1998-02-18 17:59:16 +00:00
|
|
|
/* Convert text in given files from the specified from-set to the to-set.
|
2024-01-01 18:12:26 +00:00
|
|
|
Copyright (C) 1998-2024 Free Software Foundation, Inc.
|
1998-02-18 17:59:16 +00:00
|
|
|
This file is part of the GNU C Library.
|
|
|
|
|
2005-12-07 05:49:17 +00:00
|
|
|
This program is free software; you can redistribute it and/or modify
|
2007-07-16 00:56:07 +00:00
|
|
|
it under the terms of the GNU General Public License as published
|
|
|
|
by the Free Software Foundation; version 2 of the License, or
|
|
|
|
(at your option) any later version.
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2005-12-07 05:49:17 +00:00
|
|
|
This program is distributed in the hope that it will be useful,
|
1998-02-18 17:59:16 +00:00
|
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
2005-12-07 05:49:17 +00:00
|
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
GNU General Public License for more details.
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2005-12-07 05:49:17 +00:00
|
|
|
You should have received a copy of the GNU General Public License
|
Prefer https to http for gnu.org and fsf.org URLs
Also, change sources.redhat.com to sourceware.org.
This patch was automatically generated by running the following shell
script, which uses GNU sed, and which avoids modifying files imported
from upstream:
sed -ri '
s,(http|ftp)(://(.*\.)?(gnu|fsf|sourceware)\.org($|[^.]|\.[^a-z])),https\2,g
s,(http|ftp)(://(.*\.)?)sources\.redhat\.com($|[^.]|\.[^a-z]),https\2sourceware.org\4,g
' \
$(find $(git ls-files) -prune -type f \
! -name '*.po' \
! -name 'ChangeLog*' \
! -path COPYING ! -path COPYING.LIB \
! -path manual/fdl-1.3.texi ! -path manual/lgpl-2.1.texi \
! -path manual/texinfo.tex ! -path scripts/config.guess \
! -path scripts/config.sub ! -path scripts/install-sh \
! -path scripts/mkinstalldirs ! -path scripts/move-if-change \
! -path INSTALL ! -path locale/programs/charmap-kw.h \
! -path po/libc.pot ! -path sysdeps/gnu/errlist.c \
! '(' -name configure \
-execdir test -f configure.ac -o -f configure.in ';' ')' \
! '(' -name preconfigure \
-execdir test -f preconfigure.ac ';' ')' \
-print)
and then by running 'make dist-prepare' to regenerate files built
from the altered files, and then executing the following to cleanup:
chmod a+x sysdeps/unix/sysv/linux/riscv/configure
# Omit irrelevant whitespace and comment-only changes,
# perhaps from a slightly-different Autoconf version.
git checkout -f \
sysdeps/csky/configure \
sysdeps/hppa/configure \
sysdeps/riscv/configure \
sysdeps/unix/sysv/linux/csky/configure
# Omit changes that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/powerpc/powerpc64/ppc-mcount.S: trailing lines
git checkout -f \
sysdeps/powerpc/powerpc64/ppc-mcount.S \
sysdeps/unix/sysv/linux/s390/s390-64/syscall.S
# Omit change that caused a pre-commit check to fail like this:
# remote: *** error: sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S: last line does not end in newline
git checkout -f sysdeps/sparc/sparc64/multiarch/memcpy-ultra3.S
2019-09-07 05:40:42 +00:00
|
|
|
along with this program; if not, see <https://www.gnu.org/licenses/>. */
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
#include <argp.h>
|
2001-02-04 05:49:00 +00:00
|
|
|
#include <assert.h>
|
1998-04-01 16:15:20 +00:00
|
|
|
#include <ctype.h>
|
1998-02-18 17:59:16 +00:00
|
|
|
#include <errno.h>
|
|
|
|
#include <error.h>
|
|
|
|
#include <fcntl.h>
|
|
|
|
#include <iconv.h>
|
2002-07-27 08:20:17 +00:00
|
|
|
#include <langinfo.h>
|
1998-02-18 17:59:16 +00:00
|
|
|
#include <locale.h>
|
1998-04-01 16:15:20 +00:00
|
|
|
#include <search.h>
|
2002-09-24 04:19:03 +00:00
|
|
|
#include <stdbool.h>
|
1998-02-18 17:59:16 +00:00
|
|
|
#include <stdio.h>
|
|
|
|
#include <stdlib.h>
|
|
|
|
#include <string.h>
|
|
|
|
#include <unistd.h>
|
1999-06-19 09:58:37 +00:00
|
|
|
#include <libintl.h>
|
2001-02-05 05:57:24 +00:00
|
|
|
#include <charmap.h>
|
1998-04-03 15:46:53 +00:00
|
|
|
#include <gconv_int.h>
|
2001-02-05 05:57:24 +00:00
|
|
|
#include "iconv_prog.h"
|
2001-09-01 19:03:22 +00:00
|
|
|
#include "iconvconfig.h"
|
2020-07-07 18:31:48 +00:00
|
|
|
#include "gconv_charset.h"
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
/* Get libc version number. */
|
|
|
|
#include "../version.h"
|
|
|
|
|
|
|
|
#define PACKAGE _libc_intl_domainname
|
|
|
|
|
|
|
|
|
|
|
|
/* Name and version of program. */
|
|
|
|
static void print_version (FILE *stream, struct argp_state *state);
|
|
|
|
void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version;
|
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
enum
|
|
|
|
{
|
|
|
|
OPT_VERBOSE = 1000,
|
|
|
|
OPT_BUFFER_SIZE,
|
|
|
|
};
|
2000-04-30 06:52:59 +00:00
|
|
|
#define OPT_LIST 'l'
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
/* Definitions of arguments for argp functions. */
|
|
|
|
static const struct argp_option options[] =
|
|
|
|
{
|
|
|
|
{ NULL, 0, NULL, 0, N_("Input/Output format specification:") },
|
2013-03-18 21:18:53 +00:00
|
|
|
{ "from-code", 'f', N_("NAME"), 0, N_("encoding of original text") },
|
|
|
|
{ "to-code", 't', N_("NAME"), 0, N_("encoding for output") },
|
1998-04-01 16:15:20 +00:00
|
|
|
{ NULL, 0, NULL, 0, N_("Information:") },
|
2000-04-30 06:52:59 +00:00
|
|
|
{ "list", 'l', NULL, 0, N_("list all known coded character sets") },
|
1998-02-18 17:59:16 +00:00
|
|
|
{ NULL, 0, NULL, 0, N_("Output control:") },
|
2000-04-30 06:52:59 +00:00
|
|
|
{ NULL, 'c', NULL, 0, N_("omit invalid characters from output") },
|
2013-03-18 21:18:53 +00:00
|
|
|
{ "output", 'o', N_("FILE"), 0, N_("output file") },
|
2000-09-04 09:40:57 +00:00
|
|
|
{ "silent", 's', NULL, 0, N_("suppress warnings") },
|
1998-02-18 17:59:16 +00:00
|
|
|
{ "verbose", OPT_VERBOSE, NULL, 0, N_("print progress information") },
|
2024-09-20 11:10:54 +00:00
|
|
|
/* This is an internal option intended for testing only. Very small
|
|
|
|
buffers do not work with all character sets. */
|
|
|
|
{ "buffer-size", OPT_BUFFER_SIZE, N_("BYTE-COUNT"), OPTION_HIDDEN,
|
|
|
|
N_("size of in-memory scratch buffer") },
|
1998-02-18 17:59:16 +00:00
|
|
|
{ NULL, 0, NULL, 0, NULL }
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Short description of program. */
|
|
|
|
static const char doc[] = N_("\
|
|
|
|
Convert encoding of given files from one encoding to another.");
|
|
|
|
|
|
|
|
/* Strings for arguments in help texts. */
|
|
|
|
static const char args_doc[] = N_("[FILE...]");
|
|
|
|
|
|
|
|
/* Prototype for option handler. */
|
2000-04-30 06:52:59 +00:00
|
|
|
static error_t parse_opt (int key, char *arg, struct argp_state *state);
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
/* Function to print some extra text in the help message. */
|
2000-04-30 06:52:59 +00:00
|
|
|
static char *more_help (int key, const char *text, void *input);
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
/* Data structure to communicate with argp functions. */
|
|
|
|
static struct argp argp =
|
|
|
|
{
|
|
|
|
options, parse_opt, args_doc, doc, NULL, more_help
|
|
|
|
};
|
|
|
|
|
2001-11-29 05:00:56 +00:00
|
|
|
/* Code sets to convert from and to respectively. An empty string as the
|
|
|
|
default causes the 'iconv_open' function to look up the charset of the
|
|
|
|
currently selected locale and use it. */
|
|
|
|
static const char *from_code = "";
|
|
|
|
static const char *to_code = "";
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
/* File to write output to. If NULL write to stdout. */
|
|
|
|
static const char *output_file;
|
|
|
|
|
1998-04-01 16:15:20 +00:00
|
|
|
/* Nonzero if list of all coded character sets is wanted. */
|
|
|
|
static int list;
|
|
|
|
|
2000-06-06 03:16:30 +00:00
|
|
|
/* If nonzero omit invalid character from output. */
|
2001-02-05 05:57:24 +00:00
|
|
|
int omit_invalid;
|
2000-06-06 03:16:30 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* Current index in argv (after command line processing) with the
|
|
|
|
input file name. */
|
|
|
|
static int current_input_file_index;
|
|
|
|
|
|
|
|
/* Size of the temporary, in-memory buffer. Exceeding it needs
|
|
|
|
spooling to disk in a temporary file. Controlled by --buffer_size. */
|
|
|
|
static size_t output_buffer_size = 1024 * 1024;
|
|
|
|
|
1998-02-18 17:59:16 +00:00
|
|
|
/* Prototypes for the functions doing the actual work. */
|
2024-09-20 11:10:54 +00:00
|
|
|
static void prepare_output_file (char **argv);
|
2024-09-20 11:10:54 +00:00
|
|
|
static void close_output_file (__gconv_t cd, int status);
|
|
|
|
static int process_block (iconv_t cd, char **addr, size_t *len,
|
|
|
|
off64_t file_offset, bool *incomplete);
|
2024-09-20 11:10:54 +00:00
|
|
|
static int process_fd (iconv_t cd, int fd);
|
|
|
|
static int process_file (iconv_t cd, FILE *input);
|
2017-08-15 15:52:59 +00:00
|
|
|
static void print_known_names (void);
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
|
|
|
|
int
|
|
|
|
main (int argc, char *argv[])
|
|
|
|
{
|
|
|
|
int status = EXIT_SUCCESS;
|
2020-07-07 18:31:48 +00:00
|
|
|
__gconv_t cd;
|
2001-02-05 05:57:24 +00:00
|
|
|
struct charmap_t *from_charmap = NULL;
|
|
|
|
struct charmap_t *to_charmap = NULL;
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
/* Set locale via LC_ALL. */
|
|
|
|
setlocale (LC_ALL, "");
|
|
|
|
|
|
|
|
/* Set the text message domain. */
|
|
|
|
textdomain (_libc_intl_domainname);
|
|
|
|
|
|
|
|
/* Parse and process arguments. */
|
2024-09-20 11:10:54 +00:00
|
|
|
argp_parse (&argp, argc, argv, 0, ¤t_input_file_index, NULL);
|
1998-02-18 17:59:16 +00:00
|
|
|
|
1998-04-01 16:15:20 +00:00
|
|
|
/* List all coded character sets if wanted. */
|
|
|
|
if (list)
|
|
|
|
{
|
|
|
|
print_known_names ();
|
|
|
|
exit (EXIT_SUCCESS);
|
|
|
|
}
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2001-02-05 05:57:24 +00:00
|
|
|
/* POSIX 1003.2b introduces a silly thing: the arguments to -t anf -f
|
|
|
|
can be file names of charmaps. In this case iconv will have to read
|
|
|
|
those charmaps and use them to do the conversion. But there are
|
|
|
|
holes in the specification. There is nothing said that if -f is a
|
|
|
|
charmap filename that -t must be, too. And vice versa. There is
|
|
|
|
also no word about the symbolic names used. What if they don't
|
|
|
|
match? */
|
|
|
|
if (strchr (from_code, '/') != NULL)
|
|
|
|
/* The from-name might be a charmap file name. Try reading the
|
|
|
|
file. */
|
2006-08-19 17:54:05 +00:00
|
|
|
from_charmap = charmap_read (from_code, /*0, 1*/1, 0, 0, 0);
|
2001-02-05 05:57:24 +00:00
|
|
|
|
2020-07-07 18:31:48 +00:00
|
|
|
if (strchr (to_code, '/') != NULL)
|
2001-02-05 05:57:24 +00:00
|
|
|
/* The to-name might be a charmap file name. Try reading the
|
|
|
|
file. */
|
2020-07-07 18:31:48 +00:00
|
|
|
to_charmap = charmap_read (to_code, /*0, 1,*/1, 0, 0, 0);
|
2001-02-05 05:57:24 +00:00
|
|
|
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2001-02-05 05:57:24 +00:00
|
|
|
/* At this point we have to handle two cases. The first one is
|
|
|
|
where a charmap is used for the from- or to-charset, or both. We
|
|
|
|
handle this special since it is very different from the sane way of
|
|
|
|
doing things. The other case allows converting using the iconv()
|
|
|
|
function. */
|
|
|
|
if (from_charmap != NULL || to_charmap != NULL)
|
|
|
|
/* Construct the conversion table and do the conversion. */
|
|
|
|
status = charmap_conversion (from_code, from_charmap, to_code, to_charmap,
|
2024-09-20 11:10:54 +00:00
|
|
|
argc, current_input_file_index, argv,
|
|
|
|
output_file);
|
1998-02-18 17:59:16 +00:00
|
|
|
else
|
2001-02-05 05:57:24 +00:00
|
|
|
{
|
2020-07-07 18:31:48 +00:00
|
|
|
struct gconv_spec conv_spec;
|
|
|
|
int res;
|
|
|
|
|
|
|
|
if (__gconv_create_spec (&conv_spec, from_code, to_code) == NULL)
|
|
|
|
{
|
|
|
|
error (EXIT_FAILURE, errno,
|
|
|
|
_("failed to start conversion processing"));
|
|
|
|
exit (1);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (omit_invalid)
|
|
|
|
conv_spec.ignore = true;
|
|
|
|
|
2001-02-05 05:57:24 +00:00
|
|
|
/* Let's see whether we have these coded character sets. */
|
2020-07-07 18:31:48 +00:00
|
|
|
res = __gconv_open (&conv_spec, &cd, 0);
|
|
|
|
|
2020-09-25 12:47:06 +00:00
|
|
|
__gconv_destroy_spec (&conv_spec);
|
2020-07-07 18:31:48 +00:00
|
|
|
|
|
|
|
if (res != __GCONV_OK)
|
2001-02-05 05:57:24 +00:00
|
|
|
{
|
2023-08-01 15:01:37 +00:00
|
|
|
if (res == __GCONV_NOCONV || res == __GCONV_NODB)
|
2002-09-24 04:19:03 +00:00
|
|
|
{
|
|
|
|
/* Try to be nice with the user and tell her which of the
|
|
|
|
two encoding names is wrong. This is possible because
|
|
|
|
all supported encodings can be converted from/to Unicode,
|
|
|
|
in other words, because the graph of encodings is
|
|
|
|
connected. */
|
|
|
|
bool from_wrong =
|
|
|
|
(iconv_open ("UTF-8", from_code) == (iconv_t) -1
|
|
|
|
&& errno == EINVAL);
|
|
|
|
bool to_wrong =
|
|
|
|
(iconv_open (to_code, "UTF-8") == (iconv_t) -1
|
|
|
|
&& errno == EINVAL);
|
|
|
|
const char *from_pretty =
|
|
|
|
(from_code[0] ? from_code : nl_langinfo (CODESET));
|
|
|
|
const char *to_pretty =
|
2020-07-07 18:31:48 +00:00
|
|
|
(to_code[0] ? to_code : nl_langinfo (CODESET));
|
2002-09-24 04:19:03 +00:00
|
|
|
|
|
|
|
if (from_wrong)
|
|
|
|
{
|
|
|
|
if (to_wrong)
|
2005-02-07 23:20:05 +00:00
|
|
|
error (0, 0,
|
2002-09-24 04:19:03 +00:00
|
|
|
_("\
|
2006-09-21 04:04:20 +00:00
|
|
|
conversions from `%s' and to `%s' are not supported"),
|
2002-09-24 04:19:03 +00:00
|
|
|
from_pretty, to_pretty);
|
|
|
|
else
|
2005-02-07 23:20:05 +00:00
|
|
|
error (0, 0,
|
2002-09-24 04:19:03 +00:00
|
|
|
_("conversion from `%s' is not supported"),
|
|
|
|
from_pretty);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (to_wrong)
|
2005-02-07 23:20:05 +00:00
|
|
|
error (0, 0,
|
2002-09-24 04:19:03 +00:00
|
|
|
_("conversion to `%s' is not supported"),
|
|
|
|
to_pretty);
|
|
|
|
else
|
2005-02-07 23:20:05 +00:00
|
|
|
error (0, 0,
|
2002-09-24 04:19:03 +00:00
|
|
|
_("conversion from `%s' to `%s' is not supported"),
|
|
|
|
from_pretty, to_pretty);
|
|
|
|
}
|
2005-02-07 23:20:05 +00:00
|
|
|
|
|
|
|
argp_help (&argp, stderr, ARGP_HELP_SEE,
|
|
|
|
program_invocation_short_name);
|
|
|
|
exit (1);
|
2002-09-24 04:19:03 +00:00
|
|
|
}
|
2001-02-05 05:57:24 +00:00
|
|
|
else
|
|
|
|
error (EXIT_FAILURE, errno,
|
|
|
|
_("failed to start conversion processing"));
|
|
|
|
}
|
2000-04-30 06:52:59 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
prepare_output_file (argv);
|
2008-03-11 20:14:03 +00:00
|
|
|
|
2001-02-05 05:57:24 +00:00
|
|
|
/* Now process the remaining files. Write them to stdout or the file
|
|
|
|
specified with the `-o' parameter. If we have no file given as
|
|
|
|
the parameter process all from stdin. */
|
2024-09-20 11:10:54 +00:00
|
|
|
if (current_input_file_index == argc)
|
2001-02-05 05:57:24 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
if (process_file (cd, stdin) != 0)
|
2001-02-05 05:57:24 +00:00
|
|
|
status = EXIT_FAILURE;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
do
|
|
|
|
{
|
2004-03-04 19:14:56 +00:00
|
|
|
int fd, ret;
|
2001-02-05 05:57:24 +00:00
|
|
|
|
|
|
|
if (verbose)
|
2024-09-20 11:10:54 +00:00
|
|
|
fprintf (stderr, "%s:\n", argv[current_input_file_index]);
|
|
|
|
if (strcmp (argv[current_input_file_index], "-") == 0)
|
|
|
|
fd = STDIN_FILENO;
|
2001-02-05 05:57:24 +00:00
|
|
|
else
|
2000-04-30 06:52:59 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
fd = open (argv[current_input_file_index], O_RDONLY);
|
2001-02-05 05:57:24 +00:00
|
|
|
|
|
|
|
if (fd == -1)
|
|
|
|
{
|
|
|
|
error (0, errno, _("cannot open input file `%s'"),
|
2024-09-20 11:10:54 +00:00
|
|
|
argv[current_input_file_index]);
|
2001-02-05 05:57:24 +00:00
|
|
|
status = EXIT_FAILURE;
|
|
|
|
continue;
|
|
|
|
}
|
2000-04-30 06:52:59 +00:00
|
|
|
}
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
{
|
2001-02-05 05:57:24 +00:00
|
|
|
/* Read the file in pieces. */
|
2024-09-20 11:10:54 +00:00
|
|
|
ret = process_fd (cd, fd);
|
2004-03-04 19:14:56 +00:00
|
|
|
|
|
|
|
/* Now close the file. */
|
2024-09-20 11:10:54 +00:00
|
|
|
if (fd != STDIN_FILENO)
|
|
|
|
close (fd);
|
2004-03-04 19:14:56 +00:00
|
|
|
|
|
|
|
if (ret != 0)
|
2001-02-05 05:57:24 +00:00
|
|
|
{
|
|
|
|
/* Something went wrong. */
|
|
|
|
status = EXIT_FAILURE;
|
|
|
|
|
2004-03-04 19:14:56 +00:00
|
|
|
if (ret < 0)
|
|
|
|
/* We cannot go on with producing output since it might
|
|
|
|
lead to problem because the last output might leave
|
|
|
|
the output stream in an undefined state. */
|
|
|
|
break;
|
2001-02-05 05:57:24 +00:00
|
|
|
}
|
1998-02-18 17:59:16 +00:00
|
|
|
}
|
|
|
|
}
|
2024-09-20 11:10:54 +00:00
|
|
|
while (++current_input_file_index < argc);
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* Ensure that iconv -c still exits with failure if iconv (the
|
|
|
|
function) has failed with E2BIG instead of EILSEQ. */
|
|
|
|
if (__gconv_has_illegal_input (cd))
|
|
|
|
status = EXIT_FAILURE;
|
|
|
|
|
2008-03-11 20:14:03 +00:00
|
|
|
/* Close the output file now. */
|
2024-09-20 11:10:54 +00:00
|
|
|
close_output_file (cd, status);
|
2008-03-11 20:14:03 +00:00
|
|
|
}
|
1998-02-18 17:59:16 +00:00
|
|
|
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Handle program arguments. */
|
|
|
|
static error_t
|
|
|
|
parse_opt (int key, char *arg, struct argp_state *state)
|
|
|
|
{
|
|
|
|
switch (key)
|
|
|
|
{
|
|
|
|
case 'f':
|
|
|
|
from_code = arg;
|
|
|
|
break;
|
|
|
|
case 't':
|
|
|
|
to_code = arg;
|
|
|
|
break;
|
|
|
|
case 'o':
|
|
|
|
output_file = arg;
|
|
|
|
break;
|
2000-04-30 06:52:59 +00:00
|
|
|
case 's':
|
|
|
|
/* Nothing, for now at least. We are not giving out any information
|
|
|
|
about missing character or so. */
|
|
|
|
break;
|
|
|
|
case 'c':
|
2000-06-06 03:16:30 +00:00
|
|
|
/* Omit invalid characters from output. */
|
|
|
|
omit_invalid = 1;
|
2000-04-30 06:52:59 +00:00
|
|
|
break;
|
2024-09-20 11:10:54 +00:00
|
|
|
case OPT_BUFFER_SIZE:
|
|
|
|
{
|
|
|
|
int i = atoi (arg);
|
|
|
|
if (i <= 0)
|
|
|
|
error (EXIT_FAILURE, 0, _("invalid buffer size: %s"), arg);
|
|
|
|
output_buffer_size = i;
|
|
|
|
}
|
|
|
|
break;
|
1998-02-18 17:59:16 +00:00
|
|
|
case OPT_VERBOSE:
|
|
|
|
verbose = 1;
|
|
|
|
break;
|
1998-04-01 16:15:20 +00:00
|
|
|
case OPT_LIST:
|
|
|
|
list = 1;
|
|
|
|
break;
|
1998-02-18 17:59:16 +00:00
|
|
|
default:
|
|
|
|
return ARGP_ERR_UNKNOWN;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static char *
|
|
|
|
more_help (int key, const char *text, void *input)
|
|
|
|
{
|
2012-11-09 22:13:45 +00:00
|
|
|
char *tp = NULL;
|
1998-02-18 17:59:16 +00:00
|
|
|
switch (key)
|
|
|
|
{
|
|
|
|
case ARGP_KEY_HELP_EXTRA:
|
|
|
|
/* We print some extra information. */
|
2012-11-09 22:13:45 +00:00
|
|
|
if (asprintf (&tp, gettext ("\
|
2004-05-17 18:59:35 +00:00
|
|
|
For bug reporting instructions, please see:\n\
|
2012-11-09 22:13:45 +00:00
|
|
|
%s.\n"), REPORT_BUGS_TO) < 0)
|
|
|
|
return NULL;
|
|
|
|
return tp;
|
1998-02-18 17:59:16 +00:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return (char *) text;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Print the version information. */
|
|
|
|
static void
|
|
|
|
print_version (FILE *stream, struct argp_state *state)
|
|
|
|
{
|
2012-11-09 22:13:45 +00:00
|
|
|
fprintf (stream, "iconv %s%s\n", PKGVERSION, VERSION);
|
1998-02-18 17:59:16 +00:00
|
|
|
fprintf (stream, gettext ("\
|
|
|
|
Copyright (C) %s Free Software Foundation, Inc.\n\
|
|
|
|
This is free software; see the source for copying conditions. There is NO\n\
|
|
|
|
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
|
2024-01-01 18:35:28 +00:00
|
|
|
"), "2024");
|
1998-02-18 17:59:16 +00:00
|
|
|
fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
|
|
|
|
}
|
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* Command line index of the last input file that overlaps with the
|
|
|
|
output file. Zero means no temporary file is ever required. */
|
|
|
|
static int last_overlapping_file_index;
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* This is set to true if the output is written to a temporary file. */
|
|
|
|
static bool output_using_temporary_file;
|
|
|
|
|
|
|
|
/* This is the file descriptor that will be used by write_output. */
|
|
|
|
static int output_fd = -1;
|
|
|
|
|
|
|
|
/* Pointers at the start and end of the fixed-size output buffer. */
|
|
|
|
static char *output_buffer_start;
|
|
|
|
|
|
|
|
/* Current write position in the output buffer. */
|
|
|
|
static char *output_buffer_current;
|
|
|
|
|
|
|
|
/* Remaining bytes after output_buffer_current in the output buffer. */
|
|
|
|
static size_t output_buffer_remaining;
|
|
|
|
|
|
|
|
|
|
|
|
/* Reduce the buffer size when writing directly to the output file, to
|
|
|
|
reduce cache utilization. */
|
|
|
|
static size_t copy_buffer_size = BUFSIZ;
|
|
|
|
|
|
|
|
static void
|
|
|
|
output_error (void)
|
|
|
|
{
|
|
|
|
error (EXIT_FAILURE, errno, _("cannot open output file"));
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
input_error (const char *path)
|
2008-03-11 20:14:03 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
error (0, errno, _("cannot open input file `%s'"), path);
|
|
|
|
}
|
2008-03-11 20:14:03 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* Opens output_file for writing, truncating it. */
|
|
|
|
static void
|
|
|
|
open_output_direct (void)
|
|
|
|
{
|
|
|
|
output_fd = open64 (output_file, O_WRONLY | O_CREAT | O_TRUNC, 0777);
|
|
|
|
if (output_fd < 0)
|
|
|
|
output_error ();
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
prepare_output_file (char **argv)
|
|
|
|
{
|
|
|
|
if (copy_buffer_size > output_buffer_size)
|
|
|
|
copy_buffer_size = output_buffer_size;
|
|
|
|
|
|
|
|
if (output_file == NULL || strcmp (output_file, "-") == 0)
|
2008-03-11 20:14:03 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
/* No buffering is required when writing to standard output
|
|
|
|
because input overlap is expected to be solved externally. */
|
|
|
|
output_fd = STDOUT_FILENO;
|
|
|
|
output_buffer_size = copy_buffer_size;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/* If iconv creates the output file, no overlap is possible. */
|
|
|
|
output_fd = open64 (output_file, O_WRONLY | O_CREAT | O_EXCL, 0777);
|
|
|
|
if (output_fd >= 0)
|
|
|
|
output_buffer_size = copy_buffer_size;
|
|
|
|
else
|
2008-03-11 20:14:03 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
/* Otherwise, check if any of the input files overlap with the
|
|
|
|
output file. */
|
|
|
|
struct statx st;
|
|
|
|
if (statx (AT_FDCWD, output_file, 0, STATX_INO | STATX_MODE, &st)
|
|
|
|
!= 0)
|
|
|
|
output_error ();
|
|
|
|
uint32_t out_dev_minor = st.stx_dev_minor;
|
|
|
|
uint32_t out_dev_major = st.stx_dev_major;
|
|
|
|
uint64_t out_ino = st.stx_ino;
|
|
|
|
|
|
|
|
int idx = current_input_file_index;
|
|
|
|
while (true)
|
|
|
|
{
|
|
|
|
/* Special case: no input files means standard input. */
|
|
|
|
if (argv[idx] == NULL && idx != current_input_file_index)
|
|
|
|
break;
|
|
|
|
|
|
|
|
int ret;
|
|
|
|
if (argv[idx] == NULL || strcmp (argv[idx], "-") == 0)
|
|
|
|
ret = statx (STDIN_FILENO, "", AT_EMPTY_PATH, STATX_INO, &st);
|
|
|
|
else
|
|
|
|
ret = statx (AT_FDCWD, argv[idx], 0, STATX_INO, &st);
|
|
|
|
if (ret != 0)
|
|
|
|
{
|
|
|
|
input_error (argv[idx]);
|
|
|
|
exit (EXIT_FAILURE);
|
|
|
|
}
|
|
|
|
if (out_dev_minor == st.stx_dev_minor
|
|
|
|
&& out_dev_major == st.stx_dev_major
|
|
|
|
&& out_ino == st.stx_ino)
|
|
|
|
{
|
|
|
|
if (argv[idx] == NULL)
|
|
|
|
/* Corner case: index of NULL would be larger than
|
|
|
|
idx while converting, triggering a switch away
|
|
|
|
from the temporary file. */
|
|
|
|
last_overlapping_file_index = INT_MAX;
|
|
|
|
else
|
|
|
|
last_overlapping_file_index = idx;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (argv[idx] == NULL)
|
|
|
|
break;
|
|
|
|
++idx;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If there is no overlap, avoid using a temporary file. */
|
|
|
|
if (last_overlapping_file_index == 0)
|
|
|
|
{
|
|
|
|
open_output_direct ();
|
|
|
|
output_buffer_size = copy_buffer_size;
|
|
|
|
}
|
2008-03-11 20:14:03 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
output_buffer_start = malloc (output_buffer_size);
|
|
|
|
if (output_buffer_start == NULL)
|
|
|
|
output_error ();
|
|
|
|
output_buffer_current = output_buffer_start;
|
|
|
|
output_buffer_remaining = output_buffer_size;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Write out the range [first, last), terminating the process on write
|
|
|
|
error. */
|
|
|
|
static void
|
|
|
|
write_fully (int fd, const char *first, const char *last)
|
|
|
|
{
|
|
|
|
while (first < last)
|
2008-03-11 20:14:03 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
ssize_t ret = write (fd, first, last - first);
|
|
|
|
if (ret == 0)
|
|
|
|
{
|
|
|
|
errno = ENOSPC;
|
|
|
|
output_error ();
|
|
|
|
}
|
|
|
|
if (ret < 0)
|
|
|
|
error (EXIT_FAILURE, errno, _("\
|
2008-03-11 20:14:03 +00:00
|
|
|
conversion stopped due to problem in writing the output"));
|
2024-09-20 11:10:54 +00:00
|
|
|
first += ret;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
flush_output (void)
|
|
|
|
{
|
|
|
|
bool temporary_file_not_needed
|
|
|
|
= current_input_file_index > last_overlapping_file_index;
|
|
|
|
if (output_fd < 0)
|
|
|
|
{
|
|
|
|
if (temporary_file_not_needed)
|
|
|
|
open_output_direct ();
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/* Create an anonymous temporary file. */
|
|
|
|
FILE *fp = tmpfile ();
|
|
|
|
if (fp == NULL)
|
|
|
|
output_error ();
|
|
|
|
output_fd = dup (fileno (fp));
|
|
|
|
if (output_fd < 0)
|
|
|
|
output_error ();
|
|
|
|
fclose (fp);
|
|
|
|
output_using_temporary_file = true;
|
|
|
|
}
|
|
|
|
/* Either way, no longer use a memory-only staging buffer. */
|
|
|
|
output_buffer_size = copy_buffer_size;
|
2008-03-11 20:14:03 +00:00
|
|
|
}
|
2024-09-20 11:10:54 +00:00
|
|
|
else if (output_using_temporary_file && temporary_file_not_needed)
|
|
|
|
{
|
|
|
|
/* The temporary file is no longer needed. Switch to direct
|
|
|
|
output, replacing output_fd. */
|
|
|
|
int temp_fd = output_fd;
|
|
|
|
open_output_direct ();
|
|
|
|
|
|
|
|
/* Copy over the data spooled to the temporary file. */
|
|
|
|
if (lseek (temp_fd, 0, SEEK_SET) < 0)
|
|
|
|
output_error ();
|
|
|
|
while (true)
|
|
|
|
{
|
|
|
|
char buf[BUFSIZ];
|
|
|
|
ssize_t ret = read (temp_fd, buf, sizeof (buf));
|
|
|
|
if (ret < 0)
|
|
|
|
output_error ();
|
|
|
|
if (ret == 0)
|
|
|
|
break;
|
|
|
|
write_fully (output_fd, buf, buf + ret);
|
|
|
|
}
|
|
|
|
close (temp_fd);
|
2008-03-11 20:14:03 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* No longer using a temporary file from now on. */
|
|
|
|
output_using_temporary_file = false;
|
|
|
|
output_buffer_size = copy_buffer_size;
|
|
|
|
}
|
2008-03-11 20:14:03 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
write_fully (output_fd, output_buffer_start, output_buffer_current);
|
|
|
|
output_buffer_current = output_buffer_start;
|
|
|
|
output_buffer_remaining = output_buffer_size;
|
2008-03-11 20:14:03 +00:00
|
|
|
}
|
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
static void
|
2024-09-20 11:10:54 +00:00
|
|
|
close_output_file (__gconv_t cd, int status)
|
2024-09-20 11:10:54 +00:00
|
|
|
{
|
|
|
|
/* Do not perform a flush if a temporary file or the in-memory
|
|
|
|
buffer is in use and there was an error. It would clobber the
|
|
|
|
overlapping input file. */
|
|
|
|
if (status != EXIT_SUCCESS && !omit_invalid &&
|
|
|
|
(output_using_temporary_file || output_fd < 0))
|
|
|
|
return;
|
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* All the input text is processed. For state-dependent character
|
|
|
|
sets we have to flush the state now.
|
|
|
|
|
|
|
|
The current_input_file_index variable is now larger than
|
|
|
|
last_overlapping_file_index, so the flush_output calls switch
|
2024-09-20 11:10:54 +00:00
|
|
|
away from the temporary file. */
|
2024-09-20 11:10:54 +00:00
|
|
|
size_t n = iconv (cd, NULL, NULL,
|
|
|
|
&output_buffer_current, &output_buffer_remaining);
|
|
|
|
if (n == (size_t) -1 && errno == E2BIG)
|
|
|
|
{
|
|
|
|
/* Try again if the state flush exceeded the buffer space. */
|
|
|
|
flush_output ();
|
|
|
|
n = iconv (cd, NULL, NULL,
|
|
|
|
&output_buffer_current, &output_buffer_remaining);
|
|
|
|
}
|
|
|
|
int saved_errno = errno;
|
2024-09-20 11:10:54 +00:00
|
|
|
flush_output ();
|
2024-09-20 11:10:54 +00:00
|
|
|
if (n == (size_t) -1 && !omit_invalid)
|
|
|
|
{
|
|
|
|
errno = saved_errno;
|
|
|
|
output_error ();
|
|
|
|
}
|
2024-09-20 11:10:54 +00:00
|
|
|
|
|
|
|
if (output_fd == STDOUT_FILENO)
|
|
|
|
{
|
|
|
|
/* Close standard output in safe manner, to report certain
|
|
|
|
ENOSPC errors. */
|
|
|
|
output_fd = dup (output_fd);
|
|
|
|
if (output_fd < 0)
|
|
|
|
output_error ();
|
|
|
|
}
|
|
|
|
if (close (output_fd) < 0)
|
|
|
|
output_error ();
|
|
|
|
}
|
2008-03-11 20:14:03 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
/* CD is the iconv handle. Input processing starts at *ADDR, and
|
|
|
|
consumes upto *LEN bytes. *ADDR and *LEN are updated. FILE_OFFSET
|
|
|
|
is the file offset of the data initially at ADDR. *INCOMPLETE is
|
|
|
|
set to true if conversion stops due to an incomplete input
|
|
|
|
sequence. */
|
2008-03-11 20:14:03 +00:00
|
|
|
static int
|
2024-09-20 11:10:54 +00:00
|
|
|
process_block (iconv_t cd, char **addr, size_t *len, off64_t file_offset,
|
|
|
|
bool *incomplete)
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
const char *start = *addr;
|
Update.
1998-04-20 18:00 Ulrich Drepper <drepper@cygnus.com>
* libc.map: Add __dgettext to GLIBC_2.0 and __libc_longjmp, and
__libc_siglongjmp to GLIBC_2.1.
* elf/dl-minimal.c (__assert_perror_fail): Don't use strerror, use
__strerror_r.
* iconv/Makefile: Don't run tests now.
* iconv/iconv_prog.c (process_block): If loop is repeated, call iconv
with correct output buffer.
Major rewrite of the low-level gconv functionality.
* iconv/gconv.c: Rewritten.
* iconv/gconv.h: Likewise.
* iconv/gconv_builtin.c: Likewise.
* iconv/gconv_builtin.h: Likewise.
* iconv/gconv_conf.c: Likewise.
* iconv/gconv_int.h: Likewise.
* iconv/gconv_open.c: Likewise.
* iconv/gconv_simple.c: Likewise.
* iconv/iconv.c: Likewise.
* iconvdata/8bit-gap.c: Likewise.
* iconvdata/8bit-generic.c: Likewise.
* iconvdata/Makefile: Likewise.
* iconvdata/big5.c: Likewise.
* iconvdata/cns11643.c: Likewise.
* iconvdata/cns11643.h: Likewise.
* iconvdata/cns11643l1.c: Likewise.
* iconvdata/cns11643l1.h: Likewise.
* iconvdata/ebcdic-at-de-a.c: Likewise.
* iconvdata/ebcdic-at-de.c: Likewise.
* iconvdata/ebcdic-ca-fr.c: Likewise.
* iconvdata/euccn.c: Likewise.
* iconvdata/eucjp.c: Likewise.
* iconvdata/euckr.c: Likewise.
* iconvdata/euctw.c: Likewise.
* iconvdata/gb2312.c: Likewise.
* iconvdata/gb2312.h: Likewise.
* iconvdata/hp-roman8.c: Likewise.
* iconvdata/iso646.c: Likewise.
* iconvdata/iso6937.c: Likewise.
* iconvdata/iso8859-1.c: Likewise.
* iconvdata/iso8859-10.c: Likewise.
* iconvdata/iso8859-2.c: Likewise.
* iconvdata/iso8859-3.c: Likewise.
* iconvdata/iso8859-4.c: Likewise.
* iconvdata/iso8859-5.c: Likewise.
* iconvdata/iso8859-6.c: Likewise.
* iconvdata/iso8859-7.c: Likewise.
* iconvdata/iso8859-8.c: Likewise.
* iconvdata/iso8859-9.c: Likewise.
* iconvdata/jis0201.c: Likewise.
* iconvdata/jis0201.h: Likewise.
* iconvdata/jis0208.c: Likewise.
* iconvdata/jis0208.h: Likewise.
* iconvdata/jis0212.c: Likewise.
* iconvdata/jis0212.h: Likewise.
* iconvdata/johab.c: Likewise.
* iconvdata/koi-8.c: Likewise.
* iconvdata/koi8-r.c: Likewise.
* iconvdata/ksc5601.c: Likewise.
* iconvdata/ksc5601.h: Likewise.
* iconvdata/latin-greek-1.c: Likewise.
* iconvdata/latin-greek.c: Likewise.
* iconvdata/run-iconv-test.sh: Likewise.
* iconvdata/sjis.c: Likewise.
* iconvdata/t61.c: Likewise.
* iconvdata/uhc.c: Likewise.
* wcsmbs/btowc.c: Likewise.
* wcsmbs/mbrtowc.c: Likewise.
* wcsmbs/mbsnrtowcs.c: Likewise.
* wcsmbs/mbsrtowcs.c: Likewise.
* wcsmbs/wcrtomb.c: Likewise.
* wcsmbs/wcsmbsload.c: Likewise.
* wcsmbs/wcsnrtombs.c: Likewise.
* wcsmbs/wcsrtombs.c: Likewise.
* wcsmbs/wctob.c: Likewise.
* iconv/loop.c: New file.
* iconv/skeleton.c: New file.
* stdlib/mblen.c: Handle empty input string correctly.
* stdlib/mbtowc.c: Likewise.
* posix/getopt.c: Various cleanups.
* sysdeps/arm/bits/setjmp.h: Add copyright text.
* sysdeps/i386/bits/setjmp.h: Likewise.
* sysdeps/m68k/bits/setjmp.h: Likewise.
* sysdeps/powerpc/bits/setjmp.h: Likewise.
* sysdeps/sparc/sparc32/bits/setjmp.h: Likewise.
* sysdeps/generic/longjmp.c: Rename function to __libc_siglongjmp
and make longjmp weak alias.
1998-04-18 20:29 Philip Blundell <Philip.Blundell@pobox.com>
* iconv/Makefile (routines): Only include gconv_dl if building for
an ELF system - dynamic linking is not available on a.out.
(CFLAGS-gconv_conf.c): Define STATIC_GCONV if omitting gconv_dl
due to above check.
* iconv/gconv_db.c: If STATIC_GCONV defined, don't try to call
routines from gconv_dl.
1998-04-17 Gordon Matzigkeit <gord@profitpress.com>
* csu/init.c (_IO_stdin_used): Protect with USE_IN_LIBIO so that
we can compile without libio.
1998-04-20 16:28 Ulrich Drepper <drepper@cygnus.com>
* sysdeps/mach/hurd/Subdirs: Remove login.
1998-04-11 Gordon Matzigkeit <gord@profitpress.com>
* db2/compat.h: Include <errno.h>, to make sure we get the
definition of EFTYPE before we define it ourselves.
1998-04-10 Gordon Matzigkeit <gord@profitpress.com>
* sysdeps/generic/bits/socket.h: Protect against multiple inclusion.
* sysdeps/mach/hurd/bits/ioctls.h: Likewise.
Fix typo to allow inclusion from sys/ioctl.h again.
1998-04-16 Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
* iconvdata/*.[ch]: Clean up namespace. Optimize character lookup.
1998-04-16 Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
* libc.map: Export __strerror_r. Remove _strerror_internal.
1998-04-16 Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
* sysdeps/generic/strcasestr.c: Undefine strcasestr, not strstr.
Also undefine __strcasestr.
1998-04-16 Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
* posix/regex.c: Rename __re_max_failures back to re_max_failures,
aliases do not work with global variables due to copy relocations.
1998-04-20 15:12 Ulrich Drepper <drepper@cygnus.com>
* manual/creature.texi: Fix type. Patch by Andreas Schwab.
1998-04-20 13:47 Ulrich Drepper <drepper@cygnus.com>
* signal/sighold.c: Include stddef.h for NULL definition.
* signal/sigrelse.c: Likewise.
* sysdeps/posix/sigignore.c: Likewise.
* sysdeps/posix/sigset.c: Likewise.
* sysdeps/posix/waitid.c: Likewise.
* sysdeps/unix/sysv/linux/rt_sigsuspend.c: Likewise.
* sysdeps/unix/sysv/linux/rt_sigtimedwait.c: Likewise.
* sysdeps/unix/sysv/linux/sigwaitinfo.c: Likewise.
* wcsmbs/mbsrtowcs.c: Include stdlib.h for MB_CUR_MAX.
Patch by Franz Sirl <Franz.Sirl-kernel@lauterbach.com>.
1998-04-13 Mark Kettenis <kettenis@phys.uva.nl>
* login/Makefile (headers): Remove utmpx.h and bits/utmpx.h.
* login/getutent.c (getutxent): Remove alias.
* login/getutent_r.c (setutxent, pututxline, endutxent):
Remove aliases.
* login/getutid.c (getutxid): Remove alias.
* login/getutline.c (getutxline): Remove alias.
* login/utmp.h: Add prototypes for __updwtmp, __getutent,
__getutid, __getutline and __pututline.
* login/utmpx.h: Moved to ...
* sysdeps/gnu/utmpx.h: ... here. [__USE_GNU]: Define UTMPX_FILE,
UTMPX_FILENAME, WTMPX_FILE and WTMPX_FILENAME, declare utmpxname
and updwtmpx.
* login/updwtmp.c: Moved to ...
* sysdeps/generic/updwtmp.c: ... here. (updwtmp): Generalized by
allowing file name transformation.
* sysdeps/gnu/updwtmp.c: New file. Use generic implementation with
additional file name transformation.
* sysdeps/unix/sysv/linux/updwtmp.c: Likewise.
* login/utmp_file.c: Moved to ...
* sysdeps/generic/utmp_file.c: ... here. (setutent_file):
Generalized by allowing file name transformation. Do not
print error message. Library functions should not print them.
Reported by Jim Meyering.
* sysdeps/gnu/utmp_file.c: New file. Use generic implementation
with additional file name transformation.
* sysdeps/unix/sysv/linux/utmp_file.c: Likewise.
* sysdeps/gnu/Makefile [$(subdir)=login] (sysdep_routines): Add
setutxent, getutxent, endutxent, getutxid, getutxid, getutxline,
pututxline, utmpxname and updwtmpx. (sysdep_headers): Add utmpx.h
and bits/utmpx.h.
* sysdeps/gnu/bits/utmpx.h [__USE_GNU] Include paths.h.
(_PATH_UTMPX): Define to _PATH_UTMP. (_PATH_WTMPX): Define to
_PATH_WTMPX. (RUN_LVL): Define only if __USE_GNU. (ACCOUNTING):
Define if __USE_GNU.
* sysdeps/gnu/setutxent.c: New file.
* sysdeps/gnu/getutxent.c: New file.
* sysdeps/gnu/endutxent.c: New file.
* sysdeps/gnu/getutxid.c: New file.
* sysdeps/gnu/getutxline.c: New file.
* sysdeps/gnu/pututxline.c: New file.
* sysdeps/gnu/utmpxname.c: New file.
* sysdeps/gnu/updwtmpx.c: New file.
* sysdeps/unix/sysv/linux/paths.h (_PATH_UTMP_DB): Remove.
* sysdeps/generic/bits/utmpx.h: Remove.
1998-04-20 Ulrich Drepper <drepper@cygnus.com>
* posix/wordexp-test.c (main): Initialize ifs element of ts for
~root test.
1998-04-17 07:53 H.J. Lu <hjl@gnu.org>
* sysdeps/unix/sysv/linux/i386/s_pread64.S: Fix a typo.
1998-04-17 11:32 Ulrich Drepper <drepper@cygnus.com>
* libio/oldfileops.c (_IO_old_file_seekoff): Define temporary st
variable using _G_stat64.
* libio/fileops.c: Remove macro definition of fstat, it is in the
global header.
Reported by Thorsten Kukuk <kukuk@weber.uni-paderborn.de>.
1998-04-17 Philip Blundell <pb@nexus.co.uk>
* sysdeps/arm/strlen.S: New file, based on code by Matthew Wilcox
<willy@odie.barnet.ac.uk>.
1998-04-16 Philip Blundell <Philip.Blundell@pobox.com>
* inet/netinet/in.h (IN6_IS_ADDR_MC_NODELOCAL): New macro,
required by IPv6 Basic API.
(IN6_IS_ADDR_MC_LINKLOCAL): Likewise.
(IN6_IS_ADDR_MC_SITELOCAL): Likewise.
(IN6_IS_ADDR_MC_ORGLOCAL): Likewise.
(IN6_IS_ADDR_MC_GLOBAL): Likewise.
1998-04-20 18:41:05 +00:00
|
|
|
size_t n;
|
2004-03-04 19:14:56 +00:00
|
|
|
int ret = 0;
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
while (*len > 0)
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
n = iconv (cd, addr, len,
|
2024-09-20 11:10:54 +00:00
|
|
|
&output_buffer_current, &output_buffer_remaining);
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2004-03-04 19:14:56 +00:00
|
|
|
if (n == (size_t) -1 && omit_invalid && errno == EILSEQ)
|
|
|
|
{
|
|
|
|
ret = 1;
|
2024-09-20 11:10:54 +00:00
|
|
|
if (*len == 0)
|
2004-03-04 19:14:56 +00:00
|
|
|
n = 0;
|
|
|
|
else
|
|
|
|
errno = E2BIG;
|
|
|
|
}
|
|
|
|
|
1998-02-18 17:59:16 +00:00
|
|
|
if (n != (size_t) -1)
|
2024-09-20 11:10:54 +00:00
|
|
|
break;
|
1998-02-18 17:59:16 +00:00
|
|
|
|
2024-09-20 11:10:54 +00:00
|
|
|
if (errno == E2BIG)
|
|
|
|
flush_output ();
|
|
|
|
else
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
|
|
|
/* iconv() ran into a problem. */
|
|
|
|
switch (errno)
|
|
|
|
{
|
|
|
|
case EILSEQ:
|
2003-04-06 20:36:19 +00:00
|
|
|
if (! omit_invalid)
|
2024-09-20 11:10:54 +00:00
|
|
|
error (0, 0, _("illegal input sequence at position %lld"),
|
|
|
|
(long long int) (file_offset + (*addr - start)));
|
1998-02-18 17:59:16 +00:00
|
|
|
break;
|
|
|
|
case EINVAL:
|
2024-09-20 11:10:54 +00:00
|
|
|
*incomplete = true;
|
|
|
|
return ret;
|
1998-02-18 17:59:16 +00:00
|
|
|
case EBADF:
|
|
|
|
error (0, 0, _("internal error (illegal descriptor)"));
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
error (0, 0, _("unknown iconv() error %d"), errno);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2004-03-04 19:14:56 +00:00
|
|
|
return ret;
|
1998-02-18 17:59:16 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
2024-09-20 11:10:54 +00:00
|
|
|
process_fd (iconv_t cd, int fd)
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
char inbuf[BUFSIZ];
|
|
|
|
char *inbuf_end = inbuf + sizeof (inbuf);
|
|
|
|
size_t inbuf_used = 0;
|
|
|
|
off64_t file_offset = 0;
|
|
|
|
int status = 0;
|
|
|
|
bool incomplete = false;
|
|
|
|
|
|
|
|
while (true)
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
2024-09-20 11:10:54 +00:00
|
|
|
char *p = inbuf + inbuf_used;
|
|
|
|
ssize_t read_ret = read (fd, p, inbuf_end - p);
|
|
|
|
if (read_ret == 0)
|
|
|
|
{
|
|
|
|
/* On EOF, check if the previous iconv invocation saw an
|
|
|
|
incomplete sequence. */
|
|
|
|
if (incomplete)
|
|
|
|
{
|
|
|
|
error (0, 0, _("\
|
|
|
|
incomplete character or shift sequence at end of buffer"));
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
if (read_ret < 0)
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
|
|
|
error (0, errno, _("error while reading the input"));
|
|
|
|
return -1;
|
|
|
|
}
|
2024-09-20 11:10:54 +00:00
|
|
|
inbuf_used += read_ret;
|
|
|
|
incomplete = false;
|
|
|
|
p = inbuf;
|
|
|
|
int ret = process_block (cd, &p, &inbuf_used, file_offset, &incomplete);
|
|
|
|
if (ret != 0)
|
|
|
|
{
|
|
|
|
status = ret;
|
|
|
|
if (ret < 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
/* The next loop iteration consumes the leftover bytes. */
|
|
|
|
memmove (inbuf, p, inbuf_used);
|
|
|
|
file_offset += read_ret - inbuf_used;
|
1998-02-18 17:59:16 +00:00
|
|
|
}
|
2024-09-20 11:10:54 +00:00
|
|
|
return status;
|
1998-02-18 17:59:16 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int
|
2024-09-20 11:10:54 +00:00
|
|
|
process_file (iconv_t cd, FILE *input)
|
1998-02-18 17:59:16 +00:00
|
|
|
{
|
|
|
|
/* This should be safe since we use this function only for `stdin' and
|
|
|
|
we haven't read anything so far. */
|
2024-09-20 11:10:54 +00:00
|
|
|
return process_fd (cd, fileno (input));
|
1998-02-18 17:59:16 +00:00
|
|
|
}
|
1998-04-01 16:15:20 +00:00
|
|
|
|
|
|
|
|
|
|
|
/* Print all known character sets/encodings. */
|
|
|
|
static void *printlist;
|
|
|
|
static size_t column;
|
|
|
|
static int not_first;
|
|
|
|
|
|
|
|
static void
|
|
|
|
insert_print_list (const void *nodep, VISIT value, int level)
|
|
|
|
{
|
|
|
|
if (value == leaf || value == postorder)
|
|
|
|
{
|
|
|
|
const struct gconv_alias *s = *(const struct gconv_alias **) nodep;
|
1998-04-24 07:07:59 +00:00
|
|
|
tsearch (s->fromname, &printlist, (__compar_fn_t) strverscmp);
|
1998-04-01 16:15:20 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2001-04-20 17:33:29 +00:00
|
|
|
do_print_human (const void *nodep, VISIT value, int level)
|
1998-04-01 16:15:20 +00:00
|
|
|
{
|
|
|
|
if (value == leaf || value == postorder)
|
|
|
|
{
|
|
|
|
const char *s = *(const char **) nodep;
|
|
|
|
size_t len = strlen (s);
|
|
|
|
size_t cnt;
|
|
|
|
|
|
|
|
while (len > 0 && s[len - 1] == '/')
|
|
|
|
--len;
|
|
|
|
|
|
|
|
for (cnt = 0; cnt < len; ++cnt)
|
|
|
|
if (isalnum (s[cnt]))
|
|
|
|
break;
|
|
|
|
if (cnt == len)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (not_first)
|
|
|
|
{
|
|
|
|
putchar (',');
|
|
|
|
++column;
|
|
|
|
|
|
|
|
if (column > 2 && column + len > 77)
|
|
|
|
{
|
|
|
|
fputs ("\n ", stdout);
|
|
|
|
column = 2;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
putchar (' ');
|
|
|
|
++column;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
else
|
1998-04-24 07:07:59 +00:00
|
|
|
not_first = 1;
|
1998-04-01 16:15:20 +00:00
|
|
|
|
|
|
|
fwrite (s, len, 1, stdout);
|
|
|
|
column += len;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2001-04-20 17:33:29 +00:00
|
|
|
static void
|
|
|
|
do_print (const void *nodep, VISIT value, int level)
|
|
|
|
{
|
|
|
|
if (value == leaf || value == postorder)
|
|
|
|
{
|
|
|
|
const char *s = *(const char **) nodep;
|
|
|
|
|
|
|
|
puts (s);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
1998-04-01 16:15:20 +00:00
|
|
|
static void
|
1999-01-18 23:15:16 +00:00
|
|
|
add_known_names (struct gconv_module *node)
|
|
|
|
{
|
|
|
|
if (node->left != NULL)
|
|
|
|
add_known_names (node->left);
|
|
|
|
if (node->right != NULL)
|
|
|
|
add_known_names (node->right);
|
|
|
|
do
|
|
|
|
{
|
2012-01-08 12:19:21 +00:00
|
|
|
if (strcmp (node->from_string, "INTERNAL") != 0)
|
|
|
|
tsearch (node->from_string, &printlist, (__compar_fn_t) strverscmp);
|
|
|
|
if (strcmp (node->to_string, "INTERNAL") != 0)
|
2000-06-20 00:34:21 +00:00
|
|
|
tsearch (node->to_string, &printlist, (__compar_fn_t) strverscmp);
|
1999-01-18 23:15:16 +00:00
|
|
|
|
2000-06-20 00:34:21 +00:00
|
|
|
node = node->same;
|
1999-01-18 23:15:16 +00:00
|
|
|
}
|
|
|
|
while (node != NULL);
|
|
|
|
}
|
|
|
|
|
2001-09-01 19:03:22 +00:00
|
|
|
|
|
|
|
static void
|
|
|
|
insert_cache (void)
|
|
|
|
{
|
|
|
|
const struct gconvcache_header *header;
|
|
|
|
const char *strtab;
|
|
|
|
const struct hash_entry *hashtab;
|
|
|
|
size_t cnt;
|
|
|
|
|
2002-08-04 23:32:14 +00:00
|
|
|
header = (const struct gconvcache_header *) __gconv_get_cache ();
|
|
|
|
strtab = (char *) header + header->string_offset;
|
|
|
|
hashtab = (struct hash_entry *) ((char *) header + header->hash_offset);
|
2001-09-01 19:03:22 +00:00
|
|
|
|
|
|
|
for (cnt = 0; cnt < header->hash_size; ++cnt)
|
|
|
|
if (hashtab[cnt].string_offset != 0)
|
|
|
|
{
|
|
|
|
const char *str = strtab + hashtab[cnt].string_offset;
|
|
|
|
|
2012-01-08 12:19:21 +00:00
|
|
|
if (strcmp (str, "INTERNAL") != 0)
|
2001-09-01 19:03:22 +00:00
|
|
|
tsearch (str, &printlist, (__compar_fn_t) strverscmp);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
1999-01-18 23:15:16 +00:00
|
|
|
static void
|
1998-04-01 16:15:20 +00:00
|
|
|
print_known_names (void)
|
|
|
|
{
|
|
|
|
iconv_t h;
|
2002-08-04 23:32:14 +00:00
|
|
|
void *cache;
|
1998-04-01 16:15:20 +00:00
|
|
|
|
|
|
|
/* We must initialize the internal databases first. */
|
|
|
|
h = iconv_open ("L1", "L1");
|
|
|
|
iconv_close (h);
|
|
|
|
|
2001-09-01 19:03:22 +00:00
|
|
|
/* See whether we have a cache. */
|
2002-08-04 23:32:14 +00:00
|
|
|
cache = __gconv_get_cache ();
|
|
|
|
if (cache != NULL)
|
2001-09-01 19:03:22 +00:00
|
|
|
/* Yep, use only this information. */
|
|
|
|
insert_cache ();
|
|
|
|
else
|
|
|
|
{
|
2002-08-04 23:32:14 +00:00
|
|
|
struct gconv_module *modules;
|
|
|
|
|
2001-09-01 19:03:22 +00:00
|
|
|
/* No, then use the information read from the gconv-modules file.
|
|
|
|
First add the aliases. */
|
2002-08-04 23:32:14 +00:00
|
|
|
twalk (__gconv_get_alias_db (), insert_print_list);
|
1998-04-01 16:15:20 +00:00
|
|
|
|
2001-09-01 19:03:22 +00:00
|
|
|
/* Add the from- and to-names from the known modules. */
|
2002-08-04 23:32:14 +00:00
|
|
|
modules = __gconv_get_modules_db ();
|
|
|
|
if (modules != NULL)
|
|
|
|
add_known_names (modules);
|
2001-09-01 19:03:22 +00:00
|
|
|
}
|
1998-04-01 16:15:20 +00:00
|
|
|
|
2004-07-17 17:45:03 +00:00
|
|
|
bool human_readable = isatty (fileno (stdout));
|
|
|
|
|
|
|
|
if (human_readable)
|
|
|
|
fputs (_("\
|
2013-10-12 12:21:55 +00:00
|
|
|
The following list contains all the coded character sets known. This does\n\
|
1998-04-01 16:15:20 +00:00
|
|
|
not necessarily mean that all combinations of these names can be used for\n\
|
|
|
|
the FROM and TO command line parameters. One coded character set can be\n\
|
2000-06-20 00:34:21 +00:00
|
|
|
listed with several different names (aliases).\n\n "), stdout);
|
1998-04-01 16:15:20 +00:00
|
|
|
|
|
|
|
/* Now print the collected names. */
|
|
|
|
column = 2;
|
2004-07-17 17:45:03 +00:00
|
|
|
twalk (printlist, human_readable ? do_print_human : do_print);
|
1998-04-01 16:15:20 +00:00
|
|
|
|
2004-07-17 17:45:03 +00:00
|
|
|
if (human_readable && column != 0)
|
|
|
|
puts ("");
|
1998-04-01 16:15:20 +00:00
|
|
|
}
|