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 */
68 #include "recur.h" /* for INFINITE_RECURSION */
74 extern struct cookie_jar *wget_cookie_jar;
76 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
77 otherwise, it will be performed by the shell. This variable will
78 be set by the wgetrc-reading function. */
80 static int enable_tilde_expansion;
83 #define CMD_DECLARE(func) static int func \
84 PARAMS ((const char *, const char *, void *))
86 CMD_DECLARE (cmd_boolean);
87 CMD_DECLARE (cmd_bytes);
88 CMD_DECLARE (cmd_bytes_large);
89 CMD_DECLARE (cmd_directory_vector);
90 CMD_DECLARE (cmd_lockable_boolean);
91 CMD_DECLARE (cmd_number);
92 CMD_DECLARE (cmd_number_inf);
93 CMD_DECLARE (cmd_string);
94 CMD_DECLARE (cmd_file);
95 CMD_DECLARE (cmd_directory);
96 CMD_DECLARE (cmd_time);
97 CMD_DECLARE (cmd_vector);
99 CMD_DECLARE (cmd_spec_dirstruct);
100 CMD_DECLARE (cmd_spec_header);
101 CMD_DECLARE (cmd_spec_htmlify);
102 CMD_DECLARE (cmd_spec_mirror);
103 CMD_DECLARE (cmd_spec_progress);
104 CMD_DECLARE (cmd_spec_recursive);
105 CMD_DECLARE (cmd_spec_restrict_file_names);
106 CMD_DECLARE (cmd_spec_timeout);
107 CMD_DECLARE (cmd_spec_useragent);
109 /* List of recognized commands, each consisting of name, closure and
110 function. When adding a new command, simply add it to the list,
111 but be sure to keep the list sorted alphabetically, as
112 command_by_name depends on it. Also, be sure to add any entries
113 that allocate memory (e.g. cmd_string and cmd_vector guys) to the
114 cleanup() function below. */
119 int (*action) PARAMS ((const char *, const char *, void *));
121 { "accept", &opt.accepts, cmd_vector },
122 { "addhostdir", &opt.add_hostdir, cmd_boolean },
123 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
124 { "background", &opt.background, cmd_boolean },
125 { "backupconverted", &opt.backup_converted, cmd_boolean },
126 { "backups", &opt.backups, cmd_number },
127 { "base", &opt.base_href, cmd_string },
128 { "bindaddress", &opt.bind_address, cmd_string },
129 { "cache", &opt.allow_cache, cmd_boolean },
130 { "connecttimeout", &opt.connect_timeout, cmd_time },
131 { "continue", &opt.always_rest, cmd_boolean },
132 { "convertlinks", &opt.convert_links, cmd_boolean },
133 { "cookies", &opt.cookies, cmd_boolean },
134 { "cutdirs", &opt.cut_dirs, cmd_number },
136 { "debug", &opt.debug, cmd_boolean },
138 { "deleteafter", &opt.delete_after, cmd_boolean },
139 { "dirprefix", &opt.dir_prefix, cmd_directory },
140 { "dirstruct", NULL, cmd_spec_dirstruct },
141 { "dnscache", &opt.dns_cache, cmd_boolean },
142 { "dnstimeout", &opt.dns_timeout, cmd_time },
143 { "domains", &opt.domains, cmd_vector },
144 { "dotbytes", &opt.dot_bytes, cmd_bytes },
145 { "dotsinline", &opt.dots_in_line, cmd_number },
146 { "dotspacing", &opt.dot_spacing, cmd_number },
147 { "dotstyle", &opt.dot_style, cmd_string },
149 { "egdfile", &opt.sslegdsock, cmd_file },
151 { "excludedirectories", &opt.excludes, cmd_directory_vector },
152 { "excludedomains", &opt.exclude_domains, cmd_vector },
153 { "followftp", &opt.follow_ftp, cmd_boolean },
154 { "followtags", &opt.follow_tags, cmd_vector },
155 { "forcehtml", &opt.force_html, cmd_boolean },
156 { "ftpproxy", &opt.ftp_proxy, cmd_string },
157 { "glob", &opt.ftp_glob, cmd_boolean },
158 { "header", NULL, cmd_spec_header },
159 { "htmlextension", &opt.html_extension, cmd_boolean },
160 { "htmlify", NULL, cmd_spec_htmlify },
161 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
162 { "httppasswd", &opt.http_passwd, cmd_string },
163 { "httpproxy", &opt.http_proxy, cmd_string },
164 { "httpsproxy", &opt.https_proxy, cmd_string },
165 { "httpuser", &opt.http_user, cmd_string },
166 { "ignorelength", &opt.ignore_length, cmd_boolean },
167 { "ignoretags", &opt.ignore_tags, cmd_vector },
168 { "includedirectories", &opt.includes, cmd_directory_vector },
169 { "input", &opt.input_filename, cmd_file },
170 { "killlonger", &opt.kill_longer, cmd_boolean },
171 { "limitrate", &opt.limit_rate, cmd_bytes },
172 { "loadcookies", &opt.cookies_input, cmd_file },
173 { "logfile", &opt.lfilename, cmd_file },
174 { "login", &opt.ftp_acc, cmd_string },
175 { "mirror", NULL, cmd_spec_mirror },
176 { "netrc", &opt.netrc, cmd_boolean },
177 { "noclobber", &opt.noclobber, cmd_boolean },
178 { "noparent", &opt.no_parent, cmd_boolean },
179 { "noproxy", &opt.no_proxy, cmd_vector },
180 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
181 { "outputdocument", &opt.output_document, cmd_file },
182 { "pagerequisites", &opt.page_requisites, cmd_boolean },
183 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
184 { "passwd", &opt.ftp_pass, cmd_string },
185 { "postdata", &opt.post_data, cmd_string },
186 { "postfile", &opt.post_file_name, cmd_file },
187 { "progress", &opt.progress_type, cmd_spec_progress },
188 { "proxypasswd", &opt.proxy_passwd, cmd_string },
189 { "proxyuser", &opt.proxy_user, cmd_string },
190 { "quiet", &opt.quiet, cmd_boolean },
191 { "quota", &opt.quota, cmd_bytes_large },
192 { "randomwait", &opt.random_wait, cmd_boolean },
193 { "readtimeout", &opt.read_timeout, cmd_time },
194 { "reclevel", &opt.reclevel, cmd_number_inf },
195 { "recursive", NULL, cmd_spec_recursive },
196 { "referer", &opt.referer, cmd_string },
197 { "reject", &opt.rejects, cmd_vector },
198 { "relativeonly", &opt.relative_only, cmd_boolean },
199 { "removelisting", &opt.remove_listing, cmd_boolean },
200 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
201 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
202 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
203 { "robots", &opt.use_robots, cmd_boolean },
204 { "savecookies", &opt.cookies_output, cmd_file },
205 { "saveheaders", &opt.save_headers, cmd_boolean },
206 { "serverresponse", &opt.server_response, cmd_boolean },
207 { "spanhosts", &opt.spanhost, cmd_boolean },
208 { "spider", &opt.spider, cmd_boolean },
210 { "sslcadir", &opt.sslcadir, cmd_directory },
211 { "sslcafile", &opt.sslcafile, cmd_file },
212 { "sslcertfile", &opt.sslcertfile, cmd_file },
213 { "sslcertkey", &opt.sslcertkey, cmd_file },
214 { "sslcerttype", &opt.sslcerttype, cmd_number },
215 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
216 { "sslprotocol", &opt.sslprotocol, cmd_number },
217 #endif /* HAVE_SSL */
218 { "strictcomments", &opt.strict_comments, cmd_boolean },
219 { "timeout", NULL, cmd_spec_timeout },
220 { "timestamping", &opt.timestamping, cmd_boolean },
221 { "tries", &opt.ntry, cmd_number_inf },
222 { "useproxy", &opt.use_proxy, cmd_boolean },
223 { "useragent", NULL, cmd_spec_useragent },
224 { "verbose", &opt.verbose, cmd_boolean },
225 { "wait", &opt.wait, cmd_time },
226 { "waitretry", &opt.waitretry, cmd_time }
229 /* Look up CMDNAME in the commands[] and return its position in the
230 array. If CMDNAME is not found, return -1. */
233 command_by_name (const char *cmdname)
235 /* Use binary search for speed. Wget has ~100 commands, which
236 guarantees a worst case performance of 7 string comparisons. */
237 int lo = 0, hi = countof (commands) - 1;
241 int mid = (lo + hi) >> 1;
242 int cmp = strcasecmp (cmdname, commands[mid].name);
253 /* Reset the variables to default values. */
259 /* Most of the default values are 0. Just reset everything, and
260 fill in the non-zero values. Note that initializing pointers to
261 NULL this way is technically illegal, but porting Wget to a
262 machine where NULL is not all-zero bit pattern will be the least
263 of the implementors' worries. */
271 opt.ftp_acc = xstrdup ("anonymous");
272 opt.ftp_pass = xstrdup ("-wget@");
276 opt.http_keep_alive = 1;
278 tmp = getenv ("no_proxy");
280 opt.no_proxy = sepstring (tmp);
283 opt.read_timeout = 900;
286 opt.remove_listing = 1;
288 opt.dot_bytes = 1024;
289 opt.dot_spacing = 10;
290 opt.dots_in_line = 50;
294 /* The default for file name restriction defaults to the OS type. */
295 #if !defined(WINDOWS) && !defined(__CYGWIN__)
296 opt.restrict_files_os = restrict_unix;
298 opt.restrict_files_os = restrict_windows;
300 opt.restrict_files_ctrl = 1;
303 /* Return the user's home directory (strdup-ed), or NULL if none is
308 char *home = getenv ("HOME");
313 /* If HOME is not defined, try getting it from the password
315 struct passwd *pwd = getpwuid (getuid ());
316 if (!pwd || !pwd->pw_dir)
321 /* #### Maybe I should grab home_dir from registry, but the best
322 that I could get from there is user's Start menu. It sucks! */
326 return home ? xstrdup (home) : NULL;
329 /* Return the path to the user's .wgetrc. This is either the value of
330 `WGETRC' environment variable, or `$HOME/.wgetrc'.
332 If the `WGETRC' variable exists but the file does not exist, the
333 function will exit(). */
335 wgetrc_file_name (void)
340 /* Try the environment. */
341 env = getenv ("WGETRC");
344 if (!file_exists_p (env))
346 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
350 return xstrdup (env);
354 /* If that failed, try $HOME/.wgetrc. */
358 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
359 sprintf (file, "%s/.wgetrc", home);
363 /* Under Windows, "home" is (for the purposes of this function) the
364 directory where `wget.exe' resides, and `wget.ini' will be used
365 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
367 It is not as trivial as I assumed, because on 95 argv[0] is full
368 path, but on NT you get what you typed in command line. --dbudor */
372 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
373 sprintf (file, "%swget.ini", home);
379 if (!file_exists_p (file))
387 static int parse_line PARAMS ((const char *, char **, char **, int *));
388 static int setval_internal PARAMS ((int, const char *, const char *));
390 /* Initialize variables from a wgetrc file. */
393 run_wgetrc (const char *file)
399 fp = fopen (file, "rb");
402 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
403 file, strerror (errno));
406 enable_tilde_expansion = 1;
408 while ((line = read_whole_line (fp)))
413 /* Parse the line. */
414 status = parse_line (line, &com, &val, &comind);
416 /* If everything is OK, set the value. */
419 if (!setval_internal (comind, com, val))
420 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
425 else if (status == 0)
426 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
430 enable_tilde_expansion = 0;
434 /* Initialize the defaults and run the system wgetrc and user's own
441 /* Load the hard-coded defaults. */
444 /* If SYSTEM_WGETRC is defined, use it. */
446 if (file_exists_p (SYSTEM_WGETRC))
447 run_wgetrc (SYSTEM_WGETRC);
449 /* Override it with your own, if one exists. */
450 file = wgetrc_file_name ();
453 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
454 something like realpath() before comparing them with `strcmp' */
456 if (!strcmp (file, SYSTEM_WGETRC))
458 fprintf (stderr, _("\
459 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
469 /* Remove dashes and underscores from S, modifying S in the
475 char *t = s; /* t - tortoise */
476 char *h = s; /* h - hare */
478 if (*h == '_' || *h == '-')
485 /* Parse the line pointed by line, with the syntax:
486 <sp>* command <sp>* = <sp>* value <sp>*
487 Uses malloc to allocate space for command and value.
488 If the line is invalid, data is freed and 0 is returned.
495 In case of success, *COM and *VAL point to freshly allocated
496 strings, and *COMIND points to com's index. In case of error or
497 empty line, those values are unaffected. */
500 parse_line (const char *line, char **com, char **val, int *comind)
503 const char *end = line + strlen (line);
504 const char *cmdstart, *cmdend;
505 const char *valstart, *valend;
510 /* Skip leading and trailing whitespace. */
511 while (*line && ISSPACE (*line))
513 while (end > line && ISSPACE (end[-1]))
516 /* Skip empty lines and comments. */
517 if (!*line || *line == '#')
523 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
527 /* Skip '=', as well as any space before or after it. */
528 while (p < end && ISSPACE (*p))
530 if (p == end || *p != '=')
533 while (p < end && ISSPACE (*p))
539 /* The line now known to be syntactically correct. Check whether
540 the command is valid. */
541 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
543 ind = command_by_name (cmdcopy);
547 /* The command is valid. Now fill in the values and report success
550 *com = strdupdelim (cmdstart, cmdend);
551 *val = strdupdelim (valstart, valend);
555 /* Run commands[comind].action. */
558 setval_internal (int comind, const char *com, const char *val)
560 assert (0 <= comind && comind < countof (commands));
561 return ((*commands[comind].action) (com, val, commands[comind].closure));
564 /* Run command COM with value VAL. If running the command produces an
565 error, report the error and exit.
567 This is intended to be called from main() to modify Wget's behavior
568 through command-line switches. Since COM is hard-coded in main(),
569 it is not canonicalized, and this aborts when COM is not found.
571 If COMIND's are exported to init.h, this function will be changed
572 to accept COMIND directly. */
575 setoptval (const char *com, const char *val)
577 if (!setval_internal (command_by_name (com), com, val))
581 /* Parse OPT into command and value and run it. For example,
582 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
583 This is used by the `--execute' flag in main.c. */
586 run_command (const char *opt)
590 int status = parse_line (opt, &com, &val, &comind);
593 if (!setval_internal (comind, com, val))
598 else if (status == 0)
600 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
606 /* Generic helper functions, for use with `commands'. */
608 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
610 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
611 && TOLOWER((p)[1]) == (c1) \
614 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
615 && TOLOWER((p)[1]) == (c1) \
616 && TOLOWER((p)[2]) == (c2) \
620 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
621 except for error messages. */
623 cmd_boolean (const char *com, const char *val, void *closure)
627 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
628 /* "on", "yes" and "1" mean true. */
630 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
631 /* "off", "no" and "0" mean false. */
636 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
637 exec_name, com, val);
641 *(int *)closure = bool_value;
645 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
646 COM is ignored, except for error messages. Values 2 and -1
647 indicate that once defined, the value may not be changed by
648 successive wgetrc files or command-line arguments.
650 Values: 2 - Enable a particular option for good ("always")
651 1 - Enable an option ("on")
652 0 - Disable an option ("off")
653 -1 - Disable an option for good ("never") */
655 cmd_lockable_boolean (const char *com, const char *val, void *closure)
657 int lockable_boolean_value;
659 int oldval = *(int *)closure;
662 * If a config file said "always" or "never", don't allow command line
663 * arguments to override the config file.
665 if (oldval == -1 || oldval == 2)
668 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
669 lockable_boolean_value = 2;
670 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
671 lockable_boolean_value = 1;
672 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
673 lockable_boolean_value = 0;
674 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
675 lockable_boolean_value = -1;
679 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
680 exec_name, com, val);
684 *(int *)closure = lockable_boolean_value;
688 static int simple_atoi PARAMS ((const char *, const char *, int *));
690 /* Set the non-negative integer value from VAL to CLOSURE. With
691 incorrect specification, the number remains unchanged. */
693 cmd_number (const char *com, const char *val, void *closure)
695 if (!simple_atoi (val, val + strlen (val), closure))
697 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
698 exec_name, com, val);
704 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
706 cmd_number_inf (const char *com, const char *val, void *closure)
708 if (!strcasecmp (val, "inf"))
713 return cmd_number (com, val, closure);
716 /* Copy (strdup) the string at COM to a new location and place a
717 pointer to *CLOSURE. */
719 cmd_string (const char *com, const char *val, void *closure)
721 char **pstring = (char **)closure;
723 xfree_null (*pstring);
724 *pstring = xstrdup (val);
729 # define ISSEP(c) ((c) == '/')
731 # define ISSEP(c) ((c) == '/' || (c) == '\\')
734 /* Like the above, but handles tilde-expansion when reading a user's
735 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
736 gets expanded to the user's home directory. */
738 cmd_file (const char *com, const char *val, void *closure)
740 char **pstring = (char **)closure;
742 xfree_null (*pstring);
744 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
746 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
749 *pstring = xstrdup (val);
755 char *home = home_dir ();
759 homelen = strlen (home);
760 while (homelen && ISSEP (home[homelen - 1]))
761 home[--homelen] = '\0';
763 /* Skip the leading "~/". */
764 for (++val; ISSEP (*val); val++)
767 result = xmalloc (homelen + 1 + strlen (val) + 1);
768 memcpy (result, home, homelen);
769 result[homelen] = '/';
770 strcpy (result + homelen + 1, val);
776 /* Convert "\" to "/". */
779 for (s = *pstring; *s; s++)
787 /* Like cmd_file, but strips trailing '/' characters. */
789 cmd_directory (const char *com, const char *val, void *closure)
793 /* Call cmd_file() for tilde expansion and separator
794 canonicalization (backslash -> slash under Windows). These
795 things should perhaps be in a separate function. */
796 if (!cmd_file (com, val, closure))
799 s = *(char **)closure;
801 while (t > s && *--t == '/')
807 /* Split VAL by space to a vector of values, and append those values
808 to vector pointed to by the CLOSURE argument. If VAL is empty, the
809 CLOSURE vector is cleared instead. */
812 cmd_vector (const char *com, const char *val, void *closure)
814 char ***pvec = (char ***)closure;
817 *pvec = merge_vecs (*pvec, sepstring (val));
827 cmd_directory_vector (const char *com, const char *val, void *closure)
829 char ***pvec = (char ***)closure;
833 /* Strip the trailing slashes from directories. */
836 seps = sepstring (val);
837 for (t = seps; t && *t; t++)
839 int len = strlen (*t);
840 /* Skip degenerate case of root directory. */
843 if ((*t)[len - 1] == '/')
844 (*t)[len - 1] = '\0';
847 *pvec = merge_vecs (*pvec, seps);
857 static int simple_atof PARAMS ((const char *, const char *, double *));
859 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
860 as "100k" or "2.5G" to a floating point number. */
863 parse_bytes_helper (const char *val, double *result)
866 const char *end = val + strlen (val);
868 /* Check for "inf". */
869 if (0 == strcmp (val, "inf"))
875 /* Strip trailing whitespace. */
876 while (val < end && ISSPACE (end[-1]))
881 switch (TOLOWER (end[-1]))
884 --end, mult = 1024.0;
887 --end, mult = 1048576.0;
890 --end, mult = 1073741824.0;
893 --end, mult = 1099511627776.0;
896 /* Not a recognized suffix: assume it's a digit. (If not,
897 simple_atof will raise an error.) */
901 /* Skip leading and trailing whitespace. */
902 while (val < end && ISSPACE (*val))
904 while (val < end && ISSPACE (end[-1]))
909 if (!simple_atof (val, end, &number))
912 *result = number * mult;
916 /* Parse VAL as a number and set its value to CLOSURE (which should
917 point to a long int).
919 By default, the value is assumed to be in bytes. If "K", "M", or
920 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
921 1<<30, respectively. Floating point values are allowed and are
922 cast to integer before use. The idea is to be able to use things
923 like 1.5k instead of "1536".
925 The string "inf" is returned as 0.
927 In case of error, 0 is returned and memory pointed to by CLOSURE
928 remains unmodified. */
931 cmd_bytes (const char *com, const char *val, void *closure)
934 if (!parse_bytes_helper (val, &byte_value))
936 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
937 exec_name, com, val);
940 *(long *)closure = (long)byte_value;
944 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
945 LARGE_INT. It works by converting the string to double, therefore
946 working with values up to 2^53-1 without loss of precision. This
947 value (8192 TB) is large enough to serve for a while. */
950 cmd_bytes_large (const char *com, const char *val, void *closure)
953 if (!parse_bytes_helper (val, &byte_value))
955 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
956 exec_name, com, val);
959 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
963 /* Store the value of VAL to *OUT. The value is a time period, by
964 default expressed in seconds, but also accepting suffixes "m", "h",
965 "d", and "w" for minutes, hours, days, and weeks respectively. */
968 cmd_time (const char *com, const char *val, void *closure)
971 const char *end = val + strlen (val);
973 /* Strip trailing whitespace. */
974 while (val < end && ISSPACE (end[-1]))
980 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
981 exec_name, com, val);
985 switch (TOLOWER (end[-1]))
988 --end, mult = 1; /* seconds */
991 --end, mult = 60; /* minutes */
994 --end, mult = 3600; /* hours */
997 --end, mult = 86400.0; /* days */
1000 --end, mult = 604800.0; /* weeks */
1003 /* Not a recognized suffix: assume it belongs to the number.
1004 (If not, atof simple_atof will raise an error.) */
1008 /* Skip leading and trailing whitespace. */
1009 while (val < end && ISSPACE (*val))
1011 while (val < end && ISSPACE (end[-1]))
1016 if (!simple_atof (val, end, &number))
1019 *(double *)closure = number * mult;
1023 /* Specialized helper functions, used by `commands' to handle some
1024 options specially. */
1026 static int check_user_specified_header PARAMS ((const char *));
1029 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1031 if (!cmd_boolean (com, val, &opt.dirstruct))
1033 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1034 must be affected inversely. */
1036 opt.no_dirstruct = 0;
1038 opt.no_dirstruct = 1;
1043 cmd_spec_header (const char *com, const char *val, void *closure)
1047 /* Empty header means reset headers. */
1048 xfree_null (opt.user_header);
1049 opt.user_header = NULL;
1055 if (!check_user_specified_header (val))
1057 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1058 exec_name, com, val);
1061 i = opt.user_header ? strlen (opt.user_header) : 0;
1062 opt.user_header = (char *)xrealloc (opt.user_header, i + strlen (val)
1064 strcpy (opt.user_header + i, val);
1066 opt.user_header[i++] = '\r';
1067 opt.user_header[i++] = '\n';
1068 opt.user_header[i] = '\0';
1074 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1076 int flag = cmd_boolean (com, val, &opt.htmlify);
1077 if (flag && !opt.htmlify)
1078 opt.remove_listing = 0;
1082 /* Set the "mirror" mode. It means: recursive download, timestamping,
1083 no limit on max. recursion depth, and don't remove listings. */
1086 cmd_spec_mirror (const char *com, const char *val, void *closure)
1090 if (!cmd_boolean (com, val, &mirror))
1095 if (!opt.no_dirstruct)
1097 opt.timestamping = 1;
1098 opt.reclevel = INFINITE_RECURSION;
1099 opt.remove_listing = 0;
1104 /* Set progress.type to VAL, but verify that it's a valid progress
1105 implementation before that. */
1108 cmd_spec_progress (const char *com, const char *val, void *closure)
1110 if (!valid_progress_implementation_p (val))
1112 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1113 exec_name, com, val);
1116 xfree_null (opt.progress_type);
1118 /* Don't call set_progress_implementation here. It will be called
1119 in main() when it becomes clear what the log output is. */
1120 opt.progress_type = xstrdup (val);
1124 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1125 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1129 cmd_spec_recursive (const char *com, const char *val, void *closure)
1131 if (!cmd_boolean (com, val, &opt.recursive))
1135 if (opt.recursive && !opt.no_dirstruct)
1142 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1144 int restrict_os = opt.restrict_files_os;
1145 int restrict_ctrl = opt.restrict_files_ctrl;
1147 const char *end = strchr (val, ',');
1149 end = val + strlen (val);
1151 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1153 if (VAL_IS ("unix"))
1154 restrict_os = restrict_unix;
1155 else if (VAL_IS ("windows"))
1156 restrict_os = restrict_windows;
1157 else if (VAL_IS ("nocontrol"))
1163 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1164 exec_name, com, val);
1172 if (!strcmp (end + 1, "nocontrol"))
1178 opt.restrict_files_os = restrict_os;
1179 opt.restrict_files_ctrl = restrict_ctrl;
1183 /* Set all three timeout values. */
1186 cmd_spec_timeout (const char *com, const char *val, void *closure)
1189 if (!cmd_time (com, val, &value))
1191 opt.read_timeout = value;
1192 opt.connect_timeout = value;
1193 opt.dns_timeout = value;
1198 cmd_spec_useragent (const char *com, const char *val, void *closure)
1200 /* Just check for empty string and newline, so we don't throw total
1201 junk to the server. */
1202 if (!*val || strchr (val, '\n'))
1204 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1205 exec_name, com, val);
1208 opt.useragent = xstrdup (val);
1212 /* Miscellaneous useful routines. */
1214 /* A very simple atoi clone, more portable than strtol and friends,
1215 but reports errors, unlike atoi. Returns 1 on success, 0 on
1216 failure. In case of success, stores result to *DEST. */
1219 simple_atoi (const char *beg, const char *end, int *dest)
1227 for (p = beg; p < end && ISDIGIT (*p); p++)
1228 result = (10 * result) + (*p - '0');
1237 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1238 doesn't handle exponential notation. Returns 1 on success, 0 on
1239 failure. In case of success, stores its result to *DEST. */
1242 simple_atof (const char *beg, const char *end, double *dest)
1252 for (p = beg; p < end; p++)
1258 result = (10 * result) + (ch - '0');
1260 result += (ch - '0') / (divider *= 10);
1281 check_user_specified_header (const char *s)
1285 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1286 /* The header MUST contain `:' preceded by at least one
1287 non-whitespace character. */
1288 if (*p != ':' || p == s)
1290 /* The header MUST NOT contain newlines. */
1291 if (strchr (s, '\n'))
1296 void cleanup_html_url PARAMS ((void));
1297 void res_cleanup PARAMS ((void));
1298 void downloaded_files_free PARAMS ((void));
1299 void http_cleanup PARAMS ((void));
1302 /* Free the memory allocated by global variables. */
1306 /* Free external resources, close files, etc. */
1311 /* We're exiting anyway so there's no real need to call free()
1312 hundreds of times. Skipping the frees will make Wget exit
1315 However, when detecting leaks, it's crucial to free() everything
1316 because then you can find the real leaks, i.e. the allocated
1317 memory which grows with the size of the program. */
1323 cleanup_html_url ();
1324 downloaded_files_free ();
1326 if (wget_cookie_jar)
1327 cookie_jar_delete (wget_cookie_jar);
1330 extern acc_t *netrc_list;
1331 free_netrc (netrc_list);
1333 xfree_null (opt.lfilename);
1334 xfree_null (opt.dir_prefix);
1335 xfree_null (opt.input_filename);
1336 xfree_null (opt.output_document);
1337 free_vec (opt.accepts);
1338 free_vec (opt.rejects);
1339 free_vec (opt.excludes);
1340 free_vec (opt.includes);
1341 free_vec (opt.domains);
1342 free_vec (opt.follow_tags);
1343 free_vec (opt.ignore_tags);
1344 xfree_null (opt.progress_type);
1345 xfree (opt.ftp_acc);
1346 xfree_null (opt.ftp_pass);
1347 xfree_null (opt.ftp_proxy);
1348 xfree_null (opt.https_proxy);
1349 xfree_null (opt.http_proxy);
1350 free_vec (opt.no_proxy);
1351 xfree_null (opt.useragent);
1352 xfree_null (opt.referer);
1353 xfree_null (opt.http_user);
1354 xfree_null (opt.http_passwd);
1355 xfree_null (opt.user_header);
1357 xfree_null (opt.sslcertkey);
1358 xfree_null (opt.sslcertfile);
1359 #endif /* HAVE_SSL */
1360 xfree_null (opt.bind_address);
1361 xfree_null (opt.cookies_input);
1362 xfree_null (opt.cookies_output);