1 /* Command line parsing.
2 Copyright (C) 1995, 1996, 1997, 1998, 2000 Free Software Foundation, Inc.
4 This file is part of Wget.
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
11 This program 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
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software
18 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 */
43 #define OPTIONS_DEFINED_HERE /* for options.h */
53 #ifndef PATH_SEPARATOR
54 # define PATH_SEPARATOR '/'
57 extern char *version_string;
66 void log_init PARAMS ((const char *, int));
67 void log_close PARAMS ((void));
68 void redirect_output PARAMS ((const char *));
70 static RETSIGTYPE redirect_output_signal PARAMS ((int));
72 const char *exec_name;
74 /* Initialize I18N. The initialization amounts to invoking
75 setlocale(), bindtextdomain() and textdomain().
76 Does nothing if NLS is disabled or missing. */
78 i18n_initialize (void)
80 /* If HAVE_NLS is defined, assume the existence of the three
81 functions invoked here. */
83 /* Set the current locale. */
84 /* Here we use LC_MESSAGES instead of LC_ALL, for two reasons.
85 First, message catalogs are all of I18N Wget uses anyway.
86 Second, setting LC_ALL has a dangerous potential of messing
87 things up. For example, when in a foreign locale, Solaris
88 strptime() fails to handle international dates correctly, which
89 makes http_atotm() malfunction. */
91 setlocale (LC_MESSAGES, "");
93 setlocale (LC_ALL, "");
95 /* Set the text message domain. */
96 bindtextdomain ("wget", LOCALEDIR);
101 /* It's kosher to declare these here because their interface _has_ to
102 be void foo(void). */
103 void url_init PARAMS ((void));
104 void host_init PARAMS ((void));
106 /* This just calls the various initialization functions from the
107 modules that need one-time initialization. */
109 private_initialize (void)
115 /* Print the usage message. */
119 printf (_("Usage: %s [OPTION]... [URL]...\n"), exec_name);
122 /* Print the help message, describing all the available options. If
123 you add an option, be sure to update this list. */
127 printf (_("GNU Wget %s, a non-interactive network retriever.\n"),
130 /* Had to split this in parts, so the #@@#%# Ultrix compiler and cpp
131 don't bitch. Also, it makes translation much easier. */
132 printf ("%s%s%s%s%s%s%s%s%s%s", _("\
134 Mandatory arguments to long options are mandatory for short options too.\n\
137 -V, --version display the version of Wget and exit.\n\
138 -h, --help print this help.\n\
139 -b, --background go to background after startup.\n\
140 -e, --execute=COMMAND execute a `.wgetrc\'-style command.\n\
142 Logging and input file:\n\
143 -o, --output-file=FILE log messages to FILE.\n\
144 -a, --append-output=FILE append messages to FILE.\n\
145 -d, --debug print debug output.\n\
146 -q, --quiet quiet (no output).\n\
147 -v, --verbose be verbose (this is the default).\n\
148 -nv, --non-verbose turn off verboseness, without being quiet.\n\
149 -i, --input-file=FILE download URLs found in FILE.\n\
150 -F, --force-html treat input file as HTML.\n\
151 -B, --base=URL prepends URL to relative links in -F -i file.\n\
154 --bind-address=ADDRESS bind to ADDRESS (hostname or IP) on local host.\n\
155 -t, --tries=NUMBER set number of retries to NUMBER (0 unlimits).\n\
156 -O --output-document=FILE write documents to FILE.\n\
157 -nc, --no-clobber don\'t clobber existing files or use .# suffixes.\n\
158 -c, --continue restart getting an existing file.\n\
159 --dot-style=STYLE set retrieval display style.\n\
160 -N, --timestamping don\'t retrieve files if older than local.\n\
161 -S, --server-response print server response.\n\
162 --spider don\'t download anything.\n\
163 -T, --timeout=SECONDS set the read timeout to SECONDS.\n\
164 -w, --wait=SECONDS wait SECONDS between retrievals.\n\
165 --waitretry=SECONDS wait 1...SECONDS between retries of a retrieval.\n\
166 -Y, --proxy=on/off turn proxy on or off.\n\
167 -Q, --quota=NUMBER set retrieval quota to NUMBER.\n\
170 -nd --no-directories don\'t create directories.\n\
171 -x, --force-directories force creation of directories.\n\
172 -nH, --no-host-directories don\'t create host directories.\n\
173 -P, --directory-prefix=PREFIX save files to PREFIX/...\n\
174 --cut-dirs=NUMBER ignore NUMBER remote directory components.\n\
177 --http-user=USER set http user to USER.\n\
178 --http-passwd=PASS set http password to PASS.\n\
179 -C, --cache=on/off (dis)allow server-cached data (normally allowed).\n\
180 -E, --html-extension save all text/html documents with .html extension.\n\
181 --ignore-length ignore `Content-Length\' header field.\n\
182 --header=STRING insert STRING among the headers.\n\
183 --proxy-user=USER set USER as proxy username.\n\
184 --proxy-passwd=PASS set PASS as proxy password.\n\
185 --referer=URL include `Referer: URL\' header in HTTP request.\n\
186 -s, --save-headers save the HTTP headers to file.\n\
187 -U, --user-agent=AGENT identify as AGENT instead of Wget/VERSION.\n\
190 --retr-symlinks when recursing, retrieve linked-to files (not dirs).\n\
191 -g, --glob=on/off turn file name globbing on or off.\n\
192 --passive-ftp use the \"passive\" transfer mode.\n\
194 Recursive retrieval:\n\
195 -r, --recursive recursive web-suck -- use with care!.\n\
196 -l, --level=NUMBER maximum recursion depth (inf or 0 for infinite).\n\
197 --delete-after delete files locally after downloading them.\n\
198 -k, --convert-links convert non-relative links to relative.\n\
199 -K, --backup-converted before converting file X, back up as X.orig.\n\
200 -m, --mirror shortcut option equivalent to -r -N -l inf -nr.\n\
201 -nr, --dont-remove-listing don\'t remove `.listing\' files.\n\
202 -p, --page-requisites get all images, etc. needed to display HTML page.\n\
204 Recursive accept/reject:\n\
205 -A, --accept=LIST comma-separated list of accepted extensions.\n\
206 -R, --reject=LIST comma-separated list of rejected extensions.\n\
207 -D, --domains=LIST comma-separated list of accepted domains.\n\
208 --exclude-domains=LIST comma-separated list of rejected domains.\n\
209 --follow-ftp follow FTP links from HTML documents.\n\
210 --follow-tags=LIST comma-separated list of followed HTML tags.\n\
211 -G, --ignore-tags=LIST comma-separated list of ignored HTML tags.\n\
212 -H, --span-hosts go to foreign hosts when recursive.\n\
213 -L, --relative follow relative links only.\n\
214 -I, --include-directories=LIST list of allowed directories.\n\
215 -X, --exclude-directories=LIST list of excluded directories.\n\
216 -nh, --no-host-lookup don\'t DNS-lookup hosts.\n\
217 -np, --no-parent don\'t ascend to the parent directory.\n\
218 \n"), _("Mail bug reports and suggestions to <bug-wget@gnu.org>.\n"));
222 main (int argc, char *const *argv)
225 int i, c, nurl, status, append_to_log;
228 static struct option long_options[] =
230 /* Options without arguments: */
231 { "background", no_argument, NULL, 'b' },
232 { "continue", no_argument, NULL, 'c' },
233 { "convert-links", no_argument, NULL, 'k' },
234 { "backup-converted", no_argument, NULL, 'K' },
235 { "debug", no_argument, NULL, 'd' },
236 { "dont-remove-listing", no_argument, NULL, 21 },
237 { "email-address", no_argument, NULL, 26 }, /* undocumented (debug) */
238 { "follow-ftp", no_argument, NULL, 14 },
239 { "force-directories", no_argument, NULL, 'x' },
240 { "force-hier", no_argument, NULL, 'x' }, /* obsolete */
241 { "force-html", no_argument, NULL, 'F'},
242 { "help", no_argument, NULL, 'h' },
243 { "html-extension", no_argument, NULL, 'E' },
244 { "ignore-length", no_argument, NULL, 10 },
245 { "mirror", no_argument, NULL, 'm' },
246 { "no-clobber", no_argument, NULL, 13 },
247 { "no-directories", no_argument, NULL, 19 },
248 { "no-host-directories", no_argument, NULL, 20 },
249 { "no-host-lookup", no_argument, NULL, 22 },
250 { "no-http-keep-alive", no_argument, NULL, 28 },
251 { "no-parent", no_argument, NULL, 5 },
252 { "non-verbose", no_argument, NULL, 18 },
253 { "passive-ftp", no_argument, NULL, 11 },
254 { "page-requisites", no_argument, NULL, 'p' },
255 { "quiet", no_argument, NULL, 'q' },
256 { "recursive", no_argument, NULL, 'r' },
257 { "relative", no_argument, NULL, 'L' },
258 { "retr-symlinks", no_argument, NULL, 9 },
259 { "save-headers", no_argument, NULL, 's' },
260 { "server-response", no_argument, NULL, 'S' },
261 { "span-hosts", no_argument, NULL, 'H' },
262 { "spider", no_argument, NULL, 4 },
263 { "timestamping", no_argument, NULL, 'N' },
264 { "verbose", no_argument, NULL, 'v' },
265 { "version", no_argument, NULL, 'V' },
267 /* Options accepting an argument: */
268 { "accept", required_argument, NULL, 'A' },
269 { "append-output", required_argument, NULL, 'a' },
270 { "backups", required_argument, NULL, 23 }, /* undocumented */
271 { "base", required_argument, NULL, 'B' },
272 { "bind-address", required_argument, NULL, 27 },
273 { "cache", required_argument, NULL, 'C' },
274 { "cut-dirs", required_argument, NULL, 17 },
275 { "delete-after", no_argument, NULL, 8 },
276 { "directory-prefix", required_argument, NULL, 'P' },
277 { "domains", required_argument, NULL, 'D' },
278 { "dot-style", required_argument, NULL, 6 },
279 { "execute", required_argument, NULL, 'e' },
280 { "exclude-directories", required_argument, NULL, 'X' },
281 { "exclude-domains", required_argument, NULL, 12 },
282 { "follow-tags", required_argument, NULL, 25 },
283 { "glob", required_argument, NULL, 'g' },
284 { "header", required_argument, NULL, 3 },
285 { "htmlify", required_argument, NULL, 7 },
286 { "http-passwd", required_argument, NULL, 2 },
287 { "http-user", required_argument, NULL, 1 },
288 { "ignore-tags", required_argument, NULL, 'G' },
289 { "include-directories", required_argument, NULL, 'I' },
290 { "input-file", required_argument, NULL, 'i' },
291 { "level", required_argument, NULL, 'l' },
292 { "no", required_argument, NULL, 'n' },
293 { "output-document", required_argument, NULL, 'O' },
294 { "output-file", required_argument, NULL, 'o' },
295 { "proxy", required_argument, NULL, 'Y' },
296 { "proxy-passwd", required_argument, NULL, 16 },
297 { "proxy-user", required_argument, NULL, 15 },
298 { "quota", required_argument, NULL, 'Q' },
299 { "reject", required_argument, NULL, 'R' },
300 { "timeout", required_argument, NULL, 'T' },
301 { "tries", required_argument, NULL, 't' },
302 { "user-agent", required_argument, NULL, 'U' },
303 { "referer", required_argument, NULL, 129 },
304 { "use-proxy", required_argument, NULL, 'Y' },
305 { "wait", required_argument, NULL, 'w' },
306 { "waitretry", required_argument, NULL, 24 },
311 private_initialize ();
315 /* Construct the name of the executable, without the directory part. */
316 exec_name = strrchr (argv[0], PATH_SEPARATOR);
323 windows_main_junk (&argc, (char **) argv, (char **) &exec_name);
326 initialize (); /* sets option defaults; reads the system wgetrc and .wgetrc */
328 /* [Is the order of the option letters significant? If not, they should be
329 alphabetized, like the long_options. The only thing I know for sure is
330 that the options with required arguments must be followed by a ':'.
331 -- Dan Harkless <dan-wget@dilvish.speed.net>] */
332 while ((c = getopt_long (argc, argv, "\
333 hpVqvdkKsxmNWrHSLcFbEY:G:g:T:U:O:l:n:i:o:a:t:D:A:R:P:B:e:Q:X:I:w:",
334 long_options, (int *)0)) != EOF)
338 /* Options without arguments: */
340 setval ("spider", "on");
343 setval ("noparent", "on");
346 setval ("deleteafter", "on");
349 setval ("retrsymlinks", "on");
352 setval ("ignorelength", "on");
355 setval ("passiveftp", "on");
358 setval ("noclobber", "on");
361 setval ("followftp", "on");
364 setval ("cutdirs", optarg);
367 setval ("verbose", "off");
370 setval ("dirstruct", "off");
373 setval ("addhostdir", "off");
376 setval ("removelisting", "off");
379 setval ("simplehostcheck", "on");
382 /* For debugging purposes. */
383 printf ("%s\n", ftp_getaddress ());
387 setval ("bindaddress", optarg);
390 setval ("httpkeepalive", "off");
393 setval ("background", "on");
396 setval ("continue", "on");
400 setval ("debug", "on");
401 #else /* not DEBUG */
402 fprintf (stderr, _("%s: debug support not compiled in.\n"),
404 #endif /* not DEBUG */
407 setval ("htmlextension", "on");
410 setval ("forcehtml", "on");
413 setval ("spanhosts", "on");
423 setval ("backupconverted", "on");
426 setval ("convertlinks", "on");
429 setval ("relativeonly", "on");
432 setval ("mirror", "on");
435 setval ("timestamping", "on");
438 setval ("pagerequisites", "on");
441 setval ("serverresponse", "on");
444 setval ("saveheaders", "on");
447 setval ("quiet", "on");
450 setval ("recursive", "on");
453 printf ("GNU Wget %s\n\n", version_string);
455 Copyright (C) 1995, 1996, 1997, 1998, 2000 Free Software Foundation, Inc.\n\
456 This program is distributed in the hope that it will be useful,\n\
457 but WITHOUT ANY WARRANTY; without even the implied warranty of\n\
458 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the\n\
459 GNU General Public License for more details.\n"));
460 printf (_("\nOriginally written by Hrvoje Niksic <hniksic@arsdigita.com>.\n"));
464 setval ("verbose", "on");
467 setval ("dirstruct", "on");
470 /* Options accepting an argument: */
472 setval ("httpuser", optarg);
475 setval ("httppasswd", optarg);
478 setval ("header", optarg);
481 setval ("dotstyle", optarg);
484 setval ("htmlify", optarg);
487 setval ("excludedomains", optarg);
490 setval ("proxyuser", optarg);
493 setval ("proxypasswd", optarg);
496 setval ("backups", optarg);
499 setval ("waitretry", optarg);
503 setval ("followtags", optarg);
506 setval ("referer", optarg);
509 setval ("accept", optarg);
512 setval ("logfile", optarg);
516 setval ("base", optarg);
519 setval ("cache", optarg);
522 setval ("domains", optarg);
527 if (parse_line (optarg, &com, &val))
529 if (!setval (com, val))
534 fprintf (stderr, _("%s: %s: invalid command\n"), exec_name,
543 setval ("ignoretags", optarg);
546 setval ("glob", optarg);
549 setval ("includedirectories", optarg);
552 setval ("input", optarg);
555 setval ("reclevel", optarg);
559 /* #### The n? options are utter crock! */
562 for (p = optarg; *p; p++)
566 setval ("verbose", "off");
569 setval ("simplehostcheck", "on");
572 setval ("addhostdir", "off");
575 setval ("dirstruct", "off");
578 setval ("noclobber", "on");
581 setval ("removelisting", "off");
584 setval ("noparent", "on");
587 setval ("httpkeepalive", "off");
590 printf (_("%s: illegal option -- `-n%c'\n"), exec_name, *p);
593 printf (_("Try `%s --help\' for more options.\n"), exec_name);
599 setval ("outputdocument", optarg);
602 setval ("logfile", optarg);
605 setval ("dirprefix", optarg);
608 setval ("quota", optarg);
611 setval ("reject", optarg);
614 setval ("timeout", optarg);
617 setval ("tries", optarg);
620 setval ("useragent", optarg);
623 setval ("wait", optarg);
626 setval ("excludedirectories", optarg);
629 setval ("useproxy", optarg);
635 printf (_("Try `%s --help' for more options.\n"), exec_name);
641 /* All user options have now been processed, so it's now safe to do
642 interoption dependency checks. */
644 if (opt.reclevel == 0)
645 opt.reclevel = INFINITE_RECURSION; /* see wget.h for commentary on this */
647 if (opt.page_requisites && !opt.recursive)
649 opt.recursive = TRUE;
651 if (!opt.no_dirstruct)
652 opt.dirstruct = TRUE; /* usually handled by cmd_spec_recursive() */
655 if (opt.verbose == -1)
656 opt.verbose = !opt.quiet;
658 /* Retain compatibility with previous scripts.
659 if wait has been set, but waitretry has not, give it the wait value.
660 A simple check on the values is not enough, I could have set
661 wait to n>0 and waitretry to 0 - HEH */
664 char opt_wait_str[256]; /* bigger than needed buf to prevent overflow */
666 sprintf(opt_wait_str, "%ld", opt.wait);
667 setval ("waitretry", opt_wait_str);
671 if (opt.verbose && opt.quiet)
673 printf (_("Can't be verbose and quiet at the same time.\n"));
677 if (opt.timestamping && opt.noclobber)
680 Can't timestamp and not clobber old files at the same time.\n"));
684 nurl = argc - optind;
685 if (!nurl && !opt.input_filename)
687 /* No URL specified. */
688 printf (_("%s: missing URL\n"), exec_name);
691 /* #### Something nicer should be printed here -- similar to the
692 pre-1.5 `--help' page. */
693 printf (_("Try `%s --help' for more options.\n"), exec_name);
698 fork_to_background ();
700 /* Allocate basic pointer. */
701 url = ALLOCA_ARRAY (char *, nurl + 1);
702 /* Fill in the arguments. */
703 for (i = 0; i < nurl; i++, optind++)
705 char *irix4_cc_needs_this;
706 STRDUP_ALLOCA (irix4_cc_needs_this, argv[optind]);
707 url[i] = irix4_cc_needs_this;
711 /* Change the title of console window on Windows. #### I think this
712 statement should belong to retrieve_url(). --hniksic. */
714 ws_changetitle (*url, nurl);
717 /* Initialize logging. */
718 log_init (opt.lfilename, append_to_log);
720 DEBUGP (("DEBUG output created by Wget %s on %s.\n\n", version_string,
722 /* Open the output filename if necessary. */
723 if (opt.output_document)
725 if (HYPHENP (opt.output_document))
729 opt.dfp = fopen (opt.output_document, "wb");
732 perror (opt.output_document);
742 /* Setup the signal handler to redirect output when hangup is
745 if (signal(SIGHUP, SIG_IGN) != SIG_IGN)
746 signal(SIGHUP, redirect_output_signal);
747 /* ...and do the same for SIGUSR1. */
748 signal (SIGUSR1, redirect_output_signal);
749 /* Writing to a closed socket normally signals SIGPIPE, and the
750 process exits. What we want is to ignore SIGPIPE and just check
751 for the return value of write(). */
752 signal (SIGPIPE, SIG_IGN);
753 #endif /* HAVE_SIGNAL */
755 status = RETROK; /* initialize it, just-in-case */
757 /* Retrieve the URLs from argument list. */
758 for (t = url; *t; t++)
760 char *filename, *redirected_URL;
763 status = retrieve_url (*t, &filename, &redirected_URL, NULL, &dt);
764 if (opt.recursive && status == RETROK && (dt & TEXTHTML))
765 status = recursive_retrieve (filename,
766 redirected_URL ? redirected_URL : *t);
768 if (opt.delete_after && file_exists_p(filename))
770 DEBUGP (("Removing file due to --delete-after in main():\n"));
771 logprintf (LOG_VERBOSE, _("Removing %s.\n"), filename);
772 if (unlink (filename))
773 logprintf (LOG_NOTQUIET, "unlink: %s\n", strerror (errno));
776 FREE_MAYBE (redirected_URL);
777 FREE_MAYBE (filename);
780 /* And then from the input file, if any. */
781 if (opt.input_filename)
784 status = retrieve_from_file (opt.input_filename, opt.force_html, &count);
786 logprintf (LOG_NOTQUIET, _("No URLs found in %s.\n"),
789 /* Print the downloaded sum. */
792 || (opt.input_filename && opt.downloaded != 0))
794 logprintf (LOG_NOTQUIET,
795 _("\nFINISHED --%s--\nDownloaded: %s bytes in %d files\n"),
797 (opt.downloaded_overflow ?
798 "<overflow>" : legible_very_long (opt.downloaded)),
800 /* Print quota warning, if exceeded. */
801 if (downloaded_exceeds_quota ())
802 logprintf (LOG_NOTQUIET,
803 _("Download quota (%s bytes) EXCEEDED!\n"),
804 legible (opt.quota));
806 if (opt.convert_links & !opt.delete_after)
808 convert_all_links ();
813 print_malloc_debug_stats ();
815 if (status == RETROK)
821 /* Hangup signal handler. When wget receives SIGHUP or SIGUSR1, it
822 will proceed operation as usual, trying to write into a log file.
823 If that is impossible, the output will be turned off. */
827 redirect_output_signal (int sig)
830 signal (sig, redirect_output_signal);
831 /* Please note that the double `%' in `%%s' is intentional, because
832 redirect_output passes tmp through printf. */
833 sprintf (tmp, _("%s received, redirecting output to `%%s'.\n"),
834 (sig == SIGHUP ? "SIGHUP" :
835 (sig == SIGUSR1 ? "SIGUSR1" :
837 redirect_output (tmp);
839 #endif /* HAVE_SIGNAL */