1 /* Command line parsing.
2 Copyright (C) 1995, 1996, 1997, 1998, 2000, 2001
3 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget; if not, write to the Free Software
19 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
27 #endif /* HAVE_UNISTD_H */
28 #include <sys/types.h>
33 #endif /* HAVE_STRING_H */
40 #endif /* HAVE_LOCALE_H */
44 #define OPTIONS_DEFINED_HERE /* for options.h */
54 #include "progress.h" /* for progress_handle_sigwinch */
56 /* On GNU system this will include system-wide getopt.h. */
59 #ifndef PATH_SEPARATOR
60 # define PATH_SEPARATOR '/'
63 extern char *version_string;
72 void log_init PARAMS ((const char *, int));
73 void log_close PARAMS ((void));
74 void redirect_output PARAMS ((const char *));
76 static RETSIGTYPE redirect_output_signal PARAMS ((int));
78 const char *exec_name;
80 /* Initialize I18N. The initialization amounts to invoking
81 setlocale(), bindtextdomain() and textdomain().
82 Does nothing if NLS is disabled or missing. */
84 i18n_initialize (void)
86 /* If HAVE_NLS is defined, assume the existence of the three
87 functions invoked here. */
89 /* Set the current locale. */
90 /* Here we use LC_MESSAGES instead of LC_ALL, for two reasons.
91 First, message catalogs are all of I18N Wget uses anyway.
92 Second, setting LC_ALL has a dangerous potential of messing
93 things up. For example, when in a foreign locale, Solaris
94 strptime() fails to handle international dates correctly, which
95 makes http_atotm() malfunction. */
97 setlocale (LC_MESSAGES, "");
98 setlocale (LC_CTYPE, "");
100 setlocale (LC_ALL, "");
102 /* Set the text message domain. */
103 bindtextdomain ("wget", LOCALEDIR);
105 #endif /* HAVE_NLS */
108 /* Print the usage message. */
112 printf (_("Usage: %s [OPTION]... [URL]...\n"), exec_name);
115 /* Print the help message, describing all the available options. If
116 you add an option, be sure to update this list. */
120 printf (_("GNU Wget %s, a non-interactive network retriever.\n"),
123 /* Had to split this in parts, so the #@@#%# Ultrix compiler and cpp
124 don't bitch. Also, it makes translation much easier. */
127 Mandatory arguments to long options are mandatory for short options too.\n\
131 -V, --version display the version of Wget and exit.\n\
132 -h, --help print this help.\n\
133 -b, --background go to background after startup.\n\
134 -e, --execute=COMMAND execute a `.wgetrc\'-style command.\n\
137 Logging and input file:\n\
138 -o, --output-file=FILE log messages to FILE.\n\
139 -a, --append-output=FILE append messages to FILE.\n\
140 -d, --debug print debug output.\n\
141 -q, --quiet quiet (no output).\n\
142 -v, --verbose be verbose (this is the default).\n\
143 -nv, --non-verbose turn off verboseness, without being quiet.\n\
144 -i, --input-file=FILE download URLs found in FILE.\n\
145 -F, --force-html treat input file as HTML.\n\
146 -B, --base=URL prepends URL to relative links in -F -i file.\n\
147 --sslcertfile=FILE optional client certificate.\n\
148 --sslcertkey=KEYFILE optional keyfile for this certificate.\n\
149 --sslegdsock=FILE socket to the egd daemon(random data source).\n\
153 --bind-address=ADDRESS bind to ADDRESS (hostname or IP) on local host.\n\
154 -t, --tries=NUMBER set number of retries to NUMBER (0 unlimits).\n\
155 -O --output-document=FILE write documents to FILE.\n\
156 -nc, --no-clobber don\'t clobber existing files or use .# suffixes.\n\
157 -c, --continue resume getting a partially-downloaded file.\n\
158 --progress=TYPE select progress gauge type.\n\
159 -N, --timestamping don\'t re-retrieve files unless newer than local.\n\
160 -S, --server-response print server response.\n\
161 --spider don\'t download anything.\n\
162 -T, --timeout=SECONDS set the read timeout to SECONDS.\n\
163 -w, --wait=SECONDS wait SECONDS between retrievals.\n\
164 --waitretry=SECONDS wait 1...SECONDS between retries of a retrieval.\n\
165 --random-wait wait from 0...2*WAIT secs between retrievals.\n\
166 -Y, --proxy=on/off turn proxy on or off.\n\
167 -Q, --quota=NUMBER set retrieval quota to NUMBER.\n\
168 --limit-rate=RATE limit download rate to RATE.\n\
172 -nd --no-directories don\'t create directories.\n\
173 -x, --force-directories force creation of directories.\n\
174 -nH, --no-host-directories don\'t create host directories.\n\
175 -P, --directory-prefix=PREFIX save files to PREFIX/...\n\
176 --cut-dirs=NUMBER ignore NUMBER remote directory components.\n\
180 --http-user=USER set http user to USER.\n\
181 --http-passwd=PASS set http password to PASS.\n\
182 -C, --cache=on/off (dis)allow server-cached data (normally allowed).\n\
183 -E, --html-extension save all text/html documents with .html extension.\n\
184 --ignore-length ignore `Content-Length\' header field.\n\
185 --header=STRING insert STRING among the headers.\n\
186 --proxy-user=USER set USER as proxy username.\n\
187 --proxy-passwd=PASS set PASS as proxy password.\n\
188 --referer=URL include `Referer: URL\' header in HTTP request.\n\
189 -s, --save-headers save the HTTP headers to file.\n\
190 -U, --user-agent=AGENT identify as AGENT instead of Wget/VERSION.\n\
191 --no-http-keep-alive disable HTTP keep-alive (persistent connections).\n\
192 --cookies=off don't use cookies.\n\
193 --load-cookies=FILE load cookies from FILE before session.\n\
194 --save-cookies=FILE save cookies to FILE after session.\n\
198 -nr, --dont-remove-listing don\'t remove `.listing\' files.\n\
199 -g, --glob=on/off turn file name globbing on or off.\n\
200 --passive-ftp use the \"passive\" transfer mode.\n\
201 --retr-symlinks when recursing, get linked-to files (not dirs).\n\
204 Recursive retrieval:\n\
205 -r, --recursive recursive web-suck -- use with care!\n\
206 -l, --level=NUMBER maximum recursion depth (inf or 0 for infinite).\n\
207 --delete-after delete files locally after downloading them.\n\
208 -k, --convert-links convert non-relative links to relative.\n\
209 -K, --backup-converted before converting file X, back up as X.orig.\n\
210 -m, --mirror shortcut option equivalent to -r -N -l inf -nr.\n\
211 -p, --page-requisites get all images, etc. needed to display HTML page.\n\
214 Recursive accept/reject:\n\
215 -A, --accept=LIST comma-separated list of accepted extensions.\n\
216 -R, --reject=LIST comma-separated list of rejected extensions.\n\
217 -D, --domains=LIST comma-separated list of accepted domains.\n\
218 --exclude-domains=LIST comma-separated list of rejected domains.\n\
219 --follow-ftp follow FTP links from HTML documents.\n\
220 --follow-tags=LIST comma-separated list of followed HTML tags.\n\
221 -G, --ignore-tags=LIST comma-separated list of ignored HTML tags.\n\
222 -H, --span-hosts go to foreign hosts when recursive.\n\
223 -L, --relative follow relative links only.\n\
224 -I, --include-directories=LIST list of allowed directories.\n\
225 -X, --exclude-directories=LIST list of excluded directories.\n\
226 -np, --no-parent don\'t ascend to the parent directory.\n\
228 fputs (_("Mail bug reports and suggestions to <bug-wget@gnu.org>.\n"),
233 main (int argc, char *const *argv)
236 int i, c, nurl, status, append_to_log;
238 static struct option long_options[] =
240 /* Options without arguments: */
241 { "background", no_argument, NULL, 'b' },
242 { "backup-converted", no_argument, NULL, 'K' },
243 { "continue", no_argument, NULL, 'c' },
244 { "convert-links", no_argument, NULL, 'k' },
245 { "debug", no_argument, NULL, 'd' },
246 { "delete-after", no_argument, NULL, 136 },
247 { "dont-remove-listing", no_argument, NULL, 149 },
248 { "follow-ftp", no_argument, NULL, 142 },
249 { "force-directories", no_argument, NULL, 'x' },
250 { "force-hier", no_argument, NULL, 'x' }, /* obsolete */
251 { "force-html", no_argument, NULL, 'F'},
252 { "help", no_argument, NULL, 'h' },
253 { "html-extension", no_argument, NULL, 'E' },
254 { "ignore-length", no_argument, NULL, 138 },
255 { "mirror", no_argument, NULL, 'm' },
256 { "no-clobber", no_argument, NULL, 141 },
257 { "no-directories", no_argument, NULL, 147 },
258 { "no-host-directories", no_argument, NULL, 148 },
259 { "no-host-lookup", no_argument, NULL, 150 },
260 { "no-http-keep-alive", no_argument, NULL, 156 },
261 { "no-parent", no_argument, NULL, 133 },
262 { "non-verbose", no_argument, NULL, 146 },
263 { "passive-ftp", no_argument, NULL, 139 },
264 { "page-requisites", no_argument, NULL, 'p' },
265 { "quiet", no_argument, NULL, 'q' },
266 { "random-wait", no_argument, NULL, 165 },
267 { "recursive", no_argument, NULL, 'r' },
268 { "relative", no_argument, NULL, 'L' },
269 { "retr-symlinks", no_argument, NULL, 137 },
270 { "save-headers", no_argument, NULL, 's' },
271 { "server-response", no_argument, NULL, 'S' },
272 { "span-hosts", no_argument, NULL, 'H' },
273 { "spider", no_argument, NULL, 132 },
274 { "timestamping", no_argument, NULL, 'N' },
275 { "verbose", no_argument, NULL, 'v' },
276 { "version", no_argument, NULL, 'V' },
278 /* Options accepting an argument: */
279 { "accept", required_argument, NULL, 'A' },
280 { "append-output", required_argument, NULL, 'a' },
281 { "backups", required_argument, NULL, 151 }, /* undocumented */
282 { "base", required_argument, NULL, 'B' },
283 { "bind-address", required_argument, NULL, 155 },
284 { "cache", required_argument, NULL, 'C' },
285 { "cookies", required_argument, NULL, 160 },
286 { "cut-dirs", required_argument, NULL, 145 },
287 { "directory-prefix", required_argument, NULL, 'P' },
288 { "domains", required_argument, NULL, 'D' },
289 { "dot-style", required_argument, NULL, 134 },
290 { "execute", required_argument, NULL, 'e' },
291 { "exclude-directories", required_argument, NULL, 'X' },
292 { "exclude-domains", required_argument, NULL, 140 },
293 { "follow-tags", required_argument, NULL, 153 },
294 { "glob", required_argument, NULL, 'g' },
295 { "header", required_argument, NULL, 131 },
296 { "htmlify", required_argument, NULL, 135 },
297 { "http-passwd", required_argument, NULL, 130 },
298 { "http-user", required_argument, NULL, 129 },
299 { "ignore-tags", required_argument, NULL, 'G' },
300 { "include-directories", required_argument, NULL, 'I' },
301 { "input-file", required_argument, NULL, 'i' },
302 { "level", required_argument, NULL, 'l' },
303 { "limit-rate", required_argument, NULL, 164 },
304 { "load-cookies", required_argument, NULL, 161 },
305 { "no", required_argument, NULL, 'n' },
306 { "output-document", required_argument, NULL, 'O' },
307 { "output-file", required_argument, NULL, 'o' },
308 { "progress", required_argument, NULL, 163 },
309 { "proxy", required_argument, NULL, 'Y' },
310 { "proxy-passwd", required_argument, NULL, 144 },
311 { "proxy-user", required_argument, NULL, 143 },
312 { "quota", required_argument, NULL, 'Q' },
313 { "reject", required_argument, NULL, 'R' },
314 { "save-cookies", required_argument, NULL, 162 },
315 { "timeout", required_argument, NULL, 'T' },
316 { "tries", required_argument, NULL, 't' },
317 { "user-agent", required_argument, NULL, 'U' },
318 { "referer", required_argument, NULL, 157 },
319 { "use-proxy", required_argument, NULL, 'Y' },
321 { "sslcertfile", required_argument, NULL, 158 },
322 { "sslcertkey", required_argument, NULL, 159 },
323 { "sslegdsock", required_argument, NULL, 166 },
324 #endif /* HAVE_SSL */
325 { "wait", required_argument, NULL, 'w' },
326 { "waitretry", required_argument, NULL, 152 },
334 /* Construct the name of the executable, without the directory part. */
335 exec_name = strrchr (argv[0], PATH_SEPARATOR);
342 windows_main_junk (&argc, (char **) argv, (char **) &exec_name);
345 initialize (); /* sets option defaults; reads the system wgetrc and .wgetrc */
347 /* [Is the order of the option letters significant? If not, they should be
348 alphabetized, like the long_options. The only thing I know for sure is
349 that the options with required arguments must be followed by a ':'.
350 -- Dan Harkless <wget@harkless.org>] */
351 while ((c = getopt_long (argc, argv, "\
352 hpVqvdkKsxmNWrHSLcFbEY:G:g:T:U:O:l:n:i:o:a:t:D:A:R:P:B:e:Q:X:I:w:C:",
353 long_options, (int *)0)) != EOF)
357 /* Options without arguments: */
359 setval ("spider", "on");
362 setval ("noparent", "on");
365 setval ("deleteafter", "on");
368 setval ("retrsymlinks", "on");
371 setval ("ignorelength", "on");
374 setval ("passiveftp", "on");
377 setval ("noclobber", "on");
380 setval ("followftp", "on");
383 setval ("cutdirs", optarg);
386 setval ("verbose", "off");
389 setval ("dirstruct", "off");
392 setval ("addhostdir", "off");
395 setval ("removelisting", "off");
398 setval ("bindaddress", optarg);
401 setval ("httpkeepalive", "off");
404 setval ("randomwait", "on");
407 setval ("background", "on");
410 setval ("continue", "on");
414 setval ("debug", "on");
415 #else /* not DEBUG */
416 fprintf (stderr, _("%s: debug support not compiled in.\n"),
418 #endif /* not DEBUG */
421 setval ("htmlextension", "on");
424 setval ("forcehtml", "on");
427 setval ("spanhosts", "on");
437 setval ("backupconverted", "on");
440 setval ("convertlinks", "on");
443 setval ("relativeonly", "on");
446 setval ("mirror", "on");
449 setval ("timestamping", "on");
452 setval ("pagerequisites", "on");
455 setval ("serverresponse", "on");
458 setval ("saveheaders", "on");
461 setval ("quiet", "on");
464 setval ("recursive", "on");
467 printf ("GNU Wget %s\n\n", version_string);
469 Copyright (C) 1995, 1996, 1997, 1998, 2000, 2001 Free Software Foundation, Inc.\n"));
471 This program is distributed in the hope that it will be useful,\n\
472 but WITHOUT ANY WARRANTY; without even the implied warranty of\n\
473 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n\
474 GNU General Public License for more details.\n"));
475 printf (_("\nOriginally written by Hrvoje Niksic <hniksic@arsdigita.com>.\n"));
479 setval ("verbose", "on");
482 setval ("dirstruct", "on");
485 /* Options accepting an argument: */
487 setval ("httpuser", optarg);
490 setval ("httppasswd", optarg);
493 setval ("header", optarg);
496 setval ("dotstyle", optarg);
499 setval ("htmlify", optarg);
502 setval ("excludedomains", optarg);
505 setval ("proxyuser", optarg);
508 setval ("proxypasswd", optarg);
511 setval ("backups", optarg);
514 setval ("waitretry", optarg);
517 setval ("followtags", optarg);
520 setval ("cookies", optarg);
523 setval ("loadcookies", optarg);
526 setval ("savecookies", optarg);
529 setval ("progress", optarg);
532 setval ("limitrate", optarg);
535 setval ("referer", optarg);
539 setval ("sslcertfile", optarg);
542 setval ("sslcertkey", optarg);
545 setval ("sslegdsock", optarg);
547 #endif /* HAVE_SSL */
549 setval ("accept", optarg);
552 setval ("logfile", optarg);
556 setval ("base", optarg);
559 setval ("cache", optarg);
562 setval ("domains", optarg);
567 if (parse_line (optarg, &com, &val))
569 if (!setval (com, val))
574 fprintf (stderr, _("%s: %s: invalid command\n"), exec_name,
583 setval ("ignoretags", optarg);
586 setval ("glob", optarg);
589 setval ("includedirectories", optarg);
592 setval ("input", optarg);
595 setval ("reclevel", optarg);
599 /* #### What we really want here is --no-foo. */
602 for (p = optarg; *p; p++)
606 setval ("verbose", "off");
609 setval ("addhostdir", "off");
612 setval ("dirstruct", "off");
615 setval ("noclobber", "on");
618 setval ("removelisting", "off");
621 setval ("noparent", "on");
624 setval ("httpkeepalive", "off");
627 printf (_("%s: illegal option -- `-n%c'\n"), exec_name, *p);
630 printf (_("Try `%s --help\' for more options.\n"), exec_name);
636 setval ("outputdocument", optarg);
639 setval ("logfile", optarg);
642 setval ("dirprefix", optarg);
645 setval ("quota", optarg);
648 setval ("reject", optarg);
651 setval ("timeout", optarg);
654 setval ("tries", optarg);
657 setval ("useragent", optarg);
660 setval ("wait", optarg);
663 setval ("excludedirectories", optarg);
666 setval ("useproxy", optarg);
672 printf (_("Try `%s --help' for more options.\n"), exec_name);
678 /* Initialize progress. Have to do this after the options are
679 processed so we know where the log file is. */
681 set_progress_implementation (opt.progress_type);
683 /* All user options have now been processed, so it's now safe to do
684 interoption dependency checks. */
686 if (opt.reclevel == 0)
687 opt.reclevel = INFINITE_RECURSION; /* see wget.h for commentary on this */
689 if (opt.page_requisites && !opt.recursive)
691 opt.recursive = TRUE;
693 if (!opt.no_dirstruct)
694 opt.dirstruct = TRUE; /* usually handled by cmd_spec_recursive() */
697 if (opt.verbose == -1)
698 opt.verbose = !opt.quiet;
701 if (opt.verbose && opt.quiet)
703 printf (_("Can't be verbose and quiet at the same time.\n"));
707 if (opt.timestamping && opt.noclobber)
710 Can't timestamp and not clobber old files at the same time.\n"));
714 nurl = argc - optind;
715 if (!nurl && !opt.input_filename)
717 /* No URL specified. */
718 printf (_("%s: missing URL\n"), exec_name);
721 /* #### Something nicer should be printed here -- similar to the
722 pre-1.5 `--help' page. */
723 printf (_("Try `%s --help' for more options.\n"), exec_name);
728 fork_to_background ();
730 /* Allocate basic pointer. */
731 url = ALLOCA_ARRAY (char *, nurl + 1);
732 /* Fill in the arguments. */
733 for (i = 0; i < nurl; i++, optind++)
735 char *rewritten = rewrite_shorthand_url (argv[optind]);
738 printf ("Converted %s to %s\n", argv[optind], rewritten);
742 url[i] = xstrdup (argv[optind]);
746 /* Change the title of console window on Windows. #### I think this
747 statement should belong to retrieve_url(). --hniksic. */
749 ws_changetitle (*url, nurl);
752 /* Initialize logging. */
753 log_init (opt.lfilename, append_to_log);
755 DEBUGP (("DEBUG output created by Wget %s on %s.\n\n", version_string,
758 /* Open the output filename if necessary. */
759 if (opt.output_document)
761 if (HYPHENP (opt.output_document))
766 opt.dfp = fopen (opt.output_document, opt.always_rest ? "ab" : "wb");
769 perror (opt.output_document);
772 if (fstat (fileno (opt.dfp), &st) == 0 && S_ISREG (st.st_mode))
773 opt.od_known_regular = 1;
781 /* Setup the signal handler to redirect output when hangup is
784 if (signal(SIGHUP, SIG_IGN) != SIG_IGN)
785 signal(SIGHUP, redirect_output_signal);
786 /* ...and do the same for SIGUSR1. */
787 signal (SIGUSR1, redirect_output_signal);
788 /* Writing to a closed socket normally signals SIGPIPE, and the
789 process exits. What we want is to ignore SIGPIPE and just check
790 for the return value of write(). */
791 signal (SIGPIPE, SIG_IGN);
793 signal (SIGWINCH, progress_handle_sigwinch);
795 #endif /* HAVE_SIGNAL */
797 status = RETROK; /* initialize it, just-in-case */
798 /*recursive_reset ();*/
799 /* Retrieve the URLs from argument list. */
800 for (t = url; *t; t++)
802 char *filename = NULL, *redirected_URL = NULL;
805 if (opt.recursive && url_scheme (*t) != SCHEME_FTP)
806 status = retrieve_tree (*t);
808 status = retrieve_url (*t, &filename, &redirected_URL, NULL, &dt);
810 if (opt.delete_after && file_exists_p(filename))
812 DEBUGP (("Removing file due to --delete-after in main():\n"));
813 logprintf (LOG_VERBOSE, _("Removing %s.\n"), filename);
814 if (unlink (filename))
815 logprintf (LOG_NOTQUIET, "unlink: %s\n", strerror (errno));
818 FREE_MAYBE (redirected_URL);
819 FREE_MAYBE (filename);
822 /* And then from the input file, if any. */
823 if (opt.input_filename)
826 status = retrieve_from_file (opt.input_filename, opt.force_html, &count);
828 logprintf (LOG_NOTQUIET, _("No URLs found in %s.\n"),
831 /* Print the downloaded sum. */
834 || (opt.input_filename && opt.downloaded != 0))
836 logprintf (LOG_NOTQUIET,
837 _("\nFINISHED --%s--\nDownloaded: %s bytes in %d files\n"),
839 (opt.downloaded_overflow ?
840 "<overflow>" : legible_very_long (opt.downloaded)),
842 /* Print quota warning, if exceeded. */
843 if (downloaded_exceeds_quota ())
844 logprintf (LOG_NOTQUIET,
845 _("Download quota (%s bytes) EXCEEDED!\n"),
846 legible (opt.quota));
849 if (opt.cookies_output)
850 save_cookies (opt.cookies_output);
852 if (opt.convert_links && !opt.delete_after)
854 convert_all_links ();
858 for (i = 0; i < nurl; i++)
863 print_malloc_debug_stats ();
865 if (status == RETROK)
872 /* Hangup signal handler. When wget receives SIGHUP or SIGUSR1, it
873 will proceed operation as usual, trying to write into a log file.
874 If that is impossible, the output will be turned off. */
877 redirect_output_signal (int sig)
880 signal (sig, redirect_output_signal);
881 /* Please note that the double `%' in `%%s' is intentional, because
882 redirect_output passes tmp through printf. */
883 sprintf (tmp, _("%s received, redirecting output to `%%s'.\n"),
884 (sig == SIGHUP ? "SIGHUP" :
885 (sig == SIGUSR1 ? "SIGUSR1" :
887 redirect_output (tmp);
889 #endif /* HAVE_SIGNAL */