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 },
169 { "input", &opt.input_filename, cmd_file },
170 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
171 { "killlonger", &opt.kill_longer, cmd_boolean },
172 { "limitrate", &opt.limit_rate, cmd_bytes },
173 { "loadcookies", &opt.cookies_input, cmd_file },
174 { "logfile", &opt.lfilename, cmd_file },
175 { "login", &opt.ftp_acc, cmd_string },
176 { "mirror", NULL, cmd_spec_mirror },
177 { "netrc", &opt.netrc, cmd_boolean },
178 { "noclobber", &opt.noclobber, cmd_boolean },
179 { "noparent", &opt.no_parent, cmd_boolean },
180 { "noproxy", &opt.no_proxy, cmd_vector },
181 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
182 { "outputdocument", &opt.output_document, cmd_file },
183 { "pagerequisites", &opt.page_requisites, cmd_boolean },
184 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
185 { "passwd", &opt.ftp_pass, cmd_string },
186 { "postdata", &opt.post_data, cmd_string },
187 { "postfile", &opt.post_file_name, cmd_file },
188 { "progress", &opt.progress_type, cmd_spec_progress },
189 { "proxypasswd", &opt.proxy_passwd, cmd_string },
190 { "proxyuser", &opt.proxy_user, cmd_string },
191 { "quiet", &opt.quiet, cmd_boolean },
192 { "quota", &opt.quota, cmd_bytes_large },
193 { "randomwait", &opt.random_wait, cmd_boolean },
194 { "readtimeout", &opt.read_timeout, cmd_time },
195 { "reclevel", &opt.reclevel, cmd_number_inf },
196 { "recursive", NULL, cmd_spec_recursive },
197 { "referer", &opt.referer, cmd_string },
198 { "reject", &opt.rejects, cmd_vector },
199 { "relativeonly", &opt.relative_only, cmd_boolean },
200 { "removelisting", &opt.remove_listing, cmd_boolean },
201 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
202 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
203 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
204 { "robots", &opt.use_robots, cmd_boolean },
205 { "savecookies", &opt.cookies_output, cmd_file },
206 { "saveheaders", &opt.save_headers, cmd_boolean },
207 { "serverresponse", &opt.server_response, cmd_boolean },
208 { "spanhosts", &opt.spanhost, cmd_boolean },
209 { "spider", &opt.spider, cmd_boolean },
211 { "sslcadir", &opt.sslcadir, cmd_directory },
212 { "sslcafile", &opt.sslcafile, cmd_file },
213 { "sslcertfile", &opt.sslcertfile, cmd_file },
214 { "sslcertkey", &opt.sslcertkey, cmd_file },
215 { "sslcerttype", &opt.sslcerttype, cmd_number },
216 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
217 { "sslprotocol", &opt.sslprotocol, cmd_number },
218 #endif /* HAVE_SSL */
219 { "strictcomments", &opt.strict_comments, cmd_boolean },
220 { "timeout", NULL, cmd_spec_timeout },
221 { "timestamping", &opt.timestamping, cmd_boolean },
222 { "tries", &opt.ntry, cmd_number_inf },
223 { "useproxy", &opt.use_proxy, cmd_boolean },
224 { "useragent", NULL, cmd_spec_useragent },
225 { "verbose", &opt.verbose, cmd_boolean },
226 { "wait", &opt.wait, cmd_time },
227 { "waitretry", &opt.waitretry, cmd_time }
230 /* Look up CMDNAME in the commands[] and return its position in the
231 array. If CMDNAME is not found, return -1. */
234 command_by_name (const char *cmdname)
236 /* Use binary search for speed. Wget has ~100 commands, which
237 guarantees a worst case performance of 7 string comparisons. */
238 int lo = 0, hi = countof (commands) - 1;
242 int mid = (lo + hi) >> 1;
243 int cmp = strcasecmp (cmdname, commands[mid].name);
254 /* Reset the variables to default values. */
260 /* Most of the default values are 0. Just reset everything, and
261 fill in the non-zero values. Note that initializing pointers to
262 NULL this way is technically illegal, but porting Wget to a
263 machine where NULL is not all-zero bit pattern will be the least
264 of the implementors' worries. */
272 opt.ftp_acc = xstrdup ("anonymous");
273 opt.ftp_pass = xstrdup ("-wget@");
277 opt.http_keep_alive = 1;
279 tmp = getenv ("no_proxy");
281 opt.no_proxy = sepstring (tmp);
284 opt.read_timeout = 900;
287 opt.remove_listing = 1;
289 opt.dot_bytes = 1024;
290 opt.dot_spacing = 10;
291 opt.dots_in_line = 50;
295 /* The default for file name restriction defaults to the OS type. */
296 #if !defined(WINDOWS) && !defined(__CYGWIN__)
297 opt.restrict_files_os = restrict_unix;
299 opt.restrict_files_os = restrict_windows;
301 opt.restrict_files_ctrl = 1;
304 /* Return the user's home directory (strdup-ed), or NULL if none is
309 char *home = getenv ("HOME");
314 /* If HOME is not defined, try getting it from the password
316 struct passwd *pwd = getpwuid (getuid ());
317 if (!pwd || !pwd->pw_dir)
322 /* #### Maybe I should grab home_dir from registry, but the best
323 that I could get from there is user's Start menu. It sucks! */
327 return home ? xstrdup (home) : NULL;
330 /* Return the path to the user's .wgetrc. This is either the value of
331 `WGETRC' environment variable, or `$HOME/.wgetrc'.
333 If the `WGETRC' variable exists but the file does not exist, the
334 function will exit(). */
336 wgetrc_file_name (void)
341 /* Try the environment. */
342 env = getenv ("WGETRC");
345 if (!file_exists_p (env))
347 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
351 return xstrdup (env);
355 /* If that failed, try $HOME/.wgetrc. */
359 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
360 sprintf (file, "%s/.wgetrc", home);
364 /* Under Windows, "home" is (for the purposes of this function) the
365 directory where `wget.exe' resides, and `wget.ini' will be used
366 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
368 It is not as trivial as I assumed, because on 95 argv[0] is full
369 path, but on NT you get what you typed in command line. --dbudor */
373 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
374 sprintf (file, "%swget.ini", home);
380 if (!file_exists_p (file))
388 static int parse_line PARAMS ((const char *, char **, char **, int *));
389 static int setval_internal PARAMS ((int, const char *, const char *));
391 /* Initialize variables from a wgetrc file. */
394 run_wgetrc (const char *file)
400 fp = fopen (file, "rb");
403 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
404 file, strerror (errno));
407 enable_tilde_expansion = 1;
409 while ((line = read_whole_line (fp)))
414 /* Parse the line. */
415 status = parse_line (line, &com, &val, &comind);
417 /* If everything is OK, set the value. */
420 if (!setval_internal (comind, com, val))
421 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
426 else if (status == 0)
427 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
431 enable_tilde_expansion = 0;
435 /* Initialize the defaults and run the system wgetrc and user's own
442 /* Load the hard-coded defaults. */
445 /* If SYSTEM_WGETRC is defined, use it. */
447 if (file_exists_p (SYSTEM_WGETRC))
448 run_wgetrc (SYSTEM_WGETRC);
450 /* Override it with your own, if one exists. */
451 file = wgetrc_file_name ();
454 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
455 something like realpath() before comparing them with `strcmp' */
457 if (!strcmp (file, SYSTEM_WGETRC))
459 fprintf (stderr, _("\
460 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
470 /* Remove dashes and underscores from S, modifying S in the
476 char *t = s; /* t - tortoise */
477 char *h = s; /* h - hare */
479 if (*h == '_' || *h == '-')
486 /* Parse the line pointed by line, with the syntax:
487 <sp>* command <sp>* = <sp>* value <sp>*
488 Uses malloc to allocate space for command and value.
489 If the line is invalid, data is freed and 0 is returned.
496 In case of success, *COM and *VAL point to freshly allocated
497 strings, and *COMIND points to com's index. In case of error or
498 empty line, those values are unaffected. */
501 parse_line (const char *line, char **com, char **val, int *comind)
504 const char *end = line + strlen (line);
505 const char *cmdstart, *cmdend;
506 const char *valstart, *valend;
511 /* Skip leading and trailing whitespace. */
512 while (*line && ISSPACE (*line))
514 while (end > line && ISSPACE (end[-1]))
517 /* Skip empty lines and comments. */
518 if (!*line || *line == '#')
524 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
528 /* Skip '=', as well as any space before or after it. */
529 while (p < end && ISSPACE (*p))
531 if (p == end || *p != '=')
534 while (p < end && ISSPACE (*p))
540 /* The line now known to be syntactically correct. Check whether
541 the command is valid. */
542 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
544 ind = command_by_name (cmdcopy);
548 /* The command is valid. Now fill in the values and report success
551 *com = strdupdelim (cmdstart, cmdend);
552 *val = strdupdelim (valstart, valend);
556 /* Run commands[comind].action. */
559 setval_internal (int comind, const char *com, const char *val)
561 assert (0 <= comind && comind < countof (commands));
562 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
563 return ((*commands[comind].action) (com, val, commands[comind].closure));
566 /* Run command COM with value VAL. If running the command produces an
567 error, report the error and exit.
569 This is intended to be called from main() to modify Wget's behavior
570 through command-line switches. Since COM is hard-coded in main(),
571 it is not canonicalized, and this aborts when COM is not found.
573 If COMIND's are exported to init.h, this function will be changed
574 to accept COMIND directly. */
577 setoptval (const char *com, const char *val)
579 assert (val != NULL);
580 if (!setval_internal (command_by_name (com), com, val))
584 /* Parse OPT into command and value and run it. For example,
585 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
586 This is used by the `--execute' flag in main.c. */
589 run_command (const char *opt)
593 int status = parse_line (opt, &com, &val, &comind);
596 if (!setval_internal (comind, com, val))
601 else if (status == 0)
603 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
609 /* Generic helper functions, for use with `commands'. */
611 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
613 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
614 && TOLOWER((p)[1]) == (c1) \
617 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
618 && TOLOWER((p)[1]) == (c1) \
619 && TOLOWER((p)[2]) == (c2) \
623 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
624 except for error messages. */
626 cmd_boolean (const char *com, const char *val, void *closure)
630 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
631 /* "on", "yes" and "1" mean true. */
633 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
634 /* "off", "no" and "0" mean false. */
639 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
640 exec_name, com, val);
644 *(int *)closure = bool_value;
648 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
649 COM is ignored, except for error messages. Values 2 and -1
650 indicate that once defined, the value may not be changed by
651 successive wgetrc files or command-line arguments.
653 Values: 2 - Enable a particular option for good ("always")
654 1 - Enable an option ("on")
655 0 - Disable an option ("off")
656 -1 - Disable an option for good ("never") */
658 cmd_lockable_boolean (const char *com, const char *val, void *closure)
660 int lockable_boolean_value;
662 int oldval = *(int *)closure;
665 * If a config file said "always" or "never", don't allow command line
666 * arguments to override the config file.
668 if (oldval == -1 || oldval == 2)
671 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
672 lockable_boolean_value = 2;
673 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
674 lockable_boolean_value = 1;
675 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
676 lockable_boolean_value = 0;
677 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
678 lockable_boolean_value = -1;
682 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
683 exec_name, com, val);
687 *(int *)closure = lockable_boolean_value;
691 static int simple_atoi PARAMS ((const char *, const char *, int *));
693 /* Set the non-negative integer value from VAL to CLOSURE. With
694 incorrect specification, the number remains unchanged. */
696 cmd_number (const char *com, const char *val, void *closure)
698 if (!simple_atoi (val, val + strlen (val), closure))
700 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
701 exec_name, com, val);
707 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
709 cmd_number_inf (const char *com, const char *val, void *closure)
711 if (!strcasecmp (val, "inf"))
716 return cmd_number (com, val, closure);
719 /* Copy (strdup) the string at COM to a new location and place a
720 pointer to *CLOSURE. */
722 cmd_string (const char *com, const char *val, void *closure)
724 char **pstring = (char **)closure;
726 xfree_null (*pstring);
727 *pstring = xstrdup (val);
732 # define ISSEP(c) ((c) == '/')
734 # define ISSEP(c) ((c) == '/' || (c) == '\\')
737 /* Like the above, but handles tilde-expansion when reading a user's
738 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
739 gets expanded to the user's home directory. */
741 cmd_file (const char *com, const char *val, void *closure)
743 char **pstring = (char **)closure;
745 xfree_null (*pstring);
747 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
749 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
752 *pstring = xstrdup (val);
758 char *home = home_dir ();
762 homelen = strlen (home);
763 while (homelen && ISSEP (home[homelen - 1]))
764 home[--homelen] = '\0';
766 /* Skip the leading "~/". */
767 for (++val; ISSEP (*val); val++)
770 result = xmalloc (homelen + 1 + strlen (val) + 1);
771 memcpy (result, home, homelen);
772 result[homelen] = '/';
773 strcpy (result + homelen + 1, val);
779 /* Convert "\" to "/". */
782 for (s = *pstring; *s; s++)
790 /* Like cmd_file, but strips trailing '/' characters. */
792 cmd_directory (const char *com, const char *val, void *closure)
796 /* Call cmd_file() for tilde expansion and separator
797 canonicalization (backslash -> slash under Windows). These
798 things should perhaps be in a separate function. */
799 if (!cmd_file (com, val, closure))
802 s = *(char **)closure;
804 while (t > s && *--t == '/')
810 /* Split VAL by space to a vector of values, and append those values
811 to vector pointed to by the CLOSURE argument. If VAL is empty, the
812 CLOSURE vector is cleared instead. */
815 cmd_vector (const char *com, const char *val, void *closure)
817 char ***pvec = (char ***)closure;
820 *pvec = merge_vecs (*pvec, sepstring (val));
830 cmd_directory_vector (const char *com, const char *val, void *closure)
832 char ***pvec = (char ***)closure;
836 /* Strip the trailing slashes from directories. */
839 seps = sepstring (val);
840 for (t = seps; t && *t; t++)
842 int len = strlen (*t);
843 /* Skip degenerate case of root directory. */
846 if ((*t)[len - 1] == '/')
847 (*t)[len - 1] = '\0';
850 *pvec = merge_vecs (*pvec, seps);
860 static int simple_atof PARAMS ((const char *, const char *, double *));
862 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
863 as "100k" or "2.5G" to a floating point number. */
866 parse_bytes_helper (const char *val, double *result)
869 const char *end = val + strlen (val);
871 /* Check for "inf". */
872 if (0 == strcmp (val, "inf"))
878 /* Strip trailing whitespace. */
879 while (val < end && ISSPACE (end[-1]))
884 switch (TOLOWER (end[-1]))
887 --end, mult = 1024.0;
890 --end, mult = 1048576.0;
893 --end, mult = 1073741824.0;
896 --end, mult = 1099511627776.0;
899 /* Not a recognized suffix: assume it's a digit. (If not,
900 simple_atof will raise an error.) */
904 /* Skip leading and trailing whitespace. */
905 while (val < end && ISSPACE (*val))
907 while (val < end && ISSPACE (end[-1]))
912 if (!simple_atof (val, end, &number))
915 *result = number * mult;
919 /* Parse VAL as a number and set its value to CLOSURE (which should
920 point to a long int).
922 By default, the value is assumed to be in bytes. If "K", "M", or
923 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
924 1<<30, respectively. Floating point values are allowed and are
925 cast to integer before use. The idea is to be able to use things
926 like 1.5k instead of "1536".
928 The string "inf" is returned as 0.
930 In case of error, 0 is returned and memory pointed to by CLOSURE
931 remains unmodified. */
934 cmd_bytes (const char *com, const char *val, void *closure)
937 if (!parse_bytes_helper (val, &byte_value))
939 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
940 exec_name, com, val);
943 *(long *)closure = (long)byte_value;
947 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
948 LARGE_INT. It works by converting the string to double, therefore
949 working with values up to 2^53-1 without loss of precision. This
950 value (8192 TB) is large enough to serve for a while. */
953 cmd_bytes_large (const char *com, const char *val, void *closure)
956 if (!parse_bytes_helper (val, &byte_value))
958 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
959 exec_name, com, val);
962 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
966 /* Store the value of VAL to *OUT. The value is a time period, by
967 default expressed in seconds, but also accepting suffixes "m", "h",
968 "d", and "w" for minutes, hours, days, and weeks respectively. */
971 cmd_time (const char *com, const char *val, void *closure)
974 const char *end = val + strlen (val);
976 /* Strip trailing whitespace. */
977 while (val < end && ISSPACE (end[-1]))
983 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
984 exec_name, com, val);
988 switch (TOLOWER (end[-1]))
991 --end, mult = 1; /* seconds */
994 --end, mult = 60; /* minutes */
997 --end, mult = 3600; /* hours */
1000 --end, mult = 86400.0; /* days */
1003 --end, mult = 604800.0; /* weeks */
1006 /* Not a recognized suffix: assume it belongs to the number.
1007 (If not, atof simple_atof will raise an error.) */
1011 /* Skip leading and trailing whitespace. */
1012 while (val < end && ISSPACE (*val))
1014 while (val < end && ISSPACE (end[-1]))
1019 if (!simple_atof (val, end, &number))
1022 *(double *)closure = number * mult;
1026 /* Specialized helper functions, used by `commands' to handle some
1027 options specially. */
1029 static int check_user_specified_header PARAMS ((const char *));
1032 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1034 if (!cmd_boolean (com, val, &opt.dirstruct))
1036 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1037 must be affected inversely. */
1039 opt.no_dirstruct = 0;
1041 opt.no_dirstruct = 1;
1046 cmd_spec_header (const char *com, const char *val, void *closure)
1050 /* Empty header means reset headers. */
1051 xfree_null (opt.user_header);
1052 opt.user_header = NULL;
1058 if (!check_user_specified_header (val))
1060 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1061 exec_name, com, val);
1064 i = opt.user_header ? strlen (opt.user_header) : 0;
1065 opt.user_header = (char *)xrealloc (opt.user_header, i + strlen (val)
1067 strcpy (opt.user_header + i, val);
1069 opt.user_header[i++] = '\r';
1070 opt.user_header[i++] = '\n';
1071 opt.user_header[i] = '\0';
1077 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1079 int flag = cmd_boolean (com, val, &opt.htmlify);
1080 if (flag && !opt.htmlify)
1081 opt.remove_listing = 0;
1085 /* Set the "mirror" mode. It means: recursive download, timestamping,
1086 no limit on max. recursion depth, and don't remove listings. */
1089 cmd_spec_mirror (const char *com, const char *val, void *closure)
1093 if (!cmd_boolean (com, val, &mirror))
1098 if (!opt.no_dirstruct)
1100 opt.timestamping = 1;
1101 opt.reclevel = INFINITE_RECURSION;
1102 opt.remove_listing = 0;
1107 /* Set progress.type to VAL, but verify that it's a valid progress
1108 implementation before that. */
1111 cmd_spec_progress (const char *com, const char *val, void *closure)
1113 if (!valid_progress_implementation_p (val))
1115 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1116 exec_name, com, val);
1119 xfree_null (opt.progress_type);
1121 /* Don't call set_progress_implementation here. It will be called
1122 in main() when it becomes clear what the log output is. */
1123 opt.progress_type = xstrdup (val);
1127 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1128 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1132 cmd_spec_recursive (const char *com, const char *val, void *closure)
1134 if (!cmd_boolean (com, val, &opt.recursive))
1138 if (opt.recursive && !opt.no_dirstruct)
1145 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1147 int restrict_os = opt.restrict_files_os;
1148 int restrict_ctrl = opt.restrict_files_ctrl;
1150 const char *end = strchr (val, ',');
1152 end = val + strlen (val);
1154 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1156 if (VAL_IS ("unix"))
1157 restrict_os = restrict_unix;
1158 else if (VAL_IS ("windows"))
1159 restrict_os = restrict_windows;
1160 else if (VAL_IS ("nocontrol"))
1166 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1167 exec_name, com, val);
1175 if (!strcmp (end + 1, "nocontrol"))
1181 opt.restrict_files_os = restrict_os;
1182 opt.restrict_files_ctrl = restrict_ctrl;
1186 /* Set all three timeout values. */
1189 cmd_spec_timeout (const char *com, const char *val, void *closure)
1192 if (!cmd_time (com, val, &value))
1194 opt.read_timeout = value;
1195 opt.connect_timeout = value;
1196 opt.dns_timeout = value;
1201 cmd_spec_useragent (const char *com, const char *val, void *closure)
1203 /* Just check for empty string and newline, so we don't throw total
1204 junk to the server. */
1205 if (!*val || strchr (val, '\n'))
1207 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1208 exec_name, com, val);
1211 opt.useragent = xstrdup (val);
1215 /* Miscellaneous useful routines. */
1217 /* A very simple atoi clone, more portable than strtol and friends,
1218 but reports errors, unlike atoi. Returns 1 on success, 0 on
1219 failure. In case of success, stores result to *DEST. */
1222 simple_atoi (const char *beg, const char *end, int *dest)
1230 for (p = beg; p < end && ISDIGIT (*p); p++)
1231 result = (10 * result) + (*p - '0');
1240 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1241 doesn't handle exponential notation. Returns 1 on success, 0 on
1242 failure. In case of success, stores its result to *DEST. */
1245 simple_atof (const char *beg, const char *end, double *dest)
1255 for (p = beg; p < end; p++)
1261 result = (10 * result) + (ch - '0');
1263 result += (ch - '0') / (divider *= 10);
1284 check_user_specified_header (const char *s)
1288 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1289 /* The header MUST contain `:' preceded by at least one
1290 non-whitespace character. */
1291 if (*p != ':' || p == s)
1293 /* The header MUST NOT contain newlines. */
1294 if (strchr (s, '\n'))
1299 void cleanup_html_url PARAMS ((void));
1300 void res_cleanup PARAMS ((void));
1301 void downloaded_files_free PARAMS ((void));
1302 void http_cleanup PARAMS ((void));
1305 /* Free the memory allocated by global variables. */
1309 /* Free external resources, close files, etc. */
1314 /* We're exiting anyway so there's no real need to call free()
1315 hundreds of times. Skipping the frees will make Wget exit
1318 However, when detecting leaks, it's crucial to free() everything
1319 because then you can find the real leaks, i.e. the allocated
1320 memory which grows with the size of the program. */
1326 cleanup_html_url ();
1327 downloaded_files_free ();
1329 if (wget_cookie_jar)
1330 cookie_jar_delete (wget_cookie_jar);
1333 extern acc_t *netrc_list;
1334 free_netrc (netrc_list);
1336 xfree_null (opt.lfilename);
1337 xfree_null (opt.dir_prefix);
1338 xfree_null (opt.input_filename);
1339 xfree_null (opt.output_document);
1340 free_vec (opt.accepts);
1341 free_vec (opt.rejects);
1342 free_vec (opt.excludes);
1343 free_vec (opt.includes);
1344 free_vec (opt.domains);
1345 free_vec (opt.follow_tags);
1346 free_vec (opt.ignore_tags);
1347 xfree_null (opt.progress_type);
1348 xfree (opt.ftp_acc);
1349 xfree_null (opt.ftp_pass);
1350 xfree_null (opt.ftp_proxy);
1351 xfree_null (opt.https_proxy);
1352 xfree_null (opt.http_proxy);
1353 free_vec (opt.no_proxy);
1354 xfree_null (opt.useragent);
1355 xfree_null (opt.referer);
1356 xfree_null (opt.http_user);
1357 xfree_null (opt.http_passwd);
1358 xfree_null (opt.user_header);
1360 xfree_null (opt.sslcertkey);
1361 xfree_null (opt.sslcertfile);
1362 #endif /* HAVE_SSL */
1363 xfree_null (opt.bind_address);
1364 xfree_null (opt.cookies_input);
1365 xfree_null (opt.cookies_output);