iconvdata/utf-16.c

   1 /* Conversion module for UTF-16.
   2    Copyright (C) 1999, 2000-2002, 2003 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 (inptr == inend                                            \
  55                     ? __GCONV_EMPTY_INPUT : __GCONV_INCOMPLETE_INPUT);        \
  56                                                                               \
  57           if (get16u (inptr) == BOM)                                          \
  58             /* Simply ignore the BOM character.  */                           \
  59             *inptrp = inptr += 2;                                             \
  60           else if (get16u (inptr) == BOM_OE)                                  \
  61             {                                                                 \
  62               ((struct utf16_data *) step->__data)->swap = 1;                 \
  63               *inptrp = inptr += 2;                                           \
  64             }                                                                 \
  65         }                                                                     \
  66     }                                                                         \
  67   else if (!FROM_DIRECTION && var == UTF_16 && !data->__internal_use          \
  68            && data->__invocation_counter == 0)                                \
  69     {                                                                         \
  70       /* Emit the Byte Order Mark.  */                                        \
  71       if (__builtin_expect (outbuf + 2 > outend, 0))                          \
  72         return __GCONV_FULL_OUTPUT;                                           \
  73                                                                               \
  74       put16u (outbuf, BOM);                                                   \
  75       outbuf += 2;                                                            \
  76     }                                                                         \
  77   swap = ((struct utf16_data *) step->__data)->swap;
  78 #define EXTRA_LOOP_ARGS         , var, swap
  79
  80
  81 /* Direction of the transformation.  */
  82 enum direction
  83 {
  84   illegal_dir,
  85   to_utf16,
  86   from_utf16
  87 };
  88
  89 enum variant
  90 {
  91   illegal_var,
  92   UTF_16,
  93   UTF_16LE,
  94   UTF_16BE
  95 };
  96
  97 struct utf16_data
  98 {
  99   enum direction dir;
 100   enum variant var;
 101   int swap;
 102 };
 103
 104
 105 extern int gconv_init (struct __gconv_step *step);
 106 int
 107 gconv_init (struct __gconv_step *step)
 108 {
 109   /* Determine which direction.  */
 110   struct utf16_data *new_data;
 111   enum direction dir = illegal_dir;
 112   enum variant var = illegal_var;
 113   int result;
 114
 115   if (__strcasecmp (step->__from_name, "UTF-16//") == 0)
 116     {
 117       dir = from_utf16;
 118       var = UTF_16;
 119     }
 120   else if (__strcasecmp (step->__to_name, "UTF-16//") == 0)
 121     {
 122       dir = to_utf16;
 123       var = UTF_16;
 124     }
 125   else if (__strcasecmp (step->__from_name, "UTF-16BE//") == 0)
 126     {
 127       dir = from_utf16;
 128       var = UTF_16BE;
 129     }
 130   else if (__strcasecmp (step->__to_name, "UTF-16BE//") == 0)
 131     {
 132       dir = to_utf16;
 133       var = UTF_16BE;
 134     }
 135   else if (__strcasecmp (step->__from_name, "UTF-16LE//") == 0)
 136     {
 137       dir = from_utf16;
 138       var = UTF_16LE;
 139     }
 140   else if (__strcasecmp (step->__to_name, "UTF-16LE//") == 0)
 141     {
 142       dir = to_utf16;
 143       var = UTF_16LE;
 144     }
 145
 146   result = __GCONV_NOCONV;
 147   if (__builtin_expect (dir, to_utf16) != illegal_dir)
 148     {
 149       new_data = (struct utf16_data *) malloc (sizeof (struct utf16_data));
 150
 151       result = __GCONV_NOMEM;
 152       if (new_data != NULL)
 153         {
 154           new_data->dir = dir;
 155           new_data->var = var;
 156           new_data->swap = ((var == UTF_16LE && BYTE_ORDER == BIG_ENDIAN)
 157                             || (var == UTF_16BE
 158                                 && BYTE_ORDER == LITTLE_ENDIAN));
 159           step->__data = new_data;
 160
 161           if (dir == from_utf16)
 162             {
 163               step->__min_needed_from = MIN_NEEDED_FROM;
 164               step->__max_needed_from = MAX_NEEDED_FROM;
 165               step->__min_needed_to = MIN_NEEDED_TO;
 166               step->__max_needed_to = MIN_NEEDED_TO;
 167             }
 168           else
 169             {
 170               step->__min_needed_from = MIN_NEEDED_TO;
 171               step->__max_needed_from = MIN_NEEDED_TO;
 172               step->__min_needed_to = MIN_NEEDED_FROM;
 173               step->__max_needed_to = MAX_NEEDED_FROM;
 174             }
 175
 176           step->__stateful = 0;
 177
 178           result = __GCONV_OK;
 179         }
 180     }
 181
 182   return result;
 183 }
 184
 185
 186 extern void gconv_end (struct __gconv_step *data);
 187 void
 188 gconv_end (struct __gconv_step *data)
 189 {
 190   free (data->__data);
 191 }
 192
 193
 194 /* Convert from the internal (UCS4-like) format to UTF-16.  */
 195 #define MIN_NEEDED_INPUT        MIN_NEEDED_TO
 196 #define MIN_NEEDED_OUTPUT       MIN_NEEDED_FROM
 197 #define MAX_NEEDED_OUTPUT       MAX_NEEDED_FROM
 198 #define LOOPFCT                 TO_LOOP
 199 #define BODY \
 200   {                                                                           \
 201     uint32_t c = get32 (inptr);                                               \
 202                                                                               \
 203     if (__builtin_expect (c >= 0xd800 && c < 0xe000, 0))                      \
 204       {                                                                       \
 205         /* Surrogate characters in UCS-4 input are not valid.                 \
 206            We must catch this.  If we let surrogates pass through,            \
 207            attackers could make a security hole exploit by                    \
 208            synthesizing any desired plane 1-16 character.  */                 \
 209         result = __GCONV_ILLEGAL_INPUT;                                       \
 210         if (! ignore_errors_p ())                                             \
 211           break;                                                              \
 212         inptr += 4;                                                           \
 213         ++*irreversible;                                                      \
 214         continue;                                                             \
 215       }                                                                       \
 216                                                                               \
 217     if (swap)                                                                 \
 218       {                                                                       \
 219         if (__builtin_expect (c >= 0x10000, 0))                               \
 220           {                                                                   \
 221             if (__builtin_expect (c >= 0x110000, 0))                          \
 222               {                                                               \
 223                 STANDARD_TO_LOOP_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 >= 0x10000, 0))                               \
 244           {                                                                   \
 245             if (__builtin_expect (c >= 0x110000, 0))                          \
 246               {                                                               \
 247                 STANDARD_TO_LOOP_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 < 0xd800, 1) || 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, 0)                             \
 310                 || __builtin_expect (u2 > 0xdfff, 0))                         \
 311               {                                                               \
 312                 /* This is no valid second word for a surrogate.  */          \
 313                 inptr -= 2;                                                   \
 314                 STANDARD_FROM_LOOP_ERR_HANDLER (2);                           \
 315               }                                                               \
 316                                                                               \
 317             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 318             inptr += 2;                                                       \
 319           }                                                                   \
 320       }                                                                       \
 321     else                                                                      \
 322       {                                                                       \
 323         if (__builtin_expect (u1 < 0xd800, 1) || u1 > 0xdfff)                 \
 324           {                                                                   \
 325             /* No surrogate.  */                                              \
 326             put32 (outptr, u1);                                               \
 327             inptr += 2;                                                       \
 328           }                                                                   \
 329         else                                                                  \
 330           {                                                                   \
 331             uint16_t u2;                                                      \
 332                                                                               \
 333             /* It's a surrogate character.  At least the first word says      \
 334                it is.  */                                                     \
 335             if (__builtin_expect (inptr + 4 > inend, 0))                      \
 336               {                                                               \
 337                 /* We don't have enough input for another complete input      \
 338                    character.  */                                             \
 339                 result = __GCONV_INCOMPLETE_INPUT;                            \
 340                 break;                                                        \
 341               }                                                               \
 342                                                                               \
 343             inptr += 2;                                                       \
 344             u2 = get16 (inptr);                                               \
 345             if (__builtin_expect (u2 < 0xdc00, 0)                             \
 346                 || __builtin_expect (u2 > 0xdfff, 0))                         \
 347               {                                                               \
 348                 /* This is no valid second word for a surrogate.  */          \
 349                 inptr -= 2;                                                   \
 350                 STANDARD_FROM_LOOP_ERR_HANDLER (2);                           \
 351               }                                                               \
 352                                                                               \
 353             put32 (outptr, ((u1 - 0xd7c0) << 10) + (u2 - 0xdc00));            \
 354             inptr += 2;                                                       \
 355           }                                                                   \
 356       }                                                                       \
 357     outptr += 4;                                                              \
 358   }
 359 #define LOOP_NEED_FLAGS
 360 #define EXTRA_LOOP_DECLS \
 361         , enum variant var, int swap
 362 #include <iconv/loop.c>
 363
 364
 365 /* Now define the toplevel functions.  */
 366 #include <iconv/skeleton.c>