now define the toplevel functions.  */
 365 #include <iconv/skeleton.c>