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 { "showprogress", &opt.show_progress, cmd_boolean },
272 { "spanhosts", &opt.spanhost, cmd_boolean },
273 { "spider", &opt.spider, cmd_boolean },
274 { "startpos", &opt.start_pos, cmd_bytes },
275 { "strictcomments", &opt.strict_comments, cmd_boolean },
276 { "timeout", NULL, cmd_spec_timeout },
277 { "timestamping", &opt.timestamping, cmd_boolean },
278 { "tries", &opt.ntry, cmd_number_inf },
279 { "trustservernames", &opt.trustservernames, cmd_boolean },
280 { "unlink", &opt.unlink, cmd_boolean },
281 { "useproxy", &opt.use_proxy, cmd_boolean },
282 { "user", &opt.user, cmd_string },
283 { "useragent", NULL, cmd_spec_useragent },
284 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
285 { "verbose", NULL, cmd_spec_verbose },
286 { "wait", &opt.wait, cmd_time },
287 { "waitretry", &opt.waitretry, cmd_time },
288 { "warccdx", &opt.warc_cdx_enabled, cmd_boolean },
289 { "warccdxdedup", &opt.warc_cdx_dedup_filename, cmd_file },
291 { "warccompression", &opt.warc_compression_enabled, cmd_boolean },
293 { "warcdigests", &opt.warc_digests_enabled, cmd_boolean },
294 { "warcfile", &opt.warc_filename, cmd_file },
295 { "warcheader", NULL, cmd_spec_warc_header },
296 { "warckeeplog", &opt.warc_keep_log, cmd_boolean },
297 { "warcmaxsize", &opt.warc_maxsize, cmd_bytes },
298 { "warctempdir", &opt.warc_tempdir, cmd_directory },
300 { "wdebug", &opt.wdebug, cmd_boolean },
304 /* Look up CMDNAME in the commands[] and return its position in the
305 array. If CMDNAME is not found, return -1. */
308 command_by_name (const char *cmdname)
310 /* Use binary search for speed. Wget has ~100 commands, which
311 guarantees a worst case performance of 7 string comparisons. */
312 int lo = 0, hi = countof (commands) - 1;
316 int mid = (lo + hi) >> 1;
317 int cmp = strcasecmp (cmdname, commands[mid].name);
328 /* Reset the variables to default values. */
334 /* Most of the default values are 0 (and 0.0, NULL, and false).
335 Just reset everything, and fill in the non-zero values. Note
336 that initializing pointers to NULL this way is technically
337 illegal, but porting Wget to a machine where NULL is not all-zero
338 bit pattern will be the least of the implementors' worries. */
345 opt.add_hostdir = true;
349 opt.http_keep_alive = true;
350 opt.use_proxy = true;
351 tmp = getenv ("no_proxy");
353 opt.no_proxy = sepstring (tmp);
354 opt.prefer_family = prefer_none;
355 opt.allow_cache = true;
357 opt.read_timeout = 900;
358 opt.use_robots = true;
360 opt.remove_listing = true;
362 opt.dot_bytes = 1024;
363 opt.dot_spacing = 10;
364 opt.dots_in_line = 50;
366 opt.dns_cache = true;
370 opt.check_cert = true;
373 /* The default for file name restriction defaults to the OS type. */
374 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
375 opt.restrict_files_os = restrict_windows;
377 opt.restrict_files_os = restrict_unix;
379 opt.restrict_files_ctrl = true;
380 opt.restrict_files_nonascii = false;
381 opt.restrict_files_case = restrict_no_case_restriction;
383 opt.regex_type = regex_type_posix;
385 opt.max_redirect = 20;
390 opt.enable_iri = true;
392 opt.enable_iri = false;
395 opt.encoding_remote = NULL;
397 opt.useservertimestamps = true;
398 opt.show_all_dns_entries = false;
400 opt.warc_maxsize = 0; /* 1024 * 1024 * 1024; */
402 opt.warc_compression_enabled = true;
404 opt.warc_compression_enabled = false;
406 opt.warc_digests_enabled = true;
407 opt.warc_cdx_enabled = false;
408 opt.warc_cdx_dedup_filename = NULL;
409 opt.warc_tempdir = NULL;
410 opt.warc_keep_log = true;
412 /* Use a negative value to mark the absence of --start-pos option */
414 opt.show_progress = false;
417 /* Return the user's home directory (strdup-ed), or NULL if none is
422 static char *buf = NULL;
423 static char *home, *ret;
427 home = getenv ("HOME");
433 /* Under MSDOS, if $HOME isn't defined, use the directory where
434 `wget.exe' resides. */
435 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
438 buff = _w32_get_argv0 ();
440 p = strrchr (buf, '/'); /* djgpp */
442 p = strrchr (buf, '\\'); /* others */
446 buff = malloc (len + 1);
450 strncpy (buff, _w32_get_argv0 (), len);
454 #elif !defined(WINDOWS)
455 /* If HOME is not defined, try getting it from the password
457 struct passwd *pwd = getpwuid (getuid ());
458 if (!pwd || !pwd->pw_dir)
462 /* Under Windows, if $HOME isn't defined, use the directory where
463 `wget.exe' resides. */
469 ret = home ? xstrdup (home) : NULL;
475 /* Check the 'WGETRC' environment variable and return the file name
476 if 'WGETRC' is set and is a valid file.
477 If the `WGETRC' variable exists but the file does not exist, the
478 function will exit(). */
480 wgetrc_env_file_name (void)
482 char *env = getenv ("WGETRC");
485 if (!file_exists_p (env))
487 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
491 return xstrdup (env);
496 /* Check for the existance of '$HOME/.wgetrc' and return its path
497 if it exists and is set. */
499 wgetrc_user_file_name (void)
503 /* If that failed, try $HOME/.wgetrc (or equivalent). */
506 file = "SYS$LOGIN:.wgetrc";
507 #else /* def __VMS */
510 file = aprintf ("%s/.wgetrc", home);
512 #endif /* def __VMS [else] */
516 if (!file_exists_p (file))
524 /* Return the path to the user's .wgetrc. This is either the value of
525 `WGETRC' environment variable, or `$HOME/.wgetrc'.
527 Additionally, for windows, look in the directory where wget.exe
530 wgetrc_file_name (void)
532 char *file = wgetrc_env_file_name ();
536 file = wgetrc_user_file_name ();
539 /* Under Windows, if we still haven't found .wgetrc, look for the file
540 `wget.ini' in the directory where `wget.exe' resides; we do this for
541 backward compatibility with previous versions of Wget.
542 SYSTEM_WGETRC should not be defined under WINDOWS. */
545 char *home = home_dir ();
551 file = aprintf ("%s/wget.ini", home);
552 if (!file_exists_p (file))
565 /* Return values of parse_line. */
573 static enum parse_line parse_line (const char *, char **, char **, int *);
574 static bool setval_internal (int, const char *, const char *);
575 static bool setval_internal_tilde (int, const char *, const char *);
577 /* Initialize variables from a wgetrc file. Returns zero (failure) if
578 there were errors in the file. */
581 run_wgetrc (const char *file)
589 fp = fopen (file, "r");
592 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
593 file, strerror (errno));
594 return true; /* not a fatal error */
597 while (getline (&line, &bufsize, fp) > 0)
599 char *com = NULL, *val = NULL;
602 /* Parse the line. */
603 switch (parse_line (line, &com, &val, &comind))
606 /* If everything is OK, set the value. */
607 if (!setval_internal_tilde (comind, com, val))
609 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
610 exec_name, file, ln);
614 case line_syntax_error:
615 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
616 exec_name, file, ln);
619 case line_unknown_command:
620 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
621 exec_name, quote (com), file, ln);
639 /* Initialize the defaults and run the system wgetrc and user's own
644 char *file, *env_sysrc;
647 /* Run a non-standard system rc file when the according environment
648 variable has been set. For internal testing purposes only! */
649 env_sysrc = getenv ("SYSTEM_WGETRC");
650 if (env_sysrc && file_exists_p (env_sysrc))
652 ok &= run_wgetrc (env_sysrc);
653 /* If there are any problems parsing the system wgetrc file, tell
657 fprintf (stderr, _("\
658 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
660 or specify a different file using --config.\n"), env_sysrc);
664 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
666 else if (file_exists_p (SYSTEM_WGETRC))
667 ok &= run_wgetrc (SYSTEM_WGETRC);
668 /* If there are any problems parsing the system wgetrc file, tell
672 fprintf (stderr, _("\
673 Parsing system wgetrc file failed. Please check\n\
675 or specify a different file using --config.\n"), SYSTEM_WGETRC);
679 /* Override it with your own, if one exists. */
680 file = wgetrc_file_name ();
683 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
684 something like realpath() before comparing them with `strcmp' */
686 if (!strcmp (file, SYSTEM_WGETRC))
688 fprintf (stderr, _("\
689 %s: Warning: Both system and user wgetrc point to %s.\n"),
690 exec_name, quote (file));
694 ok &= run_wgetrc (file);
696 /* If there were errors processing either `.wgetrc', abort. */
704 /* Remove dashes and underscores from S, modifying S in the
710 char *t = s; /* t - tortoise */
711 char *h = s; /* h - hare */
713 if (*h == '_' || *h == '-')
720 /* Parse the line pointed by line, with the syntax:
721 <sp>* command <sp>* = <sp>* value <sp>*
722 Uses malloc to allocate space for command and value.
724 Returns one of line_ok, line_empty, line_syntax_error, or
725 line_unknown_command.
727 In case of line_ok, *COM and *VAL point to freshly allocated
728 strings, and *COMIND points to com's index. In case of error or
729 empty line, their values are unmodified. */
731 static enum parse_line
732 parse_line (const char *line, char **com, char **val, int *comind)
735 const char *end = line + strlen (line);
736 const char *cmdstart, *cmdend;
737 const char *valstart, *valend;
742 /* Skip leading and trailing whitespace. */
743 while (*line && c_isspace (*line))
745 while (end > line && c_isspace (end[-1]))
748 /* Skip empty lines and comments. */
749 if (!*line || *line == '#')
755 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
759 /* Skip '=', as well as any space before or after it. */
760 while (p < end && c_isspace (*p))
762 if (p == end || *p != '=')
763 return line_syntax_error;
765 while (p < end && c_isspace (*p))
771 /* The syntax is valid (even though the command might not be). Fill
772 in the command name and value. */
773 *com = strdupdelim (cmdstart, cmdend);
774 *val = strdupdelim (valstart, valend);
776 /* The line now known to be syntactically correct. Check whether
777 the command is valid. */
778 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
780 ind = command_by_name (cmdcopy);
782 return line_unknown_command;
784 /* Report success to the caller. */
789 #if defined(WINDOWS) || defined(MSDOS)
790 # define ISSEP(c) ((c) == '/' || (c) == '\\')
792 # define ISSEP(c) ((c) == '/')
795 /* Run commands[comind].action. */
798 setval_internal (int comind, const char *com, const char *val)
800 assert (0 <= comind && ((size_t) comind) < countof (commands));
801 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
802 return commands[comind].action (com, val, commands[comind].place);
806 setval_internal_tilde (int comind, const char *com, const char *val)
812 ret = setval_internal (comind, com, val);
814 /* We make tilde expansion for cmd_file and cmd_directory */
815 if (((commands[comind].action == cmd_file) ||
816 (commands[comind].action == cmd_directory))
817 && ret && (*val == '~' && ISSEP (val[1])))
819 pstring = commands[comind].place;
823 homelen = strlen (home);
824 while (homelen && ISSEP (home[homelen - 1]))
825 home[--homelen] = '\0';
827 /* Skip the leading "~/". */
828 for (++val; ISSEP (*val); val++)
830 *pstring = concat_strings (home, "/", val, (char *)0);
836 /* Run command COM with value VAL. If running the command produces an
837 error, report the error and exit.
839 This is intended to be called from main() to modify Wget's behavior
840 through command-line switches. Since COM is hard-coded in main(),
841 it is not canonicalized, and this aborts when COM is not found.
843 If COMIND's are exported to init.h, this function will be changed
844 to accept COMIND directly. */
847 setoptval (const char *com, const char *val, const char *optname)
849 /* Prepend "--" to OPTNAME. */
850 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
853 strcpy (dd_optname + 2, optname);
855 assert (val != NULL);
856 if (!setval_internal (command_by_name (com), dd_optname, val))
860 /* Parse OPT into command and value and run it. For example,
861 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
862 This is used by the `--execute' flag in main.c. */
865 run_command (const char *opt)
869 switch (parse_line (opt, &com, &val, &comind))
872 if (!setval_internal (comind, com, val))
878 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
879 exec_name, quote (opt));
884 /* Generic helper functions, for use with `commands'. */
886 /* Forward declarations: */
891 static bool decode_string (const char *, const struct decode_item *, int, int *);
892 static bool simple_atoi (const char *, const char *, int *);
893 static bool simple_atof (const char *, const char *, double *);
895 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
897 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
898 && c_tolower((p)[1]) == (c1) \
901 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
902 && c_tolower((p)[1]) == (c1) \
903 && c_tolower((p)[2]) == (c2) \
907 /* Store the boolean value from VAL to PLACE. COM is ignored,
908 except for error messages. */
910 cmd_boolean (const char *com, const char *val, void *place)
914 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
915 /* "on", "yes" and "1" mean true. */
917 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
918 /* "off", "no" and "0" mean false. */
923 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
924 exec_name, com, quote (val));
928 *(bool *) place = value;
932 /* Set the non-negative integer value from VAL to PLACE. With
933 incorrect specification, the number remains unchanged. */
935 cmd_number (const char *com, const char *val, void *place)
937 if (!simple_atoi (val, val + strlen (val), place)
938 || *(int *) place < 0)
940 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
941 exec_name, com, quote (val));
947 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
949 cmd_number_inf (const char *com, const char *val, void *place)
951 if (!strcasecmp (val, "inf"))
956 return cmd_number (com, val, place);
959 /* Copy (strdup) the string at COM to a new location and place a
960 pointer to *PLACE. */
962 cmd_string (const char *com, const char *val, void *place)
964 char **pstring = (char **)place;
966 xfree_null (*pstring);
967 *pstring = xstrdup (val);
971 /* Like cmd_string but ensure the string is upper case. */
973 cmd_string_uppercase (const char *com, const char *val, void *place)
976 pstring = (char **)place;
977 xfree_null (*pstring);
979 *pstring = xmalloc (strlen (val) + 1);
981 for (q = *pstring; *val; val++, q++)
982 *q = c_toupper (*val);
989 /* Like cmd_string, but handles tilde-expansion when reading a user's
990 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
991 gets expanded to the user's home directory. */
993 cmd_file (const char *com, const char *val, void *place)
995 char **pstring = (char **)place;
997 xfree_null (*pstring);
999 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
1001 *pstring = xstrdup (val);
1003 #if defined(WINDOWS) || defined(MSDOS)
1004 /* Convert "\" to "/". */
1007 for (s = *pstring; *s; s++)
1015 /* Like cmd_file, but strips trailing '/' characters. */
1017 cmd_directory (const char *com, const char *val, void *place)
1021 /* Call cmd_file() for tilde expansion and separator
1022 canonicalization (backslash -> slash under Windows). These
1023 things should perhaps be in a separate function. */
1024 if (!cmd_file (com, val, place))
1027 s = *(char **)place;
1029 while (t > s && *--t == '/')
1035 /* Split VAL by space to a vector of values, and append those values
1036 to vector pointed to by the PLACE argument. If VAL is empty, the
1037 PLACE vector is cleared instead. */
1040 cmd_vector (const char *com, const char *val, void *place)
1042 char ***pvec = (char ***)place;
1045 *pvec = merge_vecs (*pvec, sepstring (val));
1055 cmd_directory_vector (const char *com, const char *val, void *place)
1057 char ***pvec = (char ***)place;
1061 /* Strip the trailing slashes from directories. */
1064 seps = sepstring (val);
1065 for (t = seps; t && *t; t++)
1067 int len = strlen (*t);
1068 /* Skip degenerate case of root directory. */
1071 if ((*t)[len - 1] == '/')
1072 (*t)[len - 1] = '\0';
1075 *pvec = merge_vecs (*pvec, seps);
1085 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1086 "100k" or "2.5G" to a floating point number. */
1089 parse_bytes_helper (const char *val, double *result)
1091 double number, mult;
1092 const char *end = val + strlen (val);
1094 /* Check for "inf". */
1095 if (0 == strcmp (val, "inf"))
1101 /* Strip trailing whitespace. */
1102 while (val < end && c_isspace (end[-1]))
1107 switch (c_tolower (end[-1]))
1110 --end, mult = 1024.0;
1113 --end, mult = 1048576.0;
1116 --end, mult = 1073741824.0;
1119 --end, mult = 1099511627776.0;
1122 /* Not a recognized suffix: assume it's a digit. (If not,
1123 simple_atof will raise an error.) */
1127 /* Skip leading and trailing whitespace. */
1128 while (val < end && c_isspace (*val))
1130 while (val < end && c_isspace (end[-1]))
1135 if (!simple_atof (val, end, &number) || number < 0)
1138 *result = number * mult;
1142 /* Parse VAL as a number and set its value to PLACE (which should
1145 By default, the value is assumed to be in bytes. If "K", "M", or
1146 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1147 1<<30, respectively. Floating point values are allowed and are
1148 cast to integer before use. The idea is to be able to use things
1149 like 1.5k instead of "1536".
1151 The string "inf" is returned as 0.
1153 In case of error, false is returned and memory pointed to by PLACE
1154 remains unmodified. */
1157 cmd_bytes (const char *com, const char *val, void *place)
1160 if (!parse_bytes_helper (val, &byte_value))
1162 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1163 exec_name, com, quote (val));
1166 *(wgint *)place = (wgint)byte_value;
1170 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1171 SIZE_SUM. It works by converting the string to double, therefore
1172 working with values up to 2^53-1 without loss of precision. This
1173 value (8192 TB) is large enough to serve for a while. */
1176 cmd_bytes_sum (const char *com, const char *val, void *place)
1179 if (!parse_bytes_helper (val, &byte_value))
1181 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1182 exec_name, com, quote (val));
1185 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1189 /* Store the value of VAL to *OUT. The value is a time period, by
1190 default expressed in seconds, but also accepting suffixes "m", "h",
1191 "d", and "w" for minutes, hours, days, and weeks respectively. */
1194 cmd_time (const char *com, const char *val, void *place)
1196 double number, mult;
1197 const char *end = val + strlen (val);
1199 /* Strip trailing whitespace. */
1200 while (val < end && c_isspace (end[-1]))
1206 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1207 exec_name, com, quote (val));
1211 switch (c_tolower (end[-1]))
1214 --end, mult = 1; /* seconds */
1217 --end, mult = 60; /* minutes */
1220 --end, mult = 3600; /* hours */
1223 --end, mult = 86400.0; /* days */
1226 --end, mult = 604800.0; /* weeks */
1229 /* Not a recognized suffix: assume it belongs to the number.
1230 (If not, simple_atof will raise an error.) */
1234 /* Skip leading and trailing whitespace. */
1235 while (val < end && c_isspace (*val))
1237 while (val < end && c_isspace (end[-1]))
1242 if (!simple_atof (val, end, &number))
1245 *(double *)place = number * mult;
1251 cmd_cert_type (const char *com, const char *val, void *place)
1253 static const struct decode_item choices[] = {
1254 { "pem", keyfile_pem },
1255 { "der", keyfile_asn1 },
1256 { "asn1", keyfile_asn1 },
1258 int ok = decode_string (val, choices, countof (choices), place);
1260 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1265 /* Specialized helper functions, used by `commands' to handle some
1266 options specially. */
1268 static bool check_user_specified_header (const char *);
1271 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1273 if (!cmd_boolean (com, val, &opt.dirstruct))
1275 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1276 must be affected inversely. */
1278 opt.no_dirstruct = false;
1280 opt.no_dirstruct = true;
1285 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1287 /* Empty value means reset the list of headers. */
1290 free_vec (opt.user_headers);
1291 opt.user_headers = NULL;
1295 if (!check_user_specified_header (val))
1297 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1298 exec_name, com, quote (val));
1301 opt.user_headers = vec_append (opt.user_headers, val);
1306 cmd_spec_warc_header (const char *com, const char *val, void *place_ignored)
1308 /* Empty value means reset the list of headers. */
1311 free_vec (opt.warc_user_headers);
1312 opt.warc_user_headers = NULL;
1316 if (!check_user_specified_header (val))
1318 fprintf (stderr, _("%s: %s: Invalid WARC header %s.\n"),
1319 exec_name, com, quote (val));
1322 opt.warc_user_headers = vec_append (opt.warc_user_headers, val);
1327 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1329 int flag = cmd_boolean (com, val, &opt.htmlify);
1330 if (flag && !opt.htmlify)
1331 opt.remove_listing = false;
1335 /* Set the "mirror" mode. It means: recursive download, timestamping,
1336 no limit on max. recursion depth, and don't remove listings. */
1339 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1343 if (!cmd_boolean (com, val, &mirror))
1347 opt.recursive = true;
1348 if (!opt.no_dirstruct)
1349 opt.dirstruct = true;
1350 opt.timestamping = true;
1351 opt.reclevel = INFINITE_RECURSION;
1352 opt.remove_listing = false;
1357 /* Validate --prefer-family and set the choice. Allowed values are
1358 "IPv4", "IPv6", and "none". */
1361 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1363 static const struct decode_item choices[] = {
1364 { "IPv4", prefer_ipv4 },
1365 { "IPv6", prefer_ipv6 },
1366 { "none", prefer_none },
1368 int prefer_family = prefer_none;
1369 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1371 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1372 opt.prefer_family = prefer_family;
1376 /* Set progress.type to VAL, but verify that it's a valid progress
1377 implementation before that. */
1380 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1382 if (!valid_progress_implementation_p (val))
1384 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1385 exec_name, com, quote (val));
1388 xfree_null (opt.progress_type);
1390 /* Don't call set_progress_implementation here. It will be called
1391 in main() when it becomes clear what the log output is. */
1392 opt.progress_type = xstrdup (val);
1396 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1397 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1401 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1403 if (!cmd_boolean (com, val, &opt.recursive))
1407 if (opt.recursive && !opt.no_dirstruct)
1408 opt.dirstruct = true;
1413 /* Validate --regex-type and set the choice. */
1416 cmd_spec_regex_type (const char *com, const char *val, void *place_ignored)
1418 static const struct decode_item choices[] = {
1419 { "posix", regex_type_posix },
1421 { "pcre", regex_type_pcre },
1424 int regex_type = regex_type_posix;
1425 int ok = decode_string (val, choices, countof (choices), ®ex_type);
1427 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1428 opt.regex_type = regex_type;
1433 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1435 int restrict_os = opt.restrict_files_os;
1436 int restrict_ctrl = opt.restrict_files_ctrl;
1437 int restrict_case = opt.restrict_files_case;
1438 int restrict_nonascii = opt.restrict_files_nonascii;
1442 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1446 end = strchr (val, ',');
1448 end = val + strlen (val);
1450 if (VAL_IS ("unix"))
1451 restrict_os = restrict_unix;
1452 else if (VAL_IS ("windows"))
1453 restrict_os = restrict_windows;
1454 else if (VAL_IS ("lowercase"))
1455 restrict_case = restrict_lowercase;
1456 else if (VAL_IS ("uppercase"))
1457 restrict_case = restrict_uppercase;
1458 else if (VAL_IS ("nocontrol"))
1459 restrict_ctrl = false;
1460 else if (VAL_IS ("ascii"))
1461 restrict_nonascii = true;
1464 fprintf (stderr, _("\
1465 %s: %s: Invalid restriction %s,\n\
1466 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1467 exec_name, com, quote (val));
1474 while (*val && *end);
1478 opt.restrict_files_os = restrict_os;
1479 opt.restrict_files_ctrl = restrict_ctrl;
1480 opt.restrict_files_case = restrict_case;
1481 opt.restrict_files_nonascii = restrict_nonascii;
1487 cmd_spec_report_speed (const char *com, const char *val, void *place_ignored)
1489 opt.report_bps = strcasecmp (val, "bits") == 0;
1490 if (!opt.report_bps)
1491 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1492 return opt.report_bps;
1497 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1499 static const struct decode_item choices[] = {
1500 { "auto", secure_protocol_auto },
1501 { "sslv2", secure_protocol_sslv2 },
1502 { "sslv3", secure_protocol_sslv3 },
1503 { "tlsv1", secure_protocol_tlsv1 },
1504 { "pfs", secure_protocol_pfs },
1506 int ok = decode_string (val, choices, countof (choices), place);
1508 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1513 /* Set all three timeout values. */
1516 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1519 if (!cmd_time (com, val, &value))
1521 opt.read_timeout = value;
1522 opt.connect_timeout = value;
1523 opt.dns_timeout = value;
1528 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1530 /* Disallow embedded newlines. */
1531 if (strchr (val, '\n'))
1533 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1534 exec_name, com, quote (val));
1537 xfree_null (opt.useragent);
1538 opt.useragent = xstrdup (val);
1542 /* The "verbose" option cannot be cmd_boolean because the variable is
1543 not bool -- it's of type int (-1 means uninitialized because of
1544 some random hackery for disallowing -q -v). */
1547 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1550 if (cmd_boolean (com, val, &flag))
1558 /* Miscellaneous useful routines. */
1560 /* A very simple atoi clone, more useful than atoi because it works on
1561 delimited strings, and has error reportage. Returns true on success,
1562 false on failure. If successful, stores result to *DEST. */
1565 simple_atoi (const char *beg, const char *end, int *dest)
1568 bool negative = false;
1569 const char *p = beg;
1571 while (p < end && c_isspace (*p))
1573 if (p < end && (*p == '-' || *p == '+'))
1575 negative = (*p == '-');
1581 /* Read negative numbers in a separate loop because the most
1582 negative integer cannot be represented as a positive number. */
1585 for (; p < end && c_isdigit (*p); p++)
1587 int next = (10 * result) + (*p - '0');
1589 return false; /* overflow */
1593 for (; p < end && c_isdigit (*p); p++)
1595 int next = (10 * result) - (*p - '0');
1597 return false; /* underflow */
1608 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1609 doesn't handle exponential notation. Returns true on success,
1610 false on failure. In case of success, stores its result to
1614 simple_atof (const char *beg, const char *end, double *dest)
1618 bool negative = false;
1619 bool seen_dot = false;
1620 bool seen_digit = false;
1623 const char *p = beg;
1625 while (p < end && c_isspace (*p))
1627 if (p < end && (*p == '-' || *p == '+'))
1629 negative = (*p == '-');
1633 for (; p < end; p++)
1639 result = (10 * result) + (ch - '0');
1641 result += (ch - '0') / (divider *= 10);
1663 /* Verify that the user-specified header in S is valid. It must
1664 contain a colon preceded by non-white-space characters and must not
1665 contain newlines. */
1668 check_user_specified_header (const char *s)
1672 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1674 /* The header MUST contain `:' preceded by at least one
1675 non-whitespace character. */
1676 if (*p != ':' || p == s)
1678 /* The header MUST NOT contain newlines. */
1679 if (strchr (s, '\n'))
1684 /* Decode VAL into a number, according to ITEMS. */
1687 decode_string (const char *val, const struct decode_item *items, int itemcount,
1691 for (i = 0; i < itemcount; i++)
1692 if (0 == strcasecmp (val, items[i].name))
1694 *place = items[i].code;
1701 void cleanup_html_url (void);
1702 void spider_cleanup (void);
1705 /* Free the memory allocated by global variables. */
1709 /* Free external resources, close files, etc. */
1711 /* Close WARC file. */
1712 if (opt.warc_filename != 0)
1718 if (fclose (output_stream) == EOF)
1719 inform_exit_status (CLOSEFAILED);
1721 /* No need to check for error because Wget flushes its output (and
1722 checks for errors) after any data arrives. */
1724 /* We're exiting anyway so there's no real need to call free()
1725 hundreds of times. Skipping the frees will make Wget exit
1728 However, when detecting leaks, it's crucial to free() everything
1729 because then you can find the real leaks, i.e. the allocated
1730 memory which grows with the size of the program. */
1736 cleanup_html_url ();
1741 for (i = 0; i < nurl; i++)
1745 extern acc_t *netrc_list;
1746 free_netrc (netrc_list);
1748 xfree_null (opt.choose_config);
1749 xfree_null (opt.lfilename);
1750 xfree_null (opt.dir_prefix);
1751 xfree_null (opt.input_filename);
1752 xfree_null (opt.output_document);
1753 free_vec (opt.accepts);
1754 free_vec (opt.rejects);
1755 free_vec (opt.excludes);
1756 free_vec (opt.includes);
1757 free_vec (opt.domains);
1758 free_vec (opt.follow_tags);
1759 free_vec (opt.ignore_tags);
1760 xfree_null (opt.progress_type);
1761 xfree_null (opt.ftp_user);
1762 xfree_null (opt.ftp_passwd);
1763 xfree_null (opt.ftp_proxy);
1764 xfree_null (opt.https_proxy);
1765 xfree_null (opt.http_proxy);
1766 free_vec (opt.no_proxy);
1767 xfree_null (opt.useragent);
1768 xfree_null (opt.referer);
1769 xfree_null (opt.http_user);
1770 xfree_null (opt.http_passwd);
1771 free_vec (opt.user_headers);
1772 free_vec (opt.warc_user_headers);
1774 xfree_null (opt.cert_file);
1775 xfree_null (opt.private_key);
1776 xfree_null (opt.ca_directory);
1777 xfree_null (opt.ca_cert);
1778 xfree_null (opt.random_file);
1779 xfree_null (opt.egd_file);
1781 xfree_null (opt.bind_address);
1782 xfree_null (opt.cookies_input);
1783 xfree_null (opt.cookies_output);
1784 xfree_null (opt.user);
1785 xfree_null (opt.passwd);
1786 xfree_null (opt.base_href);
1787 xfree_null (opt.method);
1789 #endif /* DEBUG_MALLOC */
1792 /* Unit testing routines. */
1797 test_commands_sorted()
1799 int prev_idx = 0, next_idx = 1;
1800 int command_count = countof (commands) - 1;
1802 while (next_idx <= command_count)
1804 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1807 mu_assert ("FAILED", false);
1820 test_cmd_spec_restrict_file_names()
1825 int expected_restrict_files_os;
1826 int expected_restrict_files_ctrl;
1827 int expected_restrict_files_case;
1830 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1831 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1832 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1833 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1836 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1841 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1844 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1845 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1846 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1847 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1849 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1850 res == test_array[i].result
1851 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1852 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1853 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1859 #endif /* TESTING */