usr.bin/iconv/iconv.c

   1 /* $FreeBSD$ */
   2 /* $NetBSD: iconv.c,v 1.16 2009/02/20 15:28:21 yamt Exp $ */
   3
   4 /*-
   5  * SPDX-License-Identifier: BSD-2-Clause
   6  *
   7  * Copyright (c)2003 Citrus Project,
   8  * All rights reserved.
   9  *
  10  * Redistribution and use in source and binary forms, with or without
  11  * modification, are permitted provided that the following conditions
  12  * are met:
  13  * 1. Redistributions of source code must retain the above copyright
  14  *    notice, this list of conditions and the following disclaimer.
  15  * 2. Redistributions in binary form must reproduce the above copyright
  16  *    notice, this list of conditions and the following disclaimer in the
  17  *    documentation and/or other materials provided with the distribution.
  18  *
  19  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  29  * SUCH DAMAGE.
  30  */
  31
  32 #include <sys/cdefs.h>
  33
  34 #include <err.h>
  35 #include <errno.h>
  36 #include <getopt.h>
  37 #include <iconv.h>
  38 #include <limits.h>
  39 #include <locale.h>
  40 #include <stdbool.h>
  41 #include <stdio.h>
  42 #include <stdlib.h>
  43 #include <string.h>
  44 #include <unistd.h>
  45
  46 static int              do_conv(FILE *, const char *, const char *, bool, bool);
  47 static int              do_list(unsigned int, const char * const *, void *);
  48 static void             usage(void) __dead2;
  49
  50 static const struct option long_options[] = {
  51         {"from-code",           required_argument,      NULL, 'f'},
  52         {"list",                no_argument,            NULL, 'l'},
  53         {"silent",              no_argument,            NULL, 's'},
  54         {"to-code",             required_argument,      NULL, 't'},
  55         {NULL,                  no_argument,            NULL, 0}
  56 };
  57
  58 static void
  59 usage(void)
  60 {
  61         (void)fprintf(stderr,
  62             "Usage:\t%1$s [-cs] -f <from_code> -t <to_code> [file ...]\n"
  63             "\t%1$s -f <from_code> [-cs] [-t <to_code>] [file ...]\n"
  64             "\t%1$s -t <to_code> [-cs] [-f <from_code>] [file ...]\n"
  65             "\t%1$s -l\n", getprogname());
  66         exit(1);
  67 }
  68
  69 #define INBUFSIZE 1024
  70 #define OUTBUFSIZE (INBUFSIZE * 2)
  71 static int
  72 do_conv(FILE *fp, const char *from, const char *to, bool silent,
  73     bool hide_invalid)
  74 {
  75         iconv_t cd;
  76         char inbuf[INBUFSIZE], outbuf[OUTBUFSIZE], *in, *out;
  77         unsigned long long invalids;
  78         size_t inbytes, outbytes, ret;
  79
  80         if ((cd = iconv_open(to, from)) == (iconv_t)-1)
  81                 err(EXIT_FAILURE, "iconv_open(%s, %s)", to, from);
  82
  83         if (hide_invalid) {
  84                 int arg = 1;
  85
  86                 if (iconvctl(cd, ICONV_SET_DISCARD_ILSEQ, (void *)&arg) == -1)
  87                         err(EXIT_FAILURE, NULL);
  88         }
  89         invalids = 0;
  90         while ((inbytes = fread(inbuf, 1, INBUFSIZE, fp)) > 0) {
  91                 in = inbuf;
  92                 while (inbytes > 0) {
  93                         size_t inval;
  94
  95                         out = outbuf;
  96                         outbytes = OUTBUFSIZE;
  97                         ret = __iconv(cd, &in, &inbytes, &out, &outbytes,
  98                             0, &inval);
  99                         invalids += inval;
 100                         if (outbytes < OUTBUFSIZE)
 101                                 (void)fwrite(outbuf, 1, OUTBUFSIZE - outbytes,
 102                                     stdout);
 103                         if (ret == (size_t)-1 && errno != E2BIG) {
 104                                 if (errno != EINVAL || in == inbuf)
 105                                         err(EXIT_FAILURE, "iconv()");
 106
 107                                 /* incomplete input character */
 108                                 (void)memmove(inbuf, in, inbytes);
 109                                 ret = fread(inbuf + inbytes, 1,
 110                                     INBUFSIZE - inbytes, fp);
 111                                 if (ret == 0) {
 112                                         fflush(stdout);
 113                                         if (feof(fp))
 114                                                 errx(EXIT_FAILURE,
 115                                                     "unexpected end of file; "
 116                                                     "the last character is "
 117                                                     "incomplete.");
 118                                         else
 119                                                 err(EXIT_FAILURE, "fread()");
 120                                 }
 121                                 in = inbuf;
 122                                 inbytes += ret;
 123                         }
 124                 }
 125         }
 126         /* reset the shift state of the output buffer */
 127         outbytes = OUTBUFSIZE;
 128         out = outbuf;
 129         ret = iconv(cd, NULL, NULL, &out, &outbytes);
 130         if (ret == (size_t)-1)
 131                 err(EXIT_FAILURE, "iconv()");
 132         if (outbytes < OUTBUFSIZE)
 133                 (void)fwrite(outbuf, 1, OUTBUFSIZE - outbytes, stdout);
 134
 135         if (invalids > 0 && !silent)
 136                 warnx("warning: invalid characters: %llu", invalids);
 137
 138         iconv_close(cd);
 139         return (invalids > 0);
 140 }
 141
 142 static int
 143 do_list(unsigned int n, const char * const *list, void *data __unused)
 144 {
 145         unsigned int i;
 146
 147         for(i = 0; i < n; i++) {
 148                 printf("%s", list[i]);
 149                 if (i < n - 1)
 150                         printf(" ");
 151         }
 152         printf("\n");
 153
 154         return (1);
 155 }
 156
 157 int
 158 main(int argc, char **argv)
 159 {
 160         FILE *fp;
 161         const char *opt_f, *opt_t;
 162         int ch, i, res;
 163         bool opt_c = false, opt_s = false;
 164
 165         opt_f = opt_t = "";
 166
 167         setlocale(LC_ALL, "");
 168         setprogname(argv[0]);
 169
 170         while ((ch = getopt_long(argc, argv, "csLlf:t:",
 171             long_options, NULL)) != -1) {
 172                 switch (ch) {
 173                 case 'c':
 174                         opt_c = true;
 175                         break;
 176                 case 's':
 177                         opt_s = true;
 178                         break;
 179                 case 'l':
 180                         /* list */
 181                         if (opt_s || opt_c || strcmp(opt_f, "") != 0 ||
 182                             strcmp(opt_t, "") != 0) {
 183                                 warnx("-l is not allowed with other flags.");
 184                                 usage();
 185                         }
 186                         iconvlist(do_list, NULL);
 187                         return (EXIT_SUCCESS);
 188                 case 'f':
 189                         /* from */
 190                         if (optarg != NULL)
 191                                 opt_f = optarg;
 192                         break;
 193                 case 't':
 194                         /* to */
 195                         if (optarg != NULL)
 196                                 opt_t = optarg;
 197                         break;
 198                 default:
 199                         usage();
 200                 }
 201         }
 202         argc -= optind;
 203         argv += optind;
 204         if ((strcmp(opt_f, "") == 0) && (strcmp(opt_t, "") == 0))
 205                 usage();
 206         if (argc == 0)
 207                 res = do_conv(stdin, opt_f, opt_t, opt_s, opt_c);
 208         else {
 209                 res = 0;
 210                 for (i = 0; i < argc; i++) {
 211                         fp = (strcmp(argv[i], "-") != 0) ?
 212                             fopen(argv[i], "r") : stdin;
 213                         if (fp == NULL)
 214                                 err(EXIT_FAILURE, "Cannot open `%s'",
 215                                     argv[i]);
 216                         res |= do_conv(fp, opt_f, opt_t, opt_s, opt_c);
 217                         (void)fclose(fp);
 218                 }
 219         }
 220         return (res == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
 221 }