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);
87 CMD_DECLARE (cmd_spec_dirstruct);
88 CMD_DECLARE (cmd_spec_header);
89 CMD_DECLARE (cmd_spec_htmlify);
90 CMD_DECLARE (cmd_spec_mirror);
91 CMD_DECLARE (cmd_spec_progress);
92 CMD_DECLARE (cmd_spec_recursive);
93 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_timeout);
95 CMD_DECLARE (cmd_spec_useragent);
97 /* List of recognized commands, each consisting of name, closure and
98 function. When adding a new command, simply add it to the list,
99 but be sure to keep the list sorted alphabetically, as
100 command_by_name depends on it. Also, be sure to add any entries
101 that allocate memory (e.g. cmd_string and cmd_vector guys) to the
102 cleanup() function below. */
107 int (*action) PARAMS ((const char *, const char *, void *));
109 { "accept", &opt.accepts, cmd_vector },
110 { "addhostdir", &opt.add_hostdir, cmd_boolean },
111 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
112 { "background", &opt.background, cmd_boolean },
113 { "backupconverted", &opt.backup_converted, cmd_boolean },
114 { "backups", &opt.backups, cmd_number },
115 { "base", &opt.base_href, cmd_string },
116 { "bindaddress", &opt.bind_address, cmd_string },
117 { "cache", &opt.allow_cache, cmd_boolean },
118 { "connecttimeout", &opt.connect_timeout, cmd_time },
119 { "continue", &opt.always_rest, cmd_boolean },
120 { "convertlinks", &opt.convert_links, cmd_boolean },
121 { "cookies", &opt.cookies, cmd_boolean },
122 { "cutdirs", &opt.cut_dirs, cmd_number },
124 { "debug", &opt.debug, cmd_boolean },
126 { "deleteafter", &opt.delete_after, cmd_boolean },
127 { "dirprefix", &opt.dir_prefix, cmd_directory },
128 { "dirstruct", NULL, cmd_spec_dirstruct },
129 { "dnscache", &opt.dns_cache, cmd_boolean },
130 { "dnstimeout", &opt.dns_timeout, cmd_time },
131 { "domains", &opt.domains, cmd_vector },
132 { "dotbytes", &opt.dot_bytes, cmd_bytes },
133 { "dotsinline", &opt.dots_in_line, cmd_number },
134 { "dotspacing", &opt.dot_spacing, cmd_number },
135 { "dotstyle", &opt.dot_style, cmd_string },
137 { "egdfile", &opt.sslegdsock, cmd_file },
139 { "excludedirectories", &opt.excludes, cmd_directory_vector },
140 { "excludedomains", &opt.exclude_domains, cmd_vector },
141 { "followftp", &opt.follow_ftp, cmd_boolean },
142 { "followtags", &opt.follow_tags, cmd_vector },
143 { "forcehtml", &opt.force_html, cmd_boolean },
144 { "ftppasswd", &opt.ftp_pass, cmd_string },
145 { "ftpproxy", &opt.ftp_proxy, cmd_string },
146 { "glob", &opt.ftp_glob, cmd_boolean },
147 { "header", &opt.user_headers, cmd_spec_header },
148 { "htmlextension", &opt.html_extension, cmd_boolean },
149 { "htmlify", NULL, cmd_spec_htmlify },
150 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
151 { "httppasswd", &opt.http_passwd, cmd_string },
152 { "httpproxy", &opt.http_proxy, cmd_string },
153 { "httpsproxy", &opt.https_proxy, cmd_string },
154 { "httpuser", &opt.http_user, cmd_string },
155 { "ignorelength", &opt.ignore_length, cmd_boolean },
156 { "ignoretags", &opt.ignore_tags, cmd_vector },
157 { "includedirectories", &opt.includes, cmd_directory_vector },
159 { "inet4only", &opt.ipv4_only, cmd_boolean },
160 { "inet6only", &opt.ipv6_only, cmd_boolean },
162 { "input", &opt.input_filename, cmd_file },
163 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
164 { "killlonger", &opt.kill_longer, cmd_boolean },
165 { "limitrate", &opt.limit_rate, cmd_bytes },
166 { "loadcookies", &opt.cookies_input, cmd_file },
167 { "logfile", &opt.lfilename, cmd_file },
168 { "login", &opt.ftp_acc, cmd_string },
169 { "mirror", NULL, cmd_spec_mirror },
170 { "netrc", &opt.netrc, cmd_boolean },
171 { "noclobber", &opt.noclobber, cmd_boolean },
172 { "noparent", &opt.no_parent, cmd_boolean },
173 { "noproxy", &opt.no_proxy, cmd_vector },
174 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
175 { "outputdocument", &opt.output_document, cmd_file },
176 { "pagerequisites", &opt.page_requisites, cmd_boolean },
177 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
178 { "postdata", &opt.post_data, cmd_string },
179 { "postfile", &opt.post_file_name, cmd_file },
180 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
181 { "progress", &opt.progress_type, cmd_spec_progress },
182 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
183 { "proxypasswd", &opt.proxy_passwd, cmd_string },
184 { "proxyuser", &opt.proxy_user, cmd_string },
185 { "quiet", &opt.quiet, cmd_boolean },
186 { "quota", &opt.quota, cmd_bytes_large },
187 { "randomwait", &opt.random_wait, cmd_boolean },
188 { "readtimeout", &opt.read_timeout, cmd_time },
189 { "reclevel", &opt.reclevel, cmd_number_inf },
190 { "recursive", NULL, cmd_spec_recursive },
191 { "referer", &opt.referer, cmd_string },
192 { "reject", &opt.rejects, cmd_vector },
193 { "relativeonly", &opt.relative_only, cmd_boolean },
194 { "removelisting", &opt.remove_listing, cmd_boolean },
195 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
196 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
197 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
198 { "robots", &opt.use_robots, cmd_boolean },
199 { "savecookies", &opt.cookies_output, cmd_file },
200 { "saveheaders", &opt.save_headers, cmd_boolean },
201 { "serverresponse", &opt.server_response, cmd_boolean },
202 { "spanhosts", &opt.spanhost, cmd_boolean },
203 { "spider", &opt.spider, cmd_boolean },
205 { "sslcadir", &opt.sslcadir, cmd_directory },
206 { "sslcafile", &opt.sslcafile, cmd_file },
207 { "sslcertfile", &opt.sslcertfile, cmd_file },
208 { "sslcertkey", &opt.sslcertkey, cmd_file },
209 { "sslcerttype", &opt.sslcerttype, cmd_number },
210 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
211 { "sslprotocol", &opt.sslprotocol, cmd_number },
212 #endif /* HAVE_SSL */
213 { "strictcomments", &opt.strict_comments, cmd_boolean },
214 { "timeout", NULL, cmd_spec_timeout },
215 { "timestamping", &opt.timestamping, cmd_boolean },
216 { "tries", &opt.ntry, cmd_number_inf },
217 { "useproxy", &opt.use_proxy, cmd_boolean },
218 { "useragent", NULL, cmd_spec_useragent },
219 { "verbose", &opt.verbose, cmd_boolean },
220 { "wait", &opt.wait, cmd_time },
221 { "waitretry", &opt.waitretry, cmd_time }
224 /* Look up CMDNAME in the commands[] and return its position in the
225 array. If CMDNAME is not found, return -1. */
228 command_by_name (const char *cmdname)
230 /* Use binary search for speed. Wget has ~100 commands, which
231 guarantees a worst case performance of 7 string comparisons. */
232 int lo = 0, hi = countof (commands) - 1;
236 int mid = (lo + hi) >> 1;
237 int cmp = strcasecmp (cmdname, commands[mid].name);
248 /* Reset the variables to default values. */
254 /* Most of the default values are 0. Just reset everything, and
255 fill in the non-zero values. Note that initializing pointers to
256 NULL this way is technically illegal, but porting Wget to a
257 machine where NULL is not all-zero bit pattern will be the least
258 of the implementors' worries. */
266 opt.ftp_acc = xstrdup ("anonymous");
267 opt.ftp_pass = xstrdup ("-wget@");
271 opt.http_keep_alive = 1;
273 tmp = getenv ("no_proxy");
275 opt.no_proxy = sepstring (tmp);
278 opt.read_timeout = 900;
281 opt.remove_listing = 1;
283 opt.dot_bytes = 1024;
284 opt.dot_spacing = 10;
285 opt.dots_in_line = 50;
290 /* The default for file name restriction defaults to the OS type. */
291 #if !defined(WINDOWS) && !defined(__CYGWIN__)
292 opt.restrict_files_os = restrict_unix;
294 opt.restrict_files_os = restrict_windows;
296 opt.restrict_files_ctrl = 1;
299 /* Return the user's home directory (strdup-ed), or NULL if none is
304 char *home = getenv ("HOME");
309 /* If HOME is not defined, try getting it from the password
311 struct passwd *pwd = getpwuid (getuid ());
312 if (!pwd || !pwd->pw_dir)
316 /* Under Windows, if $HOME isn't defined, use the directory where
317 `wget.exe' resides. */
322 return home ? xstrdup (home) : NULL;
325 /* Return the path to the user's .wgetrc. This is either the value of
326 `WGETRC' environment variable, or `$HOME/.wgetrc'.
328 If the `WGETRC' variable exists but the file does not exist, the
329 function will exit(). */
331 wgetrc_file_name (void)
336 /* Try the environment. */
337 env = getenv ("WGETRC");
340 if (!file_exists_p (env))
342 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
346 return xstrdup (env);
349 /* If that failed, try $HOME/.wgetrc. */
352 file = aprintf ("%s/.wgetrc", home);
356 /* Under Windows, if we still haven't found .wgetrc, look for the file
357 `wget.ini' in the directory where `wget.exe' resides; we do this for
358 backward compatibility with previous versions of Wget.
359 SYSTEM_WGETRC should not be defined under WINDOWS. */
360 if (!file || !file_exists_p (file))
366 file = aprintf ("%s/wget.ini", home);
372 if (!file_exists_p (file))
380 static int parse_line PARAMS ((const char *, char **, char **, int *));
381 static int setval_internal PARAMS ((int, const char *, const char *));
383 /* Initialize variables from a wgetrc file. */
386 run_wgetrc (const char *file)
392 fp = fopen (file, "rb");
395 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
396 file, strerror (errno));
399 enable_tilde_expansion = 1;
401 while ((line = read_whole_line (fp)) != NULL)
406 /* Parse the line. */
407 status = parse_line (line, &com, &val, &comind);
409 /* If everything is OK, set the value. */
412 if (!setval_internal (comind, com, val))
413 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
418 else if (status == 0)
419 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
423 enable_tilde_expansion = 0;
427 /* Initialize the defaults and run the system wgetrc and user's own
434 /* Load the hard-coded defaults. */
437 /* If SYSTEM_WGETRC is defined, use it. */
439 if (file_exists_p (SYSTEM_WGETRC))
440 run_wgetrc (SYSTEM_WGETRC);
442 /* Override it with your own, if one exists. */
443 file = wgetrc_file_name ();
446 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
447 something like realpath() before comparing them with `strcmp' */
449 if (!strcmp (file, SYSTEM_WGETRC))
451 fprintf (stderr, _("\
452 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
462 /* Remove dashes and underscores from S, modifying S in the
468 char *t = s; /* t - tortoise */
469 char *h = s; /* h - hare */
471 if (*h == '_' || *h == '-')
478 /* Parse the line pointed by line, with the syntax:
479 <sp>* command <sp>* = <sp>* value <sp>*
480 Uses malloc to allocate space for command and value.
481 If the line is invalid, data is freed and 0 is returned.
488 In case of success, *COM and *VAL point to freshly allocated
489 strings, and *COMIND points to com's index. In case of error or
490 empty line, those values are unaffected. */
493 parse_line (const char *line, char **com, char **val, int *comind)
496 const char *end = line + strlen (line);
497 const char *cmdstart, *cmdend;
498 const char *valstart, *valend;
503 /* Skip leading and trailing whitespace. */
504 while (*line && ISSPACE (*line))
506 while (end > line && ISSPACE (end[-1]))
509 /* Skip empty lines and comments. */
510 if (!*line || *line == '#')
516 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
520 /* Skip '=', as well as any space before or after it. */
521 while (p < end && ISSPACE (*p))
523 if (p == end || *p != '=')
526 while (p < end && ISSPACE (*p))
532 /* The line now known to be syntactically correct. Check whether
533 the command is valid. */
534 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
536 ind = command_by_name (cmdcopy);
540 /* The command is valid. Now fill in the values and report success
543 *com = strdupdelim (cmdstart, cmdend);
544 *val = strdupdelim (valstart, valend);
548 /* Run commands[comind].action. */
551 setval_internal (int comind, const char *com, const char *val)
553 assert (0 <= comind && comind < countof (commands));
554 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
555 return ((*commands[comind].action) (com, val, commands[comind].closure));
558 /* Run command COM with value VAL. If running the command produces an
559 error, report the error and exit.
561 This is intended to be called from main() to modify Wget's behavior
562 through command-line switches. Since COM is hard-coded in main(),
563 it is not canonicalized, and this aborts when COM is not found.
565 If COMIND's are exported to init.h, this function will be changed
566 to accept COMIND directly. */
569 setoptval (const char *com, const char *val)
571 assert (val != NULL);
572 if (!setval_internal (command_by_name (com), com, val))
576 /* Parse OPT into command and value and run it. For example,
577 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
578 This is used by the `--execute' flag in main.c. */
581 run_command (const char *opt)
585 int status = parse_line (opt, &com, &val, &comind);
588 if (!setval_internal (comind, com, val))
593 else if (status == 0)
595 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
601 /* Generic helper functions, for use with `commands'. */
603 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
605 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
606 && TOLOWER((p)[1]) == (c1) \
609 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
610 && TOLOWER((p)[1]) == (c1) \
611 && TOLOWER((p)[2]) == (c2) \
615 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
616 except for error messages. */
618 cmd_boolean (const char *com, const char *val, void *closure)
622 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
623 /* "on", "yes" and "1" mean true. */
625 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
626 /* "off", "no" and "0" mean false. */
631 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
632 exec_name, com, val);
636 *(int *)closure = bool_value;
640 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
641 COM is ignored, except for error messages. Values 2 and -1
642 indicate that once defined, the value may not be changed by
643 successive wgetrc files or command-line arguments.
645 Values: 2 - Enable a particular option for good ("always")
646 1 - Enable an option ("on")
647 0 - Disable an option ("off")
648 -1 - Disable an option for good ("never") */
650 cmd_lockable_boolean (const char *com, const char *val, void *closure)
652 int lockable_boolean_value;
654 int oldval = *(int *)closure;
657 * If a config file said "always" or "never", don't allow command line
658 * arguments to override the config file.
660 if (oldval == -1 || oldval == 2)
663 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
664 lockable_boolean_value = 2;
665 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
666 lockable_boolean_value = 1;
667 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
668 lockable_boolean_value = 0;
669 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
670 lockable_boolean_value = -1;
674 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
675 exec_name, com, val);
679 *(int *)closure = lockable_boolean_value;
683 static int simple_atoi PARAMS ((const char *, const char *, int *));
685 /* Set the non-negative integer value from VAL to CLOSURE. With
686 incorrect specification, the number remains unchanged. */
688 cmd_number (const char *com, const char *val, void *closure)
690 if (!simple_atoi (val, val + strlen (val), closure)
691 || *(int *) closure < 0)
693 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
694 exec_name, com, val);
700 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
702 cmd_number_inf (const char *com, const char *val, void *closure)
704 if (!strcasecmp (val, "inf"))
709 return cmd_number (com, val, closure);
712 /* Copy (strdup) the string at COM to a new location and place a
713 pointer to *CLOSURE. */
715 cmd_string (const char *com, const char *val, void *closure)
717 char **pstring = (char **)closure;
719 xfree_null (*pstring);
720 *pstring = xstrdup (val);
725 # define ISSEP(c) ((c) == '/')
727 # define ISSEP(c) ((c) == '/' || (c) == '\\')
730 /* Like the above, but handles tilde-expansion when reading a user's
731 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
732 gets expanded to the user's home directory. */
734 cmd_file (const char *com, const char *val, void *closure)
736 char **pstring = (char **)closure;
738 xfree_null (*pstring);
740 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
742 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
745 *pstring = xstrdup (val);
750 char *home = home_dir ();
754 homelen = strlen (home);
755 while (homelen && ISSEP (home[homelen - 1]))
756 home[--homelen] = '\0';
758 /* Skip the leading "~/". */
759 for (++val; ISSEP (*val); val++)
762 *pstring = concat_strings (home, "/", val, (char *) 0);
766 /* Convert "\" to "/". */
769 for (s = *pstring; *s; s++)
777 /* Like cmd_file, but strips trailing '/' characters. */
779 cmd_directory (const char *com, const char *val, void *closure)
783 /* Call cmd_file() for tilde expansion and separator
784 canonicalization (backslash -> slash under Windows). These
785 things should perhaps be in a separate function. */
786 if (!cmd_file (com, val, closure))
789 s = *(char **)closure;
791 while (t > s && *--t == '/')
797 /* Split VAL by space to a vector of values, and append those values
798 to vector pointed to by the CLOSURE argument. If VAL is empty, the
799 CLOSURE vector is cleared instead. */
802 cmd_vector (const char *com, const char *val, void *closure)
804 char ***pvec = (char ***)closure;
807 *pvec = merge_vecs (*pvec, sepstring (val));
817 cmd_directory_vector (const char *com, const char *val, void *closure)
819 char ***pvec = (char ***)closure;
823 /* Strip the trailing slashes from directories. */
826 seps = sepstring (val);
827 for (t = seps; t && *t; t++)
829 int len = strlen (*t);
830 /* Skip degenerate case of root directory. */
833 if ((*t)[len - 1] == '/')
834 (*t)[len - 1] = '\0';
837 *pvec = merge_vecs (*pvec, seps);
847 static int simple_atof PARAMS ((const char *, const char *, double *));
849 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
850 "100k" or "2.5G" to a floating point number. */
853 parse_bytes_helper (const char *val, double *result)
856 const char *end = val + strlen (val);
858 /* Check for "inf". */
859 if (0 == strcmp (val, "inf"))
865 /* Strip trailing whitespace. */
866 while (val < end && ISSPACE (end[-1]))
871 switch (TOLOWER (end[-1]))
874 --end, mult = 1024.0;
877 --end, mult = 1048576.0;
880 --end, mult = 1073741824.0;
883 --end, mult = 1099511627776.0;
886 /* Not a recognized suffix: assume it's a digit. (If not,
887 simple_atof will raise an error.) */
891 /* Skip leading and trailing whitespace. */
892 while (val < end && ISSPACE (*val))
894 while (val < end && ISSPACE (end[-1]))
899 if (!simple_atof (val, end, &number) || number < 0)
902 *result = number * mult;
906 /* Parse VAL as a number and set its value to CLOSURE (which should
909 By default, the value is assumed to be in bytes. If "K", "M", or
910 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
911 1<<30, respectively. Floating point values are allowed and are
912 cast to integer before use. The idea is to be able to use things
913 like 1.5k instead of "1536".
915 The string "inf" is returned as 0.
917 In case of error, 0 is returned and memory pointed to by CLOSURE
918 remains unmodified. */
921 cmd_bytes (const char *com, const char *val, void *closure)
924 if (!parse_bytes_helper (val, &byte_value))
926 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
927 exec_name, com, val);
930 *(wgint *)closure = (wgint)byte_value;
934 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
935 LARGE_INT. It works by converting the string to double, therefore
936 working with values up to 2^53-1 without loss of precision. This
937 value (8192 TB) is large enough to serve for a while. */
940 cmd_bytes_large (const char *com, const char *val, void *closure)
943 if (!parse_bytes_helper (val, &byte_value))
945 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
946 exec_name, com, val);
949 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
953 /* Store the value of VAL to *OUT. The value is a time period, by
954 default expressed in seconds, but also accepting suffixes "m", "h",
955 "d", and "w" for minutes, hours, days, and weeks respectively. */
958 cmd_time (const char *com, const char *val, void *closure)
961 const char *end = val + strlen (val);
963 /* Strip trailing whitespace. */
964 while (val < end && ISSPACE (end[-1]))
970 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
971 exec_name, com, val);
975 switch (TOLOWER (end[-1]))
978 --end, mult = 1; /* seconds */
981 --end, mult = 60; /* minutes */
984 --end, mult = 3600; /* hours */
987 --end, mult = 86400.0; /* days */
990 --end, mult = 604800.0; /* weeks */
993 /* Not a recognized suffix: assume it belongs to the number.
994 (If not, atof simple_atof will raise an error.) */
998 /* Skip leading and trailing whitespace. */
999 while (val < end && ISSPACE (*val))
1001 while (val < end && ISSPACE (end[-1]))
1006 if (!simple_atof (val, end, &number))
1009 *(double *)closure = number * mult;
1013 /* Specialized helper functions, used by `commands' to handle some
1014 options specially. */
1016 static int check_user_specified_header PARAMS ((const char *));
1019 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1021 if (!cmd_boolean (com, val, &opt.dirstruct))
1023 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1024 must be affected inversely. */
1026 opt.no_dirstruct = 0;
1028 opt.no_dirstruct = 1;
1033 cmd_spec_header (const char *com, const char *val, void *closure)
1035 if (!check_user_specified_header (val))
1037 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1038 exec_name, com, val);
1041 return cmd_vector (com, val, closure);
1045 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1047 int flag = cmd_boolean (com, val, &opt.htmlify);
1048 if (flag && !opt.htmlify)
1049 opt.remove_listing = 0;
1053 /* Set the "mirror" mode. It means: recursive download, timestamping,
1054 no limit on max. recursion depth, and don't remove listings. */
1057 cmd_spec_mirror (const char *com, const char *val, void *closure)
1061 if (!cmd_boolean (com, val, &mirror))
1066 if (!opt.no_dirstruct)
1068 opt.timestamping = 1;
1069 opt.reclevel = INFINITE_RECURSION;
1070 opt.remove_listing = 0;
1075 /* Set progress.type to VAL, but verify that it's a valid progress
1076 implementation before that. */
1079 cmd_spec_progress (const char *com, const char *val, void *closure)
1081 if (!valid_progress_implementation_p (val))
1083 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1084 exec_name, com, val);
1087 xfree_null (opt.progress_type);
1089 /* Don't call set_progress_implementation here. It will be called
1090 in main() when it becomes clear what the log output is. */
1091 opt.progress_type = xstrdup (val);
1095 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1096 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1100 cmd_spec_recursive (const char *com, const char *val, void *closure)
1102 if (!cmd_boolean (com, val, &opt.recursive))
1106 if (opt.recursive && !opt.no_dirstruct)
1113 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1115 int restrict_os = opt.restrict_files_os;
1116 int restrict_ctrl = opt.restrict_files_ctrl;
1118 const char *end = strchr (val, ',');
1120 end = val + strlen (val);
1122 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1124 if (VAL_IS ("unix"))
1125 restrict_os = restrict_unix;
1126 else if (VAL_IS ("windows"))
1127 restrict_os = restrict_windows;
1128 else if (VAL_IS ("nocontrol"))
1134 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1135 exec_name, com, val);
1143 if (!strcmp (end + 1, "nocontrol"))
1149 opt.restrict_files_os = restrict_os;
1150 opt.restrict_files_ctrl = restrict_ctrl;
1154 /* Set all three timeout values. */
1157 cmd_spec_timeout (const char *com, const char *val, void *closure)
1160 if (!cmd_time (com, val, &value))
1162 opt.read_timeout = value;
1163 opt.connect_timeout = value;
1164 opt.dns_timeout = value;
1169 cmd_spec_useragent (const char *com, const char *val, void *closure)
1171 /* Just check for empty string and newline, so we don't throw total
1172 junk to the server. */
1173 if (!*val || strchr (val, '\n'))
1175 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1176 exec_name, com, val);
1179 opt.useragent = xstrdup (val);
1183 /* Miscellaneous useful routines. */
1185 /* A very simple atoi clone, more useful than atoi because it works on
1186 delimited strings, and has error reportage. Returns 1 on success,
1187 0 on failure. If successful, stores result to *DEST. */
1190 simple_atoi (const char *beg, const char *end, int *dest)
1194 const char *p = beg;
1196 while (p < end && ISSPACE (*p))
1198 if (p < end && (*p == '-' || *p == '+'))
1200 negative = (*p == '-');
1206 /* Read negative numbers in a separate loop because the most
1207 negative integer cannot be represented as a positive number. */
1210 for (; p < end && ISDIGIT (*p); p++)
1212 int next = (10 * result) + (*p - '0');
1214 return 0; /* overflow */
1218 for (; p < end && ISDIGIT (*p); p++)
1220 int next = (10 * result) - (*p - '0');
1222 return 0; /* underflow */
1233 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1234 doesn't handle exponential notation. Returns 1 on success, 0 on
1235 failure. In case of success, stores its result to *DEST. */
1238 simple_atof (const char *beg, const char *end, double *dest)
1247 const char *p = beg;
1249 while (p < end && ISSPACE (*p))
1251 if (p < end && (*p == '-' || *p == '+'))
1253 negative = (*p == '-');
1257 for (; p < end; p++)
1263 result = (10 * result) + (ch - '0');
1265 result += (ch - '0') / (divider *= 10);
1288 check_user_specified_header (const char *s)
1292 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1293 /* The header MUST contain `:' preceded by at least one
1294 non-whitespace character. */
1295 if (*p != ':' || p == s)
1297 /* The header MUST NOT contain newlines. */
1298 if (strchr (s, '\n'))
1303 void cleanup_html_url PARAMS ((void));
1304 void http_cleanup PARAMS ((void));
1307 /* Free the memory allocated by global variables. */
1311 /* Free external resources, close files, etc. */
1314 extern FILE *output_stream;
1316 fclose (output_stream);
1317 /* No need to check for error because Wget flushes its output (and
1318 checks for errors) after any data arrives. */
1321 /* We're exiting anyway so there's no real need to call free()
1322 hundreds of times. Skipping the frees will make Wget exit
1325 However, when detecting leaks, it's crucial to free() everything
1326 because then you can find the real leaks, i.e. the allocated
1327 memory which grows with the size of the program. */
1333 cleanup_html_url ();
1338 extern acc_t *netrc_list;
1339 free_netrc (netrc_list);
1341 xfree_null (opt.lfilename);
1342 xfree_null (opt.dir_prefix);
1343 xfree_null (opt.input_filename);
1344 xfree_null (opt.output_document);
1345 free_vec (opt.accepts);
1346 free_vec (opt.rejects);
1347 free_vec (opt.excludes);
1348 free_vec (opt.includes);
1349 free_vec (opt.domains);
1350 free_vec (opt.follow_tags);
1351 free_vec (opt.ignore_tags);
1352 xfree_null (opt.progress_type);
1353 xfree (opt.ftp_acc);
1354 xfree_null (opt.ftp_pass);
1355 xfree_null (opt.ftp_proxy);
1356 xfree_null (opt.https_proxy);
1357 xfree_null (opt.http_proxy);
1358 free_vec (opt.no_proxy);
1359 xfree_null (opt.useragent);
1360 xfree_null (opt.referer);
1361 xfree_null (opt.http_user);
1362 xfree_null (opt.http_passwd);
1363 free_vec (opt.user_headers);
1365 xfree_null (opt.sslcertkey);
1366 xfree_null (opt.sslcertfile);
1367 #endif /* HAVE_SSL */
1368 xfree_null (opt.bind_address);
1369 xfree_null (opt.cookies_input);
1370 xfree_null (opt.cookies_output);