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 { "randomwait", &opt.random_wait, cmd_boolean },
209 { "readtimeout", &opt.read_timeout, cmd_time },
210 { "reclevel", &opt.reclevel, cmd_number_inf },
211 { "recursive", NULL, cmd_spec_recursive },
212 { "referer", &opt.referer, cmd_string },
213 { "reject", &opt.rejects, cmd_vector },
214 { "relativeonly", &opt.relative_only, cmd_boolean },
215 { "removelisting", &opt.remove_listing, cmd_boolean },
216 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
217 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
218 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
219 { "robots", &opt.use_robots, cmd_boolean },
220 { "savecookies", &opt.cookies_output, cmd_file },
221 { "saveheaders", &opt.save_headers, cmd_boolean },
223 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
225 { "serverresponse", &opt.server_response, cmd_boolean },
226 { "spanhosts", &opt.spanhost, cmd_boolean },
227 { "spider", &opt.spider, cmd_boolean },
228 { "strictcomments", &opt.strict_comments, cmd_boolean },
229 { "timeout", NULL, cmd_spec_timeout },
230 { "timestamping", &opt.timestamping, cmd_boolean },
231 { "tries", &opt.ntry, cmd_number_inf },
232 { "useproxy", &opt.use_proxy, cmd_boolean },
233 { "useragent", NULL, cmd_spec_useragent },
234 { "verbose", &opt.verbose, cmd_boolean },
235 { "wait", &opt.wait, cmd_time },
236 { "waitretry", &opt.waitretry, cmd_time }
239 /* Look up CMDNAME in the commands[] and return its position in the
240 array. If CMDNAME is not found, return -1. */
243 command_by_name (const char *cmdname)
245 /* Use binary search for speed. Wget has ~100 commands, which
246 guarantees a worst case performance of 7 string comparisons. */
247 int lo = 0, hi = countof (commands) - 1;
251 int mid = (lo + hi) >> 1;
252 int cmp = strcasecmp (cmdname, commands[mid].name);
263 /* Reset the variables to default values. */
269 /* Most of the default values are 0. Just reset everything, and
270 fill in the non-zero values. Note that initializing pointers to
271 NULL this way is technically illegal, but porting Wget to a
272 machine where NULL is not all-zero bit pattern will be the least
273 of the implementors' worries. */
281 opt.ftp_acc = xstrdup ("anonymous");
282 opt.ftp_pass = xstrdup ("-wget@");
286 opt.http_keep_alive = 1;
288 tmp = getenv ("no_proxy");
290 opt.no_proxy = sepstring (tmp);
293 opt.read_timeout = 900;
296 opt.remove_listing = 1;
298 opt.dot_bytes = 1024;
299 opt.dot_spacing = 10;
300 opt.dots_in_line = 50;
309 /* The default for file name restriction defaults to the OS type. */
310 #if !defined(WINDOWS) && !defined(__CYGWIN__)
311 opt.restrict_files_os = restrict_unix;
313 opt.restrict_files_os = restrict_windows;
315 opt.restrict_files_ctrl = 1;
318 /* Return the user's home directory (strdup-ed), or NULL if none is
323 char *home = getenv ("HOME");
328 /* If HOME is not defined, try getting it from the password
330 struct passwd *pwd = getpwuid (getuid ());
331 if (!pwd || !pwd->pw_dir)
335 /* Under Windows, if $HOME isn't defined, use the directory where
336 `wget.exe' resides. */
341 return home ? xstrdup (home) : NULL;
344 /* Return the path to the user's .wgetrc. This is either the value of
345 `WGETRC' environment variable, or `$HOME/.wgetrc'.
347 If the `WGETRC' variable exists but the file does not exist, the
348 function will exit(). */
350 wgetrc_file_name (void)
355 /* Try the environment. */
356 env = getenv ("WGETRC");
359 if (!file_exists_p (env))
361 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
365 return xstrdup (env);
368 /* If that failed, try $HOME/.wgetrc. */
371 file = aprintf ("%s/.wgetrc", home);
375 /* Under Windows, if we still haven't found .wgetrc, look for the file
376 `wget.ini' in the directory where `wget.exe' resides; we do this for
377 backward compatibility with previous versions of Wget.
378 SYSTEM_WGETRC should not be defined under WINDOWS. */
379 if (!file || !file_exists_p (file))
385 file = aprintf ("%s/wget.ini", home);
391 if (!file_exists_p (file))
399 static int parse_line PARAMS ((const char *, char **, char **, int *));
400 static int setval_internal PARAMS ((int, const char *, const char *));
402 /* Initialize variables from a wgetrc file. */
405 run_wgetrc (const char *file)
411 fp = fopen (file, "rb");
414 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
415 file, strerror (errno));
418 enable_tilde_expansion = 1;
420 while ((line = read_whole_line (fp)) != NULL)
425 /* Parse the line. */
426 status = parse_line (line, &com, &val, &comind);
428 /* If everything is OK, set the value. */
431 if (!setval_internal (comind, com, val))
432 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
437 else if (status == 0)
438 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
442 enable_tilde_expansion = 0;
446 /* Initialize the defaults and run the system wgetrc and user's own
453 /* Load the hard-coded defaults. */
456 /* If SYSTEM_WGETRC is defined, use it. */
458 if (file_exists_p (SYSTEM_WGETRC))
459 run_wgetrc (SYSTEM_WGETRC);
461 /* Override it with your own, if one exists. */
462 file = wgetrc_file_name ();
465 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
466 something like realpath() before comparing them with `strcmp' */
468 if (!strcmp (file, SYSTEM_WGETRC))
470 fprintf (stderr, _("\
471 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
481 /* Remove dashes and underscores from S, modifying S in the
487 char *t = s; /* t - tortoise */
488 char *h = s; /* h - hare */
490 if (*h == '_' || *h == '-')
497 /* Parse the line pointed by line, with the syntax:
498 <sp>* command <sp>* = <sp>* value <sp>*
499 Uses malloc to allocate space for command and value.
500 If the line is invalid, data is freed and 0 is returned.
507 In case of success, *COM and *VAL point to freshly allocated
508 strings, and *COMIND points to com's index. In case of error or
509 empty line, those values are unaffected. */
512 parse_line (const char *line, char **com, char **val, int *comind)
515 const char *end = line + strlen (line);
516 const char *cmdstart, *cmdend;
517 const char *valstart, *valend;
522 /* Skip leading and trailing whitespace. */
523 while (*line && ISSPACE (*line))
525 while (end > line && ISSPACE (end[-1]))
528 /* Skip empty lines and comments. */
529 if (!*line || *line == '#')
535 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
539 /* Skip '=', as well as any space before or after it. */
540 while (p < end && ISSPACE (*p))
542 if (p == end || *p != '=')
545 while (p < end && ISSPACE (*p))
551 /* The line now known to be syntactically correct. Check whether
552 the command is valid. */
553 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
555 ind = command_by_name (cmdcopy);
559 /* The command is valid. Now fill in the values and report success
562 *com = strdupdelim (cmdstart, cmdend);
563 *val = strdupdelim (valstart, valend);
567 /* Run commands[comind].action. */
570 setval_internal (int comind, const char *com, const char *val)
572 assert (0 <= comind && comind < countof (commands));
573 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
574 return ((*commands[comind].action) (com, val, commands[comind].place));
577 /* Run command COM with value VAL. If running the command produces an
578 error, report the error and exit.
580 This is intended to be called from main() to modify Wget's behavior
581 through command-line switches. Since COM is hard-coded in main(),
582 it is not canonicalized, and this aborts when COM is not found.
584 If COMIND's are exported to init.h, this function will be changed
585 to accept COMIND directly. */
588 setoptval (const char *com, const char *val)
590 assert (val != NULL);
591 if (!setval_internal (command_by_name (com), com, val))
595 /* Parse OPT into command and value and run it. For example,
596 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
597 This is used by the `--execute' flag in main.c. */
600 run_command (const char *opt)
604 int status = parse_line (opt, &com, &val, &comind);
607 if (!setval_internal (comind, com, val))
612 else if (status == 0)
614 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
620 /* Generic helper functions, for use with `commands'. */
622 /* Forward declarations: */
627 static int decode_string PARAMS ((const char *, const struct decode_item *,
629 static int simple_atoi PARAMS ((const char *, const char *, int *));
630 static int simple_atof PARAMS ((const char *, const char *, double *));
632 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
634 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
635 && TOLOWER((p)[1]) == (c1) \
638 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
639 && TOLOWER((p)[1]) == (c1) \
640 && TOLOWER((p)[2]) == (c2) \
644 /* Store the boolean value from VAL to PLACE. COM is ignored,
645 except for error messages. */
647 cmd_boolean (const char *com, const char *val, void *place)
651 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
652 /* "on", "yes" and "1" mean true. */
654 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
655 /* "off", "no" and "0" mean false. */
660 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
661 exec_name, com, val);
665 *(int *)place = bool_value;
669 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to PLACE.
670 COM is ignored, except for error messages. Values 2 and -1
671 indicate that once defined, the value may not be changed by
672 successive wgetrc files or command-line arguments.
674 Values: 2 - Enable a particular option for good ("always")
675 1 - Enable an option ("on")
676 0 - Disable an option ("off")
677 -1 - Disable an option for good ("never") */
679 cmd_lockable_boolean (const char *com, const char *val, void *place)
681 int lockable_boolean_value;
683 int oldval = *(int *)place;
686 * If a config file said "always" or "never", don't allow command line
687 * arguments to override the config file.
689 if (oldval == -1 || oldval == 2)
692 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
693 lockable_boolean_value = 2;
694 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
695 lockable_boolean_value = 1;
696 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
697 lockable_boolean_value = 0;
698 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
699 lockable_boolean_value = -1;
703 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
704 exec_name, com, val);
708 *(int *)place = lockable_boolean_value;
712 /* Set the non-negative integer value from VAL to PLACE. With
713 incorrect specification, the number remains unchanged. */
715 cmd_number (const char *com, const char *val, void *place)
717 if (!simple_atoi (val, val + strlen (val), place)
718 || *(int *) place < 0)
720 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
721 exec_name, com, val);
727 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
729 cmd_number_inf (const char *com, const char *val, void *place)
731 if (!strcasecmp (val, "inf"))
736 return cmd_number (com, val, place);
739 /* Copy (strdup) the string at COM to a new location and place a
740 pointer to *PLACE. */
742 cmd_string (const char *com, const char *val, void *place)
744 char **pstring = (char **)place;
746 xfree_null (*pstring);
747 *pstring = xstrdup (val);
752 # define ISSEP(c) ((c) == '/')
754 # define ISSEP(c) ((c) == '/' || (c) == '\\')
757 /* Like the above, but handles tilde-expansion when reading a user's
758 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
759 gets expanded to the user's home directory. */
761 cmd_file (const char *com, const char *val, void *place)
763 char **pstring = (char **)place;
765 xfree_null (*pstring);
767 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
769 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
772 *pstring = xstrdup (val);
777 char *home = home_dir ();
781 homelen = strlen (home);
782 while (homelen && ISSEP (home[homelen - 1]))
783 home[--homelen] = '\0';
785 /* Skip the leading "~/". */
786 for (++val; ISSEP (*val); val++)
789 *pstring = concat_strings (home, "/", val, (char *) 0);
793 /* Convert "\" to "/". */
796 for (s = *pstring; *s; s++)
804 /* Like cmd_file, but strips trailing '/' characters. */
806 cmd_directory (const char *com, const char *val, void *place)
810 /* Call cmd_file() for tilde expansion and separator
811 canonicalization (backslash -> slash under Windows). These
812 things should perhaps be in a separate function. */
813 if (!cmd_file (com, val, place))
818 while (t > s && *--t == '/')
824 /* Split VAL by space to a vector of values, and append those values
825 to vector pointed to by the PLACE argument. If VAL is empty, the
826 PLACE vector is cleared instead. */
829 cmd_vector (const char *com, const char *val, void *place)
831 char ***pvec = (char ***)place;
834 *pvec = merge_vecs (*pvec, sepstring (val));
844 cmd_directory_vector (const char *com, const char *val, void *place)
846 char ***pvec = (char ***)place;
850 /* Strip the trailing slashes from directories. */
853 seps = sepstring (val);
854 for (t = seps; t && *t; t++)
856 int len = strlen (*t);
857 /* Skip degenerate case of root directory. */
860 if ((*t)[len - 1] == '/')
861 (*t)[len - 1] = '\0';
864 *pvec = merge_vecs (*pvec, seps);
874 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
875 "100k" or "2.5G" to a floating point number. */
878 parse_bytes_helper (const char *val, double *result)
881 const char *end = val + strlen (val);
883 /* Check for "inf". */
884 if (0 == strcmp (val, "inf"))
890 /* Strip trailing whitespace. */
891 while (val < end && ISSPACE (end[-1]))
896 switch (TOLOWER (end[-1]))
899 --end, mult = 1024.0;
902 --end, mult = 1048576.0;
905 --end, mult = 1073741824.0;
908 --end, mult = 1099511627776.0;
911 /* Not a recognized suffix: assume it's a digit. (If not,
912 simple_atof will raise an error.) */
916 /* Skip leading and trailing whitespace. */
917 while (val < end && ISSPACE (*val))
919 while (val < end && ISSPACE (end[-1]))
924 if (!simple_atof (val, end, &number) || number < 0)
927 *result = number * mult;
931 /* Parse VAL as a number and set its value to PLACE (which should
934 By default, the value is assumed to be in bytes. If "K", "M", or
935 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
936 1<<30, respectively. Floating point values are allowed and are
937 cast to integer before use. The idea is to be able to use things
938 like 1.5k instead of "1536".
940 The string "inf" is returned as 0.
942 In case of error, 0 is returned and memory pointed to by PLACE
943 remains unmodified. */
946 cmd_bytes (const char *com, const char *val, void *place)
949 if (!parse_bytes_helper (val, &byte_value))
951 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
952 exec_name, com, val);
955 *(wgint *)place = (wgint)byte_value;
959 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
960 LARGE_INT. It works by converting the string to double, therefore
961 working with values up to 2^53-1 without loss of precision. This
962 value (8192 TB) is large enough to serve for a while. */
965 cmd_bytes_large (const char *com, const char *val, void *place)
968 if (!parse_bytes_helper (val, &byte_value))
970 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
971 exec_name, com, val);
974 *(LARGE_INT *)place = (LARGE_INT)byte_value;
978 /* Store the value of VAL to *OUT. The value is a time period, by
979 default expressed in seconds, but also accepting suffixes "m", "h",
980 "d", and "w" for minutes, hours, days, and weeks respectively. */
983 cmd_time (const char *com, const char *val, void *place)
986 const char *end = val + strlen (val);
988 /* Strip trailing whitespace. */
989 while (val < end && ISSPACE (end[-1]))
995 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
996 exec_name, com, val);
1000 switch (TOLOWER (end[-1]))
1003 --end, mult = 1; /* seconds */
1006 --end, mult = 60; /* minutes */
1009 --end, mult = 3600; /* hours */
1012 --end, mult = 86400.0; /* days */
1015 --end, mult = 604800.0; /* weeks */
1018 /* Not a recognized suffix: assume it belongs to the number.
1019 (If not, simple_atof will raise an error.) */
1023 /* Skip leading and trailing whitespace. */
1024 while (val < end && ISSPACE (*val))
1026 while (val < end && ISSPACE (end[-1]))
1031 if (!simple_atof (val, end, &number))
1034 *(double *)place = number * mult;
1040 cmd_cert_type (const char *com, const char *val, void *place)
1042 static const struct decode_item choices[] = {
1043 { "pem", keyfile_pem },
1044 { "der", keyfile_asn1 },
1045 { "asn1", keyfile_asn1 },
1047 int ok = decode_string (val, choices, countof (choices), place);
1049 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1054 /* Specialized helper functions, used by `commands' to handle some
1055 options specially. */
1057 static int check_user_specified_header PARAMS ((const char *));
1060 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1062 if (!cmd_boolean (com, val, &opt.dirstruct))
1064 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1065 must be affected inversely. */
1067 opt.no_dirstruct = 0;
1069 opt.no_dirstruct = 1;
1074 cmd_spec_header (const char *com, const char *val, void *place)
1076 if (!check_user_specified_header (val))
1078 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1079 exec_name, com, val);
1082 return cmd_vector (com, val, place);
1086 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1088 int flag = cmd_boolean (com, val, &opt.htmlify);
1089 if (flag && !opt.htmlify)
1090 opt.remove_listing = 0;
1094 /* Set the "mirror" mode. It means: recursive download, timestamping,
1095 no limit on max. recursion depth, and don't remove listings. */
1098 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1102 if (!cmd_boolean (com, val, &mirror))
1107 if (!opt.no_dirstruct)
1109 opt.timestamping = 1;
1110 opt.reclevel = INFINITE_RECURSION;
1111 opt.remove_listing = 0;
1116 /* Validate --prefer-family and set the choice. Allowed values are
1117 "IPv4", "IPv6", and "none". */
1120 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1122 static const struct decode_item choices[] = {
1123 { "IPv4", prefer_ipv4 },
1124 { "IPv6", prefer_ipv6 },
1125 { "none", prefer_none },
1127 int ok = decode_string (val, choices, countof (choices),
1128 (int *) &opt.prefer_family);
1130 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1134 /* Set progress.type to VAL, but verify that it's a valid progress
1135 implementation before that. */
1138 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1140 if (!valid_progress_implementation_p (val))
1142 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1143 exec_name, com, val);
1146 xfree_null (opt.progress_type);
1148 /* Don't call set_progress_implementation here. It will be called
1149 in main() when it becomes clear what the log output is. */
1150 opt.progress_type = xstrdup (val);
1154 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1155 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1159 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1161 if (!cmd_boolean (com, val, &opt.recursive))
1165 if (opt.recursive && !opt.no_dirstruct)
1172 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1174 int restrict_os = opt.restrict_files_os;
1175 int restrict_ctrl = opt.restrict_files_ctrl;
1177 const char *end = strchr (val, ',');
1179 end = val + strlen (val);
1181 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1183 if (VAL_IS ("unix"))
1184 restrict_os = restrict_unix;
1185 else if (VAL_IS ("windows"))
1186 restrict_os = restrict_windows;
1187 else if (VAL_IS ("nocontrol"))
1193 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1194 exec_name, com, val);
1202 if (!strcmp (end + 1, "nocontrol"))
1208 opt.restrict_files_os = restrict_os;
1209 opt.restrict_files_ctrl = restrict_ctrl;
1215 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1217 static const struct decode_item choices[] = {
1218 { "auto", secure_protocol_auto },
1219 { "sslv2", secure_protocol_sslv2 },
1220 { "sslv3", secure_protocol_sslv3 },
1221 { "tlsv1", secure_protocol_tlsv1 },
1223 int ok = decode_string (val, choices, countof (choices), place);
1225 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1230 /* Set all three timeout values. */
1233 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1236 if (!cmd_time (com, val, &value))
1238 opt.read_timeout = value;
1239 opt.connect_timeout = value;
1240 opt.dns_timeout = value;
1245 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1247 /* Just check for empty string and newline, so we don't throw total
1248 junk to the server. */
1249 if (!*val || strchr (val, '\n'))
1251 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1252 exec_name, com, val);
1255 opt.useragent = xstrdup (val);
1259 /* Miscellaneous useful routines. */
1261 /* A very simple atoi clone, more useful than atoi because it works on
1262 delimited strings, and has error reportage. Returns 1 on success,
1263 0 on failure. If successful, stores result to *DEST. */
1266 simple_atoi (const char *beg, const char *end, int *dest)
1270 const char *p = beg;
1272 while (p < end && ISSPACE (*p))
1274 if (p < end && (*p == '-' || *p == '+'))
1276 negative = (*p == '-');
1282 /* Read negative numbers in a separate loop because the most
1283 negative integer cannot be represented as a positive number. */
1286 for (; p < end && ISDIGIT (*p); p++)
1288 int next = (10 * result) + (*p - '0');
1290 return 0; /* overflow */
1294 for (; p < end && ISDIGIT (*p); p++)
1296 int next = (10 * result) - (*p - '0');
1298 return 0; /* underflow */
1309 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1310 doesn't handle exponential notation. Returns 1 on success, 0 on
1311 failure. In case of success, stores its result to *DEST. */
1314 simple_atof (const char *beg, const char *end, double *dest)
1323 const char *p = beg;
1325 while (p < end && ISSPACE (*p))
1327 if (p < end && (*p == '-' || *p == '+'))
1329 negative = (*p == '-');
1333 for (; p < end; p++)
1339 result = (10 * result) + (ch - '0');
1341 result += (ch - '0') / (divider *= 10);
1363 /* Verify that the user-specified header in S is valid. It must
1364 contain a colon preceded by non-white-space characters and must not
1365 contain newlines. */
1368 check_user_specified_header (const char *s)
1372 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1373 /* The header MUST contain `:' preceded by at least one
1374 non-whitespace character. */
1375 if (*p != ':' || p == s)
1377 /* The header MUST NOT contain newlines. */
1378 if (strchr (s, '\n'))
1383 /* Decode VAL into a number, according to ITEMS. */
1386 decode_string (const char *val, const struct decode_item *items, int itemcount,
1390 for (i = 0; i < itemcount; i++)
1391 if (0 == strcasecmp (val, items[i].name))
1393 *place = items[i].code;
1400 void cleanup_html_url PARAMS ((void));
1401 void http_cleanup PARAMS ((void));
1404 /* Free the memory allocated by global variables. */
1408 /* Free external resources, close files, etc. */
1411 extern FILE *output_stream;
1413 fclose (output_stream);
1414 /* No need to check for error because Wget flushes its output (and
1415 checks for errors) after any data arrives. */
1418 /* We're exiting anyway so there's no real need to call free()
1419 hundreds of times. Skipping the frees will make Wget exit
1422 However, when detecting leaks, it's crucial to free() everything
1423 because then you can find the real leaks, i.e. the allocated
1424 memory which grows with the size of the program. */
1430 cleanup_html_url ();
1435 extern acc_t *netrc_list;
1436 free_netrc (netrc_list);
1438 xfree_null (opt.lfilename);
1439 xfree_null (opt.dir_prefix);
1440 xfree_null (opt.input_filename);
1441 xfree_null (opt.output_document);
1442 free_vec (opt.accepts);
1443 free_vec (opt.rejects);
1444 free_vec (opt.excludes);
1445 free_vec (opt.includes);
1446 free_vec (opt.domains);
1447 free_vec (opt.follow_tags);
1448 free_vec (opt.ignore_tags);
1449 xfree_null (opt.progress_type);
1450 xfree (opt.ftp_acc);
1451 xfree_null (opt.ftp_pass);
1452 xfree_null (opt.ftp_proxy);
1453 xfree_null (opt.https_proxy);
1454 xfree_null (opt.http_proxy);
1455 free_vec (opt.no_proxy);
1456 xfree_null (opt.useragent);
1457 xfree_null (opt.referer);
1458 xfree_null (opt.http_user);
1459 xfree_null (opt.http_passwd);
1460 free_vec (opt.user_headers);
1462 xfree_null (opt.cert_file);
1463 xfree_null (opt.private_key);
1465 xfree_null (opt.bind_address);
1466 xfree_null (opt.cookies_input);
1467 xfree_null (opt.cookies_output);
1468 #endif /* DEBUG_MALLOC */