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);
77 CMD_DECLARE (cmd_directory_vector);
78 CMD_DECLARE (cmd_lockable_boolean);
79 CMD_DECLARE (cmd_number);
80 CMD_DECLARE (cmd_number_inf);
81 CMD_DECLARE (cmd_string);
82 CMD_DECLARE (cmd_file);
83 CMD_DECLARE (cmd_directory);
84 CMD_DECLARE (cmd_time);
85 CMD_DECLARE (cmd_vector);
88 CMD_DECLARE (cmd_spec_cert_type);
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, closure 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 { "certificatekey", &opt.cert_key, cmd_file },
132 { "certificatetype", &opt.cert_type, cmd_spec_cert_type },
133 { "checkcertificate", &opt.check_cert, cmd_boolean },
135 { "connecttimeout", &opt.connect_timeout, cmd_time },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "deleteafter", &opt.delete_after, cmd_boolean },
144 { "dirprefix", &opt.dir_prefix, cmd_directory },
145 { "dirstruct", NULL, cmd_spec_dirstruct },
146 { "dnscache", &opt.dns_cache, cmd_boolean },
147 { "dnstimeout", &opt.dns_timeout, cmd_time },
148 { "domains", &opt.domains, cmd_vector },
149 { "dotbytes", &opt.dot_bytes, cmd_bytes },
150 { "dotsinline", &opt.dots_in_line, cmd_number },
151 { "dotspacing", &opt.dot_spacing, cmd_number },
152 { "dotstyle", &opt.dot_style, cmd_string },
154 { "egdfile", &opt.egd_file, cmd_file },
156 { "excludedirectories", &opt.excludes, cmd_directory_vector },
157 { "excludedomains", &opt.exclude_domains, cmd_vector },
158 { "followftp", &opt.follow_ftp, cmd_boolean },
159 { "followtags", &opt.follow_tags, cmd_vector },
160 { "forcehtml", &opt.force_html, cmd_boolean },
161 { "ftppasswd", &opt.ftp_pass, cmd_string },
162 { "ftpproxy", &opt.ftp_proxy, cmd_string },
163 { "glob", &opt.ftp_glob, cmd_boolean },
164 { "header", &opt.user_headers, cmd_spec_header },
165 { "htmlextension", &opt.html_extension, cmd_boolean },
166 { "htmlify", NULL, cmd_spec_htmlify },
167 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
168 { "httppasswd", &opt.http_passwd, cmd_string },
169 { "httpproxy", &opt.http_proxy, cmd_string },
170 { "httpsproxy", &opt.https_proxy, cmd_string },
171 { "httpuser", &opt.http_user, cmd_string },
172 { "ignorelength", &opt.ignore_length, cmd_boolean },
173 { "ignoretags", &opt.ignore_tags, cmd_vector },
174 { "includedirectories", &opt.includes, cmd_directory_vector },
176 { "inet4only", &opt.ipv4_only, cmd_boolean },
177 { "inet6only", &opt.ipv6_only, cmd_boolean },
179 { "input", &opt.input_filename, cmd_file },
180 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
181 { "killlonger", &opt.kill_longer, cmd_boolean },
182 { "limitrate", &opt.limit_rate, cmd_bytes },
183 { "loadcookies", &opt.cookies_input, cmd_file },
184 { "logfile", &opt.lfilename, cmd_file },
185 { "login", &opt.ftp_acc, cmd_string },
186 { "mirror", NULL, cmd_spec_mirror },
187 { "netrc", &opt.netrc, cmd_boolean },
188 { "noclobber", &opt.noclobber, cmd_boolean },
189 { "noparent", &opt.no_parent, cmd_boolean },
190 { "noproxy", &opt.no_proxy, cmd_vector },
191 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
192 { "outputdocument", &opt.output_document, cmd_file },
193 { "pagerequisites", &opt.page_requisites, cmd_boolean },
194 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
195 { "postdata", &opt.post_data, cmd_string },
196 { "postfile", &opt.post_file_name, cmd_file },
197 { "preferfamily", NULL, cmd_spec_prefer_family },
198 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
199 { "progress", &opt.progress_type, cmd_spec_progress },
200 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
201 { "proxypasswd", &opt.proxy_passwd, cmd_string },
202 { "proxyuser", &opt.proxy_user, cmd_string },
203 { "quiet", &opt.quiet, cmd_boolean },
204 { "quota", &opt.quota, cmd_bytes_large },
205 { "randomwait", &opt.random_wait, cmd_boolean },
206 { "readtimeout", &opt.read_timeout, cmd_time },
207 { "reclevel", &opt.reclevel, cmd_number_inf },
208 { "recursive", NULL, cmd_spec_recursive },
209 { "referer", &opt.referer, cmd_string },
210 { "reject", &opt.rejects, cmd_vector },
211 { "relativeonly", &opt.relative_only, cmd_boolean },
212 { "removelisting", &opt.remove_listing, cmd_boolean },
213 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
214 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
215 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
216 { "robots", &opt.use_robots, cmd_boolean },
217 { "savecookies", &opt.cookies_output, cmd_file },
218 { "saveheaders", &opt.save_headers, cmd_boolean },
220 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
222 { "serverresponse", &opt.server_response, cmd_boolean },
223 { "spanhosts", &opt.spanhost, cmd_boolean },
224 { "spider", &opt.spider, cmd_boolean },
225 { "strictcomments", &opt.strict_comments, cmd_boolean },
226 { "timeout", NULL, cmd_spec_timeout },
227 { "timestamping", &opt.timestamping, cmd_boolean },
228 { "tries", &opt.ntry, cmd_number_inf },
229 { "useproxy", &opt.use_proxy, cmd_boolean },
230 { "useragent", NULL, cmd_spec_useragent },
231 { "verbose", &opt.verbose, cmd_boolean },
232 { "wait", &opt.wait, cmd_time },
233 { "waitretry", &opt.waitretry, cmd_time }
236 /* Look up CMDNAME in the commands[] and return its position in the
237 array. If CMDNAME is not found, return -1. */
240 command_by_name (const char *cmdname)
242 /* Use binary search for speed. Wget has ~100 commands, which
243 guarantees a worst case performance of 7 string comparisons. */
244 int lo = 0, hi = countof (commands) - 1;
248 int mid = (lo + hi) >> 1;
249 int cmp = strcasecmp (cmdname, commands[mid].name);
260 /* Reset the variables to default values. */
266 /* Most of the default values are 0. Just reset everything, and
267 fill in the non-zero values. Note that initializing pointers to
268 NULL this way is technically illegal, but porting Wget to a
269 machine where NULL is not all-zero bit pattern will be the least
270 of the implementors' worries. */
278 opt.ftp_acc = xstrdup ("anonymous");
279 opt.ftp_pass = xstrdup ("-wget@");
283 opt.http_keep_alive = 1;
285 tmp = getenv ("no_proxy");
287 opt.no_proxy = sepstring (tmp);
290 opt.read_timeout = 900;
293 opt.remove_listing = 1;
295 opt.dot_bytes = 1024;
296 opt.dot_spacing = 10;
297 opt.dots_in_line = 50;
306 /* The default for file name restriction defaults to the OS type. */
307 #if !defined(WINDOWS) && !defined(__CYGWIN__)
308 opt.restrict_files_os = restrict_unix;
310 opt.restrict_files_os = restrict_windows;
312 opt.restrict_files_ctrl = 1;
315 /* Return the user's home directory (strdup-ed), or NULL if none is
320 char *home = getenv ("HOME");
325 /* If HOME is not defined, try getting it from the password
327 struct passwd *pwd = getpwuid (getuid ());
328 if (!pwd || !pwd->pw_dir)
332 /* Under Windows, if $HOME isn't defined, use the directory where
333 `wget.exe' resides. */
338 return home ? xstrdup (home) : NULL;
341 /* Return the path to the user's .wgetrc. This is either the value of
342 `WGETRC' environment variable, or `$HOME/.wgetrc'.
344 If the `WGETRC' variable exists but the file does not exist, the
345 function will exit(). */
347 wgetrc_file_name (void)
352 /* Try the environment. */
353 env = getenv ("WGETRC");
356 if (!file_exists_p (env))
358 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
362 return xstrdup (env);
365 /* If that failed, try $HOME/.wgetrc. */
368 file = aprintf ("%s/.wgetrc", home);
372 /* Under Windows, if we still haven't found .wgetrc, look for the file
373 `wget.ini' in the directory where `wget.exe' resides; we do this for
374 backward compatibility with previous versions of Wget.
375 SYSTEM_WGETRC should not be defined under WINDOWS. */
376 if (!file || !file_exists_p (file))
382 file = aprintf ("%s/wget.ini", home);
388 if (!file_exists_p (file))
396 static int parse_line PARAMS ((const char *, char **, char **, int *));
397 static int setval_internal PARAMS ((int, const char *, const char *));
399 /* Initialize variables from a wgetrc file. */
402 run_wgetrc (const char *file)
408 fp = fopen (file, "rb");
411 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
412 file, strerror (errno));
415 enable_tilde_expansion = 1;
417 while ((line = read_whole_line (fp)) != NULL)
422 /* Parse the line. */
423 status = parse_line (line, &com, &val, &comind);
425 /* If everything is OK, set the value. */
428 if (!setval_internal (comind, com, val))
429 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
434 else if (status == 0)
435 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
439 enable_tilde_expansion = 0;
443 /* Initialize the defaults and run the system wgetrc and user's own
450 /* Load the hard-coded defaults. */
453 /* If SYSTEM_WGETRC is defined, use it. */
455 if (file_exists_p (SYSTEM_WGETRC))
456 run_wgetrc (SYSTEM_WGETRC);
458 /* Override it with your own, if one exists. */
459 file = wgetrc_file_name ();
462 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
463 something like realpath() before comparing them with `strcmp' */
465 if (!strcmp (file, SYSTEM_WGETRC))
467 fprintf (stderr, _("\
468 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
478 /* Remove dashes and underscores from S, modifying S in the
484 char *t = s; /* t - tortoise */
485 char *h = s; /* h - hare */
487 if (*h == '_' || *h == '-')
494 /* Parse the line pointed by line, with the syntax:
495 <sp>* command <sp>* = <sp>* value <sp>*
496 Uses malloc to allocate space for command and value.
497 If the line is invalid, data is freed and 0 is returned.
504 In case of success, *COM and *VAL point to freshly allocated
505 strings, and *COMIND points to com's index. In case of error or
506 empty line, those values are unaffected. */
509 parse_line (const char *line, char **com, char **val, int *comind)
512 const char *end = line + strlen (line);
513 const char *cmdstart, *cmdend;
514 const char *valstart, *valend;
519 /* Skip leading and trailing whitespace. */
520 while (*line && ISSPACE (*line))
522 while (end > line && ISSPACE (end[-1]))
525 /* Skip empty lines and comments. */
526 if (!*line || *line == '#')
532 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
536 /* Skip '=', as well as any space before or after it. */
537 while (p < end && ISSPACE (*p))
539 if (p == end || *p != '=')
542 while (p < end && ISSPACE (*p))
548 /* The line now known to be syntactically correct. Check whether
549 the command is valid. */
550 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
552 ind = command_by_name (cmdcopy);
556 /* The command is valid. Now fill in the values and report success
559 *com = strdupdelim (cmdstart, cmdend);
560 *val = strdupdelim (valstart, valend);
564 /* Run commands[comind].action. */
567 setval_internal (int comind, const char *com, const char *val)
569 assert (0 <= comind && comind < countof (commands));
570 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
571 return ((*commands[comind].action) (com, val, commands[comind].closure));
574 /* Run command COM with value VAL. If running the command produces an
575 error, report the error and exit.
577 This is intended to be called from main() to modify Wget's behavior
578 through command-line switches. Since COM is hard-coded in main(),
579 it is not canonicalized, and this aborts when COM is not found.
581 If COMIND's are exported to init.h, this function will be changed
582 to accept COMIND directly. */
585 setoptval (const char *com, const char *val)
587 assert (val != NULL);
588 if (!setval_internal (command_by_name (com), com, val))
592 /* Parse OPT into command and value and run it. For example,
593 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
594 This is used by the `--execute' flag in main.c. */
597 run_command (const char *opt)
601 int status = parse_line (opt, &com, &val, &comind);
604 if (!setval_internal (comind, com, val))
609 else if (status == 0)
611 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
617 /* Generic helper functions, for use with `commands'. */
619 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
621 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
622 && TOLOWER((p)[1]) == (c1) \
625 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
626 && TOLOWER((p)[1]) == (c1) \
627 && TOLOWER((p)[2]) == (c2) \
631 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
632 except for error messages. */
634 cmd_boolean (const char *com, const char *val, void *closure)
638 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
639 /* "on", "yes" and "1" mean true. */
641 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
642 /* "off", "no" and "0" mean false. */
647 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
648 exec_name, com, val);
652 *(int *)closure = bool_value;
656 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
657 COM is ignored, except for error messages. Values 2 and -1
658 indicate that once defined, the value may not be changed by
659 successive wgetrc files or command-line arguments.
661 Values: 2 - Enable a particular option for good ("always")
662 1 - Enable an option ("on")
663 0 - Disable an option ("off")
664 -1 - Disable an option for good ("never") */
666 cmd_lockable_boolean (const char *com, const char *val, void *closure)
668 int lockable_boolean_value;
670 int oldval = *(int *)closure;
673 * If a config file said "always" or "never", don't allow command line
674 * arguments to override the config file.
676 if (oldval == -1 || oldval == 2)
679 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
680 lockable_boolean_value = 2;
681 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
682 lockable_boolean_value = 1;
683 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
684 lockable_boolean_value = 0;
685 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
686 lockable_boolean_value = -1;
690 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
691 exec_name, com, val);
695 *(int *)closure = lockable_boolean_value;
699 static int simple_atoi PARAMS ((const char *, const char *, int *));
701 /* Set the non-negative integer value from VAL to CLOSURE. With
702 incorrect specification, the number remains unchanged. */
704 cmd_number (const char *com, const char *val, void *closure)
706 if (!simple_atoi (val, val + strlen (val), closure)
707 || *(int *) closure < 0)
709 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
710 exec_name, com, val);
716 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
718 cmd_number_inf (const char *com, const char *val, void *closure)
720 if (!strcasecmp (val, "inf"))
725 return cmd_number (com, val, closure);
728 /* Copy (strdup) the string at COM to a new location and place a
729 pointer to *CLOSURE. */
731 cmd_string (const char *com, const char *val, void *closure)
733 char **pstring = (char **)closure;
735 xfree_null (*pstring);
736 *pstring = xstrdup (val);
741 # define ISSEP(c) ((c) == '/')
743 # define ISSEP(c) ((c) == '/' || (c) == '\\')
746 /* Like the above, but handles tilde-expansion when reading a user's
747 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
748 gets expanded to the user's home directory. */
750 cmd_file (const char *com, const char *val, void *closure)
752 char **pstring = (char **)closure;
754 xfree_null (*pstring);
756 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
758 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
761 *pstring = xstrdup (val);
766 char *home = home_dir ();
770 homelen = strlen (home);
771 while (homelen && ISSEP (home[homelen - 1]))
772 home[--homelen] = '\0';
774 /* Skip the leading "~/". */
775 for (++val; ISSEP (*val); val++)
778 *pstring = concat_strings (home, "/", val, (char *) 0);
782 /* Convert "\" to "/". */
785 for (s = *pstring; *s; s++)
793 /* Like cmd_file, but strips trailing '/' characters. */
795 cmd_directory (const char *com, const char *val, void *closure)
799 /* Call cmd_file() for tilde expansion and separator
800 canonicalization (backslash -> slash under Windows). These
801 things should perhaps be in a separate function. */
802 if (!cmd_file (com, val, closure))
805 s = *(char **)closure;
807 while (t > s && *--t == '/')
813 /* Split VAL by space to a vector of values, and append those values
814 to vector pointed to by the CLOSURE argument. If VAL is empty, the
815 CLOSURE vector is cleared instead. */
818 cmd_vector (const char *com, const char *val, void *closure)
820 char ***pvec = (char ***)closure;
823 *pvec = merge_vecs (*pvec, sepstring (val));
833 cmd_directory_vector (const char *com, const char *val, void *closure)
835 char ***pvec = (char ***)closure;
839 /* Strip the trailing slashes from directories. */
842 seps = sepstring (val);
843 for (t = seps; t && *t; t++)
845 int len = strlen (*t);
846 /* Skip degenerate case of root directory. */
849 if ((*t)[len - 1] == '/')
850 (*t)[len - 1] = '\0';
853 *pvec = merge_vecs (*pvec, seps);
863 static int simple_atof PARAMS ((const char *, const char *, double *));
865 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
866 "100k" or "2.5G" to a floating point number. */
869 parse_bytes_helper (const char *val, double *result)
872 const char *end = val + strlen (val);
874 /* Check for "inf". */
875 if (0 == strcmp (val, "inf"))
881 /* Strip trailing whitespace. */
882 while (val < end && ISSPACE (end[-1]))
887 switch (TOLOWER (end[-1]))
890 --end, mult = 1024.0;
893 --end, mult = 1048576.0;
896 --end, mult = 1073741824.0;
899 --end, mult = 1099511627776.0;
902 /* Not a recognized suffix: assume it's a digit. (If not,
903 simple_atof will raise an error.) */
907 /* Skip leading and trailing whitespace. */
908 while (val < end && ISSPACE (*val))
910 while (val < end && ISSPACE (end[-1]))
915 if (!simple_atof (val, end, &number) || number < 0)
918 *result = number * mult;
922 /* Parse VAL as a number and set its value to CLOSURE (which should
925 By default, the value is assumed to be in bytes. If "K", "M", or
926 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
927 1<<30, respectively. Floating point values are allowed and are
928 cast to integer before use. The idea is to be able to use things
929 like 1.5k instead of "1536".
931 The string "inf" is returned as 0.
933 In case of error, 0 is returned and memory pointed to by CLOSURE
934 remains unmodified. */
937 cmd_bytes (const char *com, const char *val, void *closure)
940 if (!parse_bytes_helper (val, &byte_value))
942 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
943 exec_name, com, val);
946 *(wgint *)closure = (wgint)byte_value;
950 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
951 LARGE_INT. It works by converting the string to double, therefore
952 working with values up to 2^53-1 without loss of precision. This
953 value (8192 TB) is large enough to serve for a while. */
956 cmd_bytes_large (const char *com, const char *val, void *closure)
959 if (!parse_bytes_helper (val, &byte_value))
961 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
962 exec_name, com, val);
965 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
969 /* Store the value of VAL to *OUT. The value is a time period, by
970 default expressed in seconds, but also accepting suffixes "m", "h",
971 "d", and "w" for minutes, hours, days, and weeks respectively. */
974 cmd_time (const char *com, const char *val, void *closure)
977 const char *end = val + strlen (val);
979 /* Strip trailing whitespace. */
980 while (val < end && ISSPACE (end[-1]))
986 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
987 exec_name, com, val);
991 switch (TOLOWER (end[-1]))
994 --end, mult = 1; /* seconds */
997 --end, mult = 60; /* minutes */
1000 --end, mult = 3600; /* hours */
1003 --end, mult = 86400.0; /* days */
1006 --end, mult = 604800.0; /* weeks */
1009 /* Not a recognized suffix: assume it belongs to the number.
1010 (If not, atof simple_atof will raise an error.) */
1014 /* Skip leading and trailing whitespace. */
1015 while (val < end && ISSPACE (*val))
1017 while (val < end && ISSPACE (end[-1]))
1022 if (!simple_atof (val, end, &number))
1025 *(double *)closure = number * mult;
1029 /* Specialized helper functions, used by `commands' to handle some
1030 options specially. */
1032 static int check_user_specified_header PARAMS ((const char *));
1034 struct decode_item {
1038 static int decode_string PARAMS ((const char *, const struct decode_item *,
1043 cmd_spec_cert_type (const char *com, const char *val, void *closure)
1045 static const struct decode_item choices[] = {
1046 { "pem", cert_type_pem },
1047 { "asn1", cert_type_asn1 },
1049 int ok = decode_string (val, choices, countof (choices), closure);
1051 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1057 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1059 if (!cmd_boolean (com, val, &opt.dirstruct))
1061 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1062 must be affected inversely. */
1064 opt.no_dirstruct = 0;
1066 opt.no_dirstruct = 1;
1071 cmd_spec_header (const char *com, const char *val, void *closure)
1073 if (!check_user_specified_header (val))
1075 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1076 exec_name, com, val);
1079 return cmd_vector (com, val, closure);
1083 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1085 int flag = cmd_boolean (com, val, &opt.htmlify);
1086 if (flag && !opt.htmlify)
1087 opt.remove_listing = 0;
1091 /* Set the "mirror" mode. It means: recursive download, timestamping,
1092 no limit on max. recursion depth, and don't remove listings. */
1095 cmd_spec_mirror (const char *com, const char *val, void *closure)
1099 if (!cmd_boolean (com, val, &mirror))
1104 if (!opt.no_dirstruct)
1106 opt.timestamping = 1;
1107 opt.reclevel = INFINITE_RECURSION;
1108 opt.remove_listing = 0;
1113 /* Validate --prefer-family and set the choice. Allowed values are
1114 "IPv4", "IPv6", and "none". */
1117 cmd_spec_prefer_family (const char *com, const char *val, void *closure)
1119 static const struct decode_item choices[] = {
1120 { "IPv4", prefer_ipv4 },
1121 { "IPv6", prefer_ipv6 },
1122 { "none", prefer_none },
1124 int ok = decode_string (val, choices, countof (choices),
1125 (int *) &opt.prefer_family);
1127 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1131 /* Set progress.type to VAL, but verify that it's a valid progress
1132 implementation before that. */
1135 cmd_spec_progress (const char *com, const char *val, void *closure)
1137 if (!valid_progress_implementation_p (val))
1139 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1140 exec_name, com, val);
1143 xfree_null (opt.progress_type);
1145 /* Don't call set_progress_implementation here. It will be called
1146 in main() when it becomes clear what the log output is. */
1147 opt.progress_type = xstrdup (val);
1151 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1152 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1156 cmd_spec_recursive (const char *com, const char *val, void *closure)
1158 if (!cmd_boolean (com, val, &opt.recursive))
1162 if (opt.recursive && !opt.no_dirstruct)
1169 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1171 int restrict_os = opt.restrict_files_os;
1172 int restrict_ctrl = opt.restrict_files_ctrl;
1174 const char *end = strchr (val, ',');
1176 end = val + strlen (val);
1178 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1180 if (VAL_IS ("unix"))
1181 restrict_os = restrict_unix;
1182 else if (VAL_IS ("windows"))
1183 restrict_os = restrict_windows;
1184 else if (VAL_IS ("nocontrol"))
1190 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1191 exec_name, com, val);
1199 if (!strcmp (end + 1, "nocontrol"))
1205 opt.restrict_files_os = restrict_os;
1206 opt.restrict_files_ctrl = restrict_ctrl;
1212 cmd_spec_secure_protocol (const char *com, const char *val, void *closure)
1214 static const struct decode_item choices[] = {
1215 { "auto", secure_protocol_auto },
1216 { "sslv2", secure_protocol_sslv2 },
1217 { "sslv3", secure_protocol_sslv3 },
1218 { "tlsv1", secure_protocol_tlsv1 },
1220 int ok = decode_string (val, choices, countof (choices), closure);
1222 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1227 /* Set all three timeout values. */
1230 cmd_spec_timeout (const char *com, const char *val, void *closure)
1233 if (!cmd_time (com, val, &value))
1235 opt.read_timeout = value;
1236 opt.connect_timeout = value;
1237 opt.dns_timeout = value;
1242 cmd_spec_useragent (const char *com, const char *val, void *closure)
1244 /* Just check for empty string and newline, so we don't throw total
1245 junk to the server. */
1246 if (!*val || strchr (val, '\n'))
1248 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1249 exec_name, com, val);
1252 opt.useragent = xstrdup (val);
1256 /* Miscellaneous useful routines. */
1258 /* A very simple atoi clone, more useful than atoi because it works on
1259 delimited strings, and has error reportage. Returns 1 on success,
1260 0 on failure. If successful, stores result to *DEST. */
1263 simple_atoi (const char *beg, const char *end, int *dest)
1267 const char *p = beg;
1269 while (p < end && ISSPACE (*p))
1271 if (p < end && (*p == '-' || *p == '+'))
1273 negative = (*p == '-');
1279 /* Read negative numbers in a separate loop because the most
1280 negative integer cannot be represented as a positive number. */
1283 for (; p < end && ISDIGIT (*p); p++)
1285 int next = (10 * result) + (*p - '0');
1287 return 0; /* overflow */
1291 for (; p < end && ISDIGIT (*p); p++)
1293 int next = (10 * result) - (*p - '0');
1295 return 0; /* underflow */
1306 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1307 doesn't handle exponential notation. Returns 1 on success, 0 on
1308 failure. In case of success, stores its result to *DEST. */
1311 simple_atof (const char *beg, const char *end, double *dest)
1320 const char *p = beg;
1322 while (p < end && ISSPACE (*p))
1324 if (p < end && (*p == '-' || *p == '+'))
1326 negative = (*p == '-');
1330 for (; p < end; p++)
1336 result = (10 * result) + (ch - '0');
1338 result += (ch - '0') / (divider *= 10);
1360 /* Verify that the user-specified header in S is valid. It must
1361 contain a colon preceded by non-white-space characters and must not
1362 contain newlines. */
1365 check_user_specified_header (const char *s)
1369 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1370 /* The header MUST contain `:' preceded by at least one
1371 non-whitespace character. */
1372 if (*p != ':' || p == s)
1374 /* The header MUST NOT contain newlines. */
1375 if (strchr (s, '\n'))
1380 /* Decode VAL into a number, according to ITEMS. */
1383 decode_string (const char *val, const struct decode_item *items, int itemcount,
1387 for (i = 0; i < itemcount; i++)
1388 if (0 == strcasecmp (val, items[i].name))
1390 *place = items[i].code;
1397 void cleanup_html_url PARAMS ((void));
1398 void http_cleanup PARAMS ((void));
1401 /* Free the memory allocated by global variables. */
1405 /* Free external resources, close files, etc. */
1408 extern FILE *output_stream;
1410 fclose (output_stream);
1411 /* No need to check for error because Wget flushes its output (and
1412 checks for errors) after any data arrives. */
1415 /* We're exiting anyway so there's no real need to call free()
1416 hundreds of times. Skipping the frees will make Wget exit
1419 However, when detecting leaks, it's crucial to free() everything
1420 because then you can find the real leaks, i.e. the allocated
1421 memory which grows with the size of the program. */
1427 cleanup_html_url ();
1432 extern acc_t *netrc_list;
1433 free_netrc (netrc_list);
1435 xfree_null (opt.lfilename);
1436 xfree_null (opt.dir_prefix);
1437 xfree_null (opt.input_filename);
1438 xfree_null (opt.output_document);
1439 free_vec (opt.accepts);
1440 free_vec (opt.rejects);
1441 free_vec (opt.excludes);
1442 free_vec (opt.includes);
1443 free_vec (opt.domains);
1444 free_vec (opt.follow_tags);
1445 free_vec (opt.ignore_tags);
1446 xfree_null (opt.progress_type);
1447 xfree (opt.ftp_acc);
1448 xfree_null (opt.ftp_pass);
1449 xfree_null (opt.ftp_proxy);
1450 xfree_null (opt.https_proxy);
1451 xfree_null (opt.http_proxy);
1452 free_vec (opt.no_proxy);
1453 xfree_null (opt.useragent);
1454 xfree_null (opt.referer);
1455 xfree_null (opt.http_user);
1456 xfree_null (opt.http_passwd);
1457 free_vec (opt.user_headers);
1459 xfree_null (opt.sslcertkey);
1460 xfree_null (opt.sslcertfile);
1461 #endif /* HAVE_SSL */
1462 xfree_null (opt.bind_address);
1463 xfree_null (opt.cookies_input);
1464 xfree_null (opt.cookies_output);