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. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
55 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
56 otherwise, it will be performed by the shell. This variable will
57 be set by the wgetrc-reading function. */
59 static int enable_tilde_expansion;
62 #define CMD_DECLARE(func) static int func (const char *, const char *, void *)
64 CMD_DECLARE (cmd_boolean);
65 CMD_DECLARE (cmd_bytes);
66 CMD_DECLARE (cmd_bytes_large);
68 CMD_DECLARE (cmd_cert_type);
70 CMD_DECLARE (cmd_directory_vector);
71 CMD_DECLARE (cmd_lockable_boolean);
72 CMD_DECLARE (cmd_number);
73 CMD_DECLARE (cmd_number_inf);
74 CMD_DECLARE (cmd_string);
75 CMD_DECLARE (cmd_file);
76 CMD_DECLARE (cmd_directory);
77 CMD_DECLARE (cmd_time);
78 CMD_DECLARE (cmd_vector);
80 CMD_DECLARE (cmd_spec_dirstruct);
81 CMD_DECLARE (cmd_spec_header);
82 CMD_DECLARE (cmd_spec_htmlify);
83 CMD_DECLARE (cmd_spec_mirror);
84 CMD_DECLARE (cmd_spec_prefer_family);
85 CMD_DECLARE (cmd_spec_progress);
86 CMD_DECLARE (cmd_spec_recursive);
87 CMD_DECLARE (cmd_spec_restrict_file_names);
89 CMD_DECLARE (cmd_spec_secure_protocol);
91 CMD_DECLARE (cmd_spec_timeout);
92 CMD_DECLARE (cmd_spec_useragent);
94 /* List of recognized commands, each consisting of name, place and
95 function. When adding a new command, simply add it to the list,
96 but be sure to keep the list sorted alphabetically, as
97 command_by_name depends on it. Also, be sure to add any entries
98 that allocate memory (e.g. cmd_string and cmd_vector) to the
99 cleanup() function below. */
104 int (*action) (const char *, const char *, void *);
106 { "accept", &opt.accepts, cmd_vector },
107 { "addhostdir", &opt.add_hostdir, cmd_boolean },
108 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
109 { "background", &opt.background, cmd_boolean },
110 { "backupconverted", &opt.backup_converted, cmd_boolean },
111 { "backups", &opt.backups, cmd_number },
112 { "base", &opt.base_href, cmd_string },
113 { "bindaddress", &opt.bind_address, cmd_string },
115 { "cacertificate", &opt.ca_cert, cmd_file },
117 { "cache", &opt.allow_cache, cmd_boolean },
119 { "cadirectory", &opt.ca_directory, cmd_directory },
120 { "certificate", &opt.cert_file, cmd_file },
121 { "certificatetype", &opt.cert_type, cmd_cert_type },
122 { "checkcertificate", &opt.check_cert, cmd_boolean },
124 { "connecttimeout", &opt.connect_timeout, cmd_time },
125 { "continue", &opt.always_rest, cmd_boolean },
126 { "convertlinks", &opt.convert_links, cmd_boolean },
127 { "cookies", &opt.cookies, cmd_boolean },
128 { "cutdirs", &opt.cut_dirs, cmd_number },
130 { "debug", &opt.debug, cmd_boolean },
132 { "deleteafter", &opt.delete_after, cmd_boolean },
133 { "dirprefix", &opt.dir_prefix, cmd_directory },
134 { "dirstruct", NULL, cmd_spec_dirstruct },
135 { "dnscache", &opt.dns_cache, cmd_boolean },
136 { "dnstimeout", &opt.dns_timeout, cmd_time },
137 { "domains", &opt.domains, cmd_vector },
138 { "dotbytes", &opt.dot_bytes, cmd_bytes },
139 { "dotsinline", &opt.dots_in_line, cmd_number },
140 { "dotspacing", &opt.dot_spacing, cmd_number },
141 { "dotstyle", &opt.dot_style, cmd_string },
143 { "egdfile", &opt.egd_file, cmd_file },
145 { "excludedirectories", &opt.excludes, cmd_directory_vector },
146 { "excludedomains", &opt.exclude_domains, cmd_vector },
147 { "followftp", &opt.follow_ftp, cmd_boolean },
148 { "followtags", &opt.follow_tags, cmd_vector },
149 { "forcehtml", &opt.force_html, cmd_boolean },
150 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
151 { "ftppassword", &opt.ftp_passwd, cmd_string },
152 { "ftpproxy", &opt.ftp_proxy, cmd_string },
153 { "ftpuser", &opt.ftp_user, cmd_string },
154 { "glob", &opt.ftp_glob, cmd_boolean },
155 { "header", NULL, cmd_spec_header },
156 { "htmlextension", &opt.html_extension, cmd_boolean },
157 { "htmlify", NULL, cmd_spec_htmlify },
158 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
159 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
160 { "httppassword", &opt.http_passwd, cmd_string },
161 { "httpproxy", &opt.http_proxy, cmd_string },
162 { "httpsproxy", &opt.https_proxy, cmd_string },
163 { "httpuser", &opt.http_user, cmd_string },
164 { "ignorelength", &opt.ignore_length, cmd_boolean },
165 { "ignoretags", &opt.ignore_tags, cmd_vector },
166 { "includedirectories", &opt.includes, cmd_directory_vector },
168 { "inet4only", &opt.ipv4_only, cmd_boolean },
169 { "inet6only", &opt.ipv6_only, cmd_boolean },
171 { "input", &opt.input_filename, cmd_file },
172 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
173 { "killlonger", &opt.kill_longer, cmd_boolean },
174 { "limitrate", &opt.limit_rate, cmd_bytes },
175 { "loadcookies", &opt.cookies_input, cmd_file },
176 { "logfile", &opt.lfilename, cmd_file },
177 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
178 { "mirror", NULL, cmd_spec_mirror },
179 { "netrc", &opt.netrc, cmd_boolean },
180 { "noclobber", &opt.noclobber, cmd_boolean },
181 { "noparent", &opt.no_parent, cmd_boolean },
182 { "noproxy", &opt.no_proxy, cmd_vector },
183 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
184 { "outputdocument", &opt.output_document, cmd_file },
185 { "pagerequisites", &opt.page_requisites, cmd_boolean },
186 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
187 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
188 { "password", &opt.passwd, cmd_string },
189 { "postdata", &opt.post_data, cmd_string },
190 { "postfile", &opt.post_file_name, cmd_file },
191 { "preferfamily", NULL, cmd_spec_prefer_family },
192 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
194 { "privatekey", &opt.private_key, cmd_file },
195 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
197 { "progress", &opt.progress_type, cmd_spec_progress },
198 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
199 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
200 { "proxypassword", &opt.proxy_passwd, cmd_string },
201 { "proxyuser", &opt.proxy_user, cmd_string },
202 { "quiet", &opt.quiet, cmd_boolean },
203 { "quota", &opt.quota, cmd_bytes_large },
205 { "randomfile", &opt.random_file, cmd_file },
207 { "randomwait", &opt.random_wait, cmd_boolean },
208 { "readtimeout", &opt.read_timeout, cmd_time },
209 { "reclevel", &opt.reclevel, cmd_number_inf },
210 { "recursive", NULL, cmd_spec_recursive },
211 { "referer", &opt.referer, cmd_string },
212 { "reject", &opt.rejects, cmd_vector },
213 { "relativeonly", &opt.relative_only, cmd_boolean },
214 { "removelisting", &opt.remove_listing, cmd_boolean },
215 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
216 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
217 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
218 { "robots", &opt.use_robots, cmd_boolean },
219 { "savecookies", &opt.cookies_output, cmd_file },
220 { "saveheaders", &opt.save_headers, cmd_boolean },
222 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
224 { "serverresponse", &opt.server_response, cmd_boolean },
225 { "spanhosts", &opt.spanhost, cmd_boolean },
226 { "spider", &opt.spider, cmd_boolean },
227 { "strictcomments", &opt.strict_comments, cmd_boolean },
228 { "timeout", NULL, cmd_spec_timeout },
229 { "timestamping", &opt.timestamping, cmd_boolean },
230 { "tries", &opt.ntry, cmd_number_inf },
231 { "useproxy", &opt.use_proxy, cmd_boolean },
232 { "user", &opt.user, cmd_string },
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. */
284 opt.http_keep_alive = 1;
286 tmp = getenv ("no_proxy");
288 opt.no_proxy = sepstring (tmp);
291 opt.read_timeout = 900;
294 opt.remove_listing = 1;
296 opt.dot_bytes = 1024;
297 opt.dot_spacing = 10;
298 opt.dots_in_line = 50;
307 /* The default for file name restriction defaults to the OS type. */
308 #if !defined(WINDOWS) && !defined(__CYGWIN__)
309 opt.restrict_files_os = restrict_unix;
311 opt.restrict_files_os = restrict_windows;
313 opt.restrict_files_ctrl = 1;
316 /* Return the user's home directory (strdup-ed), or NULL if none is
321 char *home = getenv ("HOME");
326 /* If HOME is not defined, try getting it from the password
328 struct passwd *pwd = getpwuid (getuid ());
329 if (!pwd || !pwd->pw_dir)
333 /* Under Windows, if $HOME isn't defined, use the directory where
334 `wget.exe' resides. */
339 return home ? xstrdup (home) : NULL;
342 /* Return the path to the user's .wgetrc. This is either the value of
343 `WGETRC' environment variable, or `$HOME/.wgetrc'.
345 If the `WGETRC' variable exists but the file does not exist, the
346 function will exit(). */
348 wgetrc_file_name (void)
353 /* Try the environment. */
354 env = getenv ("WGETRC");
357 if (!file_exists_p (env))
359 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
363 return xstrdup (env);
366 /* If that failed, try $HOME/.wgetrc. */
369 file = aprintf ("%s/.wgetrc", home);
373 /* Under Windows, if we still haven't found .wgetrc, look for the file
374 `wget.ini' in the directory where `wget.exe' resides; we do this for
375 backward compatibility with previous versions of Wget.
376 SYSTEM_WGETRC should not be defined under WINDOWS. */
377 if (!file || !file_exists_p (file))
383 file = aprintf ("%s/wget.ini", home);
389 if (!file_exists_p (file))
397 /* Return values of parse_line. */
405 static enum parse_line parse_line (const char *, char **, char **, int *);
406 static int setval_internal (int, const char *, const char *);
408 /* Initialize variables from a wgetrc file. Returns zero (failure) if
409 there were errors in the file. */
412 run_wgetrc (const char *file)
419 fp = fopen (file, "rb");
422 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
423 file, strerror (errno));
424 return 1; /* not a fatal error */
426 enable_tilde_expansion = 1;
428 while ((line = read_whole_line (fp)) != NULL)
430 char *com = NULL, *val = NULL;
433 /* Parse the line. */
434 switch (parse_line (line, &com, &val, &comind))
437 /* If everything is OK, set the value. */
438 if (!setval_internal (comind, com, val))
440 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
441 exec_name, file, ln);
445 case line_syntax_error:
446 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
447 exec_name, file, ln);
450 case line_unknown_command:
451 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
452 exec_name, com, file, ln);
465 enable_tilde_expansion = 0;
471 /* Initialize the defaults and run the system wgetrc and user's own
479 /* Load the hard-coded defaults. */
482 /* If SYSTEM_WGETRC is defined, use it. */
484 if (file_exists_p (SYSTEM_WGETRC))
485 ok &= run_wgetrc (SYSTEM_WGETRC);
487 /* Override it with your own, if one exists. */
488 file = wgetrc_file_name ();
491 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
492 something like realpath() before comparing them with `strcmp' */
494 if (!strcmp (file, SYSTEM_WGETRC))
496 fprintf (stderr, _("\
497 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
502 ok &= run_wgetrc (file);
504 /* If there were errors processing either `.wgetrc', abort. */
512 /* Remove dashes and underscores from S, modifying S in the
518 char *t = s; /* t - tortoise */
519 char *h = s; /* h - hare */
521 if (*h == '_' || *h == '-')
528 /* Parse the line pointed by line, with the syntax:
529 <sp>* command <sp>* = <sp>* value <sp>*
530 Uses malloc to allocate space for command and value.
531 If the line is invalid, data is freed and 0 is returned.
533 Returns one of line_ok, line_empty, line_syntax_error, or
534 line_unknown_command.
536 In case of line_ok, *COM and *VAL point to freshly allocated
537 strings, and *COMIND points to com's index. In case of error or
538 empty line, their values are unmodified. */
540 static enum parse_line
541 parse_line (const char *line, char **com, char **val, int *comind)
544 const char *end = line + strlen (line);
545 const char *cmdstart, *cmdend;
546 const char *valstart, *valend;
551 /* Skip leading and trailing whitespace. */
552 while (*line && ISSPACE (*line))
554 while (end > line && ISSPACE (end[-1]))
557 /* Skip empty lines and comments. */
558 if (!*line || *line == '#')
564 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
568 /* Skip '=', as well as any space before or after it. */
569 while (p < end && ISSPACE (*p))
571 if (p == end || *p != '=')
572 return line_syntax_error;
574 while (p < end && ISSPACE (*p))
580 /* The syntax is valid (even though the command might not be). Fill
581 in the command name and value. */
582 *com = strdupdelim (cmdstart, cmdend);
583 *val = strdupdelim (valstart, valend);
585 /* The line now known to be syntactically correct. Check whether
586 the command is valid. */
587 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
589 ind = command_by_name (cmdcopy);
591 return line_unknown_command;
593 /* Report success to the caller. */
598 /* Run commands[comind].action. */
601 setval_internal (int comind, const char *com, const char *val)
603 assert (0 <= comind && comind < countof (commands));
604 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
605 return commands[comind].action (com, val, commands[comind].place);
608 /* Run command COM with value VAL. If running the command produces an
609 error, report the error and exit.
611 This is intended to be called from main() to modify Wget's behavior
612 through command-line switches. Since COM is hard-coded in main(),
613 it is not canonicalized, and this aborts when COM is not found.
615 If COMIND's are exported to init.h, this function will be changed
616 to accept COMIND directly. */
619 setoptval (const char *com, const char *val, const char *optname)
621 /* Prepend "--" to OPTNAME. */
622 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
625 strcpy (dd_optname + 2, optname);
627 assert (val != NULL);
628 if (!setval_internal (command_by_name (com), dd_optname, val))
632 /* Parse OPT into command and value and run it. For example,
633 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
634 This is used by the `--execute' flag in main.c. */
637 run_command (const char *opt)
641 switch (parse_line (opt, &com, &val, &comind))
644 if (!setval_internal (comind, com, val))
650 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
656 /* Generic helper functions, for use with `commands'. */
658 /* Forward declarations: */
663 static int decode_string (const char *, const struct decode_item *, int, int *);
664 static int simple_atoi (const char *, const char *, int *);
665 static int simple_atof (const char *, const char *, double *);
667 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
669 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
670 && TOLOWER((p)[1]) == (c1) \
673 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
674 && TOLOWER((p)[1]) == (c1) \
675 && TOLOWER((p)[2]) == (c2) \
679 /* Store the boolean value from VAL to PLACE. COM is ignored,
680 except for error messages. */
682 cmd_boolean (const char *com, const char *val, void *place)
686 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
687 /* "on", "yes" and "1" mean true. */
689 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
690 /* "off", "no" and "0" mean false. */
695 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
696 exec_name, com, val);
700 *(int *)place = bool_value;
704 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to PLACE.
705 COM is ignored, except for error messages. Values 2 and -1
706 indicate that once defined, the value may not be changed by
707 successive wgetrc files or command-line arguments.
709 Values: 2 - Enable a particular option for good ("always")
710 1 - Enable an option ("on")
711 0 - Disable an option ("off")
712 -1 - Disable an option for good ("never")
714 #### This hack is currently only used for passive FTP because a
715 contributor had broken scripts specify --passive-ftp where he
716 didn't want it. It should be removed because the same can now be
717 achieved by replacing the wget executable with a script containing:
719 exec wget "$@" --no-passive-ftp
723 cmd_lockable_boolean (const char *com, const char *val, void *place)
725 int lockable_boolean_value;
727 int oldval = *(int *)place;
730 * If a config file said "always" or "never", don't allow command line
731 * arguments to override the config file.
733 if (oldval == -1 || oldval == 2)
736 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, "always"))
741 lockable_boolean_value = 2;
742 else if (0 == strcasecmp (val, "never"))
743 lockable_boolean_value = -1;
747 _("%s: %s: Invalid extended boolean `%s';\n\
748 use one of `on', `off', `always', or `never'.\n"),
749 exec_name, com, val);
753 *(int *)place = lockable_boolean_value;
757 /* Set the non-negative integer value from VAL to PLACE. With
758 incorrect specification, the number remains unchanged. */
760 cmd_number (const char *com, const char *val, void *place)
762 if (!simple_atoi (val, val + strlen (val), place)
763 || *(int *) place < 0)
765 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
766 exec_name, com, val);
772 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
774 cmd_number_inf (const char *com, const char *val, void *place)
776 if (!strcasecmp (val, "inf"))
781 return cmd_number (com, val, place);
784 /* Copy (strdup) the string at COM to a new location and place a
785 pointer to *PLACE. */
787 cmd_string (const char *com, const char *val, void *place)
789 char **pstring = (char **)place;
791 xfree_null (*pstring);
792 *pstring = xstrdup (val);
797 # define ISSEP(c) ((c) == '/')
799 # define ISSEP(c) ((c) == '/' || (c) == '\\')
802 /* Like the above, but handles tilde-expansion when reading a user's
803 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
804 gets expanded to the user's home directory. */
806 cmd_file (const char *com, const char *val, void *place)
808 char **pstring = (char **)place;
810 xfree_null (*pstring);
812 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
814 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
817 *pstring = xstrdup (val);
822 char *home = home_dir ();
826 homelen = strlen (home);
827 while (homelen && ISSEP (home[homelen - 1]))
828 home[--homelen] = '\0';
830 /* Skip the leading "~/". */
831 for (++val; ISSEP (*val); val++)
834 *pstring = concat_strings (home, "/", val, (char *) 0);
838 /* Convert "\" to "/". */
841 for (s = *pstring; *s; s++)
849 /* Like cmd_file, but strips trailing '/' characters. */
851 cmd_directory (const char *com, const char *val, void *place)
855 /* Call cmd_file() for tilde expansion and separator
856 canonicalization (backslash -> slash under Windows). These
857 things should perhaps be in a separate function. */
858 if (!cmd_file (com, val, place))
863 while (t > s && *--t == '/')
869 /* Split VAL by space to a vector of values, and append those values
870 to vector pointed to by the PLACE argument. If VAL is empty, the
871 PLACE vector is cleared instead. */
874 cmd_vector (const char *com, const char *val, void *place)
876 char ***pvec = (char ***)place;
879 *pvec = merge_vecs (*pvec, sepstring (val));
889 cmd_directory_vector (const char *com, const char *val, void *place)
891 char ***pvec = (char ***)place;
895 /* Strip the trailing slashes from directories. */
898 seps = sepstring (val);
899 for (t = seps; t && *t; t++)
901 int len = strlen (*t);
902 /* Skip degenerate case of root directory. */
905 if ((*t)[len - 1] == '/')
906 (*t)[len - 1] = '\0';
909 *pvec = merge_vecs (*pvec, seps);
919 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
920 "100k" or "2.5G" to a floating point number. */
923 parse_bytes_helper (const char *val, double *result)
926 const char *end = val + strlen (val);
928 /* Check for "inf". */
929 if (0 == strcmp (val, "inf"))
935 /* Strip trailing whitespace. */
936 while (val < end && ISSPACE (end[-1]))
941 switch (TOLOWER (end[-1]))
944 --end, mult = 1024.0;
947 --end, mult = 1048576.0;
950 --end, mult = 1073741824.0;
953 --end, mult = 1099511627776.0;
956 /* Not a recognized suffix: assume it's a digit. (If not,
957 simple_atof will raise an error.) */
961 /* Skip leading and trailing whitespace. */
962 while (val < end && ISSPACE (*val))
964 while (val < end && ISSPACE (end[-1]))
969 if (!simple_atof (val, end, &number) || number < 0)
972 *result = number * mult;
976 /* Parse VAL as a number and set its value to PLACE (which should
979 By default, the value is assumed to be in bytes. If "K", "M", or
980 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
981 1<<30, respectively. Floating point values are allowed and are
982 cast to integer before use. The idea is to be able to use things
983 like 1.5k instead of "1536".
985 The string "inf" is returned as 0.
987 In case of error, 0 is returned and memory pointed to by PLACE
988 remains unmodified. */
991 cmd_bytes (const char *com, const char *val, void *place)
994 if (!parse_bytes_helper (val, &byte_value))
996 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
997 exec_name, com, val);
1000 *(wgint *)place = (wgint)byte_value;
1004 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1005 LARGE_INT. It works by converting the string to double, therefore
1006 working with values up to 2^53-1 without loss of precision. This
1007 value (8192 TB) is large enough to serve for a while. */
1010 cmd_bytes_large (const char *com, const char *val, void *place)
1013 if (!parse_bytes_helper (val, &byte_value))
1015 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
1016 exec_name, com, val);
1019 *(LARGE_INT *)place = (LARGE_INT)byte_value;
1023 /* Store the value of VAL to *OUT. The value is a time period, by
1024 default expressed in seconds, but also accepting suffixes "m", "h",
1025 "d", and "w" for minutes, hours, days, and weeks respectively. */
1028 cmd_time (const char *com, const char *val, void *place)
1030 double number, mult;
1031 const char *end = val + strlen (val);
1033 /* Strip trailing whitespace. */
1034 while (val < end && ISSPACE (end[-1]))
1040 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1041 exec_name, com, val);
1045 switch (TOLOWER (end[-1]))
1048 --end, mult = 1; /* seconds */
1051 --end, mult = 60; /* minutes */
1054 --end, mult = 3600; /* hours */
1057 --end, mult = 86400.0; /* days */
1060 --end, mult = 604800.0; /* weeks */
1063 /* Not a recognized suffix: assume it belongs to the number.
1064 (If not, simple_atof will raise an error.) */
1068 /* Skip leading and trailing whitespace. */
1069 while (val < end && ISSPACE (*val))
1071 while (val < end && ISSPACE (end[-1]))
1076 if (!simple_atof (val, end, &number))
1079 *(double *)place = number * mult;
1085 cmd_cert_type (const char *com, const char *val, void *place)
1087 static const struct decode_item choices[] = {
1088 { "pem", keyfile_pem },
1089 { "der", keyfile_asn1 },
1090 { "asn1", keyfile_asn1 },
1092 int ok = decode_string (val, choices, countof (choices), place);
1094 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1099 /* Specialized helper functions, used by `commands' to handle some
1100 options specially. */
1102 static int check_user_specified_header (const char *);
1105 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1107 if (!cmd_boolean (com, val, &opt.dirstruct))
1109 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1110 must be affected inversely. */
1112 opt.no_dirstruct = 0;
1114 opt.no_dirstruct = 1;
1119 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1121 /* Empty value means reset the list of headers. */
1124 free_vec (opt.user_headers);
1125 opt.user_headers = NULL;
1129 if (!check_user_specified_header (val))
1131 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1132 exec_name, com, val);
1135 opt.user_headers = vec_append (opt.user_headers, val);
1140 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1142 int flag = cmd_boolean (com, val, &opt.htmlify);
1143 if (flag && !opt.htmlify)
1144 opt.remove_listing = 0;
1148 /* Set the "mirror" mode. It means: recursive download, timestamping,
1149 no limit on max. recursion depth, and don't remove listings. */
1152 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1156 if (!cmd_boolean (com, val, &mirror))
1161 if (!opt.no_dirstruct)
1163 opt.timestamping = 1;
1164 opt.reclevel = INFINITE_RECURSION;
1165 opt.remove_listing = 0;
1170 /* Validate --prefer-family and set the choice. Allowed values are
1171 "IPv4", "IPv6", and "none". */
1174 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1176 static const struct decode_item choices[] = {
1177 { "IPv4", prefer_ipv4 },
1178 { "IPv6", prefer_ipv6 },
1179 { "none", prefer_none },
1181 int ok = decode_string (val, choices, countof (choices),
1182 (int *) &opt.prefer_family);
1184 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1188 /* Set progress.type to VAL, but verify that it's a valid progress
1189 implementation before that. */
1192 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1194 if (!valid_progress_implementation_p (val))
1196 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1197 exec_name, com, val);
1200 xfree_null (opt.progress_type);
1202 /* Don't call set_progress_implementation here. It will be called
1203 in main() when it becomes clear what the log output is. */
1204 opt.progress_type = xstrdup (val);
1208 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1209 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1213 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1215 if (!cmd_boolean (com, val, &opt.recursive))
1219 if (opt.recursive && !opt.no_dirstruct)
1226 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1228 int restrict_os = opt.restrict_files_os;
1229 int restrict_ctrl = opt.restrict_files_ctrl;
1231 const char *end = strchr (val, ',');
1233 end = val + strlen (val);
1235 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1237 if (VAL_IS ("unix"))
1238 restrict_os = restrict_unix;
1239 else if (VAL_IS ("windows"))
1240 restrict_os = restrict_windows;
1241 else if (VAL_IS ("nocontrol"))
1247 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1248 exec_name, com, val);
1256 if (!strcmp (end + 1, "nocontrol"))
1262 opt.restrict_files_os = restrict_os;
1263 opt.restrict_files_ctrl = restrict_ctrl;
1269 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1271 static const struct decode_item choices[] = {
1272 { "auto", secure_protocol_auto },
1273 { "sslv2", secure_protocol_sslv2 },
1274 { "sslv3", secure_protocol_sslv3 },
1275 { "tlsv1", secure_protocol_tlsv1 },
1277 int ok = decode_string (val, choices, countof (choices), place);
1279 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1284 /* Set all three timeout values. */
1287 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1290 if (!cmd_time (com, val, &value))
1292 opt.read_timeout = value;
1293 opt.connect_timeout = value;
1294 opt.dns_timeout = value;
1299 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1301 /* Disallow embedded newlines. */
1302 if (strchr (val, '\n'))
1304 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1305 exec_name, com, val);
1308 xfree_null (opt.useragent);
1309 opt.useragent = xstrdup (val);
1313 /* Miscellaneous useful routines. */
1315 /* A very simple atoi clone, more useful than atoi because it works on
1316 delimited strings, and has error reportage. Returns 1 on success,
1317 0 on failure. If successful, stores result to *DEST. */
1320 simple_atoi (const char *beg, const char *end, int *dest)
1324 const char *p = beg;
1326 while (p < end && ISSPACE (*p))
1328 if (p < end && (*p == '-' || *p == '+'))
1330 negative = (*p == '-');
1336 /* Read negative numbers in a separate loop because the most
1337 negative integer cannot be represented as a positive number. */
1340 for (; p < end && ISDIGIT (*p); p++)
1342 int next = (10 * result) + (*p - '0');
1344 return 0; /* overflow */
1348 for (; p < end && ISDIGIT (*p); p++)
1350 int next = (10 * result) - (*p - '0');
1352 return 0; /* underflow */
1363 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1364 doesn't handle exponential notation. Returns 1 on success, 0 on
1365 failure. In case of success, stores its result to *DEST. */
1368 simple_atof (const char *beg, const char *end, double *dest)
1377 const char *p = beg;
1379 while (p < end && ISSPACE (*p))
1381 if (p < end && (*p == '-' || *p == '+'))
1383 negative = (*p == '-');
1387 for (; p < end; p++)
1393 result = (10 * result) + (ch - '0');
1395 result += (ch - '0') / (divider *= 10);
1417 /* Verify that the user-specified header in S is valid. It must
1418 contain a colon preceded by non-white-space characters and must not
1419 contain newlines. */
1422 check_user_specified_header (const char *s)
1426 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1427 /* The header MUST contain `:' preceded by at least one
1428 non-whitespace character. */
1429 if (*p != ':' || p == s)
1431 /* The header MUST NOT contain newlines. */
1432 if (strchr (s, '\n'))
1437 /* Decode VAL into a number, according to ITEMS. */
1440 decode_string (const char *val, const struct decode_item *items, int itemcount,
1444 for (i = 0; i < itemcount; i++)
1445 if (0 == strcasecmp (val, items[i].name))
1447 *place = items[i].code;
1454 void cleanup_html_url (void);
1455 void http_cleanup (void);
1458 /* Free the memory allocated by global variables. */
1462 /* Free external resources, close files, etc. */
1465 extern FILE *output_stream;
1467 fclose (output_stream);
1468 /* No need to check for error because Wget flushes its output (and
1469 checks for errors) after any data arrives. */
1472 /* We're exiting anyway so there's no real need to call free()
1473 hundreds of times. Skipping the frees will make Wget exit
1476 However, when detecting leaks, it's crucial to free() everything
1477 because then you can find the real leaks, i.e. the allocated
1478 memory which grows with the size of the program. */
1484 cleanup_html_url ();
1489 extern acc_t *netrc_list;
1490 free_netrc (netrc_list);
1492 xfree_null (opt.lfilename);
1493 xfree_null (opt.dir_prefix);
1494 xfree_null (opt.input_filename);
1495 xfree_null (opt.output_document);
1496 free_vec (opt.accepts);
1497 free_vec (opt.rejects);
1498 free_vec (opt.excludes);
1499 free_vec (opt.includes);
1500 free_vec (opt.domains);
1501 free_vec (opt.follow_tags);
1502 free_vec (opt.ignore_tags);
1503 xfree_null (opt.progress_type);
1504 xfree_null (opt.ftp_user);
1505 xfree_null (opt.ftp_passwd);
1506 xfree_null (opt.ftp_proxy);
1507 xfree_null (opt.https_proxy);
1508 xfree_null (opt.http_proxy);
1509 free_vec (opt.no_proxy);
1510 xfree_null (opt.useragent);
1511 xfree_null (opt.referer);
1512 xfree_null (opt.http_user);
1513 xfree_null (opt.http_passwd);
1514 free_vec (opt.user_headers);
1516 xfree_null (opt.cert_file);
1517 xfree_null (opt.private_key);
1518 xfree_null (opt.ca_directory);
1519 xfree_null (opt.ca_cert);
1520 xfree_null (opt.random_file);
1521 xfree_null (opt.egd_file);
1523 xfree_null (opt.bind_address);
1524 xfree_null (opt.cookies_input);
1525 xfree_null (opt.cookies_output);
1526 xfree_null (opt.user);
1527 xfree_null (opt.passwd);
1528 #endif /* DEBUG_MALLOC */