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_passwd, cmd_string }, /* deprecated */
161 { "ftppassword", &opt.ftp_passwd, cmd_string },
162 { "ftpuser", &opt.ftp_user, cmd_string },
163 { "ftpproxy", &opt.ftp_proxy, cmd_string },
164 { "glob", &opt.ftp_glob, cmd_boolean },
165 { "header", &opt.user_headers, cmd_spec_header },
166 { "htmlextension", &opt.html_extension, cmd_boolean },
167 { "htmlify", NULL, cmd_spec_htmlify },
168 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
169 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
170 { "httppassword", &opt.http_passwd, cmd_string },
171 { "httpproxy", &opt.http_proxy, cmd_string },
172 { "httpsproxy", &opt.https_proxy, cmd_string },
173 { "httpuser", &opt.http_user, cmd_string },
174 { "ignorelength", &opt.ignore_length, cmd_boolean },
175 { "ignoretags", &opt.ignore_tags, cmd_vector },
176 { "includedirectories", &opt.includes, cmd_directory_vector },
178 { "inet4only", &opt.ipv4_only, cmd_boolean },
179 { "inet6only", &opt.ipv6_only, cmd_boolean },
181 { "input", &opt.input_filename, cmd_file },
182 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
183 { "killlonger", &opt.kill_longer, cmd_boolean },
184 { "limitrate", &opt.limit_rate, cmd_bytes },
185 { "loadcookies", &opt.cookies_input, cmd_file },
186 { "logfile", &opt.lfilename, cmd_file },
187 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
188 { "mirror", NULL, cmd_spec_mirror },
189 { "netrc", &opt.netrc, cmd_boolean },
190 { "noclobber", &opt.noclobber, cmd_boolean },
191 { "noparent", &opt.no_parent, cmd_boolean },
192 { "noproxy", &opt.no_proxy, cmd_vector },
193 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
194 { "outputdocument", &opt.output_document, cmd_file },
195 { "pagerequisites", &opt.page_requisites, cmd_boolean },
196 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
197 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
198 { "password", &opt.passwd, cmd_string },
199 { "postdata", &opt.post_data, cmd_string },
200 { "postfile", &opt.post_file_name, cmd_file },
201 { "preferfamily", NULL, cmd_spec_prefer_family },
202 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
204 { "privatekey", &opt.private_key, cmd_file },
205 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
207 { "progress", &opt.progress_type, cmd_spec_progress },
208 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
209 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
210 { "proxypassword", &opt.proxy_passwd, cmd_string },
211 { "proxyuser", &opt.proxy_user, cmd_string },
212 { "quiet", &opt.quiet, cmd_boolean },
213 { "quota", &opt.quota, cmd_bytes_large },
215 { "randomfile", &opt.random_file, cmd_file },
217 { "randomwait", &opt.random_wait, cmd_boolean },
218 { "readtimeout", &opt.read_timeout, cmd_time },
219 { "reclevel", &opt.reclevel, cmd_number_inf },
220 { "recursive", NULL, cmd_spec_recursive },
221 { "referer", &opt.referer, cmd_string },
222 { "reject", &opt.rejects, cmd_vector },
223 { "relativeonly", &opt.relative_only, cmd_boolean },
224 { "removelisting", &opt.remove_listing, cmd_boolean },
225 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
226 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
227 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
228 { "robots", &opt.use_robots, cmd_boolean },
229 { "savecookies", &opt.cookies_output, cmd_file },
230 { "saveheaders", &opt.save_headers, cmd_boolean },
232 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
234 { "serverresponse", &opt.server_response, cmd_boolean },
235 { "spanhosts", &opt.spanhost, cmd_boolean },
236 { "spider", &opt.spider, cmd_boolean },
237 { "strictcomments", &opt.strict_comments, cmd_boolean },
238 { "timeout", NULL, cmd_spec_timeout },
239 { "timestamping", &opt.timestamping, cmd_boolean },
240 { "tries", &opt.ntry, cmd_number_inf },
241 { "useproxy", &opt.use_proxy, cmd_boolean },
242 { "user", &opt.user, cmd_string },
243 { "useragent", NULL, cmd_spec_useragent },
244 { "verbose", &opt.verbose, cmd_boolean },
245 { "wait", &opt.wait, cmd_time },
246 { "waitretry", &opt.waitretry, cmd_time }
249 /* Look up CMDNAME in the commands[] and return its position in the
250 array. If CMDNAME is not found, return -1. */
253 command_by_name (const char *cmdname)
255 /* Use binary search for speed. Wget has ~100 commands, which
256 guarantees a worst case performance of 7 string comparisons. */
257 int lo = 0, hi = countof (commands) - 1;
261 int mid = (lo + hi) >> 1;
262 int cmp = strcasecmp (cmdname, commands[mid].name);
273 /* Reset the variables to default values. */
279 /* Most of the default values are 0. Just reset everything, and
280 fill in the non-zero values. Note that initializing pointers to
281 NULL this way is technically illegal, but porting Wget to a
282 machine where NULL is not all-zero bit pattern will be the least
283 of the implementors' worries. */
294 opt.http_keep_alive = 1;
296 tmp = getenv ("no_proxy");
298 opt.no_proxy = sepstring (tmp);
301 opt.read_timeout = 900;
304 opt.remove_listing = 1;
306 opt.dot_bytes = 1024;
307 opt.dot_spacing = 10;
308 opt.dots_in_line = 50;
317 /* The default for file name restriction defaults to the OS type. */
318 #if !defined(WINDOWS) && !defined(__CYGWIN__)
319 opt.restrict_files_os = restrict_unix;
321 opt.restrict_files_os = restrict_windows;
323 opt.restrict_files_ctrl = 1;
326 /* Return the user's home directory (strdup-ed), or NULL if none is
331 char *home = getenv ("HOME");
336 /* If HOME is not defined, try getting it from the password
338 struct passwd *pwd = getpwuid (getuid ());
339 if (!pwd || !pwd->pw_dir)
343 /* Under Windows, if $HOME isn't defined, use the directory where
344 `wget.exe' resides. */
349 return home ? xstrdup (home) : NULL;
352 /* Return the path to the user's .wgetrc. This is either the value of
353 `WGETRC' environment variable, or `$HOME/.wgetrc'.
355 If the `WGETRC' variable exists but the file does not exist, the
356 function will exit(). */
358 wgetrc_file_name (void)
363 /* Try the environment. */
364 env = getenv ("WGETRC");
367 if (!file_exists_p (env))
369 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
373 return xstrdup (env);
376 /* If that failed, try $HOME/.wgetrc. */
379 file = aprintf ("%s/.wgetrc", home);
383 /* Under Windows, if we still haven't found .wgetrc, look for the file
384 `wget.ini' in the directory where `wget.exe' resides; we do this for
385 backward compatibility with previous versions of Wget.
386 SYSTEM_WGETRC should not be defined under WINDOWS. */
387 if (!file || !file_exists_p (file))
393 file = aprintf ("%s/wget.ini", home);
399 if (!file_exists_p (file))
407 /* Return values of parse_line. */
415 static enum parse_line parse_line PARAMS ((const char *, char **,
417 static int setval_internal PARAMS ((int, const char *, const char *));
419 /* Initialize variables from a wgetrc file. Returns zero (failure) if
420 there were errors in the file. */
423 run_wgetrc (const char *file)
430 fp = fopen (file, "rb");
433 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
434 file, strerror (errno));
435 return 1; /* not a fatal error */
437 enable_tilde_expansion = 1;
439 while ((line = read_whole_line (fp)) != NULL)
441 char *com = NULL, *val = NULL;
444 /* Parse the line. */
445 switch (parse_line (line, &com, &val, &comind))
448 /* If everything is OK, set the value. */
449 if (!setval_internal (comind, com, val))
451 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
452 exec_name, file, ln);
456 case line_syntax_error:
457 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
458 exec_name, file, ln);
461 case line_unknown_command:
462 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
463 exec_name, com, file, ln);
476 enable_tilde_expansion = 0;
482 /* Initialize the defaults and run the system wgetrc and user's own
490 /* Load the hard-coded defaults. */
493 /* If SYSTEM_WGETRC is defined, use it. */
495 if (file_exists_p (SYSTEM_WGETRC))
496 ok &= run_wgetrc (SYSTEM_WGETRC);
498 /* Override it with your own, if one exists. */
499 file = wgetrc_file_name ();
502 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
503 something like realpath() before comparing them with `strcmp' */
505 if (!strcmp (file, SYSTEM_WGETRC))
507 fprintf (stderr, _("\
508 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
513 ok &= run_wgetrc (file);
515 /* If there were errors processing either `.wgetrc', abort. */
523 /* Remove dashes and underscores from S, modifying S in the
529 char *t = s; /* t - tortoise */
530 char *h = s; /* h - hare */
532 if (*h == '_' || *h == '-')
539 /* Parse the line pointed by line, with the syntax:
540 <sp>* command <sp>* = <sp>* value <sp>*
541 Uses malloc to allocate space for command and value.
542 If the line is invalid, data is freed and 0 is returned.
544 Returns one of line_ok, line_empty, line_syntax_error, or
545 line_unknown_command.
547 In case of line_ok, *COM and *VAL point to freshly allocated
548 strings, and *COMIND points to com's index. In case of error or
549 empty line, their values are unmodified. */
551 static enum parse_line
552 parse_line (const char *line, char **com, char **val, int *comind)
555 const char *end = line + strlen (line);
556 const char *cmdstart, *cmdend;
557 const char *valstart, *valend;
562 /* Skip leading and trailing whitespace. */
563 while (*line && ISSPACE (*line))
565 while (end > line && ISSPACE (end[-1]))
568 /* Skip empty lines and comments. */
569 if (!*line || *line == '#')
575 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
579 /* Skip '=', as well as any space before or after it. */
580 while (p < end && ISSPACE (*p))
582 if (p == end || *p != '=')
583 return line_syntax_error;
585 while (p < end && ISSPACE (*p))
591 /* The syntax is valid (even though the command might not be). Fill
592 in the command name and value. */
593 *com = strdupdelim (cmdstart, cmdend);
594 *val = strdupdelim (valstart, valend);
596 /* The line now known to be syntactically correct. Check whether
597 the command is valid. */
598 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
600 ind = command_by_name (cmdcopy);
602 return line_unknown_command;
604 /* Report success to the caller. */
609 /* Run commands[comind].action. */
612 setval_internal (int comind, const char *com, const char *val)
614 assert (0 <= comind && comind < countof (commands));
615 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
616 return ((*commands[comind].action) (com, val, commands[comind].place));
619 /* Run command COM with value VAL. If running the command produces an
620 error, report the error and exit.
622 This is intended to be called from main() to modify Wget's behavior
623 through command-line switches. Since COM is hard-coded in main(),
624 it is not canonicalized, and this aborts when COM is not found.
626 If COMIND's are exported to init.h, this function will be changed
627 to accept COMIND directly. */
630 setoptval (const char *com, const char *val)
632 assert (val != NULL);
633 if (!setval_internal (command_by_name (com), com, val))
637 /* Parse OPT into command and value and run it. For example,
638 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
639 This is used by the `--execute' flag in main.c. */
642 run_command (const char *opt)
646 int status = parse_line (opt, &com, &val, &comind);
649 if (!setval_internal (comind, com, val))
654 else if (status == 0)
656 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
662 /* Generic helper functions, for use with `commands'. */
664 /* Forward declarations: */
669 static int decode_string PARAMS ((const char *, const struct decode_item *,
671 static int simple_atoi PARAMS ((const char *, const char *, int *));
672 static int simple_atof PARAMS ((const char *, const char *, double *));
674 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
676 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
677 && TOLOWER((p)[1]) == (c1) \
680 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
681 && TOLOWER((p)[1]) == (c1) \
682 && TOLOWER((p)[2]) == (c2) \
686 /* Store the boolean value from VAL to PLACE. COM is ignored,
687 except for error messages. */
689 cmd_boolean (const char *com, const char *val, void *place)
693 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
694 /* "on", "yes" and "1" mean true. */
696 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
697 /* "off", "no" and "0" mean false. */
702 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
703 exec_name, com, val);
707 *(int *)place = bool_value;
711 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to PLACE.
712 COM is ignored, except for error messages. Values 2 and -1
713 indicate that once defined, the value may not be changed by
714 successive wgetrc files or command-line arguments.
716 Values: 2 - Enable a particular option for good ("always")
717 1 - Enable an option ("on")
718 0 - Disable an option ("off")
719 -1 - Disable an option for good ("never") */
721 cmd_lockable_boolean (const char *com, const char *val, void *place)
723 int lockable_boolean_value;
725 int oldval = *(int *)place;
728 * If a config file said "always" or "never", don't allow command line
729 * arguments to override the config file.
731 if (oldval == -1 || oldval == 2)
734 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
735 lockable_boolean_value = 2;
736 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
737 lockable_boolean_value = 1;
738 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
739 lockable_boolean_value = 0;
740 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
741 lockable_boolean_value = -1;
745 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
746 exec_name, com, val);
750 *(int *)place = lockable_boolean_value;
754 /* Set the non-negative integer value from VAL to PLACE. With
755 incorrect specification, the number remains unchanged. */
757 cmd_number (const char *com, const char *val, void *place)
759 if (!simple_atoi (val, val + strlen (val), place)
760 || *(int *) place < 0)
762 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
763 exec_name, com, val);
769 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
771 cmd_number_inf (const char *com, const char *val, void *place)
773 if (!strcasecmp (val, "inf"))
778 return cmd_number (com, val, place);
781 /* Copy (strdup) the string at COM to a new location and place a
782 pointer to *PLACE. */
784 cmd_string (const char *com, const char *val, void *place)
786 char **pstring = (char **)place;
788 xfree_null (*pstring);
789 *pstring = xstrdup (val);
794 # define ISSEP(c) ((c) == '/')
796 # define ISSEP(c) ((c) == '/' || (c) == '\\')
799 /* Like the above, but handles tilde-expansion when reading a user's
800 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
801 gets expanded to the user's home directory. */
803 cmd_file (const char *com, const char *val, void *place)
805 char **pstring = (char **)place;
807 xfree_null (*pstring);
809 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
811 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
814 *pstring = xstrdup (val);
819 char *home = home_dir ();
823 homelen = strlen (home);
824 while (homelen && ISSEP (home[homelen - 1]))
825 home[--homelen] = '\0';
827 /* Skip the leading "~/". */
828 for (++val; ISSEP (*val); val++)
831 *pstring = concat_strings (home, "/", val, (char *) 0);
835 /* Convert "\" to "/". */
838 for (s = *pstring; *s; s++)
846 /* Like cmd_file, but strips trailing '/' characters. */
848 cmd_directory (const char *com, const char *val, void *place)
852 /* Call cmd_file() for tilde expansion and separator
853 canonicalization (backslash -> slash under Windows). These
854 things should perhaps be in a separate function. */
855 if (!cmd_file (com, val, place))
860 while (t > s && *--t == '/')
866 /* Split VAL by space to a vector of values, and append those values
867 to vector pointed to by the PLACE argument. If VAL is empty, the
868 PLACE vector is cleared instead. */
871 cmd_vector (const char *com, const char *val, void *place)
873 char ***pvec = (char ***)place;
876 *pvec = merge_vecs (*pvec, sepstring (val));
886 cmd_directory_vector (const char *com, const char *val, void *place)
888 char ***pvec = (char ***)place;
892 /* Strip the trailing slashes from directories. */
895 seps = sepstring (val);
896 for (t = seps; t && *t; t++)
898 int len = strlen (*t);
899 /* Skip degenerate case of root directory. */
902 if ((*t)[len - 1] == '/')
903 (*t)[len - 1] = '\0';
906 *pvec = merge_vecs (*pvec, seps);
916 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
917 "100k" or "2.5G" to a floating point number. */
920 parse_bytes_helper (const char *val, double *result)
923 const char *end = val + strlen (val);
925 /* Check for "inf". */
926 if (0 == strcmp (val, "inf"))
932 /* Strip trailing whitespace. */
933 while (val < end && ISSPACE (end[-1]))
938 switch (TOLOWER (end[-1]))
941 --end, mult = 1024.0;
944 --end, mult = 1048576.0;
947 --end, mult = 1073741824.0;
950 --end, mult = 1099511627776.0;
953 /* Not a recognized suffix: assume it's a digit. (If not,
954 simple_atof will raise an error.) */
958 /* Skip leading and trailing whitespace. */
959 while (val < end && ISSPACE (*val))
961 while (val < end && ISSPACE (end[-1]))
966 if (!simple_atof (val, end, &number) || number < 0)
969 *result = number * mult;
973 /* Parse VAL as a number and set its value to PLACE (which should
976 By default, the value is assumed to be in bytes. If "K", "M", or
977 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
978 1<<30, respectively. Floating point values are allowed and are
979 cast to integer before use. The idea is to be able to use things
980 like 1.5k instead of "1536".
982 The string "inf" is returned as 0.
984 In case of error, 0 is returned and memory pointed to by PLACE
985 remains unmodified. */
988 cmd_bytes (const char *com, const char *val, void *place)
991 if (!parse_bytes_helper (val, &byte_value))
993 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
994 exec_name, com, val);
997 *(wgint *)place = (wgint)byte_value;
1001 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1002 LARGE_INT. It works by converting the string to double, therefore
1003 working with values up to 2^53-1 without loss of precision. This
1004 value (8192 TB) is large enough to serve for a while. */
1007 cmd_bytes_large (const char *com, const char *val, void *place)
1010 if (!parse_bytes_helper (val, &byte_value))
1012 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
1013 exec_name, com, val);
1016 *(LARGE_INT *)place = (LARGE_INT)byte_value;
1020 /* Store the value of VAL to *OUT. The value is a time period, by
1021 default expressed in seconds, but also accepting suffixes "m", "h",
1022 "d", and "w" for minutes, hours, days, and weeks respectively. */
1025 cmd_time (const char *com, const char *val, void *place)
1027 double number, mult;
1028 const char *end = val + strlen (val);
1030 /* Strip trailing whitespace. */
1031 while (val < end && ISSPACE (end[-1]))
1037 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1038 exec_name, com, val);
1042 switch (TOLOWER (end[-1]))
1045 --end, mult = 1; /* seconds */
1048 --end, mult = 60; /* minutes */
1051 --end, mult = 3600; /* hours */
1054 --end, mult = 86400.0; /* days */
1057 --end, mult = 604800.0; /* weeks */
1060 /* Not a recognized suffix: assume it belongs to the number.
1061 (If not, simple_atof will raise an error.) */
1065 /* Skip leading and trailing whitespace. */
1066 while (val < end && ISSPACE (*val))
1068 while (val < end && ISSPACE (end[-1]))
1073 if (!simple_atof (val, end, &number))
1076 *(double *)place = number * mult;
1082 cmd_cert_type (const char *com, const char *val, void *place)
1084 static const struct decode_item choices[] = {
1085 { "pem", keyfile_pem },
1086 { "der", keyfile_asn1 },
1087 { "asn1", keyfile_asn1 },
1089 int ok = decode_string (val, choices, countof (choices), place);
1091 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1096 /* Specialized helper functions, used by `commands' to handle some
1097 options specially. */
1099 static int check_user_specified_header PARAMS ((const char *));
1102 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1104 if (!cmd_boolean (com, val, &opt.dirstruct))
1106 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1107 must be affected inversely. */
1109 opt.no_dirstruct = 0;
1111 opt.no_dirstruct = 1;
1116 cmd_spec_header (const char *com, const char *val, void *place)
1118 if (!check_user_specified_header (val))
1120 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1121 exec_name, com, val);
1124 return cmd_vector (com, val, place);
1128 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1130 int flag = cmd_boolean (com, val, &opt.htmlify);
1131 if (flag && !opt.htmlify)
1132 opt.remove_listing = 0;
1136 /* Set the "mirror" mode. It means: recursive download, timestamping,
1137 no limit on max. recursion depth, and don't remove listings. */
1140 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1144 if (!cmd_boolean (com, val, &mirror))
1149 if (!opt.no_dirstruct)
1151 opt.timestamping = 1;
1152 opt.reclevel = INFINITE_RECURSION;
1153 opt.remove_listing = 0;
1158 /* Validate --prefer-family and set the choice. Allowed values are
1159 "IPv4", "IPv6", and "none". */
1162 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1164 static const struct decode_item choices[] = {
1165 { "IPv4", prefer_ipv4 },
1166 { "IPv6", prefer_ipv6 },
1167 { "none", prefer_none },
1169 int ok = decode_string (val, choices, countof (choices),
1170 (int *) &opt.prefer_family);
1172 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1176 /* Set progress.type to VAL, but verify that it's a valid progress
1177 implementation before that. */
1180 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1182 if (!valid_progress_implementation_p (val))
1184 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1185 exec_name, com, val);
1188 xfree_null (opt.progress_type);
1190 /* Don't call set_progress_implementation here. It will be called
1191 in main() when it becomes clear what the log output is. */
1192 opt.progress_type = xstrdup (val);
1196 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1197 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1201 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1203 if (!cmd_boolean (com, val, &opt.recursive))
1207 if (opt.recursive && !opt.no_dirstruct)
1214 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1216 int restrict_os = opt.restrict_files_os;
1217 int restrict_ctrl = opt.restrict_files_ctrl;
1219 const char *end = strchr (val, ',');
1221 end = val + strlen (val);
1223 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1225 if (VAL_IS ("unix"))
1226 restrict_os = restrict_unix;
1227 else if (VAL_IS ("windows"))
1228 restrict_os = restrict_windows;
1229 else if (VAL_IS ("nocontrol"))
1235 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1236 exec_name, com, val);
1244 if (!strcmp (end + 1, "nocontrol"))
1250 opt.restrict_files_os = restrict_os;
1251 opt.restrict_files_ctrl = restrict_ctrl;
1257 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1259 static const struct decode_item choices[] = {
1260 { "auto", secure_protocol_auto },
1261 { "sslv2", secure_protocol_sslv2 },
1262 { "sslv3", secure_protocol_sslv3 },
1263 { "tlsv1", secure_protocol_tlsv1 },
1265 int ok = decode_string (val, choices, countof (choices), place);
1267 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1272 /* Set all three timeout values. */
1275 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1278 if (!cmd_time (com, val, &value))
1280 opt.read_timeout = value;
1281 opt.connect_timeout = value;
1282 opt.dns_timeout = value;
1287 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1289 /* Just check for empty string and newline, so we don't throw total
1290 junk to the server. */
1291 if (!*val || strchr (val, '\n'))
1293 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1294 exec_name, com, val);
1297 xfree_null (opt.useragent);
1298 opt.useragent = xstrdup (val);
1302 /* Miscellaneous useful routines. */
1304 /* A very simple atoi clone, more useful than atoi because it works on
1305 delimited strings, and has error reportage. Returns 1 on success,
1306 0 on failure. If successful, stores result to *DEST. */
1309 simple_atoi (const char *beg, const char *end, int *dest)
1313 const char *p = beg;
1315 while (p < end && ISSPACE (*p))
1317 if (p < end && (*p == '-' || *p == '+'))
1319 negative = (*p == '-');
1325 /* Read negative numbers in a separate loop because the most
1326 negative integer cannot be represented as a positive number. */
1329 for (; p < end && ISDIGIT (*p); p++)
1331 int next = (10 * result) + (*p - '0');
1333 return 0; /* overflow */
1337 for (; p < end && ISDIGIT (*p); p++)
1339 int next = (10 * result) - (*p - '0');
1341 return 0; /* underflow */
1352 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1353 doesn't handle exponential notation. Returns 1 on success, 0 on
1354 failure. In case of success, stores its result to *DEST. */
1357 simple_atof (const char *beg, const char *end, double *dest)
1366 const char *p = beg;
1368 while (p < end && ISSPACE (*p))
1370 if (p < end && (*p == '-' || *p == '+'))
1372 negative = (*p == '-');
1376 for (; p < end; p++)
1382 result = (10 * result) + (ch - '0');
1384 result += (ch - '0') / (divider *= 10);
1406 /* Verify that the user-specified header in S is valid. It must
1407 contain a colon preceded by non-white-space characters and must not
1408 contain newlines. */
1411 check_user_specified_header (const char *s)
1415 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1416 /* The header MUST contain `:' preceded by at least one
1417 non-whitespace character. */
1418 if (*p != ':' || p == s)
1420 /* The header MUST NOT contain newlines. */
1421 if (strchr (s, '\n'))
1426 /* Decode VAL into a number, according to ITEMS. */
1429 decode_string (const char *val, const struct decode_item *items, int itemcount,
1433 for (i = 0; i < itemcount; i++)
1434 if (0 == strcasecmp (val, items[i].name))
1436 *place = items[i].code;
1443 void cleanup_html_url PARAMS ((void));
1444 void http_cleanup PARAMS ((void));
1447 /* Free the memory allocated by global variables. */
1451 /* Free external resources, close files, etc. */
1454 extern FILE *output_stream;
1456 fclose (output_stream);
1457 /* No need to check for error because Wget flushes its output (and
1458 checks for errors) after any data arrives. */
1461 /* We're exiting anyway so there's no real need to call free()
1462 hundreds of times. Skipping the frees will make Wget exit
1465 However, when detecting leaks, it's crucial to free() everything
1466 because then you can find the real leaks, i.e. the allocated
1467 memory which grows with the size of the program. */
1473 cleanup_html_url ();
1478 extern acc_t *netrc_list;
1479 free_netrc (netrc_list);
1481 xfree_null (opt.lfilename);
1482 xfree_null (opt.dir_prefix);
1483 xfree_null (opt.input_filename);
1484 xfree_null (opt.output_document);
1485 free_vec (opt.accepts);
1486 free_vec (opt.rejects);
1487 free_vec (opt.excludes);
1488 free_vec (opt.includes);
1489 free_vec (opt.domains);
1490 free_vec (opt.follow_tags);
1491 free_vec (opt.ignore_tags);
1492 xfree_null (opt.progress_type);
1493 xfree (opt.ftp_acc);
1494 xfree_null (opt.ftp_user);
1495 xfree_null (opt.ftp_passwd);
1496 xfree_null (opt.ftp_proxy);
1497 xfree_null (opt.https_proxy);
1498 xfree_null (opt.http_proxy);
1499 free_vec (opt.no_proxy);
1500 xfree_null (opt.useragent);
1501 xfree_null (opt.referer);
1502 xfree_null (opt.http_user);
1503 xfree_null (opt.http_passwd);
1504 free_vec (opt.user_headers);
1506 xfree_null (opt.cert_file);
1507 xfree_null (opt.private_key);
1509 xfree_null (opt.bind_address);
1510 xfree_null (opt.cookies_input);
1511 xfree_null (opt.cookies_output);
1512 xfree_null (opt.user);
1513 xfree_null (opt.passwd);
1514 #endif /* DEBUG_MALLOC */