1 | /* Convert text in given files from the specified from-set to the to-set. |
2 | Copyright (C) 1998-2016 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 verbose ouput is wanted. */ |
101 | int verbose; |
102 | |
103 | /* Nonzero if list of all coded character sets is wanted. */ |
104 | static int list; |
105 | |
106 | /* If nonzero omit invalid character from output. */ |
107 | int omit_invalid; |
108 | |
109 | /* Prototypes for the functions doing the actual work. */ |
110 | static int process_block (iconv_t cd, char *addr, size_t len, FILE **output, |
111 | const char *output_file); |
112 | static int process_fd (iconv_t cd, int fd, FILE **output, |
113 | const char *output_file); |
114 | static int process_file (iconv_t cd, FILE *input, FILE **output, |
115 | const char *output_file); |
116 | static void print_known_names (void) internal_function; |
117 | |
118 | |
119 | int |
120 | main (int argc, char *argv[]) |
121 | { |
122 | int status = EXIT_SUCCESS; |
123 | int remaining; |
124 | iconv_t cd; |
125 | const char *orig_to_code; |
126 | struct charmap_t *from_charmap = NULL; |
127 | struct charmap_t *to_charmap = NULL; |
128 | |
129 | /* Set locale via LC_ALL. */ |
130 | setlocale (LC_ALL, "" ); |
131 | |
132 | /* Set the text message domain. */ |
133 | textdomain (_libc_intl_domainname); |
134 | |
135 | /* Parse and process arguments. */ |
136 | argp_parse (&argp, argc, argv, 0, &remaining, NULL); |
137 | |
138 | /* List all coded character sets if wanted. */ |
139 | if (list) |
140 | { |
141 | print_known_names (); |
142 | exit (EXIT_SUCCESS); |
143 | } |
144 | |
145 | /* If we have to ignore errors make sure we use the appropriate name for |
146 | the to-character-set. */ |
147 | orig_to_code = to_code; |
148 | if (omit_invalid) |
149 | { |
150 | const char *errhand = strchrnul (to_code, '/'); |
151 | int nslash = 2; |
152 | char *newp; |
153 | char *cp; |
154 | |
155 | if (*errhand == '/') |
156 | { |
157 | --nslash; |
158 | errhand = strchrnul (errhand + 1, '/'); |
159 | |
160 | if (*errhand == '/') |
161 | { |
162 | --nslash; |
163 | errhand = strchr (errhand, '\0'); |
164 | } |
165 | } |
166 | |
167 | newp = (char *) alloca (errhand - to_code + nslash + 7 + 1); |
168 | cp = mempcpy (newp, to_code, errhand - to_code); |
169 | while (nslash-- > 0) |
170 | *cp++ = '/'; |
171 | if (cp[-1] != '/') |
172 | *cp++ = ','; |
173 | memcpy (cp, "IGNORE" , sizeof ("IGNORE" )); |
174 | |
175 | to_code = newp; |
176 | } |
177 | |
178 | /* POSIX 1003.2b introduces a silly thing: the arguments to -t anf -f |
179 | can be file names of charmaps. In this case iconv will have to read |
180 | those charmaps and use them to do the conversion. But there are |
181 | holes in the specification. There is nothing said that if -f is a |
182 | charmap filename that -t must be, too. And vice versa. There is |
183 | also no word about the symbolic names used. What if they don't |
184 | match? */ |
185 | if (strchr (from_code, '/') != NULL) |
186 | /* The from-name might be a charmap file name. Try reading the |
187 | file. */ |
188 | from_charmap = charmap_read (from_code, /*0, 1*/1, 0, 0, 0); |
189 | |
190 | if (strchr (orig_to_code, '/') != NULL) |
191 | /* The to-name might be a charmap file name. Try reading the |
192 | file. */ |
193 | to_charmap = charmap_read (orig_to_code, /*0, 1,*/1, 0, 0, 0); |
194 | |
195 | |
196 | /* At this point we have to handle two cases. The first one is |
197 | where a charmap is used for the from- or to-charset, or both. We |
198 | handle this special since it is very different from the sane way of |
199 | doing things. The other case allows converting using the iconv() |
200 | function. */ |
201 | if (from_charmap != NULL || to_charmap != NULL) |
202 | /* Construct the conversion table and do the conversion. */ |
203 | status = charmap_conversion (from_code, from_charmap, to_code, to_charmap, |
204 | argc, remaining, argv, output_file); |
205 | else |
206 | { |
207 | /* Let's see whether we have these coded character sets. */ |
208 | cd = iconv_open (to_code, from_code); |
209 | if (cd == (iconv_t) -1) |
210 | { |
211 | if (errno == EINVAL) |
212 | { |
213 | /* Try to be nice with the user and tell her which of the |
214 | two encoding names is wrong. This is possible because |
215 | all supported encodings can be converted from/to Unicode, |
216 | in other words, because the graph of encodings is |
217 | connected. */ |
218 | bool from_wrong = |
219 | (iconv_open ("UTF-8" , from_code) == (iconv_t) -1 |
220 | && errno == EINVAL); |
221 | bool to_wrong = |
222 | (iconv_open (to_code, "UTF-8" ) == (iconv_t) -1 |
223 | && errno == EINVAL); |
224 | const char *from_pretty = |
225 | (from_code[0] ? from_code : nl_langinfo (CODESET)); |
226 | const char *to_pretty = |
227 | (orig_to_code[0] ? orig_to_code : nl_langinfo (CODESET)); |
228 | |
229 | if (from_wrong) |
230 | { |
231 | if (to_wrong) |
232 | error (0, 0, |
233 | _("\ |
234 | conversions from `%s' and to `%s' are not supported" ), |
235 | from_pretty, to_pretty); |
236 | else |
237 | error (0, 0, |
238 | _("conversion from `%s' is not supported" ), |
239 | from_pretty); |
240 | } |
241 | else |
242 | { |
243 | if (to_wrong) |
244 | error (0, 0, |
245 | _("conversion to `%s' is not supported" ), |
246 | to_pretty); |
247 | else |
248 | error (0, 0, |
249 | _("conversion from `%s' to `%s' is not supported" ), |
250 | from_pretty, to_pretty); |
251 | } |
252 | |
253 | argp_help (&argp, stderr, ARGP_HELP_SEE, |
254 | program_invocation_short_name); |
255 | exit (1); |
256 | } |
257 | else |
258 | error (EXIT_FAILURE, errno, |
259 | _("failed to start conversion processing" )); |
260 | } |
261 | |
262 | /* The output file. Will be opened when we are ready to produce |
263 | output. */ |
264 | FILE *output = NULL; |
265 | |
266 | /* Now process the remaining files. Write them to stdout or the file |
267 | specified with the `-o' parameter. If we have no file given as |
268 | the parameter process all from stdin. */ |
269 | if (remaining == argc) |
270 | { |
271 | if (process_file (cd, stdin, &output, output_file) != 0) |
272 | status = EXIT_FAILURE; |
273 | } |
274 | else |
275 | do |
276 | { |
277 | #ifdef _POSIX_MAPPED_FILES |
278 | struct stat64 st; |
279 | char *addr; |
280 | #endif |
281 | int fd, ret; |
282 | |
283 | if (verbose) |
284 | fprintf (stderr, "%s:\n" , argv[remaining]); |
285 | if (strcmp (argv[remaining], "-" ) == 0) |
286 | fd = 0; |
287 | else |
288 | { |
289 | fd = open (argv[remaining], O_RDONLY); |
290 | |
291 | if (fd == -1) |
292 | { |
293 | error (0, errno, _("cannot open input file `%s'" ), |
294 | argv[remaining]); |
295 | status = EXIT_FAILURE; |
296 | continue; |
297 | } |
298 | } |
299 | |
300 | #ifdef _POSIX_MAPPED_FILES |
301 | /* We have possibilities for reading the input file. First try |
302 | to mmap() it since this will provide the fastest solution. */ |
303 | if (fstat64 (fd, &st) == 0 |
304 | && ((addr = mmap (NULL, st.st_size, PROT_READ, MAP_PRIVATE, |
305 | fd, 0)) != MAP_FAILED)) |
306 | { |
307 | /* Yes, we can use mmap(). The descriptor is not needed |
308 | anymore. */ |
309 | if (close (fd) != 0) |
310 | error (EXIT_FAILURE, errno, |
311 | _("error while closing input `%s'" ), |
312 | argv[remaining]); |
313 | |
314 | ret = process_block (cd, addr, st.st_size, &output, |
315 | output_file); |
316 | |
317 | /* We don't need the input data anymore. */ |
318 | munmap ((void *) addr, st.st_size); |
319 | |
320 | if (ret != 0) |
321 | { |
322 | status = EXIT_FAILURE; |
323 | |
324 | if (ret < 0) |
325 | /* We cannot go on with producing output since it might |
326 | lead to problem because the last output might leave |
327 | the output stream in an undefined state. */ |
328 | break; |
329 | } |
330 | } |
331 | else |
332 | #endif /* _POSIX_MAPPED_FILES */ |
333 | { |
334 | /* Read the file in pieces. */ |
335 | ret = process_fd (cd, fd, &output, output_file); |
336 | |
337 | /* Now close the file. */ |
338 | close (fd); |
339 | |
340 | if (ret != 0) |
341 | { |
342 | /* Something went wrong. */ |
343 | status = EXIT_FAILURE; |
344 | |
345 | if (ret < 0) |
346 | /* We cannot go on with producing output since it might |
347 | lead to problem because the last output might leave |
348 | the output stream in an undefined state. */ |
349 | break; |
350 | } |
351 | } |
352 | } |
353 | while (++remaining < argc); |
354 | |
355 | /* Close the output file now. */ |
356 | if (output != NULL && fclose (output)) |
357 | error (EXIT_FAILURE, errno, _("error while closing output file" )); |
358 | } |
359 | |
360 | return status; |
361 | } |
362 | |
363 | |
364 | /* Handle program arguments. */ |
365 | static error_t |
366 | parse_opt (int key, char *arg, struct argp_state *state) |
367 | { |
368 | switch (key) |
369 | { |
370 | case 'f': |
371 | from_code = arg; |
372 | break; |
373 | case 't': |
374 | to_code = arg; |
375 | break; |
376 | case 'o': |
377 | output_file = arg; |
378 | break; |
379 | case 's': |
380 | /* Nothing, for now at least. We are not giving out any information |
381 | about missing character or so. */ |
382 | break; |
383 | case 'c': |
384 | /* Omit invalid characters from output. */ |
385 | omit_invalid = 1; |
386 | break; |
387 | case OPT_VERBOSE: |
388 | verbose = 1; |
389 | break; |
390 | case OPT_LIST: |
391 | list = 1; |
392 | break; |
393 | default: |
394 | return ARGP_ERR_UNKNOWN; |
395 | } |
396 | return 0; |
397 | } |
398 | |
399 | |
400 | static char * |
401 | more_help (int key, const char *text, void *input) |
402 | { |
403 | char *tp = NULL; |
404 | switch (key) |
405 | { |
406 | case ARGP_KEY_HELP_EXTRA: |
407 | /* We print some extra information. */ |
408 | if (asprintf (&tp, gettext ("\ |
409 | For bug reporting instructions, please see:\n\ |
410 | %s.\n" ), REPORT_BUGS_TO) < 0) |
411 | return NULL; |
412 | return tp; |
413 | default: |
414 | break; |
415 | } |
416 | return (char *) text; |
417 | } |
418 | |
419 | |
420 | /* Print the version information. */ |
421 | static void |
422 | print_version (FILE *stream, struct argp_state *state) |
423 | { |
424 | fprintf (stream, "iconv %s%s\n" , PKGVERSION, VERSION); |
425 | fprintf (stream, gettext ("\ |
426 | Copyright (C) %s Free Software Foundation, Inc.\n\ |
427 | This is free software; see the source for copying conditions. There is NO\n\ |
428 | warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\ |
429 | " ), "2016" ); |
430 | fprintf (stream, gettext ("Written by %s.\n" ), "Ulrich Drepper" ); |
431 | } |
432 | |
433 | |
434 | static int |
435 | write_output (const char *outbuf, const char *outptr, FILE **output, |
436 | const char *output_file) |
437 | { |
438 | /* We have something to write out. */ |
439 | int errno_save = errno; |
440 | |
441 | if (*output == NULL) |
442 | { |
443 | /* Determine output file. */ |
444 | if (output_file != NULL && strcmp (output_file, "-" ) != 0) |
445 | { |
446 | *output = fopen (output_file, "w" ); |
447 | if (*output == NULL) |
448 | error (EXIT_FAILURE, errno, _("cannot open output file" )); |
449 | } |
450 | else |
451 | *output = stdout; |
452 | } |
453 | |
454 | if (fwrite (outbuf, 1, outptr - outbuf, *output) < (size_t) (outptr - outbuf) |
455 | || ferror (*output)) |
456 | { |
457 | /* Error occurred while printing the result. */ |
458 | error (0, 0, _("\ |
459 | conversion stopped due to problem in writing the output" )); |
460 | return -1; |
461 | } |
462 | |
463 | errno = errno_save; |
464 | |
465 | return 0; |
466 | } |
467 | |
468 | |
469 | static int |
470 | process_block (iconv_t cd, char *addr, size_t len, FILE **output, |
471 | const char *output_file) |
472 | { |
473 | #define OUTBUF_SIZE 32768 |
474 | const char *start = addr; |
475 | char outbuf[OUTBUF_SIZE]; |
476 | char *outptr; |
477 | size_t outlen; |
478 | size_t n; |
479 | int ret = 0; |
480 | |
481 | while (len > 0) |
482 | { |
483 | outptr = outbuf; |
484 | outlen = OUTBUF_SIZE; |
485 | n = iconv (cd, &addr, &len, &outptr, &outlen); |
486 | |
487 | if (n == (size_t) -1 && omit_invalid && errno == EILSEQ) |
488 | { |
489 | ret = 1; |
490 | if (len == 0) |
491 | n = 0; |
492 | else |
493 | errno = E2BIG; |
494 | } |
495 | |
496 | if (outptr != outbuf) |
497 | { |
498 | ret = write_output (outbuf, outptr, output, output_file); |
499 | if (ret != 0) |
500 | break; |
501 | } |
502 | |
503 | if (n != (size_t) -1) |
504 | { |
505 | /* All the input test is processed. For state-dependent |
506 | character sets we have to flush the state now. */ |
507 | outptr = outbuf; |
508 | outlen = OUTBUF_SIZE; |
509 | n = iconv (cd, NULL, NULL, &outptr, &outlen); |
510 | |
511 | if (outptr != outbuf) |
512 | { |
513 | ret = write_output (outbuf, outptr, output, output_file); |
514 | if (ret != 0) |
515 | break; |
516 | } |
517 | |
518 | if (n != (size_t) -1) |
519 | break; |
520 | |
521 | if (omit_invalid && errno == EILSEQ) |
522 | { |
523 | ret = 1; |
524 | break; |
525 | } |
526 | } |
527 | |
528 | if (errno != E2BIG) |
529 | { |
530 | /* iconv() ran into a problem. */ |
531 | switch (errno) |
532 | { |
533 | case EILSEQ: |
534 | if (! omit_invalid) |
535 | error (0, 0, _("illegal input sequence at position %ld" ), |
536 | (long int) (addr - start)); |
537 | break; |
538 | case EINVAL: |
539 | error (0, 0, _("\ |
540 | incomplete character or shift sequence at end of buffer" )); |
541 | break; |
542 | case EBADF: |
543 | error (0, 0, _("internal error (illegal descriptor)" )); |
544 | break; |
545 | default: |
546 | error (0, 0, _("unknown iconv() error %d" ), errno); |
547 | break; |
548 | } |
549 | |
550 | return -1; |
551 | } |
552 | } |
553 | |
554 | return ret; |
555 | } |
556 | |
557 | |
558 | static int |
559 | process_fd (iconv_t cd, int fd, FILE **output, const char *output_file) |
560 | { |
561 | /* we have a problem with reading from a desriptor since we must not |
562 | provide the iconv() function an incomplete character or shift |
563 | sequence at the end of the buffer. Since we have to deal with |
564 | arbitrary encodings we must read the whole text in a buffer and |
565 | process it in one step. */ |
566 | static char *inbuf = NULL; |
567 | static size_t maxlen = 0; |
568 | char *inptr = NULL; |
569 | size_t actlen = 0; |
570 | |
571 | while (actlen < maxlen) |
572 | { |
573 | ssize_t n = read (fd, inptr, maxlen - actlen); |
574 | |
575 | if (n == 0) |
576 | /* No more text to read. */ |
577 | break; |
578 | |
579 | if (n == -1) |
580 | { |
581 | /* Error while reading. */ |
582 | error (0, errno, _("error while reading the input" )); |
583 | return -1; |
584 | } |
585 | |
586 | inptr += n; |
587 | actlen += n; |
588 | } |
589 | |
590 | if (actlen == maxlen) |
591 | while (1) |
592 | { |
593 | ssize_t n; |
594 | char *new_inbuf; |
595 | |
596 | /* Increase the buffer. */ |
597 | new_inbuf = (char *) realloc (inbuf, maxlen + 32768); |
598 | if (new_inbuf == NULL) |
599 | { |
600 | error (0, errno, _("unable to allocate buffer for input" )); |
601 | return -1; |
602 | } |
603 | inbuf = new_inbuf; |
604 | maxlen += 32768; |
605 | inptr = inbuf + actlen; |
606 | |
607 | do |
608 | { |
609 | n = read (fd, inptr, maxlen - actlen); |
610 | |
611 | if (n == 0) |
612 | /* No more text to read. */ |
613 | break; |
614 | |
615 | if (n == -1) |
616 | { |
617 | /* Error while reading. */ |
618 | error (0, errno, _("error while reading the input" )); |
619 | return -1; |
620 | } |
621 | |
622 | inptr += n; |
623 | actlen += n; |
624 | } |
625 | while (actlen < maxlen); |
626 | |
627 | if (n == 0) |
628 | /* Break again so we leave both loops. */ |
629 | break; |
630 | } |
631 | |
632 | /* Now we have all the input in the buffer. Process it in one run. */ |
633 | return process_block (cd, inbuf, actlen, output, output_file); |
634 | } |
635 | |
636 | |
637 | static int |
638 | process_file (iconv_t cd, FILE *input, FILE **output, const char *output_file) |
639 | { |
640 | /* This should be safe since we use this function only for `stdin' and |
641 | we haven't read anything so far. */ |
642 | return process_fd (cd, fileno (input), output, output_file); |
643 | } |
644 | |
645 | |
646 | /* Print all known character sets/encodings. */ |
647 | static void *printlist; |
648 | static size_t column; |
649 | static int not_first; |
650 | |
651 | static void |
652 | insert_print_list (const void *nodep, VISIT value, int level) |
653 | { |
654 | if (value == leaf || value == postorder) |
655 | { |
656 | const struct gconv_alias *s = *(const struct gconv_alias **) nodep; |
657 | tsearch (s->fromname, &printlist, (__compar_fn_t) strverscmp); |
658 | } |
659 | } |
660 | |
661 | static void |
662 | do_print_human (const void *nodep, VISIT value, int level) |
663 | { |
664 | if (value == leaf || value == postorder) |
665 | { |
666 | const char *s = *(const char **) nodep; |
667 | size_t len = strlen (s); |
668 | size_t cnt; |
669 | |
670 | while (len > 0 && s[len - 1] == '/') |
671 | --len; |
672 | |
673 | for (cnt = 0; cnt < len; ++cnt) |
674 | if (isalnum (s[cnt])) |
675 | break; |
676 | if (cnt == len) |
677 | return; |
678 | |
679 | if (not_first) |
680 | { |
681 | putchar (','); |
682 | ++column; |
683 | |
684 | if (column > 2 && column + len > 77) |
685 | { |
686 | fputs ("\n " , stdout); |
687 | column = 2; |
688 | } |
689 | else |
690 | { |
691 | putchar (' '); |
692 | ++column; |
693 | } |
694 | } |
695 | else |
696 | not_first = 1; |
697 | |
698 | fwrite (s, len, 1, stdout); |
699 | column += len; |
700 | } |
701 | } |
702 | |
703 | static void |
704 | do_print (const void *nodep, VISIT value, int level) |
705 | { |
706 | if (value == leaf || value == postorder) |
707 | { |
708 | const char *s = *(const char **) nodep; |
709 | |
710 | puts (s); |
711 | } |
712 | } |
713 | |
714 | static void |
715 | internal_function |
716 | add_known_names (struct gconv_module *node) |
717 | { |
718 | if (node->left != NULL) |
719 | add_known_names (node->left); |
720 | if (node->right != NULL) |
721 | add_known_names (node->right); |
722 | do |
723 | { |
724 | if (strcmp (node->from_string, "INTERNAL" ) != 0) |
725 | tsearch (node->from_string, &printlist, (__compar_fn_t) strverscmp); |
726 | if (strcmp (node->to_string, "INTERNAL" ) != 0) |
727 | tsearch (node->to_string, &printlist, (__compar_fn_t) strverscmp); |
728 | |
729 | node = node->same; |
730 | } |
731 | while (node != NULL); |
732 | } |
733 | |
734 | |
735 | static void |
736 | insert_cache (void) |
737 | { |
738 | const struct gconvcache_header *; |
739 | const char *strtab; |
740 | const struct hash_entry *hashtab; |
741 | size_t cnt; |
742 | |
743 | header = (const struct gconvcache_header *) __gconv_get_cache (); |
744 | strtab = (char *) header + header->string_offset; |
745 | hashtab = (struct hash_entry *) ((char *) header + header->hash_offset); |
746 | |
747 | for (cnt = 0; cnt < header->hash_size; ++cnt) |
748 | if (hashtab[cnt].string_offset != 0) |
749 | { |
750 | const char *str = strtab + hashtab[cnt].string_offset; |
751 | |
752 | if (strcmp (str, "INTERNAL" ) != 0) |
753 | tsearch (str, &printlist, (__compar_fn_t) strverscmp); |
754 | } |
755 | } |
756 | |
757 | |
758 | static void |
759 | internal_function |
760 | print_known_names (void) |
761 | { |
762 | iconv_t h; |
763 | void *cache; |
764 | |
765 | /* We must initialize the internal databases first. */ |
766 | h = iconv_open ("L1" , "L1" ); |
767 | iconv_close (h); |
768 | |
769 | /* See whether we have a cache. */ |
770 | cache = __gconv_get_cache (); |
771 | if (cache != NULL) |
772 | /* Yep, use only this information. */ |
773 | insert_cache (); |
774 | else |
775 | { |
776 | struct gconv_module *modules; |
777 | |
778 | /* No, then use the information read from the gconv-modules file. |
779 | First add the aliases. */ |
780 | twalk (__gconv_get_alias_db (), insert_print_list); |
781 | |
782 | /* Add the from- and to-names from the known modules. */ |
783 | modules = __gconv_get_modules_db (); |
784 | if (modules != NULL) |
785 | add_known_names (modules); |
786 | } |
787 | |
788 | bool human_readable = isatty (fileno (stdout)); |
789 | |
790 | if (human_readable) |
791 | fputs (_("\ |
792 | The following list contains all the coded character sets known. This does\n\ |
793 | not necessarily mean that all combinations of these names can be used for\n\ |
794 | the FROM and TO command line parameters. One coded character set can be\n\ |
795 | listed with several different names (aliases).\n\n " ), stdout); |
796 | |
797 | /* Now print the collected names. */ |
798 | column = 2; |
799 | twalk (printlist, human_readable ? do_print_human : do_print); |
800 | |
801 | if (human_readable && column != 0) |
802 | puts ("" ); |
803 | } |
804 | |