1 /* Reading/parsing the initialization file.
2 Copyright (C) 2005 Free Software Foundation, Inc.
4 This file is part of GNU Wget.
6 GNU Wget is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
11 GNU Wget is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with Wget; if not, write to the Free Software
18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
20 In addition, as a special exception, the Free Software Foundation
21 gives permission to link the code of its release of Wget with the
22 OpenSSL project's "OpenSSL" library (or with modified versions of it
23 that use the same license as the "OpenSSL" library), and distribute
24 the linked executables. You must obey the GNU General Public License
25 in all respects for all of the code used other than "OpenSSL". If you
26 modify this file, you may extend this exception to your version of the
27 file, but you are not obligated to do so. If you do not wish to do
28 so, delete this exception statement from your version. */
33 #include <sys/types.h>
56 #include "recur.h" /* for INFINITE_RECURSION */
57 #include "convert.h" /* for convert_cleanup */
63 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
64 otherwise, it will be performed by the shell. This variable will
65 be set by the wgetrc-reading function. */
67 static int enable_tilde_expansion;
70 #define CMD_DECLARE(func) static int func \
71 PARAMS ((const char *, const char *, void *))
73 CMD_DECLARE (cmd_boolean);
74 CMD_DECLARE (cmd_bytes);
75 CMD_DECLARE (cmd_bytes_large);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_lockable_boolean);
78 CMD_DECLARE (cmd_number);
79 CMD_DECLARE (cmd_number_inf);
80 CMD_DECLARE (cmd_string);
81 CMD_DECLARE (cmd_file);
82 CMD_DECLARE (cmd_directory);
83 CMD_DECLARE (cmd_time);
84 CMD_DECLARE (cmd_vector);
86 CMD_DECLARE (cmd_spec_dirstruct);
87 CMD_DECLARE (cmd_spec_header);
88 CMD_DECLARE (cmd_spec_htmlify);
89 CMD_DECLARE (cmd_spec_mirror);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
93 CMD_DECLARE (cmd_spec_timeout);
94 CMD_DECLARE (cmd_spec_useragent);
96 /* List of recognized commands, each consisting of name, closure and
97 function. When adding a new command, simply add it to the list,
98 but be sure to keep the list sorted alphabetically, as
99 command_by_name depends on it. Also, be sure to add any entries
100 that allocate memory (e.g. cmd_string and cmd_vector guys) to the
101 cleanup() function below. */
106 int (*action) PARAMS ((const char *, const char *, void *));
108 { "accept", &opt.accepts, cmd_vector },
109 { "addhostdir", &opt.add_hostdir, cmd_boolean },
110 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
111 { "background", &opt.background, cmd_boolean },
112 { "backupconverted", &opt.backup_converted, cmd_boolean },
113 { "backups", &opt.backups, cmd_number },
114 { "base", &opt.base_href, cmd_string },
115 { "bindaddress", &opt.bind_address, cmd_string },
116 { "cache", &opt.allow_cache, cmd_boolean },
117 { "connecttimeout", &opt.connect_timeout, cmd_time },
118 { "continue", &opt.always_rest, cmd_boolean },
119 { "convertlinks", &opt.convert_links, cmd_boolean },
120 { "cookies", &opt.cookies, cmd_boolean },
121 { "cutdirs", &opt.cut_dirs, cmd_number },
123 { "debug", &opt.debug, cmd_boolean },
125 { "deleteafter", &opt.delete_after, cmd_boolean },
126 { "dirprefix", &opt.dir_prefix, cmd_directory },
127 { "dirstruct", NULL, cmd_spec_dirstruct },
128 { "dnscache", &opt.dns_cache, cmd_boolean },
129 { "dnstimeout", &opt.dns_timeout, cmd_time },
130 { "domains", &opt.domains, cmd_vector },
131 { "dotbytes", &opt.dot_bytes, cmd_bytes },
132 { "dotsinline", &opt.dots_in_line, cmd_number },
133 { "dotspacing", &opt.dot_spacing, cmd_number },
134 { "dotstyle", &opt.dot_style, cmd_string },
136 { "egdfile", &opt.sslegdsock, cmd_file },
138 { "excludedirectories", &opt.excludes, cmd_directory_vector },
139 { "excludedomains", &opt.exclude_domains, cmd_vector },
140 { "followftp", &opt.follow_ftp, cmd_boolean },
141 { "followtags", &opt.follow_tags, cmd_vector },
142 { "forcehtml", &opt.force_html, cmd_boolean },
143 { "ftppasswd", &opt.ftp_pass, cmd_string },
144 { "ftpproxy", &opt.ftp_proxy, cmd_string },
145 { "glob", &opt.ftp_glob, cmd_boolean },
146 { "header", &opt.user_headers, cmd_spec_header },
147 { "htmlextension", &opt.html_extension, cmd_boolean },
148 { "htmlify", NULL, cmd_spec_htmlify },
149 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
150 { "httppasswd", &opt.http_passwd, cmd_string },
151 { "httpproxy", &opt.http_proxy, cmd_string },
152 { "httpsproxy", &opt.https_proxy, cmd_string },
153 { "httpuser", &opt.http_user, cmd_string },
154 { "ignorelength", &opt.ignore_length, cmd_boolean },
155 { "ignoretags", &opt.ignore_tags, cmd_vector },
156 { "includedirectories", &opt.includes, cmd_directory_vector },
158 { "inet4only", &opt.ipv4_only, cmd_boolean },
159 { "inet6only", &opt.ipv6_only, cmd_boolean },
161 { "input", &opt.input_filename, cmd_file },
162 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
163 { "killlonger", &opt.kill_longer, cmd_boolean },
164 { "limitrate", &opt.limit_rate, cmd_bytes },
165 { "loadcookies", &opt.cookies_input, cmd_file },
166 { "logfile", &opt.lfilename, cmd_file },
167 { "login", &opt.ftp_acc, cmd_string },
168 { "mirror", NULL, cmd_spec_mirror },
169 { "netrc", &opt.netrc, cmd_boolean },
170 { "noclobber", &opt.noclobber, cmd_boolean },
171 { "noparent", &opt.no_parent, cmd_boolean },
172 { "noproxy", &opt.no_proxy, cmd_vector },
173 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
174 { "outputdocument", &opt.output_document, cmd_file },
175 { "pagerequisites", &opt.page_requisites, cmd_boolean },
176 { "passiveftp", &opt.ftp_pasv, cmd_lockable_boolean },
177 { "postdata", &opt.post_data, cmd_string },
178 { "postfile", &opt.post_file_name, cmd_file },
179 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
180 { "progress", &opt.progress_type, cmd_spec_progress },
181 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
182 { "proxypasswd", &opt.proxy_passwd, cmd_string },
183 { "proxyuser", &opt.proxy_user, cmd_string },
184 { "quiet", &opt.quiet, cmd_boolean },
185 { "quota", &opt.quota, cmd_bytes_large },
186 { "randomwait", &opt.random_wait, cmd_boolean },
187 { "readtimeout", &opt.read_timeout, cmd_time },
188 { "reclevel", &opt.reclevel, cmd_number_inf },
189 { "recursive", NULL, cmd_spec_recursive },
190 { "referer", &opt.referer, cmd_string },
191 { "reject", &opt.rejects, cmd_vector },
192 { "relativeonly", &opt.relative_only, cmd_boolean },
193 { "removelisting", &opt.remove_listing, cmd_boolean },
194 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
195 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
196 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
197 { "robots", &opt.use_robots, cmd_boolean },
198 { "savecookies", &opt.cookies_output, cmd_file },
199 { "saveheaders", &opt.save_headers, cmd_boolean },
200 { "serverresponse", &opt.server_response, cmd_boolean },
201 { "spanhosts", &opt.spanhost, cmd_boolean },
202 { "spider", &opt.spider, cmd_boolean },
204 { "sslcadir", &opt.sslcadir, cmd_directory },
205 { "sslcafile", &opt.sslcafile, cmd_file },
206 { "sslcertfile", &opt.sslcertfile, cmd_file },
207 { "sslcertkey", &opt.sslcertkey, cmd_file },
208 { "sslcerttype", &opt.sslcerttype, cmd_number },
209 { "sslcheckcert", &opt.sslcheckcert, cmd_number },
210 { "sslprotocol", &opt.sslprotocol, cmd_number },
211 #endif /* HAVE_SSL */
212 { "strictcomments", &opt.strict_comments, cmd_boolean },
213 { "timeout", NULL, cmd_spec_timeout },
214 { "timestamping", &opt.timestamping, cmd_boolean },
215 { "tries", &opt.ntry, cmd_number_inf },
216 { "useproxy", &opt.use_proxy, cmd_boolean },
217 { "useragent", NULL, cmd_spec_useragent },
218 { "verbose", &opt.verbose, cmd_boolean },
219 { "wait", &opt.wait, cmd_time },
220 { "waitretry", &opt.waitretry, cmd_time }
223 /* Look up CMDNAME in the commands[] and return its position in the
224 array. If CMDNAME is not found, return -1. */
227 command_by_name (const char *cmdname)
229 /* Use binary search for speed. Wget has ~100 commands, which
230 guarantees a worst case performance of 7 string comparisons. */
231 int lo = 0, hi = countof (commands) - 1;
235 int mid = (lo + hi) >> 1;
236 int cmp = strcasecmp (cmdname, commands[mid].name);
247 /* Reset the variables to default values. */
253 /* Most of the default values are 0. Just reset everything, and
254 fill in the non-zero values. Note that initializing pointers to
255 NULL this way is technically illegal, but porting Wget to a
256 machine where NULL is not all-zero bit pattern will be the least
257 of the implementors' worries. */
265 opt.ftp_acc = xstrdup ("anonymous");
266 opt.ftp_pass = xstrdup ("-wget@");
270 opt.http_keep_alive = 1;
272 tmp = getenv ("no_proxy");
274 opt.no_proxy = sepstring (tmp);
277 opt.read_timeout = 900;
280 opt.remove_listing = 1;
282 opt.dot_bytes = 1024;
283 opt.dot_spacing = 10;
284 opt.dots_in_line = 50;
289 /* The default for file name restriction defaults to the OS type. */
290 #if !defined(WINDOWS) && !defined(__CYGWIN__)
291 opt.restrict_files_os = restrict_unix;
293 opt.restrict_files_os = restrict_windows;
295 opt.restrict_files_ctrl = 1;
298 /* Return the user's home directory (strdup-ed), or NULL if none is
303 char *home = getenv ("HOME");
308 /* If HOME is not defined, try getting it from the password
310 struct passwd *pwd = getpwuid (getuid ());
311 if (!pwd || !pwd->pw_dir)
315 /* Under Windows, if $HOME isn't defined, use the directory where
316 `wget.exe' resides. */
321 return home ? xstrdup (home) : NULL;
324 /* Return the path to the user's .wgetrc. This is either the value of
325 `WGETRC' environment variable, or `$HOME/.wgetrc'.
327 If the `WGETRC' variable exists but the file does not exist, the
328 function will exit(). */
330 wgetrc_file_name (void)
335 /* Try the environment. */
336 env = getenv ("WGETRC");
339 if (!file_exists_p (env))
341 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
345 return xstrdup (env);
348 /* If that failed, try $HOME/.wgetrc. */
351 file = aprintf ("%s/.wgetrc", home);
355 /* Under Windows, if we still haven't found .wgetrc, look for the file
356 `wget.ini' in the directory where `wget.exe' resides; we do this for
357 backward compatibility with previous versions of Wget.
358 SYSTEM_WGETRC should not be defined under WINDOWS. */
359 if (!file || !file_exists_p (file))
365 file = aprintf ("%s/wget.ini", home);
371 if (!file_exists_p (file))
379 static int parse_line PARAMS ((const char *, char **, char **, int *));
380 static int setval_internal PARAMS ((int, const char *, const char *));
382 /* Initialize variables from a wgetrc file. */
385 run_wgetrc (const char *file)
391 fp = fopen (file, "rb");
394 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
395 file, strerror (errno));
398 enable_tilde_expansion = 1;
400 while ((line = read_whole_line (fp)) != NULL)
405 /* Parse the line. */
406 status = parse_line (line, &com, &val, &comind);
408 /* If everything is OK, set the value. */
411 if (!setval_internal (comind, com, val))
412 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
417 else if (status == 0)
418 fprintf (stderr, _("%s: Error in %s at line %d.\n"), exec_name,
422 enable_tilde_expansion = 0;
426 /* Initialize the defaults and run the system wgetrc and user's own
433 /* Load the hard-coded defaults. */
436 /* If SYSTEM_WGETRC is defined, use it. */
438 if (file_exists_p (SYSTEM_WGETRC))
439 run_wgetrc (SYSTEM_WGETRC);
441 /* Override it with your own, if one exists. */
442 file = wgetrc_file_name ();
445 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
446 something like realpath() before comparing them with `strcmp' */
448 if (!strcmp (file, SYSTEM_WGETRC))
450 fprintf (stderr, _("\
451 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
461 /* Remove dashes and underscores from S, modifying S in the
467 char *t = s; /* t - tortoise */
468 char *h = s; /* h - hare */
470 if (*h == '_' || *h == '-')
477 /* Parse the line pointed by line, with the syntax:
478 <sp>* command <sp>* = <sp>* value <sp>*
479 Uses malloc to allocate space for command and value.
480 If the line is invalid, data is freed and 0 is returned.
487 In case of success, *COM and *VAL point to freshly allocated
488 strings, and *COMIND points to com's index. In case of error or
489 empty line, those values are unaffected. */
492 parse_line (const char *line, char **com, char **val, int *comind)
495 const char *end = line + strlen (line);
496 const char *cmdstart, *cmdend;
497 const char *valstart, *valend;
502 /* Skip leading and trailing whitespace. */
503 while (*line && ISSPACE (*line))
505 while (end > line && ISSPACE (end[-1]))
508 /* Skip empty lines and comments. */
509 if (!*line || *line == '#')
515 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
519 /* Skip '=', as well as any space before or after it. */
520 while (p < end && ISSPACE (*p))
522 if (p == end || *p != '=')
525 while (p < end && ISSPACE (*p))
531 /* The line now known to be syntactically correct. Check whether
532 the command is valid. */
533 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
535 ind = command_by_name (cmdcopy);
539 /* The command is valid. Now fill in the values and report success
542 *com = strdupdelim (cmdstart, cmdend);
543 *val = strdupdelim (valstart, valend);
547 /* Run commands[comind].action. */
550 setval_internal (int comind, const char *com, const char *val)
552 assert (0 <= comind && comind < countof (commands));
553 DEBUGP (("Setting %s (%d) to %s\n", com, comind, val));
554 return ((*commands[comind].action) (com, val, commands[comind].closure));
557 /* Run command COM with value VAL. If running the command produces an
558 error, report the error and exit.
560 This is intended to be called from main() to modify Wget's behavior
561 through command-line switches. Since COM is hard-coded in main(),
562 it is not canonicalized, and this aborts when COM is not found.
564 If COMIND's are exported to init.h, this function will be changed
565 to accept COMIND directly. */
568 setoptval (const char *com, const char *val)
570 assert (val != NULL);
571 if (!setval_internal (command_by_name (com), com, val))
575 /* Parse OPT into command and value and run it. For example,
576 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
577 This is used by the `--execute' flag in main.c. */
580 run_command (const char *opt)
584 int status = parse_line (opt, &com, &val, &comind);
587 if (!setval_internal (comind, com, val))
592 else if (status == 0)
594 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
600 /* Generic helper functions, for use with `commands'. */
602 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
604 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
605 && TOLOWER((p)[1]) == (c1) \
608 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
609 && TOLOWER((p)[1]) == (c1) \
610 && TOLOWER((p)[2]) == (c2) \
614 /* Store the boolean value from VAL to CLOSURE. COM is ignored,
615 except for error messages. */
617 cmd_boolean (const char *com, const char *val, void *closure)
621 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
622 /* "on", "yes" and "1" mean true. */
624 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
625 /* "off", "no" and "0" mean false. */
630 _("%s: %s: Invalid boolean `%s', use `on' or `off'.\n"),
631 exec_name, com, val);
635 *(int *)closure = bool_value;
639 /* Store the lockable_boolean {2, 1, 0, -1} value from VAL to CLOSURE.
640 COM is ignored, except for error messages. Values 2 and -1
641 indicate that once defined, the value may not be changed by
642 successive wgetrc files or command-line arguments.
644 Values: 2 - Enable a particular option for good ("always")
645 1 - Enable an option ("on")
646 0 - Disable an option ("off")
647 -1 - Disable an option for good ("never") */
649 cmd_lockable_boolean (const char *com, const char *val, void *closure)
651 int lockable_boolean_value;
653 int oldval = *(int *)closure;
656 * If a config file said "always" or "never", don't allow command line
657 * arguments to override the config file.
659 if (oldval == -1 || oldval == 2)
662 if (0 == strcasecmp (val, "always") || CMP1 (val, '2'))
663 lockable_boolean_value = 2;
664 else if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
665 lockable_boolean_value = 1;
666 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
667 lockable_boolean_value = 0;
668 else if (0 == strcasecmp (val, "never") || CMP2 (val, '-', '1'))
669 lockable_boolean_value = -1;
673 _("%s: %s: Invalid boolean `%s', use always, on, off, or never.\n"),
674 exec_name, com, val);
678 *(int *)closure = lockable_boolean_value;
682 static int simple_atoi PARAMS ((const char *, const char *, int *));
684 /* Set the non-negative integer value from VAL to CLOSURE. With
685 incorrect specification, the number remains unchanged. */
687 cmd_number (const char *com, const char *val, void *closure)
689 if (!simple_atoi (val, val + strlen (val), closure)
690 || *(int *) closure < 0)
692 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
693 exec_name, com, val);
699 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
701 cmd_number_inf (const char *com, const char *val, void *closure)
703 if (!strcasecmp (val, "inf"))
708 return cmd_number (com, val, closure);
711 /* Copy (strdup) the string at COM to a new location and place a
712 pointer to *CLOSURE. */
714 cmd_string (const char *com, const char *val, void *closure)
716 char **pstring = (char **)closure;
718 xfree_null (*pstring);
719 *pstring = xstrdup (val);
724 # define ISSEP(c) ((c) == '/')
726 # define ISSEP(c) ((c) == '/' || (c) == '\\')
729 /* Like the above, but handles tilde-expansion when reading a user's
730 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
731 gets expanded to the user's home directory. */
733 cmd_file (const char *com, const char *val, void *closure)
735 char **pstring = (char **)closure;
737 xfree_null (*pstring);
739 /* #### If VAL is empty, perhaps should set *CLOSURE to NULL. */
741 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
744 *pstring = xstrdup (val);
749 char *home = home_dir ();
753 homelen = strlen (home);
754 while (homelen && ISSEP (home[homelen - 1]))
755 home[--homelen] = '\0';
757 /* Skip the leading "~/". */
758 for (++val; ISSEP (*val); val++)
761 *pstring = concat_strings (home, "/", val, (char *) 0);
765 /* Convert "\" to "/". */
768 for (s = *pstring; *s; s++)
776 /* Like cmd_file, but strips trailing '/' characters. */
778 cmd_directory (const char *com, const char *val, void *closure)
782 /* Call cmd_file() for tilde expansion and separator
783 canonicalization (backslash -> slash under Windows). These
784 things should perhaps be in a separate function. */
785 if (!cmd_file (com, val, closure))
788 s = *(char **)closure;
790 while (t > s && *--t == '/')
796 /* Split VAL by space to a vector of values, and append those values
797 to vector pointed to by the CLOSURE argument. If VAL is empty, the
798 CLOSURE vector is cleared instead. */
801 cmd_vector (const char *com, const char *val, void *closure)
803 char ***pvec = (char ***)closure;
806 *pvec = merge_vecs (*pvec, sepstring (val));
816 cmd_directory_vector (const char *com, const char *val, void *closure)
818 char ***pvec = (char ***)closure;
822 /* Strip the trailing slashes from directories. */
825 seps = sepstring (val);
826 for (t = seps; t && *t; t++)
828 int len = strlen (*t);
829 /* Skip degenerate case of root directory. */
832 if ((*t)[len - 1] == '/')
833 (*t)[len - 1] = '\0';
836 *pvec = merge_vecs (*pvec, seps);
846 static int simple_atof PARAMS ((const char *, const char *, double *));
848 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
849 "100k" or "2.5G" to a floating point number. */
852 parse_bytes_helper (const char *val, double *result)
855 const char *end = val + strlen (val);
857 /* Check for "inf". */
858 if (0 == strcmp (val, "inf"))
864 /* Strip trailing whitespace. */
865 while (val < end && ISSPACE (end[-1]))
870 switch (TOLOWER (end[-1]))
873 --end, mult = 1024.0;
876 --end, mult = 1048576.0;
879 --end, mult = 1073741824.0;
882 --end, mult = 1099511627776.0;
885 /* Not a recognized suffix: assume it's a digit. (If not,
886 simple_atof will raise an error.) */
890 /* Skip leading and trailing whitespace. */
891 while (val < end && ISSPACE (*val))
893 while (val < end && ISSPACE (end[-1]))
898 if (!simple_atof (val, end, &number) || number < 0)
901 *result = number * mult;
905 /* Parse VAL as a number and set its value to CLOSURE (which should
908 By default, the value is assumed to be in bytes. If "K", "M", or
909 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
910 1<<30, respectively. Floating point values are allowed and are
911 cast to integer before use. The idea is to be able to use things
912 like 1.5k instead of "1536".
914 The string "inf" is returned as 0.
916 In case of error, 0 is returned and memory pointed to by CLOSURE
917 remains unmodified. */
920 cmd_bytes (const char *com, const char *val, void *closure)
923 if (!parse_bytes_helper (val, &byte_value))
925 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
926 exec_name, com, val);
929 *(wgint *)closure = (wgint)byte_value;
933 /* Like cmd_bytes, but CLOSURE is interpreted as a pointer to
934 LARGE_INT. It works by converting the string to double, therefore
935 working with values up to 2^53-1 without loss of precision. This
936 value (8192 TB) is large enough to serve for a while. */
939 cmd_bytes_large (const char *com, const char *val, void *closure)
942 if (!parse_bytes_helper (val, &byte_value))
944 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
945 exec_name, com, val);
948 *(LARGE_INT *)closure = (LARGE_INT)byte_value;
952 /* Store the value of VAL to *OUT. The value is a time period, by
953 default expressed in seconds, but also accepting suffixes "m", "h",
954 "d", and "w" for minutes, hours, days, and weeks respectively. */
957 cmd_time (const char *com, const char *val, void *closure)
960 const char *end = val + strlen (val);
962 /* Strip trailing whitespace. */
963 while (val < end && ISSPACE (end[-1]))
969 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
970 exec_name, com, val);
974 switch (TOLOWER (end[-1]))
977 --end, mult = 1; /* seconds */
980 --end, mult = 60; /* minutes */
983 --end, mult = 3600; /* hours */
986 --end, mult = 86400.0; /* days */
989 --end, mult = 604800.0; /* weeks */
992 /* Not a recognized suffix: assume it belongs to the number.
993 (If not, atof simple_atof will raise an error.) */
997 /* Skip leading and trailing whitespace. */
998 while (val < end && ISSPACE (*val))
1000 while (val < end && ISSPACE (end[-1]))
1005 if (!simple_atof (val, end, &number))
1008 *(double *)closure = number * mult;
1012 /* Specialized helper functions, used by `commands' to handle some
1013 options specially. */
1015 static int check_user_specified_header PARAMS ((const char *));
1018 cmd_spec_dirstruct (const char *com, const char *val, void *closure)
1020 if (!cmd_boolean (com, val, &opt.dirstruct))
1022 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1023 must be affected inversely. */
1025 opt.no_dirstruct = 0;
1027 opt.no_dirstruct = 1;
1032 cmd_spec_header (const char *com, const char *val, void *closure)
1034 if (!check_user_specified_header (val))
1036 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1037 exec_name, com, val);
1040 return cmd_vector (com, val, closure);
1044 cmd_spec_htmlify (const char *com, const char *val, void *closure)
1046 int flag = cmd_boolean (com, val, &opt.htmlify);
1047 if (flag && !opt.htmlify)
1048 opt.remove_listing = 0;
1052 /* Set the "mirror" mode. It means: recursive download, timestamping,
1053 no limit on max. recursion depth, and don't remove listings. */
1056 cmd_spec_mirror (const char *com, const char *val, void *closure)
1060 if (!cmd_boolean (com, val, &mirror))
1065 if (!opt.no_dirstruct)
1067 opt.timestamping = 1;
1068 opt.reclevel = INFINITE_RECURSION;
1069 opt.remove_listing = 0;
1074 /* Set progress.type to VAL, but verify that it's a valid progress
1075 implementation before that. */
1078 cmd_spec_progress (const char *com, const char *val, void *closure)
1080 if (!valid_progress_implementation_p (val))
1082 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1083 exec_name, com, val);
1086 xfree_null (opt.progress_type);
1088 /* Don't call set_progress_implementation here. It will be called
1089 in main() when it becomes clear what the log output is. */
1090 opt.progress_type = xstrdup (val);
1094 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1095 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1099 cmd_spec_recursive (const char *com, const char *val, void *closure)
1101 if (!cmd_boolean (com, val, &opt.recursive))
1105 if (opt.recursive && !opt.no_dirstruct)
1112 cmd_spec_restrict_file_names (const char *com, const char *val, void *closure)
1114 int restrict_os = opt.restrict_files_os;
1115 int restrict_ctrl = opt.restrict_files_ctrl;
1117 const char *end = strchr (val, ',');
1119 end = val + strlen (val);
1121 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1123 if (VAL_IS ("unix"))
1124 restrict_os = restrict_unix;
1125 else if (VAL_IS ("windows"))
1126 restrict_os = restrict_windows;
1127 else if (VAL_IS ("nocontrol"))
1133 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1134 exec_name, com, val);
1142 if (!strcmp (end + 1, "nocontrol"))
1148 opt.restrict_files_os = restrict_os;
1149 opt.restrict_files_ctrl = restrict_ctrl;
1153 /* Set all three timeout values. */
1156 cmd_spec_timeout (const char *com, const char *val, void *closure)
1159 if (!cmd_time (com, val, &value))
1161 opt.read_timeout = value;
1162 opt.connect_timeout = value;
1163 opt.dns_timeout = value;
1168 cmd_spec_useragent (const char *com, const char *val, void *closure)
1170 /* Just check for empty string and newline, so we don't throw total
1171 junk to the server. */
1172 if (!*val || strchr (val, '\n'))
1174 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1175 exec_name, com, val);
1178 opt.useragent = xstrdup (val);
1182 /* Miscellaneous useful routines. */
1184 /* A very simple atoi clone, more useful than atoi because it works on
1185 delimited strings, and has error reportage. Returns 1 on success,
1186 0 on failure. If successful, stores result to *DEST. */
1189 simple_atoi (const char *beg, const char *end, int *dest)
1193 const char *p = beg;
1195 while (p < end && ISSPACE (*p))
1197 if (p < end && (*p == '-' || *p == '+'))
1199 negative = (*p == '-');
1205 /* Read negative numbers in a separate loop because the most
1206 negative integer cannot be represented as a positive number. */
1209 for (; p < end && ISDIGIT (*p); p++)
1211 int next = (10 * result) + (*p - '0');
1213 return 0; /* overflow */
1217 for (; p < end && ISDIGIT (*p); p++)
1219 int next = (10 * result) - (*p - '0');
1221 return 0; /* underflow */
1232 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1233 doesn't handle exponential notation. Returns 1 on success, 0 on
1234 failure. In case of success, stores its result to *DEST. */
1237 simple_atof (const char *beg, const char *end, double *dest)
1246 const char *p = beg;
1248 while (p < end && ISSPACE (*p))
1250 if (p < end && (*p == '-' || *p == '+'))
1252 negative = (*p == '-');
1256 for (; p < end; p++)
1262 result = (10 * result) + (ch - '0');
1264 result += (ch - '0') / (divider *= 10);
1287 check_user_specified_header (const char *s)
1291 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1292 /* The header MUST contain `:' preceded by at least one
1293 non-whitespace character. */
1294 if (*p != ':' || p == s)
1296 /* The header MUST NOT contain newlines. */
1297 if (strchr (s, '\n'))
1302 void cleanup_html_url PARAMS ((void));
1303 void res_cleanup PARAMS ((void));
1304 void downloaded_files_free PARAMS ((void));
1305 void http_cleanup PARAMS ((void));
1308 /* Free the memory allocated by global variables. */
1312 /* Free external resources, close files, etc. */
1315 extern FILE *output_stream;
1317 fclose (output_stream);
1318 /* No need to check for error because Wget flushes its output (and
1319 checks for errors) after any data arrives. */
1322 /* We're exiting anyway so there's no real need to call free()
1323 hundreds of times. Skipping the frees will make Wget exit
1326 However, when detecting leaks, it's crucial to free() everything
1327 because then you can find the real leaks, i.e. the allocated
1328 memory which grows with the size of the program. */
1334 cleanup_html_url ();
1335 downloaded_files_free ();
1340 extern acc_t *netrc_list;
1341 free_netrc (netrc_list);
1343 xfree_null (opt.lfilename);
1344 xfree_null (opt.dir_prefix);
1345 xfree_null (opt.input_filename);
1346 xfree_null (opt.output_document);
1347 free_vec (opt.accepts);
1348 free_vec (opt.rejects);
1349 free_vec (opt.excludes);
1350 free_vec (opt.includes);
1351 free_vec (opt.domains);
1352 free_vec (opt.follow_tags);
1353 free_vec (opt.ignore_tags);
1354 xfree_null (opt.progress_type);
1355 xfree (opt.ftp_acc);
1356 xfree_null (opt.ftp_pass);
1357 xfree_null (opt.ftp_proxy);
1358 xfree_null (opt.https_proxy);
1359 xfree_null (opt.http_proxy);
1360 free_vec (opt.no_proxy);
1361 xfree_null (opt.useragent);
1362 xfree_null (opt.referer);
1363 xfree_null (opt.http_user);
1364 xfree_null (opt.http_passwd);
1365 free_vec (opt.user_headers);
1367 xfree_null (opt.sslcertkey);
1368 xfree_null (opt.sslcertfile);
1369 #endif /* HAVE_SSL */
1370 xfree_null (opt.bind_address);
1371 xfree_null (opt.cookies_input);
1372 xfree_null (opt.cookies_output);