1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
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 { "connecttimeout", &opt.connect_timeout, cmd_time },
140 { "contentdisposition", &opt.content_disposition, cmd_boolean },
141 { "continue", &opt.always_rest, cmd_boolean },
142 { "convertlinks", &opt.convert_links, cmd_boolean },
143 { "cookies", &opt.cookies, cmd_boolean },
144 { "cutdirs", &opt.cut_dirs, cmd_number },
146 { "debug", &opt.debug, cmd_boolean },
148 { "defaultpage", &opt.default_page, cmd_string},
149 { "deleteafter", &opt.delete_after, cmd_boolean },
150 { "dirprefix", &opt.dir_prefix, cmd_directory },
151 { "dirstruct", NULL, cmd_spec_dirstruct },
152 { "dnscache", &opt.dns_cache, cmd_boolean },
153 { "dnstimeout", &opt.dns_timeout, cmd_time },
154 { "domains", &opt.domains, cmd_vector },
155 { "dotbytes", &opt.dot_bytes, cmd_bytes },
156 { "dotsinline", &opt.dots_in_line, cmd_number },
157 { "dotspacing", &opt.dot_spacing, cmd_number },
158 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
160 { "egdfile", &opt.egd_file, cmd_file },
162 { "excludedirectories", &opt.excludes, cmd_directory_vector },
163 { "excludedomains", &opt.exclude_domains, cmd_vector },
164 { "followftp", &opt.follow_ftp, cmd_boolean },
165 { "followtags", &opt.follow_tags, cmd_vector },
166 { "forcehtml", &opt.force_html, cmd_boolean },
167 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
168 { "ftppassword", &opt.ftp_passwd, cmd_string },
169 { "ftpproxy", &opt.ftp_proxy, cmd_string },
171 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
172 #endif /* def __VMS */
173 { "ftpuser", &opt.ftp_user, cmd_string },
174 { "glob", &opt.ftp_glob, cmd_boolean },
175 { "header", NULL, cmd_spec_header },
176 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
177 { "htmlify", NULL, cmd_spec_htmlify },
178 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
179 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
180 { "httppassword", &opt.http_passwd, cmd_string },
181 { "httpproxy", &opt.http_proxy, cmd_string },
182 { "httpsproxy", &opt.https_proxy, cmd_string },
183 { "httpuser", &opt.http_user, cmd_string },
184 { "ignorecase", &opt.ignore_case, cmd_boolean },
185 { "ignorelength", &opt.ignore_length, cmd_boolean },
186 { "ignoretags", &opt.ignore_tags, cmd_vector },
187 { "includedirectories", &opt.includes, cmd_directory_vector },
189 { "inet4only", &opt.ipv4_only, cmd_boolean },
190 { "inet6only", &opt.ipv6_only, cmd_boolean },
192 { "input", &opt.input_filename, cmd_file },
193 { "iri", &opt.enable_iri, cmd_boolean },
194 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
195 { "limitrate", &opt.limit_rate, cmd_bytes },
196 { "loadcookies", &opt.cookies_input, cmd_file },
197 { "localencoding", &opt.locale, cmd_string },
198 { "logfile", &opt.lfilename, cmd_file },
199 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
200 { "maxredirect", &opt.max_redirect, cmd_number },
201 { "mirror", NULL, cmd_spec_mirror },
202 { "netrc", &opt.netrc, cmd_boolean },
203 { "noclobber", &opt.noclobber, cmd_boolean },
204 { "noparent", &opt.no_parent, cmd_boolean },
205 { "noproxy", &opt.no_proxy, cmd_vector },
206 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
207 { "outputdocument", &opt.output_document, cmd_file },
208 { "pagerequisites", &opt.page_requisites, cmd_boolean },
209 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
210 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
211 { "password", &opt.passwd, cmd_string },
212 { "postdata", &opt.post_data, cmd_string },
213 { "postfile", &opt.post_file_name, cmd_file },
214 { "preferfamily", NULL, cmd_spec_prefer_family },
215 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
217 { "privatekey", &opt.private_key, cmd_file },
218 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
220 { "progress", &opt.progress_type, cmd_spec_progress },
221 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
222 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
223 { "proxypassword", &opt.proxy_passwd, cmd_string },
224 { "proxyuser", &opt.proxy_user, cmd_string },
225 { "quiet", &opt.quiet, cmd_boolean },
226 { "quota", &opt.quota, cmd_bytes_sum },
228 { "randomfile", &opt.random_file, cmd_file },
230 { "randomwait", &opt.random_wait, cmd_boolean },
231 { "readtimeout", &opt.read_timeout, cmd_time },
232 { "reclevel", &opt.reclevel, cmd_number_inf },
233 { "recursive", NULL, cmd_spec_recursive },
234 { "referer", &opt.referer, cmd_string },
235 { "reject", &opt.rejects, cmd_vector },
236 { "relativeonly", &opt.relative_only, cmd_boolean },
237 { "remoteencoding", &opt.encoding_remote, cmd_string },
238 { "removelisting", &opt.remove_listing, cmd_boolean },
239 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
240 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
241 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
242 { "robots", &opt.use_robots, cmd_boolean },
243 { "savecookies", &opt.cookies_output, cmd_file },
244 { "saveheaders", &opt.save_headers, cmd_boolean },
246 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
248 { "serverresponse", &opt.server_response, cmd_boolean },
249 { "spanhosts", &opt.spanhost, cmd_boolean },
250 { "spider", &opt.spider, cmd_boolean },
251 { "strictcomments", &opt.strict_comments, cmd_boolean },
252 { "timeout", NULL, cmd_spec_timeout },
253 { "timestamping", &opt.timestamping, cmd_boolean },
254 { "tries", &opt.ntry, cmd_number_inf },
255 { "useproxy", &opt.use_proxy, cmd_boolean },
256 { "user", &opt.user, cmd_string },
257 { "useragent", NULL, cmd_spec_useragent },
258 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
259 { "verbose", NULL, cmd_spec_verbose },
260 { "wait", &opt.wait, cmd_time },
261 { "waitretry", &opt.waitretry, cmd_time },
263 { "wdebug", &opt.wdebug, cmd_boolean },
267 /* Look up CMDNAME in the commands[] and return its position in the
268 array. If CMDNAME is not found, return -1. */
271 command_by_name (const char *cmdname)
273 /* Use binary search for speed. Wget has ~100 commands, which
274 guarantees a worst case performance of 7 string comparisons. */
275 int lo = 0, hi = countof (commands) - 1;
279 int mid = (lo + hi) >> 1;
280 int cmp = strcasecmp (cmdname, commands[mid].name);
291 /* Reset the variables to default values. */
297 /* Most of the default values are 0 (and 0.0, NULL, and false).
298 Just reset everything, and fill in the non-zero values. Note
299 that initializing pointers to NULL this way is technically
300 illegal, but porting Wget to a machine where NULL is not all-zero
301 bit pattern will be the least of the implementors' worries. */
308 opt.add_hostdir = true;
312 opt.http_keep_alive = true;
313 opt.use_proxy = true;
314 tmp = getenv ("no_proxy");
316 opt.no_proxy = sepstring (tmp);
317 opt.prefer_family = prefer_none;
318 opt.allow_cache = true;
320 opt.read_timeout = 900;
321 opt.use_robots = true;
323 opt.remove_listing = true;
325 opt.dot_bytes = 1024;
326 opt.dot_spacing = 10;
327 opt.dots_in_line = 50;
329 opt.dns_cache = true;
333 opt.check_cert = true;
336 /* The default for file name restriction defaults to the OS type. */
337 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
338 opt.restrict_files_os = restrict_windows;
340 opt.restrict_files_os = restrict_unix;
342 opt.restrict_files_ctrl = true;
343 opt.restrict_files_nonascii = false;
344 opt.restrict_files_case = restrict_no_case_restriction;
346 opt.max_redirect = 20;
351 opt.enable_iri = true;
353 opt.enable_iri = false;
356 opt.encoding_remote = NULL;
358 opt.useservertimestamps = true;
361 /* Return the user's home directory (strdup-ed), or NULL if none is
366 static char buf[PATH_MAX];
371 home = getenv ("HOME");
375 /* Under MSDOS, if $HOME isn't defined, use the directory where
376 `wget.exe' resides. */
377 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
380 strcpy (buf, _w32_get_argv0 ());
381 p = strrchr (buf, '/'); /* djgpp */
383 p = strrchr (buf, '\\'); /* others */
387 #elif !defined(WINDOWS)
388 /* If HOME is not defined, try getting it from the password
390 struct passwd *pwd = getpwuid (getuid ());
391 if (!pwd || !pwd->pw_dir)
393 strcpy (buf, pwd->pw_dir);
396 /* Under Windows, if $HOME isn't defined, use the directory where
397 `wget.exe' resides. */
403 return home ? xstrdup (home) : NULL;
406 /* Check the 'WGETRC' environment variable and return the file name
407 if 'WGETRC' is set and is a valid file.
408 If the `WGETRC' variable exists but the file does not exist, the
409 function will exit(). */
411 wgetrc_env_file_name (void)
413 char *env = getenv ("WGETRC");
416 if (!file_exists_p (env))
418 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
422 return xstrdup (env);
427 /* Check for the existance of '$HOME/.wgetrc' and return it's path
428 if it exists and is set. */
430 wgetrc_user_file_name (void)
432 char *home = home_dir ();
434 /* If that failed, try $HOME/.wgetrc (or equivalent). */
437 file = "SYS$LOGIN:.wgetrc";
438 #else /* def __VMS */
441 file = aprintf ("%s/.wgetrc", home);
443 #endif /* def __VMS [else] */
447 if (!file_exists_p (file))
455 /* Return the path to the user's .wgetrc. This is either the value of
456 `WGETRC' environment variable, or `$HOME/.wgetrc'.
458 Additionally, for windows, look in the directory where wget.exe
461 wgetrc_file_name (void)
463 char *file = wgetrc_env_file_name ();
467 file = wgetrc_user_file_name ();
470 /* Under Windows, if we still haven't found .wgetrc, look for the file
471 `wget.ini' in the directory where `wget.exe' resides; we do this for
472 backward compatibility with previous versions of Wget.
473 SYSTEM_WGETRC should not be defined under WINDOWS. */
476 char *home = home_dir ();
482 file = aprintf ("%s/wget.ini", home);
483 if (!file_exists_p (file))
496 /* Return values of parse_line. */
504 static enum parse_line parse_line (const char *, char **, char **, int *);
505 static bool setval_internal (int, const char *, const char *);
506 static bool setval_internal_tilde (int, const char *, const char *);
508 /* Initialize variables from a wgetrc file. Returns zero (failure) if
509 there were errors in the file. */
512 run_wgetrc (const char *file)
519 fp = fopen (file, "r");
522 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
523 file, strerror (errno));
524 return true; /* not a fatal error */
527 while ((line = read_whole_line (fp)) != NULL)
529 char *com = NULL, *val = NULL;
532 /* Parse the line. */
533 switch (parse_line (line, &com, &val, &comind))
536 /* If everything is OK, set the value. */
537 if (!setval_internal_tilde (comind, com, val))
539 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
540 exec_name, file, ln);
544 case line_syntax_error:
545 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
546 exec_name, file, ln);
549 case line_unknown_command:
550 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
551 exec_name, quote (com), file, ln);
569 /* Initialize the defaults and run the system wgetrc and user's own
574 char *file, *env_sysrc;
577 /* Load the hard-coded defaults. */
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 /* Override it with your own, if one exists. */
591 file = wgetrc_file_name ();
594 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
595 something like realpath() before comparing them with `strcmp' */
597 if (!strcmp (file, SYSTEM_WGETRC))
599 fprintf (stderr, _("\
600 %s: Warning: Both system and user wgetrc point to %s.\n"),
601 exec_name, quote (file));
605 ok &= run_wgetrc (file);
607 /* If there were errors processing either `.wgetrc', abort. */
615 /* Remove dashes and underscores from S, modifying S in the
621 char *t = s; /* t - tortoise */
622 char *h = s; /* h - hare */
624 if (*h == '_' || *h == '-')
631 /* Parse the line pointed by line, with the syntax:
632 <sp>* command <sp>* = <sp>* value <sp>*
633 Uses malloc to allocate space for command and value.
635 Returns one of line_ok, line_empty, line_syntax_error, or
636 line_unknown_command.
638 In case of line_ok, *COM and *VAL point to freshly allocated
639 strings, and *COMIND points to com's index. In case of error or
640 empty line, their values are unmodified. */
642 static enum parse_line
643 parse_line (const char *line, char **com, char **val, int *comind)
646 const char *end = line + strlen (line);
647 const char *cmdstart, *cmdend;
648 const char *valstart, *valend;
653 /* Skip leading and trailing whitespace. */
654 while (*line && c_isspace (*line))
656 while (end > line && c_isspace (end[-1]))
659 /* Skip empty lines and comments. */
660 if (!*line || *line == '#')
666 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
670 /* Skip '=', as well as any space before or after it. */
671 while (p < end && c_isspace (*p))
673 if (p == end || *p != '=')
674 return line_syntax_error;
676 while (p < end && c_isspace (*p))
682 /* The syntax is valid (even though the command might not be). Fill
683 in the command name and value. */
684 *com = strdupdelim (cmdstart, cmdend);
685 *val = strdupdelim (valstart, valend);
687 /* The line now known to be syntactically correct. Check whether
688 the command is valid. */
689 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
691 ind = command_by_name (cmdcopy);
693 return line_unknown_command;
695 /* Report success to the caller. */
700 #if defined(WINDOWS) || defined(MSDOS)
701 # define ISSEP(c) ((c) == '/' || (c) == '\\')
703 # define ISSEP(c) ((c) == '/')
706 /* Run commands[comind].action. */
709 setval_internal (int comind, const char *com, const char *val)
711 assert (0 <= comind && ((size_t) comind) < countof (commands));
712 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
713 return commands[comind].action (com, val, commands[comind].place);
717 setval_internal_tilde (int comind, const char *com, const char *val)
723 ret = setval_internal (comind, com, val);
725 /* We make tilde expansion for cmd_file and cmd_directory */
726 if (((commands[comind].action == cmd_file) ||
727 (commands[comind].action == cmd_directory))
728 && ret && (*val == '~' && ISSEP (val[1])))
730 pstring = commands[comind].place;
734 homelen = strlen (home);
735 while (homelen && ISSEP (home[homelen - 1]))
736 home[--homelen] = '\0';
738 /* Skip the leading "~/". */
739 for (++val; ISSEP (*val); val++)
741 *pstring = concat_strings (home, "/", val, (char *)0);
747 /* Run command COM with value VAL. If running the command produces an
748 error, report the error and exit.
750 This is intended to be called from main() to modify Wget's behavior
751 through command-line switches. Since COM is hard-coded in main(),
752 it is not canonicalized, and this aborts when COM is not found.
754 If COMIND's are exported to init.h, this function will be changed
755 to accept COMIND directly. */
758 setoptval (const char *com, const char *val, const char *optname)
760 /* Prepend "--" to OPTNAME. */
761 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
764 strcpy (dd_optname + 2, optname);
766 assert (val != NULL);
767 if (!setval_internal (command_by_name (com), dd_optname, val))
771 /* Parse OPT into command and value and run it. For example,
772 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
773 This is used by the `--execute' flag in main.c. */
776 run_command (const char *opt)
780 switch (parse_line (opt, &com, &val, &comind))
783 if (!setval_internal (comind, com, val))
789 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
790 exec_name, quote (opt));
795 /* Generic helper functions, for use with `commands'. */
797 /* Forward declarations: */
802 static bool decode_string (const char *, const struct decode_item *, int, int *);
803 static bool simple_atoi (const char *, const char *, int *);
804 static bool simple_atof (const char *, const char *, double *);
806 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
808 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
809 && c_tolower((p)[1]) == (c1) \
812 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
813 && c_tolower((p)[1]) == (c1) \
814 && c_tolower((p)[2]) == (c2) \
818 /* Store the boolean value from VAL to PLACE. COM is ignored,
819 except for error messages. */
821 cmd_boolean (const char *com, const char *val, void *place)
825 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
826 /* "on", "yes" and "1" mean true. */
828 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
829 /* "off", "no" and "0" mean false. */
834 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
835 exec_name, com, quote (val));
839 *(bool *) place = value;
843 /* Set the non-negative integer value from VAL to PLACE. With
844 incorrect specification, the number remains unchanged. */
846 cmd_number (const char *com, const char *val, void *place)
848 if (!simple_atoi (val, val + strlen (val), place)
849 || *(int *) place < 0)
851 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
852 exec_name, com, quote (val));
858 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
860 cmd_number_inf (const char *com, const char *val, void *place)
862 if (!strcasecmp (val, "inf"))
867 return cmd_number (com, val, place);
870 /* Copy (strdup) the string at COM to a new location and place a
871 pointer to *PLACE. */
873 cmd_string (const char *com, const char *val, void *place)
875 char **pstring = (char **)place;
877 xfree_null (*pstring);
878 *pstring = xstrdup (val);
883 /* Like the above, but handles tilde-expansion when reading a user's
884 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
885 gets expanded to the user's home directory. */
887 cmd_file (const char *com, const char *val, void *place)
889 char **pstring = (char **)place;
891 xfree_null (*pstring);
893 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
895 *pstring = xstrdup (val);
897 #if defined(WINDOWS) || defined(MSDOS)
898 /* Convert "\" to "/". */
901 for (s = *pstring; *s; s++)
909 /* Like cmd_file, but strips trailing '/' characters. */
911 cmd_directory (const char *com, const char *val, void *place)
915 /* Call cmd_file() for tilde expansion and separator
916 canonicalization (backslash -> slash under Windows). These
917 things should perhaps be in a separate function. */
918 if (!cmd_file (com, val, place))
923 while (t > s && *--t == '/')
929 /* Split VAL by space to a vector of values, and append those values
930 to vector pointed to by the PLACE argument. If VAL is empty, the
931 PLACE vector is cleared instead. */
934 cmd_vector (const char *com, const char *val, void *place)
936 char ***pvec = (char ***)place;
939 *pvec = merge_vecs (*pvec, sepstring (val));
949 cmd_directory_vector (const char *com, const char *val, void *place)
951 char ***pvec = (char ***)place;
955 /* Strip the trailing slashes from directories. */
958 seps = sepstring (val);
959 for (t = seps; t && *t; t++)
961 int len = strlen (*t);
962 /* Skip degenerate case of root directory. */
965 if ((*t)[len - 1] == '/')
966 (*t)[len - 1] = '\0';
969 *pvec = merge_vecs (*pvec, seps);
979 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
980 "100k" or "2.5G" to a floating point number. */
983 parse_bytes_helper (const char *val, double *result)
986 const char *end = val + strlen (val);
988 /* Check for "inf". */
989 if (0 == strcmp (val, "inf"))
995 /* Strip trailing whitespace. */
996 while (val < end && c_isspace (end[-1]))
1001 switch (c_tolower (end[-1]))
1004 --end, mult = 1024.0;
1007 --end, mult = 1048576.0;
1010 --end, mult = 1073741824.0;
1013 --end, mult = 1099511627776.0;
1016 /* Not a recognized suffix: assume it's a digit. (If not,
1017 simple_atof will raise an error.) */
1021 /* Skip leading and trailing whitespace. */
1022 while (val < end && c_isspace (*val))
1024 while (val < end && c_isspace (end[-1]))
1029 if (!simple_atof (val, end, &number) || number < 0)
1032 *result = number * mult;
1036 /* Parse VAL as a number and set its value to PLACE (which should
1039 By default, the value is assumed to be in bytes. If "K", "M", or
1040 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1041 1<<30, respectively. Floating point values are allowed and are
1042 cast to integer before use. The idea is to be able to use things
1043 like 1.5k instead of "1536".
1045 The string "inf" is returned as 0.
1047 In case of error, false is returned and memory pointed to by PLACE
1048 remains unmodified. */
1051 cmd_bytes (const char *com, const char *val, void *place)
1054 if (!parse_bytes_helper (val, &byte_value))
1056 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1057 exec_name, com, quote (val));
1060 *(wgint *)place = (wgint)byte_value;
1064 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1065 SIZE_SUM. It works by converting the string to double, therefore
1066 working with values up to 2^53-1 without loss of precision. This
1067 value (8192 TB) is large enough to serve for a while. */
1070 cmd_bytes_sum (const char *com, const char *val, void *place)
1073 if (!parse_bytes_helper (val, &byte_value))
1075 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1076 exec_name, com, quote (val));
1079 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1083 /* Store the value of VAL to *OUT. The value is a time period, by
1084 default expressed in seconds, but also accepting suffixes "m", "h",
1085 "d", and "w" for minutes, hours, days, and weeks respectively. */
1088 cmd_time (const char *com, const char *val, void *place)
1090 double number, mult;
1091 const char *end = val + strlen (val);
1093 /* Strip trailing whitespace. */
1094 while (val < end && c_isspace (end[-1]))
1100 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1101 exec_name, com, quote (val));
1105 switch (c_tolower (end[-1]))
1108 --end, mult = 1; /* seconds */
1111 --end, mult = 60; /* minutes */
1114 --end, mult = 3600; /* hours */
1117 --end, mult = 86400.0; /* days */
1120 --end, mult = 604800.0; /* weeks */
1123 /* Not a recognized suffix: assume it belongs to the number.
1124 (If not, simple_atof will raise an error.) */
1128 /* Skip leading and trailing whitespace. */
1129 while (val < end && c_isspace (*val))
1131 while (val < end && c_isspace (end[-1]))
1136 if (!simple_atof (val, end, &number))
1139 *(double *)place = number * mult;
1145 cmd_cert_type (const char *com, const char *val, void *place)
1147 static const struct decode_item choices[] = {
1148 { "pem", keyfile_pem },
1149 { "der", keyfile_asn1 },
1150 { "asn1", keyfile_asn1 },
1152 int ok = decode_string (val, choices, countof (choices), place);
1154 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1159 /* Specialized helper functions, used by `commands' to handle some
1160 options specially. */
1162 static bool check_user_specified_header (const char *);
1165 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1167 if (!cmd_boolean (com, val, &opt.dirstruct))
1169 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1170 must be affected inversely. */
1172 opt.no_dirstruct = false;
1174 opt.no_dirstruct = true;
1179 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1181 /* Empty value means reset the list of headers. */
1184 free_vec (opt.user_headers);
1185 opt.user_headers = NULL;
1189 if (!check_user_specified_header (val))
1191 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1192 exec_name, com, quote (val));
1195 opt.user_headers = vec_append (opt.user_headers, val);
1200 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1202 int flag = cmd_boolean (com, val, &opt.htmlify);
1203 if (flag && !opt.htmlify)
1204 opt.remove_listing = false;
1208 /* Set the "mirror" mode. It means: recursive download, timestamping,
1209 no limit on max. recursion depth, and don't remove listings. */
1212 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1216 if (!cmd_boolean (com, val, &mirror))
1220 opt.recursive = true;
1221 if (!opt.no_dirstruct)
1222 opt.dirstruct = true;
1223 opt.timestamping = true;
1224 opt.reclevel = INFINITE_RECURSION;
1225 opt.remove_listing = false;
1230 /* Validate --prefer-family and set the choice. Allowed values are
1231 "IPv4", "IPv6", and "none". */
1234 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1236 static const struct decode_item choices[] = {
1237 { "IPv4", prefer_ipv4 },
1238 { "IPv6", prefer_ipv6 },
1239 { "none", prefer_none },
1241 int prefer_family = prefer_none;
1242 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1244 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1245 opt.prefer_family = prefer_family;
1249 /* Set progress.type to VAL, but verify that it's a valid progress
1250 implementation before that. */
1253 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1255 if (!valid_progress_implementation_p (val))
1257 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1258 exec_name, com, quote (val));
1261 xfree_null (opt.progress_type);
1263 /* Don't call set_progress_implementation here. It will be called
1264 in main() when it becomes clear what the log output is. */
1265 opt.progress_type = xstrdup (val);
1269 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1270 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1274 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1276 if (!cmd_boolean (com, val, &opt.recursive))
1280 if (opt.recursive && !opt.no_dirstruct)
1281 opt.dirstruct = true;
1287 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1289 int restrict_os = opt.restrict_files_os;
1290 int restrict_ctrl = opt.restrict_files_ctrl;
1291 int restrict_case = opt.restrict_files_case;
1292 int restrict_nonascii = opt.restrict_files_nonascii;
1296 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1300 end = strchr (val, ',');
1302 end = val + strlen (val);
1304 if (VAL_IS ("unix"))
1305 restrict_os = restrict_unix;
1306 else if (VAL_IS ("windows"))
1307 restrict_os = restrict_windows;
1308 else if (VAL_IS ("lowercase"))
1309 restrict_case = restrict_lowercase;
1310 else if (VAL_IS ("uppercase"))
1311 restrict_case = restrict_uppercase;
1312 else if (VAL_IS ("nocontrol"))
1313 restrict_ctrl = false;
1314 else if (VAL_IS ("ascii"))
1315 restrict_nonascii = true;
1318 fprintf (stderr, _("\
1319 %s: %s: Invalid restriction %s,\n\
1320 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1321 exec_name, com, quote (val));
1328 while (*val && *end);
1332 opt.restrict_files_os = restrict_os;
1333 opt.restrict_files_ctrl = restrict_ctrl;
1334 opt.restrict_files_case = restrict_case;
1335 opt.restrict_files_nonascii = restrict_nonascii;
1342 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1344 static const struct decode_item choices[] = {
1345 { "auto", secure_protocol_auto },
1346 { "sslv2", secure_protocol_sslv2 },
1347 { "sslv3", secure_protocol_sslv3 },
1348 { "tlsv1", secure_protocol_tlsv1 },
1350 int ok = decode_string (val, choices, countof (choices), place);
1352 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1357 /* Set all three timeout values. */
1360 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1363 if (!cmd_time (com, val, &value))
1365 opt.read_timeout = value;
1366 opt.connect_timeout = value;
1367 opt.dns_timeout = value;
1372 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1374 /* Disallow embedded newlines. */
1375 if (strchr (val, '\n'))
1377 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1378 exec_name, com, quote (val));
1381 xfree_null (opt.useragent);
1382 opt.useragent = xstrdup (val);
1386 /* The "verbose" option cannot be cmd_boolean because the variable is
1387 not bool -- it's of type int (-1 means uninitialized because of
1388 some random hackery for disallowing -q -v). */
1391 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1394 if (cmd_boolean (com, val, &flag))
1402 /* Miscellaneous useful routines. */
1404 /* A very simple atoi clone, more useful than atoi because it works on
1405 delimited strings, and has error reportage. Returns true on success,
1406 false on failure. If successful, stores result to *DEST. */
1409 simple_atoi (const char *beg, const char *end, int *dest)
1412 bool negative = false;
1413 const char *p = beg;
1415 while (p < end && c_isspace (*p))
1417 if (p < end && (*p == '-' || *p == '+'))
1419 negative = (*p == '-');
1425 /* Read negative numbers in a separate loop because the most
1426 negative integer cannot be represented as a positive number. */
1429 for (; p < end && c_isdigit (*p); p++)
1431 int next = (10 * result) + (*p - '0');
1433 return false; /* overflow */
1437 for (; p < end && c_isdigit (*p); p++)
1439 int next = (10 * result) - (*p - '0');
1441 return false; /* underflow */
1452 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1453 doesn't handle exponential notation. Returns true on success,
1454 false on failure. In case of success, stores its result to
1458 simple_atof (const char *beg, const char *end, double *dest)
1462 bool negative = false;
1463 bool seen_dot = false;
1464 bool seen_digit = false;
1467 const char *p = beg;
1469 while (p < end && c_isspace (*p))
1471 if (p < end && (*p == '-' || *p == '+'))
1473 negative = (*p == '-');
1477 for (; p < end; p++)
1483 result = (10 * result) + (ch - '0');
1485 result += (ch - '0') / (divider *= 10);
1507 /* Verify that the user-specified header in S is valid. It must
1508 contain a colon preceded by non-white-space characters and must not
1509 contain newlines. */
1512 check_user_specified_header (const char *s)
1516 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1518 /* The header MUST contain `:' preceded by at least one
1519 non-whitespace character. */
1520 if (*p != ':' || p == s)
1522 /* The header MUST NOT contain newlines. */
1523 if (strchr (s, '\n'))
1528 /* Decode VAL into a number, according to ITEMS. */
1531 decode_string (const char *val, const struct decode_item *items, int itemcount,
1535 for (i = 0; i < itemcount; i++)
1536 if (0 == strcasecmp (val, items[i].name))
1538 *place = items[i].code;
1545 void cleanup_html_url (void);
1548 /* Free the memory allocated by global variables. */
1552 /* Free external resources, close files, etc. */
1555 fclose (output_stream);
1556 /* No need to check for error because Wget flushes its output (and
1557 checks for errors) after any data arrives. */
1559 /* We're exiting anyway so there's no real need to call free()
1560 hundreds of times. Skipping the frees will make Wget exit
1563 However, when detecting leaks, it's crucial to free() everything
1564 because then you can find the real leaks, i.e. the allocated
1565 memory which grows with the size of the program. */
1571 cleanup_html_url ();
1576 extern acc_t *netrc_list;
1577 free_netrc (netrc_list);
1579 xfree_null (opt.lfilename);
1580 xfree_null (opt.dir_prefix);
1581 xfree_null (opt.input_filename);
1582 xfree_null (opt.output_document);
1583 free_vec (opt.accepts);
1584 free_vec (opt.rejects);
1585 free_vec (opt.excludes);
1586 free_vec (opt.includes);
1587 free_vec (opt.domains);
1588 free_vec (opt.follow_tags);
1589 free_vec (opt.ignore_tags);
1590 xfree_null (opt.progress_type);
1591 xfree_null (opt.ftp_user);
1592 xfree_null (opt.ftp_passwd);
1593 xfree_null (opt.ftp_proxy);
1594 xfree_null (opt.https_proxy);
1595 xfree_null (opt.http_proxy);
1596 free_vec (opt.no_proxy);
1597 xfree_null (opt.useragent);
1598 xfree_null (opt.referer);
1599 xfree_null (opt.http_user);
1600 xfree_null (opt.http_passwd);
1601 free_vec (opt.user_headers);
1603 xfree_null (opt.cert_file);
1604 xfree_null (opt.private_key);
1605 xfree_null (opt.ca_directory);
1606 xfree_null (opt.ca_cert);
1607 xfree_null (opt.random_file);
1608 xfree_null (opt.egd_file);
1610 xfree_null (opt.bind_address);
1611 xfree_null (opt.cookies_input);
1612 xfree_null (opt.cookies_output);
1613 xfree_null (opt.user);
1614 xfree_null (opt.passwd);
1615 xfree_null (opt.base_href);
1617 #endif /* DEBUG_MALLOC */
1620 /* Unit testing routines. */
1625 test_commands_sorted()
1627 int prev_idx = 0, next_idx = 1;
1628 int command_count = countof (commands) - 1;
1630 while (next_idx <= command_count)
1632 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1635 mu_assert ("FAILED", false);
1648 test_cmd_spec_restrict_file_names()
1653 int expected_restrict_files_os;
1654 int expected_restrict_files_ctrl;
1655 int expected_restrict_files_case;
1658 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1659 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1660 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1661 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1664 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1669 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1672 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1673 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1674 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1675 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1677 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1678 res == test_array[i].result
1679 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1680 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1681 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1687 #endif /* TESTING */