| 1 | /* Convert text in given files from the specified from-set to the to-set. |
| 2 | Copyright (C) 1998-2021 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 | This program is free software; you can redistribute it and/or modify |
| 7 | it under the terms of the GNU General Public License as published |
| 8 | by the Free Software Foundation; version 2 of the License, or |
| 9 | (at your option) any later version. |
| 10 | |
| 11 | This program 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 |
| 14 | GNU General Public License for more details. |
| 15 | |
| 16 | You should have received a copy of the GNU General Public License |
| 17 | along with this program; if not, see <https://www.gnu.org/licenses/>. */ |
| 18 | |
| 19 | #include <argp.h> |
| 20 | #include <assert.h> |
| 21 | #include <ctype.h> |
| 22 | #include <errno.h> |
| 23 | #include <error.h> |
| 24 | #include <fcntl.h> |
| 25 | #include <iconv.h> |
| 26 | #include <langinfo.h> |
| 27 | #include <locale.h> |
| 28 | #include <search.h> |
| 29 | #include <stdbool.h> |
| 30 | #include <stdio.h> |
| 31 | #include <stdlib.h> |
| 32 | #include <string.h> |
| 33 | #include <unistd.h> |
| 34 | #include <libintl.h> |
| 35 | #ifdef _POSIX_MAPPED_FILES |
| 36 | # include <sys/mman.h> |
| 37 | #endif |
| 38 | #include <charmap.h> |
| 39 | #include <gconv_int.h> |
| 40 | #include "iconv_prog.h" |
| 41 | #include "iconvconfig.h" |
| 42 | #include "gconv_charset.h" |
| 43 | |
| 44 | /* Get libc version number. */ |
| 45 | #include "../version.h" |
| 46 | |
| 47 | #define PACKAGE _libc_intl_domainname |
| 48 | |
| 49 | |
| 50 | /* Name and version of program. */ |
| 51 | static void print_version (FILE *stream, struct argp_state *state); |
| 52 | void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version; |
| 53 | |
| 54 | #define OPT_VERBOSE 1000 |
| 55 | #define OPT_LIST 'l' |
| 56 | |
| 57 | /* Definitions of arguments for argp functions. */ |
| 58 | static const struct argp_option options[] = |
| 59 | { |
| 60 | { NULL, 0, NULL, 0, N_("Input/Output format specification:" ) }, |
| 61 | { "from-code" , 'f', N_("NAME" ), 0, N_("encoding of original text" ) }, |
| 62 | { "to-code" , 't', N_("NAME" ), 0, N_("encoding for output" ) }, |
| 63 | { NULL, 0, NULL, 0, N_("Information:" ) }, |
| 64 | { "list" , 'l', NULL, 0, N_("list all known coded character sets" ) }, |
| 65 | { NULL, 0, NULL, 0, N_("Output control:" ) }, |
| 66 | { NULL, 'c', NULL, 0, N_("omit invalid characters from output" ) }, |
| 67 | { "output" , 'o', N_("FILE" ), 0, N_("output file" ) }, |
| 68 | { "silent" , 's', NULL, 0, N_("suppress warnings" ) }, |
| 69 | { "verbose" , OPT_VERBOSE, NULL, 0, N_("print progress information" ) }, |
| 70 | { NULL, 0, NULL, 0, NULL } |
| 71 | }; |
| 72 | |
| 73 | /* Short description of program. */ |
| 74 | static const char doc[] = N_("\ |
| 75 | Convert encoding of given files from one encoding to another." ); |
| 76 | |
| 77 | /* Strings for arguments in help texts. */ |
| 78 | static const char args_doc[] = N_("[FILE...]" ); |
| 79 | |
| 80 | /* Prototype for option handler. */ |
| 81 | static error_t parse_opt (int key, char *arg, struct argp_state *state); |
| 82 | |
| 83 | /* Function to print some extra text in the help message. */ |
| 84 | static char *more_help (int key, const char *text, void *input); |
| 85 | |
| 86 | /* Data structure to communicate with argp functions. */ |
| 87 | static struct argp argp = |
| 88 | { |
| 89 | options, parse_opt, args_doc, doc, NULL, more_help |
| 90 | }; |
| 91 | |
| 92 | /* Code sets to convert from and to respectively. An empty string as the |
| 93 | default causes the 'iconv_open' function to look up the charset of the |
| 94 | currently selected locale and use it. */ |
| 95 | static const char *from_code = "" ; |
| 96 | static const char *to_code = "" ; |
| 97 | |
| 98 | /* File to write output to. If NULL write to stdout. */ |
| 99 | static const char *output_file; |
| 100 | |
| 101 | /* Nonzero if list of all coded character sets is wanted. */ |
| 102 | static int list; |
| 103 | |
| 104 | /* If nonzero omit invalid character from output. */ |
| 105 | int omit_invalid; |
| 106 | |
| 107 | /* Prototypes for the functions doing the actual work. */ |
| 108 | static int process_block (iconv_t cd, char *addr, size_t len, FILE **output, |
| 109 | const char *output_file); |
| 110 | static int process_fd (iconv_t cd, int fd, FILE **output, |
| 111 | const char *output_file); |
| 112 | static int process_file (iconv_t cd, FILE *input, FILE **output, |
| 113 | const char *output_file); |
| 114 | static void print_known_names (void); |
| 115 | |
| 116 | |
| 117 | int |
| 118 | main (int argc, char *argv[]) |
| 119 | { |
| 120 | int status = EXIT_SUCCESS; |
| 121 | int remaining; |
| 122 | __gconv_t cd; |
| 123 | struct charmap_t *from_charmap = NULL; |
| 124 | struct charmap_t *to_charmap = NULL; |
| 125 | |
| 126 | /* Set locale via LC_ALL. */ |
| 127 | setlocale (LC_ALL, "" ); |
| 128 | |
| 129 | /* Set the text message domain. */ |
| 130 | textdomain (_libc_intl_domainname); |
| 131 | |
| 132 | /* Parse and process arguments. */ |
| 133 | argp_parse (&argp, argc, argv, 0, &remaining, NULL); |
| 134 | |
| 135 | /* List all coded character sets if wanted. */ |
| 136 | if (list) |
| 137 | { |
| 138 | print_known_names (); |
| 139 | exit (EXIT_SUCCESS); |
| 140 | } |
| 141 | |
| 142 | /* POSIX 1003.2b introduces a silly thing: the arguments to -t anf -f |
| 143 | can be file names of charmaps. In this case iconv will have to read |
| 144 | those charmaps and use them to do the conversion. But there are |
| 145 | holes in the specification. There is nothing said that if -f is a |
| 146 | charmap filename that -t must be, too. And vice versa. There is |
| 147 | also no word about the symbolic names used. What if they don't |
| 148 | match? */ |
| 149 | if (strchr (from_code, '/') != NULL) |
| 150 | /* The from-name might be a charmap file name. Try reading the |
| 151 | file. */ |
| 152 | from_charmap = charmap_read (from_code, /*0, 1*/1, 0, 0, 0); |
| 153 | |
| 154 | if (strchr (to_code, '/') != NULL) |
| 155 | /* The to-name might be a charmap file name. Try reading the |
| 156 | file. */ |
| 157 | to_charmap = charmap_read (to_code, /*0, 1,*/1, 0, 0, 0); |
| 158 | |
| 159 | |
| 160 | /* At this point we have to handle two cases. The first one is |
| 161 | where a charmap is used for the from- or to-charset, or both. We |
| 162 | handle this special since it is very different from the sane way of |
| 163 | doing things. The other case allows converting using the iconv() |
| 164 | function. */ |
| 165 | if (from_charmap != NULL || to_charmap != NULL) |
| 166 | /* Construct the conversion table and do the conversion. */ |
| 167 | status = charmap_conversion (from_code, from_charmap, to_code, to_charmap, |
| 168 | argc, remaining, argv, output_file); |
| 169 | else |
| 170 | { |
| 171 | struct gconv_spec conv_spec; |
| 172 | int res; |
| 173 | |
| 174 | if (__gconv_create_spec (&conv_spec, from_code, to_code) == NULL) |
| 175 | { |
| 176 | error (EXIT_FAILURE, errno, |
| 177 | _("failed to start conversion processing" )); |
| 178 | exit (1); |
| 179 | } |
| 180 | |
| 181 | if (omit_invalid) |
| 182 | conv_spec.ignore = true; |
| 183 | |
| 184 | /* Let's see whether we have these coded character sets. */ |
| 185 | res = __gconv_open (&conv_spec, &cd, 0); |
| 186 | |
| 187 | __gconv_destroy_spec (&conv_spec); |
| 188 | |
| 189 | if (res != __GCONV_OK) |
| 190 | { |
| 191 | if (errno == EINVAL) |
| 192 | { |
| 193 | /* Try to be nice with the user and tell her which of the |
| 194 | two encoding names is wrong. This is possible because |
| 195 | all supported encodings can be converted from/to Unicode, |
| 196 | in other words, because the graph of encodings is |
| 197 | connected. */ |
| 198 | bool from_wrong = |
| 199 | (iconv_open ("UTF-8" , from_code) == (iconv_t) -1 |
| 200 | && errno == EINVAL); |
| 201 | bool to_wrong = |
| 202 | (iconv_open (to_code, "UTF-8" ) == (iconv_t) -1 |
| 203 | && errno == EINVAL); |
| 204 | const char *from_pretty = |
| 205 | (from_code[0] ? from_code : nl_langinfo (CODESET)); |
| 206 | const char *to_pretty = |
| 207 | (to_code[0] ? to_code : nl_langinfo (CODESET)); |
| 208 | |
| 209 | if (from_wrong) |
| 210 | { |
| 211 | if (to_wrong) |
| 212 | error (0, 0, |
| 213 | _("\ |
| 214 | conversions from `%s' and to `%s' are not supported" ), |
| 215 | from_pretty, to_pretty); |
| 216 | else |
| 217 | error (0, 0, |
| 218 | _("conversion from `%s' is not supported" ), |
| 219 | from_pretty); |
| 220 | } |
| 221 | else |
| 222 | { |
| 223 | if (to_wrong) |
| 224 | error (0, 0, |
| 225 | _("conversion to `%s' is not supported" ), |
| 226 | to_pretty); |
| 227 | else |
| 228 | error (0, 0, |
| 229 | _("conversion from `%s' to `%s' is not supported" ), |
| 230 | from_pretty, to_pretty); |
| 231 | } |
| 232 | |
| 233 | argp_help (&argp, stderr, ARGP_HELP_SEE, |
| 234 | program_invocation_short_name); |
| 235 | exit (1); |
| 236 | } |
| 237 | else |
| 238 | error (EXIT_FAILURE, errno, |
| 239 | _("failed to start conversion processing" )); |
| 240 | } |
| 241 | |
| 242 | /* The output file. Will be opened when we are ready to produce |
| 243 | output. */ |
| 244 | FILE *output = NULL; |
| 245 | |
| 246 | /* Now process the remaining files. Write them to stdout or the file |
| 247 | specified with the `-o' parameter. If we have no file given as |
| 248 | the parameter process all from stdin. */ |
| 249 | if (remaining == argc) |
| 250 | { |
| 251 | if (process_file (cd, stdin, &output, output_file) != 0) |
| 252 | status = EXIT_FAILURE; |
| 253 | } |
| 254 | else |
| 255 | do |
| 256 | { |
| 257 | #ifdef _POSIX_MAPPED_FILES |
| 258 | struct stat64 st; |
| 259 | char *addr; |
| 260 | #endif |
| 261 | int fd, ret; |
| 262 | |
| 263 | if (verbose) |
| 264 | fprintf (stderr, "%s:\n" , argv[remaining]); |
| 265 | if (strcmp (argv[remaining], "-" ) == 0) |
| 266 | fd = 0; |
| 267 | else |
| 268 | { |
| 269 | fd = open (argv[remaining], O_RDONLY); |
| 270 | |
| 271 | if (fd == -1) |
| 272 | { |
| 273 | error (0, errno, _("cannot open input file `%s'" ), |
| 274 | argv[remaining]); |
| 275 | status = EXIT_FAILURE; |
| 276 | continue; |
| 277 | } |
| 278 | } |
| 279 | |
| 280 | #ifdef _POSIX_MAPPED_FILES |
| 281 | /* We have possibilities for reading the input file. First try |
| 282 | to mmap() it since this will provide the fastest solution. */ |
| 283 | if (fstat64 (fd, &st) == 0 |
| 284 | && ((addr = mmap (NULL, st.st_size, PROT_READ, MAP_PRIVATE, |
| 285 | fd, 0)) != MAP_FAILED)) |
| 286 | { |
| 287 | /* Yes, we can use mmap(). The descriptor is not needed |
| 288 | anymore. */ |
| 289 | if (close (fd) != 0) |
| 290 | error (EXIT_FAILURE, errno, |
| 291 | _("error while closing input `%s'" ), |
| 292 | argv[remaining]); |
| 293 | |
| 294 | ret = process_block (cd, addr, st.st_size, &output, |
| 295 | output_file); |
| 296 | |
| 297 | /* We don't need the input data anymore. */ |
| 298 | munmap ((void *) addr, st.st_size); |
| 299 | |
| 300 | if (ret != 0) |
| 301 | { |
| 302 | status = EXIT_FAILURE; |
| 303 | |
| 304 | if (ret < 0) |
| 305 | /* We cannot go on with producing output since it might |
| 306 | lead to problem because the last output might leave |
| 307 | the output stream in an undefined state. */ |
| 308 | break; |
| 309 | } |
| 310 | } |
| 311 | else |
| 312 | #endif /* _POSIX_MAPPED_FILES */ |
| 313 | { |
| 314 | /* Read the file in pieces. */ |
| 315 | ret = process_fd (cd, fd, &output, output_file); |
| 316 | |
| 317 | /* Now close the file. */ |
| 318 | close (fd); |
| 319 | |
| 320 | if (ret != 0) |
| 321 | { |
| 322 | /* Something went wrong. */ |
| 323 | status = EXIT_FAILURE; |
| 324 | |
| 325 | if (ret < 0) |
| 326 | /* We cannot go on with producing output since it might |
| 327 | lead to problem because the last output might leave |
| 328 | the output stream in an undefined state. */ |
| 329 | break; |
| 330 | } |
| 331 | } |
| 332 | } |
| 333 | while (++remaining < argc); |
| 334 | |
| 335 | /* Close the output file now. */ |
| 336 | if (output != NULL && fclose (output)) |
| 337 | error (EXIT_FAILURE, errno, _("error while closing output file" )); |
| 338 | } |
| 339 | |
| 340 | return status; |
| 341 | } |
| 342 | |
| 343 | |
| 344 | /* Handle program arguments. */ |
| 345 | static error_t |
| 346 | parse_opt (int key, char *arg, struct argp_state *state) |
| 347 | { |
| 348 | switch (key) |
| 349 | { |
| 350 | case 'f': |
| 351 | from_code = arg; |
| 352 | break; |
| 353 | case 't': |
| 354 | to_code = arg; |
| 355 | break; |
| 356 | case 'o': |
| 357 | output_file = arg; |
| 358 | break; |
| 359 | case 's': |
| 360 | /* Nothing, for now at least. We are not giving out any information |
| 361 | about missing character or so. */ |
| 362 | break; |
| 363 | case 'c': |
| 364 | /* Omit invalid characters from output. */ |
| 365 | omit_invalid = 1; |
| 366 | break; |
| 367 | case OPT_VERBOSE: |
| 368 | verbose = 1; |
| 369 | break; |
| 370 | case OPT_LIST: |
| 371 | list = 1; |
| 372 | break; |
| 373 | default: |
| 374 | return ARGP_ERR_UNKNOWN; |
| 375 | } |
| 376 | return 0; |
| 377 | } |
| 378 | |
| 379 | |
| 380 | static char * |
| 381 | more_help (int key, const char *text, void *input) |
| 382 | { |
| 383 | char *tp = NULL; |
| 384 | switch (key) |
| 385 | { |
| 386 | case ARGP_KEY_HELP_EXTRA: |
| 387 | /* We print some extra information. */ |
| 388 | if (asprintf (&tp, gettext ("\ |
| 389 | For bug reporting instructions, please see:\n\ |
| 390 | %s.\n" ), REPORT_BUGS_TO) < 0) |
| 391 | return NULL; |
| 392 | return tp; |
| 393 | default: |
| 394 | break; |
| 395 | } |
| 396 | return (char *) text; |
| 397 | } |
| 398 | |
| 399 | |
| 400 | /* Print the version information. */ |
| 401 | static void |
| 402 | print_version (FILE *stream, struct argp_state *state) |
| 403 | { |
| 404 | fprintf (stream, "iconv %s%s\n" , PKGVERSION, VERSION); |
| 405 | fprintf (stream, gettext ("\ |
| 406 | Copyright (C) %s Free Software Foundation, Inc.\n\ |
| 407 | This is free software; see the source for copying conditions. There is NO\n\ |
| 408 | warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\ |
| 409 | " ), "2021" ); |
| 410 | fprintf (stream, gettext ("Written by %s.\n" ), "Ulrich Drepper" ); |
| 411 | } |
| 412 | |
| 413 | |
| 414 | static int |
| 415 | write_output (const char *outbuf, const char *outptr, FILE **output, |
| 416 | const char *output_file) |
| 417 | { |
| 418 | /* We have something to write out. */ |
| 419 | int errno_save = errno; |
| 420 | |
| 421 | if (*output == NULL) |
| 422 | { |
| 423 | /* Determine output file. */ |
| 424 | if (output_file != NULL && strcmp (output_file, "-" ) != 0) |
| 425 | { |
| 426 | *output = fopen (output_file, "w" ); |
| 427 | if (*output == NULL) |
| 428 | error (EXIT_FAILURE, errno, _("cannot open output file" )); |
| 429 | } |
| 430 | else |
| 431 | *output = stdout; |
| 432 | } |
| 433 | |
| 434 | if (fwrite (outbuf, 1, outptr - outbuf, *output) < (size_t) (outptr - outbuf) |
| 435 | || ferror (*output)) |
| 436 | { |
| 437 | /* Error occurred while printing the result. */ |
| 438 | error (0, 0, _("\ |
| 439 | conversion stopped due to problem in writing the output" )); |
| 440 | return -1; |
| 441 | } |
| 442 | |
| 443 | errno = errno_save; |
| 444 | |
| 445 | return 0; |
| 446 | } |
| 447 | |
| 448 | |
| 449 | static int |
| 450 | process_block (iconv_t cd, char *addr, size_t len, FILE **output, |
| 451 | const char *output_file) |
| 452 | { |
| 453 | #define OUTBUF_SIZE 32768 |
| 454 | const char *start = addr; |
| 455 | char outbuf[OUTBUF_SIZE]; |
| 456 | char *outptr; |
| 457 | size_t outlen; |
| 458 | size_t n; |
| 459 | int ret = 0; |
| 460 | |
| 461 | while (len > 0) |
| 462 | { |
| 463 | outptr = outbuf; |
| 464 | outlen = OUTBUF_SIZE; |
| 465 | n = iconv (cd, &addr, &len, &outptr, &outlen); |
| 466 | |
| 467 | if (n == (size_t) -1 && omit_invalid && errno == EILSEQ) |
| 468 | { |
| 469 | ret = 1; |
| 470 | if (len == 0) |
| 471 | n = 0; |
| 472 | else |
| 473 | errno = E2BIG; |
| 474 | } |
| 475 | |
| 476 | if (outptr != outbuf) |
| 477 | { |
| 478 | ret = write_output (outbuf, outptr, output, output_file); |
| 479 | if (ret != 0) |
| 480 | break; |
| 481 | } |
| 482 | |
| 483 | if (n != (size_t) -1) |
| 484 | { |
| 485 | /* All the input test is processed. For state-dependent |
| 486 | character sets we have to flush the state now. */ |
| 487 | outptr = outbuf; |
| 488 | outlen = OUTBUF_SIZE; |
| 489 | n = iconv (cd, NULL, NULL, &outptr, &outlen); |
| 490 | |
| 491 | if (outptr != outbuf) |
| 492 | { |
| 493 | ret = write_output (outbuf, outptr, output, output_file); |
| 494 | if (ret != 0) |
| 495 | break; |
| 496 | } |
| 497 | |
| 498 | if (n != (size_t) -1) |
| 499 | break; |
| 500 | |
| 501 | if (omit_invalid && errno == EILSEQ) |
| 502 | { |
| 503 | ret = 1; |
| 504 | break; |
| 505 | } |
| 506 | } |
| 507 | |
| 508 | if (errno != E2BIG) |
| 509 | { |
| 510 | /* iconv() ran into a problem. */ |
| 511 | switch (errno) |
| 512 | { |
| 513 | case EILSEQ: |
| 514 | if (! omit_invalid) |
| 515 | error (0, 0, _("illegal input sequence at position %ld" ), |
| 516 | (long int) (addr - start)); |
| 517 | break; |
| 518 | case EINVAL: |
| 519 | error (0, 0, _("\ |
| 520 | incomplete character or shift sequence at end of buffer" )); |
| 521 | break; |
| 522 | case EBADF: |
| 523 | error (0, 0, _("internal error (illegal descriptor)" )); |
| 524 | break; |
| 525 | default: |
| 526 | error (0, 0, _("unknown iconv() error %d" ), errno); |
| 527 | break; |
| 528 | } |
| 529 | |
| 530 | return -1; |
| 531 | } |
| 532 | } |
| 533 | |
| 534 | return ret; |
| 535 | } |
| 536 | |
| 537 | |
| 538 | static int |
| 539 | process_fd (iconv_t cd, int fd, FILE **output, const char *output_file) |
| 540 | { |
| 541 | /* we have a problem with reading from a desriptor since we must not |
| 542 | provide the iconv() function an incomplete character or shift |
| 543 | sequence at the end of the buffer. Since we have to deal with |
| 544 | arbitrary encodings we must read the whole text in a buffer and |
| 545 | process it in one step. */ |
| 546 | static char *inbuf = NULL; |
| 547 | static size_t maxlen = 0; |
| 548 | char *inptr = NULL; |
| 549 | size_t actlen = 0; |
| 550 | |
| 551 | while (actlen < maxlen) |
| 552 | { |
| 553 | ssize_t n = read (fd, inptr, maxlen - actlen); |
| 554 | |
| 555 | if (n == 0) |
| 556 | /* No more text to read. */ |
| 557 | break; |
| 558 | |
| 559 | if (n == -1) |
| 560 | { |
| 561 | /* Error while reading. */ |
| 562 | error (0, errno, _("error while reading the input" )); |
| 563 | return -1; |
| 564 | } |
| 565 | |
| 566 | inptr += n; |
| 567 | actlen += n; |
| 568 | } |
| 569 | |
| 570 | if (actlen == maxlen) |
| 571 | while (1) |
| 572 | { |
| 573 | ssize_t n; |
| 574 | char *new_inbuf; |
| 575 | |
| 576 | /* Increase the buffer. */ |
| 577 | new_inbuf = (char *) realloc (inbuf, maxlen + 32768); |
| 578 | if (new_inbuf == NULL) |
| 579 | { |
| 580 | error (0, errno, _("unable to allocate buffer for input" )); |
| 581 | return -1; |
| 582 | } |
| 583 | inbuf = new_inbuf; |
| 584 | maxlen += 32768; |
| 585 | inptr = inbuf + actlen; |
| 586 | |
| 587 | do |
| 588 | { |
| 589 | n = read (fd, inptr, maxlen - actlen); |
| 590 | |
| 591 | if (n == 0) |
| 592 | /* No more text to read. */ |
| 593 | break; |
| 594 | |
| 595 | if (n == -1) |
| 596 | { |
| 597 | /* Error while reading. */ |
| 598 | error (0, errno, _("error while reading the input" )); |
| 599 | return -1; |
| 600 | } |
| 601 | |
| 602 | inptr += n; |
| 603 | actlen += n; |
| 604 | } |
| 605 | while (actlen < maxlen); |
| 606 | |
| 607 | if (n == 0) |
| 608 | /* Break again so we leave both loops. */ |
| 609 | break; |
| 610 | } |
| 611 | |
| 612 | /* Now we have all the input in the buffer. Process it in one run. */ |
| 613 | return process_block (cd, inbuf, actlen, output, output_file); |
| 614 | } |
| 615 | |
| 616 | |
| 617 | static int |
| 618 | process_file (iconv_t cd, FILE *input, FILE **output, const char *output_file) |
| 619 | { |
| 620 | /* This should be safe since we use this function only for `stdin' and |
| 621 | we haven't read anything so far. */ |
| 622 | return process_fd (cd, fileno (input), output, output_file); |
| 623 | } |
| 624 | |
| 625 | |
| 626 | /* Print all known character sets/encodings. */ |
| 627 | static void *printlist; |
| 628 | static size_t column; |
| 629 | static int not_first; |
| 630 | |
| 631 | static void |
| 632 | insert_print_list (const void *nodep, VISIT value, int level) |
| 633 | { |
| 634 | if (value == leaf || value == postorder) |
| 635 | { |
| 636 | const struct gconv_alias *s = *(const struct gconv_alias **) nodep; |
| 637 | tsearch (s->fromname, &printlist, (__compar_fn_t) strverscmp); |
| 638 | } |
| 639 | } |
| 640 | |
| 641 | static void |
| 642 | do_print_human (const void *nodep, VISIT value, int level) |
| 643 | { |
| 644 | if (value == leaf || value == postorder) |
| 645 | { |
| 646 | const char *s = *(const char **) nodep; |
| 647 | size_t len = strlen (s); |
| 648 | size_t cnt; |
| 649 | |
| 650 | while (len > 0 && s[len - 1] == '/') |
| 651 | --len; |
| 652 | |
| 653 | for (cnt = 0; cnt < len; ++cnt) |
| 654 | if (isalnum (s[cnt])) |
| 655 | break; |
| 656 | if (cnt == len) |
| 657 | return; |
| 658 | |
| 659 | if (not_first) |
| 660 | { |
| 661 | putchar (','); |
| 662 | ++column; |
| 663 | |
| 664 | if (column > 2 && column + len > 77) |
| 665 | { |
| 666 | fputs ("\n " , stdout); |
| 667 | column = 2; |
| 668 | } |
| 669 | else |
| 670 | { |
| 671 | putchar (' '); |
| 672 | ++column; |
| 673 | } |
| 674 | } |
| 675 | else |
| 676 | not_first = 1; |
| 677 | |
| 678 | fwrite (s, len, 1, stdout); |
| 679 | column += len; |
| 680 | } |
| 681 | } |
| 682 | |
| 683 | static void |
| 684 | do_print (const void *nodep, VISIT value, int level) |
| 685 | { |
| 686 | if (value == leaf || value == postorder) |
| 687 | { |
| 688 | const char *s = *(const char **) nodep; |
| 689 | |
| 690 | puts (s); |
| 691 | } |
| 692 | } |
| 693 | |
| 694 | static void |
| 695 | add_known_names (struct gconv_module *node) |
| 696 | { |
| 697 | if (node->left != NULL) |
| 698 | add_known_names (node->left); |
| 699 | if (node->right != NULL) |
| 700 | add_known_names (node->right); |
| 701 | do |
| 702 | { |
| 703 | if (strcmp (node->from_string, "INTERNAL" ) != 0) |
| 704 | tsearch (node->from_string, &printlist, (__compar_fn_t) strverscmp); |
| 705 | if (strcmp (node->to_string, "INTERNAL" ) != 0) |
| 706 | tsearch (node->to_string, &printlist, (__compar_fn_t) strverscmp); |
| 707 | |
| 708 | node = node->same; |
| 709 | } |
| 710 | while (node != NULL); |
| 711 | } |
| 712 | |
| 713 | |
| 714 | static void |
| 715 | insert_cache (void) |
| 716 | { |
| 717 | const struct gconvcache_header *; |
| 718 | const char *strtab; |
| 719 | const struct hash_entry *hashtab; |
| 720 | size_t cnt; |
| 721 | |
| 722 | header = (const struct gconvcache_header *) __gconv_get_cache (); |
| 723 | strtab = (char *) header + header->string_offset; |
| 724 | hashtab = (struct hash_entry *) ((char *) header + header->hash_offset); |
| 725 | |
| 726 | for (cnt = 0; cnt < header->hash_size; ++cnt) |
| 727 | if (hashtab[cnt].string_offset != 0) |
| 728 | { |
| 729 | const char *str = strtab + hashtab[cnt].string_offset; |
| 730 | |
| 731 | if (strcmp (str, "INTERNAL" ) != 0) |
| 732 | tsearch (str, &printlist, (__compar_fn_t) strverscmp); |
| 733 | } |
| 734 | } |
| 735 | |
| 736 | |
| 737 | static void |
| 738 | print_known_names (void) |
| 739 | { |
| 740 | iconv_t h; |
| 741 | void *cache; |
| 742 | |
| 743 | /* We must initialize the internal databases first. */ |
| 744 | h = iconv_open ("L1" , "L1" ); |
| 745 | iconv_close (h); |
| 746 | |
| 747 | /* See whether we have a cache. */ |
| 748 | cache = __gconv_get_cache (); |
| 749 | if (cache != NULL) |
| 750 | /* Yep, use only this information. */ |
| 751 | insert_cache (); |
| 752 | else |
| 753 | { |
| 754 | struct gconv_module *modules; |
| 755 | |
| 756 | /* No, then use the information read from the gconv-modules file. |
| 757 | First add the aliases. */ |
| 758 | twalk (__gconv_get_alias_db (), insert_print_list); |
| 759 | |
| 760 | /* Add the from- and to-names from the known modules. */ |
| 761 | modules = __gconv_get_modules_db (); |
| 762 | if (modules != NULL) |
| 763 | add_known_names (modules); |
| 764 | } |
| 765 | |
| 766 | bool human_readable = isatty (fileno (stdout)); |
| 767 | |
| 768 | if (human_readable) |
| 769 | fputs (_("\ |
| 770 | The following list contains all the coded character sets known. This does\n\ |
| 771 | not necessarily mean that all combinations of these names can be used for\n\ |
| 772 | the FROM and TO command line parameters. One coded character set can be\n\ |
| 773 | listed with several different names (aliases).\n\n " ), stdout); |
| 774 | |
| 775 | /* Now print the collected names. */ |
| 776 | column = 2; |
| 777 | twalk (printlist, human_readable ? do_print_human : do_print); |
| 778 | |
| 779 | if (human_readable && column != 0) |
| 780 | puts ("" ); |
| 781 | } |
| 782 | |