1 /* Convert text in given files from the specified from-set to the to-set.
2 Copyright (C) 1998, 1999 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998.
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.
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.
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. */
34 #ifdef _POSIX_MAPPED_FILES
35 # include <sys/mman.h>
37 #include <gconv_int.h>
39 /* Get libc version number. */
40 #include "../version.h"
42 #define PACKAGE _libc_intl_domainname
45 /* Name and version of program. */
46 static void print_version (FILE *stream
, struct argp_state
*state
);
47 void (*argp_program_version_hook
) (FILE *, struct argp_state
*) = print_version
;
49 #define OPT_VERBOSE 1000
52 /* Definitions of arguments for argp functions. */
53 static const struct argp_option options
[] =
55 { NULL
, 0, NULL
, 0, N_("Input/Output format specification:") },
56 { "from-code", 'f', "NAME", 0, N_("encoding of original text") },
57 { "to-code", 't', "NAME", 0, N_("encoding for output") },
58 { NULL
, 0, NULL
, 0, N_("Information:") },
59 { "list", OPT_LIST
, NULL
, 0, N_("list all known coded character sets") },
60 { NULL
, 0, NULL
, 0, N_("Output control:") },
61 { "output", 'o', "FILE", 0, N_("output file") },
62 { "verbose", OPT_VERBOSE
, NULL
, 0, N_("print progress information") },
63 { NULL
, 0, NULL
, 0, NULL
}
66 /* Short description of program. */
67 static const char doc
[] = N_("\
68 Convert encoding of given files from one encoding to another.");
70 /* Strings for arguments in help texts. */
71 static const char args_doc
[] = N_("[FILE...]");
73 /* Prototype for option handler. */
74 static error_t parse_opt
__P ((int key
, char *arg
, struct argp_state
*state
));
76 /* Function to print some extra text in the help message. */
77 static char *more_help
__P ((int key
, const char *text
, void *input
));
79 /* Data structure to communicate with argp functions. */
80 static struct argp argp
=
82 options
, parse_opt
, args_doc
, doc
, NULL
, more_help
85 /* Code sets to convert from and to respectively. */
86 static const char *from_code
;
87 static const char *to_code
;
89 /* File to write output to. If NULL write to stdout. */
90 static const char *output_file
;
92 /* Nonzero if verbose ouput is wanted. */
95 /* Nonzero if list of all coded character sets is wanted. */
98 /* Prototypes for the functions doing the actual work. */
99 static int process_block (iconv_t cd
, const char *addr
, size_t len
,
101 static int process_fd (iconv_t cd
, int fd
, FILE *output
);
102 static int process_file (iconv_t cd
, FILE *input
, FILE *output
);
103 static void print_known_names (void) internal_function
;
107 main (int argc
, char *argv
[])
109 int status
= EXIT_SUCCESS
;
114 /* Set locale via LC_ALL. */
115 setlocale (LC_ALL
, "");
117 /* Set the text message domain. */
118 textdomain (_libc_intl_domainname
);
120 /* Parse and process arguments. */
121 argp_parse (&argp
, argc
, argv
, 0, &remaining
, NULL
);
123 /* List all coded character sets if wanted. */
126 print_known_names ();
130 /* If either the from- or to-code is not specified this is an error
131 since we do not know what to do. */
132 if (from_code
== NULL
&& to_code
== NULL
)
133 error (EXIT_FAILURE
, 0,
134 _("neither original nor target encoding specified"));
135 if (from_code
== NULL
)
136 error (EXIT_FAILURE
, 0, _("original encoding not specified using `-f'"));
138 error (EXIT_FAILURE
, 0, _("target encoding not specified using `-t'"));
140 /* Let's see whether we have these coded character sets. */
141 cd
= iconv_open (to_code
, from_code
);
142 if (cd
== (iconv_t
) -1)
145 error (EXIT_FAILURE
, 0, _("conversion from `%s' to `%s' not supported"),
148 error (EXIT_FAILURE
, errno
, _("failed to start conversion processing"));
151 /* Determine output file. */
152 if (output_file
!= NULL
)
154 output
= fopen (output_file
, "w");
156 error (EXIT_FAILURE
, errno
, _("cannot open output file"));
161 /* Now process the remaining files. Write them to stdout or the file
162 specified with the `-o' parameter. If we have no file given as
163 the parameter process all from stdin. */
164 if (remaining
== argc
)
166 if (process_file (cd
, stdin
, output
) != 0)
167 status
= EXIT_FAILURE
;
174 int fd
= open (argv
[remaining
], O_RDONLY
);
177 printf ("%s:\n", argv
[remaining
]);
181 error (0, errno
, _("cannot open input file `%s'"),
183 status
= EXIT_FAILURE
;
187 #ifdef _POSIX_MAPPED_FILES
188 /* We have possibilities for reading the input file. First try
189 to mmap() it since this will provide the fastest solution. */
190 if (fstat (fd
, &st
) == 0
191 && ((addr
= mmap (NULL
, st
.st_size
, PROT_READ
, MAP_PRIVATE
, fd
, 0))
194 /* Yes, we can use mmap(). The descriptor is not needed
197 error (EXIT_FAILURE
, errno
, _("error while closing input `%s'"),
200 if (process_block (cd
, addr
, st
.st_size
, output
) < 0)
202 /* Something went wrong. */
203 status
= EXIT_FAILURE
;
205 /* We don't need the input data anymore. */
206 munmap ((void *) addr
, st
.st_size
);
208 /* We cannot go on with producing output since it might
209 lead to problem because the last output might leave
210 the output stream in an undefined state. */
214 /* We don't need the input data anymore. */
215 munmap ((void *) addr
, st
.st_size
);
218 #endif /* _POSIX_MAPPED_FILES */
220 /* Read the file in pieces. */
221 if (process_fd (cd
, fd
, output
) != 0)
223 /* Something went wrong. */
224 status
= EXIT_FAILURE
;
226 /* We don't need the input file anymore. */
229 /* We cannot go on with producing output since it might
230 lead to problem because the last output might leave
231 the output stream in an undefined state. */
235 /* Now close the file. */
239 while (++remaining
< argc
);
241 /* Close the output file now. */
243 error (EXIT_FAILURE
, errno
, _("error while closing output file"));
249 /* Handle program arguments. */
251 parse_opt (int key
, char *arg
, struct argp_state
*state
)
271 return ARGP_ERR_UNKNOWN
;
278 more_help (int key
, const char *text
, void *input
)
282 case ARGP_KEY_HELP_EXTRA
:
283 /* We print some extra information. */
284 return strdup (gettext ("\
285 Report bugs using the `glibcbug' script to <bugs@gnu.org>.\n"));
289 return (char *) text
;
293 /* Print the version information. */
295 print_version (FILE *stream
, struct argp_state
*state
)
297 fprintf (stream
, "iconv (GNU %s) %s\n", PACKAGE
, VERSION
);
298 fprintf (stream
, gettext ("\
299 Copyright (C) %s Free Software Foundation, Inc.\n\
300 This is free software; see the source for copying conditions. There is NO\n\
301 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
303 fprintf (stream
, gettext ("Written by %s.\n"), "Ulrich Drepper");
308 process_block (iconv_t cd
, const char *addr
, size_t len
, FILE *output
)
310 #define OUTBUF_SIZE 32768
311 const char *start
= addr
;
312 char outbuf
[OUTBUF_SIZE
];
320 outlen
= OUTBUF_SIZE
;
321 n
= iconv (cd
, &addr
, &len
, &outptr
, &outlen
);
323 if (outptr
!= outbuf
)
325 /* We have something to write out. */
326 int errno_save
= errno
;
328 if (fwrite (outbuf
, 1, outptr
- outbuf
, output
) < outptr
- outbuf
331 /* Error occurred while printing the result. */
333 conversion stopped due to problem in writing the output"));
340 if (n
!= (size_t) -1)
341 /* Everything is processed. */
346 /* iconv() ran into a problem. */
350 error (0, 0, _("illegal input sequence at position %ld"),
355 incomplete character or shift sequence at end of buffer"));
358 error (0, 0, _("internal error (illegal descriptor)"));
361 error (0, 0, _("unknown iconv() error %d"), errno
);
374 process_fd (iconv_t cd
, int fd
, FILE *output
)
376 /* we have a problem with reading from a desriptor since we must not
377 provide the iconv() function an incomplete character or shift
378 sequence at the end of the buffer. Since we have to deal with
379 arbitrary encodings we must read the whole text in a buffer and
380 process it in one step. */
381 static char *inbuf
= NULL
;
382 static size_t maxlen
= 0;
386 while (actlen
< maxlen
)
388 size_t n
= read (fd
, inptr
, maxlen
- actlen
);
391 /* No more text to read. */
396 /* Error while reading. */
397 error (0, errno
, _("error while reading the input"));
405 if (actlen
== maxlen
)
410 /* Increase the buffer. */
412 inbuf
= realloc (inbuf
, maxlen
);
414 error (0, errno
, _("unable to allocate buffer for input"));
415 inptr
= inbuf
+ actlen
;
419 n
= read (fd
, inptr
, maxlen
- actlen
);
422 /* No more text to read. */
427 /* Error while reading. */
428 error (0, errno
, _("error while reading the input"));
435 while (actlen
< maxlen
);
438 /* Break again so we leave both loops. */
442 /* Now we have all the input in the buffer. Process it in one run. */
443 return process_block (cd
, inbuf
, actlen
, output
);
448 process_file (iconv_t cd
, FILE *input
, FILE *output
)
450 /* This should be safe since we use this function only for `stdin' and
451 we haven't read anything so far. */
452 return process_fd (cd
, fileno (input
), output
);
456 /* Print all known character sets/encodings. */
457 static void *printlist
;
458 static size_t column
;
459 static int not_first
;
462 insert_print_list (const void *nodep
, VISIT value
, int level
)
464 if (value
== leaf
|| value
== postorder
)
466 const struct gconv_alias
*s
= *(const struct gconv_alias
**) nodep
;
467 tsearch (s
->fromname
, &printlist
, (__compar_fn_t
) strverscmp
);
472 do_print (const void *nodep
, VISIT value
, int level
)
474 if (value
== leaf
|| value
== postorder
)
476 const char *s
= *(const char **) nodep
;
477 size_t len
= strlen (s
);
480 while (len
> 0 && s
[len
- 1] == '/')
483 for (cnt
= 0; cnt
< len
; ++cnt
)
484 if (isalnum (s
[cnt
]))
494 if (column
> 2 && column
+ len
> 77)
496 fputs ("\n ", stdout
);
508 fwrite (s
, len
, 1, stdout
);
515 add_known_names (struct gconv_module
*node
)
517 if (node
->left
!= NULL
)
518 add_known_names (node
->left
);
519 if (node
->right
!= NULL
)
520 add_known_names (node
->right
);
521 if (node
->same
!= NULL
)
522 add_known_names (node
->same
);
525 if (node
->from_pattern
== NULL
)
527 if (strcmp (node
->from_constpfx
, "INTERNAL"))
528 tsearch (node
->from_constpfx
, &printlist
,
529 (__compar_fn_t
) strverscmp
);
530 if (strcmp (node
->to_string
, "INTERNAL"))
531 tsearch (node
->to_string
, &printlist
, (__compar_fn_t
) strverscmp
);
534 if (strcmp (node
->from_pattern
, "INTERNAL"))
535 tsearch (node
->from_pattern
, &printlist
, (__compar_fn_t
) strverscmp
);
537 node
= node
->matching
;
539 while (node
!= NULL
);
544 print_known_names (void)
548 /* We must initialize the internal databases first. */
549 h
= iconv_open ("L1", "L1");
552 /* First add the aliases. */
553 twalk (__gconv_alias_db
, insert_print_list
);
555 /* Add the from- and to-names from the known modules. */
556 add_known_names (__gconv_modules_db
);
559 The following list contain all the coded character sets known. This does\n\
560 not necessarily mean that all combinations of these names can be used for\n\
561 the FROM and TO command line parameters. One coded character set can be\n\
562 listed with several different names (aliases).\n\
563 Some of the names are no plain strings but instead regular expressions and\n\
564 they match a variety of names which can be given as parameters to the\n\
565 program.\n\n "), stdout
);
567 /* Now print the collected names. */
569 twalk (printlist
, do_print
);