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 { "contentonerror", &opt.content_on_error, 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 },
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 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
252 { "spanhosts", &opt.spanhost, cmd_boolean },
253 { "spider", &opt.spider, cmd_boolean },
254 { "strictcomments", &opt.strict_comments, cmd_boolean },
255 { "timeout", NULL, cmd_spec_timeout },
256 { "timestamping", &opt.timestamping, cmd_boolean },
257 { "tries", &opt.ntry, cmd_number_inf },
258 { "trustservernames", &opt.trustservernames, cmd_boolean },
259 { "unlink", &opt.unlink, cmd_boolean },
260 { "useproxy", &opt.use_proxy, cmd_boolean },
261 { "user", &opt.user, cmd_string },
262 { "useragent", NULL, cmd_spec_useragent },
263 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
264 { "verbose", NULL, cmd_spec_verbose },
265 { "wait", &opt.wait, cmd_time },
266 { "waitretry", &opt.waitretry, cmd_time },
268 { "wdebug", &opt.wdebug, cmd_boolean },
272 /* Look up CMDNAME in the commands[] and return its position in the
273 array. If CMDNAME is not found, return -1. */
276 command_by_name (const char *cmdname)
278 /* Use binary search for speed. Wget has ~100 commands, which
279 guarantees a worst case performance of 7 string comparisons. */
280 int lo = 0, hi = countof (commands) - 1;
284 int mid = (lo + hi) >> 1;
285 int cmp = strcasecmp (cmdname, commands[mid].name);
296 /* Reset the variables to default values. */
302 /* Most of the default values are 0 (and 0.0, NULL, and false).
303 Just reset everything, and fill in the non-zero values. Note
304 that initializing pointers to NULL this way is technically
305 illegal, but porting Wget to a machine where NULL is not all-zero
306 bit pattern will be the least of the implementors' worries. */
313 opt.add_hostdir = true;
317 opt.http_keep_alive = true;
318 opt.use_proxy = true;
319 tmp = getenv ("no_proxy");
321 opt.no_proxy = sepstring (tmp);
322 opt.prefer_family = prefer_none;
323 opt.allow_cache = true;
325 opt.read_timeout = 900;
326 opt.use_robots = true;
328 opt.remove_listing = true;
330 opt.dot_bytes = 1024;
331 opt.dot_spacing = 10;
332 opt.dots_in_line = 50;
334 opt.dns_cache = true;
338 opt.check_cert = true;
341 /* The default for file name restriction defaults to the OS type. */
342 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
343 opt.restrict_files_os = restrict_windows;
345 opt.restrict_files_os = restrict_unix;
347 opt.restrict_files_ctrl = true;
348 opt.restrict_files_nonascii = false;
349 opt.restrict_files_case = restrict_no_case_restriction;
351 opt.max_redirect = 20;
356 opt.enable_iri = true;
358 opt.enable_iri = false;
361 opt.encoding_remote = NULL;
363 opt.useservertimestamps = true;
364 opt.show_all_dns_entries = false;
367 /* Return the user's home directory (strdup-ed), or NULL if none is
372 static char *buf = NULL;
373 static char *home, *ret;
377 home = getenv ("HOME");
383 /* Under MSDOS, if $HOME isn't defined, use the directory where
384 `wget.exe' resides. */
385 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
388 buff = _w32_get_argv0 ();
390 p = strrchr (buf, '/'); /* djgpp */
392 p = strrchr (buf, '\\'); /* others */
396 buff = malloc (len + 1);
400 strncpy (buff, _w32_get_argv0 (), len);
404 #elif !defined(WINDOWS)
405 /* If HOME is not defined, try getting it from the password
407 struct passwd *pwd = getpwuid (getuid ());
408 if (!pwd || !pwd->pw_dir)
412 /* Under Windows, if $HOME isn't defined, use the directory where
413 `wget.exe' resides. */
419 ret = home ? xstrdup (home) : NULL;
426 /* Check the 'WGETRC' environment variable and return the file name
427 if 'WGETRC' is set and is a valid file.
428 If the `WGETRC' variable exists but the file does not exist, the
429 function will exit(). */
431 wgetrc_env_file_name (void)
433 char *env = getenv ("WGETRC");
436 if (!file_exists_p (env))
438 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
442 return xstrdup (env);
447 /* Check for the existance of '$HOME/.wgetrc' and return it's path
448 if it exists and is set. */
450 wgetrc_user_file_name (void)
454 /* If that failed, try $HOME/.wgetrc (or equivalent). */
457 file = "SYS$LOGIN:.wgetrc";
458 #else /* def __VMS */
461 file = aprintf ("%s/.wgetrc", home);
463 #endif /* def __VMS [else] */
467 if (!file_exists_p (file))
475 /* Return the path to the user's .wgetrc. This is either the value of
476 `WGETRC' environment variable, or `$HOME/.wgetrc'.
478 Additionally, for windows, look in the directory where wget.exe
481 wgetrc_file_name (void)
483 char *file = wgetrc_env_file_name ();
487 file = wgetrc_user_file_name ();
490 /* Under Windows, if we still haven't found .wgetrc, look for the file
491 `wget.ini' in the directory where `wget.exe' resides; we do this for
492 backward compatibility with previous versions of Wget.
493 SYSTEM_WGETRC should not be defined under WINDOWS. */
496 char *home = home_dir ();
502 file = aprintf ("%s/wget.ini", home);
503 if (!file_exists_p (file))
516 /* Return values of parse_line. */
524 static enum parse_line parse_line (const char *, char **, char **, int *);
525 static bool setval_internal (int, const char *, const char *);
526 static bool setval_internal_tilde (int, const char *, const char *);
528 /* Initialize variables from a wgetrc file. Returns zero (failure) if
529 there were errors in the file. */
532 run_wgetrc (const char *file)
539 fp = fopen (file, "r");
542 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
543 file, strerror (errno));
544 return true; /* not a fatal error */
547 while ((line = read_whole_line (fp)) != NULL)
549 char *com = NULL, *val = NULL;
552 /* Parse the line. */
553 switch (parse_line (line, &com, &val, &comind))
556 /* If everything is OK, set the value. */
557 if (!setval_internal_tilde (comind, com, val))
559 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
560 exec_name, file, ln);
564 case line_syntax_error:
565 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
566 exec_name, file, ln);
569 case line_unknown_command:
570 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
571 exec_name, quote (com), file, ln);
589 /* Initialize the defaults and run the system wgetrc and user's own
594 char *file, *env_sysrc;
597 /* Run a non-standard system rc file when the according environment
598 variable has been set. For internal testing purposes only! */
599 env_sysrc = getenv ("SYSTEM_WGETRC");
600 if (env_sysrc && file_exists_p (env_sysrc))
602 ok &= run_wgetrc (env_sysrc);
603 /* If there are any problems parsing the system wgetrc file, tell
607 fprintf (stderr, _("\
608 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
610 or specify a different file using --config.\n"), env_sysrc);
614 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
616 else if (file_exists_p (SYSTEM_WGETRC))
617 ok &= run_wgetrc (SYSTEM_WGETRC);
618 /* If there are any problems parsing the system wgetrc file, tell
622 fprintf (stderr, _("\
623 Parsing system wgetrc file failed. Please check\n\
625 or specify a different file using --config.\n"), SYSTEM_WGETRC);
629 /* Override it with your own, if one exists. */
630 file = wgetrc_file_name ();
633 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
634 something like realpath() before comparing them with `strcmp' */
636 if (!strcmp (file, SYSTEM_WGETRC))
638 fprintf (stderr, _("\
639 %s: Warning: Both system and user wgetrc point to %s.\n"),
640 exec_name, quote (file));
644 ok &= run_wgetrc (file);
646 /* If there were errors processing either `.wgetrc', abort. */
654 /* Remove dashes and underscores from S, modifying S in the
660 char *t = s; /* t - tortoise */
661 char *h = s; /* h - hare */
663 if (*h == '_' || *h == '-')
670 /* Parse the line pointed by line, with the syntax:
671 <sp>* command <sp>* = <sp>* value <sp>*
672 Uses malloc to allocate space for command and value.
674 Returns one of line_ok, line_empty, line_syntax_error, or
675 line_unknown_command.
677 In case of line_ok, *COM and *VAL point to freshly allocated
678 strings, and *COMIND points to com's index. In case of error or
679 empty line, their values are unmodified. */
681 static enum parse_line
682 parse_line (const char *line, char **com, char **val, int *comind)
685 const char *end = line + strlen (line);
686 const char *cmdstart, *cmdend;
687 const char *valstart, *valend;
692 /* Skip leading and trailing whitespace. */
693 while (*line && c_isspace (*line))
695 while (end > line && c_isspace (end[-1]))
698 /* Skip empty lines and comments. */
699 if (!*line || *line == '#')
705 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
709 /* Skip '=', as well as any space before or after it. */
710 while (p < end && c_isspace (*p))
712 if (p == end || *p != '=')
713 return line_syntax_error;
715 while (p < end && c_isspace (*p))
721 /* The syntax is valid (even though the command might not be). Fill
722 in the command name and value. */
723 *com = strdupdelim (cmdstart, cmdend);
724 *val = strdupdelim (valstart, valend);
726 /* The line now known to be syntactically correct. Check whether
727 the command is valid. */
728 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
730 ind = command_by_name (cmdcopy);
732 return line_unknown_command;
734 /* Report success to the caller. */
739 #if defined(WINDOWS) || defined(MSDOS)
740 # define ISSEP(c) ((c) == '/' || (c) == '\\')
742 # define ISSEP(c) ((c) == '/')
745 /* Run commands[comind].action. */
748 setval_internal (int comind, const char *com, const char *val)
750 assert (0 <= comind && ((size_t) comind) < countof (commands));
751 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
752 return commands[comind].action (com, val, commands[comind].place);
756 setval_internal_tilde (int comind, const char *com, const char *val)
762 ret = setval_internal (comind, com, val);
764 /* We make tilde expansion for cmd_file and cmd_directory */
765 if (((commands[comind].action == cmd_file) ||
766 (commands[comind].action == cmd_directory))
767 && ret && (*val == '~' && ISSEP (val[1])))
769 pstring = commands[comind].place;
773 homelen = strlen (home);
774 while (homelen && ISSEP (home[homelen - 1]))
775 home[--homelen] = '\0';
777 /* Skip the leading "~/". */
778 for (++val; ISSEP (*val); val++)
780 *pstring = concat_strings (home, "/", val, (char *)0);
786 /* Run command COM with value VAL. If running the command produces an
787 error, report the error and exit.
789 This is intended to be called from main() to modify Wget's behavior
790 through command-line switches. Since COM is hard-coded in main(),
791 it is not canonicalized, and this aborts when COM is not found.
793 If COMIND's are exported to init.h, this function will be changed
794 to accept COMIND directly. */
797 setoptval (const char *com, const char *val, const char *optname)
799 /* Prepend "--" to OPTNAME. */
800 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
803 strcpy (dd_optname + 2, optname);
805 assert (val != NULL);
806 if (!setval_internal (command_by_name (com), dd_optname, val))
810 /* Parse OPT into command and value and run it. For example,
811 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
812 This is used by the `--execute' flag in main.c. */
815 run_command (const char *opt)
819 switch (parse_line (opt, &com, &val, &comind))
822 if (!setval_internal (comind, com, val))
828 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
829 exec_name, quote (opt));
834 /* Generic helper functions, for use with `commands'. */
836 /* Forward declarations: */
841 static bool decode_string (const char *, const struct decode_item *, int, int *);
842 static bool simple_atoi (const char *, const char *, int *);
843 static bool simple_atof (const char *, const char *, double *);
845 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
847 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
848 && c_tolower((p)[1]) == (c1) \
851 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
852 && c_tolower((p)[1]) == (c1) \
853 && c_tolower((p)[2]) == (c2) \
857 /* Store the boolean value from VAL to PLACE. COM is ignored,
858 except for error messages. */
860 cmd_boolean (const char *com, const char *val, void *place)
864 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
865 /* "on", "yes" and "1" mean true. */
867 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
868 /* "off", "no" and "0" mean false. */
873 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
874 exec_name, com, quote (val));
878 *(bool *) place = value;
882 /* Set the non-negative integer value from VAL to PLACE. With
883 incorrect specification, the number remains unchanged. */
885 cmd_number (const char *com, const char *val, void *place)
887 if (!simple_atoi (val, val + strlen (val), place)
888 || *(int *) place < 0)
890 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
891 exec_name, com, quote (val));
897 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
899 cmd_number_inf (const char *com, const char *val, void *place)
901 if (!strcasecmp (val, "inf"))
906 return cmd_number (com, val, place);
909 /* Copy (strdup) the string at COM to a new location and place a
910 pointer to *PLACE. */
912 cmd_string (const char *com, const char *val, void *place)
914 char **pstring = (char **)place;
916 xfree_null (*pstring);
917 *pstring = xstrdup (val);
922 /* Like the above, but handles tilde-expansion when reading a user's
923 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
924 gets expanded to the user's home directory. */
926 cmd_file (const char *com, const char *val, void *place)
928 char **pstring = (char **)place;
930 xfree_null (*pstring);
932 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
934 *pstring = xstrdup (val);
936 #if defined(WINDOWS) || defined(MSDOS)
937 /* Convert "\" to "/". */
940 for (s = *pstring; *s; s++)
948 /* Like cmd_file, but strips trailing '/' characters. */
950 cmd_directory (const char *com, const char *val, void *place)
954 /* Call cmd_file() for tilde expansion and separator
955 canonicalization (backslash -> slash under Windows). These
956 things should perhaps be in a separate function. */
957 if (!cmd_file (com, val, place))
962 while (t > s && *--t == '/')
968 /* Split VAL by space to a vector of values, and append those values
969 to vector pointed to by the PLACE argument. If VAL is empty, the
970 PLACE vector is cleared instead. */
973 cmd_vector (const char *com, const char *val, void *place)
975 char ***pvec = (char ***)place;
978 *pvec = merge_vecs (*pvec, sepstring (val));
988 cmd_directory_vector (const char *com, const char *val, void *place)
990 char ***pvec = (char ***)place;
994 /* Strip the trailing slashes from directories. */
997 seps = sepstring (val);
998 for (t = seps; t && *t; t++)
1000 int len = strlen (*t);
1001 /* Skip degenerate case of root directory. */
1004 if ((*t)[len - 1] == '/')
1005 (*t)[len - 1] = '\0';
1008 *pvec = merge_vecs (*pvec, seps);
1018 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1019 "100k" or "2.5G" to a floating point number. */
1022 parse_bytes_helper (const char *val, double *result)
1024 double number, mult;
1025 const char *end = val + strlen (val);
1027 /* Check for "inf". */
1028 if (0 == strcmp (val, "inf"))
1034 /* Strip trailing whitespace. */
1035 while (val < end && c_isspace (end[-1]))
1040 switch (c_tolower (end[-1]))
1043 --end, mult = 1024.0;
1046 --end, mult = 1048576.0;
1049 --end, mult = 1073741824.0;
1052 --end, mult = 1099511627776.0;
1055 /* Not a recognized suffix: assume it's a digit. (If not,
1056 simple_atof will raise an error.) */
1060 /* Skip leading and trailing whitespace. */
1061 while (val < end && c_isspace (*val))
1063 while (val < end && c_isspace (end[-1]))
1068 if (!simple_atof (val, end, &number) || number < 0)
1071 *result = number * mult;
1075 /* Parse VAL as a number and set its value to PLACE (which should
1078 By default, the value is assumed to be in bytes. If "K", "M", or
1079 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1080 1<<30, respectively. Floating point values are allowed and are
1081 cast to integer before use. The idea is to be able to use things
1082 like 1.5k instead of "1536".
1084 The string "inf" is returned as 0.
1086 In case of error, false is returned and memory pointed to by PLACE
1087 remains unmodified. */
1090 cmd_bytes (const char *com, const char *val, void *place)
1093 if (!parse_bytes_helper (val, &byte_value))
1095 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1096 exec_name, com, quote (val));
1099 *(wgint *)place = (wgint)byte_value;
1103 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1104 SIZE_SUM. It works by converting the string to double, therefore
1105 working with values up to 2^53-1 without loss of precision. This
1106 value (8192 TB) is large enough to serve for a while. */
1109 cmd_bytes_sum (const char *com, const char *val, void *place)
1112 if (!parse_bytes_helper (val, &byte_value))
1114 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1115 exec_name, com, quote (val));
1118 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1122 /* Store the value of VAL to *OUT. The value is a time period, by
1123 default expressed in seconds, but also accepting suffixes "m", "h",
1124 "d", and "w" for minutes, hours, days, and weeks respectively. */
1127 cmd_time (const char *com, const char *val, void *place)
1129 double number, mult;
1130 const char *end = val + strlen (val);
1132 /* Strip trailing whitespace. */
1133 while (val < end && c_isspace (end[-1]))
1139 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1140 exec_name, com, quote (val));
1144 switch (c_tolower (end[-1]))
1147 --end, mult = 1; /* seconds */
1150 --end, mult = 60; /* minutes */
1153 --end, mult = 3600; /* hours */
1156 --end, mult = 86400.0; /* days */
1159 --end, mult = 604800.0; /* weeks */
1162 /* Not a recognized suffix: assume it belongs to the number.
1163 (If not, simple_atof will raise an error.) */
1167 /* Skip leading and trailing whitespace. */
1168 while (val < end && c_isspace (*val))
1170 while (val < end && c_isspace (end[-1]))
1175 if (!simple_atof (val, end, &number))
1178 *(double *)place = number * mult;
1184 cmd_cert_type (const char *com, const char *val, void *place)
1186 static const struct decode_item choices[] = {
1187 { "pem", keyfile_pem },
1188 { "der", keyfile_asn1 },
1189 { "asn1", keyfile_asn1 },
1191 int ok = decode_string (val, choices, countof (choices), place);
1193 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1198 /* Specialized helper functions, used by `commands' to handle some
1199 options specially. */
1201 static bool check_user_specified_header (const char *);
1204 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1206 if (!cmd_boolean (com, val, &opt.dirstruct))
1208 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1209 must be affected inversely. */
1211 opt.no_dirstruct = false;
1213 opt.no_dirstruct = true;
1218 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1220 /* Empty value means reset the list of headers. */
1223 free_vec (opt.user_headers);
1224 opt.user_headers = NULL;
1228 if (!check_user_specified_header (val))
1230 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1231 exec_name, com, quote (val));
1234 opt.user_headers = vec_append (opt.user_headers, val);
1239 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1241 int flag = cmd_boolean (com, val, &opt.htmlify);
1242 if (flag && !opt.htmlify)
1243 opt.remove_listing = false;
1247 /* Set the "mirror" mode. It means: recursive download, timestamping,
1248 no limit on max. recursion depth, and don't remove listings. */
1251 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1255 if (!cmd_boolean (com, val, &mirror))
1259 opt.recursive = true;
1260 if (!opt.no_dirstruct)
1261 opt.dirstruct = true;
1262 opt.timestamping = true;
1263 opt.reclevel = INFINITE_RECURSION;
1264 opt.remove_listing = false;
1269 /* Validate --prefer-family and set the choice. Allowed values are
1270 "IPv4", "IPv6", and "none". */
1273 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1275 static const struct decode_item choices[] = {
1276 { "IPv4", prefer_ipv4 },
1277 { "IPv6", prefer_ipv6 },
1278 { "none", prefer_none },
1280 int prefer_family = prefer_none;
1281 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1283 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1284 opt.prefer_family = prefer_family;
1288 /* Set progress.type to VAL, but verify that it's a valid progress
1289 implementation before that. */
1292 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1294 if (!valid_progress_implementation_p (val))
1296 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1297 exec_name, com, quote (val));
1300 xfree_null (opt.progress_type);
1302 /* Don't call set_progress_implementation here. It will be called
1303 in main() when it becomes clear what the log output is. */
1304 opt.progress_type = xstrdup (val);
1308 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1309 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1313 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1315 if (!cmd_boolean (com, val, &opt.recursive))
1319 if (opt.recursive && !opt.no_dirstruct)
1320 opt.dirstruct = true;
1326 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1328 int restrict_os = opt.restrict_files_os;
1329 int restrict_ctrl = opt.restrict_files_ctrl;
1330 int restrict_case = opt.restrict_files_case;
1331 int restrict_nonascii = opt.restrict_files_nonascii;
1335 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1339 end = strchr (val, ',');
1341 end = val + strlen (val);
1343 if (VAL_IS ("unix"))
1344 restrict_os = restrict_unix;
1345 else if (VAL_IS ("windows"))
1346 restrict_os = restrict_windows;
1347 else if (VAL_IS ("lowercase"))
1348 restrict_case = restrict_lowercase;
1349 else if (VAL_IS ("uppercase"))
1350 restrict_case = restrict_uppercase;
1351 else if (VAL_IS ("nocontrol"))
1352 restrict_ctrl = false;
1353 else if (VAL_IS ("ascii"))
1354 restrict_nonascii = true;
1357 fprintf (stderr, _("\
1358 %s: %s: Invalid restriction %s,\n\
1359 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1360 exec_name, com, quote (val));
1367 while (*val && *end);
1371 opt.restrict_files_os = restrict_os;
1372 opt.restrict_files_ctrl = restrict_ctrl;
1373 opt.restrict_files_case = restrict_case;
1374 opt.restrict_files_nonascii = restrict_nonascii;
1381 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1383 static const struct decode_item choices[] = {
1384 { "auto", secure_protocol_auto },
1385 { "sslv2", secure_protocol_sslv2 },
1386 { "sslv3", secure_protocol_sslv3 },
1387 { "tlsv1", secure_protocol_tlsv1 },
1389 int ok = decode_string (val, choices, countof (choices), place);
1391 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1396 /* Set all three timeout values. */
1399 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1402 if (!cmd_time (com, val, &value))
1404 opt.read_timeout = value;
1405 opt.connect_timeout = value;
1406 opt.dns_timeout = value;
1411 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1413 /* Disallow embedded newlines. */
1414 if (strchr (val, '\n'))
1416 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1417 exec_name, com, quote (val));
1420 xfree_null (opt.useragent);
1421 opt.useragent = xstrdup (val);
1425 /* The "verbose" option cannot be cmd_boolean because the variable is
1426 not bool -- it's of type int (-1 means uninitialized because of
1427 some random hackery for disallowing -q -v). */
1430 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1433 if (cmd_boolean (com, val, &flag))
1441 /* Miscellaneous useful routines. */
1443 /* A very simple atoi clone, more useful than atoi because it works on
1444 delimited strings, and has error reportage. Returns true on success,
1445 false on failure. If successful, stores result to *DEST. */
1448 simple_atoi (const char *beg, const char *end, int *dest)
1451 bool negative = false;
1452 const char *p = beg;
1454 while (p < end && c_isspace (*p))
1456 if (p < end && (*p == '-' || *p == '+'))
1458 negative = (*p == '-');
1464 /* Read negative numbers in a separate loop because the most
1465 negative integer cannot be represented as a positive number. */
1468 for (; p < end && c_isdigit (*p); p++)
1470 int next = (10 * result) + (*p - '0');
1472 return false; /* overflow */
1476 for (; p < end && c_isdigit (*p); p++)
1478 int next = (10 * result) - (*p - '0');
1480 return false; /* underflow */
1491 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1492 doesn't handle exponential notation. Returns true on success,
1493 false on failure. In case of success, stores its result to
1497 simple_atof (const char *beg, const char *end, double *dest)
1501 bool negative = false;
1502 bool seen_dot = false;
1503 bool seen_digit = false;
1506 const char *p = beg;
1508 while (p < end && c_isspace (*p))
1510 if (p < end && (*p == '-' || *p == '+'))
1512 negative = (*p == '-');
1516 for (; p < end; p++)
1522 result = (10 * result) + (ch - '0');
1524 result += (ch - '0') / (divider *= 10);
1546 /* Verify that the user-specified header in S is valid. It must
1547 contain a colon preceded by non-white-space characters and must not
1548 contain newlines. */
1551 check_user_specified_header (const char *s)
1555 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1557 /* The header MUST contain `:' preceded by at least one
1558 non-whitespace character. */
1559 if (*p != ':' || p == s)
1561 /* The header MUST NOT contain newlines. */
1562 if (strchr (s, '\n'))
1567 /* Decode VAL into a number, according to ITEMS. */
1570 decode_string (const char *val, const struct decode_item *items, int itemcount,
1574 for (i = 0; i < itemcount; i++)
1575 if (0 == strcasecmp (val, items[i].name))
1577 *place = items[i].code;
1584 void cleanup_html_url (void);
1587 /* Free the memory allocated by global variables. */
1591 /* Free external resources, close files, etc. */
1594 fclose (output_stream);
1595 /* No need to check for error because Wget flushes its output (and
1596 checks for errors) after any data arrives. */
1598 /* We're exiting anyway so there's no real need to call free()
1599 hundreds of times. Skipping the frees will make Wget exit
1602 However, when detecting leaks, it's crucial to free() everything
1603 because then you can find the real leaks, i.e. the allocated
1604 memory which grows with the size of the program. */
1610 cleanup_html_url ();
1615 extern acc_t *netrc_list;
1616 free_netrc (netrc_list);
1618 xfree_null (opt.choose_config);
1619 xfree_null (opt.lfilename);
1620 xfree_null (opt.dir_prefix);
1621 xfree_null (opt.input_filename);
1622 xfree_null (opt.output_document);
1623 free_vec (opt.accepts);
1624 free_vec (opt.rejects);
1625 free_vec (opt.excludes);
1626 free_vec (opt.includes);
1627 free_vec (opt.domains);
1628 free_vec (opt.follow_tags);
1629 free_vec (opt.ignore_tags);
1630 xfree_null (opt.progress_type);
1631 xfree_null (opt.ftp_user);
1632 xfree_null (opt.ftp_passwd);
1633 xfree_null (opt.ftp_proxy);
1634 xfree_null (opt.https_proxy);
1635 xfree_null (opt.http_proxy);
1636 free_vec (opt.no_proxy);
1637 xfree_null (opt.useragent);
1638 xfree_null (opt.referer);
1639 xfree_null (opt.http_user);
1640 xfree_null (opt.http_passwd);
1641 free_vec (opt.user_headers);
1643 xfree_null (opt.cert_file);
1644 xfree_null (opt.private_key);
1645 xfree_null (opt.ca_directory);
1646 xfree_null (opt.ca_cert);
1647 xfree_null (opt.random_file);
1648 xfree_null (opt.egd_file);
1650 xfree_null (opt.bind_address);
1651 xfree_null (opt.cookies_input);
1652 xfree_null (opt.cookies_output);
1653 xfree_null (opt.user);
1654 xfree_null (opt.passwd);
1655 xfree_null (opt.base_href);
1657 #endif /* DEBUG_MALLOC */
1660 /* Unit testing routines. */
1665 test_commands_sorted()
1667 int prev_idx = 0, next_idx = 1;
1668 int command_count = countof (commands) - 1;
1670 while (next_idx <= command_count)
1672 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1675 mu_assert ("FAILED", false);
1688 test_cmd_spec_restrict_file_names()
1693 int expected_restrict_files_os;
1694 int expected_restrict_files_ctrl;
1695 int expected_restrict_files_case;
1698 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1699 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1700 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1701 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1704 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1709 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1712 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1713 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1714 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1715 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1717 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1718 res == test_array[i].result
1719 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1720 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1721 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1727 #endif /* TESTING */