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 },
140 { "cacertificate", &opt.ca_cert, cmd_file },
142 { "cache", &opt.allow_cache, cmd_boolean },
144 { "cadirectory", &opt.ca_directory, cmd_directory },
145 { "certificate", &opt.cert_file, cmd_file },
146 { "certificatetype", &opt.cert_type, cmd_cert_type },
147 { "checkcertificate", &opt.check_cert, cmd_boolean },
149 { "chooseconfig", &opt.choose_config, cmd_file },
150 { "connecttimeout", &opt.connect_timeout, cmd_time },
151 { "contentdisposition", &opt.content_disposition, cmd_boolean },
152 { "contentonerror", &opt.content_on_error, cmd_boolean },
153 { "continue", &opt.always_rest, cmd_boolean },
154 { "convertlinks", &opt.convert_links, cmd_boolean },
155 { "cookies", &opt.cookies, cmd_boolean },
156 { "cutdirs", &opt.cut_dirs, cmd_number },
158 { "debug", &opt.debug, cmd_boolean },
160 { "defaultpage", &opt.default_page, cmd_string},
161 { "deleteafter", &opt.delete_after, cmd_boolean },
162 { "dirprefix", &opt.dir_prefix, cmd_directory },
163 { "dirstruct", NULL, cmd_spec_dirstruct },
164 { "dnscache", &opt.dns_cache, cmd_boolean },
165 { "dnstimeout", &opt.dns_timeout, cmd_time },
166 { "domains", &opt.domains, cmd_vector },
167 { "dotbytes", &opt.dot_bytes, cmd_bytes },
168 { "dotsinline", &opt.dots_in_line, cmd_number },
169 { "dotspacing", &opt.dot_spacing, cmd_number },
170 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
172 { "egdfile", &opt.egd_file, cmd_file },
174 { "excludedirectories", &opt.excludes, cmd_directory_vector },
175 { "excludedomains", &opt.exclude_domains, cmd_vector },
176 { "followftp", &opt.follow_ftp, cmd_boolean },
177 { "followtags", &opt.follow_tags, cmd_vector },
178 { "forcehtml", &opt.force_html, cmd_boolean },
179 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
180 { "ftppassword", &opt.ftp_passwd, cmd_string },
181 { "ftpproxy", &opt.ftp_proxy, cmd_string },
183 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
184 #endif /* def __VMS */
185 { "ftpuser", &opt.ftp_user, cmd_string },
186 { "glob", &opt.ftp_glob, cmd_boolean },
187 { "header", NULL, cmd_spec_header },
188 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
189 { "htmlify", NULL, cmd_spec_htmlify },
190 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
191 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
192 { "httppassword", &opt.http_passwd, cmd_string },
193 { "httpproxy", &opt.http_proxy, cmd_string },
194 { "httpsproxy", &opt.https_proxy, cmd_string },
195 { "httpuser", &opt.http_user, cmd_string },
196 { "ignorecase", &opt.ignore_case, cmd_boolean },
197 { "ignorelength", &opt.ignore_length, cmd_boolean },
198 { "ignoretags", &opt.ignore_tags, cmd_vector },
199 { "includedirectories", &opt.includes, cmd_directory_vector },
201 { "inet4only", &opt.ipv4_only, cmd_boolean },
202 { "inet6only", &opt.ipv6_only, cmd_boolean },
204 { "input", &opt.input_filename, cmd_file },
205 { "iri", &opt.enable_iri, cmd_boolean },
206 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
207 { "limitrate", &opt.limit_rate, cmd_bytes },
208 { "loadcookies", &opt.cookies_input, cmd_file },
209 { "localencoding", &opt.locale, cmd_string },
210 { "logfile", &opt.lfilename, cmd_file },
211 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
212 { "maxredirect", &opt.max_redirect, cmd_number },
213 { "mirror", NULL, cmd_spec_mirror },
214 { "netrc", &opt.netrc, cmd_boolean },
215 { "noclobber", &opt.noclobber, cmd_boolean },
216 { "noparent", &opt.no_parent, cmd_boolean },
217 { "noproxy", &opt.no_proxy, cmd_vector },
218 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
219 { "outputdocument", &opt.output_document, cmd_file },
220 { "pagerequisites", &opt.page_requisites, cmd_boolean },
221 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
222 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
223 { "password", &opt.passwd, cmd_string },
224 { "postdata", &opt.post_data, cmd_string },
225 { "postfile", &opt.post_file_name, cmd_file },
226 { "preferfamily", NULL, cmd_spec_prefer_family },
227 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
229 { "privatekey", &opt.private_key, cmd_file },
230 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
232 { "progress", &opt.progress_type, cmd_spec_progress },
233 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
234 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
235 { "proxypassword", &opt.proxy_passwd, cmd_string },
236 { "proxyuser", &opt.proxy_user, cmd_string },
237 { "quiet", &opt.quiet, cmd_boolean },
238 { "quota", &opt.quota, cmd_bytes_sum },
240 { "randomfile", &opt.random_file, cmd_file },
242 { "randomwait", &opt.random_wait, cmd_boolean },
243 { "readtimeout", &opt.read_timeout, cmd_time },
244 { "reclevel", &opt.reclevel, cmd_number_inf },
245 { "recursive", NULL, cmd_spec_recursive },
246 { "referer", &opt.referer, cmd_string },
247 { "regextype", &opt.regex_type, cmd_spec_regex_type },
248 { "reject", &opt.rejects, cmd_vector },
249 { "rejectregex", &opt.rejectregex_s, cmd_string },
250 { "relativeonly", &opt.relative_only, cmd_boolean },
251 { "remoteencoding", &opt.encoding_remote, cmd_string },
252 { "removelisting", &opt.remove_listing, cmd_boolean },
253 { "reportspeed", &opt.report_bps, cmd_spec_report_speed},
254 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
255 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
256 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
257 { "robots", &opt.use_robots, cmd_boolean },
258 { "savecookies", &opt.cookies_output, cmd_file },
259 { "saveheaders", &opt.save_headers, cmd_boolean },
261 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
263 { "serverresponse", &opt.server_response, cmd_boolean },
264 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
265 { "spanhosts", &opt.spanhost, cmd_boolean },
266 { "spider", &opt.spider, cmd_boolean },
267 { "strictcomments", &opt.strict_comments, cmd_boolean },
268 { "timeout", NULL, cmd_spec_timeout },
269 { "timestamping", &opt.timestamping, cmd_boolean },
270 { "tries", &opt.ntry, cmd_number_inf },
271 { "trustservernames", &opt.trustservernames, cmd_boolean },
272 { "unlink", &opt.unlink, cmd_boolean },
273 { "useproxy", &opt.use_proxy, cmd_boolean },
274 { "user", &opt.user, cmd_string },
275 { "useragent", NULL, cmd_spec_useragent },
276 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
277 { "verbose", NULL, cmd_spec_verbose },
278 { "wait", &opt.wait, cmd_time },
279 { "waitretry", &opt.waitretry, cmd_time },
280 { "warccdx", &opt.warc_cdx_enabled, cmd_boolean },
281 { "warccdxdedup", &opt.warc_cdx_dedup_filename, cmd_file },
283 { "warccompression", &opt.warc_compression_enabled, cmd_boolean },
285 { "warcdigests", &opt.warc_digests_enabled, cmd_boolean },
286 { "warcfile", &opt.warc_filename, cmd_file },
287 { "warcheader", NULL, cmd_spec_warc_header },
288 { "warckeeplog", &opt.warc_keep_log, cmd_boolean },
289 { "warcmaxsize", &opt.warc_maxsize, cmd_bytes },
290 { "warctempdir", &opt.warc_tempdir, cmd_directory },
292 { "wdebug", &opt.wdebug, cmd_boolean },
296 /* Look up CMDNAME in the commands[] and return its position in the
297 array. If CMDNAME is not found, return -1. */
300 command_by_name (const char *cmdname)
302 /* Use binary search for speed. Wget has ~100 commands, which
303 guarantees a worst case performance of 7 string comparisons. */
304 int lo = 0, hi = countof (commands) - 1;
308 int mid = (lo + hi) >> 1;
309 int cmp = strcasecmp (cmdname, commands[mid].name);
320 /* Reset the variables to default values. */
326 /* Most of the default values are 0 (and 0.0, NULL, and false).
327 Just reset everything, and fill in the non-zero values. Note
328 that initializing pointers to NULL this way is technically
329 illegal, but porting Wget to a machine where NULL is not all-zero
330 bit pattern will be the least of the implementors' worries. */
337 opt.add_hostdir = true;
341 opt.http_keep_alive = true;
342 opt.use_proxy = true;
343 tmp = getenv ("no_proxy");
345 opt.no_proxy = sepstring (tmp);
346 opt.prefer_family = prefer_none;
347 opt.allow_cache = true;
349 opt.read_timeout = 900;
350 opt.use_robots = true;
352 opt.remove_listing = true;
354 opt.dot_bytes = 1024;
355 opt.dot_spacing = 10;
356 opt.dots_in_line = 50;
358 opt.dns_cache = true;
362 opt.check_cert = true;
365 /* The default for file name restriction defaults to the OS type. */
366 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
367 opt.restrict_files_os = restrict_windows;
369 opt.restrict_files_os = restrict_unix;
371 opt.restrict_files_ctrl = true;
372 opt.restrict_files_nonascii = false;
373 opt.restrict_files_case = restrict_no_case_restriction;
375 opt.regex_type = regex_type_posix;
377 opt.max_redirect = 20;
382 opt.enable_iri = true;
384 opt.enable_iri = false;
387 opt.encoding_remote = NULL;
389 opt.useservertimestamps = true;
390 opt.show_all_dns_entries = false;
392 opt.warc_maxsize = 0; /* 1024 * 1024 * 1024; */
394 opt.warc_compression_enabled = true;
396 opt.warc_compression_enabled = false;
398 opt.warc_digests_enabled = true;
399 opt.warc_cdx_enabled = false;
400 opt.warc_cdx_dedup_filename = NULL;
401 opt.warc_tempdir = NULL;
402 opt.warc_keep_log = true;
405 /* Return the user's home directory (strdup-ed), or NULL if none is
410 static char *buf = NULL;
411 static char *home, *ret;
415 home = getenv ("HOME");
421 /* Under MSDOS, if $HOME isn't defined, use the directory where
422 `wget.exe' resides. */
423 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
426 buff = _w32_get_argv0 ();
428 p = strrchr (buf, '/'); /* djgpp */
430 p = strrchr (buf, '\\'); /* others */
434 buff = malloc (len + 1);
438 strncpy (buff, _w32_get_argv0 (), len);
442 #elif !defined(WINDOWS)
443 /* If HOME is not defined, try getting it from the password
445 struct passwd *pwd = getpwuid (getuid ());
446 if (!pwd || !pwd->pw_dir)
450 /* Under Windows, if $HOME isn't defined, use the directory where
451 `wget.exe' resides. */
457 ret = home ? xstrdup (home) : NULL;
464 /* Check the 'WGETRC' environment variable and return the file name
465 if 'WGETRC' is set and is a valid file.
466 If the `WGETRC' variable exists but the file does not exist, the
467 function will exit(). */
469 wgetrc_env_file_name (void)
471 char *env = getenv ("WGETRC");
474 if (!file_exists_p (env))
476 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
480 return xstrdup (env);
485 /* Check for the existance of '$HOME/.wgetrc' and return its path
486 if it exists and is set. */
488 wgetrc_user_file_name (void)
492 /* If that failed, try $HOME/.wgetrc (or equivalent). */
495 file = "SYS$LOGIN:.wgetrc";
496 #else /* def __VMS */
499 file = aprintf ("%s/.wgetrc", home);
501 #endif /* def __VMS [else] */
505 if (!file_exists_p (file))
513 /* Return the path to the user's .wgetrc. This is either the value of
514 `WGETRC' environment variable, or `$HOME/.wgetrc'.
516 Additionally, for windows, look in the directory where wget.exe
519 wgetrc_file_name (void)
521 char *file = wgetrc_env_file_name ();
525 file = wgetrc_user_file_name ();
528 /* Under Windows, if we still haven't found .wgetrc, look for the file
529 `wget.ini' in the directory where `wget.exe' resides; we do this for
530 backward compatibility with previous versions of Wget.
531 SYSTEM_WGETRC should not be defined under WINDOWS. */
534 char *home = home_dir ();
540 file = aprintf ("%s/wget.ini", home);
541 if (!file_exists_p (file))
554 /* Return values of parse_line. */
562 static enum parse_line parse_line (const char *, char **, char **, int *);
563 static bool setval_internal (int, const char *, const char *);
564 static bool setval_internal_tilde (int, const char *, const char *);
566 /* Initialize variables from a wgetrc file. Returns zero (failure) if
567 there were errors in the file. */
570 run_wgetrc (const char *file)
577 fp = fopen (file, "r");
580 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
581 file, strerror (errno));
582 return true; /* not a fatal error */
585 while ((line = read_whole_line (fp)) != NULL)
587 char *com = NULL, *val = NULL;
590 /* Parse the line. */
591 switch (parse_line (line, &com, &val, &comind))
594 /* If everything is OK, set the value. */
595 if (!setval_internal_tilde (comind, com, val))
597 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
598 exec_name, file, ln);
602 case line_syntax_error:
603 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
604 exec_name, file, ln);
607 case line_unknown_command:
608 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
609 exec_name, quote (com), file, ln);
627 /* Initialize the defaults and run the system wgetrc and user's own
632 char *file, *env_sysrc;
635 /* Run a non-standard system rc file when the according environment
636 variable has been set. For internal testing purposes only! */
637 env_sysrc = getenv ("SYSTEM_WGETRC");
638 if (env_sysrc && file_exists_p (env_sysrc))
640 ok &= run_wgetrc (env_sysrc);
641 /* If there are any problems parsing the system wgetrc file, tell
645 fprintf (stderr, _("\
646 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
648 or specify a different file using --config.\n"), env_sysrc);
652 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
654 else if (file_exists_p (SYSTEM_WGETRC))
655 ok &= run_wgetrc (SYSTEM_WGETRC);
656 /* If there are any problems parsing the system wgetrc file, tell
660 fprintf (stderr, _("\
661 Parsing system wgetrc file failed. Please check\n\
663 or specify a different file using --config.\n"), SYSTEM_WGETRC);
667 /* Override it with your own, if one exists. */
668 file = wgetrc_file_name ();
671 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
672 something like realpath() before comparing them with `strcmp' */
674 if (!strcmp (file, SYSTEM_WGETRC))
676 fprintf (stderr, _("\
677 %s: Warning: Both system and user wgetrc point to %s.\n"),
678 exec_name, quote (file));
682 ok &= run_wgetrc (file);
684 /* If there were errors processing either `.wgetrc', abort. */
692 /* Remove dashes and underscores from S, modifying S in the
698 char *t = s; /* t - tortoise */
699 char *h = s; /* h - hare */
701 if (*h == '_' || *h == '-')
708 /* Parse the line pointed by line, with the syntax:
709 <sp>* command <sp>* = <sp>* value <sp>*
710 Uses malloc to allocate space for command and value.
712 Returns one of line_ok, line_empty, line_syntax_error, or
713 line_unknown_command.
715 In case of line_ok, *COM and *VAL point to freshly allocated
716 strings, and *COMIND points to com's index. In case of error or
717 empty line, their values are unmodified. */
719 static enum parse_line
720 parse_line (const char *line, char **com, char **val, int *comind)
723 const char *end = line + strlen (line);
724 const char *cmdstart, *cmdend;
725 const char *valstart, *valend;
730 /* Skip leading and trailing whitespace. */
731 while (*line && c_isspace (*line))
733 while (end > line && c_isspace (end[-1]))
736 /* Skip empty lines and comments. */
737 if (!*line || *line == '#')
743 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
747 /* Skip '=', as well as any space before or after it. */
748 while (p < end && c_isspace (*p))
750 if (p == end || *p != '=')
751 return line_syntax_error;
753 while (p < end && c_isspace (*p))
759 /* The syntax is valid (even though the command might not be). Fill
760 in the command name and value. */
761 *com = strdupdelim (cmdstart, cmdend);
762 *val = strdupdelim (valstart, valend);
764 /* The line now known to be syntactically correct. Check whether
765 the command is valid. */
766 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
768 ind = command_by_name (cmdcopy);
770 return line_unknown_command;
772 /* Report success to the caller. */
777 #if defined(WINDOWS) || defined(MSDOS)
778 # define ISSEP(c) ((c) == '/' || (c) == '\\')
780 # define ISSEP(c) ((c) == '/')
783 /* Run commands[comind].action. */
786 setval_internal (int comind, const char *com, const char *val)
788 assert (0 <= comind && ((size_t) comind) < countof (commands));
789 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
790 return commands[comind].action (com, val, commands[comind].place);
794 setval_internal_tilde (int comind, const char *com, const char *val)
800 ret = setval_internal (comind, com, val);
802 /* We make tilde expansion for cmd_file and cmd_directory */
803 if (((commands[comind].action == cmd_file) ||
804 (commands[comind].action == cmd_directory))
805 && ret && (*val == '~' && ISSEP (val[1])))
807 pstring = commands[comind].place;
811 homelen = strlen (home);
812 while (homelen && ISSEP (home[homelen - 1]))
813 home[--homelen] = '\0';
815 /* Skip the leading "~/". */
816 for (++val; ISSEP (*val); val++)
818 *pstring = concat_strings (home, "/", val, (char *)0);
824 /* Run command COM with value VAL. If running the command produces an
825 error, report the error and exit.
827 This is intended to be called from main() to modify Wget's behavior
828 through command-line switches. Since COM is hard-coded in main(),
829 it is not canonicalized, and this aborts when COM is not found.
831 If COMIND's are exported to init.h, this function will be changed
832 to accept COMIND directly. */
835 setoptval (const char *com, const char *val, const char *optname)
837 /* Prepend "--" to OPTNAME. */
838 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
841 strcpy (dd_optname + 2, optname);
843 assert (val != NULL);
844 if (!setval_internal (command_by_name (com), dd_optname, val))
848 /* Parse OPT into command and value and run it. For example,
849 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
850 This is used by the `--execute' flag in main.c. */
853 run_command (const char *opt)
857 switch (parse_line (opt, &com, &val, &comind))
860 if (!setval_internal (comind, com, val))
866 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
867 exec_name, quote (opt));
872 /* Generic helper functions, for use with `commands'. */
874 /* Forward declarations: */
879 static bool decode_string (const char *, const struct decode_item *, int, int *);
880 static bool simple_atoi (const char *, const char *, int *);
881 static bool simple_atof (const char *, const char *, double *);
883 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
885 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
886 && c_tolower((p)[1]) == (c1) \
889 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
890 && c_tolower((p)[1]) == (c1) \
891 && c_tolower((p)[2]) == (c2) \
895 /* Store the boolean value from VAL to PLACE. COM is ignored,
896 except for error messages. */
898 cmd_boolean (const char *com, const char *val, void *place)
902 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
903 /* "on", "yes" and "1" mean true. */
905 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
906 /* "off", "no" and "0" mean false. */
911 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
912 exec_name, com, quote (val));
916 *(bool *) place = value;
920 /* Set the non-negative integer value from VAL to PLACE. With
921 incorrect specification, the number remains unchanged. */
923 cmd_number (const char *com, const char *val, void *place)
925 if (!simple_atoi (val, val + strlen (val), place)
926 || *(int *) place < 0)
928 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
929 exec_name, com, quote (val));
935 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
937 cmd_number_inf (const char *com, const char *val, void *place)
939 if (!strcasecmp (val, "inf"))
944 return cmd_number (com, val, place);
947 /* Copy (strdup) the string at COM to a new location and place a
948 pointer to *PLACE. */
950 cmd_string (const char *com, const char *val, void *place)
952 char **pstring = (char **)place;
954 xfree_null (*pstring);
955 *pstring = xstrdup (val);
960 /* Like the above, but handles tilde-expansion when reading a user's
961 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
962 gets expanded to the user's home directory. */
964 cmd_file (const char *com, const char *val, void *place)
966 char **pstring = (char **)place;
968 xfree_null (*pstring);
970 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
972 *pstring = xstrdup (val);
974 #if defined(WINDOWS) || defined(MSDOS)
975 /* Convert "\" to "/". */
978 for (s = *pstring; *s; s++)
986 /* Like cmd_file, but strips trailing '/' characters. */
988 cmd_directory (const char *com, const char *val, void *place)
992 /* Call cmd_file() for tilde expansion and separator
993 canonicalization (backslash -> slash under Windows). These
994 things should perhaps be in a separate function. */
995 if (!cmd_file (com, val, place))
1000 while (t > s && *--t == '/')
1006 /* Split VAL by space to a vector of values, and append those values
1007 to vector pointed to by the PLACE argument. If VAL is empty, the
1008 PLACE vector is cleared instead. */
1011 cmd_vector (const char *com, const char *val, void *place)
1013 char ***pvec = (char ***)place;
1016 *pvec = merge_vecs (*pvec, sepstring (val));
1026 cmd_directory_vector (const char *com, const char *val, void *place)
1028 char ***pvec = (char ***)place;
1032 /* Strip the trailing slashes from directories. */
1035 seps = sepstring (val);
1036 for (t = seps; t && *t; t++)
1038 int len = strlen (*t);
1039 /* Skip degenerate case of root directory. */
1042 if ((*t)[len - 1] == '/')
1043 (*t)[len - 1] = '\0';
1046 *pvec = merge_vecs (*pvec, seps);
1056 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1057 "100k" or "2.5G" to a floating point number. */
1060 parse_bytes_helper (const char *val, double *result)
1062 double number, mult;
1063 const char *end = val + strlen (val);
1065 /* Check for "inf". */
1066 if (0 == strcmp (val, "inf"))
1072 /* Strip trailing whitespace. */
1073 while (val < end && c_isspace (end[-1]))
1078 switch (c_tolower (end[-1]))
1081 --end, mult = 1024.0;
1084 --end, mult = 1048576.0;
1087 --end, mult = 1073741824.0;
1090 --end, mult = 1099511627776.0;
1093 /* Not a recognized suffix: assume it's a digit. (If not,
1094 simple_atof will raise an error.) */
1098 /* Skip leading and trailing whitespace. */
1099 while (val < end && c_isspace (*val))
1101 while (val < end && c_isspace (end[-1]))
1106 if (!simple_atof (val, end, &number) || number < 0)
1109 *result = number * mult;
1113 /* Parse VAL as a number and set its value to PLACE (which should
1116 By default, the value is assumed to be in bytes. If "K", "M", or
1117 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1118 1<<30, respectively. Floating point values are allowed and are
1119 cast to integer before use. The idea is to be able to use things
1120 like 1.5k instead of "1536".
1122 The string "inf" is returned as 0.
1124 In case of error, false is returned and memory pointed to by PLACE
1125 remains unmodified. */
1128 cmd_bytes (const char *com, const char *val, void *place)
1131 if (!parse_bytes_helper (val, &byte_value))
1133 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1134 exec_name, com, quote (val));
1137 *(wgint *)place = (wgint)byte_value;
1141 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1142 SIZE_SUM. It works by converting the string to double, therefore
1143 working with values up to 2^53-1 without loss of precision. This
1144 value (8192 TB) is large enough to serve for a while. */
1147 cmd_bytes_sum (const char *com, const char *val, void *place)
1150 if (!parse_bytes_helper (val, &byte_value))
1152 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1153 exec_name, com, quote (val));
1156 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1160 /* Store the value of VAL to *OUT. The value is a time period, by
1161 default expressed in seconds, but also accepting suffixes "m", "h",
1162 "d", and "w" for minutes, hours, days, and weeks respectively. */
1165 cmd_time (const char *com, const char *val, void *place)
1167 double number, mult;
1168 const char *end = val + strlen (val);
1170 /* Strip trailing whitespace. */
1171 while (val < end && c_isspace (end[-1]))
1177 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1178 exec_name, com, quote (val));
1182 switch (c_tolower (end[-1]))
1185 --end, mult = 1; /* seconds */
1188 --end, mult = 60; /* minutes */
1191 --end, mult = 3600; /* hours */
1194 --end, mult = 86400.0; /* days */
1197 --end, mult = 604800.0; /* weeks */
1200 /* Not a recognized suffix: assume it belongs to the number.
1201 (If not, simple_atof will raise an error.) */
1205 /* Skip leading and trailing whitespace. */
1206 while (val < end && c_isspace (*val))
1208 while (val < end && c_isspace (end[-1]))
1213 if (!simple_atof (val, end, &number))
1216 *(double *)place = number * mult;
1222 cmd_cert_type (const char *com, const char *val, void *place)
1224 static const struct decode_item choices[] = {
1225 { "pem", keyfile_pem },
1226 { "der", keyfile_asn1 },
1227 { "asn1", keyfile_asn1 },
1229 int ok = decode_string (val, choices, countof (choices), place);
1231 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1236 /* Specialized helper functions, used by `commands' to handle some
1237 options specially. */
1239 static bool check_user_specified_header (const char *);
1242 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1244 if (!cmd_boolean (com, val, &opt.dirstruct))
1246 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1247 must be affected inversely. */
1249 opt.no_dirstruct = false;
1251 opt.no_dirstruct = true;
1256 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1258 /* Empty value means reset the list of headers. */
1261 free_vec (opt.user_headers);
1262 opt.user_headers = NULL;
1266 if (!check_user_specified_header (val))
1268 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1269 exec_name, com, quote (val));
1272 opt.user_headers = vec_append (opt.user_headers, val);
1277 cmd_spec_warc_header (const char *com, const char *val, void *place_ignored)
1279 /* Empty value means reset the list of headers. */
1282 free_vec (opt.warc_user_headers);
1283 opt.warc_user_headers = NULL;
1287 if (!check_user_specified_header (val))
1289 fprintf (stderr, _("%s: %s: Invalid WARC header %s.\n"),
1290 exec_name, com, quote (val));
1293 opt.warc_user_headers = vec_append (opt.warc_user_headers, val);
1298 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1300 int flag = cmd_boolean (com, val, &opt.htmlify);
1301 if (flag && !opt.htmlify)
1302 opt.remove_listing = false;
1306 /* Set the "mirror" mode. It means: recursive download, timestamping,
1307 no limit on max. recursion depth, and don't remove listings. */
1310 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1314 if (!cmd_boolean (com, val, &mirror))
1318 opt.recursive = true;
1319 if (!opt.no_dirstruct)
1320 opt.dirstruct = true;
1321 opt.timestamping = true;
1322 opt.reclevel = INFINITE_RECURSION;
1323 opt.remove_listing = false;
1328 /* Validate --prefer-family and set the choice. Allowed values are
1329 "IPv4", "IPv6", and "none". */
1332 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1334 static const struct decode_item choices[] = {
1335 { "IPv4", prefer_ipv4 },
1336 { "IPv6", prefer_ipv6 },
1337 { "none", prefer_none },
1339 int prefer_family = prefer_none;
1340 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1342 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1343 opt.prefer_family = prefer_family;
1347 /* Set progress.type to VAL, but verify that it's a valid progress
1348 implementation before that. */
1351 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1353 if (!valid_progress_implementation_p (val))
1355 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1356 exec_name, com, quote (val));
1359 xfree_null (opt.progress_type);
1361 /* Don't call set_progress_implementation here. It will be called
1362 in main() when it becomes clear what the log output is. */
1363 opt.progress_type = xstrdup (val);
1367 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1368 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1372 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1374 if (!cmd_boolean (com, val, &opt.recursive))
1378 if (opt.recursive && !opt.no_dirstruct)
1379 opt.dirstruct = true;
1384 /* Validate --regex-type and set the choice. */
1387 cmd_spec_regex_type (const char *com, const char *val, void *place_ignored)
1389 static const struct decode_item choices[] = {
1390 { "posix", regex_type_posix },
1392 { "pcre", regex_type_pcre },
1395 int regex_type = regex_type_posix;
1396 int ok = decode_string (val, choices, countof (choices), ®ex_type);
1398 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1399 opt.regex_type = regex_type;
1404 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1406 int restrict_os = opt.restrict_files_os;
1407 int restrict_ctrl = opt.restrict_files_ctrl;
1408 int restrict_case = opt.restrict_files_case;
1409 int restrict_nonascii = opt.restrict_files_nonascii;
1413 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1417 end = strchr (val, ',');
1419 end = val + strlen (val);
1421 if (VAL_IS ("unix"))
1422 restrict_os = restrict_unix;
1423 else if (VAL_IS ("windows"))
1424 restrict_os = restrict_windows;
1425 else if (VAL_IS ("lowercase"))
1426 restrict_case = restrict_lowercase;
1427 else if (VAL_IS ("uppercase"))
1428 restrict_case = restrict_uppercase;
1429 else if (VAL_IS ("nocontrol"))
1430 restrict_ctrl = false;
1431 else if (VAL_IS ("ascii"))
1432 restrict_nonascii = true;
1435 fprintf (stderr, _("\
1436 %s: %s: Invalid restriction %s,\n\
1437 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1438 exec_name, com, quote (val));
1445 while (*val && *end);
1449 opt.restrict_files_os = restrict_os;
1450 opt.restrict_files_ctrl = restrict_ctrl;
1451 opt.restrict_files_case = restrict_case;
1452 opt.restrict_files_nonascii = restrict_nonascii;
1458 cmd_spec_report_speed (const char *com, const char *val, void *place_ignored)
1460 opt.report_bps = strcasecmp (val, "bits") == 0;
1461 if (!opt.report_bps)
1462 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1463 return opt.report_bps;
1468 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1470 static const struct decode_item choices[] = {
1471 { "auto", secure_protocol_auto },
1472 { "sslv2", secure_protocol_sslv2 },
1473 { "sslv3", secure_protocol_sslv3 },
1474 { "tlsv1", secure_protocol_tlsv1 },
1476 int ok = decode_string (val, choices, countof (choices), place);
1478 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1483 /* Set all three timeout values. */
1486 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1489 if (!cmd_time (com, val, &value))
1491 opt.read_timeout = value;
1492 opt.connect_timeout = value;
1493 opt.dns_timeout = value;
1498 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1500 /* Disallow embedded newlines. */
1501 if (strchr (val, '\n'))
1503 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1504 exec_name, com, quote (val));
1507 xfree_null (opt.useragent);
1508 opt.useragent = xstrdup (val);
1512 /* The "verbose" option cannot be cmd_boolean because the variable is
1513 not bool -- it's of type int (-1 means uninitialized because of
1514 some random hackery for disallowing -q -v). */
1517 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1520 if (cmd_boolean (com, val, &flag))
1528 /* Miscellaneous useful routines. */
1530 /* A very simple atoi clone, more useful than atoi because it works on
1531 delimited strings, and has error reportage. Returns true on success,
1532 false on failure. If successful, stores result to *DEST. */
1535 simple_atoi (const char *beg, const char *end, int *dest)
1538 bool negative = false;
1539 const char *p = beg;
1541 while (p < end && c_isspace (*p))
1543 if (p < end && (*p == '-' || *p == '+'))
1545 negative = (*p == '-');
1551 /* Read negative numbers in a separate loop because the most
1552 negative integer cannot be represented as a positive number. */
1555 for (; p < end && c_isdigit (*p); p++)
1557 int next = (10 * result) + (*p - '0');
1559 return false; /* overflow */
1563 for (; p < end && c_isdigit (*p); p++)
1565 int next = (10 * result) - (*p - '0');
1567 return false; /* underflow */
1578 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1579 doesn't handle exponential notation. Returns true on success,
1580 false on failure. In case of success, stores its result to
1584 simple_atof (const char *beg, const char *end, double *dest)
1588 bool negative = false;
1589 bool seen_dot = false;
1590 bool seen_digit = false;
1593 const char *p = beg;
1595 while (p < end && c_isspace (*p))
1597 if (p < end && (*p == '-' || *p == '+'))
1599 negative = (*p == '-');
1603 for (; p < end; p++)
1609 result = (10 * result) + (ch - '0');
1611 result += (ch - '0') / (divider *= 10);
1633 /* Verify that the user-specified header in S is valid. It must
1634 contain a colon preceded by non-white-space characters and must not
1635 contain newlines. */
1638 check_user_specified_header (const char *s)
1642 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1644 /* The header MUST contain `:' preceded by at least one
1645 non-whitespace character. */
1646 if (*p != ':' || p == s)
1648 /* The header MUST NOT contain newlines. */
1649 if (strchr (s, '\n'))
1654 /* Decode VAL into a number, according to ITEMS. */
1657 decode_string (const char *val, const struct decode_item *items, int itemcount,
1661 for (i = 0; i < itemcount; i++)
1662 if (0 == strcasecmp (val, items[i].name))
1664 *place = items[i].code;
1671 void cleanup_html_url (void);
1674 /* Free the memory allocated by global variables. */
1678 /* Free external resources, close files, etc. */
1680 /* Close WARC file. */
1681 if (opt.warc_filename != 0)
1687 if (fclose (output_stream) == EOF)
1688 inform_exit_status (CLOSEFAILED);
1690 /* No need to check for error because Wget flushes its output (and
1691 checks for errors) after any data arrives. */
1693 /* We're exiting anyway so there's no real need to call free()
1694 hundreds of times. Skipping the frees will make Wget exit
1697 However, when detecting leaks, it's crucial to free() everything
1698 because then you can find the real leaks, i.e. the allocated
1699 memory which grows with the size of the program. */
1705 cleanup_html_url ();
1709 for (i = 0; i < nurl; i++)
1713 extern acc_t *netrc_list;
1714 free_netrc (netrc_list);
1716 xfree_null (opt.choose_config);
1717 xfree_null (opt.lfilename);
1718 xfree_null (opt.dir_prefix);
1719 xfree_null (opt.input_filename);
1720 xfree_null (opt.output_document);
1721 free_vec (opt.accepts);
1722 free_vec (opt.rejects);
1723 free_vec (opt.excludes);
1724 free_vec (opt.includes);
1725 free_vec (opt.domains);
1726 free_vec (opt.follow_tags);
1727 free_vec (opt.ignore_tags);
1728 xfree_null (opt.progress_type);
1729 xfree_null (opt.ftp_user);
1730 xfree_null (opt.ftp_passwd);
1731 xfree_null (opt.ftp_proxy);
1732 xfree_null (opt.https_proxy);
1733 xfree_null (opt.http_proxy);
1734 free_vec (opt.no_proxy);
1735 xfree_null (opt.useragent);
1736 xfree_null (opt.referer);
1737 xfree_null (opt.http_user);
1738 xfree_null (opt.http_passwd);
1739 free_vec (opt.user_headers);
1740 free_vec (opt.warc_user_headers);
1742 xfree_null (opt.cert_file);
1743 xfree_null (opt.private_key);
1744 xfree_null (opt.ca_directory);
1745 xfree_null (opt.ca_cert);
1746 xfree_null (opt.random_file);
1747 xfree_null (opt.egd_file);
1749 xfree_null (opt.bind_address);
1750 xfree_null (opt.cookies_input);
1751 xfree_null (opt.cookies_output);
1752 xfree_null (opt.user);
1753 xfree_null (opt.passwd);
1754 xfree_null (opt.base_href);
1756 #endif /* DEBUG_MALLOC */
1759 /* Unit testing routines. */
1764 test_commands_sorted()
1766 int prev_idx = 0, next_idx = 1;
1767 int command_count = countof (commands) - 1;
1769 while (next_idx <= command_count)
1771 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1774 mu_assert ("FAILED", false);
1787 test_cmd_spec_restrict_file_names()
1792 int expected_restrict_files_os;
1793 int expected_restrict_files_ctrl;
1794 int expected_restrict_files_case;
1797 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1798 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1799 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1800 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1803 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1808 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1811 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1812 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1813 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1814 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1816 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1817 res == test_array[i].result
1818 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1819 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1820 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1826 #endif /* TESTING */