1 | /* Convert text in given files from the specified from-set to the to-set. |
2 | Copyright (C) 1998-2018 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 <http://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 | |
43 | /* Get libc version number. */ |
44 | #include "../version.h" |
45 | |
46 | #define PACKAGE _libc_intl_domainname |
47 | |
48 | |
49 | /* Name and version of program. */ |
50 | static void print_version (FILE *stream, struct argp_state *state); |
51 | void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version; |
52 | |
53 | #define OPT_VERBOSE 1000 |
54 | #define OPT_LIST 'l' |
55 | |
56 | /* Definitions of arguments for argp functions. */ |
57 | static const struct argp_option options[] = |
58 | { |
59 | { NULL, 0, NULL, 0, N_("Input/Output format specification:" ) }, |
60 | { "from-code" , 'f', N_("NAME" ), 0, N_("encoding of original text" ) }, |
61 | { "to-code" , 't', N_("NAME" ), 0, N_("encoding for output" ) }, |
62 | { NULL, 0, NULL, 0, N_("Information:" ) }, |
63 | { "list" , 'l', NULL, 0, N_("list all known coded character sets" ) }, |
64 | { NULL, 0, NULL, 0, N_("Output control:" ) }, |
65 | { NULL, 'c', NULL, 0, N_("omit invalid characters from output" ) }, |
66 | { "output" , 'o', N_("FILE" ), 0, N_("output file" ) }, |
67 | { "silent" , 's', NULL, 0, N_("suppress warnings" ) }, |
68 | { "verbose" , OPT_VERBOSE, NULL, 0, N_("print progress information" ) }, |
69 | { NULL, 0, NULL, 0, NULL } |
70 | }; |
71 | |
72 | /* Short description of program. */ |
73 | static const char doc[] = N_("\ |
74 | Convert encoding of given files from one encoding to another." ); |
75 | |
76 | /* Strings for arguments in help texts. */ |
77 | static const char args_doc[] = N_("[FILE...]" ); |
78 | |
79 | /* Prototype for option handler. */ |
80 | static error_t parse_opt (int key, char *arg, struct argp_state *state); |
81 | |
82 | /* Function to print some extra text in the help message. */ |
83 | static char *more_help (int key, const char *text, void *input); |
84 | |
85 | /* Data structure to communicate with argp functions. */ |
86 | static struct argp argp = |
87 | { |
88 | options, parse_opt, args_doc, doc, NULL, more_help |
89 | }; |
90 | |
91 | /* Code sets to convert from and to respectively. An empty string as the |
92 | default causes the 'iconv_open' function to look up the charset of the |
93 | currently selected locale and use it. */ |
94 | static const char *from_code = "" ; |
95 | static const char *to_code = "" ; |
96 | |
97 | /* File to write output to. If NULL write to stdout. */ |
98 | static const char *output_file; |
99 | |
100 | /* Nonzero if list of all coded character sets is wanted. */ |
101 | static int list; |
102 | |
103 | /* If nonzero omit invalid character from output. */ |
104 | int omit_invalid; |
105 | |
106 | /* Prototypes for the functions doing the actual work. */ |
107 | static int process_block (iconv_t cd, char *addr, size_t len, FILE **output, |
108 | const char *output_file); |
109 | static int process_fd (iconv_t cd, int fd, FILE **output, |
110 | const char *output_file); |
111 | static int process_file (iconv_t cd, FILE *input, FILE **output, |
112 | const char *output_file); |
113 | static void print_known_names (void); |
114 | |
115 | |
116 | int |
117 | main (int argc, char *argv[]) |
118 | { |
119 | int status = EXIT_SUCCESS; |
120 | int remaining; |
121 | iconv_t cd; |
122 | const char *orig_to_code; |
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 | /* If we have to ignore errors make sure we use the appropriate name for |
143 | the to-character-set. */ |
144 | orig_to_code = to_code; |
145 | if (omit_invalid) |
146 | { |
147 | const char *errhand = strchrnul (to_code, '/'); |
148 | int nslash = 2; |
149 | char *newp; |
150 | char *cp; |
151 | |
152 | if (*errhand == '/') |
153 | { |
154 | --nslash; |
155 | errhand = strchrnul (errhand + 1, '/'); |
156 | |
157 | if (*errhand == '/') |
158 | { |
159 | --nslash; |
160 | errhand = strchr (errhand, '\0'); |
161 | } |
162 | } |
163 | |
164 | newp = (char *) alloca (errhand - to_code + nslash + 7 + 1); |
165 | cp = mempcpy (newp, to_code, errhand - to_code); |
166 | while (nslash-- > 0) |
167 | *cp++ = '/'; |
168 | if (cp[-1] != '/') |
169 | *cp++ = ','; |
170 | memcpy (cp, "IGNORE" , sizeof ("IGNORE" )); |
171 | |
172 | to_code = newp; |
173 | } |
174 | |
175 | /* POSIX 1003.2b introduces a silly thing: the arguments to -t anf -f |
176 | can be file names of charmaps. In this case iconv will have to read |
177 | those charmaps and use them to do the conversion. But there are |
178 | holes in the specification. There is nothing said that if -f is a |
179 | charmap filename that -t must be, too. And vice versa. There is |
180 | also no word about the symbolic names used. What if they don't |
181 | match? */ |
182 | if (strchr (from_code, '/') != NULL) |
183 | /* The from-name might be a charmap file name. Try reading the |
184 | file. */ |
185 | from_charmap = charmap_read (from_code, /*0, 1*/1, 0, 0, 0); |
186 | |
187 | if (strchr (orig_to_code, '/') != NULL) |
188 | /* The to-name might be a charmap file name. Try reading the |
189 | file. */ |
190 | to_charmap = charmap_read (orig_to_code, /*0, 1,*/1, 0, 0, 0); |
191 | |
192 | |
193 | /* At this point we have to handle two cases. The first one is |
194 | where a charmap is used for the from- or to-charset, or both. We |
195 | handle this special since it is very different from the sane way of |
196 | doing things. The other case allows converting using the iconv() |
197 | function. */ |
198 | if (from_charmap != NULL || to_charmap != NULL) |
199 | /* Construct the conversion table and do the conversion. */ |
200 | status = charmap_conversion (from_code, from_charmap, to_code, to_charmap, |
201 | argc, remaining, argv, output_file); |
202 | else |
203 | { |
204 | /* Let's see whether we have these coded character sets. */ |
205 | cd = iconv_open (to_code, from_code); |
206 | if (cd == (iconv_t) -1) |
207 | { |
208 | if (errno == EINVAL) |
209 | { |
210 | /* Try to be nice with the user and tell her which of the |
211 | two encoding names is wrong. This is possible because |
212 | all supported encodings can be converted from/to Unicode, |
213 | in other words, because the graph of encodings is |
214 | connected. */ |
215 | bool from_wrong = |
216 | (iconv_open ("UTF-8" , from_code) == (iconv_t) -1 |
217 | && errno == EINVAL); |
218 | bool to_wrong = |
219 | (iconv_open (to_code, "UTF-8" ) == (iconv_t) -1 |
220 | && errno == EINVAL); |
221 | const char *from_pretty = |
222 | (from_code[0] ? from_code : nl_langinfo (CODESET)); |
223 | const char *to_pretty = |
224 | (orig_to_code[0] ? orig_to_code : nl_langinfo (CODESET)); |
225 | |
226 | if (from_wrong) |
227 | { |
228 | if (to_wrong) |
229 | error (0, 0, |
230 | _("\ |
231 | conversions from `%s' and to `%s' are not supported" ), |
232 | from_pretty, to_pretty); |
233 | else |
234 | error (0, 0, |
235 | _("conversion from `%s' is not supported" ), |
236 | from_pretty); |
237 | } |
238 | else |
239 | { |
240 | if (to_wrong) |
241 | error (0, 0, |
242 | _("conversion to `%s' is not supported" ), |
243 | to_pretty); |
244 | else |
245 | error (0, 0, |
246 | _("conversion from `%s' to `%s' is not supported" ), |
247 | from_pretty, to_pretty); |
248 | } |
249 | |
250 | argp_help (&argp, stderr, ARGP_HELP_SEE, |
251 | program_invocation_short_name); |
252 | exit (1); |
253 | } |
254 | else |
255 | error (EXIT_FAILURE, errno, |
256 | _("failed to start conversion processing" )); |
257 | } |
258 | |
259 | /* The output file. Will be opened when we are ready to produce |
260 | output. */ |
261 | FILE *output = NULL; |
262 | |
263 | /* Now process the remaining files. Write them to stdout or the file |
264 | specified with the `-o' parameter. If we have no file given as |
265 | the parameter process all from stdin. */ |
266 | if (remaining == argc) |
267 | { |
268 | if (process_file (cd, stdin, &output, output_file) != 0) |
269 | status = EXIT_FAILURE; |
270 | } |
271 | else |
272 | do |
273 | { |
274 | #ifdef _POSIX_MAPPED_FILES |
275 | struct stat64 st; |
276 | char *addr; |
277 | #endif |
278 | int fd, ret; |
279 | |
280 | if (verbose) |
281 | fprintf (stderr, "%s:\n" , argv[remaining]); |
282 | if (strcmp (argv[remaining], "-" ) == 0) |
283 | fd = 0; |
284 | else |
285 | { |
286 | fd = open (argv[remaining], O_RDONLY); |
287 | |
288 | if (fd == -1) |
289 | { |
290 | error (0, errno, _("cannot open input file `%s'" ), |
291 | argv[remaining]); |
292 | status = EXIT_FAILURE; |
293 | continue; |
294 | } |
295 | } |
296 | |
297 | #ifdef _POSIX_MAPPED_FILES |
298 | /* We have possibilities for reading the input file. First try |
299 | to mmap() it since this will provide the fastest solution. */ |
300 | if (fstat64 (fd, &st) == 0 |
301 | && ((addr = mmap (NULL, st.st_size, PROT_READ, MAP_PRIVATE, |
302 | fd, 0)) != MAP_FAILED)) |
303 | { |
304 | /* Yes, we can use mmap(). The descriptor is not needed |
305 | anymore. */ |
306 | if (close (fd) != 0) |
307 | error (EXIT_FAILURE, errno, |
308 | _("error while closing input `%s'" ), |
309 | argv[remaining]); |
310 | |
311 | ret = process_block (cd, addr, st.st_size, &output, |
312 | output_file); |
313 | |
314 | /* We don't need the input data anymore. */ |
315 | munmap ((void *) addr, st.st_size); |
316 | |
317 | if (ret != 0) |
318 | { |
319 | status = EXIT_FAILURE; |
320 | |
321 | if (ret < 0) |
322 | /* We cannot go on with producing output since it might |
323 | lead to problem because the last output might leave |
324 | the output stream in an undefined state. */ |
325 | break; |
326 | } |
327 | } |
328 | else |
329 | #endif /* _POSIX_MAPPED_FILES */ |
330 | { |
331 | /* Read the file in pieces. */ |
332 | ret = process_fd (cd, fd, &output, output_file); |
333 | |
334 | /* Now close the file. */ |
335 | close (fd); |
336 | |
337 | if (ret != 0) |
338 | { |
339 | /* Something went wrong. */ |
340 | status = EXIT_FAILURE; |
341 | |
342 | if (ret < 0) |
343 | /* We cannot go on with producing output since it might |
344 | lead to problem because the last output might leave |
345 | the output stream in an undefined state. */ |
346 | break; |
347 | } |
348 | } |
349 | } |
350 | while (++remaining < argc); |
351 | |
352 | /* Close the output file now. */ |
353 | if (output != NULL && fclose (output)) |
354 | error (EXIT_FAILURE, errno, _("error while closing output file" )); |
355 | } |
356 | |
357 | return status; |
358 | } |
359 | |
360 | |
361 | /* Handle program arguments. */ |
362 | static error_t |
363 | parse_opt (int key, char *arg, struct argp_state *state) |
364 | { |
365 | switch (key) |
366 | { |
367 | case 'f': |
368 | from_code = arg; |
369 | break; |
370 | case 't': |
371 | to_code = arg; |
372 | break; |
373 | case 'o': |
374 | output_file = arg; |
375 | break; |
376 | case 's': |
377 | /* Nothing, for now at least. We are not giving out any information |
378 | about missing character or so. */ |
379 | break; |
380 | case 'c': |
381 | /* Omit invalid characters from output. */ |
382 | omit_invalid = 1; |
383 | break; |
384 | case OPT_VERBOSE: |
385 | verbose = 1; |
386 | break; |
387 | case OPT_LIST: |
388 | list = 1; |
389 | break; |
390 | default: |
391 | return ARGP_ERR_UNKNOWN; |
392 | } |
393 | return 0; |
394 | } |
395 | |
396 | |
397 | static char * |
398 | more_help (int key, const char *text, void *input) |
399 | { |
400 | char *tp = NULL; |
401 | switch (key) |
402 | { |
403 | case ARGP_KEY_HELP_EXTRA: |
404 | /* We print some extra information. */ |
405 | if (asprintf (&tp, gettext ("\ |
406 | For bug reporting instructions, please see:\n\ |
407 | %s.\n" ), REPORT_BUGS_TO) < 0) |
408 | return NULL; |
409 | return tp; |
410 | default: |
411 | break; |
412 | } |
413 | return (char *) text; |
414 | } |
415 | |
416 | |
417 | /* Print the version information. */ |
418 | static void |
419 | print_version (FILE *stream, struct argp_state *state) |
420 | { |
421 | fprintf (stream, "iconv %s%s\n" , PKGVERSION, VERSION); |
422 | fprintf (stream, gettext ("\ |
423 | Copyright (C) %s Free Software Foundation, Inc.\n\ |
424 | This is free software; see the source for copying conditions. There is NO\n\ |
425 | warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\ |
426 | " ), "2018" ); |
427 | fprintf (stream, gettext ("Written by %s.\n" ), "Ulrich Drepper" ); |
428 | } |
429 | |
430 | |
431 | static int |
432 | write_output (const char *outbuf, const char *outptr, FILE **output, |
433 | const char *output_file) |
434 | { |
435 | /* We have something to write out. */ |
436 | int errno_save = errno; |
437 | |
438 | if (*output == NULL) |
439 | { |
440 | /* Determine output file. */ |
441 | if (output_file != NULL && strcmp (output_file, "-" ) != 0) |
442 | { |
443 | *output = fopen (output_file, "w" ); |
444 | if (*output == NULL) |
445 | error (EXIT_FAILURE, errno, _("cannot open output file" )); |
446 | } |
447 | else |
448 | *output = stdout; |
449 | } |
450 | |
451 | if (fwrite (outbuf, 1, outptr - outbuf, *output) < (size_t) (outptr - outbuf) |
452 | || ferror (*output)) |
453 | { |
454 | /* Error occurred while printing the result. */ |
455 | error (0, 0, _("\ |
456 | conversion stopped due to problem in writing the output" )); |
457 | return -1; |
458 | } |
459 | |
460 | errno = errno_save; |
461 | |
462 | return 0; |
463 | } |
464 | |
465 | |
466 | static int |
467 | process_block (iconv_t cd, char *addr, size_t len, FILE **output, |
468 | const char *output_file) |
469 | { |
470 | #define OUTBUF_SIZE 32768 |
471 | const char *start = addr; |
472 | char outbuf[OUTBUF_SIZE]; |
473 | char *outptr; |
474 | size_t outlen; |
475 | size_t n; |
476 | int ret = 0; |
477 | |
478 | while (len > 0) |
479 | { |
480 | outptr = outbuf; |
481 | outlen = OUTBUF_SIZE; |
482 | n = iconv (cd, &addr, &len, &outptr, &outlen); |
483 | |
484 | if (n == (size_t) -1 && omit_invalid && errno == EILSEQ) |
485 | { |
486 | ret = 1; |
487 | if (len == 0) |
488 | n = 0; |
489 | else |
490 | errno = E2BIG; |
491 | } |
492 | |
493 | if (outptr != outbuf) |
494 | { |
495 | ret = write_output (outbuf, outptr, output, output_file); |
496 | if (ret != 0) |
497 | break; |
498 | } |
499 | |
500 | if (n != (size_t) -1) |
501 | { |
502 | /* All the input test is processed. For state-dependent |
503 | character sets we have to flush the state now. */ |
504 | outptr = outbuf; |
505 | outlen = OUTBUF_SIZE; |
506 | n = iconv (cd, NULL, NULL, &outptr, &outlen); |
507 | |
508 | if (outptr != outbuf) |
509 | { |
510 | ret = write_output (outbuf, outptr, output, output_file); |
511 | if (ret != 0) |
512 | break; |
513 | } |
514 | |
515 | if (n != (size_t) -1) |
516 | break; |
517 | |
518 | if (omit_invalid && errno == EILSEQ) |
519 | { |
520 | ret = 1; |
521 | break; |
522 | } |
523 | } |
524 | |
525 | if (errno != E2BIG) |
526 | { |
527 | /* iconv() ran into a problem. */ |
528 | switch (errno) |
529 | { |
530 | case EILSEQ: |
531 | if (! omit_invalid) |
532 | error (0, 0, _("illegal input sequence at position %ld" ), |
533 | (long int) (addr - start)); |
534 | break; |
535 | case EINVAL: |
536 | error (0, 0, _("\ |
537 | incomplete character or shift sequence at end of buffer" )); |
538 | break; |
539 | case EBADF: |
540 | error (0, 0, _("internal error (illegal descriptor)" )); |
541 | break; |
542 | default: |
543 | error (0, 0, _("unknown iconv() error %d" ), errno); |
544 | break; |
545 | } |
546 | |
547 | return -1; |
548 | } |
549 | } |
550 | |
551 | return ret; |
552 | } |
553 | |
554 | |
555 | static int |
556 | process_fd (iconv_t cd, int fd, FILE **output, const char *output_file) |
557 | { |
558 | /* we have a problem with reading from a desriptor since we must not |
559 | provide the iconv() function an incomplete character or shift |
560 | sequence at the end of the buffer. Since we have to deal with |
561 | arbitrary encodings we must read the whole text in a buffer and |
562 | process it in one step. */ |
563 | static char *inbuf = NULL; |
564 | static size_t maxlen = 0; |
565 | char *inptr = NULL; |
566 | size_t actlen = 0; |
567 | |
568 | while (actlen < maxlen) |
569 | { |
570 | ssize_t n = read (fd, inptr, maxlen - actlen); |
571 | |
572 | if (n == 0) |
573 | /* No more text to read. */ |
574 | break; |
575 | |
576 | if (n == -1) |
577 | { |
578 | /* Error while reading. */ |
579 | error (0, errno, _("error while reading the input" )); |
580 | return -1; |
581 | } |
582 | |
583 | inptr += n; |
584 | actlen += n; |
585 | } |
586 | |
587 | if (actlen == maxlen) |
588 | while (1) |
589 | { |
590 | ssize_t n; |
591 | char *new_inbuf; |
592 | |
593 | /* Increase the buffer. */ |
594 | new_inbuf = (char *) realloc (inbuf, maxlen + 32768); |
595 | if (new_inbuf == NULL) |
596 | { |
597 | error (0, errno, _("unable to allocate buffer for input" )); |
598 | return -1; |
599 | } |
600 | inbuf = new_inbuf; |
601 | maxlen += 32768; |
602 | inptr = inbuf + actlen; |
603 | |
604 | do |
605 | { |
606 | n = read (fd, inptr, maxlen - actlen); |
607 | |
608 | if (n == 0) |
609 | /* No more text to read. */ |
610 | break; |
611 | |
612 | if (n == -1) |
613 | { |
614 | /* Error while reading. */ |
615 | error (0, errno, _("error while reading the input" )); |
616 | return -1; |
617 | } |
618 | |
619 | inptr += n; |
620 | actlen += n; |
621 | } |
622 | while (actlen < maxlen); |
623 | |
624 | if (n == 0) |
625 | /* Break again so we leave both loops. */ |
626 | break; |
627 | } |
628 | |
629 | /* Now we have all the input in the buffer. Process it in one run. */ |
630 | return process_block (cd, inbuf, actlen, output, output_file); |
631 | } |
632 | |
633 | |
634 | static int |
635 | process_file (iconv_t cd, FILE *input, FILE **output, const char *output_file) |
636 | { |
637 | /* This should be safe since we use this function only for `stdin' and |
638 | we haven't read anything so far. */ |
639 | return process_fd (cd, fileno (input), output, output_file); |
640 | } |
641 | |
642 | |
643 | /* Print all known character sets/encodings. */ |
644 | static void *printlist; |
645 | static size_t column; |
646 | static int not_first; |
647 | |
648 | static void |
649 | insert_print_list (const void *nodep, VISIT value, int level) |
650 | { |
651 | if (value == leaf || value == postorder) |
652 | { |
653 | const struct gconv_alias *s = *(const struct gconv_alias **) nodep; |
654 | tsearch (s->fromname, &printlist, (__compar_fn_t) strverscmp); |
655 | } |
656 | } |
657 | |
658 | static void |
659 | do_print_human (const void *nodep, VISIT value, int level) |
660 | { |
661 | if (value == leaf || value == postorder) |
662 | { |
663 | const char *s = *(const char **) nodep; |
664 | size_t len = strlen (s); |
665 | size_t cnt; |
666 | |
667 | while (len > 0 && s[len - 1] == '/') |
668 | --len; |
669 | |
670 | for (cnt = 0; cnt < len; ++cnt) |
671 | if (isalnum (s[cnt])) |
672 | break; |
673 | if (cnt == len) |
674 | return; |
675 | |
676 | if (not_first) |
677 | { |
678 | putchar (','); |
679 | ++column; |
680 | |
681 | if (column > 2 && column + len > 77) |
682 | { |
683 | fputs ("\n " , stdout); |
684 | column = 2; |
685 | } |
686 | else |
687 | { |
688 | putchar (' '); |
689 | ++column; |
690 | } |
691 | } |
692 | else |
693 | not_first = 1; |
694 | |
695 | fwrite (s, len, 1, stdout); |
696 | column += len; |
697 | } |
698 | } |
699 | |
700 | static void |
701 | do_print (const void *nodep, VISIT value, int level) |
702 | { |
703 | if (value == leaf || value == postorder) |
704 | { |
705 | const char *s = *(const char **) nodep; |
706 | |
707 | puts (s); |
708 | } |
709 | } |
710 | |
711 | static void |
712 | add_known_names (struct gconv_module *node) |
713 | { |
714 | if (node->left != NULL) |
715 | add_known_names (node->left); |
716 | if (node->right != NULL) |
717 | add_known_names (node->right); |
718 | do |
719 | { |
720 | if (strcmp (node->from_string, "INTERNAL" ) != 0) |
721 | tsearch (node->from_string, &printlist, (__compar_fn_t) strverscmp); |
722 | if (strcmp (node->to_string, "INTERNAL" ) != 0) |
723 | tsearch (node->to_string, &printlist, (__compar_fn_t) strverscmp); |
724 | |
725 | node = node->same; |
726 | } |
727 | while (node != NULL); |
728 | } |
729 | |
730 | |
731 | static void |
732 | insert_cache (void) |
733 | { |
734 | const struct gconvcache_header *; |
735 | const char *strtab; |
736 | const struct hash_entry *hashtab; |
737 | size_t cnt; |
738 | |
739 | header = (const struct gconvcache_header *) __gconv_get_cache (); |
740 | strtab = (char *) header + header->string_offset; |
741 | hashtab = (struct hash_entry *) ((char *) header + header->hash_offset); |
742 | |
743 | for (cnt = 0; cnt < header->hash_size; ++cnt) |
744 | if (hashtab[cnt].string_offset != 0) |
745 | { |
746 | const char *str = strtab + hashtab[cnt].string_offset; |
747 | |
748 | if (strcmp (str, "INTERNAL" ) != 0) |
749 | tsearch (str, &printlist, (__compar_fn_t) strverscmp); |
750 | } |
751 | } |
752 | |
753 | |
754 | static void |
755 | print_known_names (void) |
756 | { |
757 | iconv_t h; |
758 | void *cache; |
759 | |
760 | /* We must initialize the internal databases first. */ |
761 | h = iconv_open ("L1" , "L1" ); |
762 | iconv_close (h); |
763 | |
764 | /* See whether we have a cache. */ |
765 | cache = __gconv_get_cache (); |
766 | if (cache != NULL) |
767 | /* Yep, use only this information. */ |
768 | insert_cache (); |
769 | else |
770 | { |
771 | struct gconv_module *modules; |
772 | |
773 | /* No, then use the information read from the gconv-modules file. |
774 | First add the aliases. */ |
775 | twalk (__gconv_get_alias_db (), insert_print_list); |
776 | |
777 | /* Add the from- and to-names from the known modules. */ |
778 | modules = __gconv_get_modules_db (); |
779 | if (modules != NULL) |
780 | add_known_names (modules); |
781 | } |
782 | |
783 | bool human_readable = isatty (fileno (stdout)); |
784 | |
785 | if (human_readable) |
786 | fputs (_("\ |
787 | The following list contains all the coded character sets known. This does\n\ |
788 | not necessarily mean that all combinations of these names can be used for\n\ |
789 | the FROM and TO command line parameters. One coded character set can be\n\ |
790 | listed with several different names (aliases).\n\n " ), stdout); |
791 | |
792 | /* Now print the collected names. */ |
793 | column = 2; |
794 | twalk (printlist, human_readable ? do_print_human : do_print); |
795 | |
796 | if (human_readable && column != 0) |
797 | puts ("" ); |
798 | } |
799 | |