1 /* Reading/parsing the initialization file.
2 Copyright (C) 1995, 1996, 1997, 1998, 2000, 2001, 2003
3 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 2 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, write to the Free Software
19 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
21 In addition, as a special exception, the Free Software Foundation
22 gives permission to link the code of its release of Wget with the
23 OpenSSL project's "OpenSSL" library (or with modified versions of it
24 that use the same license as the "OpenSSL" library), and distribute
25 the linked executables. You must obey the GNU General Public License
26 in all respects for all of the code used other than "OpenSSL". If you
27 modify this file, you may extend this exception to your version of the
28 file, but you are not obligated to do so. If you do not wish to do
29 so, delete this exception statement from your version. */
34 #include <sys/types.h>
56 #include "cookies.h" /* for cookie_jar_delete */
58 #include "recur.h" /* for INFINITE_RECURSION */
64 extern struct cookie_jar *wget_cookie_jar;
66 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
67 otherwise, it will be performed by the shell. This variable will
68 be set by the wgetrc-reading function. */
70 static int enable_tilde_expansion;
73 #define CMD_DECLARE(func) static int func \
74 PARAMS ((const char *, const char *, void *))
76 CMD_DECLARE (cmd_boolean);
77 CMD_DECLARE (cmd_bytes);
78 CMD_DECLARE (cmd_bytes_large);
79 CMD_DECLARE (cmd_directory_vector);
80 CMD_DECLARE (cmd_lockable_boolean);
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_progress);
94 CMD_DECLARE (cmd_spec_recursive);
95 CMD_DECLARE (cmd_spec_restrict_file_names);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
99 /* List of recognized commands, each consisting of name, closure and
100 function. When adding a new command, simply add it to the list,
101 but be sure to keep the list sorted alphabetically, as
102 command_by_name depends on it. Also, be sure to add any entries
103 that allocate memory (e.g. cmd_string and cmd_vector guys) to the
104 cleanup() function below. */
109 int (*action) PARAMS ((const char *, const char *, void *));
111 { "accept", &opt.accepts, cmd_vector },
112 { "addhostdir", &opt.add_hostdir, cmd_boolean },
113 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
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 },
119 { "cache", &opt.allow_cache, cmd_boolean },
120 { "connecttimeout", &opt.connect_timeout, cmd_time },
121 { "continue", &opt.always_rest, cmd_boolean },
122 { "convertlinks", &opt.convert_links, cmd_boolean },
123 { "cookies", &opt.cookies, cmd_boolean },
124 { "cutdirs", &opt.cut_dirs, cmd_number },
126 { "debug", &opt.debug, cmd_boolean },
128 { "deleteafter", &opt.delete_after, cmd_boolean },
129 { "dirprefix", &opt.dir_prefix, cmd_directory },
130 { "dirstruct", NULL, cmd_spec_dirstruct },
131 { "dnscache", &opt.dns_cache, cmd_boolean },
132 { "dnstimeout", &opt.dns_timeout, cmd_time },
133 { "domains", &opt.domains, cmd_vector },
134 { "dotbytes", &opt.dot_bytes, cmd_bytes },
135 { "dotsinline", &opt.dots_in_line, cmd_number },
136 { "dotspacing", &opt.dot_spacing, cmd_number },
137 { "dotstyle", &opt.dot_style, cmd_string },
139 { "egdfile", &opt.sslegdsock, cmd_file },
141 { "excludedirectories", &opt.excludes, cmd_directory_vector },
142 { "excludedomains", &opt.exclude_domains, cmd_vector },
143 { "followftp", &opt.follow_ftp, cmd_boolean },
144 { "followtags", &opt.follow_tags, cmd_vector },
145 { "forcehtml", &opt.force_html, cmd_boolean },
146 { "ftpproxy", &opt.ftp_proxy, cmd_string },
147 { "glob", &opt.ftp_glob, cmd_boolean },
148 { "header", &opt.user_headers, cmd_spec_header },
149 { "htmlextension", &opt.html_extension, cmd_boolean },
150 { "htmlify", NULL, cmd_spec_htmlify },
151 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
152 { "httppasswd", &opt.http_passwd, cmd_string },
153 { "httpproxy", &opt.http_proxy, cmd_string },
154 { "httpsproxy", &opt.https_proxy, cmd_string },
155 { "httpuser", &opt.http_user, cmd_string },
156 { "ignorelength", &opt.ignore_length, cmd_boolean },
157 { "ignoretags", &opt.ignore_tags, cmd_vector },
158 { "includedirectories", &opt.includes, cmd_directory_vector },
160 { "inet4only", &opt.ipv4_only, cmd_boolean },
161 { "inet6only", &opt.ipv6_only, cmd_boolean },
163 { "input", &opt.input_filename, cmd_file },
164 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
165 { "killlonger", &opt.kill_longer, cmd_boolean },
166 { "limitrate", &opt.limit_rate, cmd_bytes },
167 { "loadcookies", &opt.cookies_input, cmd_file },
168 { "logfile", &opt.lfilename, cmd_file },
169 { "login", &opt.ftp_acc, cmd_string },
170 { "mirror", NULL, cmd_spec_mirror },
171 { "netrc", &opt.netrc, cmd_boolean },
172 { "noclobber", &opt.noclobber, cmd_boolean },
173 { "noparent", &opt.no_parent, cmd_boolean },
174 { "noproxy", &opt.no_proxy, cmd_vector },
175 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
176 { "outputdocument", &opt.output_document, cmd_file },
177 { "pagerequisites", &opt.page_requisites, cmd_boolean },
178 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
179 { "passwd", &opt.ftp_pass, cmd_string },
180 { "postdata", &opt.post_data, cmd_string },
181 { "postfile", &opt.post_file_name, cmd_file },
182 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
183 { "progress", &opt.progress_type, cmd_spec_progress },
184 { "proxypasswd", &opt.proxy_passwd, cmd_string },
185 { "proxyuser", &opt.proxy_user, cmd_string },
186 { "quiet", &opt.quiet, cmd_boolean },
187 { "quota", &opt.quota, cmd_bytes_large },
188 { "randomwait", &opt.random_wait, cmd_boolean },
189 { "readtimeout", &opt.read_timeout, cmd_time },
190 { "reclevel", &opt.reclevel, cmd_number_inf },
191 { "recursive", NULL, cmd_spec_recursive },
192 { "referer", &opt.referer, cmd_string },
193 { "reject", &opt.rejects, cmd_vector },
194 { "relativeonly", &opt.relative_only, cmd_boolean },
195 { "removelisting", &opt.remove_listing, cmd_boolean },
196 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
197 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
198 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
199 { "robots", &opt.use_robots, cmd_boolean },
200 { "savecookies", &opt.cookies_output, cmd_file },
201 { "saveheaders", &opt.save_headers, cmd_boolean },
202 { "serverresponse", &opt.server_response, cmd_boolean },
203 { "spanhosts", &opt.spanhost, cmd_boolean },
204 { "spider", &opt.spider, cmd_boolean },
206 { "sslcadir", &opt.sslcadir, cmd_directory },
207 { "sslcafile", &opt.sslcafile, cmd_file },
208 { "sslcertfile", &opt.sslcertfile, cmd_file },
209 { "sslcertkey", &opt.sslcertkey, cmd_file },
210 { "sslcerttype", &opt.sslcerttype, cmd_number },
211 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
212 { "sslprotocol", &opt.sslprotocol, cmd_number },
213 #endif /* HAVE_SSL */
214 { "strictcomments", &opt.strict_comments, cmd_boolean },
215 { "timeout", NULL, cmd_spec_timeout },
216 { "timestamping", &opt.timestamping, cmd_boolean },
217 { "tries", &opt.ntry, cmd_number_inf },
218 { "useproxy", &opt.use_proxy, cmd_boolean },
219 { "useragent", NULL, cmd_spec_useragent },
220 { "verbose", &opt.verbose, cmd_boolean },
221 { "wait", &opt.wait, cmd_time },
222 { "waitretry", &opt.waitretry, cmd_time }
225 /* Look up CMDNAME in the commands[] and return its position in the
226 array. If CMDNAME is not found, return -1. */
229 command_by_name (const char *cmdname)
231 /* Use binary search for speed. Wget has ~100 commands, which
232 guarantees a worst case performance of 7 string comparisons. */
233 int lo = 0, hi = countof (commands) - 1;
237 int mid = (lo + hi) >> 1;
238 int cmp = strcasecmp (cmdname, commands[mid].name);
249 /* Reset the variables to default values. */
255 /* Most of the default values are 0. Just reset everything, and
256 fill in the non-zero values. Note that initializing pointers to
257 NULL this way is technically illegal, but porting Wget to a
258 machine where NULL is not all-zero bit pattern will be the least
259 of the implementors' worries. */
267 opt.ftp_acc = xstrdup ("anonymous");
268 opt.ftp_pass = xstrdup ("-wget@");
272 opt.http_keep_alive = 1;
274 tmp = getenv ("no_proxy");
276 opt.no_proxy = sepstring (tmp);
279 opt.read_timeout = 900;
282 opt.remove_listing = 1;
284 opt.dot_bytes = 1024;
285 opt.dot_spacing = 10;
286 opt.dots_in_line = 50;
290 /* The default for file name restriction defaults to the OS type. */
291 #if !defined(WINDOWS) && !defined(__CYGWIN__)
292 opt.restrict_files_os = restrict_unix;
294 opt.restrict_files_os = restrict_windows;
296 opt.restrict_files_ctrl = 1;
299 /* Return the user's home directory (strdup-ed), or NULL if none is
304 char *home = getenv ("HOME");
309 /* If HOME is not defined, try getting it from the password
311 struct passwd *pwd = getpwuid (getuid ());
312 if (!pwd || !pwd->pw_dir)
317 /* #### Maybe I should grab home_dir from registry, but the best
318 that I could get from there is user's Start menu. It sucks! */
322 return home ? xstrdup (home) : NULL;
325 /* Return the path to the user's .wgetrc. This is either the value of
326 `WGETRC' environment variable, or `$HOME/.wgetrc'.
328 If the `WGETRC' variable exists but the file does not exist, the
329 function will exit(). */
331 wgetrc_file_name (void)
336 /* Try the environment. */
337 env = getenv ("WGETRC");
340 if (!file_exists_p (env))
342 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
346 return xstrdup (env);
350 /* If that failed, try $HOME/.wgetrc. */
354 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
355 sprintf (file, "%s/.wgetrc", home);
359 /* Under Windows, "home" is (for the purposes of this function) the
360 directory where `wget.exe' resides, and `wget.ini' will be used
361 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
363 It is not as trivial as I assumed, because on 95 argv[0] is full
364 path, but on NT you get what you typed in command line. --dbudor */
368 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
369 sprintf (file, "%swget.ini", home);
375 if (!file_exists_p (file))
383 static int parse_line PARAMS ((const char *, char **, char **, int *));
384 static int setval_internal PARAMS ((int, const char *, const char *));
386 /* Initialize variables from a wgetrc file. */
389 run_wgetrc (const char *file)
395 fp = fopen (file, "rb");
398 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
399 file, strerror (errno));
402 enable_tilde_expansion = 1;
404 while ((line = read_whole_line (fp)))
409 /* Parse the line. */
410 status = parse_line (line, &com, &val, &comind);
412 /* If everything is OK, set the value. */
415 if (!setval_internal (comind, com, val))
416 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
421 else if (status == 0)
422 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
426 enable_tilde_expansion = 0;
430 /* Initialize the defaults and run the system wgetrc and user's own
437 /* Load the hard-coded defaults. */
440 /* If SYSTEM_WGETRC is defined, use it. */
442 if (file_exists_p (SYSTEM_WGETRC))
443 run_wgetrc (SYSTEM_WGETRC);
445 /* Override it with your own, if one exists. */
446 file = wgetrc_file_name ();
449 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
450 something like realpath() before comparing them with `strcmp' */
452 if (!strcmp (file, SYSTEM_WGETRC))
454 fprintf (stderr, _("\
455 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
465 /* Remove dashes and underscores from S, modifying S in the
471 char *t = s; /* t - tortoise */
472 char *h = s; /* h - hare */
474 if (*h == '_' || *h == '-')
481 /* Parse the line pointed by line, with the syntax:
482 <sp>* command <sp>* = <sp>* value <sp>*
483 Uses malloc to allocate space for command and value.
484 If the line is invalid, data is freed and 0 is returned.
491 In case of success, *COM and *VAL point to freshly allocated
492 strings, and *COMIND points to com's index. In case of error or
493 empty line, those values are unaffected. */
496 parse_line (const char *line, char **com, char **val, int *comind)
499 const char *end = line + strlen (line);
500 const char *cmdstart, *cmdend;
501 const char *valstart, *valend;
506 /* Skip leading and trailing whitespace. */
507 while (*line && ISSPACE (*line))
509 while (end > line && ISSPACE (end[-1]))
512 /* Skip empty lines and comments. */
513 if (!*line || *line == '#')
519 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
523 /* Skip '=', as well as any space before or after it. */
524 while (p < end && ISSPACE (*p))
526 if (p == end || *p != '=')
529 while (p < end && ISSPACE (*p))
535 /* The line now known to be syntactically correct. Check whether
536 the command is valid. */
537 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
539 ind = command_by_name (cmdcopy);
543 /* The command is valid. Now fill in the values and report success
546 *com = strdupdelim (cmdstart, cmdend);
547 *val = strdupdelim (valstart, valend);
551 /* Run commands[comind].action. */
554 setval_internal (int comind, const char *com, const char *val)
556 assert (0 <= comind && comind < countof (commands));
557 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
558 return ((*commands[comind].action) (com, val, commands[comind].closure));
561 /* Run command COM with value VAL. If running the command produces an
562 error, report the error and exit.
564 This is intended to be called from main() to modify Wget's behavior
565 through command-line switches. Since COM is hard-coded in main(),
566 it is not canonicalized, and this aborts when COM is not found.
568 If COMIND's are exported to init.h, this function will be changed
569 to accept COMIND directly. */
572 setoptval (const char *com, const char *val)
574 assert (val != NULL);
575 if (!setval_internal (command_by_name (com), com, val))
579 /* Parse OPT into command and value and run it. For example,
580 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
581 This is used by the `--execute' flag in main.c. */
584 run_command (const char *opt)
588 int status = parse_line (opt, &com, &val, &comind);
591 if (!setval_internal (comind, com, val))
596 else if (status == 0)
598 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
604 /* Generic helper functions, for use with `commands'. */
606 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
608 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
609 && TOLOWER((p)[1]) == (c1) \
612 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
613 && TOLOWER((p)[1]) == (c1) \
614 && TOLOWER((p)[2]) == (c2) \
618 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
619 except for error messages. */
621 cmd_boolean (const char *com, const char *val, void *closure)
625 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
626 /* "on", "yes" and "1" mean true. */
628 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
629 /* "off", "no" and "0" mean false. */
634 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
635 exec_name, com, val);
639 *(int *)closure = bool_value;
643 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
644 COM is ignored, except for error messages. Values 2 and -1
645 indicate that once defined, the value may not be changed by
646 successive wgetrc files or command-line arguments.
648 Values: 2 - Enable a particular option for good ("always")
649 1 - Enable an option ("on")
650 0 - Disable an option ("off")
651 -1 - Disable an option for good ("never") */
653 cmd_lockable_boolean (const char *com, const char *val, void *closure)
655 int lockable_boolean_value;
657 int oldval = *(int *)closure;
660 * If a config file said "always" or "never", don't allow command line
661 * arguments to override the config file.
663 if (oldval == -1 || oldval == 2)
666 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
667 lockable_boolean_value = 2;
668 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
669 lockable_boolean_value = 1;
670 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
671 lockable_boolean_value = 0;
672 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
673 lockable_boolean_value = -1;
677 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
678 exec_name, com, val);
682 *(int *)closure = lockable_boolean_value;
686 static int simple_atoi PARAMS ((const char *, const char *, int *));
688 /* Set the non-negative integer value from VAL to CLOSURE. With
689 incorrect specification, the number remains unchanged. */
691 cmd_number (const char *com, const char *val, void *closure)
693 if (!simple_atoi (val, val + strlen (val), closure))
695 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
696 exec_name, com, val);
702 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
704 cmd_number_inf (const char *com, const char *val, void *closure)
706 if (!strcasecmp (val, "inf"))
711 return cmd_number (com, val, closure);
714 /* Copy (strdup) the string at COM to a new location and place a
715 pointer to *CLOSURE. */
717 cmd_string (const char *com, const char *val, void *closure)
719 char **pstring = (char **)closure;
721 xfree_null (*pstring);
722 *pstring = xstrdup (val);
727 # define ISSEP(c) ((c) == '/')
729 # define ISSEP(c) ((c) == '/' || (c) == '\\')
732 /* Like the above, but handles tilde-expansion when reading a user's
733 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
734 gets expanded to the user's home directory. */
736 cmd_file (const char *com, const char *val, void *closure)
738 char **pstring = (char **)closure;
740 xfree_null (*pstring);
742 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
744 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
747 *pstring = xstrdup (val);
753 char *home = home_dir ();
757 homelen = strlen (home);
758 while (homelen && ISSEP (home[homelen - 1]))
759 home[--homelen] = '\0';
761 /* Skip the leading "~/". */
762 for (++val; ISSEP (*val); val++)
765 result = xmalloc (homelen + 1 + strlen (val) + 1);
766 memcpy (result, home, homelen);
767 result[homelen] = '/';
768 strcpy (result + homelen + 1, val);
774 /* Convert "\" to "/". */
777 for (s = *pstring; *s; s++)
785 /* Like cmd_file, but strips trailing '/' characters. */
787 cmd_directory (const char *com, const char *val, void *closure)
791 /* Call cmd_file() for tilde expansion and separator
792 canonicalization (backslash -> slash under Windows). These
793 things should perhaps be in a separate function. */
794 if (!cmd_file (com, val, closure))
797 s = *(char **)closure;
799 while (t > s && *--t == '/')
805 /* Split VAL by space to a vector of values, and append those values
806 to vector pointed to by the CLOSURE argument. If VAL is empty, the
807 CLOSURE vector is cleared instead. */
810 cmd_vector (const char *com, const char *val, void *closure)
812 char ***pvec = (char ***)closure;
815 *pvec = merge_vecs (*pvec, sepstring (val));
825 cmd_directory_vector (const char *com, const char *val, void *closure)
827 char ***pvec = (char ***)closure;
831 /* Strip the trailing slashes from directories. */
834 seps = sepstring (val);
835 for (t = seps; t && *t; t++)
837 int len = strlen (*t);
838 /* Skip degenerate case of root directory. */
841 if ((*t)[len - 1] == '/')
842 (*t)[len - 1] = '\0';
845 *pvec = merge_vecs (*pvec, seps);
855 static int simple_atof PARAMS ((const char *, const char *, double *));
857 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
858 as "100k" or "2.5G" to a floating point number. */
861 parse_bytes_helper (const char *val, double *result)
864 const char *end = val + strlen (val);
866 /* Check for "inf". */
867 if (0 == strcmp (val, "inf"))
873 /* Strip trailing whitespace. */
874 while (val < end && ISSPACE (end[-1]))
879 switch (TOLOWER (end[-1]))
882 --end, mult = 1024.0;
885 --end, mult = 1048576.0;
888 --end, mult = 1073741824.0;
891 --end, mult = 1099511627776.0;
894 /* Not a recognized suffix: assume it's a digit. (If not,
895 simple_atof will raise an error.) */
899 /* Skip leading and trailing whitespace. */
900 while (val < end && ISSPACE (*val))
902 while (val < end && ISSPACE (end[-1]))
907 if (!simple_atof (val, end, &number))
910 *result = number * mult;
914 /* Parse VAL as a number and set its value to CLOSURE (which should
915 point to a long int).
917 By default, the value is assumed to be in bytes. If "K", "M", or
918 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
919 1<<30, respectively. Floating point values are allowed and are
920 cast to integer before use. The idea is to be able to use things
921 like 1.5k instead of "1536".
923 The string "inf" is returned as 0.
925 In case of error, 0 is returned and memory pointed to by CLOSURE
926 remains unmodified. */
929 cmd_bytes (const char *com, const char *val, void *closure)
932 if (!parse_bytes_helper (val, &byte_value))
934 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
935 exec_name, com, val);
938 *(long *)closure = (long)byte_value;
942 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
943 LARGE_INT. It works by converting the string to double, therefore
944 working with values up to 2^53-1 without loss of precision. This
945 value (8192 TB) is large enough to serve for a while. */
948 cmd_bytes_large (const char *com, const char *val, void *closure)
951 if (!parse_bytes_helper (val, &byte_value))
953 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
954 exec_name, com, val);
957 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
961 /* Store the value of VAL to *OUT. The value is a time period, by
962 default expressed in seconds, but also accepting suffixes "m", "h",
963 "d", and "w" for minutes, hours, days, and weeks respectively. */
966 cmd_time (const char *com, const char *val, void *closure)
969 const char *end = val + strlen (val);
971 /* Strip trailing whitespace. */
972 while (val < end && ISSPACE (end[-1]))
978 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
979 exec_name, com, val);
983 switch (TOLOWER (end[-1]))
986 --end, mult = 1; /* seconds */
989 --end, mult = 60; /* minutes */
992 --end, mult = 3600; /* hours */
995 --end, mult = 86400.0; /* days */
998 --end, mult = 604800.0; /* weeks */
1001 /* Not a recognized suffix: assume it belongs to the number.
1002 (If not, atof simple_atof will raise an error.) */
1006 /* Skip leading and trailing whitespace. */
1007 while (val < end && ISSPACE (*val))
1009 while (val < end && ISSPACE (end[-1]))
1014 if (!simple_atof (val, end, &number))
1017 *(double *)closure = number * mult;
1021 /* Specialized helper functions, used by `commands' to handle some
1022 options specially. */
1024 static int check_user_specified_header PARAMS ((const char *));
1027 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1029 if (!cmd_boolean (com, val, &opt.dirstruct))
1031 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1032 must be affected inversely. */
1034 opt.no_dirstruct = 0;
1036 opt.no_dirstruct = 1;
1041 cmd_spec_header (const char *com, const char *val, void *closure)
1043 if (!check_user_specified_header (val))
1045 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1046 exec_name, com, val);
1049 return cmd_vector (com, val, closure);
1053 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1055 int flag = cmd_boolean (com, val, &opt.htmlify);
1056 if (flag && !opt.htmlify)
1057 opt.remove_listing = 0;
1061 /* Set the "mirror" mode. It means: recursive download, timestamping,
1062 no limit on max. recursion depth, and don't remove listings. */
1065 cmd_spec_mirror (const char *com, const char *val, void *closure)
1069 if (!cmd_boolean (com, val, &mirror))
1074 if (!opt.no_dirstruct)
1076 opt.timestamping = 1;
1077 opt.reclevel = INFINITE_RECURSION;
1078 opt.remove_listing = 0;
1083 /* Set progress.type to VAL, but verify that it's a valid progress
1084 implementation before that. */
1087 cmd_spec_progress (const char *com, const char *val, void *closure)
1089 if (!valid_progress_implementation_p (val))
1091 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1092 exec_name, com, val);
1095 xfree_null (opt.progress_type);
1097 /* Don't call set_progress_implementation here. It will be called
1098 in main() when it becomes clear what the log output is. */
1099 opt.progress_type = xstrdup (val);
1103 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1104 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1108 cmd_spec_recursive (const char *com, const char *val, void *closure)
1110 if (!cmd_boolean (com, val, &opt.recursive))
1114 if (opt.recursive && !opt.no_dirstruct)
1121 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1123 int restrict_os = opt.restrict_files_os;
1124 int restrict_ctrl = opt.restrict_files_ctrl;
1126 const char *end = strchr (val, ',');
1128 end = val + strlen (val);
1130 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1132 if (VAL_IS ("unix"))
1133 restrict_os = restrict_unix;
1134 else if (VAL_IS ("windows"))
1135 restrict_os = restrict_windows;
1136 else if (VAL_IS ("nocontrol"))
1142 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1143 exec_name, com, val);
1151 if (!strcmp (end + 1, "nocontrol"))
1157 opt.restrict_files_os = restrict_os;
1158 opt.restrict_files_ctrl = restrict_ctrl;
1162 /* Set all three timeout values. */
1165 cmd_spec_timeout (const char *com, const char *val, void *closure)
1168 if (!cmd_time (com, val, &value))
1170 opt.read_timeout = value;
1171 opt.connect_timeout = value;
1172 opt.dns_timeout = value;
1177 cmd_spec_useragent (const char *com, const char *val, void *closure)
1179 /* Just check for empty string and newline, so we don't throw total
1180 junk to the server. */
1181 if (!*val || strchr (val, '\n'))
1183 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1184 exec_name, com, val);
1187 opt.useragent = xstrdup (val);
1191 /* Miscellaneous useful routines. */
1193 /* A very simple atoi clone, more portable than strtol and friends,
1194 but reports errors, unlike atoi. Returns 1 on success, 0 on
1195 failure. In case of success, stores result to *DEST. */
1198 simple_atoi (const char *beg, const char *end, int *dest)
1206 for (p = beg; p < end && ISDIGIT (*p); p++)
1207 result = (10 * result) + (*p - '0');
1216 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1217 doesn't handle exponential notation. Returns 1 on success, 0 on
1218 failure. In case of success, stores its result to *DEST. */
1221 simple_atof (const char *beg, const char *end, double *dest)
1231 for (p = beg; p < end; p++)
1237 result = (10 * result) + (ch - '0');
1239 result += (ch - '0') / (divider *= 10);
1260 check_user_specified_header (const char *s)
1264 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1265 /* The header MUST contain `:' preceded by at least one
1266 non-whitespace character. */
1267 if (*p != ':' || p == s)
1269 /* The header MUST NOT contain newlines. */
1270 if (strchr (s, '\n'))
1275 void cleanup_html_url PARAMS ((void));
1276 void res_cleanup PARAMS ((void));
1277 void downloaded_files_free PARAMS ((void));
1278 void http_cleanup PARAMS ((void));
1281 /* Free the memory allocated by global variables. */
1285 /* Free external resources, close files, etc. */
1288 extern FILE *output_stream;
1290 fclose (output_stream);
1291 /* No need to check for error because Wget flushes its output (and
1292 checks for errors) after any data arrives. */
1295 /* We're exiting anyway so there's no real need to call free()
1296 hundreds of times. Skipping the frees will make Wget exit
1299 However, when detecting leaks, it's crucial to free() everything
1300 because then you can find the real leaks, i.e. the allocated
1301 memory which grows with the size of the program. */
1307 cleanup_html_url ();
1308 downloaded_files_free ();
1310 if (wget_cookie_jar)
1311 cookie_jar_delete (wget_cookie_jar);
1314 extern acc_t *netrc_list;
1315 free_netrc (netrc_list);
1317 xfree_null (opt.lfilename);
1318 xfree_null (opt.dir_prefix);
1319 xfree_null (opt.input_filename);
1320 xfree_null (opt.output_document);
1321 free_vec (opt.accepts);
1322 free_vec (opt.rejects);
1323 free_vec (opt.excludes);
1324 free_vec (opt.includes);
1325 free_vec (opt.domains);
1326 free_vec (opt.follow_tags);
1327 free_vec (opt.ignore_tags);
1328 xfree_null (opt.progress_type);
1329 xfree (opt.ftp_acc);
1330 xfree_null (opt.ftp_pass);
1331 xfree_null (opt.ftp_proxy);
1332 xfree_null (opt.https_proxy);
1333 xfree_null (opt.http_proxy);
1334 free_vec (opt.no_proxy);
1335 xfree_null (opt.useragent);
1336 xfree_null (opt.referer);
1337 xfree_null (opt.http_user);
1338 xfree_null (opt.http_passwd);
1339 xfree_null (opt.user_header);
1341 xfree_null (opt.sslcertkey);
1342 xfree_null (opt.sslcertfile);
1343 #endif /* HAVE_SSL */
1344 xfree_null (opt.bind_address);
1345 xfree_null (opt.cookies_input);
1346 xfree_null (opt.cookies_output);