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. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
63 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
65 CMD_DECLARE (cmd_boolean);
66 CMD_DECLARE (cmd_bytes);
67 CMD_DECLARE (cmd_bytes_sum);
69 CMD_DECLARE (cmd_cert_type);
71 CMD_DECLARE (cmd_directory_vector);
72 CMD_DECLARE (cmd_number);
73 CMD_DECLARE (cmd_number_inf);
74 CMD_DECLARE (cmd_string);
75 CMD_DECLARE (cmd_file);
76 CMD_DECLARE (cmd_directory);
77 CMD_DECLARE (cmd_time);
78 CMD_DECLARE (cmd_vector);
80 CMD_DECLARE (cmd_spec_dirstruct);
81 CMD_DECLARE (cmd_spec_header);
82 CMD_DECLARE (cmd_spec_htmlify);
83 CMD_DECLARE (cmd_spec_mirror);
84 CMD_DECLARE (cmd_spec_prefer_family);
85 CMD_DECLARE (cmd_spec_progress);
86 CMD_DECLARE (cmd_spec_recursive);
87 CMD_DECLARE (cmd_spec_restrict_file_names);
89 CMD_DECLARE (cmd_spec_secure_protocol);
91 CMD_DECLARE (cmd_spec_timeout);
92 CMD_DECLARE (cmd_spec_useragent);
93 CMD_DECLARE (cmd_spec_verbose);
95 /* List of recognized commands, each consisting of name, place and
96 function. When adding a new command, simply add it to the list,
97 but be sure to keep the list sorted alphabetically, as
98 command_by_name's binary search depends on it. Also, be sure to
99 add any entries that allocate memory (e.g. cmd_string and
100 cmd_vector) to the cleanup() function below. */
102 static const struct {
105 bool (*action) (const char *, const char *, void *);
107 /* KEEP THIS LIST ALPHABETICALLY SORTED */
108 { "accept", &opt.accepts, cmd_vector },
109 { "addhostdir", &opt.add_hostdir, cmd_boolean },
110 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
111 { "askpassword", &opt.ask_passwd, cmd_boolean },
112 { "authnochallenge", &opt.auth_without_challenge,
114 { "background", &opt.background, cmd_boolean },
115 { "backupconverted", &opt.backup_converted, cmd_boolean },
116 { "backups", &opt.backups, cmd_number },
117 { "base", &opt.base_href, cmd_string },
118 { "bindaddress", &opt.bind_address, cmd_string },
120 { "cacertificate", &opt.ca_cert, cmd_file },
122 { "cache", &opt.allow_cache, cmd_boolean },
124 { "cadirectory", &opt.ca_directory, cmd_directory },
125 { "certificate", &opt.cert_file, cmd_file },
126 { "certificatetype", &opt.cert_type, cmd_cert_type },
127 { "checkcertificate", &opt.check_cert, cmd_boolean },
129 { "connecttimeout", &opt.connect_timeout, cmd_time },
130 { "contentdisposition", &opt.content_disposition, cmd_boolean },
131 { "continue", &opt.always_rest, cmd_boolean },
132 { "convertlinks", &opt.convert_links, cmd_boolean },
133 { "cookies", &opt.cookies, cmd_boolean },
134 { "cutdirs", &opt.cut_dirs, cmd_number },
136 { "debug", &opt.debug, cmd_boolean },
138 { "defaultpage", &opt.default_page, cmd_string},
139 { "deleteafter", &opt.delete_after, cmd_boolean },
140 { "dirprefix", &opt.dir_prefix, cmd_directory },
141 { "dirstruct", NULL, cmd_spec_dirstruct },
142 { "dnscache", &opt.dns_cache, cmd_boolean },
143 { "dnstimeout", &opt.dns_timeout, cmd_time },
144 { "domains", &opt.domains, cmd_vector },
145 { "dotbytes", &opt.dot_bytes, cmd_bytes },
146 { "dotsinline", &opt.dots_in_line, cmd_number },
147 { "dotspacing", &opt.dot_spacing, cmd_number },
148 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
150 { "egdfile", &opt.egd_file, cmd_file },
152 { "excludedirectories", &opt.excludes, cmd_directory_vector },
153 { "excludedomains", &opt.exclude_domains, cmd_vector },
154 { "followftp", &opt.follow_ftp, cmd_boolean },
155 { "followtags", &opt.follow_tags, cmd_vector },
156 { "forcehtml", &opt.force_html, cmd_boolean },
157 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
158 { "ftppassword", &opt.ftp_passwd, cmd_string },
159 { "ftpproxy", &opt.ftp_proxy, cmd_string },
161 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
162 #endif /* def __VMS */
163 { "ftpuser", &opt.ftp_user, cmd_string },
164 { "glob", &opt.ftp_glob, cmd_boolean },
165 { "header", NULL, cmd_spec_header },
166 { "htmlextension", &opt.html_extension, cmd_boolean },
167 { "htmlify", NULL, cmd_spec_htmlify },
168 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
169 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
170 { "httppassword", &opt.http_passwd, cmd_string },
171 { "httpproxy", &opt.http_proxy, cmd_string },
172 { "httpsproxy", &opt.https_proxy, cmd_string },
173 { "httpuser", &opt.http_user, cmd_string },
174 { "ignorecase", &opt.ignore_case, cmd_boolean },
175 { "ignorelength", &opt.ignore_length, cmd_boolean },
176 { "ignoretags", &opt.ignore_tags, cmd_vector },
177 { "includedirectories", &opt.includes, cmd_directory_vector },
179 { "inet4only", &opt.ipv4_only, cmd_boolean },
180 { "inet6only", &opt.ipv6_only, cmd_boolean },
182 { "input", &opt.input_filename, cmd_file },
183 { "iri", &opt.enable_iri, cmd_boolean },
184 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
185 { "limitrate", &opt.limit_rate, cmd_bytes },
186 { "loadcookies", &opt.cookies_input, cmd_file },
187 { "localencoding", &opt.locale, cmd_string },
188 { "logfile", &opt.lfilename, cmd_file },
189 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
190 { "maxredirect", &opt.max_redirect, cmd_number },
191 { "mirror", NULL, cmd_spec_mirror },
192 { "netrc", &opt.netrc, cmd_boolean },
193 { "noclobber", &opt.noclobber, cmd_boolean },
194 { "noparent", &opt.no_parent, cmd_boolean },
195 { "noproxy", &opt.no_proxy, cmd_vector },
196 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
197 { "outputdocument", &opt.output_document, cmd_file },
198 { "pagerequisites", &opt.page_requisites, cmd_boolean },
199 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
200 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
201 { "password", &opt.passwd, cmd_string },
202 { "postdata", &opt.post_data, cmd_string },
203 { "postfile", &opt.post_file_name, cmd_file },
204 { "preferfamily", NULL, cmd_spec_prefer_family },
205 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
207 { "privatekey", &opt.private_key, cmd_file },
208 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
210 { "progress", &opt.progress_type, cmd_spec_progress },
211 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
212 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
213 { "proxypassword", &opt.proxy_passwd, cmd_string },
214 { "proxyuser", &opt.proxy_user, cmd_string },
215 { "quiet", &opt.quiet, cmd_boolean },
216 { "quota", &opt.quota, cmd_bytes_sum },
218 { "randomfile", &opt.random_file, cmd_file },
220 { "randomwait", &opt.random_wait, cmd_boolean },
221 { "readtimeout", &opt.read_timeout, cmd_time },
222 { "reclevel", &opt.reclevel, cmd_number_inf },
223 { "recursive", NULL, cmd_spec_recursive },
224 { "referer", &opt.referer, cmd_string },
225 { "reject", &opt.rejects, cmd_vector },
226 { "relativeonly", &opt.relative_only, cmd_boolean },
227 { "remoteencoding", &opt.encoding_remote, cmd_string },
228 { "removelisting", &opt.remove_listing, cmd_boolean },
229 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
230 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
231 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
232 { "robots", &opt.use_robots, cmd_boolean },
233 { "savecookies", &opt.cookies_output, cmd_file },
234 { "saveheaders", &opt.save_headers, cmd_boolean },
236 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
238 { "serverresponse", &opt.server_response, cmd_boolean },
239 { "spanhosts", &opt.spanhost, cmd_boolean },
240 { "spider", &opt.spider, cmd_boolean },
241 { "strictcomments", &opt.strict_comments, cmd_boolean },
242 { "timeout", NULL, cmd_spec_timeout },
243 { "timestamping", &opt.timestamping, cmd_boolean },
244 { "tries", &opt.ntry, cmd_number_inf },
245 { "useproxy", &opt.use_proxy, cmd_boolean },
246 { "user", &opt.user, cmd_string },
247 { "useragent", NULL, cmd_spec_useragent },
248 { "verbose", NULL, cmd_spec_verbose },
249 { "wait", &opt.wait, cmd_time },
250 { "waitretry", &opt.waitretry, cmd_time },
252 { "wdebug", &opt.wdebug, cmd_boolean },
256 /* Look up CMDNAME in the commands[] and return its position in the
257 array. If CMDNAME is not found, return -1. */
260 command_by_name (const char *cmdname)
262 /* Use binary search for speed. Wget has ~100 commands, which
263 guarantees a worst case performance of 7 string comparisons. */
264 int lo = 0, hi = countof (commands) - 1;
268 int mid = (lo + hi) >> 1;
269 int cmp = strcasecmp (cmdname, commands[mid].name);
280 /* Reset the variables to default values. */
286 /* Most of the default values are 0 (and 0.0, NULL, and false).
287 Just reset everything, and fill in the non-zero values. Note
288 that initializing pointers to NULL this way is technically
289 illegal, but porting Wget to a machine where NULL is not all-zero
290 bit pattern will be the least of the implementors' worries. */
297 opt.add_hostdir = true;
301 opt.http_keep_alive = true;
302 opt.use_proxy = true;
303 tmp = getenv ("no_proxy");
305 opt.no_proxy = sepstring (tmp);
306 opt.prefer_family = prefer_none;
307 opt.allow_cache = true;
309 opt.read_timeout = 900;
310 opt.use_robots = true;
312 opt.remove_listing = true;
314 opt.dot_bytes = 1024;
315 opt.dot_spacing = 10;
316 opt.dots_in_line = 50;
318 opt.dns_cache = true;
322 opt.check_cert = true;
325 /* The default for file name restriction defaults to the OS type. */
326 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
327 opt.restrict_files_os = restrict_windows;
329 opt.restrict_files_os = restrict_unix;
331 opt.restrict_files_ctrl = true;
332 opt.restrict_files_nonascii = false;
333 opt.restrict_files_case = restrict_no_case_restriction;
335 opt.max_redirect = 20;
340 opt.enable_iri = true;
342 opt.enable_iri = false;
345 opt.encoding_remote = NULL;
348 /* Return the user's home directory (strdup-ed), or NULL if none is
353 static char buf[PATH_MAX];
358 home = getenv ("HOME");
362 /* Under MSDOS, if $HOME isn't defined, use the directory where
363 `wget.exe' resides. */
364 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
367 strcpy (buf, _w32_get_argv0 ());
368 p = strrchr (buf, '/'); /* djgpp */
370 p = strrchr (buf, '\\'); /* others */
374 #elif !defined(WINDOWS)
375 /* If HOME is not defined, try getting it from the password
377 struct passwd *pwd = getpwuid (getuid ());
378 if (!pwd || !pwd->pw_dir)
380 strcpy (buf, pwd->pw_dir);
383 /* Under Windows, if $HOME isn't defined, use the directory where
384 `wget.exe' resides. */
390 return home ? xstrdup (home) : NULL;
393 /* Check the 'WGETRC' environment variable and return the file name
394 if 'WGETRC' is set and is a valid file.
395 If the `WGETRC' variable exists but the file does not exist, the
396 function will exit(). */
398 wgetrc_env_file_name (void)
400 char *env = getenv ("WGETRC");
403 if (!file_exists_p (env))
405 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
409 return xstrdup (env);
414 /* Check for the existance of '$HOME/.wgetrc' and return it's path
415 if it exists and is set. */
417 wgetrc_user_file_name (void)
419 char *home = home_dir ();
421 /* If that failed, try $HOME/.wgetrc (or equivalent). */
424 file = "SYS$LOGIN:.wgetrc";
425 #else /* def __VMS */
428 file = aprintf ("%s/.wgetrc", home);
430 #endif /* def __VMS [else] */
434 if (!file_exists_p (file))
442 /* Return the path to the user's .wgetrc. This is either the value of
443 `WGETRC' environment variable, or `$HOME/.wgetrc'.
445 Additionally, for windows, look in the directory where wget.exe
448 wgetrc_file_name (void)
450 char *file = wgetrc_env_file_name ();
454 file = wgetrc_user_file_name ();
457 /* Under Windows, if we still haven't found .wgetrc, look for the file
458 `wget.ini' in the directory where `wget.exe' resides; we do this for
459 backward compatibility with previous versions of Wget.
460 SYSTEM_WGETRC should not be defined under WINDOWS. */
463 char *home = home_dir ();
469 file = aprintf ("%s/wget.ini", home);
470 if (!file_exists_p (file))
483 /* Return values of parse_line. */
491 static enum parse_line parse_line (const char *, char **, char **, int *);
492 static bool setval_internal (int, const char *, const char *);
493 static bool setval_internal_tilde (int, const char *, const char *);
495 /* Initialize variables from a wgetrc file. Returns zero (failure) if
496 there were errors in the file. */
499 run_wgetrc (const char *file)
506 fp = fopen (file, "r");
509 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
510 file, strerror (errno));
511 return true; /* not a fatal error */
514 while ((line = read_whole_line (fp)) != NULL)
516 char *com = NULL, *val = NULL;
519 /* Parse the line. */
520 switch (parse_line (line, &com, &val, &comind))
523 /* If everything is OK, set the value. */
524 if (!setval_internal_tilde (comind, com, val))
526 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
527 exec_name, file, ln);
531 case line_syntax_error:
532 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
533 exec_name, file, ln);
536 case line_unknown_command:
537 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
538 exec_name, quote (com), file, ln);
556 /* Initialize the defaults and run the system wgetrc and user's own
561 char *file, *env_sysrc;
564 /* Load the hard-coded defaults. */
567 /* Run a non-standard system rc file when the according environment
568 variable has been set. For internal testing purposes only! */
569 env_sysrc = getenv ("SYSTEM_WGETRC");
570 if (env_sysrc && file_exists_p (env_sysrc))
571 ok &= run_wgetrc (env_sysrc);
572 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
574 else if (file_exists_p (SYSTEM_WGETRC))
575 ok &= run_wgetrc (SYSTEM_WGETRC);
577 /* Override it with your own, if one exists. */
578 file = wgetrc_file_name ();
581 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
582 something like realpath() before comparing them with `strcmp' */
584 if (!strcmp (file, SYSTEM_WGETRC))
586 fprintf (stderr, _("\
587 %s: Warning: Both system and user wgetrc point to %s.\n"),
588 exec_name, quote (file));
592 ok &= run_wgetrc (file);
594 /* If there were errors processing either `.wgetrc', abort. */
602 /* Remove dashes and underscores from S, modifying S in the
608 char *t = s; /* t - tortoise */
609 char *h = s; /* h - hare */
611 if (*h == '_' || *h == '-')
618 /* Parse the line pointed by line, with the syntax:
619 <sp>* command <sp>* = <sp>* value <sp>*
620 Uses malloc to allocate space for command and value.
622 Returns one of line_ok, line_empty, line_syntax_error, or
623 line_unknown_command.
625 In case of line_ok, *COM and *VAL point to freshly allocated
626 strings, and *COMIND points to com's index. In case of error or
627 empty line, their values are unmodified. */
629 static enum parse_line
630 parse_line (const char *line, char **com, char **val, int *comind)
633 const char *end = line + strlen (line);
634 const char *cmdstart, *cmdend;
635 const char *valstart, *valend;
640 /* Skip leading and trailing whitespace. */
641 while (*line && c_isspace (*line))
643 while (end > line && c_isspace (end[-1]))
646 /* Skip empty lines and comments. */
647 if (!*line || *line == '#')
653 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
657 /* Skip '=', as well as any space before or after it. */
658 while (p < end && c_isspace (*p))
660 if (p == end || *p != '=')
661 return line_syntax_error;
663 while (p < end && c_isspace (*p))
669 /* The syntax is valid (even though the command might not be). Fill
670 in the command name and value. */
671 *com = strdupdelim (cmdstart, cmdend);
672 *val = strdupdelim (valstart, valend);
674 /* The line now known to be syntactically correct. Check whether
675 the command is valid. */
676 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
678 ind = command_by_name (cmdcopy);
680 return line_unknown_command;
682 /* Report success to the caller. */
687 #if defined(WINDOWS) || defined(MSDOS)
688 # define ISSEP(c) ((c) == '/' || (c) == '\\')
690 # define ISSEP(c) ((c) == '/')
693 /* Run commands[comind].action. */
696 setval_internal (int comind, const char *com, const char *val)
698 assert (0 <= comind && ((size_t) comind) < countof (commands));
699 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
700 return commands[comind].action (com, val, commands[comind].place);
704 setval_internal_tilde (int comind, const char *com, const char *val)
710 ret = setval_internal (comind, com, val);
712 /* We make tilde expansion for cmd_file and cmd_directory */
713 if (((commands[comind].action == cmd_file) ||
714 (commands[comind].action == cmd_directory))
715 && ret && (*val == '~' && ISSEP (val[1])))
717 pstring = commands[comind].place;
721 homelen = strlen (home);
722 while (homelen && ISSEP (home[homelen - 1]))
723 home[--homelen] = '\0';
725 /* Skip the leading "~/". */
726 for (++val; ISSEP (*val); val++)
728 *pstring = concat_strings (home, "/", val, (char *)0);
734 /* Run command COM with value VAL. If running the command produces an
735 error, report the error and exit.
737 This is intended to be called from main() to modify Wget's behavior
738 through command-line switches. Since COM is hard-coded in main(),
739 it is not canonicalized, and this aborts when COM is not found.
741 If COMIND's are exported to init.h, this function will be changed
742 to accept COMIND directly. */
745 setoptval (const char *com, const char *val, const char *optname)
747 /* Prepend "--" to OPTNAME. */
748 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
751 strcpy (dd_optname + 2, optname);
753 assert (val != NULL);
754 if (!setval_internal (command_by_name (com), dd_optname, val))
758 /* Parse OPT into command and value and run it. For example,
759 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
760 This is used by the `--execute' flag in main.c. */
763 run_command (const char *opt)
767 switch (parse_line (opt, &com, &val, &comind))
770 if (!setval_internal (comind, com, val))
776 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
777 exec_name, quote (opt));
782 /* Generic helper functions, for use with `commands'. */
784 /* Forward declarations: */
789 static bool decode_string (const char *, const struct decode_item *, int, int *);
790 static bool simple_atoi (const char *, const char *, int *);
791 static bool simple_atof (const char *, const char *, double *);
793 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
795 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
796 && c_tolower((p)[1]) == (c1) \
799 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
800 && c_tolower((p)[1]) == (c1) \
801 && c_tolower((p)[2]) == (c2) \
805 /* Store the boolean value from VAL to PLACE. COM is ignored,
806 except for error messages. */
808 cmd_boolean (const char *com, const char *val, void *place)
812 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
813 /* "on", "yes" and "1" mean true. */
815 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
816 /* "off", "no" and "0" mean false. */
821 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
822 exec_name, com, quote (val));
826 *(bool *) place = value;
830 /* Set the non-negative integer value from VAL to PLACE. With
831 incorrect specification, the number remains unchanged. */
833 cmd_number (const char *com, const char *val, void *place)
835 if (!simple_atoi (val, val + strlen (val), place)
836 || *(int *) place < 0)
838 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
839 exec_name, com, quote (val));
845 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
847 cmd_number_inf (const char *com, const char *val, void *place)
849 if (!strcasecmp (val, "inf"))
854 return cmd_number (com, val, place);
857 /* Copy (strdup) the string at COM to a new location and place a
858 pointer to *PLACE. */
860 cmd_string (const char *com, const char *val, void *place)
862 char **pstring = (char **)place;
864 xfree_null (*pstring);
865 *pstring = xstrdup (val);
870 /* Like the above, but handles tilde-expansion when reading a user's
871 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
872 gets expanded to the user's home directory. */
874 cmd_file (const char *com, const char *val, void *place)
876 char **pstring = (char **)place;
878 xfree_null (*pstring);
880 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
882 *pstring = xstrdup (val);
884 #if defined(WINDOWS) || defined(MSDOS)
885 /* Convert "\" to "/". */
888 for (s = *pstring; *s; s++)
896 /* Like cmd_file, but strips trailing '/' characters. */
898 cmd_directory (const char *com, const char *val, void *place)
902 /* Call cmd_file() for tilde expansion and separator
903 canonicalization (backslash -> slash under Windows). These
904 things should perhaps be in a separate function. */
905 if (!cmd_file (com, val, place))
910 while (t > s && *--t == '/')
916 /* Split VAL by space to a vector of values, and append those values
917 to vector pointed to by the PLACE argument. If VAL is empty, the
918 PLACE vector is cleared instead. */
921 cmd_vector (const char *com, const char *val, void *place)
923 char ***pvec = (char ***)place;
926 *pvec = merge_vecs (*pvec, sepstring (val));
936 cmd_directory_vector (const char *com, const char *val, void *place)
938 char ***pvec = (char ***)place;
942 /* Strip the trailing slashes from directories. */
945 seps = sepstring (val);
946 for (t = seps; t && *t; t++)
948 int len = strlen (*t);
949 /* Skip degenerate case of root directory. */
952 if ((*t)[len - 1] == '/')
953 (*t)[len - 1] = '\0';
956 *pvec = merge_vecs (*pvec, seps);
966 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
967 "100k" or "2.5G" to a floating point number. */
970 parse_bytes_helper (const char *val, double *result)
973 const char *end = val + strlen (val);
975 /* Check for "inf". */
976 if (0 == strcmp (val, "inf"))
982 /* Strip trailing whitespace. */
983 while (val < end && c_isspace (end[-1]))
988 switch (c_tolower (end[-1]))
991 --end, mult = 1024.0;
994 --end, mult = 1048576.0;
997 --end, mult = 1073741824.0;
1000 --end, mult = 1099511627776.0;
1003 /* Not a recognized suffix: assume it's a digit. (If not,
1004 simple_atof will raise an error.) */
1008 /* Skip leading and trailing whitespace. */
1009 while (val < end && c_isspace (*val))
1011 while (val < end && c_isspace (end[-1]))
1016 if (!simple_atof (val, end, &number) || number < 0)
1019 *result = number * mult;
1023 /* Parse VAL as a number and set its value to PLACE (which should
1026 By default, the value is assumed to be in bytes. If "K", "M", or
1027 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1028 1<<30, respectively. Floating point values are allowed and are
1029 cast to integer before use. The idea is to be able to use things
1030 like 1.5k instead of "1536".
1032 The string "inf" is returned as 0.
1034 In case of error, false is returned and memory pointed to by PLACE
1035 remains unmodified. */
1038 cmd_bytes (const char *com, const char *val, void *place)
1041 if (!parse_bytes_helper (val, &byte_value))
1043 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1044 exec_name, com, quote (val));
1047 *(wgint *)place = (wgint)byte_value;
1051 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1052 SIZE_SUM. It works by converting the string to double, therefore
1053 working with values up to 2^53-1 without loss of precision. This
1054 value (8192 TB) is large enough to serve for a while. */
1057 cmd_bytes_sum (const char *com, const char *val, void *place)
1060 if (!parse_bytes_helper (val, &byte_value))
1062 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1063 exec_name, com, quote (val));
1066 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1070 /* Store the value of VAL to *OUT. The value is a time period, by
1071 default expressed in seconds, but also accepting suffixes "m", "h",
1072 "d", and "w" for minutes, hours, days, and weeks respectively. */
1075 cmd_time (const char *com, const char *val, void *place)
1077 double number, mult;
1078 const char *end = val + strlen (val);
1080 /* Strip trailing whitespace. */
1081 while (val < end && c_isspace (end[-1]))
1087 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1088 exec_name, com, quote (val));
1092 switch (c_tolower (end[-1]))
1095 --end, mult = 1; /* seconds */
1098 --end, mult = 60; /* minutes */
1101 --end, mult = 3600; /* hours */
1104 --end, mult = 86400.0; /* days */
1107 --end, mult = 604800.0; /* weeks */
1110 /* Not a recognized suffix: assume it belongs to the number.
1111 (If not, simple_atof will raise an error.) */
1115 /* Skip leading and trailing whitespace. */
1116 while (val < end && c_isspace (*val))
1118 while (val < end && c_isspace (end[-1]))
1123 if (!simple_atof (val, end, &number))
1126 *(double *)place = number * mult;
1132 cmd_cert_type (const char *com, const char *val, void *place)
1134 static const struct decode_item choices[] = {
1135 { "pem", keyfile_pem },
1136 { "der", keyfile_asn1 },
1137 { "asn1", keyfile_asn1 },
1139 int ok = decode_string (val, choices, countof (choices), place);
1141 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1146 /* Specialized helper functions, used by `commands' to handle some
1147 options specially. */
1149 static bool check_user_specified_header (const char *);
1152 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1154 if (!cmd_boolean (com, val, &opt.dirstruct))
1156 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1157 must be affected inversely. */
1159 opt.no_dirstruct = false;
1161 opt.no_dirstruct = true;
1166 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1168 /* Empty value means reset the list of headers. */
1171 free_vec (opt.user_headers);
1172 opt.user_headers = NULL;
1176 if (!check_user_specified_header (val))
1178 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1179 exec_name, com, quote (val));
1182 opt.user_headers = vec_append (opt.user_headers, val);
1187 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1189 int flag = cmd_boolean (com, val, &opt.htmlify);
1190 if (flag && !opt.htmlify)
1191 opt.remove_listing = false;
1195 /* Set the "mirror" mode. It means: recursive download, timestamping,
1196 no limit on max. recursion depth, and don't remove listings. */
1199 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1203 if (!cmd_boolean (com, val, &mirror))
1207 opt.recursive = true;
1208 if (!opt.no_dirstruct)
1209 opt.dirstruct = true;
1210 opt.timestamping = true;
1211 opt.reclevel = INFINITE_RECURSION;
1212 opt.remove_listing = false;
1217 /* Validate --prefer-family and set the choice. Allowed values are
1218 "IPv4", "IPv6", and "none". */
1221 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1223 static const struct decode_item choices[] = {
1224 { "IPv4", prefer_ipv4 },
1225 { "IPv6", prefer_ipv6 },
1226 { "none", prefer_none },
1228 int prefer_family = prefer_none;
1229 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1231 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1232 opt.prefer_family = prefer_family;
1236 /* Set progress.type to VAL, but verify that it's a valid progress
1237 implementation before that. */
1240 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1242 if (!valid_progress_implementation_p (val))
1244 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1245 exec_name, com, quote (val));
1248 xfree_null (opt.progress_type);
1250 /* Don't call set_progress_implementation here. It will be called
1251 in main() when it becomes clear what the log output is. */
1252 opt.progress_type = xstrdup (val);
1256 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1257 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1261 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1263 if (!cmd_boolean (com, val, &opt.recursive))
1267 if (opt.recursive && !opt.no_dirstruct)
1268 opt.dirstruct = true;
1274 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1276 int restrict_os = opt.restrict_files_os;
1277 int restrict_ctrl = opt.restrict_files_ctrl;
1278 int restrict_case = opt.restrict_files_case;
1279 int restrict_nonascii = opt.restrict_files_nonascii;
1283 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1287 end = strchr (val, ',');
1289 end = val + strlen (val);
1291 if (VAL_IS ("unix"))
1292 restrict_os = restrict_unix;
1293 else if (VAL_IS ("windows"))
1294 restrict_os = restrict_windows;
1295 else if (VAL_IS ("lowercase"))
1296 restrict_case = restrict_lowercase;
1297 else if (VAL_IS ("uppercase"))
1298 restrict_case = restrict_uppercase;
1299 else if (VAL_IS ("nocontrol"))
1300 restrict_ctrl = false;
1301 else if (VAL_IS ("ascii"))
1302 restrict_nonascii = true;
1305 fprintf (stderr, _("\
1306 %s: %s: Invalid restriction %s,\n\
1307 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1308 exec_name, com, quote (val));
1315 while (*val && *end);
1319 opt.restrict_files_os = restrict_os;
1320 opt.restrict_files_ctrl = restrict_ctrl;
1321 opt.restrict_files_case = restrict_case;
1322 opt.restrict_files_nonascii = restrict_nonascii;
1329 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1331 static const struct decode_item choices[] = {
1332 { "auto", secure_protocol_auto },
1333 { "sslv2", secure_protocol_sslv2 },
1334 { "sslv3", secure_protocol_sslv3 },
1335 { "tlsv1", secure_protocol_tlsv1 },
1337 int ok = decode_string (val, choices, countof (choices), place);
1339 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1344 /* Set all three timeout values. */
1347 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1350 if (!cmd_time (com, val, &value))
1352 opt.read_timeout = value;
1353 opt.connect_timeout = value;
1354 opt.dns_timeout = value;
1359 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1361 /* Disallow embedded newlines. */
1362 if (strchr (val, '\n'))
1364 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1365 exec_name, com, quote (val));
1368 xfree_null (opt.useragent);
1369 opt.useragent = xstrdup (val);
1373 /* The "verbose" option cannot be cmd_boolean because the variable is
1374 not bool -- it's of type int (-1 means uninitialized because of
1375 some random hackery for disallowing -q -v). */
1378 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1381 if (cmd_boolean (com, val, &flag))
1389 /* Miscellaneous useful routines. */
1391 /* A very simple atoi clone, more useful than atoi because it works on
1392 delimited strings, and has error reportage. Returns true on success,
1393 false on failure. If successful, stores result to *DEST. */
1396 simple_atoi (const char *beg, const char *end, int *dest)
1399 bool negative = false;
1400 const char *p = beg;
1402 while (p < end && c_isspace (*p))
1404 if (p < end && (*p == '-' || *p == '+'))
1406 negative = (*p == '-');
1412 /* Read negative numbers in a separate loop because the most
1413 negative integer cannot be represented as a positive number. */
1416 for (; p < end && c_isdigit (*p); p++)
1418 int next = (10 * result) + (*p - '0');
1420 return false; /* overflow */
1424 for (; p < end && c_isdigit (*p); p++)
1426 int next = (10 * result) - (*p - '0');
1428 return false; /* underflow */
1439 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1440 doesn't handle exponential notation. Returns true on success,
1441 false on failure. In case of success, stores its result to
1445 simple_atof (const char *beg, const char *end, double *dest)
1449 bool negative = false;
1450 bool seen_dot = false;
1451 bool seen_digit = false;
1454 const char *p = beg;
1456 while (p < end && c_isspace (*p))
1458 if (p < end && (*p == '-' || *p == '+'))
1460 negative = (*p == '-');
1464 for (; p < end; p++)
1470 result = (10 * result) + (ch - '0');
1472 result += (ch - '0') / (divider *= 10);
1494 /* Verify that the user-specified header in S is valid. It must
1495 contain a colon preceded by non-white-space characters and must not
1496 contain newlines. */
1499 check_user_specified_header (const char *s)
1503 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1505 /* The header MUST contain `:' preceded by at least one
1506 non-whitespace character. */
1507 if (*p != ':' || p == s)
1509 /* The header MUST NOT contain newlines. */
1510 if (strchr (s, '\n'))
1515 /* Decode VAL into a number, according to ITEMS. */
1518 decode_string (const char *val, const struct decode_item *items, int itemcount,
1522 for (i = 0; i < itemcount; i++)
1523 if (0 == strcasecmp (val, items[i].name))
1525 *place = items[i].code;
1532 void cleanup_html_url (void);
1535 /* Free the memory allocated by global variables. */
1539 /* Free external resources, close files, etc. */
1542 fclose (output_stream);
1543 /* No need to check for error because Wget flushes its output (and
1544 checks for errors) after any data arrives. */
1546 /* We're exiting anyway so there's no real need to call free()
1547 hundreds of times. Skipping the frees will make Wget exit
1550 However, when detecting leaks, it's crucial to free() everything
1551 because then you can find the real leaks, i.e. the allocated
1552 memory which grows with the size of the program. */
1558 cleanup_html_url ();
1563 extern acc_t *netrc_list;
1564 free_netrc (netrc_list);
1566 xfree_null (opt.lfilename);
1567 xfree_null (opt.dir_prefix);
1568 xfree_null (opt.input_filename);
1569 xfree_null (opt.output_document);
1570 free_vec (opt.accepts);
1571 free_vec (opt.rejects);
1572 free_vec (opt.excludes);
1573 free_vec (opt.includes);
1574 free_vec (opt.domains);
1575 free_vec (opt.follow_tags);
1576 free_vec (opt.ignore_tags);
1577 xfree_null (opt.progress_type);
1578 xfree_null (opt.ftp_user);
1579 xfree_null (opt.ftp_passwd);
1580 xfree_null (opt.ftp_proxy);
1581 xfree_null (opt.https_proxy);
1582 xfree_null (opt.http_proxy);
1583 free_vec (opt.no_proxy);
1584 xfree_null (opt.useragent);
1585 xfree_null (opt.referer);
1586 xfree_null (opt.http_user);
1587 xfree_null (opt.http_passwd);
1588 free_vec (opt.user_headers);
1590 xfree_null (opt.cert_file);
1591 xfree_null (opt.private_key);
1592 xfree_null (opt.ca_directory);
1593 xfree_null (opt.ca_cert);
1594 xfree_null (opt.random_file);
1595 xfree_null (opt.egd_file);
1597 xfree_null (opt.bind_address);
1598 xfree_null (opt.cookies_input);
1599 xfree_null (opt.cookies_output);
1600 xfree_null (opt.user);
1601 xfree_null (opt.passwd);
1602 xfree_null (opt.base_href);
1604 #endif /* DEBUG_MALLOC */
1607 /* Unit testing routines. */
1612 test_commands_sorted()
1614 int prev_idx = 0, next_idx = 1;
1615 int command_count = countof (commands) - 1;
1617 while (next_idx <= command_count)
1619 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1622 mu_assert ("FAILED", false);
1635 test_cmd_spec_restrict_file_names()
1640 int expected_restrict_files_os;
1641 int expected_restrict_files_ctrl;
1642 int expected_restrict_files_case;
1645 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1646 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1647 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1648 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1651 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1656 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1659 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1660 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1661 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1662 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1664 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1665 res == test_array[i].result
1666 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1667 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1668 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1674 #endif /* TESTING */