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 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 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
61 #include "recur.h" /* for INFINITE_RECURSION */
62 #include "convert.h" /* for convert_cleanup */
63 #include "res.h" /* for res_cleanup */
64 #include "http.h" /* for http_cleanup */
65 #include "retr.h" /* for output_stream */
73 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
75 CMD_DECLARE (cmd_boolean);
76 CMD_DECLARE (cmd_bytes);
77 CMD_DECLARE (cmd_bytes_sum);
79 CMD_DECLARE (cmd_cert_type);
81 CMD_DECLARE (cmd_directory_vector);
82 CMD_DECLARE (cmd_number);
83 CMD_DECLARE (cmd_number_inf);
84 CMD_DECLARE (cmd_string);
85 CMD_DECLARE (cmd_file);
86 CMD_DECLARE (cmd_directory);
87 CMD_DECLARE (cmd_time);
88 CMD_DECLARE (cmd_vector);
90 CMD_DECLARE (cmd_spec_dirstruct);
91 CMD_DECLARE (cmd_spec_header);
92 CMD_DECLARE (cmd_spec_htmlify);
93 CMD_DECLARE (cmd_spec_mirror);
94 CMD_DECLARE (cmd_spec_prefer_family);
95 CMD_DECLARE (cmd_spec_progress);
96 CMD_DECLARE (cmd_spec_recursive);
97 CMD_DECLARE (cmd_spec_restrict_file_names);
99 CMD_DECLARE (cmd_spec_secure_protocol);
101 CMD_DECLARE (cmd_spec_timeout);
102 CMD_DECLARE (cmd_spec_useragent);
103 CMD_DECLARE (cmd_spec_verbose);
105 /* List of recognized commands, each consisting of name, place and
106 function. When adding a new command, simply add it to the list,
107 but be sure to keep the list sorted alphabetically, as
108 command_by_name's binary search depends on it. Also, be sure to
109 add any entries that allocate memory (e.g. cmd_string and
110 cmd_vector) to the cleanup() function below. */
112 static const struct {
115 bool (*action) (const char *, const char *, void *);
117 /* KEEP THIS LIST ALPHABETICALLY SORTED */
118 { "accept", &opt.accepts, cmd_vector },
119 { "addhostdir", &opt.add_hostdir, cmd_boolean },
120 { "adjustextension", &opt.adjust_extension, cmd_boolean },
121 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
122 { "askpassword", &opt.ask_passwd, cmd_boolean },
123 { "authnochallenge", &opt.auth_without_challenge,
125 { "background", &opt.background, cmd_boolean },
126 { "backupconverted", &opt.backup_converted, cmd_boolean },
127 { "backups", &opt.backups, cmd_number },
128 { "base", &opt.base_href, cmd_string },
129 { "bindaddress", &opt.bind_address, cmd_string },
131 { "cacertificate", &opt.ca_cert, cmd_file },
133 { "cache", &opt.allow_cache, cmd_boolean },
135 { "cadirectory", &opt.ca_directory, cmd_directory },
136 { "certificate", &opt.cert_file, cmd_file },
137 { "certificatetype", &opt.cert_type, cmd_cert_type },
138 { "checkcertificate", &opt.check_cert, cmd_boolean },
140 { "chooseconfig", &opt.choose_config, cmd_file },
141 { "connecttimeout", &opt.connect_timeout, cmd_time },
142 { "contentdisposition", &opt.content_disposition, cmd_boolean },
143 { "continue", &opt.always_rest, cmd_boolean },
144 { "convertlinks", &opt.convert_links, cmd_boolean },
145 { "cookies", &opt.cookies, cmd_boolean },
146 { "cutdirs", &opt.cut_dirs, cmd_number },
148 { "debug", &opt.debug, cmd_boolean },
150 { "defaultpage", &opt.default_page, cmd_string},
151 { "deleteafter", &opt.delete_after, cmd_boolean },
152 { "dirprefix", &opt.dir_prefix, cmd_directory },
153 { "dirstruct", NULL, cmd_spec_dirstruct },
154 { "dnscache", &opt.dns_cache, cmd_boolean },
155 { "dnstimeout", &opt.dns_timeout, cmd_time },
156 { "domains", &opt.domains, cmd_vector },
157 { "dotbytes", &opt.dot_bytes, cmd_bytes },
158 { "dotsinline", &opt.dots_in_line, cmd_number },
159 { "dotspacing", &opt.dot_spacing, cmd_number },
160 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
162 { "egdfile", &opt.egd_file, cmd_file },
164 { "excludedirectories", &opt.excludes, cmd_directory_vector },
165 { "excludedomains", &opt.exclude_domains, cmd_vector },
166 { "followftp", &opt.follow_ftp, cmd_boolean },
167 { "followtags", &opt.follow_tags, cmd_vector },
168 { "forcehtml", &opt.force_html, cmd_boolean },
169 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
170 { "ftppassword", &opt.ftp_passwd, cmd_string },
171 { "ftpproxy", &opt.ftp_proxy, cmd_string },
173 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
174 #endif /* def __VMS */
175 { "ftpuser", &opt.ftp_user, cmd_string },
176 { "glob", &opt.ftp_glob, cmd_boolean },
177 { "header", NULL, cmd_spec_header },
178 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
179 { "htmlify", NULL, cmd_spec_htmlify },
180 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
181 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
182 { "httppassword", &opt.http_passwd, cmd_string },
183 { "httpproxy", &opt.http_proxy, cmd_string },
184 { "httpsproxy", &opt.https_proxy, cmd_string },
185 { "httpuser", &opt.http_user, cmd_string },
186 { "ignorecase", &opt.ignore_case, cmd_boolean },
187 { "ignorelength", &opt.ignore_length, cmd_boolean },
188 { "ignoretags", &opt.ignore_tags, cmd_vector },
189 { "includedirectories", &opt.includes, cmd_directory_vector },
191 { "inet4only", &opt.ipv4_only, cmd_boolean },
192 { "inet6only", &opt.ipv6_only, cmd_boolean },
194 { "input", &opt.input_filename, cmd_file },
195 { "iri", &opt.enable_iri, cmd_boolean },
196 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
197 { "limitrate", &opt.limit_rate, cmd_bytes },
198 { "loadcookies", &opt.cookies_input, cmd_file },
199 { "localencoding", &opt.locale, cmd_string },
200 { "logfile", &opt.lfilename, cmd_file },
201 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
202 { "maxredirect", &opt.max_redirect, cmd_number },
203 { "mirror", NULL, cmd_spec_mirror },
204 { "netrc", &opt.netrc, cmd_boolean },
205 { "noclobber", &opt.noclobber, cmd_boolean },
206 { "noparent", &opt.no_parent, cmd_boolean },
207 { "noproxy", &opt.no_proxy, cmd_vector },
208 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
209 { "outputdocument", &opt.output_document, cmd_file },
210 { "pagerequisites", &opt.page_requisites, cmd_boolean },
211 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
212 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
213 { "password", &opt.passwd, cmd_string },
214 { "postdata", &opt.post_data, cmd_string },
215 { "postfile", &opt.post_file_name, cmd_file },
216 { "preferfamily", NULL, cmd_spec_prefer_family },
217 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
219 { "privatekey", &opt.private_key, cmd_file },
220 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
222 { "progress", &opt.progress_type, cmd_spec_progress },
223 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
224 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
225 { "proxypassword", &opt.proxy_passwd, cmd_string },
226 { "proxyuser", &opt.proxy_user, cmd_string },
227 { "quiet", &opt.quiet, cmd_boolean },
228 { "quota", &opt.quota, cmd_bytes_sum },
230 { "randomfile", &opt.random_file, cmd_file },
232 { "randomwait", &opt.random_wait, cmd_boolean },
233 { "readtimeout", &opt.read_timeout, cmd_time },
234 { "reclevel", &opt.reclevel, cmd_number_inf },
235 { "recursive", NULL, cmd_spec_recursive },
236 { "referer", &opt.referer, cmd_string },
237 { "reject", &opt.rejects, cmd_vector },
238 { "relativeonly", &opt.relative_only, cmd_boolean },
239 { "remoteencoding", &opt.encoding_remote, cmd_string },
240 { "removelisting", &opt.remove_listing, cmd_boolean },
241 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
242 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
243 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
244 { "robots", &opt.use_robots, cmd_boolean },
245 { "savecookies", &opt.cookies_output, cmd_file },
246 { "saveheaders", &opt.save_headers, cmd_boolean },
248 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
250 { "serverresponse", &opt.server_response, 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[PATH_MAX];
375 home = getenv ("HOME");
379 /* Under MSDOS, if $HOME isn't defined, use the directory where
380 `wget.exe' resides. */
381 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
384 strcpy (buf, _w32_get_argv0 ());
385 p = strrchr (buf, '/'); /* djgpp */
387 p = strrchr (buf, '\\'); /* others */
391 #elif !defined(WINDOWS)
392 /* If HOME is not defined, try getting it from the password
394 struct passwd *pwd = getpwuid (getuid ());
395 if (!pwd || !pwd->pw_dir)
397 strcpy (buf, pwd->pw_dir);
400 /* Under Windows, if $HOME isn't defined, use the directory where
401 `wget.exe' resides. */
407 return home ? xstrdup (home) : NULL;
410 /* Check the 'WGETRC' environment variable and return the file name
411 if 'WGETRC' is set and is a valid file.
412 If the `WGETRC' variable exists but the file does not exist, the
413 function will exit(). */
415 wgetrc_env_file_name (void)
417 char *env = getenv ("WGETRC");
420 if (!file_exists_p (env))
422 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
426 return xstrdup (env);
431 /* Check for the existance of '$HOME/.wgetrc' and return it's path
432 if it exists and is set. */
434 wgetrc_user_file_name (void)
438 /* If that failed, try $HOME/.wgetrc (or equivalent). */
441 file = "SYS$LOGIN:.wgetrc";
442 #else /* def __VMS */
445 file = aprintf ("%s/.wgetrc", home);
447 #endif /* def __VMS [else] */
451 if (!file_exists_p (file))
459 /* Return the path to the user's .wgetrc. This is either the value of
460 `WGETRC' environment variable, or `$HOME/.wgetrc'.
462 Additionally, for windows, look in the directory where wget.exe
465 wgetrc_file_name (void)
467 char *file = wgetrc_env_file_name ();
471 file = wgetrc_user_file_name ();
474 /* Under Windows, if we still haven't found .wgetrc, look for the file
475 `wget.ini' in the directory where `wget.exe' resides; we do this for
476 backward compatibility with previous versions of Wget.
477 SYSTEM_WGETRC should not be defined under WINDOWS. */
480 char *home = home_dir ();
486 file = aprintf ("%s/wget.ini", home);
487 if (!file_exists_p (file))
500 /* Return values of parse_line. */
508 static enum parse_line parse_line (const char *, char **, char **, int *);
509 static bool setval_internal (int, const char *, const char *);
510 static bool setval_internal_tilde (int, const char *, const char *);
512 /* Initialize variables from a wgetrc file. Returns zero (failure) if
513 there were errors in the file. */
516 run_wgetrc (const char *file)
523 fp = fopen (file, "r");
526 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
527 file, strerror (errno));
528 return true; /* not a fatal error */
531 while ((line = read_whole_line (fp)) != NULL)
533 char *com = NULL, *val = NULL;
536 /* Parse the line. */
537 switch (parse_line (line, &com, &val, &comind))
540 /* If everything is OK, set the value. */
541 if (!setval_internal_tilde (comind, com, val))
543 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
544 exec_name, file, ln);
548 case line_syntax_error:
549 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
550 exec_name, file, ln);
553 case line_unknown_command:
554 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
555 exec_name, quote (com), file, ln);
573 /* Initialize the defaults and run the system wgetrc and user's own
578 char *file, *env_sysrc;
581 /* Run a non-standard system rc file when the according environment
582 variable has been set. For internal testing purposes only! */
583 env_sysrc = getenv ("SYSTEM_WGETRC");
584 if (env_sysrc && file_exists_p (env_sysrc))
585 ok &= run_wgetrc (env_sysrc);
586 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
588 else if (file_exists_p (SYSTEM_WGETRC))
589 ok &= run_wgetrc (SYSTEM_WGETRC);
591 /* Override it with your own, if one exists. */
592 file = wgetrc_file_name ();
595 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
596 something like realpath() before comparing them with `strcmp' */
598 if (!strcmp (file, SYSTEM_WGETRC))
600 fprintf (stderr, _("\
601 %s: Warning: Both system and user wgetrc point to %s.\n"),
602 exec_name, quote (file));
606 ok &= run_wgetrc (file);
608 /* If there were errors processing either `.wgetrc', abort. */
616 /* Remove dashes and underscores from S, modifying S in the
622 char *t = s; /* t - tortoise */
623 char *h = s; /* h - hare */
625 if (*h == '_' || *h == '-')
632 /* Parse the line pointed by line, with the syntax:
633 <sp>* command <sp>* = <sp>* value <sp>*
634 Uses malloc to allocate space for command and value.
636 Returns one of line_ok, line_empty, line_syntax_error, or
637 line_unknown_command.
639 In case of line_ok, *COM and *VAL point to freshly allocated
640 strings, and *COMIND points to com's index. In case of error or
641 empty line, their values are unmodified. */
643 static enum parse_line
644 parse_line (const char *line, char **com, char **val, int *comind)
647 const char *end = line + strlen (line);
648 const char *cmdstart, *cmdend;
649 const char *valstart, *valend;
654 /* Skip leading and trailing whitespace. */
655 while (*line && c_isspace (*line))
657 while (end > line && c_isspace (end[-1]))
660 /* Skip empty lines and comments. */
661 if (!*line || *line == '#')
667 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
671 /* Skip '=', as well as any space before or after it. */
672 while (p < end && c_isspace (*p))
674 if (p == end || *p != '=')
675 return line_syntax_error;
677 while (p < end && c_isspace (*p))
683 /* The syntax is valid (even though the command might not be). Fill
684 in the command name and value. */
685 *com = strdupdelim (cmdstart, cmdend);
686 *val = strdupdelim (valstart, valend);
688 /* The line now known to be syntactically correct. Check whether
689 the command is valid. */
690 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
692 ind = command_by_name (cmdcopy);
694 return line_unknown_command;
696 /* Report success to the caller. */
701 #if defined(WINDOWS) || defined(MSDOS)
702 # define ISSEP(c) ((c) == '/' || (c) == '\\')
704 # define ISSEP(c) ((c) == '/')
707 /* Run commands[comind].action. */
710 setval_internal (int comind, const char *com, const char *val)
712 assert (0 <= comind && ((size_t) comind) < countof (commands));
713 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
714 return commands[comind].action (com, val, commands[comind].place);
718 setval_internal_tilde (int comind, const char *com, const char *val)
724 ret = setval_internal (comind, com, val);
726 /* We make tilde expansion for cmd_file and cmd_directory */
727 if (((commands[comind].action == cmd_file) ||
728 (commands[comind].action == cmd_directory))
729 && ret && (*val == '~' && ISSEP (val[1])))
731 pstring = commands[comind].place;
735 homelen = strlen (home);
736 while (homelen && ISSEP (home[homelen - 1]))
737 home[--homelen] = '\0';
739 /* Skip the leading "~/". */
740 for (++val; ISSEP (*val); val++)
742 *pstring = concat_strings (home, "/", val, (char *)0);
748 /* Run command COM with value VAL. If running the command produces an
749 error, report the error and exit.
751 This is intended to be called from main() to modify Wget's behavior
752 through command-line switches. Since COM is hard-coded in main(),
753 it is not canonicalized, and this aborts when COM is not found.
755 If COMIND's are exported to init.h, this function will be changed
756 to accept COMIND directly. */
759 setoptval (const char *com, const char *val, const char *optname)
761 /* Prepend "--" to OPTNAME. */
762 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
765 strcpy (dd_optname + 2, optname);
767 assert (val != NULL);
768 if (!setval_internal (command_by_name (com), dd_optname, val))
772 /* Parse OPT into command and value and run it. For example,
773 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
774 This is used by the `--execute' flag in main.c. */
777 run_command (const char *opt)
781 switch (parse_line (opt, &com, &val, &comind))
784 if (!setval_internal (comind, com, val))
790 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
791 exec_name, quote (opt));
796 /* Generic helper functions, for use with `commands'. */
798 /* Forward declarations: */
803 static bool decode_string (const char *, const struct decode_item *, int, int *);
804 static bool simple_atoi (const char *, const char *, int *);
805 static bool simple_atof (const char *, const char *, double *);
807 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
809 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
810 && c_tolower((p)[1]) == (c1) \
813 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
814 && c_tolower((p)[1]) == (c1) \
815 && c_tolower((p)[2]) == (c2) \
819 /* Store the boolean value from VAL to PLACE. COM is ignored,
820 except for error messages. */
822 cmd_boolean (const char *com, const char *val, void *place)
826 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
827 /* "on", "yes" and "1" mean true. */
829 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
830 /* "off", "no" and "0" mean false. */
835 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
836 exec_name, com, quote (val));
840 *(bool *) place = value;
844 /* Set the non-negative integer value from VAL to PLACE. With
845 incorrect specification, the number remains unchanged. */
847 cmd_number (const char *com, const char *val, void *place)
849 if (!simple_atoi (val, val + strlen (val), place)
850 || *(int *) place < 0)
852 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
853 exec_name, com, quote (val));
859 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
861 cmd_number_inf (const char *com, const char *val, void *place)
863 if (!strcasecmp (val, "inf"))
868 return cmd_number (com, val, place);
871 /* Copy (strdup) the string at COM to a new location and place a
872 pointer to *PLACE. */
874 cmd_string (const char *com, const char *val, void *place)
876 char **pstring = (char **)place;
878 xfree_null (*pstring);
879 *pstring = xstrdup (val);
884 /* Like the above, but handles tilde-expansion when reading a user's
885 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
886 gets expanded to the user's home directory. */
888 cmd_file (const char *com, const char *val, void *place)
890 char **pstring = (char **)place;
892 xfree_null (*pstring);
894 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
896 *pstring = xstrdup (val);
898 #if defined(WINDOWS) || defined(MSDOS)
899 /* Convert "\" to "/". */
902 for (s = *pstring; *s; s++)
910 /* Like cmd_file, but strips trailing '/' characters. */
912 cmd_directory (const char *com, const char *val, void *place)
916 /* Call cmd_file() for tilde expansion and separator
917 canonicalization (backslash -> slash under Windows). These
918 things should perhaps be in a separate function. */
919 if (!cmd_file (com, val, place))
924 while (t > s && *--t == '/')
930 /* Split VAL by space to a vector of values, and append those values
931 to vector pointed to by the PLACE argument. If VAL is empty, the
932 PLACE vector is cleared instead. */
935 cmd_vector (const char *com, const char *val, void *place)
937 char ***pvec = (char ***)place;
940 *pvec = merge_vecs (*pvec, sepstring (val));
950 cmd_directory_vector (const char *com, const char *val, void *place)
952 char ***pvec = (char ***)place;
956 /* Strip the trailing slashes from directories. */
959 seps = sepstring (val);
960 for (t = seps; t && *t; t++)
962 int len = strlen (*t);
963 /* Skip degenerate case of root directory. */
966 if ((*t)[len - 1] == '/')
967 (*t)[len - 1] = '\0';
970 *pvec = merge_vecs (*pvec, seps);
980 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
981 "100k" or "2.5G" to a floating point number. */
984 parse_bytes_helper (const char *val, double *result)
987 const char *end = val + strlen (val);
989 /* Check for "inf". */
990 if (0 == strcmp (val, "inf"))
996 /* Strip trailing whitespace. */
997 while (val < end && c_isspace (end[-1]))
1002 switch (c_tolower (end[-1]))
1005 --end, mult = 1024.0;
1008 --end, mult = 1048576.0;
1011 --end, mult = 1073741824.0;
1014 --end, mult = 1099511627776.0;
1017 /* Not a recognized suffix: assume it's a digit. (If not,
1018 simple_atof will raise an error.) */
1022 /* Skip leading and trailing whitespace. */
1023 while (val < end && c_isspace (*val))
1025 while (val < end && c_isspace (end[-1]))
1030 if (!simple_atof (val, end, &number) || number < 0)
1033 *result = number * mult;
1037 /* Parse VAL as a number and set its value to PLACE (which should
1040 By default, the value is assumed to be in bytes. If "K", "M", or
1041 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1042 1<<30, respectively. Floating point values are allowed and are
1043 cast to integer before use. The idea is to be able to use things
1044 like 1.5k instead of "1536".
1046 The string "inf" is returned as 0.
1048 In case of error, false is returned and memory pointed to by PLACE
1049 remains unmodified. */
1052 cmd_bytes (const char *com, const char *val, void *place)
1055 if (!parse_bytes_helper (val, &byte_value))
1057 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1058 exec_name, com, quote (val));
1061 *(wgint *)place = (wgint)byte_value;
1065 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1066 SIZE_SUM. It works by converting the string to double, therefore
1067 working with values up to 2^53-1 without loss of precision. This
1068 value (8192 TB) is large enough to serve for a while. */
1071 cmd_bytes_sum (const char *com, const char *val, void *place)
1074 if (!parse_bytes_helper (val, &byte_value))
1076 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1077 exec_name, com, quote (val));
1080 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1084 /* Store the value of VAL to *OUT. The value is a time period, by
1085 default expressed in seconds, but also accepting suffixes "m", "h",
1086 "d", and "w" for minutes, hours, days, and weeks respectively. */
1089 cmd_time (const char *com, const char *val, void *place)
1091 double number, mult;
1092 const char *end = val + strlen (val);
1094 /* Strip trailing whitespace. */
1095 while (val < end && c_isspace (end[-1]))
1101 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1102 exec_name, com, quote (val));
1106 switch (c_tolower (end[-1]))
1109 --end, mult = 1; /* seconds */
1112 --end, mult = 60; /* minutes */
1115 --end, mult = 3600; /* hours */
1118 --end, mult = 86400.0; /* days */
1121 --end, mult = 604800.0; /* weeks */
1124 /* Not a recognized suffix: assume it belongs to the number.
1125 (If not, simple_atof will raise an error.) */
1129 /* Skip leading and trailing whitespace. */
1130 while (val < end && c_isspace (*val))
1132 while (val < end && c_isspace (end[-1]))
1137 if (!simple_atof (val, end, &number))
1140 *(double *)place = number * mult;
1146 cmd_cert_type (const char *com, const char *val, void *place)
1148 static const struct decode_item choices[] = {
1149 { "pem", keyfile_pem },
1150 { "der", keyfile_asn1 },
1151 { "asn1", keyfile_asn1 },
1153 int ok = decode_string (val, choices, countof (choices), place);
1155 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1160 /* Specialized helper functions, used by `commands' to handle some
1161 options specially. */
1163 static bool check_user_specified_header (const char *);
1166 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1168 if (!cmd_boolean (com, val, &opt.dirstruct))
1170 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1171 must be affected inversely. */
1173 opt.no_dirstruct = false;
1175 opt.no_dirstruct = true;
1180 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1182 /* Empty value means reset the list of headers. */
1185 free_vec (opt.user_headers);
1186 opt.user_headers = NULL;
1190 if (!check_user_specified_header (val))
1192 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1193 exec_name, com, quote (val));
1196 opt.user_headers = vec_append (opt.user_headers, val);
1201 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1203 int flag = cmd_boolean (com, val, &opt.htmlify);
1204 if (flag && !opt.htmlify)
1205 opt.remove_listing = false;
1209 /* Set the "mirror" mode. It means: recursive download, timestamping,
1210 no limit on max. recursion depth, and don't remove listings. */
1213 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1217 if (!cmd_boolean (com, val, &mirror))
1221 opt.recursive = true;
1222 if (!opt.no_dirstruct)
1223 opt.dirstruct = true;
1224 opt.timestamping = true;
1225 opt.reclevel = INFINITE_RECURSION;
1226 opt.remove_listing = false;
1231 /* Validate --prefer-family and set the choice. Allowed values are
1232 "IPv4", "IPv6", and "none". */
1235 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1237 static const struct decode_item choices[] = {
1238 { "IPv4", prefer_ipv4 },
1239 { "IPv6", prefer_ipv6 },
1240 { "none", prefer_none },
1242 int prefer_family = prefer_none;
1243 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1245 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1246 opt.prefer_family = prefer_family;
1250 /* Set progress.type to VAL, but verify that it's a valid progress
1251 implementation before that. */
1254 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1256 if (!valid_progress_implementation_p (val))
1258 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1259 exec_name, com, quote (val));
1262 xfree_null (opt.progress_type);
1264 /* Don't call set_progress_implementation here. It will be called
1265 in main() when it becomes clear what the log output is. */
1266 opt.progress_type = xstrdup (val);
1270 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1271 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1275 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1277 if (!cmd_boolean (com, val, &opt.recursive))
1281 if (opt.recursive && !opt.no_dirstruct)
1282 opt.dirstruct = true;
1288 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1290 int restrict_os = opt.restrict_files_os;
1291 int restrict_ctrl = opt.restrict_files_ctrl;
1292 int restrict_case = opt.restrict_files_case;
1293 int restrict_nonascii = opt.restrict_files_nonascii;
1297 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1301 end = strchr (val, ',');
1303 end = val + strlen (val);
1305 if (VAL_IS ("unix"))
1306 restrict_os = restrict_unix;
1307 else if (VAL_IS ("windows"))
1308 restrict_os = restrict_windows;
1309 else if (VAL_IS ("lowercase"))
1310 restrict_case = restrict_lowercase;
1311 else if (VAL_IS ("uppercase"))
1312 restrict_case = restrict_uppercase;
1313 else if (VAL_IS ("nocontrol"))
1314 restrict_ctrl = false;
1315 else if (VAL_IS ("ascii"))
1316 restrict_nonascii = true;
1319 fprintf (stderr, _("\
1320 %s: %s: Invalid restriction %s,\n\
1321 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1322 exec_name, com, quote (val));
1329 while (*val && *end);
1333 opt.restrict_files_os = restrict_os;
1334 opt.restrict_files_ctrl = restrict_ctrl;
1335 opt.restrict_files_case = restrict_case;
1336 opt.restrict_files_nonascii = restrict_nonascii;
1343 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1345 static const struct decode_item choices[] = {
1346 { "auto", secure_protocol_auto },
1347 { "sslv2", secure_protocol_sslv2 },
1348 { "sslv3", secure_protocol_sslv3 },
1349 { "tlsv1", secure_protocol_tlsv1 },
1351 int ok = decode_string (val, choices, countof (choices), place);
1353 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1358 /* Set all three timeout values. */
1361 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1364 if (!cmd_time (com, val, &value))
1366 opt.read_timeout = value;
1367 opt.connect_timeout = value;
1368 opt.dns_timeout = value;
1373 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1375 /* Disallow embedded newlines. */
1376 if (strchr (val, '\n'))
1378 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1379 exec_name, com, quote (val));
1382 xfree_null (opt.useragent);
1383 opt.useragent = xstrdup (val);
1387 /* The "verbose" option cannot be cmd_boolean because the variable is
1388 not bool -- it's of type int (-1 means uninitialized because of
1389 some random hackery for disallowing -q -v). */
1392 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1395 if (cmd_boolean (com, val, &flag))
1403 /* Miscellaneous useful routines. */
1405 /* A very simple atoi clone, more useful than atoi because it works on
1406 delimited strings, and has error reportage. Returns true on success,
1407 false on failure. If successful, stores result to *DEST. */
1410 simple_atoi (const char *beg, const char *end, int *dest)
1413 bool negative = false;
1414 const char *p = beg;
1416 while (p < end && c_isspace (*p))
1418 if (p < end && (*p == '-' || *p == '+'))
1420 negative = (*p == '-');
1426 /* Read negative numbers in a separate loop because the most
1427 negative integer cannot be represented as a positive number. */
1430 for (; p < end && c_isdigit (*p); p++)
1432 int next = (10 * result) + (*p - '0');
1434 return false; /* overflow */
1438 for (; p < end && c_isdigit (*p); p++)
1440 int next = (10 * result) - (*p - '0');
1442 return false; /* underflow */
1453 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1454 doesn't handle exponential notation. Returns true on success,
1455 false on failure. In case of success, stores its result to
1459 simple_atof (const char *beg, const char *end, double *dest)
1463 bool negative = false;
1464 bool seen_dot = false;
1465 bool seen_digit = false;
1468 const char *p = beg;
1470 while (p < end && c_isspace (*p))
1472 if (p < end && (*p == '-' || *p == '+'))
1474 negative = (*p == '-');
1478 for (; p < end; p++)
1484 result = (10 * result) + (ch - '0');
1486 result += (ch - '0') / (divider *= 10);
1508 /* Verify that the user-specified header in S is valid. It must
1509 contain a colon preceded by non-white-space characters and must not
1510 contain newlines. */
1513 check_user_specified_header (const char *s)
1517 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1519 /* The header MUST contain `:' preceded by at least one
1520 non-whitespace character. */
1521 if (*p != ':' || p == s)
1523 /* The header MUST NOT contain newlines. */
1524 if (strchr (s, '\n'))
1529 /* Decode VAL into a number, according to ITEMS. */
1532 decode_string (const char *val, const struct decode_item *items, int itemcount,
1536 for (i = 0; i < itemcount; i++)
1537 if (0 == strcasecmp (val, items[i].name))
1539 *place = items[i].code;
1546 void cleanup_html_url (void);
1549 /* Free the memory allocated by global variables. */
1553 /* Free external resources, close files, etc. */
1556 fclose (output_stream);
1557 /* No need to check for error because Wget flushes its output (and
1558 checks for errors) after any data arrives. */
1560 /* We're exiting anyway so there's no real need to call free()
1561 hundreds of times. Skipping the frees will make Wget exit
1564 However, when detecting leaks, it's crucial to free() everything
1565 because then you can find the real leaks, i.e. the allocated
1566 memory which grows with the size of the program. */
1572 cleanup_html_url ();
1577 extern acc_t *netrc_list;
1578 free_netrc (netrc_list);
1580 xfree_null (opt.choose_config);
1581 xfree_null (opt.lfilename);
1582 xfree_null (opt.dir_prefix);
1583 xfree_null (opt.input_filename);
1584 xfree_null (opt.output_document);
1585 free_vec (opt.accepts);
1586 free_vec (opt.rejects);
1587 free_vec (opt.excludes);
1588 free_vec (opt.includes);
1589 free_vec (opt.domains);
1590 free_vec (opt.follow_tags);
1591 free_vec (opt.ignore_tags);
1592 xfree_null (opt.progress_type);
1593 xfree_null (opt.ftp_user);
1594 xfree_null (opt.ftp_passwd);
1595 xfree_null (opt.ftp_proxy);
1596 xfree_null (opt.https_proxy);
1597 xfree_null (opt.http_proxy);
1598 free_vec (opt.no_proxy);
1599 xfree_null (opt.useragent);
1600 xfree_null (opt.referer);
1601 xfree_null (opt.http_user);
1602 xfree_null (opt.http_passwd);
1603 free_vec (opt.user_headers);
1605 xfree_null (opt.cert_file);
1606 xfree_null (opt.private_key);
1607 xfree_null (opt.ca_directory);
1608 xfree_null (opt.ca_cert);
1609 xfree_null (opt.random_file);
1610 xfree_null (opt.egd_file);
1612 xfree_null (opt.bind_address);
1613 xfree_null (opt.cookies_input);
1614 xfree_null (opt.cookies_output);
1615 xfree_null (opt.user);
1616 xfree_null (opt.passwd);
1617 xfree_null (opt.base_href);
1619 #endif /* DEBUG_MALLOC */
1622 /* Unit testing routines. */
1627 test_commands_sorted()
1629 int prev_idx = 0, next_idx = 1;
1630 int command_count = countof (commands) - 1;
1632 while (next_idx <= command_count)
1634 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1637 mu_assert ("FAILED", false);
1650 test_cmd_spec_restrict_file_names()
1655 int expected_restrict_files_os;
1656 int expected_restrict_files_ctrl;
1657 int expected_restrict_files_case;
1660 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1661 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1662 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1663 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1666 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1671 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1674 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1675 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1676 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1677 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1679 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1680 res == test_array[i].result
1681 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1682 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1683 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1689 #endif /* TESTING */