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. */
40 /* not all systems provide PATH_MAX in limits.h */
42 # include <sys/param.h>
44 # define PATH_MAX MAXPATHLEN
59 #include "recur.h" /* for INFINITE_RECURSION */
60 #include "convert.h" /* for convert_cleanup */
61 #include "res.h" /* for res_cleanup */
62 #include "http.h" /* for http_cleanup */
63 #include "retr.h" /* for output_stream */
71 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
73 CMD_DECLARE (cmd_boolean);
74 CMD_DECLARE (cmd_bytes);
75 CMD_DECLARE (cmd_bytes_sum);
77 CMD_DECLARE (cmd_cert_type);
79 CMD_DECLARE (cmd_directory_vector);
80 CMD_DECLARE (cmd_number);
81 CMD_DECLARE (cmd_number_inf);
82 CMD_DECLARE (cmd_string);
83 CMD_DECLARE (cmd_file);
84 CMD_DECLARE (cmd_directory);
85 CMD_DECLARE (cmd_time);
86 CMD_DECLARE (cmd_vector);
88 CMD_DECLARE (cmd_spec_dirstruct);
89 CMD_DECLARE (cmd_spec_header);
90 CMD_DECLARE (cmd_spec_htmlify);
91 CMD_DECLARE (cmd_spec_mirror);
92 CMD_DECLARE (cmd_spec_prefer_family);
93 CMD_DECLARE (cmd_spec_progress);
94 CMD_DECLARE (cmd_spec_recursive);
95 CMD_DECLARE (cmd_spec_restrict_file_names);
97 CMD_DECLARE (cmd_spec_secure_protocol);
99 CMD_DECLARE (cmd_spec_timeout);
100 CMD_DECLARE (cmd_spec_useragent);
101 CMD_DECLARE (cmd_spec_verbose);
103 /* List of recognized commands, each consisting of name, place and
104 function. When adding a new command, simply add it to the list,
105 but be sure to keep the list sorted alphabetically, as
106 command_by_name's binary search depends on it. Also, be sure to
107 add any entries that allocate memory (e.g. cmd_string and
108 cmd_vector) to the cleanup() function below. */
110 static const struct {
113 bool (*action) (const char *, const char *, void *);
115 /* KEEP THIS LIST ALPHABETICALLY SORTED */
116 { "accept", &opt.accepts, cmd_vector },
117 { "addhostdir", &opt.add_hostdir, cmd_boolean },
118 { "adjustextension", &opt.adjust_extension, cmd_boolean },
119 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
120 { "askpassword", &opt.ask_passwd, cmd_boolean },
121 { "authnochallenge", &opt.auth_without_challenge,
123 { "background", &opt.background, cmd_boolean },
124 { "backupconverted", &opt.backup_converted, cmd_boolean },
125 { "backups", &opt.backups, cmd_number },
126 { "base", &opt.base_href, cmd_string },
127 { "bindaddress", &opt.bind_address, cmd_string },
129 { "cacertificate", &opt.ca_cert, cmd_file },
131 { "cache", &opt.allow_cache, cmd_boolean },
133 { "cadirectory", &opt.ca_directory, cmd_directory },
134 { "certificate", &opt.cert_file, cmd_file },
135 { "certificatetype", &opt.cert_type, cmd_cert_type },
136 { "checkcertificate", &opt.check_cert, cmd_boolean },
138 { "chooseconfig", &opt.choose_config, cmd_file },
139 { "connecttimeout", &opt.connect_timeout, cmd_time },
140 { "contentdisposition", &opt.content_disposition, cmd_boolean },
141 { "continue", &opt.always_rest, cmd_boolean },
142 { "convertlinks", &opt.convert_links, cmd_boolean },
143 { "cookies", &opt.cookies, cmd_boolean },
144 { "cutdirs", &opt.cut_dirs, cmd_number },
146 { "debug", &opt.debug, cmd_boolean },
148 { "defaultpage", &opt.default_page, cmd_string},
149 { "deleteafter", &opt.delete_after, cmd_boolean },
150 { "dirprefix", &opt.dir_prefix, cmd_directory },
151 { "dirstruct", NULL, cmd_spec_dirstruct },
152 { "dnscache", &opt.dns_cache, cmd_boolean },
153 { "dnstimeout", &opt.dns_timeout, cmd_time },
154 { "domains", &opt.domains, cmd_vector },
155 { "dotbytes", &opt.dot_bytes, cmd_bytes },
156 { "dotsinline", &opt.dots_in_line, cmd_number },
157 { "dotspacing", &opt.dot_spacing, cmd_number },
158 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
160 { "egdfile", &opt.egd_file, cmd_file },
162 { "excludedirectories", &opt.excludes, cmd_directory_vector },
163 { "excludedomains", &opt.exclude_domains, cmd_vector },
164 { "followftp", &opt.follow_ftp, cmd_boolean },
165 { "followtags", &opt.follow_tags, cmd_vector },
166 { "forcehtml", &opt.force_html, cmd_boolean },
167 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
168 { "ftppassword", &opt.ftp_passwd, cmd_string },
169 { "ftpproxy", &opt.ftp_proxy, cmd_string },
171 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
172 #endif /* def __VMS */
173 { "ftpuser", &opt.ftp_user, cmd_string },
174 { "glob", &opt.ftp_glob, cmd_boolean },
175 { "header", NULL, cmd_spec_header },
176 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
177 { "htmlify", NULL, cmd_spec_htmlify },
178 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
179 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
180 { "httppassword", &opt.http_passwd, cmd_string },
181 { "httpproxy", &opt.http_proxy, cmd_string },
182 { "httpsproxy", &opt.https_proxy, cmd_string },
183 { "httpuser", &opt.http_user, cmd_string },
184 { "ignorecase", &opt.ignore_case, cmd_boolean },
185 { "ignorelength", &opt.ignore_length, cmd_boolean },
186 { "ignoretags", &opt.ignore_tags, cmd_vector },
187 { "includedirectories", &opt.includes, cmd_directory_vector },
189 { "inet4only", &opt.ipv4_only, cmd_boolean },
190 { "inet6only", &opt.ipv6_only, cmd_boolean },
192 { "input", &opt.input_filename, cmd_file },
193 { "iri", &opt.enable_iri, cmd_boolean },
194 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
195 { "limitrate", &opt.limit_rate, cmd_bytes },
196 { "loadcookies", &opt.cookies_input, cmd_file },
197 { "localencoding", &opt.locale, cmd_string },
198 { "logfile", &opt.lfilename, cmd_file },
199 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
200 { "maxredirect", &opt.max_redirect, cmd_number },
201 { "mirror", NULL, cmd_spec_mirror },
202 { "netrc", &opt.netrc, cmd_boolean },
203 { "noclobber", &opt.noclobber, cmd_boolean },
204 { "noparent", &opt.no_parent, cmd_boolean },
205 { "noproxy", &opt.no_proxy, cmd_vector },
206 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
207 { "outputdocument", &opt.output_document, cmd_file },
208 { "pagerequisites", &opt.page_requisites, cmd_boolean },
209 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
210 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
211 { "password", &opt.passwd, cmd_string },
212 { "postdata", &opt.post_data, cmd_string },
213 { "postfile", &opt.post_file_name, cmd_file },
214 { "preferfamily", NULL, cmd_spec_prefer_family },
215 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
217 { "privatekey", &opt.private_key, cmd_file },
218 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
220 { "progress", &opt.progress_type, cmd_spec_progress },
221 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
222 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
223 { "proxypassword", &opt.proxy_passwd, cmd_string },
224 { "proxyuser", &opt.proxy_user, cmd_string },
225 { "quiet", &opt.quiet, cmd_boolean },
226 { "quota", &opt.quota, cmd_bytes_sum },
228 { "randomfile", &opt.random_file, cmd_file },
230 { "randomwait", &opt.random_wait, cmd_boolean },
231 { "readtimeout", &opt.read_timeout, cmd_time },
232 { "reclevel", &opt.reclevel, cmd_number_inf },
233 { "recursive", NULL, cmd_spec_recursive },
234 { "referer", &opt.referer, cmd_string },
235 { "reject", &opt.rejects, cmd_vector },
236 { "relativeonly", &opt.relative_only, cmd_boolean },
237 { "remoteencoding", &opt.encoding_remote, cmd_string },
238 { "removelisting", &opt.remove_listing, cmd_boolean },
239 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
240 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
241 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
242 { "robots", &opt.use_robots, cmd_boolean },
243 { "savecookies", &opt.cookies_output, cmd_file },
244 { "saveheaders", &opt.save_headers, cmd_boolean },
246 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
248 { "serverresponse", &opt.server_response, cmd_boolean },
249 { "spanhosts", &opt.spanhost, cmd_boolean },
250 { "spider", &opt.spider, cmd_boolean },
251 { "strictcomments", &opt.strict_comments, cmd_boolean },
252 { "timeout", NULL, cmd_spec_timeout },
253 { "timestamping", &opt.timestamping, cmd_boolean },
254 { "tries", &opt.ntry, cmd_number_inf },
255 { "trustservernames", &opt.trustservernames, cmd_boolean },
256 { "unlink", &opt.unlink, cmd_boolean },
257 { "useproxy", &opt.use_proxy, cmd_boolean },
258 { "user", &opt.user, cmd_string },
259 { "useragent", NULL, cmd_spec_useragent },
260 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
261 { "verbose", NULL, cmd_spec_verbose },
262 { "wait", &opt.wait, cmd_time },
263 { "waitretry", &opt.waitretry, cmd_time },
265 { "wdebug", &opt.wdebug, cmd_boolean },
269 /* Look up CMDNAME in the commands[] and return its position in the
270 array. If CMDNAME is not found, return -1. */
273 command_by_name (const char *cmdname)
275 /* Use binary search for speed. Wget has ~100 commands, which
276 guarantees a worst case performance of 7 string comparisons. */
277 int lo = 0, hi = countof (commands) - 1;
281 int mid = (lo + hi) >> 1;
282 int cmp = strcasecmp (cmdname, commands[mid].name);
293 /* Reset the variables to default values. */
299 /* Most of the default values are 0 (and 0.0, NULL, and false).
300 Just reset everything, and fill in the non-zero values. Note
301 that initializing pointers to NULL this way is technically
302 illegal, but porting Wget to a machine where NULL is not all-zero
303 bit pattern will be the least of the implementors' worries. */
310 opt.add_hostdir = true;
314 opt.http_keep_alive = true;
315 opt.use_proxy = true;
316 tmp = getenv ("no_proxy");
318 opt.no_proxy = sepstring (tmp);
319 opt.prefer_family = prefer_none;
320 opt.allow_cache = true;
322 opt.read_timeout = 900;
323 opt.use_robots = true;
325 opt.remove_listing = true;
327 opt.dot_bytes = 1024;
328 opt.dot_spacing = 10;
329 opt.dots_in_line = 50;
331 opt.dns_cache = true;
335 opt.check_cert = true;
338 /* The default for file name restriction defaults to the OS type. */
339 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
340 opt.restrict_files_os = restrict_windows;
342 opt.restrict_files_os = restrict_unix;
344 opt.restrict_files_ctrl = true;
345 opt.restrict_files_nonascii = false;
346 opt.restrict_files_case = restrict_no_case_restriction;
348 opt.max_redirect = 20;
353 opt.enable_iri = true;
355 opt.enable_iri = false;
358 opt.encoding_remote = NULL;
360 opt.useservertimestamps = true;
363 /* Return the user's home directory (strdup-ed), or NULL if none is
368 static char buf[PATH_MAX];
373 home = getenv ("HOME");
377 /* Under MSDOS, if $HOME isn't defined, use the directory where
378 `wget.exe' resides. */
379 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
382 strcpy (buf, _w32_get_argv0 ());
383 p = strrchr (buf, '/'); /* djgpp */
385 p = strrchr (buf, '\\'); /* others */
389 #elif !defined(WINDOWS)
390 /* If HOME is not defined, try getting it from the password
392 struct passwd *pwd = getpwuid (getuid ());
393 if (!pwd || !pwd->pw_dir)
395 strcpy (buf, pwd->pw_dir);
398 /* Under Windows, if $HOME isn't defined, use the directory where
399 `wget.exe' resides. */
405 return home ? xstrdup (home) : NULL;
408 /* Check the 'WGETRC' environment variable and return the file name
409 if 'WGETRC' is set and is a valid file.
410 If the `WGETRC' variable exists but the file does not exist, the
411 function will exit(). */
413 wgetrc_env_file_name (void)
415 char *env = getenv ("WGETRC");
418 if (!file_exists_p (env))
420 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
424 return xstrdup (env);
429 /* Check for the existance of '$HOME/.wgetrc' and return it's path
430 if it exists and is set. */
432 wgetrc_user_file_name (void)
436 /* If that failed, try $HOME/.wgetrc (or equivalent). */
439 file = "SYS$LOGIN:.wgetrc";
440 #else /* def __VMS */
443 file = aprintf ("%s/.wgetrc", home);
445 #endif /* def __VMS [else] */
449 if (!file_exists_p (file))
457 /* Return the path to the user's .wgetrc. This is either the value of
458 `WGETRC' environment variable, or `$HOME/.wgetrc'.
460 Additionally, for windows, look in the directory where wget.exe
463 wgetrc_file_name (void)
465 char *file = wgetrc_env_file_name ();
469 file = wgetrc_user_file_name ();
472 /* Under Windows, if we still haven't found .wgetrc, look for the file
473 `wget.ini' in the directory where `wget.exe' resides; we do this for
474 backward compatibility with previous versions of Wget.
475 SYSTEM_WGETRC should not be defined under WINDOWS. */
478 char *home = home_dir ();
484 file = aprintf ("%s/wget.ini", home);
485 if (!file_exists_p (file))
498 /* Return values of parse_line. */
506 static enum parse_line parse_line (const char *, char **, char **, int *);
507 static bool setval_internal (int, const char *, const char *);
508 static bool setval_internal_tilde (int, const char *, const char *);
510 /* Initialize variables from a wgetrc file. Returns zero (failure) if
511 there were errors in the file. */
514 run_wgetrc (const char *file)
521 fp = fopen (file, "r");
524 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
525 file, strerror (errno));
526 return true; /* not a fatal error */
529 while ((line = read_whole_line (fp)) != NULL)
531 char *com = NULL, *val = NULL;
534 /* Parse the line. */
535 switch (parse_line (line, &com, &val, &comind))
538 /* If everything is OK, set the value. */
539 if (!setval_internal_tilde (comind, com, val))
541 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
542 exec_name, file, ln);
546 case line_syntax_error:
547 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
548 exec_name, file, ln);
551 case line_unknown_command:
552 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
553 exec_name, quote (com), file, ln);
571 /* Initialize the defaults and run the system wgetrc and user's own
576 char *file, *env_sysrc;
579 /* Run a non-standard system rc file when the according environment
580 variable has been set. For internal testing purposes only! */
581 env_sysrc = getenv ("SYSTEM_WGETRC");
582 if (env_sysrc && file_exists_p (env_sysrc))
583 ok &= run_wgetrc (env_sysrc);
584 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
586 else if (file_exists_p (SYSTEM_WGETRC))
587 ok &= run_wgetrc (SYSTEM_WGETRC);
589 /* If there are any problems parsing the system wgetrc file, tell
593 fprintf (stderr, _("\
594 Parsing system wgetrc file failed, please check '%s'. \
595 Or specify a different file using --config\n"), SYSTEM_WGETRC);
598 /* Override it with your own, if one exists. */
599 file = wgetrc_file_name ();
602 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
603 something like realpath() before comparing them with `strcmp' */
605 if (!strcmp (file, SYSTEM_WGETRC))
607 fprintf (stderr, _("\
608 %s: Warning: Both system and user wgetrc point to %s.\n"),
609 exec_name, quote (file));
613 ok &= run_wgetrc (file);
615 /* If there were errors processing either `.wgetrc', abort. */
623 /* Remove dashes and underscores from S, modifying S in the
629 char *t = s; /* t - tortoise */
630 char *h = s; /* h - hare */
632 if (*h == '_' || *h == '-')
639 /* Parse the line pointed by line, with the syntax:
640 <sp>* command <sp>* = <sp>* value <sp>*
641 Uses malloc to allocate space for command and value.
643 Returns one of line_ok, line_empty, line_syntax_error, or
644 line_unknown_command.
646 In case of line_ok, *COM and *VAL point to freshly allocated
647 strings, and *COMIND points to com's index. In case of error or
648 empty line, their values are unmodified. */
650 static enum parse_line
651 parse_line (const char *line, char **com, char **val, int *comind)
654 const char *end = line + strlen (line);
655 const char *cmdstart, *cmdend;
656 const char *valstart, *valend;
661 /* Skip leading and trailing whitespace. */
662 while (*line && c_isspace (*line))
664 while (end > line && c_isspace (end[-1]))
667 /* Skip empty lines and comments. */
668 if (!*line || *line == '#')
674 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
678 /* Skip '=', as well as any space before or after it. */
679 while (p < end && c_isspace (*p))
681 if (p == end || *p != '=')
682 return line_syntax_error;
684 while (p < end && c_isspace (*p))
690 /* The syntax is valid (even though the command might not be). Fill
691 in the command name and value. */
692 *com = strdupdelim (cmdstart, cmdend);
693 *val = strdupdelim (valstart, valend);
695 /* The line now known to be syntactically correct. Check whether
696 the command is valid. */
697 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
699 ind = command_by_name (cmdcopy);
701 return line_unknown_command;
703 /* Report success to the caller. */
708 #if defined(WINDOWS) || defined(MSDOS)
709 # define ISSEP(c) ((c) == '/' || (c) == '\\')
711 # define ISSEP(c) ((c) == '/')
714 /* Run commands[comind].action. */
717 setval_internal (int comind, const char *com, const char *val)
719 assert (0 <= comind && ((size_t) comind) < countof (commands));
720 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
721 return commands[comind].action (com, val, commands[comind].place);
725 setval_internal_tilde (int comind, const char *com, const char *val)
731 ret = setval_internal (comind, com, val);
733 /* We make tilde expansion for cmd_file and cmd_directory */
734 if (((commands[comind].action == cmd_file) ||
735 (commands[comind].action == cmd_directory))
736 && ret && (*val == '~' && ISSEP (val[1])))
738 pstring = commands[comind].place;
742 homelen = strlen (home);
743 while (homelen && ISSEP (home[homelen - 1]))
744 home[--homelen] = '\0';
746 /* Skip the leading "~/". */
747 for (++val; ISSEP (*val); val++)
749 *pstring = concat_strings (home, "/", val, (char *)0);
755 /* Run command COM with value VAL. If running the command produces an
756 error, report the error and exit.
758 This is intended to be called from main() to modify Wget's behavior
759 through command-line switches. Since COM is hard-coded in main(),
760 it is not canonicalized, and this aborts when COM is not found.
762 If COMIND's are exported to init.h, this function will be changed
763 to accept COMIND directly. */
766 setoptval (const char *com, const char *val, const char *optname)
768 /* Prepend "--" to OPTNAME. */
769 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
772 strcpy (dd_optname + 2, optname);
774 assert (val != NULL);
775 if (!setval_internal (command_by_name (com), dd_optname, val))
779 /* Parse OPT into command and value and run it. For example,
780 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
781 This is used by the `--execute' flag in main.c. */
784 run_command (const char *opt)
788 switch (parse_line (opt, &com, &val, &comind))
791 if (!setval_internal (comind, com, val))
797 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
798 exec_name, quote (opt));
803 /* Generic helper functions, for use with `commands'. */
805 /* Forward declarations: */
810 static bool decode_string (const char *, const struct decode_item *, int, int *);
811 static bool simple_atoi (const char *, const char *, int *);
812 static bool simple_atof (const char *, const char *, double *);
814 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
816 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
817 && c_tolower((p)[1]) == (c1) \
820 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
821 && c_tolower((p)[1]) == (c1) \
822 && c_tolower((p)[2]) == (c2) \
826 /* Store the boolean value from VAL to PLACE. COM is ignored,
827 except for error messages. */
829 cmd_boolean (const char *com, const char *val, void *place)
833 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
834 /* "on", "yes" and "1" mean true. */
836 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
837 /* "off", "no" and "0" mean false. */
842 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
843 exec_name, com, quote (val));
847 *(bool *) place = value;
851 /* Set the non-negative integer value from VAL to PLACE. With
852 incorrect specification, the number remains unchanged. */
854 cmd_number (const char *com, const char *val, void *place)
856 if (!simple_atoi (val, val + strlen (val), place)
857 || *(int *) place < 0)
859 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
860 exec_name, com, quote (val));
866 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
868 cmd_number_inf (const char *com, const char *val, void *place)
870 if (!strcasecmp (val, "inf"))
875 return cmd_number (com, val, place);
878 /* Copy (strdup) the string at COM to a new location and place a
879 pointer to *PLACE. */
881 cmd_string (const char *com, const char *val, void *place)
883 char **pstring = (char **)place;
885 xfree_null (*pstring);
886 *pstring = xstrdup (val);
891 /* Like the above, but handles tilde-expansion when reading a user's
892 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
893 gets expanded to the user's home directory. */
895 cmd_file (const char *com, const char *val, void *place)
897 char **pstring = (char **)place;
899 xfree_null (*pstring);
901 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
903 *pstring = xstrdup (val);
905 #if defined(WINDOWS) || defined(MSDOS)
906 /* Convert "\" to "/". */
909 for (s = *pstring; *s; s++)
917 /* Like cmd_file, but strips trailing '/' characters. */
919 cmd_directory (const char *com, const char *val, void *place)
923 /* Call cmd_file() for tilde expansion and separator
924 canonicalization (backslash -> slash under Windows). These
925 things should perhaps be in a separate function. */
926 if (!cmd_file (com, val, place))
931 while (t > s && *--t == '/')
937 /* Split VAL by space to a vector of values, and append those values
938 to vector pointed to by the PLACE argument. If VAL is empty, the
939 PLACE vector is cleared instead. */
942 cmd_vector (const char *com, const char *val, void *place)
944 char ***pvec = (char ***)place;
947 *pvec = merge_vecs (*pvec, sepstring (val));
957 cmd_directory_vector (const char *com, const char *val, void *place)
959 char ***pvec = (char ***)place;
963 /* Strip the trailing slashes from directories. */
966 seps = sepstring (val);
967 for (t = seps; t && *t; t++)
969 int len = strlen (*t);
970 /* Skip degenerate case of root directory. */
973 if ((*t)[len - 1] == '/')
974 (*t)[len - 1] = '\0';
977 *pvec = merge_vecs (*pvec, seps);
987 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
988 "100k" or "2.5G" to a floating point number. */
991 parse_bytes_helper (const char *val, double *result)
994 const char *end = val + strlen (val);
996 /* Check for "inf". */
997 if (0 == strcmp (val, "inf"))
1003 /* Strip trailing whitespace. */
1004 while (val < end && c_isspace (end[-1]))
1009 switch (c_tolower (end[-1]))
1012 --end, mult = 1024.0;
1015 --end, mult = 1048576.0;
1018 --end, mult = 1073741824.0;
1021 --end, mult = 1099511627776.0;
1024 /* Not a recognized suffix: assume it's a digit. (If not,
1025 simple_atof will raise an error.) */
1029 /* Skip leading and trailing whitespace. */
1030 while (val < end && c_isspace (*val))
1032 while (val < end && c_isspace (end[-1]))
1037 if (!simple_atof (val, end, &number) || number < 0)
1040 *result = number * mult;
1044 /* Parse VAL as a number and set its value to PLACE (which should
1047 By default, the value is assumed to be in bytes. If "K", "M", or
1048 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1049 1<<30, respectively. Floating point values are allowed and are
1050 cast to integer before use. The idea is to be able to use things
1051 like 1.5k instead of "1536".
1053 The string "inf" is returned as 0.
1055 In case of error, false is returned and memory pointed to by PLACE
1056 remains unmodified. */
1059 cmd_bytes (const char *com, const char *val, void *place)
1062 if (!parse_bytes_helper (val, &byte_value))
1064 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1065 exec_name, com, quote (val));
1068 *(wgint *)place = (wgint)byte_value;
1072 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1073 SIZE_SUM. It works by converting the string to double, therefore
1074 working with values up to 2^53-1 without loss of precision. This
1075 value (8192 TB) is large enough to serve for a while. */
1078 cmd_bytes_sum (const char *com, const char *val, void *place)
1081 if (!parse_bytes_helper (val, &byte_value))
1083 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1084 exec_name, com, quote (val));
1087 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1091 /* Store the value of VAL to *OUT. The value is a time period, by
1092 default expressed in seconds, but also accepting suffixes "m", "h",
1093 "d", and "w" for minutes, hours, days, and weeks respectively. */
1096 cmd_time (const char *com, const char *val, void *place)
1098 double number, mult;
1099 const char *end = val + strlen (val);
1101 /* Strip trailing whitespace. */
1102 while (val < end && c_isspace (end[-1]))
1108 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1109 exec_name, com, quote (val));
1113 switch (c_tolower (end[-1]))
1116 --end, mult = 1; /* seconds */
1119 --end, mult = 60; /* minutes */
1122 --end, mult = 3600; /* hours */
1125 --end, mult = 86400.0; /* days */
1128 --end, mult = 604800.0; /* weeks */
1131 /* Not a recognized suffix: assume it belongs to the number.
1132 (If not, simple_atof will raise an error.) */
1136 /* Skip leading and trailing whitespace. */
1137 while (val < end && c_isspace (*val))
1139 while (val < end && c_isspace (end[-1]))
1144 if (!simple_atof (val, end, &number))
1147 *(double *)place = number * mult;
1153 cmd_cert_type (const char *com, const char *val, void *place)
1155 static const struct decode_item choices[] = {
1156 { "pem", keyfile_pem },
1157 { "der", keyfile_asn1 },
1158 { "asn1", keyfile_asn1 },
1160 int ok = decode_string (val, choices, countof (choices), place);
1162 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1167 /* Specialized helper functions, used by `commands' to handle some
1168 options specially. */
1170 static bool check_user_specified_header (const char *);
1173 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1175 if (!cmd_boolean (com, val, &opt.dirstruct))
1177 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1178 must be affected inversely. */
1180 opt.no_dirstruct = false;
1182 opt.no_dirstruct = true;
1187 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1189 /* Empty value means reset the list of headers. */
1192 free_vec (opt.user_headers);
1193 opt.user_headers = NULL;
1197 if (!check_user_specified_header (val))
1199 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1200 exec_name, com, quote (val));
1203 opt.user_headers = vec_append (opt.user_headers, val);
1208 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1210 int flag = cmd_boolean (com, val, &opt.htmlify);
1211 if (flag && !opt.htmlify)
1212 opt.remove_listing = false;
1216 /* Set the "mirror" mode. It means: recursive download, timestamping,
1217 no limit on max. recursion depth, and don't remove listings. */
1220 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1224 if (!cmd_boolean (com, val, &mirror))
1228 opt.recursive = true;
1229 if (!opt.no_dirstruct)
1230 opt.dirstruct = true;
1231 opt.timestamping = true;
1232 opt.reclevel = INFINITE_RECURSION;
1233 opt.remove_listing = false;
1238 /* Validate --prefer-family and set the choice. Allowed values are
1239 "IPv4", "IPv6", and "none". */
1242 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1244 static const struct decode_item choices[] = {
1245 { "IPv4", prefer_ipv4 },
1246 { "IPv6", prefer_ipv6 },
1247 { "none", prefer_none },
1249 int prefer_family = prefer_none;
1250 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1252 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1253 opt.prefer_family = prefer_family;
1257 /* Set progress.type to VAL, but verify that it's a valid progress
1258 implementation before that. */
1261 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1263 if (!valid_progress_implementation_p (val))
1265 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1266 exec_name, com, quote (val));
1269 xfree_null (opt.progress_type);
1271 /* Don't call set_progress_implementation here. It will be called
1272 in main() when it becomes clear what the log output is. */
1273 opt.progress_type = xstrdup (val);
1277 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1278 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1282 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1284 if (!cmd_boolean (com, val, &opt.recursive))
1288 if (opt.recursive && !opt.no_dirstruct)
1289 opt.dirstruct = true;
1295 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1297 int restrict_os = opt.restrict_files_os;
1298 int restrict_ctrl = opt.restrict_files_ctrl;
1299 int restrict_case = opt.restrict_files_case;
1300 int restrict_nonascii = opt.restrict_files_nonascii;
1304 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1308 end = strchr (val, ',');
1310 end = val + strlen (val);
1312 if (VAL_IS ("unix"))
1313 restrict_os = restrict_unix;
1314 else if (VAL_IS ("windows"))
1315 restrict_os = restrict_windows;
1316 else if (VAL_IS ("lowercase"))
1317 restrict_case = restrict_lowercase;
1318 else if (VAL_IS ("uppercase"))
1319 restrict_case = restrict_uppercase;
1320 else if (VAL_IS ("nocontrol"))
1321 restrict_ctrl = false;
1322 else if (VAL_IS ("ascii"))
1323 restrict_nonascii = true;
1326 fprintf (stderr, _("\
1327 %s: %s: Invalid restriction %s,\n\
1328 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1329 exec_name, com, quote (val));
1336 while (*val && *end);
1340 opt.restrict_files_os = restrict_os;
1341 opt.restrict_files_ctrl = restrict_ctrl;
1342 opt.restrict_files_case = restrict_case;
1343 opt.restrict_files_nonascii = restrict_nonascii;
1350 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1352 static const struct decode_item choices[] = {
1353 { "auto", secure_protocol_auto },
1354 { "sslv2", secure_protocol_sslv2 },
1355 { "sslv3", secure_protocol_sslv3 },
1356 { "tlsv1", secure_protocol_tlsv1 },
1358 int ok = decode_string (val, choices, countof (choices), place);
1360 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1365 /* Set all three timeout values. */
1368 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1371 if (!cmd_time (com, val, &value))
1373 opt.read_timeout = value;
1374 opt.connect_timeout = value;
1375 opt.dns_timeout = value;
1380 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1382 /* Disallow embedded newlines. */
1383 if (strchr (val, '\n'))
1385 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1386 exec_name, com, quote (val));
1389 xfree_null (opt.useragent);
1390 opt.useragent = xstrdup (val);
1394 /* The "verbose" option cannot be cmd_boolean because the variable is
1395 not bool -- it's of type int (-1 means uninitialized because of
1396 some random hackery for disallowing -q -v). */
1399 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1402 if (cmd_boolean (com, val, &flag))
1410 /* Miscellaneous useful routines. */
1412 /* A very simple atoi clone, more useful than atoi because it works on
1413 delimited strings, and has error reportage. Returns true on success,
1414 false on failure. If successful, stores result to *DEST. */
1417 simple_atoi (const char *beg, const char *end, int *dest)
1420 bool negative = false;
1421 const char *p = beg;
1423 while (p < end && c_isspace (*p))
1425 if (p < end && (*p == '-' || *p == '+'))
1427 negative = (*p == '-');
1433 /* Read negative numbers in a separate loop because the most
1434 negative integer cannot be represented as a positive number. */
1437 for (; p < end && c_isdigit (*p); p++)
1439 int next = (10 * result) + (*p - '0');
1441 return false; /* overflow */
1445 for (; p < end && c_isdigit (*p); p++)
1447 int next = (10 * result) - (*p - '0');
1449 return false; /* underflow */
1460 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1461 doesn't handle exponential notation. Returns true on success,
1462 false on failure. In case of success, stores its result to
1466 simple_atof (const char *beg, const char *end, double *dest)
1470 bool negative = false;
1471 bool seen_dot = false;
1472 bool seen_digit = false;
1475 const char *p = beg;
1477 while (p < end && c_isspace (*p))
1479 if (p < end && (*p == '-' || *p == '+'))
1481 negative = (*p == '-');
1485 for (; p < end; p++)
1491 result = (10 * result) + (ch - '0');
1493 result += (ch - '0') / (divider *= 10);
1515 /* Verify that the user-specified header in S is valid. It must
1516 contain a colon preceded by non-white-space characters and must not
1517 contain newlines. */
1520 check_user_specified_header (const char *s)
1524 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1526 /* The header MUST contain `:' preceded by at least one
1527 non-whitespace character. */
1528 if (*p != ':' || p == s)
1530 /* The header MUST NOT contain newlines. */
1531 if (strchr (s, '\n'))
1536 /* Decode VAL into a number, according to ITEMS. */
1539 decode_string (const char *val, const struct decode_item *items, int itemcount,
1543 for (i = 0; i < itemcount; i++)
1544 if (0 == strcasecmp (val, items[i].name))
1546 *place = items[i].code;
1553 void cleanup_html_url (void);
1556 /* Free the memory allocated by global variables. */
1560 /* Free external resources, close files, etc. */
1563 fclose (output_stream);
1564 /* No need to check for error because Wget flushes its output (and
1565 checks for errors) after any data arrives. */
1567 /* We're exiting anyway so there's no real need to call free()
1568 hundreds of times. Skipping the frees will make Wget exit
1571 However, when detecting leaks, it's crucial to free() everything
1572 because then you can find the real leaks, i.e. the allocated
1573 memory which grows with the size of the program. */
1579 cleanup_html_url ();
1584 extern acc_t *netrc_list;
1585 free_netrc (netrc_list);
1587 xfree_null (opt.choose_config);
1588 xfree_null (opt.lfilename);
1589 xfree_null (opt.dir_prefix);
1590 xfree_null (opt.input_filename);
1591 xfree_null (opt.output_document);
1592 free_vec (opt.accepts);
1593 free_vec (opt.rejects);
1594 free_vec (opt.excludes);
1595 free_vec (opt.includes);
1596 free_vec (opt.domains);
1597 free_vec (opt.follow_tags);
1598 free_vec (opt.ignore_tags);
1599 xfree_null (opt.progress_type);
1600 xfree_null (opt.ftp_user);
1601 xfree_null (opt.ftp_passwd);
1602 xfree_null (opt.ftp_proxy);
1603 xfree_null (opt.https_proxy);
1604 xfree_null (opt.http_proxy);
1605 free_vec (opt.no_proxy);
1606 xfree_null (opt.useragent);
1607 xfree_null (opt.referer);
1608 xfree_null (opt.http_user);
1609 xfree_null (opt.http_passwd);
1610 free_vec (opt.user_headers);
1612 xfree_null (opt.cert_file);
1613 xfree_null (opt.private_key);
1614 xfree_null (opt.ca_directory);
1615 xfree_null (opt.ca_cert);
1616 xfree_null (opt.random_file);
1617 xfree_null (opt.egd_file);
1619 xfree_null (opt.bind_address);
1620 xfree_null (opt.cookies_input);
1621 xfree_null (opt.cookies_output);
1622 xfree_null (opt.user);
1623 xfree_null (opt.passwd);
1624 xfree_null (opt.base_href);
1626 #endif /* DEBUG_MALLOC */
1629 /* Unit testing routines. */
1634 test_commands_sorted()
1636 int prev_idx = 0, next_idx = 1;
1637 int command_count = countof (commands) - 1;
1639 while (next_idx <= command_count)
1641 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1644 mu_assert ("FAILED", false);
1657 test_cmd_spec_restrict_file_names()
1662 int expected_restrict_files_os;
1663 int expected_restrict_files_ctrl;
1664 int expected_restrict_files_case;
1667 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1668 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1669 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1670 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1673 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1678 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1681 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1682 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1683 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1684 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1686 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1687 res == test_array[i].result
1688 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1689 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1690 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1696 #endif /* TESTING */