iconv/iconv_prog.c

   1 /* Convert text in given files from the specified from-set to the to-set.
   2    Copyright (C) 1998 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 #include <argp.h>
  22 #include <ctype.h>
  23 #include <errno.h>
  24 #include <error.h>
  25 #include <fcntl.h>
  26 #include <iconv.h>
  27 #include <locale.h>
  28 #include <search.h>
  29 #include <stdio.h>
  30 #include <stdlib.h>
  31 #include <string.h>
  32 #include <unistd.h>
  33 #include <sys/mman.h>
  34 #include <gconv_int.h>
  35
  36 /* Get libc version number.  */
  37 #include "../version.h"
  38
  39 #define PACKAGE _libc_intl_domainname
  40
  41
  42 /* Name and version of program.  */
  43 static void print_version (FILE *stream, struct argp_state *state);
  44 void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version;
  45
  46 #define OPT_VERBOSE     1000
  47 #define OPT_LIST        1001
  48
  49 /* Definitions of arguments for argp functions.  */
  50 static const struct argp_option options[] =
  51 {
  52   { NULL, 0, NULL, 0, N_("Input/Output format specification:") },
  53   { "from-code", 'f', "NAME", 0, N_("encoding of original text") },
  54   { "to-code", 't', "NAME", 0, N_("encoding for output") },
  55   { NULL, 0, NULL, 0, N_("Information:") },
  56   { "list", OPT_LIST, NULL, 0, N_("list all known coded character sets") },
  57   { NULL, 0, NULL, 0, N_("Output control:") },
  58   { "output", 'o', "FILE", 0, N_("output file") },
  59   { "verbose", OPT_VERBOSE, NULL, 0, N_("print progress information") },
  60   { NULL, 0, NULL, 0, NULL }
  61 };
  62
  63 /* Short description of program.  */
  64 static const char doc[] = N_("\
  65 Convert encoding of given files from one encoding to another.");
  66
  67 /* Strings for arguments in help texts.  */
  68 static const char args_doc[] = N_("[FILE...]");
  69
  70 /* Prototype for option handler.  */
  71 static error_t parse_opt __P ((int key, char *arg, struct argp_state *state));
  72
  73 /* Function to print some extra text in the help message.  */
  74 static char *more_help __P ((int key, const char *text, void *input));
  75
  76 /* Data structure to communicate with argp functions.  */
  77 static struct argp argp =
  78 {
  79   options, parse_opt, args_doc, doc, NULL, more_help
  80 };
  81
  82 /* Code sets to convert from and to respectively.  */
  83 static const char *from_code;
  84 static const char *to_code;
  85
  86 /* File to write output to.  If NULL write to stdout.  */
  87 static const char *output_file;
  88
  89 /* Nonzero if verbose ouput is wanted.  */
  90 static int verbose;
  91
  92 /* Nonzero if list of all coded character sets is wanted.  */
  93 static int list;
  94
  95 /* Prototypes for the functions doing the actual work.  */
  96 static int process_block (iconv_t cd, const char *addr, size_t len,
  97                           FILE *output);
  98 static int process_fd (iconv_t cd, int fd, FILE *output);
  99 static int process_file (iconv_t cd, FILE *input, FILE *output);
 100 static void print_known_names (void);
 101
 102
 103 int
 104 main (int argc, char *argv[])
 105 {
 106   int status = EXIT_SUCCESS;
 107   int remaining;
 108   FILE *output;
 109   iconv_t cd;
 110
 111   /* Set locale via LC_ALL.  */
 112   setlocale (LC_ALL, "");
 113
 114   /* Set the text message domain.  */
 115   textdomain (_libc_intl_domainname);
 116
 117   /* Parse and process arguments.  */
 118   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 119
 120   /* List all coded character sets if wanted.  */
 121   if (list)
 122     {
 123       print_known_names ();
 124       exit (EXIT_SUCCESS);
 125     }
 126
 127   /* If either the from- or to-code is not specified this is an error
 128      since we do not know what to do.  */
 129   if (from_code == NULL && to_code == NULL)
 130     error (EXIT_FAILURE, 0,
 131            _("neither original nor target encoding specified"));
 132   if (from_code == NULL)
 133     error (EXIT_FAILURE, 0, _("original encoding not specified using `-f'"));
 134   if (to_code == NULL)
 135     error (EXIT_FAILURE, 0, _("target encoding not specified using `-t'"));
 136
 137   /* Let's see whether we have these coded character sets.  */
 138   cd = iconv_open (to_code, from_code);
 139   if (cd == (iconv_t) -1)
 140     if (errno == EINVAL)
 141       error (EXIT_FAILURE, 0, _("conversion from `%s' to `%s' not supported"),
 142              from_code, to_code);
 143     else
 144       error (EXIT_FAILURE, errno, _("failed to start conversion processing"));
 145
 146   /* Determine output file.  */
 147   if (output_file != NULL)
 148     {
 149       output = fopen (output_file, "w");
 150       if (output == NULL)
 151         error (EXIT_FAILURE, errno, _("cannot open output file"));
 152     }
 153   else
 154     output = stdout;
 155
 156   /* Now process the remaining files.  Write them to stdout or the file
 157      specified with the `-o' parameter.  If we have no file given as
 158      the parameter process all from stdin.  */
 159   if (remaining == argc)
 160     process_file (cd, stdin, output);
 161   else
 162     do
 163       {
 164         struct stat st;
 165         const char *addr;
 166         int fd = open (argv[remaining], O_RDONLY);
 167
 168         if (verbose)
 169           printf ("%s:\n", argv[remaining]);
 170
 171         if (fd == -1)
 172           {
 173             error (0, errno, _("cannot open input file `%s'"),
 174                    argv[remaining]);
 175             status = EXIT_FAILURE;
 176             continue;
 177           }
 178
 179         /* We have possibilities for reading the input file.  First try
 180            to mmap() it since this will provide the fastest solution.  */
 181         if (fstat (fd, &st) == 0
 182             && ((addr = mmap (NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0))
 183                 != MAP_FAILED))
 184           {
 185             /* Yes, we can use mmap().  The descriptor is not needed
 186                anymore.  */
 187             if (close (fd) != 0)
 188               error (EXIT_FAILURE, errno, _("error while closing input `%s'"),
 189                      argv[remaining]);
 190
 191             if (process_block (cd, addr, st.st_size, output) < 0)
 192               {
 193                 /* Something went wrong.  */
 194                 status = EXIT_FAILURE;
 195
 196                 /* We don't need the input data anymore.  */
 197                 munmap ((void *) addr, st.st_size);
 198
 199                 /* We cannot go on with producing output since it might
 200                    lead to problem because the last output might leave
 201                    the output stream in an undefined state.  */
 202                 break;
 203               }
 204
 205             /* We don't need the input data anymore.  */
 206             munmap ((void *) addr, st.st_size);
 207           }
 208         else
 209           {
 210             /* Read the file in pieces.  */
 211             if (process_fd (cd, fd, output) != 0)
 212               {
 213                 /* Something went wrong.  */
 214                 status = EXIT_FAILURE;
 215
 216                 /* We don't need the input file anymore.  */
 217                 close (fd);
 218
 219                 /* We cannot go on with producing output since it might
 220                    lead to problem because the last output might leave
 221                    the output stream in an undefined state.  */
 222                 break;
 223               }
 224
 225             /* Now close the file.  */
 226             close (fd);
 227           }
 228       }
 229     while (++remaining < argc);
 230
 231   /* Close the output file now.  */
 232   if (fclose (output))
 233     error (EXIT_FAILURE, errno, _("error while closing output file"));
 234
 235   return status;
 236 }
 237
 238
 239 /* Handle program arguments.  */
 240 static error_t
 241 parse_opt (int key, char *arg, struct argp_state *state)
 242 {
 243   switch (key)
 244     {
 245     case 'f':
 246       from_code = arg;
 247       break;
 248     case 't':
 249       to_code = arg;
 250       break;
 251     case 'o':
 252       output_file = arg;
 253       break;
 254     case OPT_VERBOSE:
 255       verbose = 1;
 256       break;
 257     case OPT_LIST:
 258       list = 1;
 259       break;
 260     default:
 261       return ARGP_ERR_UNKNOWN;
 262     }
 263   return 0;
 264 }
 265
 266
 267 static char *
 268 more_help (int key, const char *text, void *input)
 269 {
 270   switch (key)
 271     {
 272     case ARGP_KEY_HELP_EXTRA:
 273       /* We print some extra information.  */
 274       return strdup (gettext ("\
 275 Report bugs using the `glibcbug' script to <bugs@gnu.org>.\n"));
 276     default:
 277       break;
 278     }
 279   return (char *) text;
 280 }
 281
 282
 283 /* Print the version information.  */
 284 static void
 285 print_version (FILE *stream, struct argp_state *state)
 286 {
 287   fprintf (stream, "iconv (GNU %s) %s\n", PACKAGE, VERSION);
 288   fprintf (stream, gettext ("\
 289 Copyright (C) %s Free Software Foundation, Inc.\n\
 290 This is free software; see the source for copying conditions.  There is NO\n\
 291 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 292 "), "1998");
 293   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 294 }
 295
 296
 297 static int
 298 process_block (iconv_t cd, const char *addr, size_t len, FILE *output)
 299 {
 300 #define OUTBUF_SIZE     32768
 301   const char *start = addr;
 302   char outbuf[OUTBUF_SIZE];
 303   char *outptr;
 304   size_t outlen;
 305   size_t n;
 306
 307   while (len > 0)
 308     {
 309       outptr = outbuf;
 310       outlen = OUTBUF_SIZE;
 311       n = iconv (cd, &addr, &len, &outptr, &outlen);
 312
 313       if (outptr != outbuf)
 314         {
 315           /* We have something to write out.  */
 316           if (fwrite (outbuf, 1, outptr - outbuf, output)  < outptr - outbuf
 317               || ferror (output))
 318             {
 319               /* Error occurred while printing the result.  */
 320               error (0, 0, _("\
 321 conversion stopped due to problem in writing the output"));
 322               return -1;
 323             }
 324         }
 325
 326       if (n != (size_t) -1)
 327         /* Everything is processed.  */
 328         break;
 329
 330       if (errno != E2BIG)
 331         {
 332           /* iconv() ran into a problem.  */
 333           switch (errno)
 334             {
 335             case EILSEQ:
 336               error (0, 0, _("illegal input sequence at position %ld"),
 337                      addr - start);
 338               break;
 339             case EINVAL:
 340               error (0, 0, _("\
 341 incomplete character or shift sequence at end of buffer"));
 342               break;
 343             case EBADF:
 344               error (0, 0, _("internal error (illegal descriptor)"));
 345               break;
 346             default:
 347               error (0, 0, _("unknown iconv() error %d"), errno);
 348               break;
 349             }
 350
 351           return -1;
 352         }
 353     }
 354
 355   return 0;
 356 }
 357
 358
 359 static int
 360 process_fd (iconv_t cd, int fd, FILE *output)
 361 {
 362   /* we have a problem with reading from a desriptor since we must not
 363      provide the iconv() function an incomplete character or shift
 364      sequence at the end of the buffer.  Since we have to deal with
 365      arbitrary encodings we must read the whole text in a buffer and
 366      process it in one step.  */
 367   static char *inbuf = NULL;
 368   static size_t maxlen = 0;
 369   char *inptr = NULL;
 370   size_t actlen = 0;
 371
 372   while (actlen < maxlen)
 373     {
 374       size_t n = read (fd, inptr, maxlen - actlen);
 375
 376       if (n == 0)
 377         /* No more text to read.  */
 378         break;
 379
 380       if (n == -1)
 381         {
 382           /* Error while reading.  */
 383           error (0, errno, _("error while reading the input"));
 384           return -1;
 385         }
 386
 387       inptr += n;
 388       actlen += n;
 389     }
 390
 391   if (actlen == maxlen)
 392     while (1)
 393       {
 394         size_t n;
 395
 396         /* Increase the buffer.  */
 397         maxlen += 32768;
 398         inbuf = realloc (inbuf, maxlen);
 399         if (inbuf == NULL)
 400           error (0, errno, _("unable to allocate buffer for input"));
 401         inptr = inbuf + actlen;
 402
 403         do
 404           {
 405             n = read (fd, inptr, maxlen - actlen);
 406
 407             if (n == 0)
 408               /* No more text to read.  */
 409               break;
 410
 411             if (n == -1)
 412               {
 413                 /* Error while reading.  */
 414                 error (0, errno, _("error while reading the input"));
 415                 return -1;
 416               }
 417
 418             inptr += n;
 419             actlen += n;
 420           }
 421         while (actlen < maxlen);
 422
 423         if (n == 0)
 424           /* Break again so we leave both loops.  */
 425           break;
 426       }
 427
 428   /* Now we have all the input in the buffer.  Process it in one run.  */
 429   return process_block (cd, inbuf, actlen, output);
 430 }
 431
 432
 433 static int
 434 process_file (iconv_t cd, FILE *input, FILE *output)
 435 {
 436   /* This should be safe since we use this function only for `stdin' and
 437      we haven't read anything so far.  */
 438   return process_fd (cd, fileno (input), output);
 439 }
 440
 441
 442 /* Print all known character sets/encodings.  */
 443 static void *printlist;
 444 static size_t column;
 445 static int not_first;
 446
 447 static void
 448 insert_print_list (const void *nodep, VISIT value, int level)
 449 {
 450   if (value == leaf || value == postorder)
 451     {
 452       const struct gconv_alias *s = *(const struct gconv_alias **) nodep;
 453       tsearch (s->fromname, &printlist, (__compar_fn_t) strverscmp);
 454     }
 455 }
 456
 457 static void
 458 do_print  (const void *nodep, VISIT value, int level)
 459 {
 460   if (value == leaf || value == postorder)
 461     {
 462       const char *s = *(const char **) nodep;
 463       size_t len = strlen (s);
 464       size_t cnt;
 465
 466       while (len > 0 && s[len - 1] == '/')
 467         --len;
 468
 469       for (cnt = 0; cnt < len; ++cnt)
 470         if (isalnum (s[cnt]))
 471           break;
 472       if (cnt == len)
 473         return;
 474
 475       if (not_first)
 476         {
 477           putchar (',');
 478           ++column;
 479
 480           if (column > 2 && column + len > 77)
 481             {
 482               fputs ("\n  ", stdout);
 483               column = 2;
 484             }
 485           else
 486             {
 487               putchar (' ');
 488               ++column;
 489             }
 490         }
 491       else
 492         not_first = 1;
 493
 494       fwrite (s, len, 1, stdout);
 495       column += len;
 496     }
 497 }
 498
 499 static void
 500 print_known_names (void)
 501 {
 502   size_t cnt;
 503   iconv_t h;
 504
 505   /* We must initialize the internal databases first.  */
 506   h = iconv_open ("L1", "L1");
 507   iconv_close (h);
 508
 509   /* First add the aliases.  */
 510   twalk (__gconv_alias_db, insert_print_list);
 511
 512   /* Add the from- and to-names from the known modules.  */
 513   for (cnt = 0; cnt < __gconv_nmodules; ++cnt)
 514     {
 515       if (__gconv_modules_db[cnt]->from_pattern == NULL)
 516         {
 517           if (strcmp (__gconv_modules_db[cnt]->from_constpfx, "INTERNAL"))
 518             tsearch (__gconv_modules_db[cnt]->from_constpfx, &printlist,
 519                      (__compar_fn_t) strcoll);
 520           if (strcmp (__gconv_modules_db[cnt]->to_string, "INTERNAL"))
 521             tsearch (__gconv_modules_db[cnt]->to_string, &printlist,
 522                      (__compar_fn_t) strcoll);
 523         }
 524       else
 525         if (strcmp (__gconv_modules_db[cnt]->from_pattern, "INTERNAL"))
 526           tsearch (__gconv_modules_db[cnt]->from_pattern, &printlist,
 527                    (__compar_fn_t) strcoll);
 528     }
 529
 530   fputs (_("\
 531 The following list contain all the coded character sets known.  This does\n\
 532 not necessarily mean that all combinations of these names can be used for\n\
 533 the FROM and TO command line parameters.  One coded character set can be\n\
 534 listed with several different names (aliases).\n\
 535   Some of the names are no plain strings but instead regular expressions and\n\
 536 they match a variety of names which can be given as parameters to the\n\
 537 program.\n\n  "), stdout);
 538
 539   /* Now print the collected names.  */
 540   column = 2;
 541   twalk (printlist, do_print);
 542
 543   if (column != 0)
 544     puts ("");
 545 }