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 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. */
41 /* not all systems provide PATH_MAX in limits.h */
43 # include <sys/param.h>
45 # define PATH_MAX MAXPATHLEN
60 #include "recur.h" /* for INFINITE_RECURSION */
61 #include "convert.h" /* for convert_cleanup */
62 #include "res.h" /* for res_cleanup */
63 #include "http.h" /* for http_cleanup */
64 #include "retr.h" /* for output_stream */
72 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
74 CMD_DECLARE (cmd_boolean);
75 CMD_DECLARE (cmd_bytes);
76 CMD_DECLARE (cmd_bytes_sum);
78 CMD_DECLARE (cmd_cert_type);
80 CMD_DECLARE (cmd_directory_vector);
81 CMD_DECLARE (cmd_number);
82 CMD_DECLARE (cmd_number_inf);
83 CMD_DECLARE (cmd_string);
84 CMD_DECLARE (cmd_file);
85 CMD_DECLARE (cmd_directory);
86 CMD_DECLARE (cmd_time);
87 CMD_DECLARE (cmd_vector);
89 CMD_DECLARE (cmd_spec_dirstruct);
90 CMD_DECLARE (cmd_spec_header);
91 CMD_DECLARE (cmd_spec_htmlify);
92 CMD_DECLARE (cmd_spec_mirror);
93 CMD_DECLARE (cmd_spec_prefer_family);
94 CMD_DECLARE (cmd_spec_progress);
95 CMD_DECLARE (cmd_spec_recursive);
96 CMD_DECLARE (cmd_spec_restrict_file_names);
98 CMD_DECLARE (cmd_spec_secure_protocol);
100 CMD_DECLARE (cmd_spec_timeout);
101 CMD_DECLARE (cmd_spec_useragent);
102 CMD_DECLARE (cmd_spec_verbose);
104 /* List of recognized commands, each consisting of name, place and
105 function. When adding a new command, simply add it to the list,
106 but be sure to keep the list sorted alphabetically, as
107 command_by_name's binary search depends on it. Also, be sure to
108 add any entries that allocate memory (e.g. cmd_string and
109 cmd_vector) to the cleanup() function below. */
111 static const struct {
114 bool (*action) (const char *, const char *, void *);
116 /* KEEP THIS LIST ALPHABETICALLY SORTED */
117 { "accept", &opt.accepts, cmd_vector },
118 { "addhostdir", &opt.add_hostdir, cmd_boolean },
119 { "adjustextension", &opt.adjust_extension, cmd_boolean },
120 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
121 { "askpassword", &opt.ask_passwd, cmd_boolean },
122 { "authnochallenge", &opt.auth_without_challenge,
124 { "background", &opt.background, cmd_boolean },
125 { "backupconverted", &opt.backup_converted, cmd_boolean },
126 { "backups", &opt.backups, cmd_number },
127 { "base", &opt.base_href, cmd_string },
128 { "bindaddress", &opt.bind_address, cmd_string },
130 { "cacertificate", &opt.ca_cert, cmd_file },
132 { "cache", &opt.allow_cache, cmd_boolean },
134 { "cadirectory", &opt.ca_directory, cmd_directory },
135 { "certificate", &opt.cert_file, cmd_file },
136 { "certificatetype", &opt.cert_type, cmd_cert_type },
137 { "checkcertificate", &opt.check_cert, cmd_boolean },
139 { "chooseconfig", &opt.choose_config, cmd_file },
140 { "connecttimeout", &opt.connect_timeout, cmd_time },
141 { "contentdisposition", &opt.content_disposition, cmd_boolean },
142 { "continue", &opt.always_rest, cmd_boolean },
143 { "convertlinks", &opt.convert_links, cmd_boolean },
144 { "cookies", &opt.cookies, cmd_boolean },
145 { "cutdirs", &opt.cut_dirs, cmd_number },
147 { "debug", &opt.debug, cmd_boolean },
149 { "defaultpage", &opt.default_page, cmd_string},
150 { "deleteafter", &opt.delete_after, cmd_boolean },
151 { "dirprefix", &opt.dir_prefix, cmd_directory },
152 { "dirstruct", NULL, cmd_spec_dirstruct },
153 { "dnscache", &opt.dns_cache, cmd_boolean },
154 { "dnstimeout", &opt.dns_timeout, cmd_time },
155 { "domains", &opt.domains, cmd_vector },
156 { "dotbytes", &opt.dot_bytes, cmd_bytes },
157 { "dotsinline", &opt.dots_in_line, cmd_number },
158 { "dotspacing", &opt.dot_spacing, cmd_number },
159 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
161 { "egdfile", &opt.egd_file, cmd_file },
163 { "excludedirectories", &opt.excludes, cmd_directory_vector },
164 { "excludedomains", &opt.exclude_domains, cmd_vector },
165 { "followftp", &opt.follow_ftp, cmd_boolean },
166 { "followtags", &opt.follow_tags, cmd_vector },
167 { "forcehtml", &opt.force_html, cmd_boolean },
168 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
169 { "ftppassword", &opt.ftp_passwd, cmd_string },
170 { "ftpproxy", &opt.ftp_proxy, cmd_string },
172 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
173 #endif /* def __VMS */
174 { "ftpuser", &opt.ftp_user, cmd_string },
175 { "glob", &opt.ftp_glob, cmd_boolean },
176 { "header", NULL, cmd_spec_header },
177 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
178 { "htmlify", NULL, cmd_spec_htmlify },
179 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
180 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
181 { "httppassword", &opt.http_passwd, cmd_string },
182 { "httpproxy", &opt.http_proxy, cmd_string },
183 { "httpsproxy", &opt.https_proxy, cmd_string },
184 { "httpuser", &opt.http_user, cmd_string },
185 { "ignorecase", &opt.ignore_case, cmd_boolean },
186 { "ignorelength", &opt.ignore_length, cmd_boolean },
187 { "ignoretags", &opt.ignore_tags, cmd_vector },
188 { "includedirectories", &opt.includes, cmd_directory_vector },
190 { "inet4only", &opt.ipv4_only, cmd_boolean },
191 { "inet6only", &opt.ipv6_only, cmd_boolean },
193 { "input", &opt.input_filename, cmd_file },
194 { "iri", &opt.enable_iri, cmd_boolean },
195 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
196 { "limitrate", &opt.limit_rate, cmd_bytes },
197 { "loadcookies", &opt.cookies_input, cmd_file },
198 { "localencoding", &opt.locale, cmd_string },
199 { "logfile", &opt.lfilename, cmd_file },
200 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
201 { "maxredirect", &opt.max_redirect, cmd_number },
202 { "mirror", NULL, cmd_spec_mirror },
203 { "netrc", &opt.netrc, cmd_boolean },
204 { "noclobber", &opt.noclobber, cmd_boolean },
205 { "noparent", &opt.no_parent, cmd_boolean },
206 { "noproxy", &opt.no_proxy, cmd_vector },
207 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
208 { "outputdocument", &opt.output_document, cmd_file },
209 { "pagerequisites", &opt.page_requisites, cmd_boolean },
210 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
211 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
212 { "password", &opt.passwd, cmd_string },
213 { "postdata", &opt.post_data, cmd_string },
214 { "postfile", &opt.post_file_name, cmd_file },
215 { "preferfamily", NULL, cmd_spec_prefer_family },
216 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
218 { "privatekey", &opt.private_key, cmd_file },
219 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
221 { "progress", &opt.progress_type, cmd_spec_progress },
222 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
223 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
224 { "proxypassword", &opt.proxy_passwd, cmd_string },
225 { "proxyuser", &opt.proxy_user, cmd_string },
226 { "quiet", &opt.quiet, cmd_boolean },
227 { "quota", &opt.quota, cmd_bytes_sum },
229 { "randomfile", &opt.random_file, cmd_file },
231 { "randomwait", &opt.random_wait, cmd_boolean },
232 { "readtimeout", &opt.read_timeout, cmd_time },
233 { "reclevel", &opt.reclevel, cmd_number_inf },
234 { "recursive", NULL, cmd_spec_recursive },
235 { "referer", &opt.referer, cmd_string },
236 { "reject", &opt.rejects, cmd_vector },
237 { "relativeonly", &opt.relative_only, cmd_boolean },
238 { "remoteencoding", &opt.encoding_remote, cmd_string },
239 { "removelisting", &opt.remove_listing, cmd_boolean },
240 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
241 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
242 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
243 { "robots", &opt.use_robots, cmd_boolean },
244 { "savecookies", &opt.cookies_output, cmd_file },
245 { "saveheaders", &opt.save_headers, cmd_boolean },
247 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
249 { "serverresponse", &opt.server_response, cmd_boolean },
250 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
251 { "spanhosts", &opt.spanhost, cmd_boolean },
252 { "spider", &opt.spider, cmd_boolean },
253 { "strictcomments", &opt.strict_comments, cmd_boolean },
254 { "timeout", NULL, cmd_spec_timeout },
255 { "timestamping", &opt.timestamping, cmd_boolean },
256 { "tries", &opt.ntry, cmd_number_inf },
257 { "trustservernames", &opt.trustservernames, cmd_boolean },
258 { "unlink", &opt.unlink, cmd_boolean },
259 { "useproxy", &opt.use_proxy, cmd_boolean },
260 { "user", &opt.user, cmd_string },
261 { "useragent", NULL, cmd_spec_useragent },
262 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
263 { "verbose", NULL, cmd_spec_verbose },
264 { "wait", &opt.wait, cmd_time },
265 { "waitretry", &opt.waitretry, cmd_time },
267 { "wdebug", &opt.wdebug, cmd_boolean },
271 /* Look up CMDNAME in the commands[] and return its position in the
272 array. If CMDNAME is not found, return -1. */
275 command_by_name (const char *cmdname)
277 /* Use binary search for speed. Wget has ~100 commands, which
278 guarantees a worst case performance of 7 string comparisons. */
279 int lo = 0, hi = countof (commands) - 1;
283 int mid = (lo + hi) >> 1;
284 int cmp = strcasecmp (cmdname, commands[mid].name);
295 /* Reset the variables to default values. */
301 /* Most of the default values are 0 (and 0.0, NULL, and false).
302 Just reset everything, and fill in the non-zero values. Note
303 that initializing pointers to NULL this way is technically
304 illegal, but porting Wget to a machine where NULL is not all-zero
305 bit pattern will be the least of the implementors' worries. */
312 opt.add_hostdir = true;
316 opt.http_keep_alive = true;
317 opt.use_proxy = true;
318 tmp = getenv ("no_proxy");
320 opt.no_proxy = sepstring (tmp);
321 opt.prefer_family = prefer_none;
322 opt.allow_cache = true;
324 opt.read_timeout = 900;
325 opt.use_robots = true;
327 opt.remove_listing = true;
329 opt.dot_bytes = 1024;
330 opt.dot_spacing = 10;
331 opt.dots_in_line = 50;
333 opt.dns_cache = true;
337 opt.check_cert = true;
340 /* The default for file name restriction defaults to the OS type. */
341 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
342 opt.restrict_files_os = restrict_windows;
344 opt.restrict_files_os = restrict_unix;
346 opt.restrict_files_ctrl = true;
347 opt.restrict_files_nonascii = false;
348 opt.restrict_files_case = restrict_no_case_restriction;
350 opt.max_redirect = 20;
355 opt.enable_iri = true;
357 opt.enable_iri = false;
360 opt.encoding_remote = NULL;
362 opt.useservertimestamps = true;
363 opt.show_all_dns_entries = false;
366 /* Return the user's home directory (strdup-ed), or NULL if none is
371 static char *buf = NULL;
372 static char *home, *ret;
377 home = getenv ("HOME");
381 /* Under MSDOS, if $HOME isn't defined, use the directory where
382 `wget.exe' resides. */
383 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
386 buff = _w32_get_argv0 ();
388 p = strrchr (buf, '/'); /* djgpp */
390 p = strrchr (buf, '\\'); /* others */
394 buff = malloc (len + 1);
398 strncpy (buff, _w32_get_argv0 (), len);
402 #elif !defined(WINDOWS)
403 /* If HOME is not defined, try getting it from the password
405 struct passwd *pwd = getpwuid (getuid ());
406 if (!pwd || !pwd->pw_dir)
410 /* Under Windows, if $HOME isn't defined, use the directory where
411 `wget.exe' resides. */
417 ret = home ? xstrdup (home) : NULL;
424 /* Check the 'WGETRC' environment variable and return the file name
425 if 'WGETRC' is set and is a valid file.
426 If the `WGETRC' variable exists but the file does not exist, the
427 function will exit(). */
429 wgetrc_env_file_name (void)
431 char *env = getenv ("WGETRC");
434 if (!file_exists_p (env))
436 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
440 return xstrdup (env);
445 /* Check for the existance of '$HOME/.wgetrc' and return it's path
446 if it exists and is set. */
448 wgetrc_user_file_name (void)
452 /* If that failed, try $HOME/.wgetrc (or equivalent). */
455 file = "SYS$LOGIN:.wgetrc";
456 #else /* def __VMS */
459 file = aprintf ("%s/.wgetrc", home);
461 #endif /* def __VMS [else] */
465 if (!file_exists_p (file))
473 /* Return the path to the user's .wgetrc. This is either the value of
474 `WGETRC' environment variable, or `$HOME/.wgetrc'.
476 Additionally, for windows, look in the directory where wget.exe
479 wgetrc_file_name (void)
481 char *file = wgetrc_env_file_name ();
485 file = wgetrc_user_file_name ();
488 /* Under Windows, if we still haven't found .wgetrc, look for the file
489 `wget.ini' in the directory where `wget.exe' resides; we do this for
490 backward compatibility with previous versions of Wget.
491 SYSTEM_WGETRC should not be defined under WINDOWS. */
494 char *home = home_dir ();
500 file = aprintf ("%s/wget.ini", home);
501 if (!file_exists_p (file))
514 /* Return values of parse_line. */
522 static enum parse_line parse_line (const char *, char **, char **, int *);
523 static bool setval_internal (int, const char *, const char *);
524 static bool setval_internal_tilde (int, const char *, const char *);
526 /* Initialize variables from a wgetrc file. Returns zero (failure) if
527 there were errors in the file. */
530 run_wgetrc (const char *file)
537 fp = fopen (file, "r");
540 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
541 file, strerror (errno));
542 return true; /* not a fatal error */
545 while ((line = read_whole_line (fp)) != NULL)
547 char *com = NULL, *val = NULL;
550 /* Parse the line. */
551 switch (parse_line (line, &com, &val, &comind))
554 /* If everything is OK, set the value. */
555 if (!setval_internal_tilde (comind, com, val))
557 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
558 exec_name, file, ln);
562 case line_syntax_error:
563 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
564 exec_name, file, ln);
567 case line_unknown_command:
568 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
569 exec_name, quote (com), file, ln);
587 /* Initialize the defaults and run the system wgetrc and user's own
592 char *file, *env_sysrc;
595 /* Run a non-standard system rc file when the according environment
596 variable has been set. For internal testing purposes only! */
597 env_sysrc = getenv ("SYSTEM_WGETRC");
598 if (env_sysrc && file_exists_p (env_sysrc))
599 ok &= run_wgetrc (env_sysrc);
600 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
602 else if (file_exists_p (SYSTEM_WGETRC))
603 ok &= run_wgetrc (SYSTEM_WGETRC);
605 /* If there are any problems parsing the system wgetrc file, tell
609 fprintf (stderr, _("\
610 Parsing system wgetrc file failed, please check '%s'. \
611 Or specify a different file using --config\n"), SYSTEM_WGETRC);
614 /* Override it with your own, if one exists. */
615 file = wgetrc_file_name ();
618 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
619 something like realpath() before comparing them with `strcmp' */
621 if (!strcmp (file, SYSTEM_WGETRC))
623 fprintf (stderr, _("\
624 %s: Warning: Both system and user wgetrc point to %s.\n"),
625 exec_name, quote (file));
629 ok &= run_wgetrc (file);
631 /* If there were errors processing either `.wgetrc', abort. */
639 /* Remove dashes and underscores from S, modifying S in the
645 char *t = s; /* t - tortoise */
646 char *h = s; /* h - hare */
648 if (*h == '_' || *h == '-')
655 /* Parse the line pointed by line, with the syntax:
656 <sp>* command <sp>* = <sp>* value <sp>*
657 Uses malloc to allocate space for command and value.
659 Returns one of line_ok, line_empty, line_syntax_error, or
660 line_unknown_command.
662 In case of line_ok, *COM and *VAL point to freshly allocated
663 strings, and *COMIND points to com's index. In case of error or
664 empty line, their values are unmodified. */
666 static enum parse_line
667 parse_line (const char *line, char **com, char **val, int *comind)
670 const char *end = line + strlen (line);
671 const char *cmdstart, *cmdend;
672 const char *valstart, *valend;
677 /* Skip leading and trailing whitespace. */
678 while (*line && c_isspace (*line))
680 while (end > line && c_isspace (end[-1]))
683 /* Skip empty lines and comments. */
684 if (!*line || *line == '#')
690 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
694 /* Skip '=', as well as any space before or after it. */
695 while (p < end && c_isspace (*p))
697 if (p == end || *p != '=')
698 return line_syntax_error;
700 while (p < end && c_isspace (*p))
706 /* The syntax is valid (even though the command might not be). Fill
707 in the command name and value. */
708 *com = strdupdelim (cmdstart, cmdend);
709 *val = strdupdelim (valstart, valend);
711 /* The line now known to be syntactically correct. Check whether
712 the command is valid. */
713 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
715 ind = command_by_name (cmdcopy);
717 return line_unknown_command;
719 /* Report success to the caller. */
724 #if defined(WINDOWS) || defined(MSDOS)
725 # define ISSEP(c) ((c) == '/' || (c) == '\\')
727 # define ISSEP(c) ((c) == '/')
730 /* Run commands[comind].action. */
733 setval_internal (int comind, const char *com, const char *val)
735 assert (0 <= comind && ((size_t) comind) < countof (commands));
736 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
737 return commands[comind].action (com, val, commands[comind].place);
741 setval_internal_tilde (int comind, const char *com, const char *val)
747 ret = setval_internal (comind, com, val);
749 /* We make tilde expansion for cmd_file and cmd_directory */
750 if (((commands[comind].action == cmd_file) ||
751 (commands[comind].action == cmd_directory))
752 && ret && (*val == '~' && ISSEP (val[1])))
754 pstring = commands[comind].place;
758 homelen = strlen (home);
759 while (homelen && ISSEP (home[homelen - 1]))
760 home[--homelen] = '\0';
762 /* Skip the leading "~/". */
763 for (++val; ISSEP (*val); val++)
765 *pstring = concat_strings (home, "/", val, (char *)0);
771 /* Run command COM with value VAL. If running the command produces an
772 error, report the error and exit.
774 This is intended to be called from main() to modify Wget's behavior
775 through command-line switches. Since COM is hard-coded in main(),
776 it is not canonicalized, and this aborts when COM is not found.
778 If COMIND's are exported to init.h, this function will be changed
779 to accept COMIND directly. */
782 setoptval (const char *com, const char *val, const char *optname)
784 /* Prepend "--" to OPTNAME. */
785 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
788 strcpy (dd_optname + 2, optname);
790 assert (val != NULL);
791 if (!setval_internal (command_by_name (com), dd_optname, val))
795 /* Parse OPT into command and value and run it. For example,
796 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
797 This is used by the `--execute' flag in main.c. */
800 run_command (const char *opt)
804 switch (parse_line (opt, &com, &val, &comind))
807 if (!setval_internal (comind, com, val))
813 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
814 exec_name, quote (opt));
819 /* Generic helper functions, for use with `commands'. */
821 /* Forward declarations: */
826 static bool decode_string (const char *, const struct decode_item *, int, int *);
827 static bool simple_atoi (const char *, const char *, int *);
828 static bool simple_atof (const char *, const char *, double *);
830 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
832 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
833 && c_tolower((p)[1]) == (c1) \
836 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
837 && c_tolower((p)[1]) == (c1) \
838 && c_tolower((p)[2]) == (c2) \
842 /* Store the boolean value from VAL to PLACE. COM is ignored,
843 except for error messages. */
845 cmd_boolean (const char *com, const char *val, void *place)
849 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
850 /* "on", "yes" and "1" mean true. */
852 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
853 /* "off", "no" and "0" mean false. */
858 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
859 exec_name, com, quote (val));
863 *(bool *) place = value;
867 /* Set the non-negative integer value from VAL to PLACE. With
868 incorrect specification, the number remains unchanged. */
870 cmd_number (const char *com, const char *val, void *place)
872 if (!simple_atoi (val, val + strlen (val), place)
873 || *(int *) place < 0)
875 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
876 exec_name, com, quote (val));
882 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
884 cmd_number_inf (const char *com, const char *val, void *place)
886 if (!strcasecmp (val, "inf"))
891 return cmd_number (com, val, place);
894 /* Copy (strdup) the string at COM to a new location and place a
895 pointer to *PLACE. */
897 cmd_string (const char *com, const char *val, void *place)
899 char **pstring = (char **)place;
901 xfree_null (*pstring);
902 *pstring = xstrdup (val);
907 /* Like the above, but handles tilde-expansion when reading a user's
908 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
909 gets expanded to the user's home directory. */
911 cmd_file (const char *com, const char *val, void *place)
913 char **pstring = (char **)place;
915 xfree_null (*pstring);
917 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
919 *pstring = xstrdup (val);
921 #if defined(WINDOWS) || defined(MSDOS)
922 /* Convert "\" to "/". */
925 for (s = *pstring; *s; s++)
933 /* Like cmd_file, but strips trailing '/' characters. */
935 cmd_directory (const char *com, const char *val, void *place)
939 /* Call cmd_file() for tilde expansion and separator
940 canonicalization (backslash -> slash under Windows). These
941 things should perhaps be in a separate function. */
942 if (!cmd_file (com, val, place))
947 while (t > s && *--t == '/')
953 /* Split VAL by space to a vector of values, and append those values
954 to vector pointed to by the PLACE argument. If VAL is empty, the
955 PLACE vector is cleared instead. */
958 cmd_vector (const char *com, const char *val, void *place)
960 char ***pvec = (char ***)place;
963 *pvec = merge_vecs (*pvec, sepstring (val));
973 cmd_directory_vector (const char *com, const char *val, void *place)
975 char ***pvec = (char ***)place;
979 /* Strip the trailing slashes from directories. */
982 seps = sepstring (val);
983 for (t = seps; t && *t; t++)
985 int len = strlen (*t);
986 /* Skip degenerate case of root directory. */
989 if ((*t)[len - 1] == '/')
990 (*t)[len - 1] = '\0';
993 *pvec = merge_vecs (*pvec, seps);
1003 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1004 "100k" or "2.5G" to a floating point number. */
1007 parse_bytes_helper (const char *val, double *result)
1009 double number, mult;
1010 const char *end = val + strlen (val);
1012 /* Check for "inf". */
1013 if (0 == strcmp (val, "inf"))
1019 /* Strip trailing whitespace. */
1020 while (val < end && c_isspace (end[-1]))
1025 switch (c_tolower (end[-1]))
1028 --end, mult = 1024.0;
1031 --end, mult = 1048576.0;
1034 --end, mult = 1073741824.0;
1037 --end, mult = 1099511627776.0;
1040 /* Not a recognized suffix: assume it's a digit. (If not,
1041 simple_atof will raise an error.) */
1045 /* Skip leading and trailing whitespace. */
1046 while (val < end && c_isspace (*val))
1048 while (val < end && c_isspace (end[-1]))
1053 if (!simple_atof (val, end, &number) || number < 0)
1056 *result = number * mult;
1060 /* Parse VAL as a number and set its value to PLACE (which should
1063 By default, the value is assumed to be in bytes. If "K", "M", or
1064 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1065 1<<30, respectively. Floating point values are allowed and are
1066 cast to integer before use. The idea is to be able to use things
1067 like 1.5k instead of "1536".
1069 The string "inf" is returned as 0.
1071 In case of error, false is returned and memory pointed to by PLACE
1072 remains unmodified. */
1075 cmd_bytes (const char *com, const char *val, void *place)
1078 if (!parse_bytes_helper (val, &byte_value))
1080 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1081 exec_name, com, quote (val));
1084 *(wgint *)place = (wgint)byte_value;
1088 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1089 SIZE_SUM. It works by converting the string to double, therefore
1090 working with values up to 2^53-1 without loss of precision. This
1091 value (8192 TB) is large enough to serve for a while. */
1094 cmd_bytes_sum (const char *com, const char *val, void *place)
1097 if (!parse_bytes_helper (val, &byte_value))
1099 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1100 exec_name, com, quote (val));
1103 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1107 /* Store the value of VAL to *OUT. The value is a time period, by
1108 default expressed in seconds, but also accepting suffixes "m", "h",
1109 "d", and "w" for minutes, hours, days, and weeks respectively. */
1112 cmd_time (const char *com, const char *val, void *place)
1114 double number, mult;
1115 const char *end = val + strlen (val);
1117 /* Strip trailing whitespace. */
1118 while (val < end && c_isspace (end[-1]))
1124 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1125 exec_name, com, quote (val));
1129 switch (c_tolower (end[-1]))
1132 --end, mult = 1; /* seconds */
1135 --end, mult = 60; /* minutes */
1138 --end, mult = 3600; /* hours */
1141 --end, mult = 86400.0; /* days */
1144 --end, mult = 604800.0; /* weeks */
1147 /* Not a recognized suffix: assume it belongs to the number.
1148 (If not, simple_atof will raise an error.) */
1152 /* Skip leading and trailing whitespace. */
1153 while (val < end && c_isspace (*val))
1155 while (val < end && c_isspace (end[-1]))
1160 if (!simple_atof (val, end, &number))
1163 *(double *)place = number * mult;
1169 cmd_cert_type (const char *com, const char *val, void *place)
1171 static const struct decode_item choices[] = {
1172 { "pem", keyfile_pem },
1173 { "der", keyfile_asn1 },
1174 { "asn1", keyfile_asn1 },
1176 int ok = decode_string (val, choices, countof (choices), place);
1178 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1183 /* Specialized helper functions, used by `commands' to handle some
1184 options specially. */
1186 static bool check_user_specified_header (const char *);
1189 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1191 if (!cmd_boolean (com, val, &opt.dirstruct))
1193 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1194 must be affected inversely. */
1196 opt.no_dirstruct = false;
1198 opt.no_dirstruct = true;
1203 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1205 /* Empty value means reset the list of headers. */
1208 free_vec (opt.user_headers);
1209 opt.user_headers = NULL;
1213 if (!check_user_specified_header (val))
1215 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1216 exec_name, com, quote (val));
1219 opt.user_headers = vec_append (opt.user_headers, val);
1224 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1226 int flag = cmd_boolean (com, val, &opt.htmlify);
1227 if (flag && !opt.htmlify)
1228 opt.remove_listing = false;
1232 /* Set the "mirror" mode. It means: recursive download, timestamping,
1233 no limit on max. recursion depth, and don't remove listings. */
1236 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1240 if (!cmd_boolean (com, val, &mirror))
1244 opt.recursive = true;
1245 if (!opt.no_dirstruct)
1246 opt.dirstruct = true;
1247 opt.timestamping = true;
1248 opt.reclevel = INFINITE_RECURSION;
1249 opt.remove_listing = false;
1254 /* Validate --prefer-family and set the choice. Allowed values are
1255 "IPv4", "IPv6", and "none". */
1258 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1260 static const struct decode_item choices[] = {
1261 { "IPv4", prefer_ipv4 },
1262 { "IPv6", prefer_ipv6 },
1263 { "none", prefer_none },
1265 int prefer_family = prefer_none;
1266 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1268 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1269 opt.prefer_family = prefer_family;
1273 /* Set progress.type to VAL, but verify that it's a valid progress
1274 implementation before that. */
1277 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1279 if (!valid_progress_implementation_p (val))
1281 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1282 exec_name, com, quote (val));
1285 xfree_null (opt.progress_type);
1287 /* Don't call set_progress_implementation here. It will be called
1288 in main() when it becomes clear what the log output is. */
1289 opt.progress_type = xstrdup (val);
1293 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1294 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1298 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1300 if (!cmd_boolean (com, val, &opt.recursive))
1304 if (opt.recursive && !opt.no_dirstruct)
1305 opt.dirstruct = true;
1311 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1313 int restrict_os = opt.restrict_files_os;
1314 int restrict_ctrl = opt.restrict_files_ctrl;
1315 int restrict_case = opt.restrict_files_case;
1316 int restrict_nonascii = opt.restrict_files_nonascii;
1320 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1324 end = strchr (val, ',');
1326 end = val + strlen (val);
1328 if (VAL_IS ("unix"))
1329 restrict_os = restrict_unix;
1330 else if (VAL_IS ("windows"))
1331 restrict_os = restrict_windows;
1332 else if (VAL_IS ("lowercase"))
1333 restrict_case = restrict_lowercase;
1334 else if (VAL_IS ("uppercase"))
1335 restrict_case = restrict_uppercase;
1336 else if (VAL_IS ("nocontrol"))
1337 restrict_ctrl = false;
1338 else if (VAL_IS ("ascii"))
1339 restrict_nonascii = true;
1342 fprintf (stderr, _("\
1343 %s: %s: Invalid restriction %s,\n\
1344 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1345 exec_name, com, quote (val));
1352 while (*val && *end);
1356 opt.restrict_files_os = restrict_os;
1357 opt.restrict_files_ctrl = restrict_ctrl;
1358 opt.restrict_files_case = restrict_case;
1359 opt.restrict_files_nonascii = restrict_nonascii;
1366 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1368 static const struct decode_item choices[] = {
1369 { "auto", secure_protocol_auto },
1370 { "sslv2", secure_protocol_sslv2 },
1371 { "sslv3", secure_protocol_sslv3 },
1372 { "tlsv1", secure_protocol_tlsv1 },
1374 int ok = decode_string (val, choices, countof (choices), place);
1376 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1381 /* Set all three timeout values. */
1384 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1387 if (!cmd_time (com, val, &value))
1389 opt.read_timeout = value;
1390 opt.connect_timeout = value;
1391 opt.dns_timeout = value;
1396 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1398 /* Disallow embedded newlines. */
1399 if (strchr (val, '\n'))
1401 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1402 exec_name, com, quote (val));
1405 xfree_null (opt.useragent);
1406 opt.useragent = xstrdup (val);
1410 /* The "verbose" option cannot be cmd_boolean because the variable is
1411 not bool -- it's of type int (-1 means uninitialized because of
1412 some random hackery for disallowing -q -v). */
1415 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1418 if (cmd_boolean (com, val, &flag))
1426 /* Miscellaneous useful routines. */
1428 /* A very simple atoi clone, more useful than atoi because it works on
1429 delimited strings, and has error reportage. Returns true on success,
1430 false on failure. If successful, stores result to *DEST. */
1433 simple_atoi (const char *beg, const char *end, int *dest)
1436 bool negative = false;
1437 const char *p = beg;
1439 while (p < end && c_isspace (*p))
1441 if (p < end && (*p == '-' || *p == '+'))
1443 negative = (*p == '-');
1449 /* Read negative numbers in a separate loop because the most
1450 negative integer cannot be represented as a positive number. */
1453 for (; p < end && c_isdigit (*p); p++)
1455 int next = (10 * result) + (*p - '0');
1457 return false; /* overflow */
1461 for (; p < end && c_isdigit (*p); p++)
1463 int next = (10 * result) - (*p - '0');
1465 return false; /* underflow */
1476 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1477 doesn't handle exponential notation. Returns true on success,
1478 false on failure. In case of success, stores its result to
1482 simple_atof (const char *beg, const char *end, double *dest)
1486 bool negative = false;
1487 bool seen_dot = false;
1488 bool seen_digit = false;
1491 const char *p = beg;
1493 while (p < end && c_isspace (*p))
1495 if (p < end && (*p == '-' || *p == '+'))
1497 negative = (*p == '-');
1501 for (; p < end; p++)
1507 result = (10 * result) + (ch - '0');
1509 result += (ch - '0') / (divider *= 10);
1531 /* Verify that the user-specified header in S is valid. It must
1532 contain a colon preceded by non-white-space characters and must not
1533 contain newlines. */
1536 check_user_specified_header (const char *s)
1540 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1542 /* The header MUST contain `:' preceded by at least one
1543 non-whitespace character. */
1544 if (*p != ':' || p == s)
1546 /* The header MUST NOT contain newlines. */
1547 if (strchr (s, '\n'))
1552 /* Decode VAL into a number, according to ITEMS. */
1555 decode_string (const char *val, const struct decode_item *items, int itemcount,
1559 for (i = 0; i < itemcount; i++)
1560 if (0 == strcasecmp (val, items[i].name))
1562 *place = items[i].code;
1569 void cleanup_html_url (void);
1572 /* Free the memory allocated by global variables. */
1576 /* Free external resources, close files, etc. */
1579 fclose (output_stream);
1580 /* No need to check for error because Wget flushes its output (and
1581 checks for errors) after any data arrives. */
1583 /* We're exiting anyway so there's no real need to call free()
1584 hundreds of times. Skipping the frees will make Wget exit
1587 However, when detecting leaks, it's crucial to free() everything
1588 because then you can find the real leaks, i.e. the allocated
1589 memory which grows with the size of the program. */
1595 cleanup_html_url ();
1600 extern acc_t *netrc_list;
1601 free_netrc (netrc_list);
1603 xfree_null (opt.choose_config);
1604 xfree_null (opt.lfilename);
1605 xfree_null (opt.dir_prefix);
1606 xfree_null (opt.input_filename);
1607 xfree_null (opt.output_document);
1608 free_vec (opt.accepts);
1609 free_vec (opt.rejects);
1610 free_vec (opt.excludes);
1611 free_vec (opt.includes);
1612 free_vec (opt.domains);
1613 free_vec (opt.follow_tags);
1614 free_vec (opt.ignore_tags);
1615 xfree_null (opt.progress_type);
1616 xfree_null (opt.ftp_user);
1617 xfree_null (opt.ftp_passwd);
1618 xfree_null (opt.ftp_proxy);
1619 xfree_null (opt.https_proxy);
1620 xfree_null (opt.http_proxy);
1621 free_vec (opt.no_proxy);
1622 xfree_null (opt.useragent);
1623 xfree_null (opt.referer);
1624 xfree_null (opt.http_user);
1625 xfree_null (opt.http_passwd);
1626 free_vec (opt.user_headers);
1628 xfree_null (opt.cert_file);
1629 xfree_null (opt.private_key);
1630 xfree_null (opt.ca_directory);
1631 xfree_null (opt.ca_cert);
1632 xfree_null (opt.random_file);
1633 xfree_null (opt.egd_file);
1635 xfree_null (opt.bind_address);
1636 xfree_null (opt.cookies_input);
1637 xfree_null (opt.cookies_output);
1638 xfree_null (opt.user);
1639 xfree_null (opt.passwd);
1640 xfree_null (opt.base_href);
1642 #endif /* DEBUG_MALLOC */
1645 /* Unit testing routines. */
1650 test_commands_sorted()
1652 int prev_idx = 0, next_idx = 1;
1653 int command_count = countof (commands) - 1;
1655 while (next_idx <= command_count)
1657 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1660 mu_assert ("FAILED", false);
1673 test_cmd_spec_restrict_file_names()
1678 int expected_restrict_files_os;
1679 int expected_restrict_files_ctrl;
1680 int expected_restrict_files_case;
1683 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1684 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1685 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1686 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1689 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1694 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1697 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1698 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1699 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1700 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1702 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1703 res == test_array[i].result
1704 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1705 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1706 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1712 #endif /* TESTING */