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>
49 # include <sys/socket.h>
50 # include <netinet/in.h>
52 # include <arpa/inet.h>
66 #include "cookies.h" /* for cookie_jar_delete */
73 extern struct cookie_jar *wget_cookie_jar;
75 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
76 otherwise, it will be performed by the shell. This variable will
77 be set by the wgetrc-reading function. */
79 static int enable_tilde_expansion;
82 #define CMD_DECLARE(func) static int func \
83 PARAMS ((const char *, const char *, void *))
85 CMD_DECLARE (cmd_boolean);
86 CMD_DECLARE (cmd_bytes);
87 CMD_DECLARE (cmd_bytes_large);
88 CMD_DECLARE (cmd_directory_vector);
89 CMD_DECLARE (cmd_lockable_boolean);
90 CMD_DECLARE (cmd_number);
91 CMD_DECLARE (cmd_number_inf);
92 CMD_DECLARE (cmd_string);
93 CMD_DECLARE (cmd_file);
94 CMD_DECLARE (cmd_directory);
95 CMD_DECLARE (cmd_time);
96 CMD_DECLARE (cmd_vector);
98 CMD_DECLARE (cmd_spec_dirstruct);
99 CMD_DECLARE (cmd_spec_header);
100 CMD_DECLARE (cmd_spec_htmlify);
101 CMD_DECLARE (cmd_spec_mirror);
102 CMD_DECLARE (cmd_spec_progress);
103 CMD_DECLARE (cmd_spec_recursive);
104 CMD_DECLARE (cmd_spec_restrict_file_names);
105 CMD_DECLARE (cmd_spec_timeout);
106 CMD_DECLARE (cmd_spec_useragent);
108 /* List of recognized commands, each consisting of name, closure and function.
109 When adding a new command, simply add it to the list, but be sure to keep the
110 list sorted alphabetically, as findcmd() depends on it. Also, be sure to add
111 any entries that allocate memory (e.g. cmd_string and cmd_vector guys) to the
112 cleanup() function below. */
116 int (*action) PARAMS ((const char *, const char *, void *));
118 { "accept", &opt.accepts, cmd_vector },
119 { "addhostdir", &opt.add_hostdir, cmd_boolean },
120 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
121 { "background", &opt.background, cmd_boolean },
122 { "backupconverted", &opt.backup_converted, cmd_boolean },
123 { "backups", &opt.backups, cmd_number },
124 { "base", &opt.base_href, cmd_string },
125 { "bindaddress", &opt.bind_address, cmd_string },
126 { "cache", &opt.allow_cache, cmd_boolean },
127 { "connecttimeout", &opt.connect_timeout, cmd_time },
128 { "continue", &opt.always_rest, cmd_boolean },
129 { "convertlinks", &opt.convert_links, cmd_boolean },
130 { "cookies", &opt.cookies, cmd_boolean },
131 { "cutdirs", &opt.cut_dirs, cmd_number },
133 { "debug", &opt.debug, cmd_boolean },
135 { "deleteafter", &opt.delete_after, cmd_boolean },
136 { "dirprefix", &opt.dir_prefix, cmd_directory },
137 { "dirstruct", NULL, cmd_spec_dirstruct },
138 { "dnscache", &opt.dns_cache, cmd_boolean },
139 { "dnstimeout", &opt.dns_timeout, cmd_time },
140 { "domains", &opt.domains, cmd_vector },
141 { "dotbytes", &opt.dot_bytes, cmd_bytes },
142 { "dotsinline", &opt.dots_in_line, cmd_number },
143 { "dotspacing", &opt.dot_spacing, cmd_number },
144 { "dotstyle", &opt.dot_style, cmd_string },
146 { "egdfile", &opt.sslegdsock, cmd_file },
148 { "excludedirectories", &opt.excludes, cmd_directory_vector },
149 { "excludedomains", &opt.exclude_domains, cmd_vector },
150 { "followftp", &opt.follow_ftp, cmd_boolean },
151 { "followtags", &opt.follow_tags, cmd_vector },
152 { "forcehtml", &opt.force_html, cmd_boolean },
153 { "ftpproxy", &opt.ftp_proxy, cmd_string },
154 { "glob", &opt.ftp_glob, cmd_boolean },
155 { "header", NULL, cmd_spec_header },
156 { "htmlextension", &opt.html_extension, cmd_boolean },
157 { "htmlify", NULL, cmd_spec_htmlify },
158 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
159 { "httppasswd", &opt.http_passwd, cmd_string },
160 { "httpproxy", &opt.http_proxy, cmd_string },
161 { "httpsproxy", &opt.https_proxy, cmd_string },
162 { "httpuser", &opt.http_user, cmd_string },
163 { "ignorelength", &opt.ignore_length, cmd_boolean },
164 { "ignoretags", &opt.ignore_tags, cmd_vector },
165 { "includedirectories", &opt.includes, cmd_directory_vector },
166 { "input", &opt.input_filename, cmd_file },
167 { "killlonger", &opt.kill_longer, cmd_boolean },
168 { "limitrate", &opt.limit_rate, cmd_bytes },
169 { "loadcookies", &opt.cookies_input, cmd_file },
170 { "logfile", &opt.lfilename, cmd_file },
171 { "login", &opt.ftp_acc, cmd_string },
172 { "mirror", NULL, cmd_spec_mirror },
173 { "netrc", &opt.netrc, cmd_boolean },
174 { "noclobber", &opt.noclobber, cmd_boolean },
175 { "noparent", &opt.no_parent, cmd_boolean },
176 { "noproxy", &opt.no_proxy, cmd_vector },
177 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
178 { "outputdocument", &opt.output_document, cmd_file },
179 { "pagerequisites", &opt.page_requisites, cmd_boolean },
180 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
181 { "passwd", &opt.ftp_pass, cmd_string },
182 { "postdata", &opt.post_data, cmd_string },
183 { "postfile", &opt.post_file_name, cmd_file },
184 { "progress", &opt.progress_type, cmd_spec_progress },
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 COM in the commands[] array and return its index. If COM
227 is not found, -1 is returned. This function uses binary search. */
230 findcmd (const char *com)
232 int lo = 0, hi = countof (commands) - 1;
236 int mid = (lo + hi) >> 1;
237 int cmp = strcasecmp (com, commands[mid].name);
248 /* Reset the variables to default values. */
254 /* Most of the default values are 0. Just reset everything, and
255 fill in the non-zero values. Note that initializing pointers to
256 NULL this way is technically illegal, but porting Wget to a
257 machine where NULL is not all-zero bit pattern will be the least
258 of the implementors' worries. */
266 opt.ftp_acc = xstrdup ("anonymous");
267 opt.ftp_pass = xstrdup ("-wget@");
271 opt.http_keep_alive = 1;
273 tmp = getenv ("no_proxy");
275 opt.no_proxy = sepstring (tmp);
278 opt.read_timeout = 900;
281 opt.remove_listing = 1;
283 opt.dot_bytes = 1024;
284 opt.dot_spacing = 10;
285 opt.dots_in_line = 50;
289 /* The default for file name restriction defaults to the OS type. */
290 #if !defined(WINDOWS) && !defined(__CYGWIN__)
291 opt.restrict_files_os = restrict_unix;
293 opt.restrict_files_os = restrict_windows;
295 opt.restrict_files_ctrl = 1;
298 /* Return the user's home directory (strdup-ed), or NULL if none is
303 char *home = getenv ("HOME");
308 /* If HOME is not defined, try getting it from the password
310 struct passwd *pwd = getpwuid (getuid ());
311 if (!pwd || !pwd->pw_dir)
316 /* #### Maybe I should grab home_dir from registry, but the best
317 that I could get from there is user's Start menu. It sucks! */
321 return home ? xstrdup (home) : NULL;
324 /* Return the path to the user's .wgetrc. This is either the value of
325 `WGETRC' environment variable, or `$HOME/.wgetrc'.
327 If the `WGETRC' variable exists but the file does not exist, the
328 function will exit(). */
330 wgetrc_file_name (void)
335 /* Try the environment. */
336 env = getenv ("WGETRC");
339 if (!file_exists_p (env))
341 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
345 return xstrdup (env);
349 /* If that failed, try $HOME/.wgetrc. */
353 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
354 sprintf (file, "%s/.wgetrc", home);
358 /* Under Windows, "home" is (for the purposes of this function) the
359 directory where `wget.exe' resides, and `wget.ini' will be used
360 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
362 It is not as trivial as I assumed, because on 95 argv[0] is full
363 path, but on NT you get what you typed in command line. --dbudor */
367 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
368 sprintf (file, "%swget.ini", home);
374 if (!file_exists_p (file))
382 static int parse_line PARAMS ((const char *, char **, char **, int *));
383 static int setval_internal PARAMS ((int, const char *, const char *));
385 /* Initialize variables from a wgetrc file. */
388 run_wgetrc (const char *file)
394 fp = fopen (file, "rb");
397 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
398 file, strerror (errno));
401 enable_tilde_expansion = 1;
403 while ((line = read_whole_line (fp)))
408 /* Parse the line. */
409 status = parse_line (line, &com, &val, &comind);
411 /* If everything is OK, set the value. */
414 if (!setval_internal (comind, com, val))
415 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
420 else if (status == 0)
421 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
425 enable_tilde_expansion = 0;
429 /* Initialize the defaults and run the system wgetrc and user's own
436 /* Load the hard-coded defaults. */
439 /* If SYSTEM_WGETRC is defined, use it. */
441 if (file_exists_p (SYSTEM_WGETRC))
442 run_wgetrc (SYSTEM_WGETRC);
444 /* Override it with your own, if one exists. */
445 file = wgetrc_file_name ();
448 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
449 something like realpath() before comparing them with `strcmp' */
451 if (!strcmp (file, SYSTEM_WGETRC))
453 fprintf (stderr, _("\
454 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
464 /* Remove dashes and underscores from S, modifying S in the
470 char *t = s; /* t - tortoise */
471 char *h = s; /* h - hare */
473 if (*h == '_' || *h == '-')
480 /* Parse the line pointed by line, with the syntax:
481 <sp>* command <sp>* = <sp>* value <sp>*
482 Uses malloc to allocate space for command and value.
483 If the line is invalid, data is freed and 0 is returned.
490 In case of success, *COM and *VAL point to freshly allocated
491 strings, and *COMIND points to com's index. In case of error or
492 empty line, those values are unaffected. */
495 parse_line (const char *line, char **com, char **val, int *comind)
498 const char *end = line + strlen (line);
499 const char *cmdstart, *cmdend;
500 const char *valstart, *valend;
505 /* Skip leading and trailing whitespace. */
506 while (*line && ISSPACE (*line))
508 while (end > line && ISSPACE (end[-1]))
511 /* Skip empty lines and comments. */
512 if (!*line || *line == '#')
518 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
522 /* Skip '=', as well as any space before or after it. */
523 while (p < end && ISSPACE (*p))
525 if (p == end || *p != '=')
528 while (p < end && ISSPACE (*p))
534 /* The line now known to be syntactically correct. Check whether
535 the command is valid. */
536 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
538 ind = findcmd (cmdcopy);
542 /* The command is valid. Now fill in the values and report success
545 *com = strdupdelim (cmdstart, cmdend);
546 *val = strdupdelim (valstart, valend);
550 /* Run commands[comind].action. */
553 setval_internal (int comind, const char *com, const char *val)
555 assert (0 <= comind && comind < countof (commands));
556 return ((*commands[comind].action) (com, val, commands[comind].closure));
559 /* Run command COM with value VAL. If running the command produces an
560 error, report the error and exit.
562 This is intended to be called from main() to modify Wget's behavior
563 through command-line switches. Since COM is hard-coded in main(),
564 it is not canonicalized, and this aborts when COM is not found.
566 If COMIND's are exported to init.h, this function will be changed
567 to accept COMIND directly. */
570 setoptval (const char *com, const char *val)
572 if (!setval_internal (findcmd (com), com, val))
576 /* Parse OPT into command and value and run it. For example,
577 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
578 This is used by the `--execute' flag in main.c. */
581 run_command (const char *opt)
585 int status = parse_line (opt, &com, &val, &comind);
588 if (!setval_internal (comind, com, val))
593 else if (status == 0)
595 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
601 /* Generic helper functions, for use with `commands'. */
603 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
605 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
606 && TOLOWER((p)[1]) == (c1) \
609 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
610 && TOLOWER((p)[1]) == (c1) \
611 && TOLOWER((p)[2]) == (c2) \
615 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
616 except for error messages. */
618 cmd_boolean (const char *com, const char *val, void *closure)
622 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
623 /* "on", "yes" and "1" mean true. */
625 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
626 /* "off", "no" and "0" mean false. */
631 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
632 exec_name, com, val);
636 *(int *)closure = bool_value;
640 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
641 COM is ignored, except for error messages. Values 2 and -1
642 indicate that once defined, the value may not be changed by
643 successive wgetrc files or command-line arguments.
645 Values: 2 - Enable a particular option for good ("always")
646 1 - Enable an option ("on")
647 0 - Disable an option ("off")
648 -1 - Disable an option for good ("never") */
650 cmd_lockable_boolean (const char *com, const char *val, void *closure)
652 int lockable_boolean_value;
654 int oldval = *(int *)closure;
657 * If a config file said "always" or "never", don't allow command line
658 * arguments to override the config file.
660 if (oldval == -1 || oldval == 2)
663 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
664 lockable_boolean_value = 2;
665 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
666 lockable_boolean_value = 1;
667 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
668 lockable_boolean_value = 0;
669 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
670 lockable_boolean_value = -1;
674 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
675 exec_name, com, val);
679 *(int *)closure = lockable_boolean_value;
683 static int simple_atoi PARAMS ((const char *, const char *, int *));
685 /* Set the non-negative integer value from VAL to CLOSURE. With
686 incorrect specification, the number remains unchanged. */
688 cmd_number (const char *com, const char *val, void *closure)
690 if (!simple_atoi (val, val + strlen (val), closure))
692 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
693 exec_name, com, val);
699 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
701 cmd_number_inf (const char *com, const char *val, void *closure)
703 if (!strcasecmp (val, "inf"))
708 return cmd_number (com, val, closure);
711 /* Copy (strdup) the string at COM to a new location and place a
712 pointer to *CLOSURE. */
714 cmd_string (const char *com, const char *val, void *closure)
716 char **pstring = (char **)closure;
718 FREE_MAYBE (*pstring);
719 *pstring = xstrdup (val);
724 # define ISSEP(c) ((c) == '/')
726 # define ISSEP(c) ((c) == '/' || (c) == '\\')
729 /* Like the above, but handles tilde-expansion when reading a user's
730 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
731 gets expanded to the user's home directory. */
733 cmd_file (const char *com, const char *val, void *closure)
735 char **pstring = (char **)closure;
737 FREE_MAYBE (*pstring);
739 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
741 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
744 *pstring = xstrdup (val);
750 char *home = home_dir ();
754 homelen = strlen (home);
755 while (homelen && ISSEP (home[homelen - 1]))
756 home[--homelen] = '\0';
758 /* Skip the leading "~/". */
759 for (++val; ISSEP (*val); val++)
762 result = xmalloc (homelen + 1 + strlen (val) + 1);
763 memcpy (result, home, homelen);
764 result[homelen] = '/';
765 strcpy (result + homelen + 1, val);
771 /* Convert "\" to "/". */
774 for (s = *pstring; *s; s++)
782 /* Like cmd_file, but strips trailing '/' characters. */
784 cmd_directory (const char *com, const char *val, void *closure)
788 /* Call cmd_file() for tilde expansion and separator
789 canonicalization (backslash -> slash under Windows). These
790 things should perhaps be in a separate function. */
791 if (!cmd_file (com, val, closure))
794 s = *(char **)closure;
796 while (t > s && *--t == '/')
802 /* Split VAL by space to a vector of values, and append those values
803 to vector pointed to by the CLOSURE argument. If VAL is empty, the
804 CLOSURE vector is cleared instead. */
807 cmd_vector (const char *com, const char *val, void *closure)
809 char ***pvec = (char ***)closure;
812 *pvec = merge_vecs (*pvec, sepstring (val));
822 cmd_directory_vector (const char *com, const char *val, void *closure)
824 char ***pvec = (char ***)closure;
828 /* Strip the trailing slashes from directories. */
831 seps = sepstring (val);
832 for (t = seps; t && *t; t++)
834 int len = strlen (*t);
835 /* Skip degenerate case of root directory. */
838 if ((*t)[len - 1] == '/')
839 (*t)[len - 1] = '\0';
842 *pvec = merge_vecs (*pvec, seps);
852 static int simple_atof PARAMS ((const char *, const char *, double *));
854 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
855 as "100k" or "2.5G" to a floating point number. */
858 parse_bytes_helper (const char *val, double *result)
861 const char *end = val + strlen (val);
863 /* Check for "inf". */
864 if (0 == strcmp (val, "inf"))
870 /* Strip trailing whitespace. */
871 while (val < end && ISSPACE (end[-1]))
876 switch (TOLOWER (end[-1]))
879 --end, mult = 1024.0;
882 --end, mult = 1048576.0;
885 --end, mult = 1073741824.0;
888 --end, mult = 1099511627776.0;
891 /* Not a recognized suffix: assume it's a digit. (If not,
892 simple_atof will raise an error.) */
896 /* Skip leading and trailing whitespace. */
897 while (val < end && ISSPACE (*val))
899 while (val < end && ISSPACE (end[-1]))
904 if (!simple_atof (val, end, &number))
907 *result = number * mult;
911 /* Parse VAL as a number and set its value to CLOSURE (which should
912 point to a long int).
914 By default, the value is assumed to be in bytes. If "K", "M", or
915 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
916 1<<30, respectively. Floating point values are allowed and are
917 cast to integer before use. The idea is to be able to use things
918 like 1.5k instead of "1536".
920 The string "inf" is returned as 0.
922 In case of error, 0 is returned and memory pointed to by CLOSURE
923 remains unmodified. */
926 cmd_bytes (const char *com, const char *val, void *closure)
929 if (!parse_bytes_helper (val, &byte_value))
931 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
932 exec_name, com, val);
935 *(long *)closure = (long)byte_value;
939 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
940 LARGE_INT. It works by converting the string to double, therefore
941 working with values up to 2^53-1 without loss of precision. This
942 value (8192 TB) is large enough to serve for a while. */
945 cmd_bytes_large (const char *com, const char *val, void *closure)
948 if (!parse_bytes_helper (val, &byte_value))
950 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
951 exec_name, com, val);
954 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
958 /* Store the value of VAL to *OUT. The value is a time period, by
959 default expressed in seconds, but also accepting suffixes "m", "h",
960 "d", and "w" for minutes, hours, days, and weeks respectively. */
963 cmd_time (const char *com, const char *val, void *closure)
966 const char *end = val + strlen (val);
968 /* Strip trailing whitespace. */
969 while (val < end && ISSPACE (end[-1]))
975 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
976 exec_name, com, val);
980 switch (TOLOWER (end[-1]))
983 --end, mult = 1; /* seconds */
986 --end, mult = 60; /* minutes */
989 --end, mult = 3600; /* hours */
992 --end, mult = 86400.0; /* days */
995 --end, mult = 604800.0; /* weeks */
998 /* Not a recognized suffix: assume it belongs to the number.
999 (If not, atof simple_atof will raise an error.) */
1003 /* Skip leading and trailing whitespace. */
1004 while (val < end && ISSPACE (*val))
1006 while (val < end && ISSPACE (end[-1]))
1011 if (!simple_atof (val, end, &number))
1014 *(double *)closure = number * mult;
1018 /* Specialized helper functions, used by `commands' to handle some
1019 options specially. */
1021 static int check_user_specified_header PARAMS ((const char *));
1024 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1026 if (!cmd_boolean (com, val, &opt.dirstruct))
1028 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1029 must be affected inversely. */
1031 opt.no_dirstruct = 0;
1033 opt.no_dirstruct = 1;
1038 cmd_spec_header (const char *com, const char *val, void *closure)
1042 /* Empty header means reset headers. */
1043 FREE_MAYBE (opt.user_header);
1044 opt.user_header = NULL;
1050 if (!check_user_specified_header (val))
1052 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1053 exec_name, com, val);
1056 i = opt.user_header ? strlen (opt.user_header) : 0;
1057 opt.user_header = (char *)xrealloc (opt.user_header, i + strlen (val)
1059 strcpy (opt.user_header + i, val);
1061 opt.user_header[i++] = '\r';
1062 opt.user_header[i++] = '\n';
1063 opt.user_header[i] = '\0';
1069 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1071 int flag = cmd_boolean (com, val, &opt.htmlify);
1072 if (flag && !opt.htmlify)
1073 opt.remove_listing = 0;
1077 /* Set the "mirror" mode. It means: recursive download, timestamping,
1078 no limit on max. recursion depth, and don't remove listings. */
1081 cmd_spec_mirror (const char *com, const char *val, void *closure)
1085 if (!cmd_boolean (com, val, &mirror))
1090 if (!opt.no_dirstruct)
1092 opt.timestamping = 1;
1093 opt.reclevel = INFINITE_RECURSION;
1094 opt.remove_listing = 0;
1099 /* Set progress.type to VAL, but verify that it's a valid progress
1100 implementation before that. */
1103 cmd_spec_progress (const char *com, const char *val, void *closure)
1105 if (!valid_progress_implementation_p (val))
1107 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1108 exec_name, com, val);
1111 FREE_MAYBE (opt.progress_type);
1113 /* Don't call set_progress_implementation here. It will be called
1114 in main() when it becomes clear what the log output is. */
1115 opt.progress_type = xstrdup (val);
1119 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1120 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1124 cmd_spec_recursive (const char *com, const char *val, void *closure)
1126 if (!cmd_boolean (com, val, &opt.recursive))
1130 if (opt.recursive && !opt.no_dirstruct)
1137 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1139 int restrict_os = opt.restrict_files_os;
1140 int restrict_ctrl = opt.restrict_files_ctrl;
1142 const char *end = strchr (val, ',');
1144 end = val + strlen (val);
1146 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1148 if (VAL_IS ("unix"))
1149 restrict_os = restrict_unix;
1150 else if (VAL_IS ("windows"))
1151 restrict_os = restrict_windows;
1152 else if (VAL_IS ("nocontrol"))
1158 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1159 exec_name, com, val);
1167 if (!strcmp (end + 1, "nocontrol"))
1173 opt.restrict_files_os = restrict_os;
1174 opt.restrict_files_ctrl = restrict_ctrl;
1178 /* Set all three timeout values. */
1181 cmd_spec_timeout (const char *com, const char *val, void *closure)
1184 if (!cmd_time (com, val, &value))
1186 opt.read_timeout = value;
1187 opt.connect_timeout = value;
1188 opt.dns_timeout = value;
1193 cmd_spec_useragent (const char *com, const char *val, void *closure)
1195 /* Just check for empty string and newline, so we don't throw total
1196 junk to the server. */
1197 if (!*val || strchr (val, '\n'))
1199 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1200 exec_name, com, val);
1203 opt.useragent = xstrdup (val);
1207 /* Miscellaneous useful routines. */
1209 /* A very simple atoi clone, more portable than strtol and friends,
1210 but reports errors, unlike atoi. Returns 1 on success, 0 on
1211 failure. In case of success, stores result to *DEST. */
1214 simple_atoi (const char *beg, const char *end, int *dest)
1222 for (p = beg; p < end && ISDIGIT (*p); p++)
1223 result = (10 * result) + (*p - '0');
1232 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1233 doesn't handle exponential notation. Returns 1 on success, 0 on
1234 failure. In case of success, stores its result to *DEST. */
1237 simple_atof (const char *beg, const char *end, double *dest)
1247 for (p = beg; p < end; p++)
1253 result = (10 * result) + (ch - '0');
1255 result += (ch - '0') / (divider *= 10);
1276 check_user_specified_header (const char *s)
1280 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1281 /* The header MUST contain `:' preceded by at least one
1282 non-whitespace character. */
1283 if (*p != ':' || p == s)
1285 /* The header MUST NOT contain newlines. */
1286 if (strchr (s, '\n'))
1291 void cleanup_html_url PARAMS ((void));
1292 void res_cleanup PARAMS ((void));
1293 void downloaded_files_free PARAMS ((void));
1294 void http_cleanup PARAMS ((void));
1297 /* Free the memory allocated by global variables. */
1301 /* Free external resources, close files, etc. */
1306 /* We're exiting anyway so there's no real need to call free()
1307 hundreds of times. Skipping the frees will make Wget exit
1310 However, when detecting leaks, it's crucial to free() everything
1311 because then you can find the real leaks, i.e. the allocated
1312 memory which grows with the size of the program. */
1318 cleanup_html_url ();
1319 downloaded_files_free ();
1321 cookie_jar_delete (wget_cookie_jar);
1324 extern acc_t *netrc_list;
1325 free_netrc (netrc_list);
1327 FREE_MAYBE (opt.lfilename);
1328 FREE_MAYBE (opt.dir_prefix);
1329 FREE_MAYBE (opt.input_filename);
1330 FREE_MAYBE (opt.output_document);
1331 free_vec (opt.accepts);
1332 free_vec (opt.rejects);
1333 free_vec (opt.excludes);
1334 free_vec (opt.includes);
1335 free_vec (opt.domains);
1336 free_vec (opt.follow_tags);
1337 free_vec (opt.ignore_tags);
1338 FREE_MAYBE (opt.progress_type);
1339 xfree (opt.ftp_acc);
1340 FREE_MAYBE (opt.ftp_pass);
1341 FREE_MAYBE (opt.ftp_proxy);
1342 FREE_MAYBE (opt.https_proxy);
1343 FREE_MAYBE (opt.http_proxy);
1344 free_vec (opt.no_proxy);
1345 FREE_MAYBE (opt.useragent);
1346 FREE_MAYBE (opt.referer);
1347 FREE_MAYBE (opt.http_user);
1348 FREE_MAYBE (opt.http_passwd);
1349 FREE_MAYBE (opt.user_header);
1351 FREE_MAYBE (opt.sslcertkey);
1352 FREE_MAYBE (opt.sslcertfile);
1353 #endif /* HAVE_SSL */
1354 FREE_MAYBE (opt.bind_address);
1355 FREE_MAYBE (opt.cookies_input);
1356 FREE_MAYBE (opt.cookies_output);