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 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
189 { "progress", &opt.progress_type, cmd_spec_progress },
190 { "proxypasswd", &opt.proxy_passwd, cmd_string },
191 { "proxyuser", &opt.proxy_user, cmd_string },
192 { "quiet", &opt.quiet, cmd_boolean },
193 { "quota", &opt.quota, cmd_bytes_large },
194 { "randomwait", &opt.random_wait, cmd_boolean },
195 { "readtimeout", &opt.read_timeout, cmd_time },
196 { "reclevel", &opt.reclevel, cmd_number_inf },
197 { "recursive", NULL, cmd_spec_recursive },
198 { "referer", &opt.referer, cmd_string },
199 { "reject", &opt.rejects, cmd_vector },
200 { "relativeonly", &opt.relative_only, cmd_boolean },
201 { "removelisting", &opt.remove_listing, cmd_boolean },
202 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
203 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
204 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
205 { "robots", &opt.use_robots, cmd_boolean },
206 { "savecookies", &opt.cookies_output, cmd_file },
207 { "saveheaders", &opt.save_headers, cmd_boolean },
208 { "serverresponse", &opt.server_response, cmd_boolean },
209 { "spanhosts", &opt.spanhost, cmd_boolean },
210 { "spider", &opt.spider, cmd_boolean },
212 { "sslcadir", &opt.sslcadir, cmd_directory },
213 { "sslcafile", &opt.sslcafile, cmd_file },
214 { "sslcertfile", &opt.sslcertfile, cmd_file },
215 { "sslcertkey", &opt.sslcertkey, cmd_file },
216 { "sslcerttype", &opt.sslcerttype, cmd_number },
217 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
218 { "sslprotocol", &opt.sslprotocol, cmd_number },
219 #endif /* HAVE_SSL */
220 { "strictcomments", &opt.strict_comments, cmd_boolean },
221 { "timeout", NULL, cmd_spec_timeout },
222 { "timestamping", &opt.timestamping, cmd_boolean },
223 { "tries", &opt.ntry, cmd_number_inf },
224 { "useproxy", &opt.use_proxy, cmd_boolean },
225 { "useragent", NULL, cmd_spec_useragent },
226 { "verbose", &opt.verbose, cmd_boolean },
227 { "wait", &opt.wait, cmd_time },
228 { "waitretry", &opt.waitretry, cmd_time }
231 /* Look up CMDNAME in the commands[] and return its position in the
232 array. If CMDNAME is not found, return -1. */
235 command_by_name (const char *cmdname)
237 /* Use binary search for speed. Wget has ~100 commands, which
238 guarantees a worst case performance of 7 string comparisons. */
239 int lo = 0, hi = countof (commands) - 1;
243 int mid = (lo + hi) >> 1;
244 int cmp = strcasecmp (cmdname, commands[mid].name);
255 /* Reset the variables to default values. */
261 /* Most of the default values are 0. Just reset everything, and
262 fill in the non-zero values. Note that initializing pointers to
263 NULL this way is technically illegal, but porting Wget to a
264 machine where NULL is not all-zero bit pattern will be the least
265 of the implementors' worries. */
273 opt.ftp_acc = xstrdup ("anonymous");
274 opt.ftp_pass = xstrdup ("-wget@");
278 opt.http_keep_alive = 1;
280 tmp = getenv ("no_proxy");
282 opt.no_proxy = sepstring (tmp);
285 opt.read_timeout = 900;
288 opt.remove_listing = 1;
290 opt.dot_bytes = 1024;
291 opt.dot_spacing = 10;
292 opt.dots_in_line = 50;
296 /* The default for file name restriction defaults to the OS type. */
297 #if !defined(WINDOWS) && !defined(__CYGWIN__)
298 opt.restrict_files_os = restrict_unix;
300 opt.restrict_files_os = restrict_windows;
302 opt.restrict_files_ctrl = 1;
305 /* Return the user's home directory (strdup-ed), or NULL if none is
310 char *home = getenv ("HOME");
315 /* If HOME is not defined, try getting it from the password
317 struct passwd *pwd = getpwuid (getuid ());
318 if (!pwd || !pwd->pw_dir)
323 /* #### Maybe I should grab home_dir from registry, but the best
324 that I could get from there is user's Start menu. It sucks! */
328 return home ? xstrdup (home) : NULL;
331 /* Return the path to the user's .wgetrc. This is either the value of
332 `WGETRC' environment variable, or `$HOME/.wgetrc'.
334 If the `WGETRC' variable exists but the file does not exist, the
335 function will exit(). */
337 wgetrc_file_name (void)
342 /* Try the environment. */
343 env = getenv ("WGETRC");
346 if (!file_exists_p (env))
348 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
352 return xstrdup (env);
356 /* If that failed, try $HOME/.wgetrc. */
360 file = (char *)xmalloc (strlen (home) + 1 + strlen (".wgetrc") + 1);
361 sprintf (file, "%s/.wgetrc", home);
365 /* Under Windows, "home" is (for the purposes of this function) the
366 directory where `wget.exe' resides, and `wget.ini' will be used
367 as file name. SYSTEM_WGETRC should not be defined under WINDOWS.
369 It is not as trivial as I assumed, because on 95 argv[0] is full
370 path, but on NT you get what you typed in command line. --dbudor */
374 file = (char *)xmalloc (strlen (home) + strlen ("wget.ini") + 1);
375 sprintf (file, "%swget.ini", home);
381 if (!file_exists_p (file))
389 static int parse_line PARAMS ((const char *, char **, char **, int *));
390 static int setval_internal PARAMS ((int, const char *, const char *));
392 /* Initialize variables from a wgetrc file. */
395 run_wgetrc (const char *file)
401 fp = fopen (file, "rb");
404 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
405 file, strerror (errno));
408 enable_tilde_expansion = 1;
410 while ((line = read_whole_line (fp)))
415 /* Parse the line. */
416 status = parse_line (line, &com, &val, &comind);
418 /* If everything is OK, set the value. */
421 if (!setval_internal (comind, com, val))
422 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
427 else if (status == 0)
428 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
432 enable_tilde_expansion = 0;
436 /* Initialize the defaults and run the system wgetrc and user's own
443 /* Load the hard-coded defaults. */
446 /* If SYSTEM_WGETRC is defined, use it. */
448 if (file_exists_p (SYSTEM_WGETRC))
449 run_wgetrc (SYSTEM_WGETRC);
451 /* Override it with your own, if one exists. */
452 file = wgetrc_file_name ();
455 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
456 something like realpath() before comparing them with `strcmp' */
458 if (!strcmp (file, SYSTEM_WGETRC))
460 fprintf (stderr, _("\
461 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
471 /* Remove dashes and underscores from S, modifying S in the
477 char *t = s; /* t - tortoise */
478 char *h = s; /* h - hare */
480 if (*h == '_' || *h == '-')
487 /* Parse the line pointed by line, with the syntax:
488 <sp>* command <sp>* = <sp>* value <sp>*
489 Uses malloc to allocate space for command and value.
490 If the line is invalid, data is freed and 0 is returned.
497 In case of success, *COM and *VAL point to freshly allocated
498 strings, and *COMIND points to com's index. In case of error or
499 empty line, those values are unaffected. */
502 parse_line (const char *line, char **com, char **val, int *comind)
505 const char *end = line + strlen (line);
506 const char *cmdstart, *cmdend;
507 const char *valstart, *valend;
512 /* Skip leading and trailing whitespace. */
513 while (*line && ISSPACE (*line))
515 while (end > line && ISSPACE (end[-1]))
518 /* Skip empty lines and comments. */
519 if (!*line || *line == '#')
525 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
529 /* Skip '=', as well as any space before or after it. */
530 while (p < end && ISSPACE (*p))
532 if (p == end || *p != '=')
535 while (p < end && ISSPACE (*p))
541 /* The line now known to be syntactically correct. Check whether
542 the command is valid. */
543 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
545 ind = command_by_name (cmdcopy);
549 /* The command is valid. Now fill in the values and report success
552 *com = strdupdelim (cmdstart, cmdend);
553 *val = strdupdelim (valstart, valend);
557 /* Run commands[comind].action. */
560 setval_internal (int comind, const char *com, const char *val)
562 assert (0 <= comind && comind < countof (commands));
563 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
564 return ((*commands[comind].action) (com, val, commands[comind].closure));
567 /* Run command COM with value VAL. If running the command produces an
568 error, report the error and exit.
570 This is intended to be called from main() to modify Wget's behavior
571 through command-line switches. Since COM is hard-coded in main(),
572 it is not canonicalized, and this aborts when COM is not found.
574 If COMIND's are exported to init.h, this function will be changed
575 to accept COMIND directly. */
578 setoptval (const char *com, const char *val)
580 assert (val != NULL);
581 if (!setval_internal (command_by_name (com), com, val))
585 /* Parse OPT into command and value and run it. For example,
586 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
587 This is used by the `--execute' flag in main.c. */
590 run_command (const char *opt)
594 int status = parse_line (opt, &com, &val, &comind);
597 if (!setval_internal (comind, com, val))
602 else if (status == 0)
604 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
610 /* Generic helper functions, for use with `commands'. */
612 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
614 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
615 && TOLOWER((p)[1]) == (c1) \
618 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
619 && TOLOWER((p)[1]) == (c1) \
620 && TOLOWER((p)[2]) == (c2) \
624 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
625 except for error messages. */
627 cmd_boolean (const char *com, const char *val, void *closure)
631 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
632 /* "on", "yes" and "1" mean true. */
634 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
635 /* "off", "no" and "0" mean false. */
640 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
641 exec_name, com, val);
645 *(int *)closure = bool_value;
649 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
650 COM is ignored, except for error messages. Values 2 and -1
651 indicate that once defined, the value may not be changed by
652 successive wgetrc files or command-line arguments.
654 Values: 2 - Enable a particular option for good ("always")
655 1 - Enable an option ("on")
656 0 - Disable an option ("off")
657 -1 - Disable an option for good ("never") */
659 cmd_lockable_boolean (const char *com, const char *val, void *closure)
661 int lockable_boolean_value;
663 int oldval = *(int *)closure;
666 * If a config file said "always" or "never", don't allow command line
667 * arguments to override the config file.
669 if (oldval == -1 || oldval == 2)
672 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
673 lockable_boolean_value = 2;
674 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
675 lockable_boolean_value = 1;
676 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
677 lockable_boolean_value = 0;
678 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
679 lockable_boolean_value = -1;
683 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
684 exec_name, com, val);
688 *(int *)closure = lockable_boolean_value;
692 static int simple_atoi PARAMS ((const char *, const char *, int *));
694 /* Set the non-negative integer value from VAL to CLOSURE. With
695 incorrect specification, the number remains unchanged. */
697 cmd_number (const char *com, const char *val, void *closure)
699 if (!simple_atoi (val, val + strlen (val), closure))
701 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
702 exec_name, com, val);
708 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
710 cmd_number_inf (const char *com, const char *val, void *closure)
712 if (!strcasecmp (val, "inf"))
717 return cmd_number (com, val, closure);
720 /* Copy (strdup) the string at COM to a new location and place a
721 pointer to *CLOSURE. */
723 cmd_string (const char *com, const char *val, void *closure)
725 char **pstring = (char **)closure;
727 xfree_null (*pstring);
728 *pstring = xstrdup (val);
733 # define ISSEP(c) ((c) == '/')
735 # define ISSEP(c) ((c) == '/' || (c) == '\\')
738 /* Like the above, but handles tilde-expansion when reading a user's
739 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
740 gets expanded to the user's home directory. */
742 cmd_file (const char *com, const char *val, void *closure)
744 char **pstring = (char **)closure;
746 xfree_null (*pstring);
748 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
750 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
753 *pstring = xstrdup (val);
759 char *home = home_dir ();
763 homelen = strlen (home);
764 while (homelen && ISSEP (home[homelen - 1]))
765 home[--homelen] = '\0';
767 /* Skip the leading "~/". */
768 for (++val; ISSEP (*val); val++)
771 result = xmalloc (homelen + 1 + strlen (val) + 1);
772 memcpy (result, home, homelen);
773 result[homelen] = '/';
774 strcpy (result + homelen + 1, val);
780 /* Convert "\" to "/". */
783 for (s = *pstring; *s; s++)
791 /* Like cmd_file, but strips trailing '/' characters. */
793 cmd_directory (const char *com, const char *val, void *closure)
797 /* Call cmd_file() for tilde expansion and separator
798 canonicalization (backslash -> slash under Windows). These
799 things should perhaps be in a separate function. */
800 if (!cmd_file (com, val, closure))
803 s = *(char **)closure;
805 while (t > s && *--t == '/')
811 /* Split VAL by space to a vector of values, and append those values
812 to vector pointed to by the CLOSURE argument. If VAL is empty, the
813 CLOSURE vector is cleared instead. */
816 cmd_vector (const char *com, const char *val, void *closure)
818 char ***pvec = (char ***)closure;
821 *pvec = merge_vecs (*pvec, sepstring (val));
831 cmd_directory_vector (const char *com, const char *val, void *closure)
833 char ***pvec = (char ***)closure;
837 /* Strip the trailing slashes from directories. */
840 seps = sepstring (val);
841 for (t = seps; t && *t; t++)
843 int len = strlen (*t);
844 /* Skip degenerate case of root directory. */
847 if ((*t)[len - 1] == '/')
848 (*t)[len - 1] = '\0';
851 *pvec = merge_vecs (*pvec, seps);
861 static int simple_atof PARAMS ((const char *, const char *, double *));
863 /* Enginge for cmd_bytes and cmd_bytes_large: converts a string such
864 as "100k" or "2.5G" to a floating point number. */
867 parse_bytes_helper (const char *val, double *result)
870 const char *end = val + strlen (val);
872 /* Check for "inf". */
873 if (0 == strcmp (val, "inf"))
879 /* Strip trailing whitespace. */
880 while (val < end && ISSPACE (end[-1]))
885 switch (TOLOWER (end[-1]))
888 --end, mult = 1024.0;
891 --end, mult = 1048576.0;
894 --end, mult = 1073741824.0;
897 --end, mult = 1099511627776.0;
900 /* Not a recognized suffix: assume it's a digit. (If not,
901 simple_atof will raise an error.) */
905 /* Skip leading and trailing whitespace. */
906 while (val < end && ISSPACE (*val))
908 while (val < end && ISSPACE (end[-1]))
913 if (!simple_atof (val, end, &number))
916 *result = number * mult;
920 /* Parse VAL as a number and set its value to CLOSURE (which should
921 point to a long int).
923 By default, the value is assumed to be in bytes. If "K", "M", or
924 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
925 1<<30, respectively. Floating point values are allowed and are
926 cast to integer before use. The idea is to be able to use things
927 like 1.5k instead of "1536".
929 The string "inf" is returned as 0.
931 In case of error, 0 is returned and memory pointed to by CLOSURE
932 remains unmodified. */
935 cmd_bytes (const char *com, const char *val, void *closure)
938 if (!parse_bytes_helper (val, &byte_value))
940 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
941 exec_name, com, val);
944 *(long *)closure = (long)byte_value;
948 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
949 LARGE_INT. It works by converting the string to double, therefore
950 working with values up to 2^53-1 without loss of precision. This
951 value (8192 TB) is large enough to serve for a while. */
954 cmd_bytes_large (const char *com, const char *val, void *closure)
957 if (!parse_bytes_helper (val, &byte_value))
959 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
960 exec_name, com, val);
963 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
967 /* Store the value of VAL to *OUT. The value is a time period, by
968 default expressed in seconds, but also accepting suffixes "m", "h",
969 "d", and "w" for minutes, hours, days, and weeks respectively. */
972 cmd_time (const char *com, const char *val, void *closure)
975 const char *end = val + strlen (val);
977 /* Strip trailing whitespace. */
978 while (val < end && ISSPACE (end[-1]))
984 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
985 exec_name, com, val);
989 switch (TOLOWER (end[-1]))
992 --end, mult = 1; /* seconds */
995 --end, mult = 60; /* minutes */
998 --end, mult = 3600; /* hours */
1001 --end, mult = 86400.0; /* days */
1004 --end, mult = 604800.0; /* weeks */
1007 /* Not a recognized suffix: assume it belongs to the number.
1008 (If not, atof simple_atof will raise an error.) */
1012 /* Skip leading and trailing whitespace. */
1013 while (val < end && ISSPACE (*val))
1015 while (val < end && ISSPACE (end[-1]))
1020 if (!simple_atof (val, end, &number))
1023 *(double *)closure = number * mult;
1027 /* Specialized helper functions, used by `commands' to handle some
1028 options specially. */
1030 static int check_user_specified_header PARAMS ((const char *));
1033 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1035 if (!cmd_boolean (com, val, &opt.dirstruct))
1037 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1038 must be affected inversely. */
1040 opt.no_dirstruct = 0;
1042 opt.no_dirstruct = 1;
1047 cmd_spec_header (const char *com, const char *val, void *closure)
1051 /* Empty header means reset headers. */
1052 xfree_null (opt.user_header);
1053 opt.user_header = NULL;
1059 if (!check_user_specified_header (val))
1061 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1062 exec_name, com, val);
1065 i = opt.user_header ? strlen (opt.user_header) : 0;
1066 opt.user_header = (char *)xrealloc (opt.user_header, i + strlen (val)
1068 strcpy (opt.user_header + i, val);
1070 opt.user_header[i++] = '\r';
1071 opt.user_header[i++] = '\n';
1072 opt.user_header[i] = '\0';
1078 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1080 int flag = cmd_boolean (com, val, &opt.htmlify);
1081 if (flag && !opt.htmlify)
1082 opt.remove_listing = 0;
1086 /* Set the "mirror" mode. It means: recursive download, timestamping,
1087 no limit on max. recursion depth, and don't remove listings. */
1090 cmd_spec_mirror (const char *com, const char *val, void *closure)
1094 if (!cmd_boolean (com, val, &mirror))
1099 if (!opt.no_dirstruct)
1101 opt.timestamping = 1;
1102 opt.reclevel = INFINITE_RECURSION;
1103 opt.remove_listing = 0;
1108 /* Set progress.type to VAL, but verify that it's a valid progress
1109 implementation before that. */
1112 cmd_spec_progress (const char *com, const char *val, void *closure)
1114 if (!valid_progress_implementation_p (val))
1116 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1117 exec_name, com, val);
1120 xfree_null (opt.progress_type);
1122 /* Don't call set_progress_implementation here. It will be called
1123 in main() when it becomes clear what the log output is. */
1124 opt.progress_type = xstrdup (val);
1128 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1129 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1133 cmd_spec_recursive (const char *com, const char *val, void *closure)
1135 if (!cmd_boolean (com, val, &opt.recursive))
1139 if (opt.recursive && !opt.no_dirstruct)
1146 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1148 int restrict_os = opt.restrict_files_os;
1149 int restrict_ctrl = opt.restrict_files_ctrl;
1151 const char *end = strchr (val, ',');
1153 end = val + strlen (val);
1155 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1157 if (VAL_IS ("unix"))
1158 restrict_os = restrict_unix;
1159 else if (VAL_IS ("windows"))
1160 restrict_os = restrict_windows;
1161 else if (VAL_IS ("nocontrol"))
1167 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1168 exec_name, com, val);
1176 if (!strcmp (end + 1, "nocontrol"))
1182 opt.restrict_files_os = restrict_os;
1183 opt.restrict_files_ctrl = restrict_ctrl;
1187 /* Set all three timeout values. */
1190 cmd_spec_timeout (const char *com, const char *val, void *closure)
1193 if (!cmd_time (com, val, &value))
1195 opt.read_timeout = value;
1196 opt.connect_timeout = value;
1197 opt.dns_timeout = value;
1202 cmd_spec_useragent (const char *com, const char *val, void *closure)
1204 /* Just check for empty string and newline, so we don't throw total
1205 junk to the server. */
1206 if (!*val || strchr (val, '\n'))
1208 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1209 exec_name, com, val);
1212 opt.useragent = xstrdup (val);
1216 /* Miscellaneous useful routines. */
1218 /* A very simple atoi clone, more portable than strtol and friends,
1219 but reports errors, unlike atoi. Returns 1 on success, 0 on
1220 failure. In case of success, stores result to *DEST. */
1223 simple_atoi (const char *beg, const char *end, int *dest)
1231 for (p = beg; p < end && ISDIGIT (*p); p++)
1232 result = (10 * result) + (*p - '0');
1241 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1242 doesn't handle exponential notation. Returns 1 on success, 0 on
1243 failure. In case of success, stores its result to *DEST. */
1246 simple_atof (const char *beg, const char *end, double *dest)
1256 for (p = beg; p < end; p++)
1262 result = (10 * result) + (ch - '0');
1264 result += (ch - '0') / (divider *= 10);
1285 check_user_specified_header (const char *s)
1289 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1290 /* The header MUST contain `:' preceded by at least one
1291 non-whitespace character. */
1292 if (*p != ':' || p == s)
1294 /* The header MUST NOT contain newlines. */
1295 if (strchr (s, '\n'))
1300 void cleanup_html_url PARAMS ((void));
1301 void res_cleanup PARAMS ((void));
1302 void downloaded_files_free PARAMS ((void));
1303 void http_cleanup PARAMS ((void));
1306 /* Free the memory allocated by global variables. */
1310 /* Free external resources, close files, etc. */
1315 /* We're exiting anyway so there's no real need to call free()
1316 hundreds of times. Skipping the frees will make Wget exit
1319 However, when detecting leaks, it's crucial to free() everything
1320 because then you can find the real leaks, i.e. the allocated
1321 memory which grows with the size of the program. */
1327 cleanup_html_url ();
1328 downloaded_files_free ();
1330 if (wget_cookie_jar)
1331 cookie_jar_delete (wget_cookie_jar);
1334 extern acc_t *netrc_list;
1335 free_netrc (netrc_list);
1337 xfree_null (opt.lfilename);
1338 xfree_null (opt.dir_prefix);
1339 xfree_null (opt.input_filename);
1340 xfree_null (opt.output_document);
1341 free_vec (opt.accepts);
1342 free_vec (opt.rejects);
1343 free_vec (opt.excludes);
1344 free_vec (opt.includes);
1345 free_vec (opt.domains);
1346 free_vec (opt.follow_tags);
1347 free_vec (opt.ignore_tags);
1348 xfree_null (opt.progress_type);
1349 xfree (opt.ftp_acc);
1350 xfree_null (opt.ftp_pass);
1351 xfree_null (opt.ftp_proxy);
1352 xfree_null (opt.https_proxy);
1353 xfree_null (opt.http_proxy);
1354 free_vec (opt.no_proxy);
1355 xfree_null (opt.useragent);
1356 xfree_null (opt.referer);
1357 xfree_null (opt.http_user);
1358 xfree_null (opt.http_passwd);
1359 xfree_null (opt.user_header);
1361 xfree_null (opt.sslcertkey);
1362 xfree_null (opt.sslcertfile);
1363 #endif /* HAVE_SSL */
1364 xfree_null (opt.bind_address);
1365 xfree_null (opt.cookies_input);
1366 xfree_null (opt.cookies_output);