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;
365 /* Return the user's home directory (strdup-ed), or NULL if none is
370 static char *buf = NULL;
371 static char *home, *ret;
376 home = getenv ("HOME");
380 /* Under MSDOS, if $HOME isn't defined, use the directory where
381 `wget.exe' resides. */
382 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
385 buff = _w32_get_argv0 ();
387 p = strrchr (buf, '/'); /* djgpp */
389 p = strrchr (buf, '\\'); /* others */
393 buff = malloc (len + 1);
397 strncpy (buff, _w32_get_argv0 (), len);
401 #elif !defined(WINDOWS)
402 /* If HOME is not defined, try getting it from the password
404 struct passwd *pwd = getpwuid (getuid ());
405 if (!pwd || !pwd->pw_dir)
409 /* Under Windows, if $HOME isn't defined, use the directory where
410 `wget.exe' resides. */
416 ret = home ? xstrdup (home) : NULL;
423 /* Check the 'WGETRC' environment variable and return the file name
424 if 'WGETRC' is set and is a valid file.
425 If the `WGETRC' variable exists but the file does not exist, the
426 function will exit(). */
428 wgetrc_env_file_name (void)
430 char *env = getenv ("WGETRC");
433 if (!file_exists_p (env))
435 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
439 return xstrdup (env);
444 /* Check for the existance of '$HOME/.wgetrc' and return it's path
445 if it exists and is set. */
447 wgetrc_user_file_name (void)
451 /* If that failed, try $HOME/.wgetrc (or equivalent). */
454 file = "SYS$LOGIN:.wgetrc";
455 #else /* def __VMS */
458 file = aprintf ("%s/.wgetrc", home);
460 #endif /* def __VMS [else] */
464 if (!file_exists_p (file))
472 /* Return the path to the user's .wgetrc. This is either the value of
473 `WGETRC' environment variable, or `$HOME/.wgetrc'.
475 Additionally, for windows, look in the directory where wget.exe
478 wgetrc_file_name (void)
480 char *file = wgetrc_env_file_name ();
484 file = wgetrc_user_file_name ();
487 /* Under Windows, if we still haven't found .wgetrc, look for the file
488 `wget.ini' in the directory where `wget.exe' resides; we do this for
489 backward compatibility with previous versions of Wget.
490 SYSTEM_WGETRC should not be defined under WINDOWS. */
493 char *home = home_dir ();
499 file = aprintf ("%s/wget.ini", home);
500 if (!file_exists_p (file))
513 /* Return values of parse_line. */
521 static enum parse_line parse_line (const char *, char **, char **, int *);
522 static bool setval_internal (int, const char *, const char *);
523 static bool setval_internal_tilde (int, const char *, const char *);
525 /* Initialize variables from a wgetrc file. Returns zero (failure) if
526 there were errors in the file. */
529 run_wgetrc (const char *file)
536 fp = fopen (file, "r");
539 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
540 file, strerror (errno));
541 return true; /* not a fatal error */
544 while ((line = read_whole_line (fp)) != NULL)
546 char *com = NULL, *val = NULL;
549 /* Parse the line. */
550 switch (parse_line (line, &com, &val, &comind))
553 /* If everything is OK, set the value. */
554 if (!setval_internal_tilde (comind, com, val))
556 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
557 exec_name, file, ln);
561 case line_syntax_error:
562 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
563 exec_name, file, ln);
566 case line_unknown_command:
567 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
568 exec_name, quote (com), file, ln);
586 /* Initialize the defaults and run the system wgetrc and user's own
591 char *file, *env_sysrc;
594 /* Run a non-standard system rc file when the according environment
595 variable has been set. For internal testing purposes only! */
596 env_sysrc = getenv ("SYSTEM_WGETRC");
597 if (env_sysrc && file_exists_p (env_sysrc))
598 ok &= run_wgetrc (env_sysrc);
599 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
601 else if (file_exists_p (SYSTEM_WGETRC))
602 ok &= run_wgetrc (SYSTEM_WGETRC);
604 /* If there are any problems parsing the system wgetrc file, tell
608 fprintf (stderr, _("\
609 Parsing system wgetrc file failed, please check '%s'. \
610 Or specify a different file using --config\n"), SYSTEM_WGETRC);
613 /* Override it with your own, if one exists. */
614 file = wgetrc_file_name ();
617 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
618 something like realpath() before comparing them with `strcmp' */
620 if (!strcmp (file, SYSTEM_WGETRC))
622 fprintf (stderr, _("\
623 %s: Warning: Both system and user wgetrc point to %s.\n"),
624 exec_name, quote (file));
628 ok &= run_wgetrc (file);
630 /* If there were errors processing either `.wgetrc', abort. */
638 /* Remove dashes and underscores from S, modifying S in the
644 char *t = s; /* t - tortoise */
645 char *h = s; /* h - hare */
647 if (*h == '_' || *h == '-')
654 /* Parse the line pointed by line, with the syntax:
655 <sp>* command <sp>* = <sp>* value <sp>*
656 Uses malloc to allocate space for command and value.
658 Returns one of line_ok, line_empty, line_syntax_error, or
659 line_unknown_command.
661 In case of line_ok, *COM and *VAL point to freshly allocated
662 strings, and *COMIND points to com's index. In case of error or
663 empty line, their values are unmodified. */
665 static enum parse_line
666 parse_line (const char *line, char **com, char **val, int *comind)
669 const char *end = line + strlen (line);
670 const char *cmdstart, *cmdend;
671 const char *valstart, *valend;
676 /* Skip leading and trailing whitespace. */
677 while (*line && c_isspace (*line))
679 while (end > line && c_isspace (end[-1]))
682 /* Skip empty lines and comments. */
683 if (!*line || *line == '#')
689 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
693 /* Skip '=', as well as any space before or after it. */
694 while (p < end && c_isspace (*p))
696 if (p == end || *p != '=')
697 return line_syntax_error;
699 while (p < end && c_isspace (*p))
705 /* The syntax is valid (even though the command might not be). Fill
706 in the command name and value. */
707 *com = strdupdelim (cmdstart, cmdend);
708 *val = strdupdelim (valstart, valend);
710 /* The line now known to be syntactically correct. Check whether
711 the command is valid. */
712 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
714 ind = command_by_name (cmdcopy);
716 return line_unknown_command;
718 /* Report success to the caller. */
723 #if defined(WINDOWS) || defined(MSDOS)
724 # define ISSEP(c) ((c) == '/' || (c) == '\\')
726 # define ISSEP(c) ((c) == '/')
729 /* Run commands[comind].action. */
732 setval_internal (int comind, const char *com, const char *val)
734 assert (0 <= comind && ((size_t) comind) < countof (commands));
735 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
736 return commands[comind].action (com, val, commands[comind].place);
740 setval_internal_tilde (int comind, const char *com, const char *val)
746 ret = setval_internal (comind, com, val);
748 /* We make tilde expansion for cmd_file and cmd_directory */
749 if (((commands[comind].action == cmd_file) ||
750 (commands[comind].action == cmd_directory))
751 && ret && (*val == '~' && ISSEP (val[1])))
753 pstring = commands[comind].place;
757 homelen = strlen (home);
758 while (homelen && ISSEP (home[homelen - 1]))
759 home[--homelen] = '\0';
761 /* Skip the leading "~/". */
762 for (++val; ISSEP (*val); val++)
764 *pstring = concat_strings (home, "/", val, (char *)0);
770 /* Run command COM with value VAL. If running the command produces an
771 error, report the error and exit.
773 This is intended to be called from main() to modify Wget's behavior
774 through command-line switches. Since COM is hard-coded in main(),
775 it is not canonicalized, and this aborts when COM is not found.
777 If COMIND's are exported to init.h, this function will be changed
778 to accept COMIND directly. */
781 setoptval (const char *com, const char *val, const char *optname)
783 /* Prepend "--" to OPTNAME. */
784 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
787 strcpy (dd_optname + 2, optname);
789 assert (val != NULL);
790 if (!setval_internal (command_by_name (com), dd_optname, val))
794 /* Parse OPT into command and value and run it. For example,
795 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
796 This is used by the `--execute' flag in main.c. */
799 run_command (const char *opt)
803 switch (parse_line (opt, &com, &val, &comind))
806 if (!setval_internal (comind, com, val))
812 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
813 exec_name, quote (opt));
818 /* Generic helper functions, for use with `commands'. */
820 /* Forward declarations: */
825 static bool decode_string (const char *, const struct decode_item *, int, int *);
826 static bool simple_atoi (const char *, const char *, int *);
827 static bool simple_atof (const char *, const char *, double *);
829 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
831 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
832 && c_tolower((p)[1]) == (c1) \
835 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
836 && c_tolower((p)[1]) == (c1) \
837 && c_tolower((p)[2]) == (c2) \
841 /* Store the boolean value from VAL to PLACE. COM is ignored,
842 except for error messages. */
844 cmd_boolean (const char *com, const char *val, void *place)
848 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
849 /* "on", "yes" and "1" mean true. */
851 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
852 /* "off", "no" and "0" mean false. */
857 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
858 exec_name, com, quote (val));
862 *(bool *) place = value;
866 /* Set the non-negative integer value from VAL to PLACE. With
867 incorrect specification, the number remains unchanged. */
869 cmd_number (const char *com, const char *val, void *place)
871 if (!simple_atoi (val, val + strlen (val), place)
872 || *(int *) place < 0)
874 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
875 exec_name, com, quote (val));
881 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
883 cmd_number_inf (const char *com, const char *val, void *place)
885 if (!strcasecmp (val, "inf"))
890 return cmd_number (com, val, place);
893 /* Copy (strdup) the string at COM to a new location and place a
894 pointer to *PLACE. */
896 cmd_string (const char *com, const char *val, void *place)
898 char **pstring = (char **)place;
900 xfree_null (*pstring);
901 *pstring = xstrdup (val);
906 /* Like the above, but handles tilde-expansion when reading a user's
907 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
908 gets expanded to the user's home directory. */
910 cmd_file (const char *com, const char *val, void *place)
912 char **pstring = (char **)place;
914 xfree_null (*pstring);
916 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
918 *pstring = xstrdup (val);
920 #if defined(WINDOWS) || defined(MSDOS)
921 /* Convert "\" to "/". */
924 for (s = *pstring; *s; s++)
932 /* Like cmd_file, but strips trailing '/' characters. */
934 cmd_directory (const char *com, const char *val, void *place)
938 /* Call cmd_file() for tilde expansion and separator
939 canonicalization (backslash -> slash under Windows). These
940 things should perhaps be in a separate function. */
941 if (!cmd_file (com, val, place))
946 while (t > s && *--t == '/')
952 /* Split VAL by space to a vector of values, and append those values
953 to vector pointed to by the PLACE argument. If VAL is empty, the
954 PLACE vector is cleared instead. */
957 cmd_vector (const char *com, const char *val, void *place)
959 char ***pvec = (char ***)place;
962 *pvec = merge_vecs (*pvec, sepstring (val));
972 cmd_directory_vector (const char *com, const char *val, void *place)
974 char ***pvec = (char ***)place;
978 /* Strip the trailing slashes from directories. */
981 seps = sepstring (val);
982 for (t = seps; t && *t; t++)
984 int len = strlen (*t);
985 /* Skip degenerate case of root directory. */
988 if ((*t)[len - 1] == '/')
989 (*t)[len - 1] = '\0';
992 *pvec = merge_vecs (*pvec, seps);
1002 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1003 "100k" or "2.5G" to a floating point number. */
1006 parse_bytes_helper (const char *val, double *result)
1008 double number, mult;
1009 const char *end = val + strlen (val);
1011 /* Check for "inf". */
1012 if (0 == strcmp (val, "inf"))
1018 /* Strip trailing whitespace. */
1019 while (val < end && c_isspace (end[-1]))
1024 switch (c_tolower (end[-1]))
1027 --end, mult = 1024.0;
1030 --end, mult = 1048576.0;
1033 --end, mult = 1073741824.0;
1036 --end, mult = 1099511627776.0;
1039 /* Not a recognized suffix: assume it's a digit. (If not,
1040 simple_atof will raise an error.) */
1044 /* Skip leading and trailing whitespace. */
1045 while (val < end && c_isspace (*val))
1047 while (val < end && c_isspace (end[-1]))
1052 if (!simple_atof (val, end, &number) || number < 0)
1055 *result = number * mult;
1059 /* Parse VAL as a number and set its value to PLACE (which should
1062 By default, the value is assumed to be in bytes. If "K", "M", or
1063 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1064 1<<30, respectively. Floating point values are allowed and are
1065 cast to integer before use. The idea is to be able to use things
1066 like 1.5k instead of "1536".
1068 The string "inf" is returned as 0.
1070 In case of error, false is returned and memory pointed to by PLACE
1071 remains unmodified. */
1074 cmd_bytes (const char *com, const char *val, void *place)
1077 if (!parse_bytes_helper (val, &byte_value))
1079 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1080 exec_name, com, quote (val));
1083 *(wgint *)place = (wgint)byte_value;
1087 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1088 SIZE_SUM. It works by converting the string to double, therefore
1089 working with values up to 2^53-1 without loss of precision. This
1090 value (8192 TB) is large enough to serve for a while. */
1093 cmd_bytes_sum (const char *com, const char *val, void *place)
1096 if (!parse_bytes_helper (val, &byte_value))
1098 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1099 exec_name, com, quote (val));
1102 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1106 /* Store the value of VAL to *OUT. The value is a time period, by
1107 default expressed in seconds, but also accepting suffixes "m", "h",
1108 "d", and "w" for minutes, hours, days, and weeks respectively. */
1111 cmd_time (const char *com, const char *val, void *place)
1113 double number, mult;
1114 const char *end = val + strlen (val);
1116 /* Strip trailing whitespace. */
1117 while (val < end && c_isspace (end[-1]))
1123 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1124 exec_name, com, quote (val));
1128 switch (c_tolower (end[-1]))
1131 --end, mult = 1; /* seconds */
1134 --end, mult = 60; /* minutes */
1137 --end, mult = 3600; /* hours */
1140 --end, mult = 86400.0; /* days */
1143 --end, mult = 604800.0; /* weeks */
1146 /* Not a recognized suffix: assume it belongs to the number.
1147 (If not, simple_atof will raise an error.) */
1151 /* Skip leading and trailing whitespace. */
1152 while (val < end && c_isspace (*val))
1154 while (val < end && c_isspace (end[-1]))
1159 if (!simple_atof (val, end, &number))
1162 *(double *)place = number * mult;
1168 cmd_cert_type (const char *com, const char *val, void *place)
1170 static const struct decode_item choices[] = {
1171 { "pem", keyfile_pem },
1172 { "der", keyfile_asn1 },
1173 { "asn1", keyfile_asn1 },
1175 int ok = decode_string (val, choices, countof (choices), place);
1177 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1182 /* Specialized helper functions, used by `commands' to handle some
1183 options specially. */
1185 static bool check_user_specified_header (const char *);
1188 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1190 if (!cmd_boolean (com, val, &opt.dirstruct))
1192 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1193 must be affected inversely. */
1195 opt.no_dirstruct = false;
1197 opt.no_dirstruct = true;
1202 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1204 /* Empty value means reset the list of headers. */
1207 free_vec (opt.user_headers);
1208 opt.user_headers = NULL;
1212 if (!check_user_specified_header (val))
1214 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1215 exec_name, com, quote (val));
1218 opt.user_headers = vec_append (opt.user_headers, val);
1223 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1225 int flag = cmd_boolean (com, val, &opt.htmlify);
1226 if (flag && !opt.htmlify)
1227 opt.remove_listing = false;
1231 /* Set the "mirror" mode. It means: recursive download, timestamping,
1232 no limit on max. recursion depth, and don't remove listings. */
1235 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1239 if (!cmd_boolean (com, val, &mirror))
1243 opt.recursive = true;
1244 if (!opt.no_dirstruct)
1245 opt.dirstruct = true;
1246 opt.timestamping = true;
1247 opt.reclevel = INFINITE_RECURSION;
1248 opt.remove_listing = false;
1253 /* Validate --prefer-family and set the choice. Allowed values are
1254 "IPv4", "IPv6", and "none". */
1257 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1259 static const struct decode_item choices[] = {
1260 { "IPv4", prefer_ipv4 },
1261 { "IPv6", prefer_ipv6 },
1262 { "none", prefer_none },
1264 int prefer_family = prefer_none;
1265 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1267 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1268 opt.prefer_family = prefer_family;
1272 /* Set progress.type to VAL, but verify that it's a valid progress
1273 implementation before that. */
1276 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1278 if (!valid_progress_implementation_p (val))
1280 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1281 exec_name, com, quote (val));
1284 xfree_null (opt.progress_type);
1286 /* Don't call set_progress_implementation here. It will be called
1287 in main() when it becomes clear what the log output is. */
1288 opt.progress_type = xstrdup (val);
1292 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1293 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1297 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1299 if (!cmd_boolean (com, val, &opt.recursive))
1303 if (opt.recursive && !opt.no_dirstruct)
1304 opt.dirstruct = true;
1310 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1312 int restrict_os = opt.restrict_files_os;
1313 int restrict_ctrl = opt.restrict_files_ctrl;
1314 int restrict_case = opt.restrict_files_case;
1315 int restrict_nonascii = opt.restrict_files_nonascii;
1319 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1323 end = strchr (val, ',');
1325 end = val + strlen (val);
1327 if (VAL_IS ("unix"))
1328 restrict_os = restrict_unix;
1329 else if (VAL_IS ("windows"))
1330 restrict_os = restrict_windows;
1331 else if (VAL_IS ("lowercase"))
1332 restrict_case = restrict_lowercase;
1333 else if (VAL_IS ("uppercase"))
1334 restrict_case = restrict_uppercase;
1335 else if (VAL_IS ("nocontrol"))
1336 restrict_ctrl = false;
1337 else if (VAL_IS ("ascii"))
1338 restrict_nonascii = true;
1341 fprintf (stderr, _("\
1342 %s: %s: Invalid restriction %s,\n\
1343 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1344 exec_name, com, quote (val));
1351 while (*val && *end);
1355 opt.restrict_files_os = restrict_os;
1356 opt.restrict_files_ctrl = restrict_ctrl;
1357 opt.restrict_files_case = restrict_case;
1358 opt.restrict_files_nonascii = restrict_nonascii;
1365 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1367 static const struct decode_item choices[] = {
1368 { "auto", secure_protocol_auto },
1369 { "sslv2", secure_protocol_sslv2 },
1370 { "sslv3", secure_protocol_sslv3 },
1371 { "tlsv1", secure_protocol_tlsv1 },
1373 int ok = decode_string (val, choices, countof (choices), place);
1375 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1380 /* Set all three timeout values. */
1383 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1386 if (!cmd_time (com, val, &value))
1388 opt.read_timeout = value;
1389 opt.connect_timeout = value;
1390 opt.dns_timeout = value;
1395 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1397 /* Disallow embedded newlines. */
1398 if (strchr (val, '\n'))
1400 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1401 exec_name, com, quote (val));
1404 xfree_null (opt.useragent);
1405 opt.useragent = xstrdup (val);
1409 /* The "verbose" option cannot be cmd_boolean because the variable is
1410 not bool -- it's of type int (-1 means uninitialized because of
1411 some random hackery for disallowing -q -v). */
1414 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1417 if (cmd_boolean (com, val, &flag))
1425 /* Miscellaneous useful routines. */
1427 /* A very simple atoi clone, more useful than atoi because it works on
1428 delimited strings, and has error reportage. Returns true on success,
1429 false on failure. If successful, stores result to *DEST. */
1432 simple_atoi (const char *beg, const char *end, int *dest)
1435 bool negative = false;
1436 const char *p = beg;
1438 while (p < end && c_isspace (*p))
1440 if (p < end && (*p == '-' || *p == '+'))
1442 negative = (*p == '-');
1448 /* Read negative numbers in a separate loop because the most
1449 negative integer cannot be represented as a positive number. */
1452 for (; p < end && c_isdigit (*p); p++)
1454 int next = (10 * result) + (*p - '0');
1456 return false; /* overflow */
1460 for (; p < end && c_isdigit (*p); p++)
1462 int next = (10 * result) - (*p - '0');
1464 return false; /* underflow */
1475 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1476 doesn't handle exponential notation. Returns true on success,
1477 false on failure. In case of success, stores its result to
1481 simple_atof (const char *beg, const char *end, double *dest)
1485 bool negative = false;
1486 bool seen_dot = false;
1487 bool seen_digit = false;
1490 const char *p = beg;
1492 while (p < end && c_isspace (*p))
1494 if (p < end && (*p == '-' || *p == '+'))
1496 negative = (*p == '-');
1500 for (; p < end; p++)
1506 result = (10 * result) + (ch - '0');
1508 result += (ch - '0') / (divider *= 10);
1530 /* Verify that the user-specified header in S is valid. It must
1531 contain a colon preceded by non-white-space characters and must not
1532 contain newlines. */
1535 check_user_specified_header (const char *s)
1539 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1541 /* The header MUST contain `:' preceded by at least one
1542 non-whitespace character. */
1543 if (*p != ':' || p == s)
1545 /* The header MUST NOT contain newlines. */
1546 if (strchr (s, '\n'))
1551 /* Decode VAL into a number, according to ITEMS. */
1554 decode_string (const char *val, const struct decode_item *items, int itemcount,
1558 for (i = 0; i < itemcount; i++)
1559 if (0 == strcasecmp (val, items[i].name))
1561 *place = items[i].code;
1568 void cleanup_html_url (void);
1571 /* Free the memory allocated by global variables. */
1575 /* Free external resources, close files, etc. */
1578 fclose (output_stream);
1579 /* No need to check for error because Wget flushes its output (and
1580 checks for errors) after any data arrives. */
1582 /* We're exiting anyway so there's no real need to call free()
1583 hundreds of times. Skipping the frees will make Wget exit
1586 However, when detecting leaks, it's crucial to free() everything
1587 because then you can find the real leaks, i.e. the allocated
1588 memory which grows with the size of the program. */
1594 cleanup_html_url ();
1599 extern acc_t *netrc_list;
1600 free_netrc (netrc_list);
1602 xfree_null (opt.choose_config);
1603 xfree_null (opt.lfilename);
1604 xfree_null (opt.dir_prefix);
1605 xfree_null (opt.input_filename);
1606 xfree_null (opt.output_document);
1607 free_vec (opt.accepts);
1608 free_vec (opt.rejects);
1609 free_vec (opt.excludes);
1610 free_vec (opt.includes);
1611 free_vec (opt.domains);
1612 free_vec (opt.follow_tags);
1613 free_vec (opt.ignore_tags);
1614 xfree_null (opt.progress_type);
1615 xfree_null (opt.ftp_user);
1616 xfree_null (opt.ftp_passwd);
1617 xfree_null (opt.ftp_proxy);
1618 xfree_null (opt.https_proxy);
1619 xfree_null (opt.http_proxy);
1620 free_vec (opt.no_proxy);
1621 xfree_null (opt.useragent);
1622 xfree_null (opt.referer);
1623 xfree_null (opt.http_user);
1624 xfree_null (opt.http_passwd);
1625 free_vec (opt.user_headers);
1627 xfree_null (opt.cert_file);
1628 xfree_null (opt.private_key);
1629 xfree_null (opt.ca_directory);
1630 xfree_null (opt.ca_cert);
1631 xfree_null (opt.random_file);
1632 xfree_null (opt.egd_file);
1634 xfree_null (opt.bind_address);
1635 xfree_null (opt.cookies_input);
1636 xfree_null (opt.cookies_output);
1637 xfree_null (opt.user);
1638 xfree_null (opt.passwd);
1639 xfree_null (opt.base_href);
1641 #endif /* DEBUG_MALLOC */
1644 /* Unit testing routines. */
1649 test_commands_sorted()
1651 int prev_idx = 0, next_idx = 1;
1652 int command_count = countof (commands) - 1;
1654 while (next_idx <= command_count)
1656 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1659 mu_assert ("FAILED", false);
1672 test_cmd_spec_restrict_file_names()
1677 int expected_restrict_files_os;
1678 int expected_restrict_files_ctrl;
1679 int expected_restrict_files_case;
1682 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1683 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1684 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1685 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1688 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1693 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1696 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1697 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1698 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1699 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1701 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1702 res == test_array[i].result
1703 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1704 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1705 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1711 #endif /* TESTING */