1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
3 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012 Free Software Foundation,
6 This file is part of GNU Wget.
8 GNU Wget is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 3 of the License, or
11 (at your option) any later version.
13 GNU Wget is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with Wget. If not, see <http://www.gnu.org/licenses/>.
21 Additional permission under GNU GPL version 3 section 7
23 If you modify this program, or any covered work, by linking or
24 combining it with the OpenSSL project's OpenSSL library (or a
25 modified version of that library), containing parts covered by the
26 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
27 grants you additional permission to convey the resulting work.
28 Corresponding Source for a non-source form of such a combination
29 shall include the source code for the parts of OpenSSL used as well
30 as that of the covered work. */
42 /* not all systems provide PATH_MAX in limits.h */
44 # include <sys/param.h>
46 # define PATH_MAX MAXPATHLEN
65 #include "recur.h" /* for INFINITE_RECURSION */
66 #include "convert.h" /* for convert_cleanup */
67 #include "res.h" /* for res_cleanup */
68 #include "http.h" /* for http_cleanup */
69 #include "retr.h" /* for output_stream */
70 #include "warc.h" /* for warc_close */
78 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
80 CMD_DECLARE (cmd_boolean);
81 CMD_DECLARE (cmd_bytes);
82 CMD_DECLARE (cmd_bytes_sum);
84 CMD_DECLARE (cmd_cert_type);
86 CMD_DECLARE (cmd_directory_vector);
87 CMD_DECLARE (cmd_number);
88 CMD_DECLARE (cmd_number_inf);
89 CMD_DECLARE (cmd_string);
90 CMD_DECLARE (cmd_string_uppercase);
91 CMD_DECLARE (cmd_file);
92 CMD_DECLARE (cmd_directory);
93 CMD_DECLARE (cmd_time);
94 CMD_DECLARE (cmd_vector);
96 CMD_DECLARE (cmd_spec_dirstruct);
97 CMD_DECLARE (cmd_spec_header);
98 CMD_DECLARE (cmd_spec_warc_header);
99 CMD_DECLARE (cmd_spec_htmlify);
100 CMD_DECLARE (cmd_spec_mirror);
101 CMD_DECLARE (cmd_spec_prefer_family);
102 CMD_DECLARE (cmd_spec_progress);
103 CMD_DECLARE (cmd_spec_recursive);
104 CMD_DECLARE (cmd_spec_regex_type);
105 CMD_DECLARE (cmd_spec_restrict_file_names);
106 CMD_DECLARE (cmd_spec_report_speed);
108 CMD_DECLARE (cmd_spec_secure_protocol);
110 CMD_DECLARE (cmd_spec_timeout);
111 CMD_DECLARE (cmd_spec_useragent);
112 CMD_DECLARE (cmd_spec_verbose);
114 /* List of recognized commands, each consisting of name, place and
115 function. When adding a new command, simply add it to the list,
116 but be sure to keep the list sorted alphabetically, as
117 command_by_name's binary search depends on it. Also, be sure to
118 add any entries that allocate memory (e.g. cmd_string and
119 cmd_vector) to the cleanup() function below. */
121 static const struct {
124 bool (*action) (const char *, const char *, void *);
126 /* KEEP THIS LIST ALPHABETICALLY SORTED */
127 { "accept", &opt.accepts, cmd_vector },
128 { "acceptregex", &opt.acceptregex_s, cmd_string },
129 { "addhostdir", &opt.add_hostdir, cmd_boolean },
130 { "adjustextension", &opt.adjust_extension, cmd_boolean },
131 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
132 { "askpassword", &opt.ask_passwd, cmd_boolean },
133 { "authnochallenge", &opt.auth_without_challenge,
135 { "background", &opt.background, cmd_boolean },
136 { "backupconverted", &opt.backup_converted, cmd_boolean },
137 { "backups", &opt.backups, cmd_number },
138 { "base", &opt.base_href, cmd_string },
139 { "bindaddress", &opt.bind_address, cmd_string },
140 { "bodydata", &opt.body_data, cmd_string },
141 { "bodyfile", &opt.body_file, cmd_string },
143 { "cacertificate", &opt.ca_cert, cmd_file },
145 { "cache", &opt.allow_cache, cmd_boolean },
147 { "cadirectory", &opt.ca_directory, cmd_directory },
148 { "certificate", &opt.cert_file, cmd_file },
149 { "certificatetype", &opt.cert_type, cmd_cert_type },
150 { "checkcertificate", &opt.check_cert, cmd_boolean },
152 { "chooseconfig", &opt.choose_config, cmd_file },
153 { "connecttimeout", &opt.connect_timeout, cmd_time },
154 { "contentdisposition", &opt.content_disposition, cmd_boolean },
155 { "contentonerror", &opt.content_on_error, cmd_boolean },
156 { "continue", &opt.always_rest, cmd_boolean },
157 { "convertlinks", &opt.convert_links, cmd_boolean },
158 { "cookies", &opt.cookies, cmd_boolean },
159 { "cutdirs", &opt.cut_dirs, cmd_number },
160 { "debug", &opt.debug, cmd_boolean },
161 { "defaultpage", &opt.default_page, cmd_string },
162 { "deleteafter", &opt.delete_after, cmd_boolean },
163 { "dirprefix", &opt.dir_prefix, cmd_directory },
164 { "dirstruct", NULL, cmd_spec_dirstruct },
165 { "dnscache", &opt.dns_cache, cmd_boolean },
166 { "dnstimeout", &opt.dns_timeout, cmd_time },
167 { "domains", &opt.domains, cmd_vector },
168 { "dotbytes", &opt.dot_bytes, cmd_bytes },
169 { "dotsinline", &opt.dots_in_line, cmd_number },
170 { "dotspacing", &opt.dot_spacing, cmd_number },
171 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
173 { "egdfile", &opt.egd_file, cmd_file },
175 { "excludedirectories", &opt.excludes, cmd_directory_vector },
176 { "excludedomains", &opt.exclude_domains, cmd_vector },
177 { "followftp", &opt.follow_ftp, cmd_boolean },
178 { "followtags", &opt.follow_tags, cmd_vector },
179 { "forcehtml", &opt.force_html, cmd_boolean },
180 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
181 { "ftppassword", &opt.ftp_passwd, cmd_string },
182 { "ftpproxy", &opt.ftp_proxy, cmd_string },
184 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
185 #endif /* def __VMS */
186 { "ftpuser", &opt.ftp_user, cmd_string },
187 { "glob", &opt.ftp_glob, cmd_boolean },
188 { "header", NULL, cmd_spec_header },
189 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
190 { "htmlify", NULL, cmd_spec_htmlify },
191 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
192 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
193 { "httppassword", &opt.http_passwd, cmd_string },
194 { "httpproxy", &opt.http_proxy, cmd_string },
196 { "httpsonly", &opt.https_only, cmd_boolean },
198 { "httpsproxy", &opt.https_proxy, cmd_string },
199 { "httpuser", &opt.http_user, cmd_string },
200 { "ignorecase", &opt.ignore_case, cmd_boolean },
201 { "ignorelength", &opt.ignore_length, cmd_boolean },
202 { "ignoretags", &opt.ignore_tags, cmd_vector },
203 { "includedirectories", &opt.includes, cmd_directory_vector },
205 { "inet4only", &opt.ipv4_only, cmd_boolean },
206 { "inet6only", &opt.ipv6_only, cmd_boolean },
208 { "input", &opt.input_filename, cmd_file },
209 { "iri", &opt.enable_iri, cmd_boolean },
210 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
211 { "limitrate", &opt.limit_rate, cmd_bytes },
212 { "loadcookies", &opt.cookies_input, cmd_file },
213 { "localencoding", &opt.locale, cmd_string },
214 { "logfile", &opt.lfilename, cmd_file },
215 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
216 { "maxredirect", &opt.max_redirect, cmd_number },
217 { "method", &opt.method, cmd_string_uppercase },
218 { "mirror", NULL, cmd_spec_mirror },
219 { "netrc", &opt.netrc, cmd_boolean },
220 { "noclobber", &opt.noclobber, cmd_boolean },
221 { "noconfig", &opt.noconfig, cmd_boolean },
222 { "noparent", &opt.no_parent, cmd_boolean },
223 { "noproxy", &opt.no_proxy, cmd_vector },
224 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
225 { "outputdocument", &opt.output_document, cmd_file },
226 { "pagerequisites", &opt.page_requisites, cmd_boolean },
227 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
228 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
229 { "password", &opt.passwd, cmd_string },
230 { "postdata", &opt.post_data, cmd_string },
231 { "postfile", &opt.post_file_name, cmd_file },
232 { "preferfamily", NULL, cmd_spec_prefer_family },
233 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
235 { "privatekey", &opt.private_key, cmd_file },
236 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
238 { "progress", &opt.progress_type, cmd_spec_progress },
239 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
240 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
241 { "proxypassword", &opt.proxy_passwd, cmd_string },
242 { "proxyuser", &opt.proxy_user, cmd_string },
243 { "quiet", &opt.quiet, cmd_boolean },
244 { "quota", &opt.quota, cmd_bytes_sum },
246 { "randomfile", &opt.random_file, cmd_file },
248 { "randomwait", &opt.random_wait, cmd_boolean },
249 { "readtimeout", &opt.read_timeout, cmd_time },
250 { "reclevel", &opt.reclevel, cmd_number_inf },
251 { "recursive", NULL, cmd_spec_recursive },
252 { "referer", &opt.referer, cmd_string },
253 { "regextype", &opt.regex_type, cmd_spec_regex_type },
254 { "reject", &opt.rejects, cmd_vector },
255 { "rejectregex", &opt.rejectregex_s, cmd_string },
256 { "relativeonly", &opt.relative_only, cmd_boolean },
257 { "remoteencoding", &opt.encoding_remote, cmd_string },
258 { "removelisting", &opt.remove_listing, cmd_boolean },
259 { "reportspeed", &opt.report_bps, cmd_spec_report_speed},
260 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
261 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
262 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
263 { "robots", &opt.use_robots, cmd_boolean },
264 { "savecookies", &opt.cookies_output, cmd_file },
265 { "saveheaders", &opt.save_headers, cmd_boolean },
267 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
269 { "serverresponse", &opt.server_response, cmd_boolean },
270 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
271 { "spanhosts", &opt.spanhost, cmd_boolean },
272 { "spider", &opt.spider, cmd_boolean },
273 { "startpos", &opt.start_pos, cmd_bytes },
274 { "strictcomments", &opt.strict_comments, cmd_boolean },
275 { "timeout", NULL, cmd_spec_timeout },
276 { "timestamping", &opt.timestamping, cmd_boolean },
277 { "tries", &opt.ntry, cmd_number_inf },
278 { "trustservernames", &opt.trustservernames, cmd_boolean },
279 { "unlink", &opt.unlink, cmd_boolean },
280 { "useproxy", &opt.use_proxy, cmd_boolean },
281 { "user", &opt.user, cmd_string },
282 { "useragent", NULL, cmd_spec_useragent },
283 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
284 { "verbose", NULL, cmd_spec_verbose },
285 { "wait", &opt.wait, cmd_time },
286 { "waitretry", &opt.waitretry, cmd_time },
287 { "warccdx", &opt.warc_cdx_enabled, cmd_boolean },
288 { "warccdxdedup", &opt.warc_cdx_dedup_filename, cmd_file },
290 { "warccompression", &opt.warc_compression_enabled, cmd_boolean },
292 { "warcdigests", &opt.warc_digests_enabled, cmd_boolean },
293 { "warcfile", &opt.warc_filename, cmd_file },
294 { "warcheader", NULL, cmd_spec_warc_header },
295 { "warckeeplog", &opt.warc_keep_log, cmd_boolean },
296 { "warcmaxsize", &opt.warc_maxsize, cmd_bytes },
297 { "warctempdir", &opt.warc_tempdir, cmd_directory },
299 { "wdebug", &opt.wdebug, cmd_boolean },
303 /* Look up CMDNAME in the commands[] and return its position in the
304 array. If CMDNAME is not found, return -1. */
307 command_by_name (const char *cmdname)
309 /* Use binary search for speed. Wget has ~100 commands, which
310 guarantees a worst case performance of 7 string comparisons. */
311 int lo = 0, hi = countof (commands) - 1;
315 int mid = (lo + hi) >> 1;
316 int cmp = strcasecmp (cmdname, commands[mid].name);
327 /* Reset the variables to default values. */
333 /* Most of the default values are 0 (and 0.0, NULL, and false).
334 Just reset everything, and fill in the non-zero values. Note
335 that initializing pointers to NULL this way is technically
336 illegal, but porting Wget to a machine where NULL is not all-zero
337 bit pattern will be the least of the implementors' worries. */
344 opt.add_hostdir = true;
348 opt.http_keep_alive = true;
349 opt.use_proxy = true;
350 tmp = getenv ("no_proxy");
352 opt.no_proxy = sepstring (tmp);
353 opt.prefer_family = prefer_none;
354 opt.allow_cache = true;
356 opt.read_timeout = 900;
357 opt.use_robots = true;
359 opt.remove_listing = true;
361 opt.dot_bytes = 1024;
362 opt.dot_spacing = 10;
363 opt.dots_in_line = 50;
365 opt.dns_cache = true;
369 opt.check_cert = true;
372 /* The default for file name restriction defaults to the OS type. */
373 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
374 opt.restrict_files_os = restrict_windows;
376 opt.restrict_files_os = restrict_unix;
378 opt.restrict_files_ctrl = true;
379 opt.restrict_files_nonascii = false;
380 opt.restrict_files_case = restrict_no_case_restriction;
382 opt.regex_type = regex_type_posix;
384 opt.max_redirect = 20;
389 opt.enable_iri = true;
391 opt.enable_iri = false;
394 opt.encoding_remote = NULL;
396 opt.useservertimestamps = true;
397 opt.show_all_dns_entries = false;
399 opt.warc_maxsize = 0; /* 1024 * 1024 * 1024; */
401 opt.warc_compression_enabled = true;
403 opt.warc_compression_enabled = false;
405 opt.warc_digests_enabled = true;
406 opt.warc_cdx_enabled = false;
407 opt.warc_cdx_dedup_filename = NULL;
408 opt.warc_tempdir = NULL;
409 opt.warc_keep_log = true;
411 /* Use a negative value to mark the absence of --start-pos option */
415 /* Return the user's home directory (strdup-ed), or NULL if none is
420 static char *buf = NULL;
421 static char *home, *ret;
425 home = getenv ("HOME");
431 /* Under MSDOS, if $HOME isn't defined, use the directory where
432 `wget.exe' resides. */
433 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
436 buff = _w32_get_argv0 ();
438 p = strrchr (buf, '/'); /* djgpp */
440 p = strrchr (buf, '\\'); /* others */
444 buff = malloc (len + 1);
448 strncpy (buff, _w32_get_argv0 (), len);
452 #elif !defined(WINDOWS)
453 /* If HOME is not defined, try getting it from the password
455 struct passwd *pwd = getpwuid (getuid ());
456 if (!pwd || !pwd->pw_dir)
460 /* Under Windows, if $HOME isn't defined, use the directory where
461 `wget.exe' resides. */
467 ret = home ? xstrdup (home) : NULL;
473 /* Check the 'WGETRC' environment variable and return the file name
474 if 'WGETRC' is set and is a valid file.
475 If the `WGETRC' variable exists but the file does not exist, the
476 function will exit(). */
478 wgetrc_env_file_name (void)
480 char *env = getenv ("WGETRC");
483 if (!file_exists_p (env))
485 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
489 return xstrdup (env);
494 /* Check for the existance of '$HOME/.wgetrc' and return its path
495 if it exists and is set. */
497 wgetrc_user_file_name (void)
501 /* If that failed, try $HOME/.wgetrc (or equivalent). */
504 file = "SYS$LOGIN:.wgetrc";
505 #else /* def __VMS */
508 file = aprintf ("%s/.wgetrc", home);
510 #endif /* def __VMS [else] */
514 if (!file_exists_p (file))
522 /* Return the path to the user's .wgetrc. This is either the value of
523 `WGETRC' environment variable, or `$HOME/.wgetrc'.
525 Additionally, for windows, look in the directory where wget.exe
528 wgetrc_file_name (void)
530 char *file = wgetrc_env_file_name ();
534 file = wgetrc_user_file_name ();
537 /* Under Windows, if we still haven't found .wgetrc, look for the file
538 `wget.ini' in the directory where `wget.exe' resides; we do this for
539 backward compatibility with previous versions of Wget.
540 SYSTEM_WGETRC should not be defined under WINDOWS. */
543 char *home = home_dir ();
549 file = aprintf ("%s/wget.ini", home);
550 if (!file_exists_p (file))
563 /* Return values of parse_line. */
571 static enum parse_line parse_line (const char *, char **, char **, int *);
572 static bool setval_internal (int, const char *, const char *);
573 static bool setval_internal_tilde (int, const char *, const char *);
575 /* Initialize variables from a wgetrc file. Returns zero (failure) if
576 there were errors in the file. */
579 run_wgetrc (const char *file)
587 fp = fopen (file, "r");
590 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
591 file, strerror (errno));
592 return true; /* not a fatal error */
595 while (getline (&line, &bufsize, fp) > 0)
597 char *com = NULL, *val = NULL;
600 /* Parse the line. */
601 switch (parse_line (line, &com, &val, &comind))
604 /* If everything is OK, set the value. */
605 if (!setval_internal_tilde (comind, com, val))
607 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
608 exec_name, file, ln);
612 case line_syntax_error:
613 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
614 exec_name, file, ln);
617 case line_unknown_command:
618 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
619 exec_name, quote (com), file, ln);
637 /* Initialize the defaults and run the system wgetrc and user's own
642 char *file, *env_sysrc;
645 /* Run a non-standard system rc file when the according environment
646 variable has been set. For internal testing purposes only! */
647 env_sysrc = getenv ("SYSTEM_WGETRC");
648 if (env_sysrc && file_exists_p (env_sysrc))
650 ok &= run_wgetrc (env_sysrc);
651 /* If there are any problems parsing the system wgetrc file, tell
655 fprintf (stderr, _("\
656 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
658 or specify a different file using --config.\n"), env_sysrc);
662 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
664 else if (file_exists_p (SYSTEM_WGETRC))
665 ok &= run_wgetrc (SYSTEM_WGETRC);
666 /* If there are any problems parsing the system wgetrc file, tell
670 fprintf (stderr, _("\
671 Parsing system wgetrc file failed. Please check\n\
673 or specify a different file using --config.\n"), SYSTEM_WGETRC);
677 /* Override it with your own, if one exists. */
678 file = wgetrc_file_name ();
681 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
682 something like realpath() before comparing them with `strcmp' */
684 if (!strcmp (file, SYSTEM_WGETRC))
686 fprintf (stderr, _("\
687 %s: Warning: Both system and user wgetrc point to %s.\n"),
688 exec_name, quote (file));
692 ok &= run_wgetrc (file);
694 /* If there were errors processing either `.wgetrc', abort. */
702 /* Remove dashes and underscores from S, modifying S in the
708 char *t = s; /* t - tortoise */
709 char *h = s; /* h - hare */
711 if (*h == '_' || *h == '-')
718 /* Parse the line pointed by line, with the syntax:
719 <sp>* command <sp>* = <sp>* value <sp>*
720 Uses malloc to allocate space for command and value.
722 Returns one of line_ok, line_empty, line_syntax_error, or
723 line_unknown_command.
725 In case of line_ok, *COM and *VAL point to freshly allocated
726 strings, and *COMIND points to com's index. In case of error or
727 empty line, their values are unmodified. */
729 static enum parse_line
730 parse_line (const char *line, char **com, char **val, int *comind)
733 const char *end = line + strlen (line);
734 const char *cmdstart, *cmdend;
735 const char *valstart, *valend;
740 /* Skip leading and trailing whitespace. */
741 while (*line && c_isspace (*line))
743 while (end > line && c_isspace (end[-1]))
746 /* Skip empty lines and comments. */
747 if (!*line || *line == '#')
753 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
757 /* Skip '=', as well as any space before or after it. */
758 while (p < end && c_isspace (*p))
760 if (p == end || *p != '=')
761 return line_syntax_error;
763 while (p < end && c_isspace (*p))
769 /* The syntax is valid (even though the command might not be). Fill
770 in the command name and value. */
771 *com = strdupdelim (cmdstart, cmdend);
772 *val = strdupdelim (valstart, valend);
774 /* The line now known to be syntactically correct. Check whether
775 the command is valid. */
776 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
778 ind = command_by_name (cmdcopy);
780 return line_unknown_command;
782 /* Report success to the caller. */
787 #if defined(WINDOWS) || defined(MSDOS)
788 # define ISSEP(c) ((c) == '/' || (c) == '\\')
790 # define ISSEP(c) ((c) == '/')
793 /* Run commands[comind].action. */
796 setval_internal (int comind, const char *com, const char *val)
798 assert (0 <= comind && ((size_t) comind) < countof (commands));
799 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
800 return commands[comind].action (com, val, commands[comind].place);
804 setval_internal_tilde (int comind, const char *com, const char *val)
810 ret = setval_internal (comind, com, val);
812 /* We make tilde expansion for cmd_file and cmd_directory */
813 if (((commands[comind].action == cmd_file) ||
814 (commands[comind].action == cmd_directory))
815 && ret && (*val == '~' && ISSEP (val[1])))
817 pstring = commands[comind].place;
821 homelen = strlen (home);
822 while (homelen && ISSEP (home[homelen - 1]))
823 home[--homelen] = '\0';
825 /* Skip the leading "~/". */
826 for (++val; ISSEP (*val); val++)
828 *pstring = concat_strings (home, "/", val, (char *)0);
834 /* Run command COM with value VAL. If running the command produces an
835 error, report the error and exit.
837 This is intended to be called from main() to modify Wget's behavior
838 through command-line switches. Since COM is hard-coded in main(),
839 it is not canonicalized, and this aborts when COM is not found.
841 If COMIND's are exported to init.h, this function will be changed
842 to accept COMIND directly. */
845 setoptval (const char *com, const char *val, const char *optname)
847 /* Prepend "--" to OPTNAME. */
848 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
851 strcpy (dd_optname + 2, optname);
853 assert (val != NULL);
854 if (!setval_internal (command_by_name (com), dd_optname, val))
858 /* Parse OPT into command and value and run it. For example,
859 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
860 This is used by the `--execute' flag in main.c. */
863 run_command (const char *opt)
867 switch (parse_line (opt, &com, &val, &comind))
870 if (!setval_internal (comind, com, val))
876 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
877 exec_name, quote (opt));
882 /* Generic helper functions, for use with `commands'. */
884 /* Forward declarations: */
889 static bool decode_string (const char *, const struct decode_item *, int, int *);
890 static bool simple_atoi (const char *, const char *, int *);
891 static bool simple_atof (const char *, const char *, double *);
893 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
895 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
896 && c_tolower((p)[1]) == (c1) \
899 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
900 && c_tolower((p)[1]) == (c1) \
901 && c_tolower((p)[2]) == (c2) \
905 /* Store the boolean value from VAL to PLACE. COM is ignored,
906 except for error messages. */
908 cmd_boolean (const char *com, const char *val, void *place)
912 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
913 /* "on", "yes" and "1" mean true. */
915 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
916 /* "off", "no" and "0" mean false. */
921 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
922 exec_name, com, quote (val));
926 *(bool *) place = value;
930 /* Set the non-negative integer value from VAL to PLACE. With
931 incorrect specification, the number remains unchanged. */
933 cmd_number (const char *com, const char *val, void *place)
935 if (!simple_atoi (val, val + strlen (val), place)
936 || *(int *) place < 0)
938 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
939 exec_name, com, quote (val));
945 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
947 cmd_number_inf (const char *com, const char *val, void *place)
949 if (!strcasecmp (val, "inf"))
954 return cmd_number (com, val, place);
957 /* Copy (strdup) the string at COM to a new location and place a
958 pointer to *PLACE. */
960 cmd_string (const char *com, const char *val, void *place)
962 char **pstring = (char **)place;
964 xfree_null (*pstring);
965 *pstring = xstrdup (val);
969 /* Like cmd_string but ensure the string is upper case. */
971 cmd_string_uppercase (const char *com, const char *val, void *place)
974 pstring = (char **)place;
975 xfree_null (*pstring);
977 *pstring = xmalloc (strlen (val) + 1);
979 for (q = *pstring; *val; val++, q++)
980 *q = c_toupper (*val);
987 /* Like cmd_string, but handles tilde-expansion when reading a user's
988 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
989 gets expanded to the user's home directory. */
991 cmd_file (const char *com, const char *val, void *place)
993 char **pstring = (char **)place;
995 xfree_null (*pstring);
997 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
999 *pstring = xstrdup (val);
1001 #if defined(WINDOWS) || defined(MSDOS)
1002 /* Convert "\" to "/". */
1005 for (s = *pstring; *s; s++)
1013 /* Like cmd_file, but strips trailing '/' characters. */
1015 cmd_directory (const char *com, const char *val, void *place)
1019 /* Call cmd_file() for tilde expansion and separator
1020 canonicalization (backslash -> slash under Windows). These
1021 things should perhaps be in a separate function. */
1022 if (!cmd_file (com, val, place))
1025 s = *(char **)place;
1027 while (t > s && *--t == '/')
1033 /* Split VAL by space to a vector of values, and append those values
1034 to vector pointed to by the PLACE argument. If VAL is empty, the
1035 PLACE vector is cleared instead. */
1038 cmd_vector (const char *com, const char *val, void *place)
1040 char ***pvec = (char ***)place;
1043 *pvec = merge_vecs (*pvec, sepstring (val));
1053 cmd_directory_vector (const char *com, const char *val, void *place)
1055 char ***pvec = (char ***)place;
1059 /* Strip the trailing slashes from directories. */
1062 seps = sepstring (val);
1063 for (t = seps; t && *t; t++)
1065 int len = strlen (*t);
1066 /* Skip degenerate case of root directory. */
1069 if ((*t)[len - 1] == '/')
1070 (*t)[len - 1] = '\0';
1073 *pvec = merge_vecs (*pvec, seps);
1083 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1084 "100k" or "2.5G" to a floating point number. */
1087 parse_bytes_helper (const char *val, double *result)
1089 double number, mult;
1090 const char *end = val + strlen (val);
1092 /* Check for "inf". */
1093 if (0 == strcmp (val, "inf"))
1099 /* Strip trailing whitespace. */
1100 while (val < end && c_isspace (end[-1]))
1105 switch (c_tolower (end[-1]))
1108 --end, mult = 1024.0;
1111 --end, mult = 1048576.0;
1114 --end, mult = 1073741824.0;
1117 --end, mult = 1099511627776.0;
1120 /* Not a recognized suffix: assume it's a digit. (If not,
1121 simple_atof will raise an error.) */
1125 /* Skip leading and trailing whitespace. */
1126 while (val < end && c_isspace (*val))
1128 while (val < end && c_isspace (end[-1]))
1133 if (!simple_atof (val, end, &number) || number < 0)
1136 *result = number * mult;
1140 /* Parse VAL as a number and set its value to PLACE (which should
1143 By default, the value is assumed to be in bytes. If "K", "M", or
1144 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1145 1<<30, respectively. Floating point values are allowed and are
1146 cast to integer before use. The idea is to be able to use things
1147 like 1.5k instead of "1536".
1149 The string "inf" is returned as 0.
1151 In case of error, false is returned and memory pointed to by PLACE
1152 remains unmodified. */
1155 cmd_bytes (const char *com, const char *val, void *place)
1158 if (!parse_bytes_helper (val, &byte_value))
1160 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1161 exec_name, com, quote (val));
1164 *(wgint *)place = (wgint)byte_value;
1168 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1169 SIZE_SUM. It works by converting the string to double, therefore
1170 working with values up to 2^53-1 without loss of precision. This
1171 value (8192 TB) is large enough to serve for a while. */
1174 cmd_bytes_sum (const char *com, const char *val, void *place)
1177 if (!parse_bytes_helper (val, &byte_value))
1179 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1180 exec_name, com, quote (val));
1183 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1187 /* Store the value of VAL to *OUT. The value is a time period, by
1188 default expressed in seconds, but also accepting suffixes "m", "h",
1189 "d", and "w" for minutes, hours, days, and weeks respectively. */
1192 cmd_time (const char *com, const char *val, void *place)
1194 double number, mult;
1195 const char *end = val + strlen (val);
1197 /* Strip trailing whitespace. */
1198 while (val < end && c_isspace (end[-1]))
1204 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1205 exec_name, com, quote (val));
1209 switch (c_tolower (end[-1]))
1212 --end, mult = 1; /* seconds */
1215 --end, mult = 60; /* minutes */
1218 --end, mult = 3600; /* hours */
1221 --end, mult = 86400.0; /* days */
1224 --end, mult = 604800.0; /* weeks */
1227 /* Not a recognized suffix: assume it belongs to the number.
1228 (If not, simple_atof will raise an error.) */
1232 /* Skip leading and trailing whitespace. */
1233 while (val < end && c_isspace (*val))
1235 while (val < end && c_isspace (end[-1]))
1240 if (!simple_atof (val, end, &number))
1243 *(double *)place = number * mult;
1249 cmd_cert_type (const char *com, const char *val, void *place)
1251 static const struct decode_item choices[] = {
1252 { "pem", keyfile_pem },
1253 { "der", keyfile_asn1 },
1254 { "asn1", keyfile_asn1 },
1256 int ok = decode_string (val, choices, countof (choices), place);
1258 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1263 /* Specialized helper functions, used by `commands' to handle some
1264 options specially. */
1266 static bool check_user_specified_header (const char *);
1269 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1271 if (!cmd_boolean (com, val, &opt.dirstruct))
1273 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1274 must be affected inversely. */
1276 opt.no_dirstruct = false;
1278 opt.no_dirstruct = true;
1283 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1285 /* Empty value means reset the list of headers. */
1288 free_vec (opt.user_headers);
1289 opt.user_headers = NULL;
1293 if (!check_user_specified_header (val))
1295 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1296 exec_name, com, quote (val));
1299 opt.user_headers = vec_append (opt.user_headers, val);
1304 cmd_spec_warc_header (const char *com, const char *val, void *place_ignored)
1306 /* Empty value means reset the list of headers. */
1309 free_vec (opt.warc_user_headers);
1310 opt.warc_user_headers = NULL;
1314 if (!check_user_specified_header (val))
1316 fprintf (stderr, _("%s: %s: Invalid WARC header %s.\n"),
1317 exec_name, com, quote (val));
1320 opt.warc_user_headers = vec_append (opt.warc_user_headers, val);
1325 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1327 int flag = cmd_boolean (com, val, &opt.htmlify);
1328 if (flag && !opt.htmlify)
1329 opt.remove_listing = false;
1333 /* Set the "mirror" mode. It means: recursive download, timestamping,
1334 no limit on max. recursion depth, and don't remove listings. */
1337 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1341 if (!cmd_boolean (com, val, &mirror))
1345 opt.recursive = true;
1346 if (!opt.no_dirstruct)
1347 opt.dirstruct = true;
1348 opt.timestamping = true;
1349 opt.reclevel = INFINITE_RECURSION;
1350 opt.remove_listing = false;
1355 /* Validate --prefer-family and set the choice. Allowed values are
1356 "IPv4", "IPv6", and "none". */
1359 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1361 static const struct decode_item choices[] = {
1362 { "IPv4", prefer_ipv4 },
1363 { "IPv6", prefer_ipv6 },
1364 { "none", prefer_none },
1366 int prefer_family = prefer_none;
1367 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1369 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1370 opt.prefer_family = prefer_family;
1374 /* Set progress.type to VAL, but verify that it's a valid progress
1375 implementation before that. */
1378 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1380 if (!valid_progress_implementation_p (val))
1382 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1383 exec_name, com, quote (val));
1386 xfree_null (opt.progress_type);
1388 /* Don't call set_progress_implementation here. It will be called
1389 in main() when it becomes clear what the log output is. */
1390 opt.progress_type = xstrdup (val);
1394 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1395 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1399 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1401 if (!cmd_boolean (com, val, &opt.recursive))
1405 if (opt.recursive && !opt.no_dirstruct)
1406 opt.dirstruct = true;
1411 /* Validate --regex-type and set the choice. */
1414 cmd_spec_regex_type (const char *com, const char *val, void *place_ignored)
1416 static const struct decode_item choices[] = {
1417 { "posix", regex_type_posix },
1419 { "pcre", regex_type_pcre },
1422 int regex_type = regex_type_posix;
1423 int ok = decode_string (val, choices, countof (choices), ®ex_type);
1425 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1426 opt.regex_type = regex_type;
1431 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1433 int restrict_os = opt.restrict_files_os;
1434 int restrict_ctrl = opt.restrict_files_ctrl;
1435 int restrict_case = opt.restrict_files_case;
1436 int restrict_nonascii = opt.restrict_files_nonascii;
1440 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1444 end = strchr (val, ',');
1446 end = val + strlen (val);
1448 if (VAL_IS ("unix"))
1449 restrict_os = restrict_unix;
1450 else if (VAL_IS ("windows"))
1451 restrict_os = restrict_windows;
1452 else if (VAL_IS ("lowercase"))
1453 restrict_case = restrict_lowercase;
1454 else if (VAL_IS ("uppercase"))
1455 restrict_case = restrict_uppercase;
1456 else if (VAL_IS ("nocontrol"))
1457 restrict_ctrl = false;
1458 else if (VAL_IS ("ascii"))
1459 restrict_nonascii = true;
1462 fprintf (stderr, _("\
1463 %s: %s: Invalid restriction %s,\n\
1464 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1465 exec_name, com, quote (val));
1472 while (*val && *end);
1476 opt.restrict_files_os = restrict_os;
1477 opt.restrict_files_ctrl = restrict_ctrl;
1478 opt.restrict_files_case = restrict_case;
1479 opt.restrict_files_nonascii = restrict_nonascii;
1485 cmd_spec_report_speed (const char *com, const char *val, void *place_ignored)
1487 opt.report_bps = strcasecmp (val, "bits") == 0;
1488 if (!opt.report_bps)
1489 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1490 return opt.report_bps;
1495 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1497 static const struct decode_item choices[] = {
1498 { "auto", secure_protocol_auto },
1499 { "sslv2", secure_protocol_sslv2 },
1500 { "sslv3", secure_protocol_sslv3 },
1501 { "tlsv1", secure_protocol_tlsv1 },
1502 { "pfs", secure_protocol_pfs },
1504 int ok = decode_string (val, choices, countof (choices), place);
1506 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1511 /* Set all three timeout values. */
1514 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1517 if (!cmd_time (com, val, &value))
1519 opt.read_timeout = value;
1520 opt.connect_timeout = value;
1521 opt.dns_timeout = value;
1526 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1528 /* Disallow embedded newlines. */
1529 if (strchr (val, '\n'))
1531 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1532 exec_name, com, quote (val));
1535 xfree_null (opt.useragent);
1536 opt.useragent = xstrdup (val);
1540 /* The "verbose" option cannot be cmd_boolean because the variable is
1541 not bool -- it's of type int (-1 means uninitialized because of
1542 some random hackery for disallowing -q -v). */
1545 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1548 if (cmd_boolean (com, val, &flag))
1556 /* Miscellaneous useful routines. */
1558 /* A very simple atoi clone, more useful than atoi because it works on
1559 delimited strings, and has error reportage. Returns true on success,
1560 false on failure. If successful, stores result to *DEST. */
1563 simple_atoi (const char *beg, const char *end, int *dest)
1566 bool negative = false;
1567 const char *p = beg;
1569 while (p < end && c_isspace (*p))
1571 if (p < end && (*p == '-' || *p == '+'))
1573 negative = (*p == '-');
1579 /* Read negative numbers in a separate loop because the most
1580 negative integer cannot be represented as a positive number. */
1583 for (; p < end && c_isdigit (*p); p++)
1585 int next = (10 * result) + (*p - '0');
1587 return false; /* overflow */
1591 for (; p < end && c_isdigit (*p); p++)
1593 int next = (10 * result) - (*p - '0');
1595 return false; /* underflow */
1606 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1607 doesn't handle exponential notation. Returns true on success,
1608 false on failure. In case of success, stores its result to
1612 simple_atof (const char *beg, const char *end, double *dest)
1616 bool negative = false;
1617 bool seen_dot = false;
1618 bool seen_digit = false;
1621 const char *p = beg;
1623 while (p < end && c_isspace (*p))
1625 if (p < end && (*p == '-' || *p == '+'))
1627 negative = (*p == '-');
1631 for (; p < end; p++)
1637 result = (10 * result) + (ch - '0');
1639 result += (ch - '0') / (divider *= 10);
1661 /* Verify that the user-specified header in S is valid. It must
1662 contain a colon preceded by non-white-space characters and must not
1663 contain newlines. */
1666 check_user_specified_header (const char *s)
1670 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1672 /* The header MUST contain `:' preceded by at least one
1673 non-whitespace character. */
1674 if (*p != ':' || p == s)
1676 /* The header MUST NOT contain newlines. */
1677 if (strchr (s, '\n'))
1682 /* Decode VAL into a number, according to ITEMS. */
1685 decode_string (const char *val, const struct decode_item *items, int itemcount,
1689 for (i = 0; i < itemcount; i++)
1690 if (0 == strcasecmp (val, items[i].name))
1692 *place = items[i].code;
1699 void cleanup_html_url (void);
1700 void spider_cleanup (void);
1703 /* Free the memory allocated by global variables. */
1707 /* Free external resources, close files, etc. */
1709 /* Close WARC file. */
1710 if (opt.warc_filename != 0)
1716 if (fclose (output_stream) == EOF)
1717 inform_exit_status (CLOSEFAILED);
1719 /* No need to check for error because Wget flushes its output (and
1720 checks for errors) after any data arrives. */
1722 /* We're exiting anyway so there's no real need to call free()
1723 hundreds of times. Skipping the frees will make Wget exit
1726 However, when detecting leaks, it's crucial to free() everything
1727 because then you can find the real leaks, i.e. the allocated
1728 memory which grows with the size of the program. */
1734 cleanup_html_url ();
1739 for (i = 0; i < nurl; i++)
1743 extern acc_t *netrc_list;
1744 free_netrc (netrc_list);
1746 xfree_null (opt.choose_config);
1747 xfree_null (opt.lfilename);
1748 xfree_null (opt.dir_prefix);
1749 xfree_null (opt.input_filename);
1750 xfree_null (opt.output_document);
1751 free_vec (opt.accepts);
1752 free_vec (opt.rejects);
1753 free_vec (opt.excludes);
1754 free_vec (opt.includes);
1755 free_vec (opt.domains);
1756 free_vec (opt.follow_tags);
1757 free_vec (opt.ignore_tags);
1758 xfree_null (opt.progress_type);
1759 xfree_null (opt.ftp_user);
1760 xfree_null (opt.ftp_passwd);
1761 xfree_null (opt.ftp_proxy);
1762 xfree_null (opt.https_proxy);
1763 xfree_null (opt.http_proxy);
1764 free_vec (opt.no_proxy);
1765 xfree_null (opt.useragent);
1766 xfree_null (opt.referer);
1767 xfree_null (opt.http_user);
1768 xfree_null (opt.http_passwd);
1769 free_vec (opt.user_headers);
1770 free_vec (opt.warc_user_headers);
1772 xfree_null (opt.cert_file);
1773 xfree_null (opt.private_key);
1774 xfree_null (opt.ca_directory);
1775 xfree_null (opt.ca_cert);
1776 xfree_null (opt.random_file);
1777 xfree_null (opt.egd_file);
1779 xfree_null (opt.bind_address);
1780 xfree_null (opt.cookies_input);
1781 xfree_null (opt.cookies_output);
1782 xfree_null (opt.user);
1783 xfree_null (opt.passwd);
1784 xfree_null (opt.base_href);
1785 xfree_null (opt.method);
1787 #endif /* DEBUG_MALLOC */
1790 /* Unit testing routines. */
1795 test_commands_sorted()
1797 int prev_idx = 0, next_idx = 1;
1798 int command_count = countof (commands) - 1;
1800 while (next_idx <= command_count)
1802 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1805 mu_assert ("FAILED", false);
1818 test_cmd_spec_restrict_file_names()
1823 int expected_restrict_files_os;
1824 int expected_restrict_files_ctrl;
1825 int expected_restrict_files_case;
1828 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1829 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1830 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1831 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1834 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1839 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1842 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1843 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1844 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1845 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1847 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1848 res == test_array[i].result
1849 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1850 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1851 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1857 #endif /* TESTING */