1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996-2005 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 Foundation, Inc.,
18 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 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. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
57 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
58 otherwise, it will be performed by the shell. This variable will
59 be set by the wgetrc-reading function. */
61 static bool enable_tilde_expansion;
64 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
66 CMD_DECLARE (cmd_boolean);
67 CMD_DECLARE (cmd_bytes);
68 CMD_DECLARE (cmd_bytes_sum);
70 CMD_DECLARE (cmd_cert_type);
72 CMD_DECLARE (cmd_directory_vector);
73 CMD_DECLARE (cmd_number);
74 CMD_DECLARE (cmd_number_inf);
75 CMD_DECLARE (cmd_string);
76 CMD_DECLARE (cmd_file);
77 CMD_DECLARE (cmd_directory);
78 CMD_DECLARE (cmd_time);
79 CMD_DECLARE (cmd_vector);
81 CMD_DECLARE (cmd_spec_dirstruct);
82 CMD_DECLARE (cmd_spec_header);
83 CMD_DECLARE (cmd_spec_htmlify);
84 CMD_DECLARE (cmd_spec_mirror);
85 CMD_DECLARE (cmd_spec_prefer_family);
86 CMD_DECLARE (cmd_spec_progress);
87 CMD_DECLARE (cmd_spec_recursive);
88 CMD_DECLARE (cmd_spec_restrict_file_names);
90 CMD_DECLARE (cmd_spec_secure_protocol);
92 CMD_DECLARE (cmd_spec_timeout);
93 CMD_DECLARE (cmd_spec_useragent);
94 CMD_DECLARE (cmd_spec_verbose);
96 /* List of recognized commands, each consisting of name, place 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) to the
101 cleanup() function below. */
106 bool (*action) (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 },
117 { "cacertificate", &opt.ca_cert, cmd_file },
119 { "cache", &opt.allow_cache, cmd_boolean },
121 { "cadirectory", &opt.ca_directory, cmd_directory },
122 { "certificate", &opt.cert_file, cmd_file },
123 { "certificatetype", &opt.cert_type, cmd_cert_type },
124 { "checkcertificate", &opt.check_cert, cmd_boolean },
126 { "connecttimeout", &opt.connect_timeout, cmd_time },
127 { "continue", &opt.always_rest, cmd_boolean },
128 { "convertlinks", &opt.convert_links, cmd_boolean },
129 { "cookies", &opt.cookies, cmd_boolean },
130 { "cutdirs", &opt.cut_dirs, cmd_number },
132 { "debug", &opt.debug, cmd_boolean },
134 { "deleteafter", &opt.delete_after, cmd_boolean },
135 { "dirprefix", &opt.dir_prefix, cmd_directory },
136 { "dirstruct", NULL, cmd_spec_dirstruct },
137 { "dnscache", &opt.dns_cache, cmd_boolean },
138 { "dnstimeout", &opt.dns_timeout, cmd_time },
139 { "domains", &opt.domains, cmd_vector },
140 { "dotbytes", &opt.dot_bytes, cmd_bytes },
141 { "dotsinline", &opt.dots_in_line, cmd_number },
142 { "dotspacing", &opt.dot_spacing, cmd_number },
143 { "dotstyle", &opt.dot_style, cmd_string },
145 { "egdfile", &opt.egd_file, cmd_file },
147 { "excludedirectories", &opt.excludes, cmd_directory_vector },
148 { "excludedomains", &opt.exclude_domains, cmd_vector },
149 { "followftp", &opt.follow_ftp, cmd_boolean },
150 { "followtags", &opt.follow_tags, cmd_vector },
151 { "forcehtml", &opt.force_html, cmd_boolean },
152 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
153 { "ftppassword", &opt.ftp_passwd, cmd_string },
154 { "ftpproxy", &opt.ftp_proxy, cmd_string },
155 { "ftpuser", &opt.ftp_user, cmd_string },
156 { "glob", &opt.ftp_glob, cmd_boolean },
157 { "header", NULL, cmd_spec_header },
158 { "htmlextension", &opt.html_extension, cmd_boolean },
159 { "htmlify", NULL, cmd_spec_htmlify },
160 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
161 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
162 { "httppassword", &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 { "ignorecase", &opt.ignore_case, cmd_boolean },
167 { "ignorelength", &opt.ignore_length, cmd_boolean },
168 { "ignoretags", &opt.ignore_tags, cmd_vector },
169 { "includedirectories", &opt.includes, cmd_directory_vector },
171 { "inet4only", &opt.ipv4_only, cmd_boolean },
172 { "inet6only", &opt.ipv6_only, cmd_boolean },
174 { "input", &opt.input_filename, cmd_file },
175 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
176 { "killlonger", &opt.kill_longer, cmd_boolean },
177 { "limitrate", &opt.limit_rate, cmd_bytes },
178 { "loadcookies", &opt.cookies_input, cmd_file },
179 { "logfile", &opt.lfilename, cmd_file },
180 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
181 { "mirror", NULL, cmd_spec_mirror },
182 { "netrc", &opt.netrc, cmd_boolean },
183 { "noclobber", &opt.noclobber, cmd_boolean },
184 { "noparent", &opt.no_parent, cmd_boolean },
185 { "noproxy", &opt.no_proxy, cmd_vector },
186 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
187 { "outputdocument", &opt.output_document, cmd_file },
188 { "pagerequisites", &opt.page_requisites, cmd_boolean },
189 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
190 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
191 { "password", &opt.passwd, cmd_string },
192 { "postdata", &opt.post_data, cmd_string },
193 { "postfile", &opt.post_file_name, cmd_file },
194 { "preferfamily", NULL, cmd_spec_prefer_family },
195 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
197 { "privatekey", &opt.private_key, cmd_file },
198 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
200 { "progress", &opt.progress_type, cmd_spec_progress },
201 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
202 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
203 { "proxypassword", &opt.proxy_passwd, cmd_string },
204 { "proxyuser", &opt.proxy_user, cmd_string },
205 { "quiet", &opt.quiet, cmd_boolean },
206 { "quota", &opt.quota, cmd_bytes_sum },
208 { "randomfile", &opt.random_file, cmd_file },
210 { "randomwait", &opt.random_wait, cmd_boolean },
211 { "readtimeout", &opt.read_timeout, cmd_time },
212 { "reclevel", &opt.reclevel, cmd_number_inf },
213 { "recursive", NULL, cmd_spec_recursive },
214 { "referer", &opt.referer, cmd_string },
215 { "reject", &opt.rejects, cmd_vector },
216 { "relativeonly", &opt.relative_only, cmd_boolean },
217 { "removelisting", &opt.remove_listing, cmd_boolean },
218 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
219 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
220 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
221 { "robots", &opt.use_robots, cmd_boolean },
222 { "savecookies", &opt.cookies_output, cmd_file },
223 { "saveheaders", &opt.save_headers, cmd_boolean },
225 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
227 { "serverresponse", &opt.server_response, cmd_boolean },
228 { "spanhosts", &opt.spanhost, cmd_boolean },
229 { "spider", &opt.spider, cmd_boolean },
230 { "strictcomments", &opt.strict_comments, cmd_boolean },
231 { "timeout", NULL, cmd_spec_timeout },
232 { "timestamping", &opt.timestamping, cmd_boolean },
233 { "tries", &opt.ntry, cmd_number_inf },
234 { "useproxy", &opt.use_proxy, cmd_boolean },
235 { "user", &opt.user, cmd_string },
236 { "useragent", NULL, cmd_spec_useragent },
237 { "verbose", NULL, cmd_spec_verbose },
238 { "wait", &opt.wait, cmd_time },
239 { "waitretry", &opt.waitretry, cmd_time }
242 /* Look up CMDNAME in the commands[] and return its position in the
243 array. If CMDNAME is not found, return -1. */
246 command_by_name (const char *cmdname)
248 /* Use binary search for speed. Wget has ~100 commands, which
249 guarantees a worst case performance of 7 string comparisons. */
250 int lo = 0, hi = countof (commands) - 1;
254 int mid = (lo + hi) >> 1;
255 int cmp = strcasecmp (cmdname, commands[mid].name);
266 /* Reset the variables to default values. */
272 /* Most of the default values are 0 (and 0.0, NULL, and false).
273 Just reset everything, and fill in the non-zero values. Note
274 that initializing pointers to NULL this way is technically
275 illegal, but porting Wget to a machine where NULL is not all-zero
276 bit pattern will be the least of the implementors' worries. */
283 opt.add_hostdir = true;
287 opt.http_keep_alive = true;
288 opt.use_proxy = true;
289 tmp = getenv ("no_proxy");
291 opt.no_proxy = sepstring (tmp);
292 opt.allow_cache = true;
294 opt.read_timeout = 900;
295 opt.use_robots = true;
297 opt.remove_listing = true;
299 opt.dot_bytes = 1024;
300 opt.dot_spacing = 10;
301 opt.dots_in_line = 50;
303 opt.dns_cache = true;
307 opt.check_cert = true;
310 /* The default for file name restriction defaults to the OS type. */
311 #if !defined(WINDOWS) && !defined(__CYGWIN__)
312 opt.restrict_files_os = restrict_unix;
314 opt.restrict_files_os = restrict_windows;
316 opt.restrict_files_ctrl = true;
319 /* Return the user's home directory (strdup-ed), or NULL if none is
324 char *home = getenv ("HOME");
329 /* If HOME is not defined, try getting it from the password
331 struct passwd *pwd = getpwuid (getuid ());
332 if (!pwd || !pwd->pw_dir)
336 /* Under Windows, if $HOME isn't defined, use the directory where
337 `wget.exe' resides. */
342 return home ? xstrdup (home) : NULL;
345 /* Return the path to the user's .wgetrc. This is either the value of
346 `WGETRC' environment variable, or `$HOME/.wgetrc'.
348 If the `WGETRC' variable exists but the file does not exist, the
349 function will exit(). */
351 wgetrc_file_name (void)
356 /* Try the environment. */
357 env = getenv ("WGETRC");
360 if (!file_exists_p (env))
362 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
366 return xstrdup (env);
369 /* If that failed, try $HOME/.wgetrc. */
372 file = aprintf ("%s/.wgetrc", home);
376 /* Under Windows, if we still haven't found .wgetrc, look for the file
377 `wget.ini' in the directory where `wget.exe' resides; we do this for
378 backward compatibility with previous versions of Wget.
379 SYSTEM_WGETRC should not be defined under WINDOWS. */
380 if (!file || !file_exists_p (file))
386 file = aprintf ("%s/wget.ini", home);
392 if (!file_exists_p (file))
400 /* Return values of parse_line. */
408 static enum parse_line parse_line (const char *, char **, char **, int *);
409 static bool setval_internal (int, const char *, const char *);
411 /* Initialize variables from a wgetrc file. Returns zero (failure) if
412 there were errors in the file. */
415 run_wgetrc (const char *file)
422 fp = fopen (file, "rb");
425 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
426 file, strerror (errno));
427 return true; /* not a fatal error */
429 enable_tilde_expansion = true;
431 while ((line = read_whole_line (fp)) != NULL)
433 char *com = NULL, *val = NULL;
436 /* Parse the line. */
437 switch (parse_line (line, &com, &val, &comind))
440 /* If everything is OK, set the value. */
441 if (!setval_internal (comind, com, val))
443 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
444 exec_name, file, ln);
448 case line_syntax_error:
449 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
450 exec_name, file, ln);
453 case line_unknown_command:
454 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
455 exec_name, com, file, ln);
468 enable_tilde_expansion = false;
474 /* Initialize the defaults and run the system wgetrc and user's own
482 /* Load the hard-coded defaults. */
485 /* If SYSTEM_WGETRC is defined, use it. */
487 if (file_exists_p (SYSTEM_WGETRC))
488 ok &= run_wgetrc (SYSTEM_WGETRC);
490 /* Override it with your own, if one exists. */
491 file = wgetrc_file_name ();
494 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
495 something like realpath() before comparing them with `strcmp' */
497 if (!strcmp (file, SYSTEM_WGETRC))
499 fprintf (stderr, _("\
500 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
505 ok &= run_wgetrc (file);
507 /* If there were errors processing either `.wgetrc', abort. */
515 /* Remove dashes and underscores from S, modifying S in the
521 char *t = s; /* t - tortoise */
522 char *h = s; /* h - hare */
524 if (*h == '_' || *h == '-')
531 /* Parse the line pointed by line, with the syntax:
532 <sp>* command <sp>* = <sp>* value <sp>*
533 Uses malloc to allocate space for command and value.
535 Returns one of line_ok, line_empty, line_syntax_error, or
536 line_unknown_command.
538 In case of line_ok, *COM and *VAL point to freshly allocated
539 strings, and *COMIND points to com's index. In case of error or
540 empty line, their values are unmodified. */
542 static enum parse_line
543 parse_line (const char *line, char **com, char **val, int *comind)
546 const char *end = line + strlen (line);
547 const char *cmdstart, *cmdend;
548 const char *valstart, *valend;
553 /* Skip leading and trailing whitespace. */
554 while (*line && ISSPACE (*line))
556 while (end > line && ISSPACE (end[-1]))
559 /* Skip empty lines and comments. */
560 if (!*line || *line == '#')
566 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
570 /* Skip '=', as well as any space before or after it. */
571 while (p < end && ISSPACE (*p))
573 if (p == end || *p != '=')
574 return line_syntax_error;
576 while (p < end && ISSPACE (*p))
582 /* The syntax is valid (even though the command might not be). Fill
583 in the command name and value. */
584 *com = strdupdelim (cmdstart, cmdend);
585 *val = strdupdelim (valstart, valend);
587 /* The line now known to be syntactically correct. Check whether
588 the command is valid. */
589 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
591 ind = command_by_name (cmdcopy);
593 return line_unknown_command;
595 /* Report success to the caller. */
600 /* Run commands[comind].action. */
603 setval_internal (int comind, const char *com, const char *val)
605 assert (0 <= comind && comind < countof (commands));
606 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
607 return commands[comind].action (com, val, commands[comind].place);
610 /* Run command COM with value VAL. If running the command produces an
611 error, report the error and exit.
613 This is intended to be called from main() to modify Wget's behavior
614 through command-line switches. Since COM is hard-coded in main(),
615 it is not canonicalized, and this aborts when COM is not found.
617 If COMIND's are exported to init.h, this function will be changed
618 to accept COMIND directly. */
621 setoptval (const char *com, const char *val, const char *optname)
623 /* Prepend "--" to OPTNAME. */
624 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
627 strcpy (dd_optname + 2, optname);
629 assert (val != NULL);
630 if (!setval_internal (command_by_name (com), dd_optname, val))
634 /* Parse OPT into command and value and run it. For example,
635 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
636 This is used by the `--execute' flag in main.c. */
639 run_command (const char *opt)
643 switch (parse_line (opt, &com, &val, &comind))
646 if (!setval_internal (comind, com, val))
652 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
658 /* Generic helper functions, for use with `commands'. */
660 /* Forward declarations: */
665 static bool decode_string (const char *, const struct decode_item *, int, int *);
666 static bool simple_atoi (const char *, const char *, int *);
667 static bool simple_atof (const char *, const char *, double *);
669 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
671 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
672 && TOLOWER((p)[1]) == (c1) \
675 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
676 && TOLOWER((p)[1]) == (c1) \
677 && TOLOWER((p)[2]) == (c2) \
681 /* Store the boolean value from VAL to PLACE. COM is ignored,
682 except for error messages. */
684 cmd_boolean (const char *com, const char *val, void *place)
688 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
689 /* "on", "yes" and "1" mean true. */
691 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
692 /* "off", "no" and "0" mean false. */
697 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
698 exec_name, com, val);
702 *(bool *) place = value;
706 /* Set the non-negative integer value from VAL to PLACE. With
707 incorrect specification, the number remains unchanged. */
709 cmd_number (const char *com, const char *val, void *place)
711 if (!simple_atoi (val, val + strlen (val), place)
712 || *(int *) place < 0)
714 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
715 exec_name, com, val);
721 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
723 cmd_number_inf (const char *com, const char *val, void *place)
725 if (!strcasecmp (val, "inf"))
730 return cmd_number (com, val, place);
733 /* Copy (strdup) the string at COM to a new location and place a
734 pointer to *PLACE. */
736 cmd_string (const char *com, const char *val, void *place)
738 char **pstring = (char **)place;
740 xfree_null (*pstring);
741 *pstring = xstrdup (val);
746 # define ISSEP(c) ((c) == '/')
748 # define ISSEP(c) ((c) == '/' || (c) == '\\')
751 /* Like the above, but handles tilde-expansion when reading a user's
752 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
753 gets expanded to the user's home directory. */
755 cmd_file (const char *com, const char *val, void *place)
757 char **pstring = (char **)place;
759 xfree_null (*pstring);
761 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
763 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
766 *pstring = xstrdup (val);
771 char *home = home_dir ();
775 homelen = strlen (home);
776 while (homelen && ISSEP (home[homelen - 1]))
777 home[--homelen] = '\0';
779 /* Skip the leading "~/". */
780 for (++val; ISSEP (*val); val++)
783 *pstring = concat_strings (home, "/", val, (char *) 0);
787 /* Convert "\" to "/". */
790 for (s = *pstring; *s; s++)
798 /* Like cmd_file, but strips trailing '/' characters. */
800 cmd_directory (const char *com, const char *val, void *place)
804 /* Call cmd_file() for tilde expansion and separator
805 canonicalization (backslash -> slash under Windows). These
806 things should perhaps be in a separate function. */
807 if (!cmd_file (com, val, place))
812 while (t > s && *--t == '/')
818 /* Split VAL by space to a vector of values, and append those values
819 to vector pointed to by the PLACE argument. If VAL is empty, the
820 PLACE vector is cleared instead. */
823 cmd_vector (const char *com, const char *val, void *place)
825 char ***pvec = (char ***)place;
828 *pvec = merge_vecs (*pvec, sepstring (val));
838 cmd_directory_vector (const char *com, const char *val, void *place)
840 char ***pvec = (char ***)place;
844 /* Strip the trailing slashes from directories. */
847 seps = sepstring (val);
848 for (t = seps; t && *t; t++)
850 int len = strlen (*t);
851 /* Skip degenerate case of root directory. */
854 if ((*t)[len - 1] == '/')
855 (*t)[len - 1] = '\0';
858 *pvec = merge_vecs (*pvec, seps);
868 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
869 "100k" or "2.5G" to a floating point number. */
872 parse_bytes_helper (const char *val, double *result)
875 const char *end = val + strlen (val);
877 /* Check for "inf". */
878 if (0 == strcmp (val, "inf"))
884 /* Strip trailing whitespace. */
885 while (val < end && ISSPACE (end[-1]))
890 switch (TOLOWER (end[-1]))
893 --end, mult = 1024.0;
896 --end, mult = 1048576.0;
899 --end, mult = 1073741824.0;
902 --end, mult = 1099511627776.0;
905 /* Not a recognized suffix: assume it's a digit. (If not,
906 simple_atof will raise an error.) */
910 /* Skip leading and trailing whitespace. */
911 while (val < end && ISSPACE (*val))
913 while (val < end && ISSPACE (end[-1]))
918 if (!simple_atof (val, end, &number) || number < 0)
921 *result = number * mult;
925 /* Parse VAL as a number and set its value to PLACE (which should
928 By default, the value is assumed to be in bytes. If "K", "M", or
929 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
930 1<<30, respectively. Floating point values are allowed and are
931 cast to integer before use. The idea is to be able to use things
932 like 1.5k instead of "1536".
934 The string "inf" is returned as 0.
936 In case of error, false is returned and memory pointed to by PLACE
937 remains unmodified. */
940 cmd_bytes (const char *com, const char *val, void *place)
943 if (!parse_bytes_helper (val, &byte_value))
945 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
946 exec_name, com, val);
949 *(wgint *)place = (wgint)byte_value;
953 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
954 SIZE_SUM. It works by converting the string to double, therefore
955 working with values up to 2^53-1 without loss of precision. This
956 value (8192 TB) is large enough to serve for a while. */
959 cmd_bytes_sum (const char *com, const char *val, void *place)
962 if (!parse_bytes_helper (val, &byte_value))
964 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
965 exec_name, com, val);
968 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
972 /* Store the value of VAL to *OUT. The value is a time period, by
973 default expressed in seconds, but also accepting suffixes "m", "h",
974 "d", and "w" for minutes, hours, days, and weeks respectively. */
977 cmd_time (const char *com, const char *val, void *place)
980 const char *end = val + strlen (val);
982 /* Strip trailing whitespace. */
983 while (val < end && ISSPACE (end[-1]))
989 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
990 exec_name, com, val);
994 switch (TOLOWER (end[-1]))
997 --end, mult = 1; /* seconds */
1000 --end, mult = 60; /* minutes */
1003 --end, mult = 3600; /* hours */
1006 --end, mult = 86400.0; /* days */
1009 --end, mult = 604800.0; /* weeks */
1012 /* Not a recognized suffix: assume it belongs to the number.
1013 (If not, simple_atof will raise an error.) */
1017 /* Skip leading and trailing whitespace. */
1018 while (val < end && ISSPACE (*val))
1020 while (val < end && ISSPACE (end[-1]))
1025 if (!simple_atof (val, end, &number))
1028 *(double *)place = number * mult;
1034 cmd_cert_type (const char *com, const char *val, void *place)
1036 static const struct decode_item choices[] = {
1037 { "pem", keyfile_pem },
1038 { "der", keyfile_asn1 },
1039 { "asn1", keyfile_asn1 },
1041 int ok = decode_string (val, choices, countof (choices), place);
1043 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1048 /* Specialized helper functions, used by `commands' to handle some
1049 options specially. */
1051 static bool check_user_specified_header (const char *);
1054 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1056 if (!cmd_boolean (com, val, &opt.dirstruct))
1058 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1059 must be affected inversely. */
1061 opt.no_dirstruct = false;
1063 opt.no_dirstruct = true;
1068 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1070 /* Empty value means reset the list of headers. */
1073 free_vec (opt.user_headers);
1074 opt.user_headers = NULL;
1078 if (!check_user_specified_header (val))
1080 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1081 exec_name, com, val);
1084 opt.user_headers = vec_append (opt.user_headers, val);
1089 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1091 int flag = cmd_boolean (com, val, &opt.htmlify);
1092 if (flag && !opt.htmlify)
1093 opt.remove_listing = false;
1097 /* Set the "mirror" mode. It means: recursive download, timestamping,
1098 no limit on max. recursion depth, and don't remove listings. */
1101 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1105 if (!cmd_boolean (com, val, &mirror))
1109 opt.recursive = true;
1110 if (!opt.no_dirstruct)
1111 opt.dirstruct = true;
1112 opt.timestamping = true;
1113 opt.reclevel = INFINITE_RECURSION;
1114 opt.remove_listing = false;
1119 /* Validate --prefer-family and set the choice. Allowed values are
1120 "IPv4", "IPv6", and "none". */
1123 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1125 static const struct decode_item choices[] = {
1126 { "IPv4", prefer_ipv4 },
1127 { "IPv6", prefer_ipv6 },
1128 { "none", prefer_none },
1130 int ok = decode_string (val, choices, countof (choices),
1131 (int *) &opt.prefer_family);
1133 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1137 /* Set progress.type to VAL, but verify that it's a valid progress
1138 implementation before that. */
1141 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1143 if (!valid_progress_implementation_p (val))
1145 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1146 exec_name, com, val);
1149 xfree_null (opt.progress_type);
1151 /* Don't call set_progress_implementation here. It will be called
1152 in main() when it becomes clear what the log output is. */
1153 opt.progress_type = xstrdup (val);
1157 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1158 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1162 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1164 if (!cmd_boolean (com, val, &opt.recursive))
1168 if (opt.recursive && !opt.no_dirstruct)
1169 opt.dirstruct = true;
1175 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1177 int restrict_os = opt.restrict_files_os;
1178 int restrict_ctrl = opt.restrict_files_ctrl;
1180 const char *end = strchr (val, ',');
1182 end = val + strlen (val);
1184 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1186 if (VAL_IS ("unix"))
1187 restrict_os = restrict_unix;
1188 else if (VAL_IS ("windows"))
1189 restrict_os = restrict_windows;
1190 else if (VAL_IS ("nocontrol"))
1196 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1197 exec_name, com, val);
1205 if (!strcmp (end + 1, "nocontrol"))
1206 restrict_ctrl = false;
1211 opt.restrict_files_os = restrict_os;
1212 opt.restrict_files_ctrl = restrict_ctrl;
1218 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1220 static const struct decode_item choices[] = {
1221 { "auto", secure_protocol_auto },
1222 { "sslv2", secure_protocol_sslv2 },
1223 { "sslv3", secure_protocol_sslv3 },
1224 { "tlsv1", secure_protocol_tlsv1 },
1226 int ok = decode_string (val, choices, countof (choices), place);
1228 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1233 /* Set all three timeout values. */
1236 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1239 if (!cmd_time (com, val, &value))
1241 opt.read_timeout = value;
1242 opt.connect_timeout = value;
1243 opt.dns_timeout = value;
1248 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1250 /* Disallow embedded newlines. */
1251 if (strchr (val, '\n'))
1253 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1254 exec_name, com, val);
1257 xfree_null (opt.useragent);
1258 opt.useragent = xstrdup (val);
1262 /* The "verbose" option cannot be cmd_boolean because the variable is
1263 not bool -- it's of type int (-1 means uninitialized because of
1264 some random hackery for disallowing -q -v). */
1267 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1270 if (cmd_boolean (com, val, &flag))
1278 /* Miscellaneous useful routines. */
1280 /* A very simple atoi clone, more useful than atoi because it works on
1281 delimited strings, and has error reportage. Returns true on success,
1282 false on failure. If successful, stores result to *DEST. */
1285 simple_atoi (const char *beg, const char *end, int *dest)
1288 bool negative = false;
1289 const char *p = beg;
1291 while (p < end && ISSPACE (*p))
1293 if (p < end && (*p == '-' || *p == '+'))
1295 negative = (*p == '-');
1301 /* Read negative numbers in a separate loop because the most
1302 negative integer cannot be represented as a positive number. */
1305 for (; p < end && ISDIGIT (*p); p++)
1307 int next = (10 * result) + (*p - '0');
1309 return false; /* overflow */
1313 for (; p < end && ISDIGIT (*p); p++)
1315 int next = (10 * result) - (*p - '0');
1317 return false; /* underflow */
1328 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1329 doesn't handle exponential notation. Returns true on success,
1330 false on failure. In case of success, stores its result to
1334 simple_atof (const char *beg, const char *end, double *dest)
1338 bool negative = false;
1339 bool seen_dot = false;
1340 bool seen_digit = false;
1343 const char *p = beg;
1345 while (p < end && ISSPACE (*p))
1347 if (p < end && (*p == '-' || *p == '+'))
1349 negative = (*p == '-');
1353 for (; p < end; p++)
1359 result = (10 * result) + (ch - '0');
1361 result += (ch - '0') / (divider *= 10);
1383 /* Verify that the user-specified header in S is valid. It must
1384 contain a colon preceded by non-white-space characters and must not
1385 contain newlines. */
1388 check_user_specified_header (const char *s)
1392 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1393 /* The header MUST contain `:' preceded by at least one
1394 non-whitespace character. */
1395 if (*p != ':' || p == s)
1397 /* The header MUST NOT contain newlines. */
1398 if (strchr (s, '\n'))
1403 /* Decode VAL into a number, according to ITEMS. */
1406 decode_string (const char *val, const struct decode_item *items, int itemcount,
1410 for (i = 0; i < itemcount; i++)
1411 if (0 == strcasecmp (val, items[i].name))
1413 *place = items[i].code;
1420 void cleanup_html_url (void);
1423 /* Free the memory allocated by global variables. */
1427 /* Free external resources, close files, etc. */
1430 fclose (output_stream);
1431 /* No need to check for error because Wget flushes its output (and
1432 checks for errors) after any data arrives. */
1434 /* We're exiting anyway so there's no real need to call free()
1435 hundreds of times. Skipping the frees will make Wget exit
1438 However, when detecting leaks, it's crucial to free() everything
1439 because then you can find the real leaks, i.e. the allocated
1440 memory which grows with the size of the program. */
1446 cleanup_html_url ();
1451 extern acc_t *netrc_list;
1452 free_netrc (netrc_list);
1454 xfree_null (opt.lfilename);
1455 xfree_null (opt.dir_prefix);
1456 xfree_null (opt.input_filename);
1457 xfree_null (opt.output_document);
1458 free_vec (opt.accepts);
1459 free_vec (opt.rejects);
1460 free_vec (opt.excludes);
1461 free_vec (opt.includes);
1462 free_vec (opt.domains);
1463 free_vec (opt.follow_tags);
1464 free_vec (opt.ignore_tags);
1465 xfree_null (opt.progress_type);
1466 xfree_null (opt.ftp_user);
1467 xfree_null (opt.ftp_passwd);
1468 xfree_null (opt.ftp_proxy);
1469 xfree_null (opt.https_proxy);
1470 xfree_null (opt.http_proxy);
1471 free_vec (opt.no_proxy);
1472 xfree_null (opt.useragent);
1473 xfree_null (opt.referer);
1474 xfree_null (opt.http_user);
1475 xfree_null (opt.http_passwd);
1476 free_vec (opt.user_headers);
1478 xfree_null (opt.cert_file);
1479 xfree_null (opt.private_key);
1480 xfree_null (opt.ca_directory);
1481 xfree_null (opt.ca_cert);
1482 xfree_null (opt.random_file);
1483 xfree_null (opt.egd_file);
1485 xfree_null (opt.bind_address);
1486 xfree_null (opt.cookies_input);
1487 xfree_null (opt.cookies_output);
1488 xfree_null (opt.user);
1489 xfree_null (opt.passwd);
1490 #endif /* DEBUG_MALLOC */