1 /* Reading/parsing the initialization file.
2 Copyright (C) 2005 Free Software Foundation, Inc.
4 This file is part of GNU Wget.
6 GNU Wget is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
11 GNU Wget is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with Wget; if not, write to the Free Software
18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
20 In addition, as a special exception, the Free Software Foundation
21 gives permission to link the code of its release of Wget with the
22 OpenSSL project's "OpenSSL" library (or with modified versions of it
23 that use the same license as the "OpenSSL" library), and distribute
24 the linked executables. You must obey the GNU General Public License
25 in all respects for all of the code used other than "OpenSSL". If you
26 modify this file, you may extend this exception to your version of the
27 file, but you are not obligated to do so. If you do not wish to do
28 so, delete this exception statement from your version. */
33 #include <sys/types.h>
56 #include "recur.h" /* for INFINITE_RECURSION */
57 #include "convert.h" /* for convert_cleanup */
58 #include "res.h" /* for res_cleanup */
64 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
65 otherwise, it will be performed by the shell. This variable will
66 be set by the wgetrc-reading function. */
68 static int enable_tilde_expansion;
71 #define CMD_DECLARE(func) static int func \
72 PARAMS ((const char *, const char *, void *))
74 CMD_DECLARE (cmd_boolean);
75 CMD_DECLARE (cmd_bytes);
76 CMD_DECLARE (cmd_bytes_large);
78 CMD_DECLARE (cmd_cert_type);
80 CMD_DECLARE (cmd_directory_vector);
81 CMD_DECLARE (cmd_lockable_boolean);
82 CMD_DECLARE (cmd_number);
83 CMD_DECLARE (cmd_number_inf);
84 CMD_DECLARE (cmd_string);
85 CMD_DECLARE (cmd_file);
86 CMD_DECLARE (cmd_directory);
87 CMD_DECLARE (cmd_time);
88 CMD_DECLARE (cmd_vector);
90 CMD_DECLARE (cmd_spec_dirstruct);
91 CMD_DECLARE (cmd_spec_header);
92 CMD_DECLARE (cmd_spec_htmlify);
93 CMD_DECLARE (cmd_spec_mirror);
94 CMD_DECLARE (cmd_spec_prefer_family);
95 CMD_DECLARE (cmd_spec_progress);
96 CMD_DECLARE (cmd_spec_recursive);
97 CMD_DECLARE (cmd_spec_restrict_file_names);
99 CMD_DECLARE (cmd_spec_secure_protocol);
101 CMD_DECLARE (cmd_spec_timeout);
102 CMD_DECLARE (cmd_spec_useragent);
104 /* List of recognized commands, each consisting of name, place and
105 function. When adding a new command, simply add it to the list,
106 but be sure to keep the list sorted alphabetically, as
107 command_by_name depends on it. Also, be sure to add any entries
108 that allocate memory (e.g. cmd_string and cmd_vector guys) to the
109 cleanup() function below. */
114 int (*action) PARAMS ((const char *, const char *, void *));
116 { "accept", &opt.accepts, cmd_vector },
117 { "addhostdir", &opt.add_hostdir, cmd_boolean },
118 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "continue", &opt.always_rest, cmd_boolean },
136 { "convertlinks", &opt.convert_links, cmd_boolean },
137 { "cookies", &opt.cookies, cmd_boolean },
138 { "cutdirs", &opt.cut_dirs, cmd_number },
140 { "debug", &opt.debug, cmd_boolean },
142 { "deleteafter", &opt.delete_after, cmd_boolean },
143 { "dirprefix", &opt.dir_prefix, cmd_directory },
144 { "dirstruct", NULL, cmd_spec_dirstruct },
145 { "dnscache", &opt.dns_cache, cmd_boolean },
146 { "dnstimeout", &opt.dns_timeout, cmd_time },
147 { "domains", &opt.domains, cmd_vector },
148 { "dotbytes", &opt.dot_bytes, cmd_bytes },
149 { "dotsinline", &opt.dots_in_line, cmd_number },
150 { "dotspacing", &opt.dot_spacing, cmd_number },
151 { "dotstyle", &opt.dot_style, cmd_string },
153 { "egdfile", &opt.egd_file, cmd_file },
155 { "excludedirectories", &opt.excludes, cmd_directory_vector },
156 { "excludedomains", &opt.exclude_domains, cmd_vector },
157 { "followftp", &opt.follow_ftp, cmd_boolean },
158 { "followtags", &opt.follow_tags, cmd_vector },
159 { "forcehtml", &opt.force_html, cmd_boolean },
160 { "ftppasswd", &opt.ftp_pass, cmd_string },
161 { "ftpproxy", &opt.ftp_proxy, cmd_string },
162 { "glob", &opt.ftp_glob, cmd_boolean },
163 { "header", &opt.user_headers, cmd_spec_header },
164 { "htmlextension", &opt.html_extension, cmd_boolean },
165 { "htmlify", NULL, cmd_spec_htmlify },
166 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
167 { "httppasswd", &opt.http_passwd, cmd_string },
168 { "httpproxy", &opt.http_proxy, cmd_string },
169 { "httpsproxy", &opt.https_proxy, cmd_string },
170 { "httpuser", &opt.http_user, cmd_string },
171 { "ignorelength", &opt.ignore_length, cmd_boolean },
172 { "ignoretags", &opt.ignore_tags, cmd_vector },
173 { "includedirectories", &opt.includes, cmd_directory_vector },
175 { "inet4only", &opt.ipv4_only, cmd_boolean },
176 { "inet6only", &opt.ipv6_only, cmd_boolean },
178 { "input", &opt.input_filename, cmd_file },
179 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
180 { "killlonger", &opt.kill_longer, cmd_boolean },
181 { "limitrate", &opt.limit_rate, cmd_bytes },
182 { "loadcookies", &opt.cookies_input, cmd_file },
183 { "logfile", &opt.lfilename, cmd_file },
184 { "login", &opt.ftp_acc, cmd_string },
185 { "mirror", NULL, cmd_spec_mirror },
186 { "netrc", &opt.netrc, cmd_boolean },
187 { "noclobber", &opt.noclobber, cmd_boolean },
188 { "noparent", &opt.no_parent, cmd_boolean },
189 { "noproxy", &opt.no_proxy, cmd_vector },
190 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
191 { "outputdocument", &opt.output_document, cmd_file },
192 { "pagerequisites", &opt.page_requisites, cmd_boolean },
193 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
194 { "postdata", &opt.post_data, cmd_string },
195 { "postfile", &opt.post_file_name, cmd_file },
196 { "preferfamily", NULL, cmd_spec_prefer_family },
197 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
199 { "privatekey", &opt.private_key, cmd_file },
200 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
202 { "progress", &opt.progress_type, cmd_spec_progress },
203 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
204 { "proxypasswd", &opt.proxy_passwd, cmd_string },
205 { "proxyuser", &opt.proxy_user, cmd_string },
206 { "quiet", &opt.quiet, cmd_boolean },
207 { "quota", &opt.quota, cmd_bytes_large },
208 { "randomfile", &opt.random_file, cmd_file },
209 { "randomwait", &opt.random_wait, cmd_boolean },
210 { "readtimeout", &opt.read_timeout, cmd_time },
211 { "reclevel", &opt.reclevel, cmd_number_inf },
212 { "recursive", NULL, cmd_spec_recursive },
213 { "referer", &opt.referer, cmd_string },
214 { "reject", &opt.rejects, cmd_vector },
215 { "relativeonly", &opt.relative_only, cmd_boolean },
216 { "removelisting", &opt.remove_listing, cmd_boolean },
217 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
218 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
219 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
220 { "robots", &opt.use_robots, cmd_boolean },
221 { "savecookies", &opt.cookies_output, cmd_file },
222 { "saveheaders", &opt.save_headers, cmd_boolean },
224 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
226 { "serverresponse", &opt.server_response, cmd_boolean },
227 { "spanhosts", &opt.spanhost, cmd_boolean },
228 { "spider", &opt.spider, cmd_boolean },
229 { "strictcomments", &opt.strict_comments, cmd_boolean },
230 { "timeout", NULL, cmd_spec_timeout },
231 { "timestamping", &opt.timestamping, cmd_boolean },
232 { "tries", &opt.ntry, cmd_number_inf },
233 { "useproxy", &opt.use_proxy, cmd_boolean },
234 { "useragent", NULL, cmd_spec_useragent },
235 { "verbose", &opt.verbose, cmd_boolean },
236 { "wait", &opt.wait, cmd_time },
237 { "waitretry", &opt.waitretry, cmd_time }
240 /* Look up CMDNAME in the commands[] and return its position in the
241 array. If CMDNAME is not found, return -1. */
244 command_by_name (const char *cmdname)
246 /* Use binary search for speed. Wget has ~100 commands, which
247 guarantees a worst case performance of 7 string comparisons. */
248 int lo = 0, hi = countof (commands) - 1;
252 int mid = (lo + hi) >> 1;
253 int cmp = strcasecmp (cmdname, commands[mid].name);
264 /* Reset the variables to default values. */
270 /* Most of the default values are 0. Just reset everything, and
271 fill in the non-zero values. Note that initializing pointers to
272 NULL this way is technically illegal, but porting Wget to a
273 machine where NULL is not all-zero bit pattern will be the least
274 of the implementors' worries. */
282 opt.ftp_acc = xstrdup ("anonymous");
283 opt.ftp_pass = xstrdup ("-wget@");
287 opt.http_keep_alive = 1;
289 tmp = getenv ("no_proxy");
291 opt.no_proxy = sepstring (tmp);
294 opt.read_timeout = 900;
297 opt.remove_listing = 1;
299 opt.dot_bytes = 1024;
300 opt.dot_spacing = 10;
301 opt.dots_in_line = 50;
310 /* The default for file name restriction defaults to the OS type. */
311 #if !defined(WINDOWS) && !defined(__CYGWIN__)
312 opt.restrict_files_os = restrict_unix;
314 opt.restrict_files_os = restrict_windows;
316 opt.restrict_files_ctrl = 1;
319 /* Return the user's home directory (strdup-ed), or NULL if none is
324 char *home = getenv ("HOME");
329 /* If HOME is not defined, try getting it from the password
331 struct passwd *pwd = getpwuid (getuid ());
332 if (!pwd || !pwd->pw_dir)
336 /* Under Windows, if $HOME isn't defined, use the directory where
337 `wget.exe' resides. */
342 return home ? xstrdup (home) : NULL;
345 /* Return the path to the user's .wgetrc. This is either the value of
346 `WGETRC' environment variable, or `$HOME/.wgetrc'.
348 If the `WGETRC' variable exists but the file does not exist, the
349 function will exit(). */
351 wgetrc_file_name (void)
356 /* Try the environment. */
357 env = getenv ("WGETRC");
360 if (!file_exists_p (env))
362 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
366 return xstrdup (env);
369 /* If that failed, try $HOME/.wgetrc. */
372 file = aprintf ("%s/.wgetrc", home);
376 /* Under Windows, if we still haven't found .wgetrc, look for the file
377 `wget.ini' in the directory where `wget.exe' resides; we do this for
378 backward compatibility with previous versions of Wget.
379 SYSTEM_WGETRC should not be defined under WINDOWS. */
380 if (!file || !file_exists_p (file))
386 file = aprintf ("%s/wget.ini", home);
392 if (!file_exists_p (file))
400 static int parse_line PARAMS ((const char *, char **, char **, int *));
401 static int setval_internal PARAMS ((int, const char *, const char *));
403 /* Initialize variables from a wgetrc file. */
406 run_wgetrc (const char *file)
412 fp = fopen (file, "rb");
415 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
416 file, strerror (errno));
419 enable_tilde_expansion = 1;
421 while ((line = read_whole_line (fp)) != NULL)
426 /* Parse the line. */
427 status = parse_line (line, &com, &val, &comind);
429 /* If everything is OK, set the value. */
432 if (!setval_internal (comind, com, val))
433 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
438 else if (status == 0)
439 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
443 enable_tilde_expansion = 0;
447 /* Initialize the defaults and run the system wgetrc and user's own
454 /* Load the hard-coded defaults. */
457 /* If SYSTEM_WGETRC is defined, use it. */
459 if (file_exists_p (SYSTEM_WGETRC))
460 run_wgetrc (SYSTEM_WGETRC);
462 /* Override it with your own, if one exists. */
463 file = wgetrc_file_name ();
466 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
467 something like realpath() before comparing them with `strcmp' */
469 if (!strcmp (file, SYSTEM_WGETRC))
471 fprintf (stderr, _("\
472 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
482 /* Remove dashes and underscores from S, modifying S in the
488 char *t = s; /* t - tortoise */
489 char *h = s; /* h - hare */
491 if (*h == '_' || *h == '-')
498 /* Parse the line pointed by line, with the syntax:
499 <sp>* command <sp>* = <sp>* value <sp>*
500 Uses malloc to allocate space for command and value.
501 If the line is invalid, data is freed and 0 is returned.
508 In case of success, *COM and *VAL point to freshly allocated
509 strings, and *COMIND points to com's index. In case of error or
510 empty line, those values are unaffected. */
513 parse_line (const char *line, char **com, char **val, int *comind)
516 const char *end = line + strlen (line);
517 const char *cmdstart, *cmdend;
518 const char *valstart, *valend;
523 /* Skip leading and trailing whitespace. */
524 while (*line && ISSPACE (*line))
526 while (end > line && ISSPACE (end[-1]))
529 /* Skip empty lines and comments. */
530 if (!*line || *line == '#')
536 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
540 /* Skip '=', as well as any space before or after it. */
541 while (p < end && ISSPACE (*p))
543 if (p == end || *p != '=')
546 while (p < end && ISSPACE (*p))
552 /* The line now known to be syntactically correct. Check whether
553 the command is valid. */
554 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
556 ind = command_by_name (cmdcopy);
560 /* The command is valid. Now fill in the values and report success
563 *com = strdupdelim (cmdstart, cmdend);
564 *val = strdupdelim (valstart, valend);
568 /* Run commands[comind].action. */
571 setval_internal (int comind, const char *com, const char *val)
573 assert (0 <= comind && comind < countof (commands));
574 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
575 return ((*commands[comind].action) (com, val, commands[comind].place));
578 /* Run command COM with value VAL. If running the command produces an
579 error, report the error and exit.
581 This is intended to be called from main() to modify Wget's behavior
582 through command-line switches. Since COM is hard-coded in main(),
583 it is not canonicalized, and this aborts when COM is not found.
585 If COMIND's are exported to init.h, this function will be changed
586 to accept COMIND directly. */
589 setoptval (const char *com, const char *val)
591 assert (val != NULL);
592 if (!setval_internal (command_by_name (com), com, val))
596 /* Parse OPT into command and value and run it. For example,
597 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
598 This is used by the `--execute' flag in main.c. */
601 run_command (const char *opt)
605 int status = parse_line (opt, &com, &val, &comind);
608 if (!setval_internal (comind, com, val))
613 else if (status == 0)
615 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
621 /* Generic helper functions, for use with `commands'. */
623 /* Forward declarations: */
628 static int decode_string PARAMS ((const char *, const struct decode_item *,
630 static int simple_atoi PARAMS ((const char *, const char *, int *));
631 static int simple_atof PARAMS ((const char *, const char *, double *));
633 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
635 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
636 && TOLOWER((p)[1]) == (c1) \
639 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
640 && TOLOWER((p)[1]) == (c1) \
641 && TOLOWER((p)[2]) == (c2) \
645 /* Store the boolean value from VAL to PLACE. COM is ignored,
646 except for error messages. */
648 cmd_boolean (const char *com, const char *val, void *place)
652 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
653 /* "on", "yes" and "1" mean true. */
655 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
656 /* "off", "no" and "0" mean false. */
661 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
662 exec_name, com, val);
666 *(int *)place = bool_value;
670 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to PLACE.
671 COM is ignored, except for error messages. Values 2 and -1
672 indicate that once defined, the value may not be changed by
673 successive wgetrc files or command-line arguments.
675 Values: 2 - Enable a particular option for good ("always")
676 1 - Enable an option ("on")
677 0 - Disable an option ("off")
678 -1 - Disable an option for good ("never") */
680 cmd_lockable_boolean (const char *com, const char *val, void *place)
682 int lockable_boolean_value;
684 int oldval = *(int *)place;
687 * If a config file said "always" or "never", don't allow command line
688 * arguments to override the config file.
690 if (oldval == -1 || oldval == 2)
693 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
694 lockable_boolean_value = 2;
695 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
696 lockable_boolean_value = 1;
697 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
698 lockable_boolean_value = 0;
699 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
700 lockable_boolean_value = -1;
704 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
705 exec_name, com, val);
709 *(int *)place = lockable_boolean_value;
713 /* Set the non-negative integer value from VAL to PLACE. With
714 incorrect specification, the number remains unchanged. */
716 cmd_number (const char *com, const char *val, void *place)
718 if (!simple_atoi (val, val + strlen (val), place)
719 || *(int *) place < 0)
721 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
722 exec_name, com, val);
728 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
730 cmd_number_inf (const char *com, const char *val, void *place)
732 if (!strcasecmp (val, "inf"))
737 return cmd_number (com, val, place);
740 /* Copy (strdup) the string at COM to a new location and place a
741 pointer to *PLACE. */
743 cmd_string (const char *com, const char *val, void *place)
745 char **pstring = (char **)place;
747 xfree_null (*pstring);
748 *pstring = xstrdup (val);
753 # define ISSEP(c) ((c) == '/')
755 # define ISSEP(c) ((c) == '/' || (c) == '\\')
758 /* Like the above, but handles tilde-expansion when reading a user's
759 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
760 gets expanded to the user's home directory. */
762 cmd_file (const char *com, const char *val, void *place)
764 char **pstring = (char **)place;
766 xfree_null (*pstring);
768 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
770 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
773 *pstring = xstrdup (val);
778 char *home = home_dir ();
782 homelen = strlen (home);
783 while (homelen && ISSEP (home[homelen - 1]))
784 home[--homelen] = '\0';
786 /* Skip the leading "~/". */
787 for (++val; ISSEP (*val); val++)
790 *pstring = concat_strings (home, "/", val, (char *) 0);
794 /* Convert "\" to "/". */
797 for (s = *pstring; *s; s++)
805 /* Like cmd_file, but strips trailing '/' characters. */
807 cmd_directory (const char *com, const char *val, void *place)
811 /* Call cmd_file() for tilde expansion and separator
812 canonicalization (backslash -> slash under Windows). These
813 things should perhaps be in a separate function. */
814 if (!cmd_file (com, val, place))
819 while (t > s && *--t == '/')
825 /* Split VAL by space to a vector of values, and append those values
826 to vector pointed to by the PLACE argument. If VAL is empty, the
827 PLACE vector is cleared instead. */
830 cmd_vector (const char *com, const char *val, void *place)
832 char ***pvec = (char ***)place;
835 *pvec = merge_vecs (*pvec, sepstring (val));
845 cmd_directory_vector (const char *com, const char *val, void *place)
847 char ***pvec = (char ***)place;
851 /* Strip the trailing slashes from directories. */
854 seps = sepstring (val);
855 for (t = seps; t && *t; t++)
857 int len = strlen (*t);
858 /* Skip degenerate case of root directory. */
861 if ((*t)[len - 1] == '/')
862 (*t)[len - 1] = '\0';
865 *pvec = merge_vecs (*pvec, seps);
875 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
876 "100k" or "2.5G" to a floating point number. */
879 parse_bytes_helper (const char *val, double *result)
882 const char *end = val + strlen (val);
884 /* Check for "inf". */
885 if (0 == strcmp (val, "inf"))
891 /* Strip trailing whitespace. */
892 while (val < end && ISSPACE (end[-1]))
897 switch (TOLOWER (end[-1]))
900 --end, mult = 1024.0;
903 --end, mult = 1048576.0;
906 --end, mult = 1073741824.0;
909 --end, mult = 1099511627776.0;
912 /* Not a recognized suffix: assume it's a digit. (If not,
913 simple_atof will raise an error.) */
917 /* Skip leading and trailing whitespace. */
918 while (val < end && ISSPACE (*val))
920 while (val < end && ISSPACE (end[-1]))
925 if (!simple_atof (val, end, &number) || number < 0)
928 *result = number * mult;
932 /* Parse VAL as a number and set its value to PLACE (which should
935 By default, the value is assumed to be in bytes. If "K", "M", or
936 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
937 1<<30, respectively. Floating point values are allowed and are
938 cast to integer before use. The idea is to be able to use things
939 like 1.5k instead of "1536".
941 The string "inf" is returned as 0.
943 In case of error, 0 is returned and memory pointed to by PLACE
944 remains unmodified. */
947 cmd_bytes (const char *com, const char *val, void *place)
950 if (!parse_bytes_helper (val, &byte_value))
952 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
953 exec_name, com, val);
956 *(wgint *)place = (wgint)byte_value;
960 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
961 LARGE_INT. It works by converting the string to double, therefore
962 working with values up to 2^53-1 without loss of precision. This
963 value (8192 TB) is large enough to serve for a while. */
966 cmd_bytes_large (const char *com, const char *val, void *place)
969 if (!parse_bytes_helper (val, &byte_value))
971 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
972 exec_name, com, val);
975 *(LARGE_INT *)place = (LARGE_INT)byte_value;
979 /* Store the value of VAL to *OUT. The value is a time period, by
980 default expressed in seconds, but also accepting suffixes "m", "h",
981 "d", and "w" for minutes, hours, days, and weeks respectively. */
984 cmd_time (const char *com, const char *val, void *place)
987 const char *end = val + strlen (val);
989 /* Strip trailing whitespace. */
990 while (val < end && ISSPACE (end[-1]))
996 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
997 exec_name, com, val);
1001 switch (TOLOWER (end[-1]))
1004 --end, mult = 1; /* seconds */
1007 --end, mult = 60; /* minutes */
1010 --end, mult = 3600; /* hours */
1013 --end, mult = 86400.0; /* days */
1016 --end, mult = 604800.0; /* weeks */
1019 /* Not a recognized suffix: assume it belongs to the number.
1020 (If not, simple_atof will raise an error.) */
1024 /* Skip leading and trailing whitespace. */
1025 while (val < end && ISSPACE (*val))
1027 while (val < end && ISSPACE (end[-1]))
1032 if (!simple_atof (val, end, &number))
1035 *(double *)place = number * mult;
1041 cmd_cert_type (const char *com, const char *val, void *place)
1043 static const struct decode_item choices[] = {
1044 { "pem", keyfile_pem },
1045 { "der", keyfile_asn1 },
1046 { "asn1", keyfile_asn1 },
1048 int ok = decode_string (val, choices, countof (choices), place);
1050 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1055 /* Specialized helper functions, used by `commands' to handle some
1056 options specially. */
1058 static int check_user_specified_header PARAMS ((const char *));
1061 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1063 if (!cmd_boolean (com, val, &opt.dirstruct))
1065 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1066 must be affected inversely. */
1068 opt.no_dirstruct = 0;
1070 opt.no_dirstruct = 1;
1075 cmd_spec_header (const char *com, const char *val, void *place)
1077 if (!check_user_specified_header (val))
1079 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1080 exec_name, com, val);
1083 return cmd_vector (com, val, place);
1087 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1089 int flag = cmd_boolean (com, val, &opt.htmlify);
1090 if (flag && !opt.htmlify)
1091 opt.remove_listing = 0;
1095 /* Set the "mirror" mode. It means: recursive download, timestamping,
1096 no limit on max. recursion depth, and don't remove listings. */
1099 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1103 if (!cmd_boolean (com, val, &mirror))
1108 if (!opt.no_dirstruct)
1110 opt.timestamping = 1;
1111 opt.reclevel = INFINITE_RECURSION;
1112 opt.remove_listing = 0;
1117 /* Validate --prefer-family and set the choice. Allowed values are
1118 "IPv4", "IPv6", and "none". */
1121 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1123 static const struct decode_item choices[] = {
1124 { "IPv4", prefer_ipv4 },
1125 { "IPv6", prefer_ipv6 },
1126 { "none", prefer_none },
1128 int ok = decode_string (val, choices, countof (choices),
1129 (int *) &opt.prefer_family);
1131 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1135 /* Set progress.type to VAL, but verify that it's a valid progress
1136 implementation before that. */
1139 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1141 if (!valid_progress_implementation_p (val))
1143 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1144 exec_name, com, val);
1147 xfree_null (opt.progress_type);
1149 /* Don't call set_progress_implementation here. It will be called
1150 in main() when it becomes clear what the log output is. */
1151 opt.progress_type = xstrdup (val);
1155 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1156 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1160 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1162 if (!cmd_boolean (com, val, &opt.recursive))
1166 if (opt.recursive && !opt.no_dirstruct)
1173 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1175 int restrict_os = opt.restrict_files_os;
1176 int restrict_ctrl = opt.restrict_files_ctrl;
1178 const char *end = strchr (val, ',');
1180 end = val + strlen (val);
1182 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1184 if (VAL_IS ("unix"))
1185 restrict_os = restrict_unix;
1186 else if (VAL_IS ("windows"))
1187 restrict_os = restrict_windows;
1188 else if (VAL_IS ("nocontrol"))
1194 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1195 exec_name, com, val);
1203 if (!strcmp (end + 1, "nocontrol"))
1209 opt.restrict_files_os = restrict_os;
1210 opt.restrict_files_ctrl = restrict_ctrl;
1216 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1218 static const struct decode_item choices[] = {
1219 { "auto", secure_protocol_auto },
1220 { "sslv2", secure_protocol_sslv2 },
1221 { "sslv3", secure_protocol_sslv3 },
1222 { "tlsv1", secure_protocol_tlsv1 },
1224 int ok = decode_string (val, choices, countof (choices), place);
1226 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1231 /* Set all three timeout values. */
1234 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1237 if (!cmd_time (com, val, &value))
1239 opt.read_timeout = value;
1240 opt.connect_timeout = value;
1241 opt.dns_timeout = value;
1246 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1248 /* Just check for empty string and newline, so we don't throw total
1249 junk to the server. */
1250 if (!*val || strchr (val, '\n'))
1252 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1253 exec_name, com, val);
1256 opt.useragent = xstrdup (val);
1260 /* Miscellaneous useful routines. */
1262 /* A very simple atoi clone, more useful than atoi because it works on
1263 delimited strings, and has error reportage. Returns 1 on success,
1264 0 on failure. If successful, stores result to *DEST. */
1267 simple_atoi (const char *beg, const char *end, int *dest)
1271 const char *p = beg;
1273 while (p < end && ISSPACE (*p))
1275 if (p < end && (*p == '-' || *p == '+'))
1277 negative = (*p == '-');
1283 /* Read negative numbers in a separate loop because the most
1284 negative integer cannot be represented as a positive number. */
1287 for (; p < end && ISDIGIT (*p); p++)
1289 int next = (10 * result) + (*p - '0');
1291 return 0; /* overflow */
1295 for (; p < end && ISDIGIT (*p); p++)
1297 int next = (10 * result) - (*p - '0');
1299 return 0; /* underflow */
1310 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1311 doesn't handle exponential notation. Returns 1 on success, 0 on
1312 failure. In case of success, stores its result to *DEST. */
1315 simple_atof (const char *beg, const char *end, double *dest)
1324 const char *p = beg;
1326 while (p < end && ISSPACE (*p))
1328 if (p < end && (*p == '-' || *p == '+'))
1330 negative = (*p == '-');
1334 for (; p < end; p++)
1340 result = (10 * result) + (ch - '0');
1342 result += (ch - '0') / (divider *= 10);
1364 /* Verify that the user-specified header in S is valid. It must
1365 contain a colon preceded by non-white-space characters and must not
1366 contain newlines. */
1369 check_user_specified_header (const char *s)
1373 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1374 /* The header MUST contain `:' preceded by at least one
1375 non-whitespace character. */
1376 if (*p != ':' || p == s)
1378 /* The header MUST NOT contain newlines. */
1379 if (strchr (s, '\n'))
1384 /* Decode VAL into a number, according to ITEMS. */
1387 decode_string (const char *val, const struct decode_item *items, int itemcount,
1391 for (i = 0; i < itemcount; i++)
1392 if (0 == strcasecmp (val, items[i].name))
1394 *place = items[i].code;
1401 void cleanup_html_url PARAMS ((void));
1402 void http_cleanup PARAMS ((void));
1405 /* Free the memory allocated by global variables. */
1409 /* Free external resources, close files, etc. */
1412 extern FILE *output_stream;
1414 fclose (output_stream);
1415 /* No need to check for error because Wget flushes its output (and
1416 checks for errors) after any data arrives. */
1419 /* We're exiting anyway so there's no real need to call free()
1420 hundreds of times. Skipping the frees will make Wget exit
1423 However, when detecting leaks, it's crucial to free() everything
1424 because then you can find the real leaks, i.e. the allocated
1425 memory which grows with the size of the program. */
1431 cleanup_html_url ();
1436 extern acc_t *netrc_list;
1437 free_netrc (netrc_list);
1439 xfree_null (opt.lfilename);
1440 xfree_null (opt.dir_prefix);
1441 xfree_null (opt.input_filename);
1442 xfree_null (opt.output_document);
1443 free_vec (opt.accepts);
1444 free_vec (opt.rejects);
1445 free_vec (opt.excludes);
1446 free_vec (opt.includes);
1447 free_vec (opt.domains);
1448 free_vec (opt.follow_tags);
1449 free_vec (opt.ignore_tags);
1450 xfree_null (opt.progress_type);
1451 xfree (opt.ftp_acc);
1452 xfree_null (opt.ftp_pass);
1453 xfree_null (opt.ftp_proxy);
1454 xfree_null (opt.https_proxy);
1455 xfree_null (opt.http_proxy);
1456 free_vec (opt.no_proxy);
1457 xfree_null (opt.useragent);
1458 xfree_null (opt.referer);
1459 xfree_null (opt.http_user);
1460 xfree_null (opt.http_passwd);
1461 free_vec (opt.user_headers);
1463 xfree_null (opt.cert_file);
1464 xfree_null (opt.private_key);
1466 xfree_null (opt.bind_address);
1467 xfree_null (opt.cookies_input);
1468 xfree_null (opt.cookies_output);
1469 #endif /* DEBUG_MALLOC */