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. */
259 memset (&opt, 0, sizeof (opt));
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 = findcmd (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 return ((*commands[comind].action) (com, val, commands[comind].closure));
560 /* Run command COM with value VAL. If running the command produces an
561 error, report the error and exit.
563 This is intended to be called from main() to modify Wget's behavior
564 through command-line switches. Since COM is hard-coded in main(),
565 it is not canonicalized, and this aborts when COM is not found.
567 If COMIND's are exported to init.h, this function will be changed
568 to accept COMIND directly. */
571 setoptval (const char *com, const char *val)
573 if (!setval_internal (findcmd (com), com, val))
577 /* Parse OPT into command and value and run it. For example,
578 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
579 This is used by the `--execute' flag in main.c. */
582 run_command (const char *opt)
586 int status = parse_line (opt, &com, &val, &comind);
589 if (!setval_internal (comind, com, val))
594 else if (status == 0)
596 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
602 /* Generic helper functions, for use with `commands'. */
604 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
606 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
607 && TOLOWER((p)[1]) == (c1) \
610 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
611 && TOLOWER((p)[1]) == (c1) \
612 && TOLOWER((p)[2]) == (c2) \
616 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
617 except for error messages. */
619 cmd_boolean (const char *com, const char *val, void *closure)
623 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
624 /* "on", "yes" and "1" mean true. */
626 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
627 /* "off", "no" and "0" mean false. */
632 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
633 exec_name, com, val);
637 *(int *)closure = bool_value;
641 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
642 COM is ignored, except for error messages. Values 2 and -1
643 indicate that once defined, the value may not be changed by
644 successive wgetrc files or command-line arguments.
646 Values: 2 - Enable a particular option for good ("always")
647 1 - Enable an option ("on")
648 0 - Disable an option ("off")
649 -1 - Disable an option for good ("never") */
651 cmd_lockable_boolean (const char *com, const char *val, void *closure)
653 int lockable_boolean_value;
655 int oldval = *(int *)closure;
658 * If a config file said "always" or "never", don't allow command line
659 * arguments to override the config file.
661 if (oldval == -1 || oldval == 2)
664 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
665 lockable_boolean_value = 2;
666 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
667 lockable_boolean_value = 1;
668 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
669 lockable_boolean_value = 0;
670 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
671 lockable_boolean_value = -1;
675 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
676 exec_name, com, val);
680 *(int *)closure = lockable_boolean_value;
684 static int simple_atoi PARAMS ((const char *, const char *, int *));
686 /* Set the non-negative integer value from VAL to CLOSURE. With
687 incorrect specification, the number remains unchanged. */
689 cmd_number (const char *com, const char *val, void *closure)
691 if (!simple_atoi (val, val + strlen (val), closure))
693 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
694 exec_name, com, val);
700 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
702 cmd_number_inf (const char *com, const char *val, void *closure)
704 if (!strcasecmp (val, "inf"))
709 return cmd_number (com, val, closure);
712 /* Copy (strdup) the string at COM to a new location and place a
713 pointer to *CLOSURE. */
715 cmd_string (const char *com, const char *val, void *closure)
717 char **pstring = (char **)closure;
719 FREE_MAYBE (*pstring);
720 *pstring = xstrdup (val);
725 # define ISSEP(c) ((c) == '/')
727 # define ISSEP(c) ((c) == '/' || (c) == '\\')
730 /* Like the above, but handles tilde-expansion when reading a user's
731 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
732 gets expanded to the user's home directory. */
734 cmd_file (const char *com, const char *val, void *closure)
736 char **pstring = (char **)closure;
738 FREE_MAYBE (*pstring);
740 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
742 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
745 *pstring = xstrdup (val);
751 char *home = home_dir ();
755 homelen = strlen (home);
756 while (homelen && ISSEP (home[homelen - 1]))
757 home[--homelen] = '\0';
759 /* Skip the leading "~/". */
760 for (++val; ISSEP (*val); val++)
763 result = xmalloc (homelen + 1 + strlen (val) + 1);
764 memcpy (result, home, homelen);
765 result[homelen] = '/';
766 strcpy (result + homelen + 1, val);
772 /* Convert "\" to "/". */
775 for (s = *pstring; *s; s++)
783 /* Like cmd_file, but strips trailing '/' characters. */
785 cmd_directory (const char *com, const char *val, void *closure)
789 /* Call cmd_file() for tilde expansion and separator
790 canonicalization (backslash -> slash under Windows). These
791 things should perhaps be in a separate function. */
792 if (!cmd_file (com, val, closure))
795 s = *(char **)closure;
797 while (t > s && *--t == '/')
803 /* Split VAL by space to a vector of values, and append those values
804 to vector pointed to by the CLOSURE argument. If VAL is empty, the
805 CLOSURE vector is cleared instead. */
808 cmd_vector (const char *com, const char *val, void *closure)
810 char ***pvec = (char ***)closure;
813 *pvec = merge_vecs (*pvec, sepstring (val));
823 cmd_directory_vector (const char *com, const char *val, void *closure)
825 char ***pvec = (char ***)closure;
829 /* Strip the trailing slashes from directories. */
832 seps = sepstring (val);
833 for (t = seps; t && *t; t++)
835 int len = strlen (*t);
836 /* Skip degenerate case of root directory. */
839 if ((*t)[len - 1] == '/')
840 (*t)[len - 1] = '\0';
843 *pvec = merge_vecs (*pvec, seps);
853 static int simple_atof PARAMS ((const char *, const char *, double *));
855 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
856 as "100k" or "2.5G" to a floating point number. */
859 parse_bytes_helper (const char *val, double *result)
862 const char *end = val + strlen (val);
864 /* Check for "inf". */
865 if (0 == strcmp (val, "inf"))
871 /* Strip trailing whitespace. */
872 while (val < end && ISSPACE (end[-1]))
877 switch (TOLOWER (end[-1]))
880 --end, mult = 1024.0;
883 --end, mult = 1048576.0;
886 --end, mult = 1073741824.0;
889 --end, mult = 1099511627776.0;
892 /* Not a recognized suffix: assume it's a digit. (If not,
893 simple_atof will raise an error.) */
897 /* Skip leading and trailing whitespace. */
898 while (val < end && ISSPACE (*val))
900 while (val < end && ISSPACE (end[-1]))
905 if (!simple_atof (val, end, &number))
908 *result = number * mult;
912 /* Parse VAL as a number and set its value to CLOSURE (which should
913 point to a long int).
915 By default, the value is assumed to be in bytes. If "K", "M", or
916 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
917 1<<30, respectively. Floating point values are allowed and are
918 cast to integer before use. The idea is to be able to use things
919 like 1.5k instead of "1536".
921 The string "inf" is returned as 0.
923 In case of error, 0 is returned and memory pointed to by CLOSURE
924 remains unmodified. */
927 cmd_bytes (const char *com, const char *val, void *closure)
930 if (!parse_bytes_helper (val, &byte_value))
932 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
933 exec_name, com, val);
936 *(long *)closure = (long)byte_value;
940 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
941 LARGE_INT. It works by converting the string to double, therefore
942 working with values up to 2^53-1 without loss of precision. This
943 value (8192 TB) is large enough to serve for a while. */
946 cmd_bytes_large (const char *com, const char *val, void *closure)
949 if (!parse_bytes_helper (val, &byte_value))
951 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
952 exec_name, com, val);
955 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
959 /* Store the value of VAL to *OUT. The value is a time period, by
960 default expressed in seconds, but also accepting suffixes "m", "h",
961 "d", and "w" for minutes, hours, days, and weeks respectively. */
964 cmd_time (const char *com, const char *val, void *closure)
967 const char *end = val + strlen (val);
969 /* Strip trailing whitespace. */
970 while (val < end && ISSPACE (end[-1]))
976 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
977 exec_name, com, val);
981 switch (TOLOWER (end[-1]))
984 --end, mult = 1; /* seconds */
987 --end, mult = 60; /* minutes */
990 --end, mult = 3600; /* hours */
993 --end, mult = 86400.0; /* days */
996 --end, mult = 604800.0; /* weeks */
999 /* Not a recognized suffix: assume it belongs to the number.
1000 (If not, atof simple_atof will raise an error.) */
1004 /* Skip leading and trailing whitespace. */
1005 while (val < end && ISSPACE (*val))
1007 while (val < end && ISSPACE (end[-1]))
1012 if (!simple_atof (val, end, &number))
1015 *(double *)closure = number * mult;
1019 /* Specialized helper functions, used by `commands' to handle some
1020 options specially. */
1022 static int check_user_specified_header PARAMS ((const char *));
1025 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1027 if (!cmd_boolean (com, val, &opt.dirstruct))
1029 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1030 must be affected inversely. */
1032 opt.no_dirstruct = 0;
1034 opt.no_dirstruct = 1;
1039 cmd_spec_header (const char *com, const char *val, void *closure)
1043 /* Empty header means reset headers. */
1044 FREE_MAYBE (opt.user_header);
1045 opt.user_header = NULL;
1051 if (!check_user_specified_header (val))
1053 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1054 exec_name, com, val);
1057 i = opt.user_header ? strlen (opt.user_header) : 0;
1058 opt.user_header = (char *)xrealloc (opt.user_header, i + strlen (val)
1060 strcpy (opt.user_header + i, val);
1062 opt.user_header[i++] = '\r';
1063 opt.user_header[i++] = '\n';
1064 opt.user_header[i] = '\0';
1070 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1072 int flag = cmd_boolean (com, val, &opt.htmlify);
1073 if (flag && !opt.htmlify)
1074 opt.remove_listing = 0;
1078 /* Set the "mirror" mode. It means: recursive download, timestamping,
1079 no limit on max. recursion depth, and don't remove listings. */
1082 cmd_spec_mirror (const char *com, const char *val, void *closure)
1086 if (!cmd_boolean (com, val, &mirror))
1091 if (!opt.no_dirstruct)
1093 opt.timestamping = 1;
1094 opt.reclevel = INFINITE_RECURSION;
1095 opt.remove_listing = 0;
1100 /* Set progress.type to VAL, but verify that it's a valid progress
1101 implementation before that. */
1104 cmd_spec_progress (const char *com, const char *val, void *closure)
1106 if (!valid_progress_implementation_p (val))
1108 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1109 exec_name, com, val);
1112 FREE_MAYBE (opt.progress_type);
1114 /* Don't call set_progress_implementation here. It will be called
1115 in main() when it becomes clear what the log output is. */
1116 opt.progress_type = xstrdup (val);
1120 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1121 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1125 cmd_spec_recursive (const char *com, const char *val, void *closure)
1127 if (!cmd_boolean (com, val, &opt.recursive))
1131 if (opt.recursive && !opt.no_dirstruct)
1138 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1140 int restrict_os = opt.restrict_files_os;
1141 int restrict_ctrl = opt.restrict_files_ctrl;
1143 const char *end = strchr (val, ',');
1145 end = val + strlen (val);
1147 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1149 if (VAL_IS ("unix"))
1150 restrict_os = restrict_unix;
1151 else if (VAL_IS ("windows"))
1152 restrict_os = restrict_windows;
1153 else if (VAL_IS ("nocontrol"))
1159 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1160 exec_name, com, val);
1168 if (!strcmp (end + 1, "nocontrol"))
1174 opt.restrict_files_os = restrict_os;
1175 opt.restrict_files_ctrl = restrict_ctrl;
1179 /* Set all three timeout values. */
1182 cmd_spec_timeout (const char *com, const char *val, void *closure)
1185 if (!cmd_time (com, val, &value))
1187 opt.read_timeout = value;
1188 opt.connect_timeout = value;
1189 opt.dns_timeout = value;
1194 cmd_spec_useragent (const char *com, const char *val, void *closure)
1196 /* Just check for empty string and newline, so we don't throw total
1197 junk to the server. */
1198 if (!*val || strchr (val, '\n'))
1200 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1201 exec_name, com, val);
1204 opt.useragent = xstrdup (val);
1208 /* Miscellaneous useful routines. */
1210 /* A very simple atoi clone, more portable than strtol and friends,
1211 but reports errors, unlike atoi. Returns 1 on success, 0 on
1212 failure. In case of success, stores result to *DEST. */
1215 simple_atoi (const char *beg, const char *end, int *dest)
1223 for (p = beg; p < end && ISDIGIT (*p); p++)
1224 result = (10 * result) + (*p - '0');
1233 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1234 doesn't handle exponential notation. Returns 1 on success, 0 on
1235 failure. In case of success, stores its result to *DEST. */
1238 simple_atof (const char *beg, const char *end, double *dest)
1248 for (p = beg; p < end; p++)
1254 result = (10 * result) + (ch - '0');
1256 result += (ch - '0') / (divider *= 10);
1277 check_user_specified_header (const char *s)
1281 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1282 /* The header MUST contain `:' preceded by at least one
1283 non-whitespace character. */
1284 if (*p != ':' || p == s)
1286 /* The header MUST NOT contain newlines. */
1287 if (strchr (s, '\n'))
1292 void cleanup_html_url PARAMS ((void));
1293 void res_cleanup PARAMS ((void));
1294 void downloaded_files_free PARAMS ((void));
1295 void http_cleanup PARAMS ((void));
1298 /* Free the memory allocated by global variables. */
1302 /* Free external resources, close files, etc. */
1307 /* We're exiting anyway so there's no real need to call free()
1308 hundreds of times. Skipping the frees will make Wget exit
1311 However, when detecting leaks, it's crucial to free() everything
1312 because then you can find the real leaks, i.e. the allocated
1313 memory which grows with the size of the program. */
1319 cleanup_html_url ();
1320 downloaded_files_free ();
1322 cookie_jar_delete (wget_cookie_jar);
1325 extern acc_t *netrc_list;
1326 free_netrc (netrc_list);
1328 FREE_MAYBE (opt.lfilename);
1329 FREE_MAYBE (opt.dir_prefix);
1330 FREE_MAYBE (opt.input_filename);
1331 FREE_MAYBE (opt.output_document);
1332 free_vec (opt.accepts);
1333 free_vec (opt.rejects);
1334 free_vec (opt.excludes);
1335 free_vec (opt.includes);
1336 free_vec (opt.domains);
1337 free_vec (opt.follow_tags);
1338 free_vec (opt.ignore_tags);
1339 FREE_MAYBE (opt.progress_type);
1340 xfree (opt.ftp_acc);
1341 FREE_MAYBE (opt.ftp_pass);
1342 FREE_MAYBE (opt.ftp_proxy);
1343 FREE_MAYBE (opt.https_proxy);
1344 FREE_MAYBE (opt.http_proxy);
1345 free_vec (opt.no_proxy);
1346 FREE_MAYBE (opt.useragent);
1347 FREE_MAYBE (opt.referer);
1348 FREE_MAYBE (opt.http_user);
1349 FREE_MAYBE (opt.http_passwd);
1350 FREE_MAYBE (opt.user_header);
1352 FREE_MAYBE (opt.sslcertkey);
1353 FREE_MAYBE (opt.sslcertfile);
1354 #endif /* HAVE_SSL */
1355 FREE_MAYBE (opt.bind_address);
1356 FREE_MAYBE (opt.cookies_input);
1357 FREE_MAYBE (opt.cookies_output);