gnu/dist/gettext/gettext-tools/src/urlget.c

   1 /* Get the contents of an URL.
   2    Copyright (C) 2001-2003, 2005 Free Software Foundation, Inc.
   3    Written by Bruno Haible <haible@clisp.cons.org>, 2001.
   4
   5    This program is free software; you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published by
   7    the Free Software Foundation; either version 2, or (at your option)
   8    any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program; if not, write to the Free Software Foundation,
  17    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  18
  19
  20 #ifdef HAVE_CONFIG_H
  21 # include "config.h"
  22 #endif
  23
  24 #include <errno.h>
  25 #include <fcntl.h>
  26 #include <getopt.h>
  27 #include <stdbool.h>
  28 #include <stdio.h>
  29 #include <stdlib.h>
  30 #include <locale.h>
  31
  32 #ifdef HAVE_UNISTD_H
  33 # include <unistd.h>
  34 #endif
  35
  36 #include "closeout.h"
  37 #include "error.h"
  38 #include "error-progname.h"
  39 #include "progname.h"
  40 #include "relocatable.h"
  41 #include "basename.h"
  42 #include "full-write.h"
  43 #include "execute.h"
  44 #include "javaexec.h"
  45 #include "exit.h"
  46 #include "binary-io.h"
  47 #include "gettext.h"
  48
  49 #define _(str) gettext (str)
  50
  51 #ifndef STDOUT_FILENO
  52 # define STDOUT_FILENO 1
  53 #endif
  54
  55
  56 /* Only high-level toolkits, written in languages with exception handling,
  57    have an URL datatype and operations to fetch an URL's contents.  Such
  58    toolkits are Java (class java.net.URL), Qt (classes QUrl and QUrlOperator).
  59    We use the Java toolkit.
  60    Note that this program doesn't handle redirection pages; programs which
  61    wish to process HTML redirection tags need to include a HTML parser,
  62    and only full-fledged browsers like w3m, lynx, links have have both
  63    an URL fetcher (which covers at least the protocols "http", "ftp", "file")
  64    and a HTML parser.  */
  65
  66
  67 /* Long options.  */
  68 static const struct option long_options[] =
  69 {
  70   { "help", no_argument, NULL, 'h' },
  71   { "version", no_argument, NULL, 'V' },
  72   { NULL, 0, NULL, 0 }
  73 };
  74
  75
  76 /* Forward declaration of local functions.  */
  77 static void usage (int status)
  78 #if defined __GNUC__ && ((__GNUC__ == 2 && __GNUC_MINOR__ >= 5) || __GNUC__ > 2)
  79      __attribute__ ((noreturn))
  80 #endif
  81 ;
  82 static void fetch (const char *url, const char *file);
  83
  84
  85 int
  86 main (int argc, char *argv[])
  87 {
  88   int optchar;
  89   bool do_help;
  90   bool do_version;
  91
  92   /* Set program name for messages.  */
  93   set_program_name (argv[0]);
  94   error_print_progname = maybe_print_progname;
  95
  96 #ifdef HAVE_SETLOCALE
  97   /* Set locale via LC_ALL.  */
  98   setlocale (LC_ALL, "");
  99 #endif
 100
 101   /* Set the text message domain.  */
 102   bindtextdomain (PACKAGE, relocate (LOCALEDIR));
 103   textdomain (PACKAGE);
 104
 105   /* Ensure that write errors on stdout are detected.  */
 106   atexit (close_stdout);
 107
 108   /* Set default values for variables.  */
 109   do_help = false;
 110   do_version = false;
 111
 112   /* Parse command line options.  */
 113   while ((optchar = getopt_long (argc, argv, "hV", long_options, NULL)) != EOF)
 114     switch (optchar)
 115     {
 116     case '\0':          /* Long option.  */
 117       break;
 118     case 'h':
 119       do_help = true;
 120       break;
 121     case 'V':
 122       do_version = true;
 123       break;
 124     default:
 125       usage (EXIT_FAILURE);
 126       /* NOTREACHED */
 127     }
 128
 129   /* Version information requested.  */
 130   if (do_version)
 131     {
 132       printf ("%s (GNU %s) %s\n", basename (program_name), PACKAGE, VERSION);
 133       /* xgettext: no-wrap */
 134       printf (_("Copyright (C) %s Free Software Foundation, Inc.\n\
 135 This is free software; see the source for copying conditions.  There is NO\n\
 136 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 137 "),
 138               "2001-2003");
 139       printf (_("Written by %s.\n"), "Bruno Haible");
 140       exit (EXIT_SUCCESS);
 141     }
 142
 143   /* Help is requested.  */
 144   if (do_help)
 145     usage (EXIT_SUCCESS);
 146
 147   /* Test argument count.  */
 148   if (optind + 2 != argc)
 149     error (EXIT_FAILURE, 0, _("expected two arguments"));
 150
 151   /* Fetch the contents.  */
 152   fetch (argv[optind], argv[optind + 1]);
 153
 154   exit (EXIT_SUCCESS);
 155 }
 156
 157 /* Display usage information and exit.  */
 158 static void
 159 usage (int status)
 160 {
 161   if (status != EXIT_SUCCESS)
 162     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 163              program_name);
 164   else
 165     {
 166       printf (_("\
 167 Usage: %s [OPTION] URL FILE\n\
 168 "), program_name);
 169       printf ("\n");
 170       /* xgettext: no-wrap */
 171       printf (_("\
 172 Fetches and outputs the contents of an URL.  If the URL cannot be accessed,\n\
 173 the locally accessible FILE is used instead.\n\
 174 "));
 175       printf ("\n");
 176       printf (_("\
 177 Informative output:\n"));
 178       printf (_("\
 179   -h, --help                  display this help and exit\n"));
 180       printf (_("\
 181   -V, --version               output version information and exit\n"));
 182       printf ("\n");
 183       fputs (_("Report bugs to <bug-gnu-gettext@gnu.org>.\n"),
 184              stdout);
 185     }
 186
 187   exit (status);
 188 }
 189
 190 /* Copy a file's contents to stdout.  */
 191 static void
 192 cat_file (const char *src_filename)
 193 {
 194   int src_fd;
 195   char buf[4096];
 196   const int buf_size = sizeof (buf);
 197
 198   src_fd = open (src_filename, O_RDONLY | O_BINARY);
 199   if (src_fd < 0)
 200     error (EXIT_FAILURE, errno, _("error while opening \"%s\" for reading"),
 201            src_filename);
 202
 203   for (;;)
 204     {
 205       ssize_t n_read = read (src_fd, buf, buf_size);
 206       if (n_read < 0)
 207         {
 208 #ifdef EINTR
 209           if (errno == EINTR)
 210             continue;
 211 #endif
 212           error (EXIT_FAILURE, errno, _("error reading \"%s\""), src_filename);
 213         }
 214       if (n_read == 0)
 215         break;
 216
 217       if (full_write (STDOUT_FILENO, buf, n_read) < n_read)
 218         error (EXIT_FAILURE, errno, _("error writing stdout"));
 219     }
 220
 221   if (close (src_fd) < 0)
 222     error (EXIT_FAILURE, errno, _("error after reading \"%s\""), src_filename);
 223 }
 224
 225 static bool
 226 execute_it (const char *progname,
 227             const char *prog_path, char **prog_argv,
 228             void *private_data)
 229 {
 230   (void) private_data;
 231
 232   return execute (progname, prog_path, prog_argv, true, true, false, false,
 233                   true, false)
 234          != 0;
 235 }
 236
 237 /* Fetch the URL.  Upon error, use the FILE as fallback.  */
 238 static void
 239 fetch (const char *url, const char *file)
 240 {
 241   /* First try: using Java.  */
 242   {
 243     const char *class_name = "gnu.gettext.GetURL";
 244     const char *gettextjexedir;
 245     const char *gettextjar;
 246     const char *args[2];
 247
 248 #if USEJEXE
 249     /* Make it possible to override the executable's location.  This is
 250        necessary for running the testsuite before "make install".  */
 251     gettextjexedir = getenv ("GETTEXTJEXEDIR");
 252     if (gettextjexedir == NULL || gettextjexedir[0] == '\0')
 253       gettextjexedir = relocate (GETTEXTJEXEDIR);
 254 #else
 255     gettextjexedir = NULL;
 256 #endif
 257
 258     /* Make it possible to override the gettext.jar location.  This is
 259        necessary for running the testsuite before "make install".  */
 260     gettextjar = getenv ("GETTEXTJAR");
 261     if (gettextjar == NULL || gettextjar[0] == '\0')
 262       gettextjar = relocate (GETTEXTJAR);
 263
 264     /* Prepare arguments.  */
 265     args[0] = url;
 266     args[1] = NULL;
 267
 268     /* Fetch the URL's contents.  */
 269     if (execute_java_class (class_name, &gettextjar, 1, true, gettextjexedir,
 270                             args,
 271                             false, true,
 272                             execute_it, NULL) == 0)
 273       return;
 274   }
 275
 276   /* Second try: using "wget -q -O - url".  */
 277   {
 278     static bool wget_tested;
 279     static bool wget_present;
 280
 281     if (!wget_tested)
 282       {
 283         /* Test for presence of wget: "wget --version > /dev/null"  */
 284         char *argv[3];
 285         int exitstatus;
 286
 287         argv[0] = "wget";
 288         argv[1] = "--version";
 289         argv[2] = NULL;
 290         exitstatus = execute ("wget", "wget", argv, false, false, true, true,
 291                               true, false);
 292         wget_present = (exitstatus == 0);
 293         wget_tested = true;
 294       }
 295
 296     if (wget_present)
 297       {
 298         char *argv[8];
 299         int exitstatus;
 300
 301         argv[0] = "wget";
 302         argv[1] = "-q";
 303         argv[2] = "-O"; argv[3] = "-";
 304         argv[4] = "-T"; argv[5] = "30";
 305         argv[6] = (char *) url;
 306         argv[7] = NULL;
 307         exitstatus = execute ("wget", "wget", argv, true, false, false, false,
 308                               true, false);
 309         if (exitstatus != 127)
 310           {
 311             if (exitstatus != 0)
 312               /* Use the file as fallback.  */
 313               cat_file (file);
 314             return;
 315           }
 316       }
 317   }
 318
 319   /* Third try: using "lynx -source url".  */
 320   {
 321     static bool lynx_tested;
 322     static bool lynx_present;
 323
 324     if (!lynx_tested)
 325       {
 326         /* Test for presence of lynx: "lynx --version > /dev/null"  */
 327         char *argv[3];
 328         int exitstatus;
 329
 330         argv[0] = "lynx";
 331         argv[1] = "--version";
 332         argv[2] = NULL;
 333         exitstatus = execute ("lynx", "lynx", argv, false, false, true, true,
 334                               true, false);
 335         lynx_present = (exitstatus == 0);
 336         lynx_tested = true;
 337       }
 338
 339     if (lynx_present)
 340       {
 341         char *argv[4];
 342         int exitstatus;
 343
 344         argv[0] = "lynx";
 345         argv[1] = "-source";
 346         argv[2] = (char *) url;
 347         argv[3] = NULL;
 348         exitstatus = execute ("lynx", "lynx", argv, true, false, false, false,
 349                               true, false);
 350         if (exitstatus != 127)
 351           {
 352             if (exitstatus != 0)
 353               /* Use the file as fallback.  */
 354               cat_file (file);
 355             return;
 356           }
 357       }
 358   }
 359
 360   /* Fourth try: using "curl --silent url".  */
 361   {
 362     static bool curl_tested;
 363     static bool curl_present;
 364
 365     if (!curl_tested)
 366       {
 367         /* Test for presence of curl: "curl --version > /dev/null"  */
 368         char *argv[3];
 369         int exitstatus;
 370
 371         argv[0] = "curl";
 372         argv[1] = "--version";
 373         argv[2] = NULL;
 374         exitstatus = execute ("curl", "curl", argv, false, false, true, true,
 375                               true, false);
 376         curl_present = (exitstatus == 0 || exitstatus == 2);
 377         curl_tested = true;
 378       }
 379
 380     if (curl_present)
 381       {
 382         char *argv[4];
 383         int exitstatus;
 384
 385         argv[0] = "curl";
 386         argv[1] = "--silent";
 387         argv[2] = (char *) url;
 388         argv[3] = NULL;
 389         exitstatus = execute ("curl", "curl", argv, true, false, false, false,
 390                               true, false);
 391         if (exitstatus != 127)
 392           {
 393             if (exitstatus != 0)
 394               /* Use the file as fallback.  */
 395               cat_file (file);
 396             return;
 397           }
 398       }
 399   }
 400
 401   /* Use the file as fallback.  */
 402   cat_file (file);
 403 }