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_file);
91 CMD_DECLARE (cmd_directory);
92 CMD_DECLARE (cmd_time);
93 CMD_DECLARE (cmd_vector);
95 CMD_DECLARE (cmd_spec_dirstruct);
96 CMD_DECLARE (cmd_spec_header);
97 CMD_DECLARE (cmd_spec_warc_header);
98 CMD_DECLARE (cmd_spec_htmlify);
99 CMD_DECLARE (cmd_spec_mirror);
100 CMD_DECLARE (cmd_spec_prefer_family);
101 CMD_DECLARE (cmd_spec_progress);
102 CMD_DECLARE (cmd_spec_recursive);
103 CMD_DECLARE (cmd_spec_regex_type);
104 CMD_DECLARE (cmd_spec_restrict_file_names);
105 CMD_DECLARE (cmd_spec_report_speed);
107 CMD_DECLARE (cmd_spec_secure_protocol);
109 CMD_DECLARE (cmd_spec_timeout);
110 CMD_DECLARE (cmd_spec_useragent);
111 CMD_DECLARE (cmd_spec_verbose);
113 /* List of recognized commands, each consisting of name, place and
114 function. When adding a new command, simply add it to the list,
115 but be sure to keep the list sorted alphabetically, as
116 command_by_name's binary search depends on it. Also, be sure to
117 add any entries that allocate memory (e.g. cmd_string and
118 cmd_vector) to the cleanup() function below. */
120 static const struct {
123 bool (*action) (const char *, const char *, void *);
125 /* KEEP THIS LIST ALPHABETICALLY SORTED */
126 { "accept", &opt.accepts, cmd_vector },
127 { "acceptregex", &opt.acceptregex_s, cmd_string },
128 { "addhostdir", &opt.add_hostdir, cmd_boolean },
129 { "adjustextension", &opt.adjust_extension, cmd_boolean },
130 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
131 { "askpassword", &opt.ask_passwd, cmd_boolean },
132 { "authnochallenge", &opt.auth_without_challenge,
134 { "background", &opt.background, cmd_boolean },
135 { "backupconverted", &opt.backup_converted, cmd_boolean },
136 { "backups", &opt.backups, cmd_number },
137 { "base", &opt.base_href, cmd_string },
138 { "bindaddress", &opt.bind_address, cmd_string },
139 { "bodydata", &opt.body_data, cmd_string },
140 { "bodyfile", &opt.body_file, cmd_string },
142 { "cacertificate", &opt.ca_cert, cmd_file },
144 { "cache", &opt.allow_cache, cmd_boolean },
146 { "cadirectory", &opt.ca_directory, cmd_directory },
147 { "certificate", &opt.cert_file, cmd_file },
148 { "certificatetype", &opt.cert_type, cmd_cert_type },
149 { "checkcertificate", &opt.check_cert, cmd_boolean },
151 { "chooseconfig", &opt.choose_config, cmd_file },
152 { "connecttimeout", &opt.connect_timeout, cmd_time },
153 { "contentdisposition", &opt.content_disposition, cmd_boolean },
154 { "contentonerror", &opt.content_on_error, cmd_boolean },
155 { "continue", &opt.always_rest, cmd_boolean },
156 { "convertlinks", &opt.convert_links, cmd_boolean },
157 { "cookies", &opt.cookies, cmd_boolean },
158 { "cutdirs", &opt.cut_dirs, cmd_number },
160 { "debug", &opt.debug, cmd_boolean },
162 { "defaultpage", &opt.default_page, cmd_string },
163 { "deleteafter", &opt.delete_after, cmd_boolean },
164 { "dirprefix", &opt.dir_prefix, cmd_directory },
165 { "dirstruct", NULL, cmd_spec_dirstruct },
166 { "dnscache", &opt.dns_cache, cmd_boolean },
167 { "dnstimeout", &opt.dns_timeout, cmd_time },
168 { "domains", &opt.domains, cmd_vector },
169 { "dotbytes", &opt.dot_bytes, cmd_bytes },
170 { "dotsinline", &opt.dots_in_line, cmd_number },
171 { "dotspacing", &opt.dot_spacing, cmd_number },
172 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
174 { "egdfile", &opt.egd_file, cmd_file },
176 { "excludedirectories", &opt.excludes, cmd_directory_vector },
177 { "excludedomains", &opt.exclude_domains, cmd_vector },
178 { "followftp", &opt.follow_ftp, cmd_boolean },
179 { "followtags", &opt.follow_tags, cmd_vector },
180 { "forcehtml", &opt.force_html, cmd_boolean },
181 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
182 { "ftppassword", &opt.ftp_passwd, cmd_string },
183 { "ftpproxy", &opt.ftp_proxy, cmd_string },
185 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
186 #endif /* def __VMS */
187 { "ftpuser", &opt.ftp_user, cmd_string },
188 { "glob", &opt.ftp_glob, cmd_boolean },
189 { "header", NULL, cmd_spec_header },
190 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
191 { "htmlify", NULL, cmd_spec_htmlify },
192 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
193 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
194 { "httppassword", &opt.http_passwd, cmd_string },
195 { "httpproxy", &opt.http_proxy, cmd_string },
196 { "httpsproxy", &opt.https_proxy, cmd_string },
197 { "httpuser", &opt.http_user, cmd_string },
198 { "ignorecase", &opt.ignore_case, cmd_boolean },
199 { "ignorelength", &opt.ignore_length, cmd_boolean },
200 { "ignoretags", &opt.ignore_tags, cmd_vector },
201 { "includedirectories", &opt.includes, cmd_directory_vector },
203 { "inet4only", &opt.ipv4_only, cmd_boolean },
204 { "inet6only", &opt.ipv6_only, cmd_boolean },
206 { "input", &opt.input_filename, cmd_file },
207 { "iri", &opt.enable_iri, cmd_boolean },
208 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
209 { "limitrate", &opt.limit_rate, cmd_bytes },
210 { "loadcookies", &opt.cookies_input, cmd_file },
211 { "localencoding", &opt.locale, cmd_string },
212 { "logfile", &opt.lfilename, cmd_file },
213 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
214 { "maxredirect", &opt.max_redirect, cmd_number },
215 { "method", &opt.method, cmd_string },
216 { "mirror", NULL, cmd_spec_mirror },
217 { "netrc", &opt.netrc, cmd_boolean },
218 { "noclobber", &opt.noclobber, cmd_boolean },
219 { "noparent", &opt.no_parent, cmd_boolean },
220 { "noproxy", &opt.no_proxy, cmd_vector },
221 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
222 { "outputdocument", &opt.output_document, cmd_file },
223 { "pagerequisites", &opt.page_requisites, cmd_boolean },
224 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
225 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
226 { "password", &opt.passwd, cmd_string },
227 { "postdata", &opt.post_data, cmd_string },
228 { "postfile", &opt.post_file_name, cmd_file },
229 { "preferfamily", NULL, cmd_spec_prefer_family },
230 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
232 { "privatekey", &opt.private_key, cmd_file },
233 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
235 { "progress", &opt.progress_type, cmd_spec_progress },
236 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
237 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
238 { "proxypassword", &opt.proxy_passwd, cmd_string },
239 { "proxyuser", &opt.proxy_user, cmd_string },
240 { "quiet", &opt.quiet, cmd_boolean },
241 { "quota", &opt.quota, cmd_bytes_sum },
243 { "randomfile", &opt.random_file, cmd_file },
245 { "randomwait", &opt.random_wait, cmd_boolean },
246 { "readtimeout", &opt.read_timeout, cmd_time },
247 { "reclevel", &opt.reclevel, cmd_number_inf },
248 { "recursive", NULL, cmd_spec_recursive },
249 { "referer", &opt.referer, cmd_string },
250 { "regextype", &opt.regex_type, cmd_spec_regex_type },
251 { "reject", &opt.rejects, cmd_vector },
252 { "rejectregex", &opt.rejectregex_s, cmd_string },
253 { "relativeonly", &opt.relative_only, cmd_boolean },
254 { "remoteencoding", &opt.encoding_remote, cmd_string },
255 { "removelisting", &opt.remove_listing, cmd_boolean },
256 { "reportspeed", &opt.report_bps, cmd_spec_report_speed},
257 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
258 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
259 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
260 { "robots", &opt.use_robots, cmd_boolean },
261 { "savecookies", &opt.cookies_output, cmd_file },
262 { "saveheaders", &opt.save_headers, cmd_boolean },
264 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
266 { "serverresponse", &opt.server_response, cmd_boolean },
267 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
268 { "spanhosts", &opt.spanhost, cmd_boolean },
269 { "spider", &opt.spider, cmd_boolean },
270 { "strictcomments", &opt.strict_comments, cmd_boolean },
271 { "timeout", NULL, cmd_spec_timeout },
272 { "timestamping", &opt.timestamping, cmd_boolean },
273 { "tries", &opt.ntry, cmd_number_inf },
274 { "trustservernames", &opt.trustservernames, cmd_boolean },
275 { "unlink", &opt.unlink, cmd_boolean },
276 { "useproxy", &opt.use_proxy, cmd_boolean },
277 { "user", &opt.user, cmd_string },
278 { "useragent", NULL, cmd_spec_useragent },
279 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
280 { "verbose", NULL, cmd_spec_verbose },
281 { "wait", &opt.wait, cmd_time },
282 { "waitretry", &opt.waitretry, cmd_time },
283 { "warccdx", &opt.warc_cdx_enabled, cmd_boolean },
284 { "warccdxdedup", &opt.warc_cdx_dedup_filename, cmd_file },
286 { "warccompression", &opt.warc_compression_enabled, cmd_boolean },
288 { "warcdigests", &opt.warc_digests_enabled, cmd_boolean },
289 { "warcfile", &opt.warc_filename, cmd_file },
290 { "warcheader", NULL, cmd_spec_warc_header },
291 { "warckeeplog", &opt.warc_keep_log, cmd_boolean },
292 { "warcmaxsize", &opt.warc_maxsize, cmd_bytes },
293 { "warctempdir", &opt.warc_tempdir, cmd_directory },
295 { "wdebug", &opt.wdebug, cmd_boolean },
299 /* Look up CMDNAME in the commands[] and return its position in the
300 array. If CMDNAME is not found, return -1. */
303 command_by_name (const char *cmdname)
305 /* Use binary search for speed. Wget has ~100 commands, which
306 guarantees a worst case performance of 7 string comparisons. */
307 int lo = 0, hi = countof (commands) - 1;
311 int mid = (lo + hi) >> 1;
312 int cmp = strcasecmp (cmdname, commands[mid].name);
323 /* Reset the variables to default values. */
329 /* Most of the default values are 0 (and 0.0, NULL, and false).
330 Just reset everything, and fill in the non-zero values. Note
331 that initializing pointers to NULL this way is technically
332 illegal, but porting Wget to a machine where NULL is not all-zero
333 bit pattern will be the least of the implementors' worries. */
340 opt.add_hostdir = true;
344 opt.http_keep_alive = true;
345 opt.use_proxy = true;
346 tmp = getenv ("no_proxy");
348 opt.no_proxy = sepstring (tmp);
349 opt.prefer_family = prefer_none;
350 opt.allow_cache = true;
352 opt.read_timeout = 900;
353 opt.use_robots = true;
355 opt.remove_listing = true;
357 opt.dot_bytes = 1024;
358 opt.dot_spacing = 10;
359 opt.dots_in_line = 50;
361 opt.dns_cache = true;
365 opt.check_cert = true;
368 /* The default for file name restriction defaults to the OS type. */
369 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
370 opt.restrict_files_os = restrict_windows;
372 opt.restrict_files_os = restrict_unix;
374 opt.restrict_files_ctrl = true;
375 opt.restrict_files_nonascii = false;
376 opt.restrict_files_case = restrict_no_case_restriction;
378 opt.regex_type = regex_type_posix;
380 opt.max_redirect = 20;
385 opt.enable_iri = true;
387 opt.enable_iri = false;
390 opt.encoding_remote = NULL;
392 opt.useservertimestamps = true;
393 opt.show_all_dns_entries = false;
395 opt.warc_maxsize = 0; /* 1024 * 1024 * 1024; */
397 opt.warc_compression_enabled = true;
399 opt.warc_compression_enabled = false;
401 opt.warc_digests_enabled = true;
402 opt.warc_cdx_enabled = false;
403 opt.warc_cdx_dedup_filename = NULL;
404 opt.warc_tempdir = NULL;
405 opt.warc_keep_log = true;
408 /* Return the user's home directory (strdup-ed), or NULL if none is
413 static char *buf = NULL;
414 static char *home, *ret;
418 home = getenv ("HOME");
424 /* Under MSDOS, if $HOME isn't defined, use the directory where
425 `wget.exe' resides. */
426 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
429 buff = _w32_get_argv0 ();
431 p = strrchr (buf, '/'); /* djgpp */
433 p = strrchr (buf, '\\'); /* others */
437 buff = malloc (len + 1);
441 strncpy (buff, _w32_get_argv0 (), len);
445 #elif !defined(WINDOWS)
446 /* If HOME is not defined, try getting it from the password
448 struct passwd *pwd = getpwuid (getuid ());
449 if (!pwd || !pwd->pw_dir)
453 /* Under Windows, if $HOME isn't defined, use the directory where
454 `wget.exe' resides. */
460 ret = home ? xstrdup (home) : NULL;
467 /* Check the 'WGETRC' environment variable and return the file name
468 if 'WGETRC' is set and is a valid file.
469 If the `WGETRC' variable exists but the file does not exist, the
470 function will exit(). */
472 wgetrc_env_file_name (void)
474 char *env = getenv ("WGETRC");
477 if (!file_exists_p (env))
479 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
483 return xstrdup (env);
488 /* Check for the existance of '$HOME/.wgetrc' and return its path
489 if it exists and is set. */
491 wgetrc_user_file_name (void)
495 /* If that failed, try $HOME/.wgetrc (or equivalent). */
498 file = "SYS$LOGIN:.wgetrc";
499 #else /* def __VMS */
502 file = aprintf ("%s/.wgetrc", home);
504 #endif /* def __VMS [else] */
508 if (!file_exists_p (file))
516 /* Return the path to the user's .wgetrc. This is either the value of
517 `WGETRC' environment variable, or `$HOME/.wgetrc'.
519 Additionally, for windows, look in the directory where wget.exe
522 wgetrc_file_name (void)
524 char *file = wgetrc_env_file_name ();
528 file = wgetrc_user_file_name ();
531 /* Under Windows, if we still haven't found .wgetrc, look for the file
532 `wget.ini' in the directory where `wget.exe' resides; we do this for
533 backward compatibility with previous versions of Wget.
534 SYSTEM_WGETRC should not be defined under WINDOWS. */
537 char *home = home_dir ();
543 file = aprintf ("%s/wget.ini", home);
544 if (!file_exists_p (file))
557 /* Return values of parse_line. */
565 static enum parse_line parse_line (const char *, char **, char **, int *);
566 static bool setval_internal (int, const char *, const char *);
567 static bool setval_internal_tilde (int, const char *, const char *);
569 /* Initialize variables from a wgetrc file. Returns zero (failure) if
570 there were errors in the file. */
573 run_wgetrc (const char *file)
580 fp = fopen (file, "r");
583 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
584 file, strerror (errno));
585 return true; /* not a fatal error */
588 while ((line = read_whole_line (fp)) != NULL)
590 char *com = NULL, *val = NULL;
593 /* Parse the line. */
594 switch (parse_line (line, &com, &val, &comind))
597 /* If everything is OK, set the value. */
598 if (!setval_internal_tilde (comind, com, val))
600 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
601 exec_name, file, ln);
605 case line_syntax_error:
606 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
607 exec_name, file, ln);
610 case line_unknown_command:
611 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
612 exec_name, quote (com), file, ln);
630 /* Initialize the defaults and run the system wgetrc and user's own
635 char *file, *env_sysrc;
638 /* Run a non-standard system rc file when the according environment
639 variable has been set. For internal testing purposes only! */
640 env_sysrc = getenv ("SYSTEM_WGETRC");
641 if (env_sysrc && file_exists_p (env_sysrc))
643 ok &= run_wgetrc (env_sysrc);
644 /* If there are any problems parsing the system wgetrc file, tell
648 fprintf (stderr, _("\
649 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
651 or specify a different file using --config.\n"), env_sysrc);
655 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
657 else if (file_exists_p (SYSTEM_WGETRC))
658 ok &= run_wgetrc (SYSTEM_WGETRC);
659 /* If there are any problems parsing the system wgetrc file, tell
663 fprintf (stderr, _("\
664 Parsing system wgetrc file failed. Please check\n\
666 or specify a different file using --config.\n"), SYSTEM_WGETRC);
670 /* Override it with your own, if one exists. */
671 file = wgetrc_file_name ();
674 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
675 something like realpath() before comparing them with `strcmp' */
677 if (!strcmp (file, SYSTEM_WGETRC))
679 fprintf (stderr, _("\
680 %s: Warning: Both system and user wgetrc point to %s.\n"),
681 exec_name, quote (file));
685 ok &= run_wgetrc (file);
687 /* If there were errors processing either `.wgetrc', abort. */
695 /* Remove dashes and underscores from S, modifying S in the
701 char *t = s; /* t - tortoise */
702 char *h = s; /* h - hare */
704 if (*h == '_' || *h == '-')
711 /* Parse the line pointed by line, with the syntax:
712 <sp>* command <sp>* = <sp>* value <sp>*
713 Uses malloc to allocate space for command and value.
715 Returns one of line_ok, line_empty, line_syntax_error, or
716 line_unknown_command.
718 In case of line_ok, *COM and *VAL point to freshly allocated
719 strings, and *COMIND points to com's index. In case of error or
720 empty line, their values are unmodified. */
722 static enum parse_line
723 parse_line (const char *line, char **com, char **val, int *comind)
726 const char *end = line + strlen (line);
727 const char *cmdstart, *cmdend;
728 const char *valstart, *valend;
733 /* Skip leading and trailing whitespace. */
734 while (*line && c_isspace (*line))
736 while (end > line && c_isspace (end[-1]))
739 /* Skip empty lines and comments. */
740 if (!*line || *line == '#')
746 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
750 /* Skip '=', as well as any space before or after it. */
751 while (p < end && c_isspace (*p))
753 if (p == end || *p != '=')
754 return line_syntax_error;
756 while (p < end && c_isspace (*p))
762 /* The syntax is valid (even though the command might not be). Fill
763 in the command name and value. */
764 *com = strdupdelim (cmdstart, cmdend);
765 *val = strdupdelim (valstart, valend);
767 /* The line now known to be syntactically correct. Check whether
768 the command is valid. */
769 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
771 ind = command_by_name (cmdcopy);
773 return line_unknown_command;
775 /* Report success to the caller. */
780 #if defined(WINDOWS) || defined(MSDOS)
781 # define ISSEP(c) ((c) == '/' || (c) == '\\')
783 # define ISSEP(c) ((c) == '/')
786 /* Run commands[comind].action. */
789 setval_internal (int comind, const char *com, const char *val)
791 assert (0 <= comind && ((size_t) comind) < countof (commands));
792 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
793 return commands[comind].action (com, val, commands[comind].place);
797 setval_internal_tilde (int comind, const char *com, const char *val)
803 ret = setval_internal (comind, com, val);
805 /* We make tilde expansion for cmd_file and cmd_directory */
806 if (((commands[comind].action == cmd_file) ||
807 (commands[comind].action == cmd_directory))
808 && ret && (*val == '~' && ISSEP (val[1])))
810 pstring = commands[comind].place;
814 homelen = strlen (home);
815 while (homelen && ISSEP (home[homelen - 1]))
816 home[--homelen] = '\0';
818 /* Skip the leading "~/". */
819 for (++val; ISSEP (*val); val++)
821 *pstring = concat_strings (home, "/", val, (char *)0);
827 /* Run command COM with value VAL. If running the command produces an
828 error, report the error and exit.
830 This is intended to be called from main() to modify Wget's behavior
831 through command-line switches. Since COM is hard-coded in main(),
832 it is not canonicalized, and this aborts when COM is not found.
834 If COMIND's are exported to init.h, this function will be changed
835 to accept COMIND directly. */
838 setoptval (const char *com, const char *val, const char *optname)
840 /* Prepend "--" to OPTNAME. */
841 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
844 strcpy (dd_optname + 2, optname);
846 assert (val != NULL);
847 if (!setval_internal (command_by_name (com), dd_optname, val))
851 /* Parse OPT into command and value and run it. For example,
852 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
853 This is used by the `--execute' flag in main.c. */
856 run_command (const char *opt)
860 switch (parse_line (opt, &com, &val, &comind))
863 if (!setval_internal (comind, com, val))
869 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
870 exec_name, quote (opt));
875 /* Generic helper functions, for use with `commands'. */
877 /* Forward declarations: */
882 static bool decode_string (const char *, const struct decode_item *, int, int *);
883 static bool simple_atoi (const char *, const char *, int *);
884 static bool simple_atof (const char *, const char *, double *);
886 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
888 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
889 && c_tolower((p)[1]) == (c1) \
892 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
893 && c_tolower((p)[1]) == (c1) \
894 && c_tolower((p)[2]) == (c2) \
898 /* Store the boolean value from VAL to PLACE. COM is ignored,
899 except for error messages. */
901 cmd_boolean (const char *com, const char *val, void *place)
905 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
906 /* "on", "yes" and "1" mean true. */
908 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
909 /* "off", "no" and "0" mean false. */
914 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
915 exec_name, com, quote (val));
919 *(bool *) place = value;
923 /* Set the non-negative integer value from VAL to PLACE. With
924 incorrect specification, the number remains unchanged. */
926 cmd_number (const char *com, const char *val, void *place)
928 if (!simple_atoi (val, val + strlen (val), place)
929 || *(int *) place < 0)
931 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
932 exec_name, com, quote (val));
938 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
940 cmd_number_inf (const char *com, const char *val, void *place)
942 if (!strcasecmp (val, "inf"))
947 return cmd_number (com, val, place);
950 /* Copy (strdup) the string at COM to a new location and place a
951 pointer to *PLACE. */
953 cmd_string (const char *com, const char *val, void *place)
955 char **pstring = (char **)place;
957 xfree_null (*pstring);
958 *pstring = xstrdup (val);
963 /* Like the above, but handles tilde-expansion when reading a user's
964 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
965 gets expanded to the user's home directory. */
967 cmd_file (const char *com, const char *val, void *place)
969 char **pstring = (char **)place;
971 xfree_null (*pstring);
973 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
975 *pstring = xstrdup (val);
977 #if defined(WINDOWS) || defined(MSDOS)
978 /* Convert "\" to "/". */
981 for (s = *pstring; *s; s++)
989 /* Like cmd_file, but strips trailing '/' characters. */
991 cmd_directory (const char *com, const char *val, void *place)
995 /* Call cmd_file() for tilde expansion and separator
996 canonicalization (backslash -> slash under Windows). These
997 things should perhaps be in a separate function. */
998 if (!cmd_file (com, val, place))
1001 s = *(char **)place;
1003 while (t > s && *--t == '/')
1009 /* Split VAL by space to a vector of values, and append those values
1010 to vector pointed to by the PLACE argument. If VAL is empty, the
1011 PLACE vector is cleared instead. */
1014 cmd_vector (const char *com, const char *val, void *place)
1016 char ***pvec = (char ***)place;
1019 *pvec = merge_vecs (*pvec, sepstring (val));
1029 cmd_directory_vector (const char *com, const char *val, void *place)
1031 char ***pvec = (char ***)place;
1035 /* Strip the trailing slashes from directories. */
1038 seps = sepstring (val);
1039 for (t = seps; t && *t; t++)
1041 int len = strlen (*t);
1042 /* Skip degenerate case of root directory. */
1045 if ((*t)[len - 1] == '/')
1046 (*t)[len - 1] = '\0';
1049 *pvec = merge_vecs (*pvec, seps);
1059 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1060 "100k" or "2.5G" to a floating point number. */
1063 parse_bytes_helper (const char *val, double *result)
1065 double number, mult;
1066 const char *end = val + strlen (val);
1068 /* Check for "inf". */
1069 if (0 == strcmp (val, "inf"))
1075 /* Strip trailing whitespace. */
1076 while (val < end && c_isspace (end[-1]))
1081 switch (c_tolower (end[-1]))
1084 --end, mult = 1024.0;
1087 --end, mult = 1048576.0;
1090 --end, mult = 1073741824.0;
1093 --end, mult = 1099511627776.0;
1096 /* Not a recognized suffix: assume it's a digit. (If not,
1097 simple_atof will raise an error.) */
1101 /* Skip leading and trailing whitespace. */
1102 while (val < end && c_isspace (*val))
1104 while (val < end && c_isspace (end[-1]))
1109 if (!simple_atof (val, end, &number) || number < 0)
1112 *result = number * mult;
1116 /* Parse VAL as a number and set its value to PLACE (which should
1119 By default, the value is assumed to be in bytes. If "K", "M", or
1120 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1121 1<<30, respectively. Floating point values are allowed and are
1122 cast to integer before use. The idea is to be able to use things
1123 like 1.5k instead of "1536".
1125 The string "inf" is returned as 0.
1127 In case of error, false is returned and memory pointed to by PLACE
1128 remains unmodified. */
1131 cmd_bytes (const char *com, const char *val, void *place)
1134 if (!parse_bytes_helper (val, &byte_value))
1136 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1137 exec_name, com, quote (val));
1140 *(wgint *)place = (wgint)byte_value;
1144 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1145 SIZE_SUM. It works by converting the string to double, therefore
1146 working with values up to 2^53-1 without loss of precision. This
1147 value (8192 TB) is large enough to serve for a while. */
1150 cmd_bytes_sum (const char *com, const char *val, void *place)
1153 if (!parse_bytes_helper (val, &byte_value))
1155 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1156 exec_name, com, quote (val));
1159 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1163 /* Store the value of VAL to *OUT. The value is a time period, by
1164 default expressed in seconds, but also accepting suffixes "m", "h",
1165 "d", and "w" for minutes, hours, days, and weeks respectively. */
1168 cmd_time (const char *com, const char *val, void *place)
1170 double number, mult;
1171 const char *end = val + strlen (val);
1173 /* Strip trailing whitespace. */
1174 while (val < end && c_isspace (end[-1]))
1180 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1181 exec_name, com, quote (val));
1185 switch (c_tolower (end[-1]))
1188 --end, mult = 1; /* seconds */
1191 --end, mult = 60; /* minutes */
1194 --end, mult = 3600; /* hours */
1197 --end, mult = 86400.0; /* days */
1200 --end, mult = 604800.0; /* weeks */
1203 /* Not a recognized suffix: assume it belongs to the number.
1204 (If not, simple_atof will raise an error.) */
1208 /* Skip leading and trailing whitespace. */
1209 while (val < end && c_isspace (*val))
1211 while (val < end && c_isspace (end[-1]))
1216 if (!simple_atof (val, end, &number))
1219 *(double *)place = number * mult;
1225 cmd_cert_type (const char *com, const char *val, void *place)
1227 static const struct decode_item choices[] = {
1228 { "pem", keyfile_pem },
1229 { "der", keyfile_asn1 },
1230 { "asn1", keyfile_asn1 },
1232 int ok = decode_string (val, choices, countof (choices), place);
1234 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1239 /* Specialized helper functions, used by `commands' to handle some
1240 options specially. */
1242 static bool check_user_specified_header (const char *);
1245 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1247 if (!cmd_boolean (com, val, &opt.dirstruct))
1249 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1250 must be affected inversely. */
1252 opt.no_dirstruct = false;
1254 opt.no_dirstruct = true;
1259 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1261 /* Empty value means reset the list of headers. */
1264 free_vec (opt.user_headers);
1265 opt.user_headers = NULL;
1269 if (!check_user_specified_header (val))
1271 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1272 exec_name, com, quote (val));
1275 opt.user_headers = vec_append (opt.user_headers, val);
1280 cmd_spec_warc_header (const char *com, const char *val, void *place_ignored)
1282 /* Empty value means reset the list of headers. */
1285 free_vec (opt.warc_user_headers);
1286 opt.warc_user_headers = NULL;
1290 if (!check_user_specified_header (val))
1292 fprintf (stderr, _("%s: %s: Invalid WARC header %s.\n"),
1293 exec_name, com, quote (val));
1296 opt.warc_user_headers = vec_append (opt.warc_user_headers, val);
1301 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1303 int flag = cmd_boolean (com, val, &opt.htmlify);
1304 if (flag && !opt.htmlify)
1305 opt.remove_listing = false;
1309 /* Set the "mirror" mode. It means: recursive download, timestamping,
1310 no limit on max. recursion depth, and don't remove listings. */
1313 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1317 if (!cmd_boolean (com, val, &mirror))
1321 opt.recursive = true;
1322 if (!opt.no_dirstruct)
1323 opt.dirstruct = true;
1324 opt.timestamping = true;
1325 opt.reclevel = INFINITE_RECURSION;
1326 opt.remove_listing = false;
1331 /* Validate --prefer-family and set the choice. Allowed values are
1332 "IPv4", "IPv6", and "none". */
1335 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1337 static const struct decode_item choices[] = {
1338 { "IPv4", prefer_ipv4 },
1339 { "IPv6", prefer_ipv6 },
1340 { "none", prefer_none },
1342 int prefer_family = prefer_none;
1343 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1345 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1346 opt.prefer_family = prefer_family;
1350 /* Set progress.type to VAL, but verify that it's a valid progress
1351 implementation before that. */
1354 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1356 if (!valid_progress_implementation_p (val))
1358 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1359 exec_name, com, quote (val));
1362 xfree_null (opt.progress_type);
1364 /* Don't call set_progress_implementation here. It will be called
1365 in main() when it becomes clear what the log output is. */
1366 opt.progress_type = xstrdup (val);
1370 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1371 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1375 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1377 if (!cmd_boolean (com, val, &opt.recursive))
1381 if (opt.recursive && !opt.no_dirstruct)
1382 opt.dirstruct = true;
1387 /* Validate --regex-type and set the choice. */
1390 cmd_spec_regex_type (const char *com, const char *val, void *place_ignored)
1392 static const struct decode_item choices[] = {
1393 { "posix", regex_type_posix },
1395 { "pcre", regex_type_pcre },
1398 int regex_type = regex_type_posix;
1399 int ok = decode_string (val, choices, countof (choices), ®ex_type);
1401 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1402 opt.regex_type = regex_type;
1407 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1409 int restrict_os = opt.restrict_files_os;
1410 int restrict_ctrl = opt.restrict_files_ctrl;
1411 int restrict_case = opt.restrict_files_case;
1412 int restrict_nonascii = opt.restrict_files_nonascii;
1416 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1420 end = strchr (val, ',');
1422 end = val + strlen (val);
1424 if (VAL_IS ("unix"))
1425 restrict_os = restrict_unix;
1426 else if (VAL_IS ("windows"))
1427 restrict_os = restrict_windows;
1428 else if (VAL_IS ("lowercase"))
1429 restrict_case = restrict_lowercase;
1430 else if (VAL_IS ("uppercase"))
1431 restrict_case = restrict_uppercase;
1432 else if (VAL_IS ("nocontrol"))
1433 restrict_ctrl = false;
1434 else if (VAL_IS ("ascii"))
1435 restrict_nonascii = true;
1438 fprintf (stderr, _("\
1439 %s: %s: Invalid restriction %s,\n\
1440 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1441 exec_name, com, quote (val));
1448 while (*val && *end);
1452 opt.restrict_files_os = restrict_os;
1453 opt.restrict_files_ctrl = restrict_ctrl;
1454 opt.restrict_files_case = restrict_case;
1455 opt.restrict_files_nonascii = restrict_nonascii;
1461 cmd_spec_report_speed (const char *com, const char *val, void *place_ignored)
1463 opt.report_bps = strcasecmp (val, "bits") == 0;
1464 if (!opt.report_bps)
1465 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1466 return opt.report_bps;
1471 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1473 static const struct decode_item choices[] = {
1474 { "auto", secure_protocol_auto },
1475 { "sslv2", secure_protocol_sslv2 },
1476 { "sslv3", secure_protocol_sslv3 },
1477 { "tlsv1", secure_protocol_tlsv1 },
1479 int ok = decode_string (val, choices, countof (choices), place);
1481 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1486 /* Set all three timeout values. */
1489 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1492 if (!cmd_time (com, val, &value))
1494 opt.read_timeout = value;
1495 opt.connect_timeout = value;
1496 opt.dns_timeout = value;
1501 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1503 /* Disallow embedded newlines. */
1504 if (strchr (val, '\n'))
1506 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1507 exec_name, com, quote (val));
1510 xfree_null (opt.useragent);
1511 opt.useragent = xstrdup (val);
1515 /* The "verbose" option cannot be cmd_boolean because the variable is
1516 not bool -- it's of type int (-1 means uninitialized because of
1517 some random hackery for disallowing -q -v). */
1520 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1523 if (cmd_boolean (com, val, &flag))
1531 /* Miscellaneous useful routines. */
1533 /* A very simple atoi clone, more useful than atoi because it works on
1534 delimited strings, and has error reportage. Returns true on success,
1535 false on failure. If successful, stores result to *DEST. */
1538 simple_atoi (const char *beg, const char *end, int *dest)
1541 bool negative = false;
1542 const char *p = beg;
1544 while (p < end && c_isspace (*p))
1546 if (p < end && (*p == '-' || *p == '+'))
1548 negative = (*p == '-');
1554 /* Read negative numbers in a separate loop because the most
1555 negative integer cannot be represented as a positive number. */
1558 for (; p < end && c_isdigit (*p); p++)
1560 int next = (10 * result) + (*p - '0');
1562 return false; /* overflow */
1566 for (; p < end && c_isdigit (*p); p++)
1568 int next = (10 * result) - (*p - '0');
1570 return false; /* underflow */
1581 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1582 doesn't handle exponential notation. Returns true on success,
1583 false on failure. In case of success, stores its result to
1587 simple_atof (const char *beg, const char *end, double *dest)
1591 bool negative = false;
1592 bool seen_dot = false;
1593 bool seen_digit = false;
1596 const char *p = beg;
1598 while (p < end && c_isspace (*p))
1600 if (p < end && (*p == '-' || *p == '+'))
1602 negative = (*p == '-');
1606 for (; p < end; p++)
1612 result = (10 * result) + (ch - '0');
1614 result += (ch - '0') / (divider *= 10);
1636 /* Verify that the user-specified header in S is valid. It must
1637 contain a colon preceded by non-white-space characters and must not
1638 contain newlines. */
1641 check_user_specified_header (const char *s)
1645 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1647 /* The header MUST contain `:' preceded by at least one
1648 non-whitespace character. */
1649 if (*p != ':' || p == s)
1651 /* The header MUST NOT contain newlines. */
1652 if (strchr (s, '\n'))
1657 /* Decode VAL into a number, according to ITEMS. */
1660 decode_string (const char *val, const struct decode_item *items, int itemcount,
1664 for (i = 0; i < itemcount; i++)
1665 if (0 == strcasecmp (val, items[i].name))
1667 *place = items[i].code;
1674 void cleanup_html_url (void);
1675 void spider_cleanup (void);
1678 /* Free the memory allocated by global variables. */
1682 /* Free external resources, close files, etc. */
1684 /* Close WARC file. */
1685 if (opt.warc_filename != 0)
1691 if (fclose (output_stream) == EOF)
1692 inform_exit_status (CLOSEFAILED);
1694 /* No need to check for error because Wget flushes its output (and
1695 checks for errors) after any data arrives. */
1697 /* We're exiting anyway so there's no real need to call free()
1698 hundreds of times. Skipping the frees will make Wget exit
1701 However, when detecting leaks, it's crucial to free() everything
1702 because then you can find the real leaks, i.e. the allocated
1703 memory which grows with the size of the program. */
1709 cleanup_html_url ();
1714 for (i = 0; i < nurl; i++)
1718 extern acc_t *netrc_list;
1719 free_netrc (netrc_list);
1721 xfree_null (opt.choose_config);
1722 xfree_null (opt.lfilename);
1723 xfree_null (opt.dir_prefix);
1724 xfree_null (opt.input_filename);
1725 xfree_null (opt.output_document);
1726 free_vec (opt.accepts);
1727 free_vec (opt.rejects);
1728 free_vec (opt.excludes);
1729 free_vec (opt.includes);
1730 free_vec (opt.domains);
1731 free_vec (opt.follow_tags);
1732 free_vec (opt.ignore_tags);
1733 xfree_null (opt.progress_type);
1734 xfree_null (opt.ftp_user);
1735 xfree_null (opt.ftp_passwd);
1736 xfree_null (opt.ftp_proxy);
1737 xfree_null (opt.https_proxy);
1738 xfree_null (opt.http_proxy);
1739 free_vec (opt.no_proxy);
1740 xfree_null (opt.useragent);
1741 xfree_null (opt.referer);
1742 xfree_null (opt.http_user);
1743 xfree_null (opt.http_passwd);
1744 free_vec (opt.user_headers);
1745 free_vec (opt.warc_user_headers);
1747 xfree_null (opt.cert_file);
1748 xfree_null (opt.private_key);
1749 xfree_null (opt.ca_directory);
1750 xfree_null (opt.ca_cert);
1751 xfree_null (opt.random_file);
1752 xfree_null (opt.egd_file);
1754 xfree_null (opt.bind_address);
1755 xfree_null (opt.cookies_input);
1756 xfree_null (opt.cookies_output);
1757 xfree_null (opt.user);
1758 xfree_null (opt.passwd);
1759 xfree_null (opt.base_href);
1760 xfree_null (opt.method);
1762 #endif /* DEBUG_MALLOC */
1765 /* Unit testing routines. */
1770 test_commands_sorted()
1772 int prev_idx = 0, next_idx = 1;
1773 int command_count = countof (commands) - 1;
1775 while (next_idx <= command_count)
1777 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1780 mu_assert ("FAILED", false);
1793 test_cmd_spec_restrict_file_names()
1798 int expected_restrict_files_os;
1799 int expected_restrict_files_ctrl;
1800 int expected_restrict_files_case;
1803 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1804 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1805 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1806 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1809 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1814 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1817 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1818 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1819 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1820 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1822 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1823 res == test_array[i].result
1824 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1825 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1826 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1832 #endif /* TESTING */