1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
3 2005, 2006, 2007, 2008, 2009, 2010, 2011 Free Software Foundation,
6 This file is part of GNU Wget.
8 GNU Wget is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 3 of the License, or
11 (at your option) any later version.
13 GNU Wget is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with Wget. If not, see <http://www.gnu.org/licenses/>.
21 Additional permission under GNU GPL version 3 section 7
23 If you modify this program, or any covered work, by linking or
24 combining it with the OpenSSL project's OpenSSL library (or a
25 modified version of that library), containing parts covered by the
26 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
27 grants you additional permission to convey the resulting work.
28 Corresponding Source for a non-source form of such a combination
29 shall include the source code for the parts of OpenSSL used as well
30 as that of the covered work. */
41 /* not all systems provide PATH_MAX in limits.h */
43 # include <sys/param.h>
45 # define PATH_MAX MAXPATHLEN
60 #include "recur.h" /* for INFINITE_RECURSION */
61 #include "convert.h" /* for convert_cleanup */
62 #include "res.h" /* for res_cleanup */
63 #include "http.h" /* for http_cleanup */
64 #include "retr.h" /* for output_stream */
72 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
74 CMD_DECLARE (cmd_boolean);
75 CMD_DECLARE (cmd_bytes);
76 CMD_DECLARE (cmd_bytes_sum);
78 CMD_DECLARE (cmd_cert_type);
80 CMD_DECLARE (cmd_directory_vector);
81 CMD_DECLARE (cmd_number);
82 CMD_DECLARE (cmd_number_inf);
83 CMD_DECLARE (cmd_string);
84 CMD_DECLARE (cmd_file);
85 CMD_DECLARE (cmd_directory);
86 CMD_DECLARE (cmd_time);
87 CMD_DECLARE (cmd_vector);
89 CMD_DECLARE (cmd_spec_dirstruct);
90 CMD_DECLARE (cmd_spec_header);
91 CMD_DECLARE (cmd_spec_htmlify);
92 CMD_DECLARE (cmd_spec_mirror);
93 CMD_DECLARE (cmd_spec_prefer_family);
94 CMD_DECLARE (cmd_spec_progress);
95 CMD_DECLARE (cmd_spec_recursive);
96 CMD_DECLARE (cmd_spec_restrict_file_names);
98 CMD_DECLARE (cmd_spec_secure_protocol);
100 CMD_DECLARE (cmd_spec_timeout);
101 CMD_DECLARE (cmd_spec_useragent);
102 CMD_DECLARE (cmd_spec_verbose);
104 /* List of recognized commands, each consisting of name, place and
105 function. When adding a new command, simply add it to the list,
106 but be sure to keep the list sorted alphabetically, as
107 command_by_name's binary search depends on it. Also, be sure to
108 add any entries that allocate memory (e.g. cmd_string and
109 cmd_vector) to the cleanup() function below. */
111 static const struct {
114 bool (*action) (const char *, const char *, void *);
116 /* KEEP THIS LIST ALPHABETICALLY SORTED */
117 { "accept", &opt.accepts, cmd_vector },
118 { "addhostdir", &opt.add_hostdir, cmd_boolean },
119 { "adjustextension", &opt.adjust_extension, cmd_boolean },
120 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
121 { "askpassword", &opt.ask_passwd, cmd_boolean },
122 { "authnochallenge", &opt.auth_without_challenge,
124 { "background", &opt.background, cmd_boolean },
125 { "backupconverted", &opt.backup_converted, cmd_boolean },
126 { "backups", &opt.backups, cmd_number },
127 { "base", &opt.base_href, cmd_string },
128 { "bindaddress", &opt.bind_address, cmd_string },
130 { "cacertificate", &opt.ca_cert, cmd_file },
132 { "cache", &opt.allow_cache, cmd_boolean },
134 { "cadirectory", &opt.ca_directory, cmd_directory },
135 { "certificate", &opt.cert_file, cmd_file },
136 { "certificatetype", &opt.cert_type, cmd_cert_type },
137 { "checkcertificate", &opt.check_cert, cmd_boolean },
139 { "chooseconfig", &opt.choose_config, cmd_file },
140 { "connecttimeout", &opt.connect_timeout, cmd_time },
141 { "contentdisposition", &opt.content_disposition, cmd_boolean },
142 { "continue", &opt.always_rest, cmd_boolean },
143 { "convertlinks", &opt.convert_links, cmd_boolean },
144 { "cookies", &opt.cookies, cmd_boolean },
145 { "cutdirs", &opt.cut_dirs, cmd_number },
147 { "debug", &opt.debug, cmd_boolean },
149 { "defaultpage", &opt.default_page, cmd_string},
150 { "deleteafter", &opt.delete_after, cmd_boolean },
151 { "dirprefix", &opt.dir_prefix, cmd_directory },
152 { "dirstruct", NULL, cmd_spec_dirstruct },
153 { "dnscache", &opt.dns_cache, cmd_boolean },
154 { "dnstimeout", &opt.dns_timeout, cmd_time },
155 { "domains", &opt.domains, cmd_vector },
156 { "dotbytes", &opt.dot_bytes, cmd_bytes },
157 { "dotsinline", &opt.dots_in_line, cmd_number },
158 { "dotspacing", &opt.dot_spacing, cmd_number },
159 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
161 { "egdfile", &opt.egd_file, cmd_file },
163 { "excludedirectories", &opt.excludes, cmd_directory_vector },
164 { "excludedomains", &opt.exclude_domains, cmd_vector },
165 { "followftp", &opt.follow_ftp, cmd_boolean },
166 { "followtags", &opt.follow_tags, cmd_vector },
167 { "forcehtml", &opt.force_html, cmd_boolean },
168 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
169 { "ftppassword", &opt.ftp_passwd, cmd_string },
170 { "ftpproxy", &opt.ftp_proxy, cmd_string },
172 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
173 #endif /* def __VMS */
174 { "ftpuser", &opt.ftp_user, cmd_string },
175 { "glob", &opt.ftp_glob, cmd_boolean },
176 { "header", NULL, cmd_spec_header },
177 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
178 { "htmlify", NULL, cmd_spec_htmlify },
179 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
180 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
181 { "httppassword", &opt.http_passwd, cmd_string },
182 { "httpproxy", &opt.http_proxy, cmd_string },
183 { "httpsproxy", &opt.https_proxy, cmd_string },
184 { "httpuser", &opt.http_user, cmd_string },
185 { "ignorecase", &opt.ignore_case, cmd_boolean },
186 { "ignorelength", &opt.ignore_length, cmd_boolean },
187 { "ignoretags", &opt.ignore_tags, cmd_vector },
188 { "includedirectories", &opt.includes, cmd_directory_vector },
190 { "inet4only", &opt.ipv4_only, cmd_boolean },
191 { "inet6only", &opt.ipv6_only, cmd_boolean },
193 { "input", &opt.input_filename, cmd_file },
194 { "iri", &opt.enable_iri, cmd_boolean },
195 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
196 { "limitrate", &opt.limit_rate, cmd_bytes },
197 { "loadcookies", &opt.cookies_input, cmd_file },
198 { "localencoding", &opt.locale, cmd_string },
199 { "logfile", &opt.lfilename, cmd_file },
200 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
201 { "maxredirect", &opt.max_redirect, cmd_number },
202 { "mirror", NULL, cmd_spec_mirror },
203 { "netrc", &opt.netrc, cmd_boolean },
204 { "noclobber", &opt.noclobber, cmd_boolean },
205 { "noparent", &opt.no_parent, cmd_boolean },
206 { "noproxy", &opt.no_proxy, cmd_vector },
207 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
208 { "outputdocument", &opt.output_document, cmd_file },
209 { "pagerequisites", &opt.page_requisites, cmd_boolean },
210 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
211 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
212 { "password", &opt.passwd, cmd_string },
213 { "postdata", &opt.post_data, cmd_string },
214 { "postfile", &opt.post_file_name, cmd_file },
215 { "preferfamily", NULL, cmd_spec_prefer_family },
216 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
218 { "privatekey", &opt.private_key, cmd_file },
219 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
221 { "progress", &opt.progress_type, cmd_spec_progress },
222 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
223 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
224 { "proxypassword", &opt.proxy_passwd, cmd_string },
225 { "proxyuser", &opt.proxy_user, cmd_string },
226 { "quiet", &opt.quiet, cmd_boolean },
227 { "quota", &opt.quota, cmd_bytes_sum },
229 { "randomfile", &opt.random_file, cmd_file },
231 { "randomwait", &opt.random_wait, cmd_boolean },
232 { "readtimeout", &opt.read_timeout, cmd_time },
233 { "reclevel", &opt.reclevel, cmd_number_inf },
234 { "recursive", NULL, cmd_spec_recursive },
235 { "referer", &opt.referer, cmd_string },
236 { "reject", &opt.rejects, cmd_vector },
237 { "relativeonly", &opt.relative_only, cmd_boolean },
238 { "remoteencoding", &opt.encoding_remote, cmd_string },
239 { "removelisting", &opt.remove_listing, cmd_boolean },
240 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
241 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
242 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
243 { "robots", &opt.use_robots, cmd_boolean },
244 { "savecookies", &opt.cookies_output, cmd_file },
245 { "saveheaders", &opt.save_headers, cmd_boolean },
247 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
249 { "serverresponse", &opt.server_response, cmd_boolean },
250 { "spanhosts", &opt.spanhost, cmd_boolean },
251 { "spider", &opt.spider, cmd_boolean },
252 { "strictcomments", &opt.strict_comments, cmd_boolean },
253 { "timeout", NULL, cmd_spec_timeout },
254 { "timestamping", &opt.timestamping, cmd_boolean },
255 { "tries", &opt.ntry, cmd_number_inf },
256 { "trustservernames", &opt.trustservernames, cmd_boolean },
257 { "unlink", &opt.unlink, cmd_boolean },
258 { "useproxy", &opt.use_proxy, cmd_boolean },
259 { "user", &opt.user, cmd_string },
260 { "useragent", NULL, cmd_spec_useragent },
261 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
262 { "verbose", NULL, cmd_spec_verbose },
263 { "wait", &opt.wait, cmd_time },
264 { "waitretry", &opt.waitretry, cmd_time },
266 { "wdebug", &opt.wdebug, cmd_boolean },
270 /* Look up CMDNAME in the commands[] and return its position in the
271 array. If CMDNAME is not found, return -1. */
274 command_by_name (const char *cmdname)
276 /* Use binary search for speed. Wget has ~100 commands, which
277 guarantees a worst case performance of 7 string comparisons. */
278 int lo = 0, hi = countof (commands) - 1;
282 int mid = (lo + hi) >> 1;
283 int cmp = strcasecmp (cmdname, commands[mid].name);
294 /* Reset the variables to default values. */
300 /* Most of the default values are 0 (and 0.0, NULL, and false).
301 Just reset everything, and fill in the non-zero values. Note
302 that initializing pointers to NULL this way is technically
303 illegal, but porting Wget to a machine where NULL is not all-zero
304 bit pattern will be the least of the implementors' worries. */
311 opt.add_hostdir = true;
315 opt.http_keep_alive = true;
316 opt.use_proxy = true;
317 tmp = getenv ("no_proxy");
319 opt.no_proxy = sepstring (tmp);
320 opt.prefer_family = prefer_none;
321 opt.allow_cache = true;
323 opt.read_timeout = 900;
324 opt.use_robots = true;
326 opt.remove_listing = true;
328 opt.dot_bytes = 1024;
329 opt.dot_spacing = 10;
330 opt.dots_in_line = 50;
332 opt.dns_cache = true;
336 opt.check_cert = true;
339 /* The default for file name restriction defaults to the OS type. */
340 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
341 opt.restrict_files_os = restrict_windows;
343 opt.restrict_files_os = restrict_unix;
345 opt.restrict_files_ctrl = true;
346 opt.restrict_files_nonascii = false;
347 opt.restrict_files_case = restrict_no_case_restriction;
349 opt.max_redirect = 20;
354 opt.enable_iri = true;
356 opt.enable_iri = false;
359 opt.encoding_remote = NULL;
361 opt.useservertimestamps = true;
364 /* Return the user's home directory (strdup-ed), or NULL if none is
369 static char buf[PATH_MAX];
374 home = getenv ("HOME");
378 /* Under MSDOS, if $HOME isn't defined, use the directory where
379 `wget.exe' resides. */
380 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
383 strcpy (buf, _w32_get_argv0 ());
384 p = strrchr (buf, '/'); /* djgpp */
386 p = strrchr (buf, '\\'); /* others */
390 #elif !defined(WINDOWS)
391 /* If HOME is not defined, try getting it from the password
393 struct passwd *pwd = getpwuid (getuid ());
394 if (!pwd || !pwd->pw_dir)
396 strcpy (buf, pwd->pw_dir);
399 /* Under Windows, if $HOME isn't defined, use the directory where
400 `wget.exe' resides. */
406 return home ? xstrdup (home) : NULL;
409 /* Check the 'WGETRC' environment variable and return the file name
410 if 'WGETRC' is set and is a valid file.
411 If the `WGETRC' variable exists but the file does not exist, the
412 function will exit(). */
414 wgetrc_env_file_name (void)
416 char *env = getenv ("WGETRC");
419 if (!file_exists_p (env))
421 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
425 return xstrdup (env);
430 /* Check for the existance of '$HOME/.wgetrc' and return it's path
431 if it exists and is set. */
433 wgetrc_user_file_name (void)
437 /* If that failed, try $HOME/.wgetrc (or equivalent). */
440 file = "SYS$LOGIN:.wgetrc";
441 #else /* def __VMS */
444 file = aprintf ("%s/.wgetrc", home);
446 #endif /* def __VMS [else] */
450 if (!file_exists_p (file))
458 /* Return the path to the user's .wgetrc. This is either the value of
459 `WGETRC' environment variable, or `$HOME/.wgetrc'.
461 Additionally, for windows, look in the directory where wget.exe
464 wgetrc_file_name (void)
466 char *file = wgetrc_env_file_name ();
470 file = wgetrc_user_file_name ();
473 /* Under Windows, if we still haven't found .wgetrc, look for the file
474 `wget.ini' in the directory where `wget.exe' resides; we do this for
475 backward compatibility with previous versions of Wget.
476 SYSTEM_WGETRC should not be defined under WINDOWS. */
479 char *home = home_dir ();
485 file = aprintf ("%s/wget.ini", home);
486 if (!file_exists_p (file))
499 /* Return values of parse_line. */
507 static enum parse_line parse_line (const char *, char **, char **, int *);
508 static bool setval_internal (int, const char *, const char *);
509 static bool setval_internal_tilde (int, const char *, const char *);
511 /* Initialize variables from a wgetrc file. Returns zero (failure) if
512 there were errors in the file. */
515 run_wgetrc (const char *file)
522 fp = fopen (file, "r");
525 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
526 file, strerror (errno));
527 return true; /* not a fatal error */
530 while ((line = read_whole_line (fp)) != NULL)
532 char *com = NULL, *val = NULL;
535 /* Parse the line. */
536 switch (parse_line (line, &com, &val, &comind))
539 /* If everything is OK, set the value. */
540 if (!setval_internal_tilde (comind, com, val))
542 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
543 exec_name, file, ln);
547 case line_syntax_error:
548 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
549 exec_name, file, ln);
552 case line_unknown_command:
553 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
554 exec_name, quote (com), file, ln);
572 /* Initialize the defaults and run the system wgetrc and user's own
577 char *file, *env_sysrc;
580 /* Run a non-standard system rc file when the according environment
581 variable has been set. For internal testing purposes only! */
582 env_sysrc = getenv ("SYSTEM_WGETRC");
583 if (env_sysrc && file_exists_p (env_sysrc))
584 ok &= run_wgetrc (env_sysrc);
585 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
587 else if (file_exists_p (SYSTEM_WGETRC))
588 ok &= run_wgetrc (SYSTEM_WGETRC);
590 /* If there are any problems parsing the system wgetrc file, tell
594 fprintf (stderr, _("\
595 Parsing system wgetrc file failed, please check '%s'. \
596 Or specify a different file using --config\n"), SYSTEM_WGETRC);
599 /* Override it with your own, if one exists. */
600 file = wgetrc_file_name ();
603 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
604 something like realpath() before comparing them with `strcmp' */
606 if (!strcmp (file, SYSTEM_WGETRC))
608 fprintf (stderr, _("\
609 %s: Warning: Both system and user wgetrc point to %s.\n"),
610 exec_name, quote (file));
614 ok &= run_wgetrc (file);
616 /* If there were errors processing either `.wgetrc', abort. */
624 /* Remove dashes and underscores from S, modifying S in the
630 char *t = s; /* t - tortoise */
631 char *h = s; /* h - hare */
633 if (*h == '_' || *h == '-')
640 /* Parse the line pointed by line, with the syntax:
641 <sp>* command <sp>* = <sp>* value <sp>*
642 Uses malloc to allocate space for command and value.
644 Returns one of line_ok, line_empty, line_syntax_error, or
645 line_unknown_command.
647 In case of line_ok, *COM and *VAL point to freshly allocated
648 strings, and *COMIND points to com's index. In case of error or
649 empty line, their values are unmodified. */
651 static enum parse_line
652 parse_line (const char *line, char **com, char **val, int *comind)
655 const char *end = line + strlen (line);
656 const char *cmdstart, *cmdend;
657 const char *valstart, *valend;
662 /* Skip leading and trailing whitespace. */
663 while (*line && c_isspace (*line))
665 while (end > line && c_isspace (end[-1]))
668 /* Skip empty lines and comments. */
669 if (!*line || *line == '#')
675 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
679 /* Skip '=', as well as any space before or after it. */
680 while (p < end && c_isspace (*p))
682 if (p == end || *p != '=')
683 return line_syntax_error;
685 while (p < end && c_isspace (*p))
691 /* The syntax is valid (even though the command might not be). Fill
692 in the command name and value. */
693 *com = strdupdelim (cmdstart, cmdend);
694 *val = strdupdelim (valstart, valend);
696 /* The line now known to be syntactically correct. Check whether
697 the command is valid. */
698 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
700 ind = command_by_name (cmdcopy);
702 return line_unknown_command;
704 /* Report success to the caller. */
709 #if defined(WINDOWS) || defined(MSDOS)
710 # define ISSEP(c) ((c) == '/' || (c) == '\\')
712 # define ISSEP(c) ((c) == '/')
715 /* Run commands[comind].action. */
718 setval_internal (int comind, const char *com, const char *val)
720 assert (0 <= comind && ((size_t) comind) < countof (commands));
721 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
722 return commands[comind].action (com, val, commands[comind].place);
726 setval_internal_tilde (int comind, const char *com, const char *val)
732 ret = setval_internal (comind, com, val);
734 /* We make tilde expansion for cmd_file and cmd_directory */
735 if (((commands[comind].action == cmd_file) ||
736 (commands[comind].action == cmd_directory))
737 && ret && (*val == '~' && ISSEP (val[1])))
739 pstring = commands[comind].place;
743 homelen = strlen (home);
744 while (homelen && ISSEP (home[homelen - 1]))
745 home[--homelen] = '\0';
747 /* Skip the leading "~/". */
748 for (++val; ISSEP (*val); val++)
750 *pstring = concat_strings (home, "/", val, (char *)0);
756 /* Run command COM with value VAL. If running the command produces an
757 error, report the error and exit.
759 This is intended to be called from main() to modify Wget's behavior
760 through command-line switches. Since COM is hard-coded in main(),
761 it is not canonicalized, and this aborts when COM is not found.
763 If COMIND's are exported to init.h, this function will be changed
764 to accept COMIND directly. */
767 setoptval (const char *com, const char *val, const char *optname)
769 /* Prepend "--" to OPTNAME. */
770 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
773 strcpy (dd_optname + 2, optname);
775 assert (val != NULL);
776 if (!setval_internal (command_by_name (com), dd_optname, val))
780 /* Parse OPT into command and value and run it. For example,
781 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
782 This is used by the `--execute' flag in main.c. */
785 run_command (const char *opt)
789 switch (parse_line (opt, &com, &val, &comind))
792 if (!setval_internal (comind, com, val))
798 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
799 exec_name, quote (opt));
804 /* Generic helper functions, for use with `commands'. */
806 /* Forward declarations: */
811 static bool decode_string (const char *, const struct decode_item *, int, int *);
812 static bool simple_atoi (const char *, const char *, int *);
813 static bool simple_atof (const char *, const char *, double *);
815 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
817 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
818 && c_tolower((p)[1]) == (c1) \
821 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
822 && c_tolower((p)[1]) == (c1) \
823 && c_tolower((p)[2]) == (c2) \
827 /* Store the boolean value from VAL to PLACE. COM is ignored,
828 except for error messages. */
830 cmd_boolean (const char *com, const char *val, void *place)
834 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
835 /* "on", "yes" and "1" mean true. */
837 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
838 /* "off", "no" and "0" mean false. */
843 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
844 exec_name, com, quote (val));
848 *(bool *) place = value;
852 /* Set the non-negative integer value from VAL to PLACE. With
853 incorrect specification, the number remains unchanged. */
855 cmd_number (const char *com, const char *val, void *place)
857 if (!simple_atoi (val, val + strlen (val), place)
858 || *(int *) place < 0)
860 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
861 exec_name, com, quote (val));
867 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
869 cmd_number_inf (const char *com, const char *val, void *place)
871 if (!strcasecmp (val, "inf"))
876 return cmd_number (com, val, place);
879 /* Copy (strdup) the string at COM to a new location and place a
880 pointer to *PLACE. */
882 cmd_string (const char *com, const char *val, void *place)
884 char **pstring = (char **)place;
886 xfree_null (*pstring);
887 *pstring = xstrdup (val);
892 /* Like the above, but handles tilde-expansion when reading a user's
893 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
894 gets expanded to the user's home directory. */
896 cmd_file (const char *com, const char *val, void *place)
898 char **pstring = (char **)place;
900 xfree_null (*pstring);
902 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
904 *pstring = xstrdup (val);
906 #if defined(WINDOWS) || defined(MSDOS)
907 /* Convert "\" to "/". */
910 for (s = *pstring; *s; s++)
918 /* Like cmd_file, but strips trailing '/' characters. */
920 cmd_directory (const char *com, const char *val, void *place)
924 /* Call cmd_file() for tilde expansion and separator
925 canonicalization (backslash -> slash under Windows). These
926 things should perhaps be in a separate function. */
927 if (!cmd_file (com, val, place))
932 while (t > s && *--t == '/')
938 /* Split VAL by space to a vector of values, and append those values
939 to vector pointed to by the PLACE argument. If VAL is empty, the
940 PLACE vector is cleared instead. */
943 cmd_vector (const char *com, const char *val, void *place)
945 char ***pvec = (char ***)place;
948 *pvec = merge_vecs (*pvec, sepstring (val));
958 cmd_directory_vector (const char *com, const char *val, void *place)
960 char ***pvec = (char ***)place;
964 /* Strip the trailing slashes from directories. */
967 seps = sepstring (val);
968 for (t = seps; t && *t; t++)
970 int len = strlen (*t);
971 /* Skip degenerate case of root directory. */
974 if ((*t)[len - 1] == '/')
975 (*t)[len - 1] = '\0';
978 *pvec = merge_vecs (*pvec, seps);
988 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
989 "100k" or "2.5G" to a floating point number. */
992 parse_bytes_helper (const char *val, double *result)
995 const char *end = val + strlen (val);
997 /* Check for "inf". */
998 if (0 == strcmp (val, "inf"))
1004 /* Strip trailing whitespace. */
1005 while (val < end && c_isspace (end[-1]))
1010 switch (c_tolower (end[-1]))
1013 --end, mult = 1024.0;
1016 --end, mult = 1048576.0;
1019 --end, mult = 1073741824.0;
1022 --end, mult = 1099511627776.0;
1025 /* Not a recognized suffix: assume it's a digit. (If not,
1026 simple_atof will raise an error.) */
1030 /* Skip leading and trailing whitespace. */
1031 while (val < end && c_isspace (*val))
1033 while (val < end && c_isspace (end[-1]))
1038 if (!simple_atof (val, end, &number) || number < 0)
1041 *result = number * mult;
1045 /* Parse VAL as a number and set its value to PLACE (which should
1048 By default, the value is assumed to be in bytes. If "K", "M", or
1049 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1050 1<<30, respectively. Floating point values are allowed and are
1051 cast to integer before use. The idea is to be able to use things
1052 like 1.5k instead of "1536".
1054 The string "inf" is returned as 0.
1056 In case of error, false is returned and memory pointed to by PLACE
1057 remains unmodified. */
1060 cmd_bytes (const char *com, const char *val, void *place)
1063 if (!parse_bytes_helper (val, &byte_value))
1065 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1066 exec_name, com, quote (val));
1069 *(wgint *)place = (wgint)byte_value;
1073 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1074 SIZE_SUM. It works by converting the string to double, therefore
1075 working with values up to 2^53-1 without loss of precision. This
1076 value (8192 TB) is large enough to serve for a while. */
1079 cmd_bytes_sum (const char *com, const char *val, void *place)
1082 if (!parse_bytes_helper (val, &byte_value))
1084 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1085 exec_name, com, quote (val));
1088 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1092 /* Store the value of VAL to *OUT. The value is a time period, by
1093 default expressed in seconds, but also accepting suffixes "m", "h",
1094 "d", and "w" for minutes, hours, days, and weeks respectively. */
1097 cmd_time (const char *com, const char *val, void *place)
1099 double number, mult;
1100 const char *end = val + strlen (val);
1102 /* Strip trailing whitespace. */
1103 while (val < end && c_isspace (end[-1]))
1109 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1110 exec_name, com, quote (val));
1114 switch (c_tolower (end[-1]))
1117 --end, mult = 1; /* seconds */
1120 --end, mult = 60; /* minutes */
1123 --end, mult = 3600; /* hours */
1126 --end, mult = 86400.0; /* days */
1129 --end, mult = 604800.0; /* weeks */
1132 /* Not a recognized suffix: assume it belongs to the number.
1133 (If not, simple_atof will raise an error.) */
1137 /* Skip leading and trailing whitespace. */
1138 while (val < end && c_isspace (*val))
1140 while (val < end && c_isspace (end[-1]))
1145 if (!simple_atof (val, end, &number))
1148 *(double *)place = number * mult;
1154 cmd_cert_type (const char *com, const char *val, void *place)
1156 static const struct decode_item choices[] = {
1157 { "pem", keyfile_pem },
1158 { "der", keyfile_asn1 },
1159 { "asn1", keyfile_asn1 },
1161 int ok = decode_string (val, choices, countof (choices), place);
1163 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1168 /* Specialized helper functions, used by `commands' to handle some
1169 options specially. */
1171 static bool check_user_specified_header (const char *);
1174 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1176 if (!cmd_boolean (com, val, &opt.dirstruct))
1178 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1179 must be affected inversely. */
1181 opt.no_dirstruct = false;
1183 opt.no_dirstruct = true;
1188 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1190 /* Empty value means reset the list of headers. */
1193 free_vec (opt.user_headers);
1194 opt.user_headers = NULL;
1198 if (!check_user_specified_header (val))
1200 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1201 exec_name, com, quote (val));
1204 opt.user_headers = vec_append (opt.user_headers, val);
1209 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1211 int flag = cmd_boolean (com, val, &opt.htmlify);
1212 if (flag && !opt.htmlify)
1213 opt.remove_listing = false;
1217 /* Set the "mirror" mode. It means: recursive download, timestamping,
1218 no limit on max. recursion depth, and don't remove listings. */
1221 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1225 if (!cmd_boolean (com, val, &mirror))
1229 opt.recursive = true;
1230 if (!opt.no_dirstruct)
1231 opt.dirstruct = true;
1232 opt.timestamping = true;
1233 opt.reclevel = INFINITE_RECURSION;
1234 opt.remove_listing = false;
1239 /* Validate --prefer-family and set the choice. Allowed values are
1240 "IPv4", "IPv6", and "none". */
1243 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1245 static const struct decode_item choices[] = {
1246 { "IPv4", prefer_ipv4 },
1247 { "IPv6", prefer_ipv6 },
1248 { "none", prefer_none },
1250 int prefer_family = prefer_none;
1251 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1253 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1254 opt.prefer_family = prefer_family;
1258 /* Set progress.type to VAL, but verify that it's a valid progress
1259 implementation before that. */
1262 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1264 if (!valid_progress_implementation_p (val))
1266 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1267 exec_name, com, quote (val));
1270 xfree_null (opt.progress_type);
1272 /* Don't call set_progress_implementation here. It will be called
1273 in main() when it becomes clear what the log output is. */
1274 opt.progress_type = xstrdup (val);
1278 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1279 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1283 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1285 if (!cmd_boolean (com, val, &opt.recursive))
1289 if (opt.recursive && !opt.no_dirstruct)
1290 opt.dirstruct = true;
1296 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1298 int restrict_os = opt.restrict_files_os;
1299 int restrict_ctrl = opt.restrict_files_ctrl;
1300 int restrict_case = opt.restrict_files_case;
1301 int restrict_nonascii = opt.restrict_files_nonascii;
1305 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1309 end = strchr (val, ',');
1311 end = val + strlen (val);
1313 if (VAL_IS ("unix"))
1314 restrict_os = restrict_unix;
1315 else if (VAL_IS ("windows"))
1316 restrict_os = restrict_windows;
1317 else if (VAL_IS ("lowercase"))
1318 restrict_case = restrict_lowercase;
1319 else if (VAL_IS ("uppercase"))
1320 restrict_case = restrict_uppercase;
1321 else if (VAL_IS ("nocontrol"))
1322 restrict_ctrl = false;
1323 else if (VAL_IS ("ascii"))
1324 restrict_nonascii = true;
1327 fprintf (stderr, _("\
1328 %s: %s: Invalid restriction %s,\n\
1329 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1330 exec_name, com, quote (val));
1337 while (*val && *end);
1341 opt.restrict_files_os = restrict_os;
1342 opt.restrict_files_ctrl = restrict_ctrl;
1343 opt.restrict_files_case = restrict_case;
1344 opt.restrict_files_nonascii = restrict_nonascii;
1351 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1353 static const struct decode_item choices[] = {
1354 { "auto", secure_protocol_auto },
1355 { "sslv2", secure_protocol_sslv2 },
1356 { "sslv3", secure_protocol_sslv3 },
1357 { "tlsv1", secure_protocol_tlsv1 },
1359 int ok = decode_string (val, choices, countof (choices), place);
1361 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1366 /* Set all three timeout values. */
1369 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1372 if (!cmd_time (com, val, &value))
1374 opt.read_timeout = value;
1375 opt.connect_timeout = value;
1376 opt.dns_timeout = value;
1381 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1383 /* Disallow embedded newlines. */
1384 if (strchr (val, '\n'))
1386 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1387 exec_name, com, quote (val));
1390 xfree_null (opt.useragent);
1391 opt.useragent = xstrdup (val);
1395 /* The "verbose" option cannot be cmd_boolean because the variable is
1396 not bool -- it's of type int (-1 means uninitialized because of
1397 some random hackery for disallowing -q -v). */
1400 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1403 if (cmd_boolean (com, val, &flag))
1411 /* Miscellaneous useful routines. */
1413 /* A very simple atoi clone, more useful than atoi because it works on
1414 delimited strings, and has error reportage. Returns true on success,
1415 false on failure. If successful, stores result to *DEST. */
1418 simple_atoi (const char *beg, const char *end, int *dest)
1421 bool negative = false;
1422 const char *p = beg;
1424 while (p < end && c_isspace (*p))
1426 if (p < end && (*p == '-' || *p == '+'))
1428 negative = (*p == '-');
1434 /* Read negative numbers in a separate loop because the most
1435 negative integer cannot be represented as a positive number. */
1438 for (; p < end && c_isdigit (*p); p++)
1440 int next = (10 * result) + (*p - '0');
1442 return false; /* overflow */
1446 for (; p < end && c_isdigit (*p); p++)
1448 int next = (10 * result) - (*p - '0');
1450 return false; /* underflow */
1461 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1462 doesn't handle exponential notation. Returns true on success,
1463 false on failure. In case of success, stores its result to
1467 simple_atof (const char *beg, const char *end, double *dest)
1471 bool negative = false;
1472 bool seen_dot = false;
1473 bool seen_digit = false;
1476 const char *p = beg;
1478 while (p < end && c_isspace (*p))
1480 if (p < end && (*p == '-' || *p == '+'))
1482 negative = (*p == '-');
1486 for (; p < end; p++)
1492 result = (10 * result) + (ch - '0');
1494 result += (ch - '0') / (divider *= 10);
1516 /* Verify that the user-specified header in S is valid. It must
1517 contain a colon preceded by non-white-space characters and must not
1518 contain newlines. */
1521 check_user_specified_header (const char *s)
1525 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1527 /* The header MUST contain `:' preceded by at least one
1528 non-whitespace character. */
1529 if (*p != ':' || p == s)
1531 /* The header MUST NOT contain newlines. */
1532 if (strchr (s, '\n'))
1537 /* Decode VAL into a number, according to ITEMS. */
1540 decode_string (const char *val, const struct decode_item *items, int itemcount,
1544 for (i = 0; i < itemcount; i++)
1545 if (0 == strcasecmp (val, items[i].name))
1547 *place = items[i].code;
1554 void cleanup_html_url (void);
1557 /* Free the memory allocated by global variables. */
1561 /* Free external resources, close files, etc. */
1564 fclose (output_stream);
1565 /* No need to check for error because Wget flushes its output (and
1566 checks for errors) after any data arrives. */
1568 /* We're exiting anyway so there's no real need to call free()
1569 hundreds of times. Skipping the frees will make Wget exit
1572 However, when detecting leaks, it's crucial to free() everything
1573 because then you can find the real leaks, i.e. the allocated
1574 memory which grows with the size of the program. */
1580 cleanup_html_url ();
1585 extern acc_t *netrc_list;
1586 free_netrc (netrc_list);
1588 xfree_null (opt.choose_config);
1589 xfree_null (opt.lfilename);
1590 xfree_null (opt.dir_prefix);
1591 xfree_null (opt.input_filename);
1592 xfree_null (opt.output_document);
1593 free_vec (opt.accepts);
1594 free_vec (opt.rejects);
1595 free_vec (opt.excludes);
1596 free_vec (opt.includes);
1597 free_vec (opt.domains);
1598 free_vec (opt.follow_tags);
1599 free_vec (opt.ignore_tags);
1600 xfree_null (opt.progress_type);
1601 xfree_null (opt.ftp_user);
1602 xfree_null (opt.ftp_passwd);
1603 xfree_null (opt.ftp_proxy);
1604 xfree_null (opt.https_proxy);
1605 xfree_null (opt.http_proxy);
1606 free_vec (opt.no_proxy);
1607 xfree_null (opt.useragent);
1608 xfree_null (opt.referer);
1609 xfree_null (opt.http_user);
1610 xfree_null (opt.http_passwd);
1611 free_vec (opt.user_headers);
1613 xfree_null (opt.cert_file);
1614 xfree_null (opt.private_key);
1615 xfree_null (opt.ca_directory);
1616 xfree_null (opt.ca_cert);
1617 xfree_null (opt.random_file);
1618 xfree_null (opt.egd_file);
1620 xfree_null (opt.bind_address);
1621 xfree_null (opt.cookies_input);
1622 xfree_null (opt.cookies_output);
1623 xfree_null (opt.user);
1624 xfree_null (opt.passwd);
1625 xfree_null (opt.base_href);
1627 #endif /* DEBUG_MALLOC */
1630 /* Unit testing routines. */
1635 test_commands_sorted()
1637 int prev_idx = 0, next_idx = 1;
1638 int command_count = countof (commands) - 1;
1640 while (next_idx <= command_count)
1642 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1645 mu_assert ("FAILED", false);
1658 test_cmd_spec_restrict_file_names()
1663 int expected_restrict_files_os;
1664 int expected_restrict_files_ctrl;
1665 int expected_restrict_files_case;
1668 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1669 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1670 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1671 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1674 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1679 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1682 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1683 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1684 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1685 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1687 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1688 res == test_array[i].result
1689 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1690 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1691 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1697 #endif /* TESTING */