1 /* Reading/parsing the initialization file.
2 Copyright (C) 1995, 1996, 1997, 1998, 2000, 2001, 2003
3 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget; if not, write to the Free Software
19 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
21 In addition, as a special exception, the Free Software Foundation
22 gives permission to link the code of its release of Wget with the
23 OpenSSL project's "OpenSSL" library (or with modified versions of it
24 that use the same license as the "OpenSSL" library), and distribute
25 the linked executables. You must obey the GNU General Public License
26 in all respects for all of the code used other than "OpenSSL". If you
27 modify this file, you may extend this exception to your version of the
28 file, but you are not obligated to do so. If you do not wish to do
29 so, delete this exception statement from your version. */
34 #include <sys/types.h>
49 # include <sys/socket.h>
50 # include <netinet/in.h>
52 # include <arpa/inet.h>
66 #include "cookies.h" /* for cookie_jar_delete */
68 #include "recur.h" /* for INFINITE_RECURSION */
74 extern struct cookie_jar *wget_cookie_jar;
76 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
77 otherwise, it will be performed by the shell. This variable will
78 be set by the wgetrc-reading function. */
80 static int enable_tilde_expansion;
83 #define CMD_DECLARE(func) static int func \
84 PARAMS ((const char *, const char *, void *))
86 CMD_DECLARE (cmd_boolean);
87 CMD_DECLARE (cmd_bytes);
88 CMD_DECLARE (cmd_bytes_large);
89 CMD_DECLARE (cmd_directory_vector);
90 CMD_DECLARE (cmd_lockable_boolean);
91 CMD_DECLARE (cmd_number);
92 CMD_DECLARE (cmd_number_inf);
93 CMD_DECLARE (cmd_string);
94 CMD_DECLARE (cmd_file);
95 CMD_DECLARE (cmd_directory);
96 CMD_DECLARE (cmd_time);
97 CMD_DECLARE (cmd_vector);
99 CMD_DECLARE (cmd_spec_dirstruct);
100 CMD_DECLARE (cmd_spec_header);
101 CMD_DECLARE (cmd_spec_htmlify);
102 CMD_DECLARE (cmd_spec_mirror);
103 CMD_DECLARE (cmd_spec_progress);
104 CMD_DECLARE (cmd_spec_recursive);
105 CMD_DECLARE (cmd_spec_restrict_file_names);
106 CMD_DECLARE (cmd_spec_timeout);
107 CMD_DECLARE (cmd_spec_useragent);
109 /* List of recognized commands, each consisting of name, closure and
110 function. When adding a new command, simply add it to the list,
111 but be sure to keep the list sorted alphabetically, as
112 command_by_name depends on it. Also, be sure to add any entries
113 that allocate memory (e.g. cmd_string and cmd_vector guys) to the
114 cleanup() function below. */
119 int (*action) PARAMS ((const char *, const char *, void *));
121 { "accept", &opt.accepts, cmd_vector },
122 { "addhostdir", &opt.add_hostdir, cmd_boolean },
123 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
124 { "background", &opt.background, cmd_boolean },
125 { "backupconverted", &opt.backup_converted, cmd_boolean },
126 { "backups", &opt.backups, cmd_number },
127 { "base", &opt.base_href, cmd_string },
128 { "bindaddress", &opt.bind_address, cmd_string },
129 { "cache", &opt.allow_cache, cmd_boolean },
130 { "connecttimeout", &opt.connect_timeout, cmd_time },
131 { "continue", &opt.always_rest, cmd_boolean },
132 { "convertlinks", &opt.convert_links, cmd_boolean },
133 { "cookies", &opt.cookies, cmd_boolean },
134 { "cutdirs", &opt.cut_dirs, cmd_number },
136 { "debug", &opt.debug, cmd_boolean },
138 { "deleteafter", &opt.delete_after, cmd_boolean },
139 { "dirprefix", &opt.dir_prefix, cmd_directory },
140 { "dirstruct", NULL, cmd_spec_dirstruct },
141 { "dnscache", &opt.dns_cache, cmd_boolean },
142 { "dnstimeout", &opt.dns_timeout, cmd_time },
143 { "domains", &opt.domains, cmd_vector },
144 { "dotbytes", &opt.dot_bytes, cmd_bytes },
145 { "dotsinline", &opt.dots_in_line, cmd_number },
146 { "dotspacing", &opt.dot_spacing, cmd_number },
147 { "dotstyle", &opt.dot_style, cmd_string },
149 { "egdfile", &opt.sslegdsock, cmd_file },
151 { "excludedirectories", &opt.excludes, cmd_directory_vector },
152 { "excludedomains", &opt.exclude_domains, cmd_vector },
153 { "followftp", &opt.follow_ftp, cmd_boolean },
154 { "followtags", &opt.follow_tags, cmd_vector },
155 { "forcehtml", &opt.force_html, cmd_boolean },
156 { "ftpproxy", &opt.ftp_proxy, cmd_string },
157 { "glob", &opt.ftp_glob, cmd_boolean },
158 { "header", NULL, cmd_spec_header },
159 { "htmlextension", &opt.html_extension, cmd_boolean },
160 { "htmlify", NULL, cmd_spec_htmlify },
161 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
162 { "httppasswd", &opt.http_passwd, cmd_string },
163 { "httpproxy", &opt.http_proxy, cmd_string },
164 { "httpsproxy", &opt.https_proxy, cmd_string },
165 { "httpuser", &opt.http_user, cmd_string },
166 { "ignorelength", &opt.ignore_length, cmd_boolean },
167 { "ignoretags", &opt.ignore_tags, cmd_vector },
168 { "includedirectories", &opt.includes, cmd_directory_vector },
170 { "inet4only", &opt.ipv4_only, cmd_boolean },
171 { "inet6only", &opt.ipv6_only, cmd_boolean },
173 { "input", &opt.input_filename, cmd_file },
174 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
175 { "killlonger", &opt.kill_longer, cmd_boolean },
176 { "limitrate", &opt.limit_rate, cmd_bytes },
177 { "loadcookies", &opt.cookies_input, cmd_file },
178 { "logfile", &opt.lfilename, cmd_file },
179 { "login", &opt.ftp_acc, cmd_string },
180 { "mirror", NULL, cmd_spec_mirror },
181 { "netrc", &opt.netrc, cmd_boolean },
182 { "noclobber", &opt.noclobber, cmd_boolean },
183 { "noparent", &opt.no_parent, cmd_boolean },
184 { "noproxy", &opt.no_proxy, cmd_vector },
185 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
186 { "outputdocument", &opt.output_document, cmd_file },
187 { "pagerequisites", &opt.page_requisites, cmd_boolean },
188 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
189 { "passwd", &opt.ftp_pass, cmd_string },
190 { "postdata", &opt.post_data, cmd_string },
191 { "postfile", &opt.post_file_name, cmd_file },
192 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
193 { "progress", &opt.progress_type, cmd_spec_progress },
194 { "proxypasswd", &opt.proxy_passwd, cmd_string },
195 { "proxyuser", &opt.proxy_user, cmd_string },
196 { "quiet", &opt.quiet, cmd_boolean },
197 { "quota", &opt.quota, cmd_bytes_large },
198 { "randomwait", &opt.random_wait, cmd_boolean },
199 { "readtimeout", &opt.read_timeout, cmd_time },
200 { "reclevel", &opt.reclevel, cmd_number_inf },
201 { "recursive", NULL, cmd_spec_recursive },
202 { "referer", &opt.referer, cmd_string },
203 { "reject", &opt.rejects, cmd_vector },
204 { "relativeonly", &opt.relative_only, cmd_boolean },
205 { "removelisting", &opt.remove_listing, cmd_boolean },
206 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
207 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
208 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
209 { "robots", &opt.use_robots, cmd_boolean },
210 { "savecookies", &opt.cookies_output, cmd_file },
211 { "saveheaders", &opt.save_headers, cmd_boolean },
212 { "serverresponse", &opt.server_response, cmd_boolean },
213 { "spanhosts", &opt.spanhost, cmd_boolean },
214 { "spider", &opt.spider, cmd_boolean },
216 { "sslcadir", &opt.sslcadir, cmd_directory },
217 { "sslcafile", &opt.sslcafile, cmd_file },
218 { "sslcertfile", &opt.sslcertfile, cmd_file },
219 { "sslcertkey", &opt.sslcertkey, cmd_file },
220 { "sslcerttype", &opt.sslcerttype, cmd_number },
221 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
222 { "sslprotocol", &opt.sslprotocol, cmd_number },
223 #endif /* HAVE_SSL */
224 { "strictcomments", &opt.strict_comments, cmd_boolean },
225 { "timeout", NULL, cmd_spec_timeout },
226 { "timestamping", &opt.timestamping, cmd_boolean },
227 { "tries", &opt.ntry, cmd_number_inf },
228 { "useproxy", &opt.use_proxy, cmd_boolean },
229 { "useragent", NULL, cmd_spec_useragent },
230 { "verbose", &opt.verbose, cmd_boolean },
231 { "wait", &opt.wait, cmd_time },
232 { "waitretry", &opt.waitretry, cmd_time }
235 /* Look up CMDNAME in the commands[] and return its position in the
236 array. If CMDNAME is not found, return -1. */
239 command_by_name (const char *cmdname)
241 /* Use binary search for speed. Wget has ~100 commands, which
242 guarantees a worst case performance of 7 string comparisons. */
243 int lo = 0, hi = countof (commands) - 1;
247 int mid = (lo + hi) >> 1;
248 int cmp = strcasecmp (cmdname, commands[mid].name);
259 /* Reset the variables to default values. */
265 /* Most of the default values are 0. Just reset everything, and
266 fill in the non-zero values. Note that initializing pointers to
267 NULL this way is technically illegal, but porting Wget to a
268 machine where NULL is not all-zero bit pattern will be the least
269 of the implementors' worries. */
277 opt.ftp_acc = xstrdup ("anonymous");
278 opt.ftp_pass = xstrdup ("-wget@");
282 opt.http_keep_alive = 1;
284 tmp = getenv ("no_proxy");
286 opt.no_proxy = sepstring (tmp);
289 opt.read_timeout = 900;
292 opt.remove_listing = 1;
294 opt.dot_bytes = 1024;
295 opt.dot_spacing = 10;
296 opt.dots_in_line = 50;
300 /* The default for file name restriction defaults to the OS type. */
301 #if !defined(WINDOWS) && !defined(__CYGWIN__)
302 opt.restrict_files_os = restrict_unix;
304 opt.restrict_files_os = restrict_windows;
306 opt.restrict_files_ctrl = 1;
309 # ifndef HAVE_GETADDRINFO_AI_ADDRCONFIG
310 /* If IPv6 is enabled, but AI_ADDRCONFIG is missing, check whether
311 we can create AF_INET6 sockets. If we can't, turn on the
312 --inet4-only setting. This is necessary because on some systems
313 (e.g. RHL 9) getaddrinfo resolves AAAA records, but socket()
314 can't even create an AF_INET6 socket, let alone connect to IPv6
315 hosts. To avoid "address family not supported" error messages,
318 We do it as early as here, so that the user can revert the
319 settingn using --no-inet4-only, in case he wants to see the error
320 messages, for whatever reason. */
322 int sock = socket (AF_INET6, SOCK_STREAM, 0);
324 opt.ipv4_only = -1; /* special value -1 because the option
325 was not specified by the user. */
329 # endif /* not HAVE_GETADDRINFO_AI_ADDRCONFIG */
330 #endif /* ENABLE_IPV6 */
333 /* Return the user's home directory (strdup-ed), or NULL if none is
338 char *home = getenv ("HOME");
343 /* If HOME is not defined, try getting it from the password
345 struct passwd *pwd = getpwuid (getuid ());
346 if (!pwd || !pwd->pw_dir)
351 /* #### Maybe I should grab home_dir from registry, but the best
352 that I could get from there is user's Start menu. It sucks! */
356 return home ? xstrdup (home) : NULL;
359 /* Return the path to the user's .wgetrc. This is either the value of
360 `WGETRC' environment variable, or `$HOME/.wgetrc'.
362 If the `WGETRC' variable exists but the file does not exist, the
363 function will exit(). */
365 wgetrc_file_name (void)
370 /* Try the environment. */
371 env = getenv ("WGETRC");
374 if (!file_exists_p (env))
376 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
380 return xstrdup (env);
384 /* If that failed, try $HOME/.wgetrc. */
388 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
389 sprintf (file, "%s/.wgetrc", home);
393 /* Under Windows, "home" is (for the purposes of this function) the
394 directory where `wget.exe' resides, and `wget.ini' will be used
395 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
397 It is not as trivial as I assumed, because on 95 argv[0] is full
398 path, but on NT you get what you typed in command line. --dbudor */
402 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
403 sprintf (file, "%swget.ini", home);
409 if (!file_exists_p (file))
417 static int parse_line PARAMS ((const char *, char **, char **, int *));
418 static int setval_internal PARAMS ((int, const char *, const char *));
420 /* Initialize variables from a wgetrc file. */
423 run_wgetrc (const char *file)
429 fp = fopen (file, "rb");
432 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
433 file, strerror (errno));
436 enable_tilde_expansion = 1;
438 while ((line = read_whole_line (fp)))
443 /* Parse the line. */
444 status = parse_line (line, &com, &val, &comind);
446 /* If everything is OK, set the value. */
449 if (!setval_internal (comind, com, val))
450 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
455 else if (status == 0)
456 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
460 enable_tilde_expansion = 0;
464 /* Initialize the defaults and run the system wgetrc and user's own
471 /* Load the hard-coded defaults. */
474 /* If SYSTEM_WGETRC is defined, use it. */
476 if (file_exists_p (SYSTEM_WGETRC))
477 run_wgetrc (SYSTEM_WGETRC);
479 /* Override it with your own, if one exists. */
480 file = wgetrc_file_name ();
483 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
484 something like realpath() before comparing them with `strcmp' */
486 if (!strcmp (file, SYSTEM_WGETRC))
488 fprintf (stderr, _("\
489 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
499 /* Remove dashes and underscores from S, modifying S in the
505 char *t = s; /* t - tortoise */
506 char *h = s; /* h - hare */
508 if (*h == '_' || *h == '-')
515 /* Parse the line pointed by line, with the syntax:
516 <sp>* command <sp>* = <sp>* value <sp>*
517 Uses malloc to allocate space for command and value.
518 If the line is invalid, data is freed and 0 is returned.
525 In case of success, *COM and *VAL point to freshly allocated
526 strings, and *COMIND points to com's index. In case of error or
527 empty line, those values are unaffected. */
530 parse_line (const char *line, char **com, char **val, int *comind)
533 const char *end = line + strlen (line);
534 const char *cmdstart, *cmdend;
535 const char *valstart, *valend;
540 /* Skip leading and trailing whitespace. */
541 while (*line && ISSPACE (*line))
543 while (end > line && ISSPACE (end[-1]))
546 /* Skip empty lines and comments. */
547 if (!*line || *line == '#')
553 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
557 /* Skip '=', as well as any space before or after it. */
558 while (p < end && ISSPACE (*p))
560 if (p == end || *p != '=')
563 while (p < end && ISSPACE (*p))
569 /* The line now known to be syntactically correct. Check whether
570 the command is valid. */
571 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
573 ind = command_by_name (cmdcopy);
577 /* The command is valid. Now fill in the values and report success
580 *com = strdupdelim (cmdstart, cmdend);
581 *val = strdupdelim (valstart, valend);
585 /* Run commands[comind].action. */
588 setval_internal (int comind, const char *com, const char *val)
590 assert (0 <= comind && comind < countof (commands));
591 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
592 return ((*commands[comind].action) (com, val, commands[comind].closure));
595 /* Run command COM with value VAL. If running the command produces an
596 error, report the error and exit.
598 This is intended to be called from main() to modify Wget's behavior
599 through command-line switches. Since COM is hard-coded in main(),
600 it is not canonicalized, and this aborts when COM is not found.
602 If COMIND's are exported to init.h, this function will be changed
603 to accept COMIND directly. */
606 setoptval (const char *com, const char *val)
608 assert (val != NULL);
609 if (!setval_internal (command_by_name (com), com, val))
613 /* Parse OPT into command and value and run it. For example,
614 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
615 This is used by the `--execute' flag in main.c. */
618 run_command (const char *opt)
622 int status = parse_line (opt, &com, &val, &comind);
625 if (!setval_internal (comind, com, val))
630 else if (status == 0)
632 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
638 /* Generic helper functions, for use with `commands'. */
640 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
642 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
643 && TOLOWER((p)[1]) == (c1) \
646 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
647 && TOLOWER((p)[1]) == (c1) \
648 && TOLOWER((p)[2]) == (c2) \
652 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
653 except for error messages. */
655 cmd_boolean (const char *com, const char *val, void *closure)
659 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
660 /* "on", "yes" and "1" mean true. */
662 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
663 /* "off", "no" and "0" mean false. */
668 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
669 exec_name, com, val);
673 *(int *)closure = bool_value;
677 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
678 COM is ignored, except for error messages. Values 2 and -1
679 indicate that once defined, the value may not be changed by
680 successive wgetrc files or command-line arguments.
682 Values: 2 - Enable a particular option for good ("always")
683 1 - Enable an option ("on")
684 0 - Disable an option ("off")
685 -1 - Disable an option for good ("never") */
687 cmd_lockable_boolean (const char *com, const char *val, void *closure)
689 int lockable_boolean_value;
691 int oldval = *(int *)closure;
694 * If a config file said "always" or "never", don't allow command line
695 * arguments to override the config file.
697 if (oldval == -1 || oldval == 2)
700 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
701 lockable_boolean_value = 2;
702 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
703 lockable_boolean_value = 1;
704 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
705 lockable_boolean_value = 0;
706 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
707 lockable_boolean_value = -1;
711 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
712 exec_name, com, val);
716 *(int *)closure = lockable_boolean_value;
720 static int simple_atoi PARAMS ((const char *, const char *, int *));
722 /* Set the non-negative integer value from VAL to CLOSURE. With
723 incorrect specification, the number remains unchanged. */
725 cmd_number (const char *com, const char *val, void *closure)
727 if (!simple_atoi (val, val + strlen (val), closure))
729 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
730 exec_name, com, val);
736 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
738 cmd_number_inf (const char *com, const char *val, void *closure)
740 if (!strcasecmp (val, "inf"))
745 return cmd_number (com, val, closure);
748 /* Copy (strdup) the string at COM to a new location and place a
749 pointer to *CLOSURE. */
751 cmd_string (const char *com, const char *val, void *closure)
753 char **pstring = (char **)closure;
755 xfree_null (*pstring);
756 *pstring = xstrdup (val);
761 # define ISSEP(c) ((c) == '/')
763 # define ISSEP(c) ((c) == '/' || (c) == '\\')
766 /* Like the above, but handles tilde-expansion when reading a user's
767 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
768 gets expanded to the user's home directory. */
770 cmd_file (const char *com, const char *val, void *closure)
772 char **pstring = (char **)closure;
774 xfree_null (*pstring);
776 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
778 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
781 *pstring = xstrdup (val);
787 char *home = home_dir ();
791 homelen = strlen (home);
792 while (homelen && ISSEP (home[homelen - 1]))
793 home[--homelen] = '\0';
795 /* Skip the leading "~/". */
796 for (++val; ISSEP (*val); val++)
799 result = xmalloc (homelen + 1 + strlen (val) + 1);
800 memcpy (result, home, homelen);
801 result[homelen] = '/';
802 strcpy (result + homelen + 1, val);
808 /* Convert "\" to "/". */
811 for (s = *pstring; *s; s++)
819 /* Like cmd_file, but strips trailing '/' characters. */
821 cmd_directory (const char *com, const char *val, void *closure)
825 /* Call cmd_file() for tilde expansion and separator
826 canonicalization (backslash -> slash under Windows). These
827 things should perhaps be in a separate function. */
828 if (!cmd_file (com, val, closure))
831 s = *(char **)closure;
833 while (t > s && *--t == '/')
839 /* Split VAL by space to a vector of values, and append those values
840 to vector pointed to by the CLOSURE argument. If VAL is empty, the
841 CLOSURE vector is cleared instead. */
844 cmd_vector (const char *com, const char *val, void *closure)
846 char ***pvec = (char ***)closure;
849 *pvec = merge_vecs (*pvec, sepstring (val));
859 cmd_directory_vector (const char *com, const char *val, void *closure)
861 char ***pvec = (char ***)closure;
865 /* Strip the trailing slashes from directories. */
868 seps = sepstring (val);
869 for (t = seps; t && *t; t++)
871 int len = strlen (*t);
872 /* Skip degenerate case of root directory. */
875 if ((*t)[len - 1] == '/')
876 (*t)[len - 1] = '\0';
879 *pvec = merge_vecs (*pvec, seps);
889 static int simple_atof PARAMS ((const char *, const char *, double *));
891 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
892 as "100k" or "2.5G" to a floating point number. */
895 parse_bytes_helper (const char *val, double *result)
898 const char *end = val + strlen (val);
900 /* Check for "inf". */
901 if (0 == strcmp (val, "inf"))
907 /* Strip trailing whitespace. */
908 while (val < end && ISSPACE (end[-1]))
913 switch (TOLOWER (end[-1]))
916 --end, mult = 1024.0;
919 --end, mult = 1048576.0;
922 --end, mult = 1073741824.0;
925 --end, mult = 1099511627776.0;
928 /* Not a recognized suffix: assume it's a digit. (If not,
929 simple_atof will raise an error.) */
933 /* Skip leading and trailing whitespace. */
934 while (val < end && ISSPACE (*val))
936 while (val < end && ISSPACE (end[-1]))
941 if (!simple_atof (val, end, &number))
944 *result = number * mult;
948 /* Parse VAL as a number and set its value to CLOSURE (which should
949 point to a long int).
951 By default, the value is assumed to be in bytes. If "K", "M", or
952 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
953 1<<30, respectively. Floating point values are allowed and are
954 cast to integer before use. The idea is to be able to use things
955 like 1.5k instead of "1536".
957 The string "inf" is returned as 0.
959 In case of error, 0 is returned and memory pointed to by CLOSURE
960 remains unmodified. */
963 cmd_bytes (const char *com, const char *val, void *closure)
966 if (!parse_bytes_helper (val, &byte_value))
968 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
969 exec_name, com, val);
972 *(long *)closure = (long)byte_value;
976 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
977 LARGE_INT. It works by converting the string to double, therefore
978 working with values up to 2^53-1 without loss of precision. This
979 value (8192 TB) is large enough to serve for a while. */
982 cmd_bytes_large (const char *com, const char *val, void *closure)
985 if (!parse_bytes_helper (val, &byte_value))
987 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
988 exec_name, com, val);
991 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
995 /* Store the value of VAL to *OUT. The value is a time period, by
996 default expressed in seconds, but also accepting suffixes "m", "h",
997 "d", and "w" for minutes, hours, days, and weeks respectively. */
1000 cmd_time (const char *com, const char *val, void *closure)
1002 double number, mult;
1003 const char *end = val + strlen (val);
1005 /* Strip trailing whitespace. */
1006 while (val < end && ISSPACE (end[-1]))
1012 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1013 exec_name, com, val);
1017 switch (TOLOWER (end[-1]))
1020 --end, mult = 1; /* seconds */
1023 --end, mult = 60; /* minutes */
1026 --end, mult = 3600; /* hours */
1029 --end, mult = 86400.0; /* days */
1032 --end, mult = 604800.0; /* weeks */
1035 /* Not a recognized suffix: assume it belongs to the number.
1036 (If not, atof simple_atof will raise an error.) */
1040 /* Skip leading and trailing whitespace. */
1041 while (val < end && ISSPACE (*val))
1043 while (val < end && ISSPACE (end[-1]))
1048 if (!simple_atof (val, end, &number))
1051 *(double *)closure = number * mult;
1055 /* Specialized helper functions, used by `commands' to handle some
1056 options specially. */
1058 static int check_user_specified_header PARAMS ((const char *));
1061 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1063 if (!cmd_boolean (com, val, &opt.dirstruct))
1065 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1066 must be affected inversely. */
1068 opt.no_dirstruct = 0;
1070 opt.no_dirstruct = 1;
1075 cmd_spec_header (const char *com, const char *val, void *closure)
1079 /* Empty header means reset headers. */
1080 xfree_null (opt.user_header);
1081 opt.user_header = NULL;
1087 if (!check_user_specified_header (val))
1089 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1090 exec_name, com, val);
1093 i = opt.user_header ? strlen (opt.user_header) : 0;
1094 opt.user_header = (char *)xrealloc (opt.user_header, i + strlen (val)
1096 strcpy (opt.user_header + i, val);
1098 opt.user_header[i++] = '\r';
1099 opt.user_header[i++] = '\n';
1100 opt.user_header[i] = '\0';
1106 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1108 int flag = cmd_boolean (com, val, &opt.htmlify);
1109 if (flag && !opt.htmlify)
1110 opt.remove_listing = 0;
1114 /* Set the "mirror" mode. It means: recursive download, timestamping,
1115 no limit on max. recursion depth, and don't remove listings. */
1118 cmd_spec_mirror (const char *com, const char *val, void *closure)
1122 if (!cmd_boolean (com, val, &mirror))
1127 if (!opt.no_dirstruct)
1129 opt.timestamping = 1;
1130 opt.reclevel = INFINITE_RECURSION;
1131 opt.remove_listing = 0;
1136 /* Set progress.type to VAL, but verify that it's a valid progress
1137 implementation before that. */
1140 cmd_spec_progress (const char *com, const char *val, void *closure)
1142 if (!valid_progress_implementation_p (val))
1144 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1145 exec_name, com, val);
1148 xfree_null (opt.progress_type);
1150 /* Don't call set_progress_implementation here. It will be called
1151 in main() when it becomes clear what the log output is. */
1152 opt.progress_type = xstrdup (val);
1156 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1157 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1161 cmd_spec_recursive (const char *com, const char *val, void *closure)
1163 if (!cmd_boolean (com, val, &opt.recursive))
1167 if (opt.recursive && !opt.no_dirstruct)
1174 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1176 int restrict_os = opt.restrict_files_os;
1177 int restrict_ctrl = opt.restrict_files_ctrl;
1179 const char *end = strchr (val, ',');
1181 end = val + strlen (val);
1183 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1185 if (VAL_IS ("unix"))
1186 restrict_os = restrict_unix;
1187 else if (VAL_IS ("windows"))
1188 restrict_os = restrict_windows;
1189 else if (VAL_IS ("nocontrol"))
1195 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1196 exec_name, com, val);
1204 if (!strcmp (end + 1, "nocontrol"))
1210 opt.restrict_files_os = restrict_os;
1211 opt.restrict_files_ctrl = restrict_ctrl;
1215 /* Set all three timeout values. */
1218 cmd_spec_timeout (const char *com, const char *val, void *closure)
1221 if (!cmd_time (com, val, &value))
1223 opt.read_timeout = value;
1224 opt.connect_timeout = value;
1225 opt.dns_timeout = value;
1230 cmd_spec_useragent (const char *com, const char *val, void *closure)
1232 /* Just check for empty string and newline, so we don't throw total
1233 junk to the server. */
1234 if (!*val || strchr (val, '\n'))
1236 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1237 exec_name, com, val);
1240 opt.useragent = xstrdup (val);
1244 /* Miscellaneous useful routines. */
1246 /* A very simple atoi clone, more portable than strtol and friends,
1247 but reports errors, unlike atoi. Returns 1 on success, 0 on
1248 failure. In case of success, stores result to *DEST. */
1251 simple_atoi (const char *beg, const char *end, int *dest)
1259 for (p = beg; p < end && ISDIGIT (*p); p++)
1260 result = (10 * result) + (*p - '0');
1269 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1270 doesn't handle exponential notation. Returns 1 on success, 0 on
1271 failure. In case of success, stores its result to *DEST. */
1274 simple_atof (const char *beg, const char *end, double *dest)
1284 for (p = beg; p < end; p++)
1290 result = (10 * result) + (ch - '0');
1292 result += (ch - '0') / (divider *= 10);
1313 check_user_specified_header (const char *s)
1317 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1318 /* The header MUST contain `:' preceded by at least one
1319 non-whitespace character. */
1320 if (*p != ':' || p == s)
1322 /* The header MUST NOT contain newlines. */
1323 if (strchr (s, '\n'))
1328 void cleanup_html_url PARAMS ((void));
1329 void res_cleanup PARAMS ((void));
1330 void downloaded_files_free PARAMS ((void));
1331 void http_cleanup PARAMS ((void));
1334 /* Free the memory allocated by global variables. */
1338 /* Free external resources, close files, etc. */
1343 /* We're exiting anyway so there's no real need to call free()
1344 hundreds of times. Skipping the frees will make Wget exit
1347 However, when detecting leaks, it's crucial to free() everything
1348 because then you can find the real leaks, i.e. the allocated
1349 memory which grows with the size of the program. */
1355 cleanup_html_url ();
1356 downloaded_files_free ();
1358 if (wget_cookie_jar)
1359 cookie_jar_delete (wget_cookie_jar);
1362 extern acc_t *netrc_list;
1363 free_netrc (netrc_list);
1365 xfree_null (opt.lfilename);
1366 xfree_null (opt.dir_prefix);
1367 xfree_null (opt.input_filename);
1368 xfree_null (opt.output_document);
1369 free_vec (opt.accepts);
1370 free_vec (opt.rejects);
1371 free_vec (opt.excludes);
1372 free_vec (opt.includes);
1373 free_vec (opt.domains);
1374 free_vec (opt.follow_tags);
1375 free_vec (opt.ignore_tags);
1376 xfree_null (opt.progress_type);
1377 xfree (opt.ftp_acc);
1378 xfree_null (opt.ftp_pass);
1379 xfree_null (opt.ftp_proxy);
1380 xfree_null (opt.https_proxy);
1381 xfree_null (opt.http_proxy);
1382 free_vec (opt.no_proxy);
1383 xfree_null (opt.useragent);
1384 xfree_null (opt.referer);
1385 xfree_null (opt.http_user);
1386 xfree_null (opt.http_passwd);
1387 xfree_null (opt.user_header);
1389 xfree_null (opt.sslcertkey);
1390 xfree_null (opt.sslcertfile);
1391 #endif /* HAVE_SSL */
1392 xfree_null (opt.bind_address);
1393 xfree_null (opt.cookies_input);
1394 xfree_null (opt.cookies_output);