iconv/skeleton.c

   1 /* Skeleton for a conversion module.
   2    Copyright (C) 1998, 1999, 2000 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998.
   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 /* This file can be included to provide definitions of several things
  22    many modules have in common.  It can be customized using the following
  23    macros:
  24
  25      DEFINE_INIT        define the default initializer.  This requires the
  26                         following symbol to be defined.
  27
  28      CHARSET_NAME       string with official name of the coded character
  29                         set (in all-caps)
  30
  31      DEFINE_FINI        define the default destructor function.
  32
  33      MIN_NEEDED_FROM    minimal number of bytes needed for the from-charset.
  34      MIN_NEEDED_TO      likewise for the to-charset.
  35
  36      MAX_NEEDED_FROM    maximal number of bytes needed for the from-charset.
  37                         This macro is optional, it defaults to MIN_NEEDED_FROM.
  38      MAX_NEEDED_TO      likewise for the to-charset.
  39
  40      DEFINE_DIRECTION_OBJECTS
  41                         two objects will be defined to be used when the
  42                         `gconv' function must only distinguish two
  43                         directions.  This is implied by DEFINE_INIT.
  44                         If this macro is not defined the following
  45                         macro must be available.
  46
  47      FROM_DIRECTION     this macro is supposed to return a value != 0
  48                         if we convert from the current character set,
  49                         otherwise it return 0.
  50
  51      EMIT_SHIFT_TO_INIT this symbol is optional.  If it is defined it
  52                         defines some code which writes out a sequence
  53                         of characters which bring the current state into
  54                         the initial state.
  55
  56      FROM_LOOP          name of the function implementing the conversion
  57                         from the current characters.
  58      TO_LOOP            likewise for the other direction
  59
  60      RESET_STATE        in case of an error we must reset the state for
  61                         the rerun so this macro must be defined for
  62                         stateful encodings.  It takes an argument which
  63                         is nonzero when saving.
  64
  65      RESET_INPUT_BUFFER If the input character sets allow this the macro
  66                         can be defined to reset the input buffer pointers
  67                         to cover only those characters up to the error.
  68
  69      FUNCTION_NAME      if not set the conversion function is named `gconv'.
  70
  71      PREPARE_LOOP       optional code preparing the conversion loop.  Can
  72                         contain variable definitions.
  73      END_LOOP           also optional, may be used to store information
  74
  75      EXTRA_LOOP_ARGS    optional macro specifying extra arguments passed
  76                         to loop function.
  77  */
  78
  79 #include <assert.h>
  80 #include <gconv.h>
  81 #include <string.h>
  82 #define __need_size_t
  83 #define __need_NULL
  84 #include <stddef.h>
  85
  86 #ifndef STATIC_GCONV
  87 # include <dlfcn.h>
  88 #endif
  89
  90 #ifndef DL_CALL_FCT
  91 # define DL_CALL_FCT(fct, args) fct args
  92 #endif
  93
  94 /* The direction objects.  */
  95 #if DEFINE_DIRECTION_OBJECTS || DEFINE_INIT
  96 static int from_object;
  97 static int to_object;
  98
  99 # ifndef FROM_DIRECTION
 100 #  define FROM_DIRECTION (step->__data == &from_object)
 101 # endif
 102 #else
 103 # ifndef FROM_DIRECTION
 104 #  error "FROM_DIRECTION must be provided if direction objects are not used"
 105 # endif
 106 #endif
 107
 108
 109 /* How many bytes are needed at most for the from-charset.  */
 110 #ifndef MAX_NEEDED_FROM
 111 # define MAX_NEEDED_FROM        MIN_NEEDED_FROM
 112 #endif
 113
 114 /* Same for the to-charset.  */
 115 #ifndef MAX_NEEDED_TO
 116 # define MAX_NEEDED_TO          MIN_NEEDED_TO
 117 #endif
 118
 119
 120 /* Define macros which can access unaligned buffers.  These macros are
 121    supposed to be used only in code outside the inner loops.  For the inner
 122    loops we have other definitions which allow optimized access.  */
 123 #ifdef _STRING_ARCH_unaligned
 124 /* We can handle unaligned memory access.  */
 125 # define get16u(addr) *((uint16_t *) (addr))
 126 # define get32u(addr) *((uint32_t *) (addr))
 127
 128 /* We need no special support for writing values either.  */
 129 # define put16u(addr, val) *((uint16_t *) (addr)) = (val)
 130 # define put32u(addr, val) *((uint32_t *) (addr)) = (val)
 131 #else
 132 /* Distinguish between big endian and little endian.  */
 133 # if __BYTE_ORDER == __LITTLE_ENDIAN
 134 #  define get16u(addr) \
 135      (((__const unsigned char *) (addr))[1] << 8                              \
 136       | ((__const unsigned char *) (addr))[0])
 137 #  define get32u(addr) \
 138      (((((__const unsigned char *) (addr))[3] << 8                            \
 139         | ((__const unsigned char *) (addr))[2]) << 8                         \
 140        | ((__const unsigned char *) (addr))[1]) << 8                          \
 141       | ((__const unsigned char *) (addr))[0])
 142
 143 #  define put16u(addr, val) \
 144      ({ uint16_t __val = (val);                                               \
 145         ((unsigned char *) (addr))[0] = __val;                                \
 146         ((unsigned char *) (addr))[1] = __val >> 8;                           \
 147         (void) 0; })
 148 #  define put32u(addr, val) \
 149      ({ uint32_t __val = (val);                                               \
 150         ((unsigned char *) (addr))[0] = __val;                                \
 151         __val >>= 8;                                                          \
 152         ((unsigned char *) (addr))[1] = __val;                                \
 153         __val >>= 8;                                                          \
 154         ((unsigned char *) (addr))[2] = __val;                                \
 155         __val >>= 8;                                                          \
 156         ((unsigned char *) (addr))[3] = __val;                                \
 157         (void) 0; })
 158 # else
 159 #  define get16u(addr) \
 160      (((__const unsigned char *) (addr))[0] << 8                              \
 161       | ((__const unsigned char *) (addr))[1])
 162 #  define get32u(addr) \
 163      (((((__const unsigned char *) (addr))[0] << 8                            \
 164         | ((__const unsigned char *) (addr))[1]) << 8                         \
 165        | ((__const unsigned char *) (addr))[2]) << 8                          \
 166       | ((__const unsigned char *) (addr))[3])
 167
 168 #  define put16u(addr, val) \
 169      ({ uint16_t __val = (val);                                               \
 170         ((unsigned char *) (addr))[1] = __val;                                \
 171         ((unsigned char *) (addr))[0] = __val >> 8;                           \
 172         (void) 0; })
 173 #  define put32u(addr, val) \
 174      ({ uint32_t __val = (val);                                               \
 175         ((unsigned char *) (addr))[3] = __val;                                \
 176         __val >>= 8;                                                          \
 177         ((unsigned char *) (addr))[2] = __val;                                \
 178         __val >>= 8;                                                          \
 179         ((unsigned char *) (addr))[1] = __val;                                \
 180         __val >>= 8;                                                          \
 181         ((unsigned char *) (addr))[0] = __val;                                \
 182         (void) 0; })
 183 # endif
 184 #endif
 185
 186
 187 /* For conversions from a fixed width character sets to another fixed width
 188    character set we we can define RESET_INPUT_BUFFER is necessary.  */
 189 #if !defined RESET_INPUT_BUFFER && !defined SAVE_RESET_STATE
 190 # if MIN_NEEDED_FROM == MAX_NEEDED_FROM && MIN_NEEDED_TO == MAX_NEEDED_TO
 191 /* We have to use these `if's here since the compiler cannot know that
 192    (outbuf - outerr) is always divisible by MIN_NEEDED_TO.  */
 193 #  define RESET_INPUT_BUFFER \
 194   if (MIN_NEEDED_FROM % MIN_NEEDED_TO == 0)                                   \
 195     *inptrp -= (outbuf - outerr) * (MIN_NEEDED_FROM / MIN_NEEDED_TO);         \
 196   else if (MIN_NEEDED_TO % MIN_NEEDED_FROM == 0)                              \
 197     *inptrp -= (outbuf - outerr) / (MIN_NEEDED_TO / MIN_NEEDED_FROM);         \
 198   else                                                                        \
 199     *inptrp -= ((outbuf - outerr) / MIN_NEEDED_TO) * MIN_NEEDED_FROM
 200 # endif
 201 #endif
 202
 203
 204 /* The default init function.  It simply matches the name and initializes
 205    the step data to point to one of the objects above.  */
 206 #if DEFINE_INIT
 207 # ifndef CHARSET_NAME
 208 #  error "CHARSET_NAME not defined"
 209 # endif
 210
 211 int
 212 gconv_init (struct __gconv_step *step)
 213 {
 214   /* Determine which direction.  */
 215   if (strcmp (step->__from_name, CHARSET_NAME) == 0)
 216     {
 217       step->__data = &from_object;
 218
 219       step->__min_needed_from = MIN_NEEDED_FROM;
 220       step->__max_needed_from = MAX_NEEDED_FROM;
 221       step->__min_needed_to = MIN_NEEDED_TO;
 222       step->__max_needed_to = MAX_NEEDED_TO;
 223     }
 224   else if (strcmp (step->__to_name, CHARSET_NAME) == 0)
 225     {
 226       step->__data = &to_object;
 227
 228       step->__min_needed_from = MIN_NEEDED_TO;
 229       step->__max_needed_from = MAX_NEEDED_TO;
 230       step->__min_needed_to = MIN_NEEDED_FROM;
 231       step->__max_needed_to = MAX_NEEDED_FROM;
 232     }
 233   else
 234     return __GCONV_NOCONV;
 235
 236 #ifdef RESET_STATE
 237   step->__stateful = 1;
 238 #else
 239   step->__stateful = 0;
 240 #endif
 241
 242   return __GCONV_OK;
 243 }
 244 #endif
 245
 246
 247 /* The default destructor function does nothing in the moment and so
 248    be define it at all.  But we still provide the macro just in case
 249    we need it some day.  */
 250 #if DEFINE_FINI
 251 #endif
 252
 253
 254 /* If no arguments have to passed to the loop function define the macro
 255    as empty.  */
 256 #ifndef EXTRA_LOOP_ARGS
 257 # define EXTRA_LOOP_ARGS
 258 #endif
 259
 260
 261 /* This is the actual conversion function.  */
 262 #ifndef FUNCTION_NAME
 263 # define FUNCTION_NAME  gconv
 264 #endif
 265
 266 /* The macros are used to access the function to convert single characters.  */
 267 #define SINGLE(fct) SINGLE2 (fct)
 268 #define SINGLE2(fct) fct##_single
 269
 270
 271 int
 272 FUNCTION_NAME (struct __gconv_step *step, struct __gconv_step_data *data,
 273                const unsigned char **inptrp, const unsigned char *inend,
 274                size_t *written, int do_flush, int consume_incomplete)
 275 {
 276   struct __gconv_step *next_step = step + 1;
 277   struct __gconv_step_data *next_data = data + 1;
 278   __gconv_fct fct = data->__is_last ? NULL : next_step->__fct;
 279   int status;
 280
 281   /* If the function is called with no input this means we have to reset
 282      to the initial state.  The possibly partly converted input is
 283      dropped.  */
 284   if (__builtin_expect (do_flush, 0))
 285     {
 286       status = __GCONV_OK;
 287
 288 #ifdef EMIT_SHIFT_TO_INIT
 289       /* Emit the escape sequence to reset the state.  */
 290       EMIT_SHIFT_TO_INIT;
 291 #endif
 292       /* Call the steps down the chain if there are any but only if we
 293          successfully emitted the escape sequence.  */
 294       if (status == __GCONV_OK && ! data->__is_last)
 295         status = DL_CALL_FCT (fct, (next_step, next_data, NULL, NULL,
 296                                     written, 1, consume_incomplete));
 297     }
 298   else
 299     {
 300       /* We preserve the initial values of the pointer variables.  */
 301       const unsigned char *inptr = *inptrp;
 302       unsigned char *outbuf = data->__outbuf;
 303       unsigned char *outend = data->__outbufend;
 304       unsigned char *outstart;
 305       /* This variable is used to count the number of characters we
 306          actually converted.  */
 307       size_t converted = 0;
 308 #if defined _STRING_ARCH_unaligned \
 309     || MIN_NEEDED_FROM == 1 || MAX_NEEDED_FROM % MIN_NEEDED_FROM != 0 \
 310     || MIN_NEEDED_TO == 1 || MAX_NEEDED_TO % MIN_NEEDED_TO != 0
 311 # define unaligned 0
 312 #else
 313       int unaligned;
 314 # define GEN_unaligned(name) GEN_unaligned2 (name)
 315 # define GEN_unaligned2(name) name##_unaligned
 316 #endif
 317
 318 #ifdef PREPARE_LOOP
 319       PREPARE_LOOP
 320 #endif
 321
 322 #if MAX_NEEDED_FROM > 1 || MAX_NEEDED_TO > 1
 323       /* If the function is used to implement the mb*towc*() or wc*tomb*()
 324          functions we must test whether any bytes from the last call are
 325          stored in the `state' object.  */
 326       if (((MAX_NEEDED_FROM > 1 && FROM_DIRECTION)
 327            || (MAX_NEEDED_TO > 1 && !FROM_DIRECTION))
 328           && consume_incomplete && (data->__statep->__count & 7) != 0)
 329         {
 330           /* Yep, we have some bytes left over.  Process them now.  */
 331
 332 # if MAX_NEEDED_FROM > 1
 333           if (MAX_NEEDED_TO == 1 || FROM_DIRECTION)
 334             status = SINGLE(FROM_LOOP) (inptrp, inend, &outbuf, outend,
 335                                         data->__statep, step->__data,
 336                                         &converted EXTRA_LOOP_ARGS);
 337 # endif
 338 # if MAX_NEEDED_FROM > 1 && MAX_NEEDED_TO > 1 && !ONE_DIRECTION
 339           else
 340 # endif
 341 # if MAX_NEEDED_TO > 1 && !ONE_DIRECTION
 342             status = SINGLE(TO_LOOP) (inptrp, inend, &outbuf, outend,
 343                                       data->__statep, step->__data,
 344                                       &converted EXTRA_LOOP_ARGS);
 345 # endif
 346
 347           if (status != __GCONV_OK)
 348             return status;
 349         }
 350 #endif
 351
 352 #if !defined _STRING_ARCH_unaligned \
 353     && MIN_NEEDED_FROM != 1 && MAX_NEEDED_FROM % MIN_NEEDED_FROM == 0 \
 354     && MIN_NEEDED_TO != 1 && MAX_NEEDED_TO % MIN_NEEDED_TO == 0
 355       /* The following assumes that encodings, which have a variable length
 356          what might unalign a buffer even though it is a aligned in the
 357          beginning, either don't have the minimal number of bytes as a divisor
 358          of the maximum length or have a minimum length of 1.  This is true
 359          for all known and supported encodings.  */
 360       unaligned = ((FROM_DIRECTION
 361                     && ((uintptr_t) inptr % MIN_NEEDED_FROM != 0
 362                         || (data->__is_last
 363                             && (uintptr_t) outbuf % MIN_NEEDED_TO != 0)))
 364                    || (!FROM_DIRECTION
 365                        && ((data->__is_last
 366                             && (uintptr_t) outbuf % MIN_NEEDED_FROM != 0)
 367                            || (uintptr_t) inptr % MIN_NEEDED_TO != 0)));
 368 #endif
 369
 370       do
 371         {
 372           /* Remember the start value for this round.  */
 373           inptr = *inptrp;
 374           /* The outbuf buffer is empty.  */
 375           outstart = outbuf;
 376
 377 #ifdef SAVE_RESET_STATE
 378           SAVE_RESET_STATE (1);
 379 #endif
 380
 381           if (!unaligned)
 382             {
 383               if (FROM_DIRECTION)
 384                 /* Run the conversion loop.  */
 385                 status = FROM_LOOP (inptrp, inend, &outbuf, outend,
 386                                     data->__statep, step->__data, &converted
 387                                     EXTRA_LOOP_ARGS);
 388               else
 389                 /* Run the conversion loop.  */
 390                 status = TO_LOOP (inptrp, inend, &outbuf, outend,
 391                                   data->__statep, step->__data, &converted
 392                                   EXTRA_LOOP_ARGS);
 393             }
 394 #if !defined _STRING_ARCH_unaligned \
 395     && MIN_NEEDED_FROM != 1 && MAX_NEEDED_FROM % MIN_NEEDED_FROM == 0 \
 396     && MIN_NEEDED_TO != 1 && MAX_NEEDED_TO % MIN_NEEDED_TO == 0
 397           else
 398             {
 399               if (FROM_DIRECTION)
 400                 /* Run the conversion loop.  */
 401                 status = GEN_unaligned (FROM_LOOP) (inptrp, inend, &outbuf,
 402                                                     outend, data->__statep,
 403                                                     step->__data, &converted
 404                                                     EXTRA_LOOP_ARGS);
 405               else
 406                 /* Run the conversion loop.  */
 407                 status = GEN_unaligned (TO_LOOP) (inptrp, inend, &outbuf,
 408                                                   outend, data->__statep,
 409                                                   step->__data, &converted
 410                                                   EXTRA_LOOP_ARGS);
 411             }
 412 #endif
 413
 414           /* We finished one use of the loops.  */
 415           ++data->__invocation_counter;
 416
 417           /* If this is the last step leave the loop, there is nothing
 418              we can do.  */
 419           if (data->__is_last)
 420             {
 421               /* Store information about how many bytes are available.  */
 422               data->__outbuf = outbuf;
 423
 424               /* Remember how many non-identical characters we converted.  */
 425               *written += converted;
 426
 427               break;
 428             }
 429
 430           /* Write out all output which was produced.  */
 431           if (outbuf > outstart)
 432             {
 433               const unsigned char *outerr = data->__outbuf;
 434               int result;
 435
 436               result = DL_CALL_FCT (fct, (next_step, next_data, &outerr,
 437                                           outbuf, written, 0,
 438                                           consume_incomplete));
 439
 440               if (result != __GCONV_EMPTY_INPUT)
 441                 {
 442                   if (__builtin_expect (outerr != outbuf, 0))
 443                     {
 444 #ifdef RESET_INPUT_BUFFER
 445                       RESET_INPUT_BUFFER;
 446 #else
 447                       /* We have a problem with the in on of the functions
 448                          below.  Undo the conversion upto the error point.  */
 449                       size_t nstatus;
 450
 451                       /* Reload the pointers.  */
 452                       *inptrp = inptr;
 453                       outbuf = outstart;
 454
 455                       /* Reset the state.  */
 456 # ifdef SAVE_RESET_STATE
 457                       SAVE_RESET_STATE (0);
 458 # endif
 459
 460                       if (FROM_DIRECTION)
 461                         /* Run the conversion loop.  */
 462                         nstatus = FROM_LOOP ((const unsigned char **) inptrp,
 463                                              (const unsigned char *) inend,
 464                                              (unsigned char **) &outbuf,
 465                                              (unsigned char *) outerr,
 466                                              data->__statep, step->__data,
 467                                              &converted EXTRA_LOOP_ARGS);
 468                       else
 469                         /* Run the conversion loop.  */
 470                         nstatus = TO_LOOP ((const unsigned char **) inptrp,
 471                                            (const unsigned char *) inend,
 472                                            (unsigned char **) &outbuf,
 473                                            (unsigned char *) outerr,
 474                                            data->__statep, step->__data,
 475                                            &converted EXTRA_LOOP_ARGS);
 476
 477                       /* We must run out of output buffer space in this
 478                          rerun.  */
 479                       assert (outbuf == outerr);
 480                       assert (nstatus == __GCONV_FULL_OUTPUT);
 481
 482                       /* If we haven't consumed a single byte decrement
 483                          the invocation counter.  */
 484                       if (outbuf == outstart)
 485                         --data->__invocation_counter;
 486 #endif  /* reset input buffer */
 487                     }
 488
 489                   /* Change the status.  */
 490                   status = result;
 491                 }
 492               else
 493                 /* All the output is consumed, we can make another run
 494                    if everything was ok.  */
 495                 if (status == __GCONV_FULL_OUTPUT)
 496                   status = __GCONV_OK;
 497             }
 498         }
 499       while (status == __GCONV_OK);
 500
 501 #ifdef END_LOOP
 502       END_LOOP
 503 #endif
 504
 505       /* If we are supposed to consume all character store now all of the
 506          remaining characters in the `state' object.  */
 507 #if MAX_NEEDED_FROM > 1 || MAX_NEEDED_TO > 1
 508       if (((MAX_NEEDED_FROM > 1 && FROM_DIRECTION)
 509            || (MAX_NEEDED_TO > 1 && !FROM_DIRECTION))
 510           && consume_incomplete && status == __GCONV_INCOMPLETE_INPUT)
 511         {
 512 # ifdef STORE_REST
 513           mbstate_t *state = data->__statep;
 514
 515           STORE_REST
 516 # else
 517           size_t cnt;
 518
 519           /* Make sure the remaining bytes fit into the state objects
 520              buffer.  */
 521           assert (inend - *inptrp < 4);
 522
 523           for (cnt = 0; *inptrp < inend; ++cnt)
 524             data->__statep->__value.__wchb[cnt] = *(*inptrp)++;
 525           data->__statep->__count &= ~7;
 526           data->__statep->__count |= cnt;
 527 # endif
 528         }
 529 #endif
 530     }
 531
 532   return status;
 533 }
 534
 535 #undef DEFINE_INIT
 536 #undef CHARSET_NAME
 537 #undef DEFINE_FINI
 538 #undef MIN_NEEDED_FROM
 539 #undef MIN_NEEDED_TO
 540 #undef MAX_NEEDED_FROM
 541 #undef MAX_NEEDED_TO
 542 #undef DEFINE_DIRECTION_OBJECTS
 543 #undef FROM_DIRECTION
 544 #undef EMIT_SHIFT_TO_INIT
 545 #undef FROM_LOOP
 546 #undef TO_LOOP
 547 #undef RESET_STATE
 548 #undef RESET_INPUT_BUFFER
 549 #undef FUNCTION_NAME
 550 #undef PREPARE_LOOP
 551 #undef END_LOOP
 552 #undef ONE_DIRECTION
 553 #undef STORE_REST