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, 2012 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. */
42 /* not all systems provide PATH_MAX in limits.h */
44 # include <sys/param.h>
46 # define PATH_MAX MAXPATHLEN
65 #include "recur.h" /* for INFINITE_RECURSION */
66 #include "convert.h" /* for convert_cleanup */
67 #include "res.h" /* for res_cleanup */
68 #include "http.h" /* for http_cleanup */
69 #include "retr.h" /* for output_stream */
77 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
79 CMD_DECLARE (cmd_boolean);
80 CMD_DECLARE (cmd_bytes);
81 CMD_DECLARE (cmd_bytes_sum);
83 CMD_DECLARE (cmd_cert_type);
85 CMD_DECLARE (cmd_directory_vector);
86 CMD_DECLARE (cmd_number);
87 CMD_DECLARE (cmd_number_inf);
88 CMD_DECLARE (cmd_string);
89 CMD_DECLARE (cmd_file);
90 CMD_DECLARE (cmd_directory);
91 CMD_DECLARE (cmd_time);
92 CMD_DECLARE (cmd_vector);
94 CMD_DECLARE (cmd_spec_dirstruct);
95 CMD_DECLARE (cmd_spec_header);
96 CMD_DECLARE (cmd_spec_warc_header);
97 CMD_DECLARE (cmd_spec_htmlify);
98 CMD_DECLARE (cmd_spec_mirror);
99 CMD_DECLARE (cmd_spec_prefer_family);
100 CMD_DECLARE (cmd_spec_progress);
101 CMD_DECLARE (cmd_spec_recursive);
102 CMD_DECLARE (cmd_spec_regex_type);
103 CMD_DECLARE (cmd_spec_restrict_file_names);
104 CMD_DECLARE (cmd_spec_report_speed);
106 CMD_DECLARE (cmd_spec_secure_protocol);
108 CMD_DECLARE (cmd_spec_timeout);
109 CMD_DECLARE (cmd_spec_useragent);
110 CMD_DECLARE (cmd_spec_verbose);
112 /* List of recognized commands, each consisting of name, place and
113 function. When adding a new command, simply add it to the list,
114 but be sure to keep the list sorted alphabetically, as
115 command_by_name's binary search depends on it. Also, be sure to
116 add any entries that allocate memory (e.g. cmd_string and
117 cmd_vector) to the cleanup() function below. */
119 static const struct {
122 bool (*action) (const char *, const char *, void *);
124 /* KEEP THIS LIST ALPHABETICALLY SORTED */
125 { "accept", &opt.accepts, cmd_vector },
126 { "acceptregex", &opt.acceptregex_s, cmd_string },
127 { "addhostdir", &opt.add_hostdir, cmd_boolean },
128 { "adjustextension", &opt.adjust_extension, cmd_boolean },
129 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
130 { "askpassword", &opt.ask_passwd, cmd_boolean },
131 { "authnochallenge", &opt.auth_without_challenge,
133 { "background", &opt.background, cmd_boolean },
134 { "backupconverted", &opt.backup_converted, cmd_boolean },
135 { "backups", &opt.backups, cmd_number },
136 { "base", &opt.base_href, cmd_string },
137 { "bindaddress", &opt.bind_address, cmd_string },
139 { "cacertificate", &opt.ca_cert, cmd_file },
141 { "cache", &opt.allow_cache, cmd_boolean },
143 { "cadirectory", &opt.ca_directory, cmd_directory },
144 { "certificate", &opt.cert_file, cmd_file },
145 { "certificatetype", &opt.cert_type, cmd_cert_type },
146 { "checkcertificate", &opt.check_cert, cmd_boolean },
148 { "chooseconfig", &opt.choose_config, cmd_file },
149 { "connecttimeout", &opt.connect_timeout, cmd_time },
150 { "contentdisposition", &opt.content_disposition, cmd_boolean },
151 { "contentonerror", &opt.content_on_error, cmd_boolean },
152 { "continue", &opt.always_rest, cmd_boolean },
153 { "convertlinks", &opt.convert_links, cmd_boolean },
154 { "cookies", &opt.cookies, cmd_boolean },
155 { "cutdirs", &opt.cut_dirs, cmd_number },
157 { "debug", &opt.debug, cmd_boolean },
159 { "defaultpage", &opt.default_page, cmd_string},
160 { "deleteafter", &opt.delete_after, cmd_boolean },
161 { "dirprefix", &opt.dir_prefix, cmd_directory },
162 { "dirstruct", NULL, cmd_spec_dirstruct },
163 { "dnscache", &opt.dns_cache, cmd_boolean },
164 { "dnstimeout", &opt.dns_timeout, cmd_time },
165 { "domains", &opt.domains, cmd_vector },
166 { "dotbytes", &opt.dot_bytes, cmd_bytes },
167 { "dotsinline", &opt.dots_in_line, cmd_number },
168 { "dotspacing", &opt.dot_spacing, cmd_number },
169 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
171 { "egdfile", &opt.egd_file, cmd_file },
173 { "excludedirectories", &opt.excludes, cmd_directory_vector },
174 { "excludedomains", &opt.exclude_domains, cmd_vector },
175 { "followftp", &opt.follow_ftp, cmd_boolean },
176 { "followtags", &opt.follow_tags, cmd_vector },
177 { "forcehtml", &opt.force_html, cmd_boolean },
178 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
179 { "ftppassword", &opt.ftp_passwd, cmd_string },
180 { "ftpproxy", &opt.ftp_proxy, cmd_string },
182 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
183 #endif /* def __VMS */
184 { "ftpuser", &opt.ftp_user, cmd_string },
185 { "glob", &opt.ftp_glob, cmd_boolean },
186 { "header", NULL, cmd_spec_header },
187 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
188 { "htmlify", NULL, cmd_spec_htmlify },
189 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
190 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
191 { "httppassword", &opt.http_passwd, cmd_string },
192 { "httpproxy", &opt.http_proxy, cmd_string },
193 { "httpsproxy", &opt.https_proxy, cmd_string },
194 { "httpuser", &opt.http_user, cmd_string },
195 { "ignorecase", &opt.ignore_case, cmd_boolean },
196 { "ignorelength", &opt.ignore_length, cmd_boolean },
197 { "ignoretags", &opt.ignore_tags, cmd_vector },
198 { "includedirectories", &opt.includes, cmd_directory_vector },
200 { "inet4only", &opt.ipv4_only, cmd_boolean },
201 { "inet6only", &opt.ipv6_only, cmd_boolean },
203 { "input", &opt.input_filename, cmd_file },
204 { "iri", &opt.enable_iri, cmd_boolean },
205 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
206 { "limitrate", &opt.limit_rate, cmd_bytes },
207 { "loadcookies", &opt.cookies_input, cmd_file },
208 { "localencoding", &opt.locale, cmd_string },
209 { "logfile", &opt.lfilename, cmd_file },
210 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
211 { "maxredirect", &opt.max_redirect, cmd_number },
212 { "mirror", NULL, cmd_spec_mirror },
213 { "netrc", &opt.netrc, cmd_boolean },
214 { "noclobber", &opt.noclobber, cmd_boolean },
215 { "noparent", &opt.no_parent, cmd_boolean },
216 { "noproxy", &opt.no_proxy, cmd_vector },
217 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
218 { "outputdocument", &opt.output_document, cmd_file },
219 { "pagerequisites", &opt.page_requisites, cmd_boolean },
220 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
221 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
222 { "password", &opt.passwd, cmd_string },
223 { "postdata", &opt.post_data, cmd_string },
224 { "postfile", &opt.post_file_name, cmd_file },
225 { "preferfamily", NULL, cmd_spec_prefer_family },
226 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
228 { "privatekey", &opt.private_key, cmd_file },
229 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
231 { "progress", &opt.progress_type, cmd_spec_progress },
232 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
233 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
234 { "proxypassword", &opt.proxy_passwd, cmd_string },
235 { "proxyuser", &opt.proxy_user, cmd_string },
236 { "quiet", &opt.quiet, cmd_boolean },
237 { "quota", &opt.quota, cmd_bytes_sum },
239 { "randomfile", &opt.random_file, cmd_file },
241 { "randomwait", &opt.random_wait, cmd_boolean },
242 { "readtimeout", &opt.read_timeout, cmd_time },
243 { "reclevel", &opt.reclevel, cmd_number_inf },
244 { "recursive", NULL, cmd_spec_recursive },
245 { "referer", &opt.referer, cmd_string },
246 { "regextype", &opt.regex_type, cmd_spec_regex_type },
247 { "reject", &opt.rejects, cmd_vector },
248 { "rejectregex", &opt.rejectregex_s, cmd_string },
249 { "relativeonly", &opt.relative_only, cmd_boolean },
250 { "remoteencoding", &opt.encoding_remote, cmd_string },
251 { "removelisting", &opt.remove_listing, cmd_boolean },
252 { "reportspeed", &opt.report_bps, cmd_spec_report_speed},
253 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
254 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
255 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
256 { "robots", &opt.use_robots, cmd_boolean },
257 { "savecookies", &opt.cookies_output, cmd_file },
258 { "saveheaders", &opt.save_headers, cmd_boolean },
260 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
262 { "serverresponse", &opt.server_response, cmd_boolean },
263 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
264 { "spanhosts", &opt.spanhost, cmd_boolean },
265 { "spider", &opt.spider, cmd_boolean },
266 { "strictcomments", &opt.strict_comments, cmd_boolean },
267 { "timeout", NULL, cmd_spec_timeout },
268 { "timestamping", &opt.timestamping, cmd_boolean },
269 { "tries", &opt.ntry, cmd_number_inf },
270 { "trustservernames", &opt.trustservernames, cmd_boolean },
271 { "unlink", &opt.unlink, cmd_boolean },
272 { "useproxy", &opt.use_proxy, cmd_boolean },
273 { "user", &opt.user, cmd_string },
274 { "useragent", NULL, cmd_spec_useragent },
275 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
276 { "verbose", NULL, cmd_spec_verbose },
277 { "wait", &opt.wait, cmd_time },
278 { "waitretry", &opt.waitretry, cmd_time },
279 { "warccdx", &opt.warc_cdx_enabled, cmd_boolean },
280 { "warccdxdedup", &opt.warc_cdx_dedup_filename, cmd_file },
282 { "warccompression", &opt.warc_compression_enabled, cmd_boolean },
284 { "warcdigests", &opt.warc_digests_enabled, cmd_boolean },
285 { "warcfile", &opt.warc_filename, cmd_file },
286 { "warcheader", NULL, cmd_spec_warc_header },
287 { "warckeeplog", &opt.warc_keep_log, cmd_boolean },
288 { "warcmaxsize", &opt.warc_maxsize, cmd_bytes },
289 { "warctempdir", &opt.warc_tempdir, cmd_directory },
291 { "wdebug", &opt.wdebug, cmd_boolean },
295 /* Look up CMDNAME in the commands[] and return its position in the
296 array. If CMDNAME is not found, return -1. */
299 command_by_name (const char *cmdname)
301 /* Use binary search for speed. Wget has ~100 commands, which
302 guarantees a worst case performance of 7 string comparisons. */
303 int lo = 0, hi = countof (commands) - 1;
307 int mid = (lo + hi) >> 1;
308 int cmp = strcasecmp (cmdname, commands[mid].name);
319 /* Reset the variables to default values. */
325 /* Most of the default values are 0 (and 0.0, NULL, and false).
326 Just reset everything, and fill in the non-zero values. Note
327 that initializing pointers to NULL this way is technically
328 illegal, but porting Wget to a machine where NULL is not all-zero
329 bit pattern will be the least of the implementors' worries. */
336 opt.add_hostdir = true;
340 opt.http_keep_alive = true;
341 opt.use_proxy = true;
342 tmp = getenv ("no_proxy");
344 opt.no_proxy = sepstring (tmp);
345 opt.prefer_family = prefer_none;
346 opt.allow_cache = true;
348 opt.read_timeout = 900;
349 opt.use_robots = true;
351 opt.remove_listing = true;
353 opt.dot_bytes = 1024;
354 opt.dot_spacing = 10;
355 opt.dots_in_line = 50;
357 opt.dns_cache = true;
361 opt.check_cert = true;
364 /* The default for file name restriction defaults to the OS type. */
365 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
366 opt.restrict_files_os = restrict_windows;
368 opt.restrict_files_os = restrict_unix;
370 opt.restrict_files_ctrl = true;
371 opt.restrict_files_nonascii = false;
372 opt.restrict_files_case = restrict_no_case_restriction;
374 opt.regex_type = regex_type_posix;
376 opt.max_redirect = 20;
381 opt.enable_iri = true;
383 opt.enable_iri = false;
386 opt.encoding_remote = NULL;
388 opt.useservertimestamps = true;
389 opt.show_all_dns_entries = false;
391 opt.warc_maxsize = 0; /* 1024 * 1024 * 1024; */
393 opt.warc_compression_enabled = true;
395 opt.warc_compression_enabled = false;
397 opt.warc_digests_enabled = true;
398 opt.warc_cdx_enabled = false;
399 opt.warc_cdx_dedup_filename = NULL;
400 opt.warc_tempdir = NULL;
401 opt.warc_keep_log = true;
404 /* Return the user's home directory (strdup-ed), or NULL if none is
409 static char *buf = NULL;
410 static char *home, *ret;
414 home = getenv ("HOME");
420 /* Under MSDOS, if $HOME isn't defined, use the directory where
421 `wget.exe' resides. */
422 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
425 buff = _w32_get_argv0 ();
427 p = strrchr (buf, '/'); /* djgpp */
429 p = strrchr (buf, '\\'); /* others */
433 buff = malloc (len + 1);
437 strncpy (buff, _w32_get_argv0 (), len);
441 #elif !defined(WINDOWS)
442 /* If HOME is not defined, try getting it from the password
444 struct passwd *pwd = getpwuid (getuid ());
445 if (!pwd || !pwd->pw_dir)
449 /* Under Windows, if $HOME isn't defined, use the directory where
450 `wget.exe' resides. */
456 ret = home ? xstrdup (home) : NULL;
463 /* Check the 'WGETRC' environment variable and return the file name
464 if 'WGETRC' is set and is a valid file.
465 If the `WGETRC' variable exists but the file does not exist, the
466 function will exit(). */
468 wgetrc_env_file_name (void)
470 char *env = getenv ("WGETRC");
473 if (!file_exists_p (env))
475 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
479 return xstrdup (env);
484 /* Check for the existance of '$HOME/.wgetrc' and return its path
485 if it exists and is set. */
487 wgetrc_user_file_name (void)
491 /* If that failed, try $HOME/.wgetrc (or equivalent). */
494 file = "SYS$LOGIN:.wgetrc";
495 #else /* def __VMS */
498 file = aprintf ("%s/.wgetrc", home);
500 #endif /* def __VMS [else] */
504 if (!file_exists_p (file))
512 /* Return the path to the user's .wgetrc. This is either the value of
513 `WGETRC' environment variable, or `$HOME/.wgetrc'.
515 Additionally, for windows, look in the directory where wget.exe
518 wgetrc_file_name (void)
520 char *file = wgetrc_env_file_name ();
524 file = wgetrc_user_file_name ();
527 /* Under Windows, if we still haven't found .wgetrc, look for the file
528 `wget.ini' in the directory where `wget.exe' resides; we do this for
529 backward compatibility with previous versions of Wget.
530 SYSTEM_WGETRC should not be defined under WINDOWS. */
533 char *home = home_dir ();
539 file = aprintf ("%s/wget.ini", home);
540 if (!file_exists_p (file))
553 /* Return values of parse_line. */
561 static enum parse_line parse_line (const char *, char **, char **, int *);
562 static bool setval_internal (int, const char *, const char *);
563 static bool setval_internal_tilde (int, const char *, const char *);
565 /* Initialize variables from a wgetrc file. Returns zero (failure) if
566 there were errors in the file. */
569 run_wgetrc (const char *file)
576 fp = fopen (file, "r");
579 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
580 file, strerror (errno));
581 return true; /* not a fatal error */
584 while ((line = read_whole_line (fp)) != NULL)
586 char *com = NULL, *val = NULL;
589 /* Parse the line. */
590 switch (parse_line (line, &com, &val, &comind))
593 /* If everything is OK, set the value. */
594 if (!setval_internal_tilde (comind, com, val))
596 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
597 exec_name, file, ln);
601 case line_syntax_error:
602 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
603 exec_name, file, ln);
606 case line_unknown_command:
607 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
608 exec_name, quote (com), file, ln);
626 /* Initialize the defaults and run the system wgetrc and user's own
631 char *file, *env_sysrc;
634 /* Run a non-standard system rc file when the according environment
635 variable has been set. For internal testing purposes only! */
636 env_sysrc = getenv ("SYSTEM_WGETRC");
637 if (env_sysrc && file_exists_p (env_sysrc))
639 ok &= run_wgetrc (env_sysrc);
640 /* If there are any problems parsing the system wgetrc file, tell
644 fprintf (stderr, _("\
645 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
647 or specify a different file using --config.\n"), env_sysrc);
651 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
653 else if (file_exists_p (SYSTEM_WGETRC))
654 ok &= run_wgetrc (SYSTEM_WGETRC);
655 /* If there are any problems parsing the system wgetrc file, tell
659 fprintf (stderr, _("\
660 Parsing system wgetrc file failed. Please check\n\
662 or specify a different file using --config.\n"), SYSTEM_WGETRC);
666 /* Override it with your own, if one exists. */
667 file = wgetrc_file_name ();
670 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
671 something like realpath() before comparing them with `strcmp' */
673 if (!strcmp (file, SYSTEM_WGETRC))
675 fprintf (stderr, _("\
676 %s: Warning: Both system and user wgetrc point to %s.\n"),
677 exec_name, quote (file));
681 ok &= run_wgetrc (file);
683 /* If there were errors processing either `.wgetrc', abort. */
691 /* Remove dashes and underscores from S, modifying S in the
697 char *t = s; /* t - tortoise */
698 char *h = s; /* h - hare */
700 if (*h == '_' || *h == '-')
707 /* Parse the line pointed by line, with the syntax:
708 <sp>* command <sp>* = <sp>* value <sp>*
709 Uses malloc to allocate space for command and value.
711 Returns one of line_ok, line_empty, line_syntax_error, or
712 line_unknown_command.
714 In case of line_ok, *COM and *VAL point to freshly allocated
715 strings, and *COMIND points to com's index. In case of error or
716 empty line, their values are unmodified. */
718 static enum parse_line
719 parse_line (const char *line, char **com, char **val, int *comind)
722 const char *end = line + strlen (line);
723 const char *cmdstart, *cmdend;
724 const char *valstart, *valend;
729 /* Skip leading and trailing whitespace. */
730 while (*line && c_isspace (*line))
732 while (end > line && c_isspace (end[-1]))
735 /* Skip empty lines and comments. */
736 if (!*line || *line == '#')
742 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
746 /* Skip '=', as well as any space before or after it. */
747 while (p < end && c_isspace (*p))
749 if (p == end || *p != '=')
750 return line_syntax_error;
752 while (p < end && c_isspace (*p))
758 /* The syntax is valid (even though the command might not be). Fill
759 in the command name and value. */
760 *com = strdupdelim (cmdstart, cmdend);
761 *val = strdupdelim (valstart, valend);
763 /* The line now known to be syntactically correct. Check whether
764 the command is valid. */
765 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
767 ind = command_by_name (cmdcopy);
769 return line_unknown_command;
771 /* Report success to the caller. */
776 #if defined(WINDOWS) || defined(MSDOS)
777 # define ISSEP(c) ((c) == '/' || (c) == '\\')
779 # define ISSEP(c) ((c) == '/')
782 /* Run commands[comind].action. */
785 setval_internal (int comind, const char *com, const char *val)
787 assert (0 <= comind && ((size_t) comind) < countof (commands));
788 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
789 return commands[comind].action (com, val, commands[comind].place);
793 setval_internal_tilde (int comind, const char *com, const char *val)
799 ret = setval_internal (comind, com, val);
801 /* We make tilde expansion for cmd_file and cmd_directory */
802 if (((commands[comind].action == cmd_file) ||
803 (commands[comind].action == cmd_directory))
804 && ret && (*val == '~' && ISSEP (val[1])))
806 pstring = commands[comind].place;
810 homelen = strlen (home);
811 while (homelen && ISSEP (home[homelen - 1]))
812 home[--homelen] = '\0';
814 /* Skip the leading "~/". */
815 for (++val; ISSEP (*val); val++)
817 *pstring = concat_strings (home, "/", val, (char *)0);
823 /* Run command COM with value VAL. If running the command produces an
824 error, report the error and exit.
826 This is intended to be called from main() to modify Wget's behavior
827 through command-line switches. Since COM is hard-coded in main(),
828 it is not canonicalized, and this aborts when COM is not found.
830 If COMIND's are exported to init.h, this function will be changed
831 to accept COMIND directly. */
834 setoptval (const char *com, const char *val, const char *optname)
836 /* Prepend "--" to OPTNAME. */
837 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
840 strcpy (dd_optname + 2, optname);
842 assert (val != NULL);
843 if (!setval_internal (command_by_name (com), dd_optname, val))
847 /* Parse OPT into command and value and run it. For example,
848 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
849 This is used by the `--execute' flag in main.c. */
852 run_command (const char *opt)
856 switch (parse_line (opt, &com, &val, &comind))
859 if (!setval_internal (comind, com, val))
865 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
866 exec_name, quote (opt));
871 /* Generic helper functions, for use with `commands'. */
873 /* Forward declarations: */
878 static bool decode_string (const char *, const struct decode_item *, int, int *);
879 static bool simple_atoi (const char *, const char *, int *);
880 static bool simple_atof (const char *, const char *, double *);
882 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
884 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
885 && c_tolower((p)[1]) == (c1) \
888 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
889 && c_tolower((p)[1]) == (c1) \
890 && c_tolower((p)[2]) == (c2) \
894 /* Store the boolean value from VAL to PLACE. COM is ignored,
895 except for error messages. */
897 cmd_boolean (const char *com, const char *val, void *place)
901 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
902 /* "on", "yes" and "1" mean true. */
904 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
905 /* "off", "no" and "0" mean false. */
910 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
911 exec_name, com, quote (val));
915 *(bool *) place = value;
919 /* Set the non-negative integer value from VAL to PLACE. With
920 incorrect specification, the number remains unchanged. */
922 cmd_number (const char *com, const char *val, void *place)
924 if (!simple_atoi (val, val + strlen (val), place)
925 || *(int *) place < 0)
927 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
928 exec_name, com, quote (val));
934 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
936 cmd_number_inf (const char *com, const char *val, void *place)
938 if (!strcasecmp (val, "inf"))
943 return cmd_number (com, val, place);
946 /* Copy (strdup) the string at COM to a new location and place a
947 pointer to *PLACE. */
949 cmd_string (const char *com, const char *val, void *place)
951 char **pstring = (char **)place;
953 xfree_null (*pstring);
954 *pstring = xstrdup (val);
959 /* Like the above, but handles tilde-expansion when reading a user's
960 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
961 gets expanded to the user's home directory. */
963 cmd_file (const char *com, const char *val, void *place)
965 char **pstring = (char **)place;
967 xfree_null (*pstring);
969 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
971 *pstring = xstrdup (val);
973 #if defined(WINDOWS) || defined(MSDOS)
974 /* Convert "\" to "/". */
977 for (s = *pstring; *s; s++)
985 /* Like cmd_file, but strips trailing '/' characters. */
987 cmd_directory (const char *com, const char *val, void *place)
991 /* Call cmd_file() for tilde expansion and separator
992 canonicalization (backslash -> slash under Windows). These
993 things should perhaps be in a separate function. */
994 if (!cmd_file (com, val, place))
999 while (t > s && *--t == '/')
1005 /* Split VAL by space to a vector of values, and append those values
1006 to vector pointed to by the PLACE argument. If VAL is empty, the
1007 PLACE vector is cleared instead. */
1010 cmd_vector (const char *com, const char *val, void *place)
1012 char ***pvec = (char ***)place;
1015 *pvec = merge_vecs (*pvec, sepstring (val));
1025 cmd_directory_vector (const char *com, const char *val, void *place)
1027 char ***pvec = (char ***)place;
1031 /* Strip the trailing slashes from directories. */
1034 seps = sepstring (val);
1035 for (t = seps; t && *t; t++)
1037 int len = strlen (*t);
1038 /* Skip degenerate case of root directory. */
1041 if ((*t)[len - 1] == '/')
1042 (*t)[len - 1] = '\0';
1045 *pvec = merge_vecs (*pvec, seps);
1055 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1056 "100k" or "2.5G" to a floating point number. */
1059 parse_bytes_helper (const char *val, double *result)
1061 double number, mult;
1062 const char *end = val + strlen (val);
1064 /* Check for "inf". */
1065 if (0 == strcmp (val, "inf"))
1071 /* Strip trailing whitespace. */
1072 while (val < end && c_isspace (end[-1]))
1077 switch (c_tolower (end[-1]))
1080 --end, mult = 1024.0;
1083 --end, mult = 1048576.0;
1086 --end, mult = 1073741824.0;
1089 --end, mult = 1099511627776.0;
1092 /* Not a recognized suffix: assume it's a digit. (If not,
1093 simple_atof will raise an error.) */
1097 /* Skip leading and trailing whitespace. */
1098 while (val < end && c_isspace (*val))
1100 while (val < end && c_isspace (end[-1]))
1105 if (!simple_atof (val, end, &number) || number < 0)
1108 *result = number * mult;
1112 /* Parse VAL as a number and set its value to PLACE (which should
1115 By default, the value is assumed to be in bytes. If "K", "M", or
1116 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1117 1<<30, respectively. Floating point values are allowed and are
1118 cast to integer before use. The idea is to be able to use things
1119 like 1.5k instead of "1536".
1121 The string "inf" is returned as 0.
1123 In case of error, false is returned and memory pointed to by PLACE
1124 remains unmodified. */
1127 cmd_bytes (const char *com, const char *val, void *place)
1130 if (!parse_bytes_helper (val, &byte_value))
1132 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1133 exec_name, com, quote (val));
1136 *(wgint *)place = (wgint)byte_value;
1140 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1141 SIZE_SUM. It works by converting the string to double, therefore
1142 working with values up to 2^53-1 without loss of precision. This
1143 value (8192 TB) is large enough to serve for a while. */
1146 cmd_bytes_sum (const char *com, const char *val, void *place)
1149 if (!parse_bytes_helper (val, &byte_value))
1151 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1152 exec_name, com, quote (val));
1155 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1159 /* Store the value of VAL to *OUT. The value is a time period, by
1160 default expressed in seconds, but also accepting suffixes "m", "h",
1161 "d", and "w" for minutes, hours, days, and weeks respectively. */
1164 cmd_time (const char *com, const char *val, void *place)
1166 double number, mult;
1167 const char *end = val + strlen (val);
1169 /* Strip trailing whitespace. */
1170 while (val < end && c_isspace (end[-1]))
1176 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1177 exec_name, com, quote (val));
1181 switch (c_tolower (end[-1]))
1184 --end, mult = 1; /* seconds */
1187 --end, mult = 60; /* minutes */
1190 --end, mult = 3600; /* hours */
1193 --end, mult = 86400.0; /* days */
1196 --end, mult = 604800.0; /* weeks */
1199 /* Not a recognized suffix: assume it belongs to the number.
1200 (If not, simple_atof will raise an error.) */
1204 /* Skip leading and trailing whitespace. */
1205 while (val < end && c_isspace (*val))
1207 while (val < end && c_isspace (end[-1]))
1212 if (!simple_atof (val, end, &number))
1215 *(double *)place = number * mult;
1221 cmd_cert_type (const char *com, const char *val, void *place)
1223 static const struct decode_item choices[] = {
1224 { "pem", keyfile_pem },
1225 { "der", keyfile_asn1 },
1226 { "asn1", keyfile_asn1 },
1228 int ok = decode_string (val, choices, countof (choices), place);
1230 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1235 /* Specialized helper functions, used by `commands' to handle some
1236 options specially. */
1238 static bool check_user_specified_header (const char *);
1241 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1243 if (!cmd_boolean (com, val, &opt.dirstruct))
1245 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1246 must be affected inversely. */
1248 opt.no_dirstruct = false;
1250 opt.no_dirstruct = true;
1255 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1257 /* Empty value means reset the list of headers. */
1260 free_vec (opt.user_headers);
1261 opt.user_headers = NULL;
1265 if (!check_user_specified_header (val))
1267 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1268 exec_name, com, quote (val));
1271 opt.user_headers = vec_append (opt.user_headers, val);
1276 cmd_spec_warc_header (const char *com, const char *val, void *place_ignored)
1278 /* Empty value means reset the list of headers. */
1281 free_vec (opt.warc_user_headers);
1282 opt.warc_user_headers = NULL;
1286 if (!check_user_specified_header (val))
1288 fprintf (stderr, _("%s: %s: Invalid WARC header %s.\n"),
1289 exec_name, com, quote (val));
1292 opt.warc_user_headers = vec_append (opt.warc_user_headers, val);
1297 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1299 int flag = cmd_boolean (com, val, &opt.htmlify);
1300 if (flag && !opt.htmlify)
1301 opt.remove_listing = false;
1305 /* Set the "mirror" mode. It means: recursive download, timestamping,
1306 no limit on max. recursion depth, and don't remove listings. */
1309 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1313 if (!cmd_boolean (com, val, &mirror))
1317 opt.recursive = true;
1318 if (!opt.no_dirstruct)
1319 opt.dirstruct = true;
1320 opt.timestamping = true;
1321 opt.reclevel = INFINITE_RECURSION;
1322 opt.remove_listing = false;
1327 /* Validate --prefer-family and set the choice. Allowed values are
1328 "IPv4", "IPv6", and "none". */
1331 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1333 static const struct decode_item choices[] = {
1334 { "IPv4", prefer_ipv4 },
1335 { "IPv6", prefer_ipv6 },
1336 { "none", prefer_none },
1338 int prefer_family = prefer_none;
1339 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1341 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1342 opt.prefer_family = prefer_family;
1346 /* Set progress.type to VAL, but verify that it's a valid progress
1347 implementation before that. */
1350 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1352 if (!valid_progress_implementation_p (val))
1354 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1355 exec_name, com, quote (val));
1358 xfree_null (opt.progress_type);
1360 /* Don't call set_progress_implementation here. It will be called
1361 in main() when it becomes clear what the log output is. */
1362 opt.progress_type = xstrdup (val);
1366 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1367 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1371 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1373 if (!cmd_boolean (com, val, &opt.recursive))
1377 if (opt.recursive && !opt.no_dirstruct)
1378 opt.dirstruct = true;
1383 /* Validate --regex-type and set the choice. */
1386 cmd_spec_regex_type (const char *com, const char *val, void *place_ignored)
1388 static const struct decode_item choices[] = {
1389 { "posix", regex_type_posix },
1391 { "pcre", regex_type_pcre },
1394 int regex_type = regex_type_posix;
1395 int ok = decode_string (val, choices, countof (choices), ®ex_type);
1397 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1398 opt.regex_type = regex_type;
1403 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1405 int restrict_os = opt.restrict_files_os;
1406 int restrict_ctrl = opt.restrict_files_ctrl;
1407 int restrict_case = opt.restrict_files_case;
1408 int restrict_nonascii = opt.restrict_files_nonascii;
1412 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1416 end = strchr (val, ',');
1418 end = val + strlen (val);
1420 if (VAL_IS ("unix"))
1421 restrict_os = restrict_unix;
1422 else if (VAL_IS ("windows"))
1423 restrict_os = restrict_windows;
1424 else if (VAL_IS ("lowercase"))
1425 restrict_case = restrict_lowercase;
1426 else if (VAL_IS ("uppercase"))
1427 restrict_case = restrict_uppercase;
1428 else if (VAL_IS ("nocontrol"))
1429 restrict_ctrl = false;
1430 else if (VAL_IS ("ascii"))
1431 restrict_nonascii = true;
1434 fprintf (stderr, _("\
1435 %s: %s: Invalid restriction %s,\n\
1436 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1437 exec_name, com, quote (val));
1444 while (*val && *end);
1448 opt.restrict_files_os = restrict_os;
1449 opt.restrict_files_ctrl = restrict_ctrl;
1450 opt.restrict_files_case = restrict_case;
1451 opt.restrict_files_nonascii = restrict_nonascii;
1457 cmd_spec_report_speed (const char *com, const char *val, void *place_ignored)
1459 opt.report_bps = strcasecmp (val, "bits") == 0;
1460 if (!opt.report_bps)
1461 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1462 return opt.report_bps;
1467 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1469 static const struct decode_item choices[] = {
1470 { "auto", secure_protocol_auto },
1471 { "sslv2", secure_protocol_sslv2 },
1472 { "sslv3", secure_protocol_sslv3 },
1473 { "tlsv1", secure_protocol_tlsv1 },
1475 int ok = decode_string (val, choices, countof (choices), place);
1477 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1482 /* Set all three timeout values. */
1485 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1488 if (!cmd_time (com, val, &value))
1490 opt.read_timeout = value;
1491 opt.connect_timeout = value;
1492 opt.dns_timeout = value;
1497 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1499 /* Disallow embedded newlines. */
1500 if (strchr (val, '\n'))
1502 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1503 exec_name, com, quote (val));
1506 xfree_null (opt.useragent);
1507 opt.useragent = xstrdup (val);
1511 /* The "verbose" option cannot be cmd_boolean because the variable is
1512 not bool -- it's of type int (-1 means uninitialized because of
1513 some random hackery for disallowing -q -v). */
1516 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1519 if (cmd_boolean (com, val, &flag))
1527 /* Miscellaneous useful routines. */
1529 /* A very simple atoi clone, more useful than atoi because it works on
1530 delimited strings, and has error reportage. Returns true on success,
1531 false on failure. If successful, stores result to *DEST. */
1534 simple_atoi (const char *beg, const char *end, int *dest)
1537 bool negative = false;
1538 const char *p = beg;
1540 while (p < end && c_isspace (*p))
1542 if (p < end && (*p == '-' || *p == '+'))
1544 negative = (*p == '-');
1550 /* Read negative numbers in a separate loop because the most
1551 negative integer cannot be represented as a positive number. */
1554 for (; p < end && c_isdigit (*p); p++)
1556 int next = (10 * result) + (*p - '0');
1558 return false; /* overflow */
1562 for (; p < end && c_isdigit (*p); p++)
1564 int next = (10 * result) - (*p - '0');
1566 return false; /* underflow */
1577 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1578 doesn't handle exponential notation. Returns true on success,
1579 false on failure. In case of success, stores its result to
1583 simple_atof (const char *beg, const char *end, double *dest)
1587 bool negative = false;
1588 bool seen_dot = false;
1589 bool seen_digit = false;
1592 const char *p = beg;
1594 while (p < end && c_isspace (*p))
1596 if (p < end && (*p == '-' || *p == '+'))
1598 negative = (*p == '-');
1602 for (; p < end; p++)
1608 result = (10 * result) + (ch - '0');
1610 result += (ch - '0') / (divider *= 10);
1632 /* Verify that the user-specified header in S is valid. It must
1633 contain a colon preceded by non-white-space characters and must not
1634 contain newlines. */
1637 check_user_specified_header (const char *s)
1641 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1643 /* The header MUST contain `:' preceded by at least one
1644 non-whitespace character. */
1645 if (*p != ':' || p == s)
1647 /* The header MUST NOT contain newlines. */
1648 if (strchr (s, '\n'))
1653 /* Decode VAL into a number, according to ITEMS. */
1656 decode_string (const char *val, const struct decode_item *items, int itemcount,
1660 for (i = 0; i < itemcount; i++)
1661 if (0 == strcasecmp (val, items[i].name))
1663 *place = items[i].code;
1670 void cleanup_html_url (void);
1673 /* Free the memory allocated by global variables. */
1677 /* Free external resources, close files, etc. */
1679 /* Close WARC file. */
1680 if (opt.warc_filename != 0)
1686 if (fclose (output_stream) == EOF)
1687 inform_exit_status (CLOSEFAILED);
1689 /* No need to check for error because Wget flushes its output (and
1690 checks for errors) after any data arrives. */
1692 /* We're exiting anyway so there's no real need to call free()
1693 hundreds of times. Skipping the frees will make Wget exit
1696 However, when detecting leaks, it's crucial to free() everything
1697 because then you can find the real leaks, i.e. the allocated
1698 memory which grows with the size of the program. */
1704 cleanup_html_url ();
1708 for (i = 0; i < nurl; i++)
1712 extern acc_t *netrc_list;
1713 free_netrc (netrc_list);
1715 xfree_null (opt.choose_config);
1716 xfree_null (opt.lfilename);
1717 xfree_null (opt.dir_prefix);
1718 xfree_null (opt.input_filename);
1719 xfree_null (opt.output_document);
1720 free_vec (opt.accepts);
1721 free_vec (opt.rejects);
1722 free_vec (opt.excludes);
1723 free_vec (opt.includes);
1724 free_vec (opt.domains);
1725 free_vec (opt.follow_tags);
1726 free_vec (opt.ignore_tags);
1727 xfree_null (opt.progress_type);
1728 xfree_null (opt.ftp_user);
1729 xfree_null (opt.ftp_passwd);
1730 xfree_null (opt.ftp_proxy);
1731 xfree_null (opt.https_proxy);
1732 xfree_null (opt.http_proxy);
1733 free_vec (opt.no_proxy);
1734 xfree_null (opt.useragent);
1735 xfree_null (opt.referer);
1736 xfree_null (opt.http_user);
1737 xfree_null (opt.http_passwd);
1738 free_vec (opt.user_headers);
1739 free_vec (opt.warc_user_headers);
1741 xfree_null (opt.cert_file);
1742 xfree_null (opt.private_key);
1743 xfree_null (opt.ca_directory);
1744 xfree_null (opt.ca_cert);
1745 xfree_null (opt.random_file);
1746 xfree_null (opt.egd_file);
1748 xfree_null (opt.bind_address);
1749 xfree_null (opt.cookies_input);
1750 xfree_null (opt.cookies_output);
1751 xfree_null (opt.user);
1752 xfree_null (opt.passwd);
1753 xfree_null (opt.base_href);
1755 #endif /* DEBUG_MALLOC */
1758 /* Unit testing routines. */
1763 test_commands_sorted()
1765 int prev_idx = 0, next_idx = 1;
1766 int command_count = countof (commands) - 1;
1768 while (next_idx <= command_count)
1770 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1773 mu_assert ("FAILED", false);
1786 test_cmd_spec_restrict_file_names()
1791 int expected_restrict_files_os;
1792 int expected_restrict_files_ctrl;
1793 int expected_restrict_files_case;
1796 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1797 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1798 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1799 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1802 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1807 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1810 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1811 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1812 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1813 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1815 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1816 res == test_array[i].result
1817 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1818 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1819 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1825 #endif /* TESTING */