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 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
185 { "proxypasswd", &opt.proxy_passwd, cmd_string },
186 { "proxyuser", &opt.proxy_user, cmd_string },
187 { "quiet", &opt.quiet, cmd_boolean },
188 { "quota", &opt.quota, cmd_bytes_large },
189 { "randomwait", &opt.random_wait, cmd_boolean },
190 { "readtimeout", &opt.read_timeout, cmd_time },
191 { "reclevel", &opt.reclevel, cmd_number_inf },
192 { "recursive", NULL, cmd_spec_recursive },
193 { "referer", &opt.referer, cmd_string },
194 { "reject", &opt.rejects, cmd_vector },
195 { "relativeonly", &opt.relative_only, cmd_boolean },
196 { "removelisting", &opt.remove_listing, cmd_boolean },
197 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
198 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
199 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
200 { "robots", &opt.use_robots, cmd_boolean },
201 { "savecookies", &opt.cookies_output, cmd_file },
202 { "saveheaders", &opt.save_headers, cmd_boolean },
203 { "serverresponse", &opt.server_response, cmd_boolean },
204 { "spanhosts", &opt.spanhost, cmd_boolean },
205 { "spider", &opt.spider, cmd_boolean },
207 { "sslcadir", &opt.sslcadir, cmd_directory },
208 { "sslcafile", &opt.sslcafile, cmd_file },
209 { "sslcertfile", &opt.sslcertfile, cmd_file },
210 { "sslcertkey", &opt.sslcertkey, cmd_file },
211 { "sslcerttype", &opt.sslcerttype, cmd_number },
212 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
213 { "sslprotocol", &opt.sslprotocol, cmd_number },
214 #endif /* HAVE_SSL */
215 { "strictcomments", &opt.strict_comments, cmd_boolean },
216 { "timeout", NULL, cmd_spec_timeout },
217 { "timestamping", &opt.timestamping, cmd_boolean },
218 { "tries", &opt.ntry, cmd_number_inf },
219 { "useproxy", &opt.use_proxy, cmd_boolean },
220 { "useragent", NULL, cmd_spec_useragent },
221 { "verbose", &opt.verbose, cmd_boolean },
222 { "wait", &opt.wait, cmd_time },
223 { "waitretry", &opt.waitretry, cmd_time }
226 /* Look up CMDNAME in the commands[] and return its position in the
227 array. If CMDNAME is not found, return -1. */
230 command_by_name (const char *cmdname)
232 /* Use binary search for speed. Wget has ~100 commands, which
233 guarantees a worst case performance of 7 string comparisons. */
234 int lo = 0, hi = countof (commands) - 1;
238 int mid = (lo + hi) >> 1;
239 int cmp = strcasecmp (cmdname, commands[mid].name);
250 /* Reset the variables to default values. */
256 /* Most of the default values are 0. Just reset everything, and
257 fill in the non-zero values. Note that initializing pointers to
258 NULL this way is technically illegal, but porting Wget to a
259 machine where NULL is not all-zero bit pattern will be the least
260 of the implementors' worries. */
268 opt.ftp_acc = xstrdup ("anonymous");
269 opt.ftp_pass = xstrdup ("-wget@");
273 opt.http_keep_alive = 1;
275 tmp = getenv ("no_proxy");
277 opt.no_proxy = sepstring (tmp);
280 opt.read_timeout = 900;
283 opt.remove_listing = 1;
285 opt.dot_bytes = 1024;
286 opt.dot_spacing = 10;
287 opt.dots_in_line = 50;
291 /* The default for file name restriction defaults to the OS type. */
292 #if !defined(WINDOWS) && !defined(__CYGWIN__)
293 opt.restrict_files_os = restrict_unix;
295 opt.restrict_files_os = restrict_windows;
297 opt.restrict_files_ctrl = 1;
300 /* Return the user's home directory (strdup-ed), or NULL if none is
305 char *home = getenv ("HOME");
310 /* If HOME is not defined, try getting it from the password
312 struct passwd *pwd = getpwuid (getuid ());
313 if (!pwd || !pwd->pw_dir)
318 /* #### Maybe I should grab home_dir from registry, but the best
319 that I could get from there is user's Start menu. It sucks! */
323 return home ? xstrdup (home) : NULL;
326 /* Return the path to the user's .wgetrc. This is either the value of
327 `WGETRC' environment variable, or `$HOME/.wgetrc'.
329 If the `WGETRC' variable exists but the file does not exist, the
330 function will exit(). */
332 wgetrc_file_name (void)
337 /* Try the environment. */
338 env = getenv ("WGETRC");
341 if (!file_exists_p (env))
343 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
347 return xstrdup (env);
351 /* If that failed, try $HOME/.wgetrc. */
355 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
356 sprintf (file, "%s/.wgetrc", home);
360 /* Under Windows, "home" is (for the purposes of this function) the
361 directory where `wget.exe' resides, and `wget.ini' will be used
362 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
364 It is not as trivial as I assumed, because on 95 argv[0] is full
365 path, but on NT you get what you typed in command line. --dbudor */
369 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
370 sprintf (file, "%swget.ini", home);
376 if (!file_exists_p (file))
384 static int parse_line PARAMS ((const char *, char **, char **, int *));
385 static int setval_internal PARAMS ((int, const char *, const char *));
387 /* Initialize variables from a wgetrc file. */
390 run_wgetrc (const char *file)
396 fp = fopen (file, "rb");
399 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
400 file, strerror (errno));
403 enable_tilde_expansion = 1;
405 while ((line = read_whole_line (fp)))
410 /* Parse the line. */
411 status = parse_line (line, &com, &val, &comind);
413 /* If everything is OK, set the value. */
416 if (!setval_internal (comind, com, val))
417 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
422 else if (status == 0)
423 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
427 enable_tilde_expansion = 0;
431 /* Initialize the defaults and run the system wgetrc and user's own
438 /* Load the hard-coded defaults. */
441 /* If SYSTEM_WGETRC is defined, use it. */
443 if (file_exists_p (SYSTEM_WGETRC))
444 run_wgetrc (SYSTEM_WGETRC);
446 /* Override it with your own, if one exists. */
447 file = wgetrc_file_name ();
450 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
451 something like realpath() before comparing them with `strcmp' */
453 if (!strcmp (file, SYSTEM_WGETRC))
455 fprintf (stderr, _("\
456 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
466 /* Remove dashes and underscores from S, modifying S in the
472 char *t = s; /* t - tortoise */
473 char *h = s; /* h - hare */
475 if (*h == '_' || *h == '-')
482 /* Parse the line pointed by line, with the syntax:
483 <sp>* command <sp>* = <sp>* value <sp>*
484 Uses malloc to allocate space for command and value.
485 If the line is invalid, data is freed and 0 is returned.
492 In case of success, *COM and *VAL point to freshly allocated
493 strings, and *COMIND points to com's index. In case of error or
494 empty line, those values are unaffected. */
497 parse_line (const char *line, char **com, char **val, int *comind)
500 const char *end = line + strlen (line);
501 const char *cmdstart, *cmdend;
502 const char *valstart, *valend;
507 /* Skip leading and trailing whitespace. */
508 while (*line && ISSPACE (*line))
510 while (end > line && ISSPACE (end[-1]))
513 /* Skip empty lines and comments. */
514 if (!*line || *line == '#')
520 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
524 /* Skip '=', as well as any space before or after it. */
525 while (p < end && ISSPACE (*p))
527 if (p == end || *p != '=')
530 while (p < end && ISSPACE (*p))
536 /* The line now known to be syntactically correct. Check whether
537 the command is valid. */
538 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
540 ind = command_by_name (cmdcopy);
544 /* The command is valid. Now fill in the values and report success
547 *com = strdupdelim (cmdstart, cmdend);
548 *val = strdupdelim (valstart, valend);
552 /* Run commands[comind].action. */
555 setval_internal (int comind, const char *com, const char *val)
557 assert (0 <= comind && comind < countof (commands));
558 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
559 return ((*commands[comind].action) (com, val, commands[comind].closure));
562 /* Run command COM with value VAL. If running the command produces an
563 error, report the error and exit.
565 This is intended to be called from main() to modify Wget's behavior
566 through command-line switches. Since COM is hard-coded in main(),
567 it is not canonicalized, and this aborts when COM is not found.
569 If COMIND's are exported to init.h, this function will be changed
570 to accept COMIND directly. */
573 setoptval (const char *com, const char *val)
575 assert (val != NULL);
576 if (!setval_internal (command_by_name (com), com, val))
580 /* Parse OPT into command and value and run it. For example,
581 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
582 This is used by the `--execute' flag in main.c. */
585 run_command (const char *opt)
589 int status = parse_line (opt, &com, &val, &comind);
592 if (!setval_internal (comind, com, val))
597 else if (status == 0)
599 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
605 /* Generic helper functions, for use with `commands'. */
607 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
609 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
610 && TOLOWER((p)[1]) == (c1) \
613 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
614 && TOLOWER((p)[1]) == (c1) \
615 && TOLOWER((p)[2]) == (c2) \
619 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
620 except for error messages. */
622 cmd_boolean (const char *com, const char *val, void *closure)
626 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
627 /* "on", "yes" and "1" mean true. */
629 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
630 /* "off", "no" and "0" mean false. */
635 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
636 exec_name, com, val);
640 *(int *)closure = bool_value;
644 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
645 COM is ignored, except for error messages. Values 2 and -1
646 indicate that once defined, the value may not be changed by
647 successive wgetrc files or command-line arguments.
649 Values: 2 - Enable a particular option for good ("always")
650 1 - Enable an option ("on")
651 0 - Disable an option ("off")
652 -1 - Disable an option for good ("never") */
654 cmd_lockable_boolean (const char *com, const char *val, void *closure)
656 int lockable_boolean_value;
658 int oldval = *(int *)closure;
661 * If a config file said "always" or "never", don't allow command line
662 * arguments to override the config file.
664 if (oldval == -1 || oldval == 2)
667 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
668 lockable_boolean_value = 2;
669 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
670 lockable_boolean_value = 1;
671 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
672 lockable_boolean_value = 0;
673 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
674 lockable_boolean_value = -1;
678 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
679 exec_name, com, val);
683 *(int *)closure = lockable_boolean_value;
687 static int simple_atoi PARAMS ((const char *, const char *, int *));
689 /* Set the non-negative integer value from VAL to CLOSURE. With
690 incorrect specification, the number remains unchanged. */
692 cmd_number (const char *com, const char *val, void *closure)
694 if (!simple_atoi (val, val + strlen (val), closure))
696 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
697 exec_name, com, val);
703 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
705 cmd_number_inf (const char *com, const char *val, void *closure)
707 if (!strcasecmp (val, "inf"))
712 return cmd_number (com, val, closure);
715 /* Copy (strdup) the string at COM to a new location and place a
716 pointer to *CLOSURE. */
718 cmd_string (const char *com, const char *val, void *closure)
720 char **pstring = (char **)closure;
722 xfree_null (*pstring);
723 *pstring = xstrdup (val);
728 # define ISSEP(c) ((c) == '/')
730 # define ISSEP(c) ((c) == '/' || (c) == '\\')
733 /* Like the above, but handles tilde-expansion when reading a user's
734 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
735 gets expanded to the user's home directory. */
737 cmd_file (const char *com, const char *val, void *closure)
739 char **pstring = (char **)closure;
741 xfree_null (*pstring);
743 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
745 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
748 *pstring = xstrdup (val);
754 char *home = home_dir ();
758 homelen = strlen (home);
759 while (homelen && ISSEP (home[homelen - 1]))
760 home[--homelen] = '\0';
762 /* Skip the leading "~/". */
763 for (++val; ISSEP (*val); val++)
766 result = xmalloc (homelen + 1 + strlen (val) + 1);
767 memcpy (result, home, homelen);
768 result[homelen] = '/';
769 strcpy (result + homelen + 1, val);
775 /* Convert "\" to "/". */
778 for (s = *pstring; *s; s++)
786 /* Like cmd_file, but strips trailing '/' characters. */
788 cmd_directory (const char *com, const char *val, void *closure)
792 /* Call cmd_file() for tilde expansion and separator
793 canonicalization (backslash -> slash under Windows). These
794 things should perhaps be in a separate function. */
795 if (!cmd_file (com, val, closure))
798 s = *(char **)closure;
800 while (t > s && *--t == '/')
806 /* Split VAL by space to a vector of values, and append those values
807 to vector pointed to by the CLOSURE argument. If VAL is empty, the
808 CLOSURE vector is cleared instead. */
811 cmd_vector (const char *com, const char *val, void *closure)
813 char ***pvec = (char ***)closure;
816 *pvec = merge_vecs (*pvec, sepstring (val));
826 cmd_directory_vector (const char *com, const char *val, void *closure)
828 char ***pvec = (char ***)closure;
832 /* Strip the trailing slashes from directories. */
835 seps = sepstring (val);
836 for (t = seps; t && *t; t++)
838 int len = strlen (*t);
839 /* Skip degenerate case of root directory. */
842 if ((*t)[len - 1] == '/')
843 (*t)[len - 1] = '\0';
846 *pvec = merge_vecs (*pvec, seps);
856 static int simple_atof PARAMS ((const char *, const char *, double *));
858 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
859 as "100k" or "2.5G" to a floating point number. */
862 parse_bytes_helper (const char *val, double *result)
865 const char *end = val + strlen (val);
867 /* Check for "inf". */
868 if (0 == strcmp (val, "inf"))
874 /* Strip trailing whitespace. */
875 while (val < end && ISSPACE (end[-1]))
880 switch (TOLOWER (end[-1]))
883 --end, mult = 1024.0;
886 --end, mult = 1048576.0;
889 --end, mult = 1073741824.0;
892 --end, mult = 1099511627776.0;
895 /* Not a recognized suffix: assume it's a digit. (If not,
896 simple_atof will raise an error.) */
900 /* Skip leading and trailing whitespace. */
901 while (val < end && ISSPACE (*val))
903 while (val < end && ISSPACE (end[-1]))
908 if (!simple_atof (val, end, &number))
911 *result = number * mult;
915 /* Parse VAL as a number and set its value to CLOSURE (which should
916 point to a long int).
918 By default, the value is assumed to be in bytes. If "K", "M", or
919 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
920 1<<30, respectively. Floating point values are allowed and are
921 cast to integer before use. The idea is to be able to use things
922 like 1.5k instead of "1536".
924 The string "inf" is returned as 0.
926 In case of error, 0 is returned and memory pointed to by CLOSURE
927 remains unmodified. */
930 cmd_bytes (const char *com, const char *val, void *closure)
933 if (!parse_bytes_helper (val, &byte_value))
935 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
936 exec_name, com, val);
939 *(long *)closure = (long)byte_value;
943 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
944 LARGE_INT. It works by converting the string to double, therefore
945 working with values up to 2^53-1 without loss of precision. This
946 value (8192 TB) is large enough to serve for a while. */
949 cmd_bytes_large (const char *com, const char *val, void *closure)
952 if (!parse_bytes_helper (val, &byte_value))
954 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
955 exec_name, com, val);
958 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
962 /* Store the value of VAL to *OUT. The value is a time period, by
963 default expressed in seconds, but also accepting suffixes "m", "h",
964 "d", and "w" for minutes, hours, days, and weeks respectively. */
967 cmd_time (const char *com, const char *val, void *closure)
970 const char *end = val + strlen (val);
972 /* Strip trailing whitespace. */
973 while (val < end && ISSPACE (end[-1]))
979 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
980 exec_name, com, val);
984 switch (TOLOWER (end[-1]))
987 --end, mult = 1; /* seconds */
990 --end, mult = 60; /* minutes */
993 --end, mult = 3600; /* hours */
996 --end, mult = 86400.0; /* days */
999 --end, mult = 604800.0; /* weeks */
1002 /* Not a recognized suffix: assume it belongs to the number.
1003 (If not, atof simple_atof will raise an error.) */
1007 /* Skip leading and trailing whitespace. */
1008 while (val < end && ISSPACE (*val))
1010 while (val < end && ISSPACE (end[-1]))
1015 if (!simple_atof (val, end, &number))
1018 *(double *)closure = number * mult;
1022 /* Specialized helper functions, used by `commands' to handle some
1023 options specially. */
1025 static int check_user_specified_header PARAMS ((const char *));
1028 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1030 if (!cmd_boolean (com, val, &opt.dirstruct))
1032 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1033 must be affected inversely. */
1035 opt.no_dirstruct = 0;
1037 opt.no_dirstruct = 1;
1042 cmd_spec_header (const char *com, const char *val, void *closure)
1044 if (!check_user_specified_header (val))
1046 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1047 exec_name, com, val);
1050 return cmd_vector (com, val, closure);
1054 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1056 int flag = cmd_boolean (com, val, &opt.htmlify);
1057 if (flag && !opt.htmlify)
1058 opt.remove_listing = 0;
1062 /* Set the "mirror" mode. It means: recursive download, timestamping,
1063 no limit on max. recursion depth, and don't remove listings. */
1066 cmd_spec_mirror (const char *com, const char *val, void *closure)
1070 if (!cmd_boolean (com, val, &mirror))
1075 if (!opt.no_dirstruct)
1077 opt.timestamping = 1;
1078 opt.reclevel = INFINITE_RECURSION;
1079 opt.remove_listing = 0;
1084 /* Set progress.type to VAL, but verify that it's a valid progress
1085 implementation before that. */
1088 cmd_spec_progress (const char *com, const char *val, void *closure)
1090 if (!valid_progress_implementation_p (val))
1092 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1093 exec_name, com, val);
1096 xfree_null (opt.progress_type);
1098 /* Don't call set_progress_implementation here. It will be called
1099 in main() when it becomes clear what the log output is. */
1100 opt.progress_type = xstrdup (val);
1104 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1105 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1109 cmd_spec_recursive (const char *com, const char *val, void *closure)
1111 if (!cmd_boolean (com, val, &opt.recursive))
1115 if (opt.recursive && !opt.no_dirstruct)
1122 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1124 int restrict_os = opt.restrict_files_os;
1125 int restrict_ctrl = opt.restrict_files_ctrl;
1127 const char *end = strchr (val, ',');
1129 end = val + strlen (val);
1131 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1133 if (VAL_IS ("unix"))
1134 restrict_os = restrict_unix;
1135 else if (VAL_IS ("windows"))
1136 restrict_os = restrict_windows;
1137 else if (VAL_IS ("nocontrol"))
1143 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1144 exec_name, com, val);
1152 if (!strcmp (end + 1, "nocontrol"))
1158 opt.restrict_files_os = restrict_os;
1159 opt.restrict_files_ctrl = restrict_ctrl;
1163 /* Set all three timeout values. */
1166 cmd_spec_timeout (const char *com, const char *val, void *closure)
1169 if (!cmd_time (com, val, &value))
1171 opt.read_timeout = value;
1172 opt.connect_timeout = value;
1173 opt.dns_timeout = value;
1178 cmd_spec_useragent (const char *com, const char *val, void *closure)
1180 /* Just check for empty string and newline, so we don't throw total
1181 junk to the server. */
1182 if (!*val || strchr (val, '\n'))
1184 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1185 exec_name, com, val);
1188 opt.useragent = xstrdup (val);
1192 /* Miscellaneous useful routines. */
1194 /* A very simple atoi clone, more portable than strtol and friends,
1195 but reports errors, unlike atoi. Returns 1 on success, 0 on
1196 failure. In case of success, stores result to *DEST. */
1199 simple_atoi (const char *beg, const char *end, int *dest)
1207 for (p = beg; p < end && ISDIGIT (*p); p++)
1208 result = (10 * result) + (*p - '0');
1217 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1218 doesn't handle exponential notation. Returns 1 on success, 0 on
1219 failure. In case of success, stores its result to *DEST. */
1222 simple_atof (const char *beg, const char *end, double *dest)
1232 for (p = beg; p < end; p++)
1238 result = (10 * result) + (ch - '0');
1240 result += (ch - '0') / (divider *= 10);
1261 check_user_specified_header (const char *s)
1265 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1266 /* The header MUST contain `:' preceded by at least one
1267 non-whitespace character. */
1268 if (*p != ':' || p == s)
1270 /* The header MUST NOT contain newlines. */
1271 if (strchr (s, '\n'))
1276 void cleanup_html_url PARAMS ((void));
1277 void res_cleanup PARAMS ((void));
1278 void downloaded_files_free PARAMS ((void));
1279 void http_cleanup PARAMS ((void));
1282 /* Free the memory allocated by global variables. */
1286 /* Free external resources, close files, etc. */
1289 extern FILE *output_stream;
1291 fclose (output_stream);
1292 /* No need to check for error because Wget flushes its output (and
1293 checks for errors) after any data arrives. */
1296 /* We're exiting anyway so there's no real need to call free()
1297 hundreds of times. Skipping the frees will make Wget exit
1300 However, when detecting leaks, it's crucial to free() everything
1301 because then you can find the real leaks, i.e. the allocated
1302 memory which grows with the size of the program. */
1308 cleanup_html_url ();
1309 downloaded_files_free ();
1311 if (wget_cookie_jar)
1312 cookie_jar_delete (wget_cookie_jar);
1315 extern acc_t *netrc_list;
1316 free_netrc (netrc_list);
1318 xfree_null (opt.lfilename);
1319 xfree_null (opt.dir_prefix);
1320 xfree_null (opt.input_filename);
1321 xfree_null (opt.output_document);
1322 free_vec (opt.accepts);
1323 free_vec (opt.rejects);
1324 free_vec (opt.excludes);
1325 free_vec (opt.includes);
1326 free_vec (opt.domains);
1327 free_vec (opt.follow_tags);
1328 free_vec (opt.ignore_tags);
1329 xfree_null (opt.progress_type);
1330 xfree (opt.ftp_acc);
1331 xfree_null (opt.ftp_pass);
1332 xfree_null (opt.ftp_proxy);
1333 xfree_null (opt.https_proxy);
1334 xfree_null (opt.http_proxy);
1335 free_vec (opt.no_proxy);
1336 xfree_null (opt.useragent);
1337 xfree_null (opt.referer);
1338 xfree_null (opt.http_user);
1339 xfree_null (opt.http_passwd);
1340 xfree_null (opt.user_header);
1342 xfree_null (opt.sslcertkey);
1343 xfree_null (opt.sslcertfile);
1344 #endif /* HAVE_SSL */
1345 xfree_null (opt.bind_address);
1346 xfree_null (opt.cookies_input);
1347 xfree_null (opt.cookies_output);