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 { "contentdisposition", &opt.content_disposition, cmd_boolean },
130 { "cookies", &opt.cookies, cmd_boolean },
131 { "cutdirs", &opt.cut_dirs, cmd_number },
133 { "debug", &opt.debug, cmd_boolean },
135 { "deleteafter", &opt.delete_after, cmd_boolean },
136 { "dirprefix", &opt.dir_prefix, cmd_directory },
137 { "dirstruct", NULL, cmd_spec_dirstruct },
138 { "dnscache", &opt.dns_cache, cmd_boolean },
139 { "dnstimeout", &opt.dns_timeout, cmd_time },
140 { "domains", &opt.domains, cmd_vector },
141 { "dotbytes", &opt.dot_bytes, cmd_bytes },
142 { "dotsinline", &opt.dots_in_line, cmd_number },
143 { "dotspacing", &opt.dot_spacing, cmd_number },
144 { "dotstyle", &opt.dot_style, cmd_string },
146 { "egdfile", &opt.egd_file, cmd_file },
148 { "excludedirectories", &opt.excludes, cmd_directory_vector },
149 { "excludedomains", &opt.exclude_domains, cmd_vector },
150 { "followftp", &opt.follow_ftp, cmd_boolean },
151 { "followtags", &opt.follow_tags, cmd_vector },
152 { "forcehtml", &opt.force_html, cmd_boolean },
153 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
154 { "ftppassword", &opt.ftp_passwd, cmd_string },
155 { "ftpproxy", &opt.ftp_proxy, cmd_string },
156 { "ftpuser", &opt.ftp_user, 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 }, /* deprecated */
163 { "httppassword", &opt.http_passwd, cmd_string },
164 { "httpproxy", &opt.http_proxy, cmd_string },
165 { "httpsproxy", &opt.https_proxy, cmd_string },
166 { "httpuser", &opt.http_user, cmd_string },
167 { "ignorecase", &opt.ignore_case, cmd_boolean },
168 { "ignorelength", &opt.ignore_length, cmd_boolean },
169 { "ignoretags", &opt.ignore_tags, cmd_vector },
170 { "includedirectories", &opt.includes, cmd_directory_vector },
172 { "inet4only", &opt.ipv4_only, cmd_boolean },
173 { "inet6only", &opt.ipv6_only, cmd_boolean },
175 { "input", &opt.input_filename, cmd_file },
176 { "keepsessioncookies", &opt.keep_session_cookies, 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;
318 opt.content_disposition = true;
321 /* Return the user's home directory (strdup-ed), or NULL if none is
326 char *home = getenv ("HOME");
331 /* If HOME is not defined, try getting it from the password
333 struct passwd *pwd = getpwuid (getuid ());
334 if (!pwd || !pwd->pw_dir)
338 /* Under Windows, if $HOME isn't defined, use the directory where
339 `wget.exe' resides. */
344 return home ? xstrdup (home) : NULL;
347 /* Return the path to the user's .wgetrc. This is either the value of
348 `WGETRC' environment variable, or `$HOME/.wgetrc'.
350 If the `WGETRC' variable exists but the file does not exist, the
351 function will exit(). */
353 wgetrc_file_name (void)
358 /* Try the environment. */
359 env = getenv ("WGETRC");
362 if (!file_exists_p (env))
364 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
368 return xstrdup (env);
371 /* If that failed, try $HOME/.wgetrc. */
374 file = aprintf ("%s/.wgetrc", home);
378 /* Under Windows, if we still haven't found .wgetrc, look for the file
379 `wget.ini' in the directory where `wget.exe' resides; we do this for
380 backward compatibility with previous versions of Wget.
381 SYSTEM_WGETRC should not be defined under WINDOWS. */
382 if (!file || !file_exists_p (file))
388 file = aprintf ("%s/wget.ini", home);
394 if (!file_exists_p (file))
402 /* Return values of parse_line. */
410 static enum parse_line parse_line (const char *, char **, char **, int *);
411 static bool setval_internal (int, const char *, const char *);
413 /* Initialize variables from a wgetrc file. Returns zero (failure) if
414 there were errors in the file. */
417 run_wgetrc (const char *file)
424 fp = fopen (file, "rb");
427 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
428 file, strerror (errno));
429 return true; /* not a fatal error */
431 enable_tilde_expansion = true;
433 while ((line = read_whole_line (fp)) != NULL)
435 char *com = NULL, *val = NULL;
438 /* Parse the line. */
439 switch (parse_line (line, &com, &val, &comind))
442 /* If everything is OK, set the value. */
443 if (!setval_internal (comind, com, val))
445 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
446 exec_name, file, ln);
450 case line_syntax_error:
451 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
452 exec_name, file, ln);
455 case line_unknown_command:
456 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
457 exec_name, com, file, ln);
470 enable_tilde_expansion = false;
476 /* Initialize the defaults and run the system wgetrc and user's own
484 /* Load the hard-coded defaults. */
487 /* If SYSTEM_WGETRC is defined, use it. */
489 if (file_exists_p (SYSTEM_WGETRC))
490 ok &= run_wgetrc (SYSTEM_WGETRC);
492 /* Override it with your own, if one exists. */
493 file = wgetrc_file_name ();
496 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
497 something like realpath() before comparing them with `strcmp' */
499 if (!strcmp (file, SYSTEM_WGETRC))
501 fprintf (stderr, _("\
502 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
507 ok &= run_wgetrc (file);
509 /* If there were errors processing either `.wgetrc', abort. */
517 /* Remove dashes and underscores from S, modifying S in the
523 char *t = s; /* t - tortoise */
524 char *h = s; /* h - hare */
526 if (*h == '_' || *h == '-')
533 /* Parse the line pointed by line, with the syntax:
534 <sp>* command <sp>* = <sp>* value <sp>*
535 Uses malloc to allocate space for command and value.
537 Returns one of line_ok, line_empty, line_syntax_error, or
538 line_unknown_command.
540 In case of line_ok, *COM and *VAL point to freshly allocated
541 strings, and *COMIND points to com's index. In case of error or
542 empty line, their values are unmodified. */
544 static enum parse_line
545 parse_line (const char *line, char **com, char **val, int *comind)
548 const char *end = line + strlen (line);
549 const char *cmdstart, *cmdend;
550 const char *valstart, *valend;
555 /* Skip leading and trailing whitespace. */
556 while (*line && ISSPACE (*line))
558 while (end > line && ISSPACE (end[-1]))
561 /* Skip empty lines and comments. */
562 if (!*line || *line == '#')
568 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
572 /* Skip '=', as well as any space before or after it. */
573 while (p < end && ISSPACE (*p))
575 if (p == end || *p != '=')
576 return line_syntax_error;
578 while (p < end && ISSPACE (*p))
584 /* The syntax is valid (even though the command might not be). Fill
585 in the command name and value. */
586 *com = strdupdelim (cmdstart, cmdend);
587 *val = strdupdelim (valstart, valend);
589 /* The line now known to be syntactically correct. Check whether
590 the command is valid. */
591 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
593 ind = command_by_name (cmdcopy);
595 return line_unknown_command;
597 /* Report success to the caller. */
602 /* Run commands[comind].action. */
605 setval_internal (int comind, const char *com, const char *val)
607 assert (0 <= comind && comind < countof (commands));
608 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
609 return commands[comind].action (com, val, commands[comind].place);
612 /* Run command COM with value VAL. If running the command produces an
613 error, report the error and exit.
615 This is intended to be called from main() to modify Wget's behavior
616 through command-line switches. Since COM is hard-coded in main(),
617 it is not canonicalized, and this aborts when COM is not found.
619 If COMIND's are exported to init.h, this function will be changed
620 to accept COMIND directly. */
623 setoptval (const char *com, const char *val, const char *optname)
625 /* Prepend "--" to OPTNAME. */
626 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
629 strcpy (dd_optname + 2, optname);
631 assert (val != NULL);
632 if (!setval_internal (command_by_name (com), dd_optname, val))
636 /* Parse OPT into command and value and run it. For example,
637 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
638 This is used by the `--execute' flag in main.c. */
641 run_command (const char *opt)
645 switch (parse_line (opt, &com, &val, &comind))
648 if (!setval_internal (comind, com, val))
654 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
660 /* Generic helper functions, for use with `commands'. */
662 /* Forward declarations: */
667 static bool decode_string (const char *, const struct decode_item *, int, int *);
668 static bool simple_atoi (const char *, const char *, int *);
669 static bool simple_atof (const char *, const char *, double *);
671 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
673 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
674 && TOLOWER((p)[1]) == (c1) \
677 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
678 && TOLOWER((p)[1]) == (c1) \
679 && TOLOWER((p)[2]) == (c2) \
683 /* Store the boolean value from VAL to PLACE. COM is ignored,
684 except for error messages. */
686 cmd_boolean (const char *com, const char *val, void *place)
690 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
691 /* "on", "yes" and "1" mean true. */
693 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
694 /* "off", "no" and "0" mean false. */
699 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
700 exec_name, com, val);
704 *(bool *) place = value;
708 /* Set the non-negative integer value from VAL to PLACE. With
709 incorrect specification, the number remains unchanged. */
711 cmd_number (const char *com, const char *val, void *place)
713 if (!simple_atoi (val, val + strlen (val), place)
714 || *(int *) place < 0)
716 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
717 exec_name, com, val);
723 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
725 cmd_number_inf (const char *com, const char *val, void *place)
727 if (!strcasecmp (val, "inf"))
732 return cmd_number (com, val, place);
735 /* Copy (strdup) the string at COM to a new location and place a
736 pointer to *PLACE. */
738 cmd_string (const char *com, const char *val, void *place)
740 char **pstring = (char **)place;
742 xfree_null (*pstring);
743 *pstring = xstrdup (val);
748 # define ISSEP(c) ((c) == '/')
750 # define ISSEP(c) ((c) == '/' || (c) == '\\')
753 /* Like the above, but handles tilde-expansion when reading a user's
754 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
755 gets expanded to the user's home directory. */
757 cmd_file (const char *com, const char *val, void *place)
759 char **pstring = (char **)place;
761 xfree_null (*pstring);
763 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
765 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
768 *pstring = xstrdup (val);
773 char *home = home_dir ();
777 homelen = strlen (home);
778 while (homelen && ISSEP (home[homelen - 1]))
779 home[--homelen] = '\0';
781 /* Skip the leading "~/". */
782 for (++val; ISSEP (*val); val++)
785 *pstring = concat_strings (home, "/", val, (char *) 0);
789 /* Convert "\" to "/". */
792 for (s = *pstring; *s; s++)
800 /* Like cmd_file, but strips trailing '/' characters. */
802 cmd_directory (const char *com, const char *val, void *place)
806 /* Call cmd_file() for tilde expansion and separator
807 canonicalization (backslash -> slash under Windows). These
808 things should perhaps be in a separate function. */
809 if (!cmd_file (com, val, place))
814 while (t > s && *--t == '/')
820 /* Split VAL by space to a vector of values, and append those values
821 to vector pointed to by the PLACE argument. If VAL is empty, the
822 PLACE vector is cleared instead. */
825 cmd_vector (const char *com, const char *val, void *place)
827 char ***pvec = (char ***)place;
830 *pvec = merge_vecs (*pvec, sepstring (val));
840 cmd_directory_vector (const char *com, const char *val, void *place)
842 char ***pvec = (char ***)place;
846 /* Strip the trailing slashes from directories. */
849 seps = sepstring (val);
850 for (t = seps; t && *t; t++)
852 int len = strlen (*t);
853 /* Skip degenerate case of root directory. */
856 if ((*t)[len - 1] == '/')
857 (*t)[len - 1] = '\0';
860 *pvec = merge_vecs (*pvec, seps);
870 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
871 "100k" or "2.5G" to a floating point number. */
874 parse_bytes_helper (const char *val, double *result)
877 const char *end = val + strlen (val);
879 /* Check for "inf". */
880 if (0 == strcmp (val, "inf"))
886 /* Strip trailing whitespace. */
887 while (val < end && ISSPACE (end[-1]))
892 switch (TOLOWER (end[-1]))
895 --end, mult = 1024.0;
898 --end, mult = 1048576.0;
901 --end, mult = 1073741824.0;
904 --end, mult = 1099511627776.0;
907 /* Not a recognized suffix: assume it's a digit. (If not,
908 simple_atof will raise an error.) */
912 /* Skip leading and trailing whitespace. */
913 while (val < end && ISSPACE (*val))
915 while (val < end && ISSPACE (end[-1]))
920 if (!simple_atof (val, end, &number) || number < 0)
923 *result = number * mult;
927 /* Parse VAL as a number and set its value to PLACE (which should
930 By default, the value is assumed to be in bytes. If "K", "M", or
931 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
932 1<<30, respectively. Floating point values are allowed and are
933 cast to integer before use. The idea is to be able to use things
934 like 1.5k instead of "1536".
936 The string "inf" is returned as 0.
938 In case of error, false is returned and memory pointed to by PLACE
939 remains unmodified. */
942 cmd_bytes (const char *com, const char *val, void *place)
945 if (!parse_bytes_helper (val, &byte_value))
947 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
948 exec_name, com, val);
951 *(wgint *)place = (wgint)byte_value;
955 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
956 SIZE_SUM. It works by converting the string to double, therefore
957 working with values up to 2^53-1 without loss of precision. This
958 value (8192 TB) is large enough to serve for a while. */
961 cmd_bytes_sum (const char *com, const char *val, void *place)
964 if (!parse_bytes_helper (val, &byte_value))
966 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
967 exec_name, com, val);
970 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
974 /* Store the value of VAL to *OUT. The value is a time period, by
975 default expressed in seconds, but also accepting suffixes "m", "h",
976 "d", and "w" for minutes, hours, days, and weeks respectively. */
979 cmd_time (const char *com, const char *val, void *place)
982 const char *end = val + strlen (val);
984 /* Strip trailing whitespace. */
985 while (val < end && ISSPACE (end[-1]))
991 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
992 exec_name, com, val);
996 switch (TOLOWER (end[-1]))
999 --end, mult = 1; /* seconds */
1002 --end, mult = 60; /* minutes */
1005 --end, mult = 3600; /* hours */
1008 --end, mult = 86400.0; /* days */
1011 --end, mult = 604800.0; /* weeks */
1014 /* Not a recognized suffix: assume it belongs to the number.
1015 (If not, simple_atof will raise an error.) */
1019 /* Skip leading and trailing whitespace. */
1020 while (val < end && ISSPACE (*val))
1022 while (val < end && ISSPACE (end[-1]))
1027 if (!simple_atof (val, end, &number))
1030 *(double *)place = number * mult;
1036 cmd_cert_type (const char *com, const char *val, void *place)
1038 static const struct decode_item choices[] = {
1039 { "pem", keyfile_pem },
1040 { "der", keyfile_asn1 },
1041 { "asn1", keyfile_asn1 },
1043 int ok = decode_string (val, choices, countof (choices), place);
1045 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1050 /* Specialized helper functions, used by `commands' to handle some
1051 options specially. */
1053 static bool check_user_specified_header (const char *);
1056 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1058 if (!cmd_boolean (com, val, &opt.dirstruct))
1060 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1061 must be affected inversely. */
1063 opt.no_dirstruct = false;
1065 opt.no_dirstruct = true;
1070 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1072 /* Empty value means reset the list of headers. */
1075 free_vec (opt.user_headers);
1076 opt.user_headers = NULL;
1080 if (!check_user_specified_header (val))
1082 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1083 exec_name, com, val);
1086 opt.user_headers = vec_append (opt.user_headers, val);
1091 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1093 int flag = cmd_boolean (com, val, &opt.htmlify);
1094 if (flag && !opt.htmlify)
1095 opt.remove_listing = false;
1099 /* Set the "mirror" mode. It means: recursive download, timestamping,
1100 no limit on max. recursion depth, and don't remove listings. */
1103 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1107 if (!cmd_boolean (com, val, &mirror))
1111 opt.recursive = true;
1112 if (!opt.no_dirstruct)
1113 opt.dirstruct = true;
1114 opt.timestamping = true;
1115 opt.reclevel = INFINITE_RECURSION;
1116 opt.remove_listing = false;
1121 /* Validate --prefer-family and set the choice. Allowed values are
1122 "IPv4", "IPv6", and "none". */
1125 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1127 static const struct decode_item choices[] = {
1128 { "IPv4", prefer_ipv4 },
1129 { "IPv6", prefer_ipv6 },
1130 { "none", prefer_none },
1132 int ok = decode_string (val, choices, countof (choices),
1133 (int *) &opt.prefer_family);
1135 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1139 /* Set progress.type to VAL, but verify that it's a valid progress
1140 implementation before that. */
1143 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1145 if (!valid_progress_implementation_p (val))
1147 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1148 exec_name, com, val);
1151 xfree_null (opt.progress_type);
1153 /* Don't call set_progress_implementation here. It will be called
1154 in main() when it becomes clear what the log output is. */
1155 opt.progress_type = xstrdup (val);
1159 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1160 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1164 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1166 if (!cmd_boolean (com, val, &opt.recursive))
1170 if (opt.recursive && !opt.no_dirstruct)
1171 opt.dirstruct = true;
1177 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1179 int restrict_os = opt.restrict_files_os;
1180 int restrict_ctrl = opt.restrict_files_ctrl;
1182 const char *end = strchr (val, ',');
1184 end = val + strlen (val);
1186 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1188 if (VAL_IS ("unix"))
1189 restrict_os = restrict_unix;
1190 else if (VAL_IS ("windows"))
1191 restrict_os = restrict_windows;
1192 else if (VAL_IS ("nocontrol"))
1198 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1199 exec_name, com, val);
1207 if (!strcmp (end + 1, "nocontrol"))
1208 restrict_ctrl = false;
1213 opt.restrict_files_os = restrict_os;
1214 opt.restrict_files_ctrl = restrict_ctrl;
1220 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1222 static const struct decode_item choices[] = {
1223 { "auto", secure_protocol_auto },
1224 { "sslv2", secure_protocol_sslv2 },
1225 { "sslv3", secure_protocol_sslv3 },
1226 { "tlsv1", secure_protocol_tlsv1 },
1228 int ok = decode_string (val, choices, countof (choices), place);
1230 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1235 /* Set all three timeout values. */
1238 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1241 if (!cmd_time (com, val, &value))
1243 opt.read_timeout = value;
1244 opt.connect_timeout = value;
1245 opt.dns_timeout = value;
1250 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1252 /* Disallow embedded newlines. */
1253 if (strchr (val, '\n'))
1255 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1256 exec_name, com, val);
1259 xfree_null (opt.useragent);
1260 opt.useragent = xstrdup (val);
1264 /* The "verbose" option cannot be cmd_boolean because the variable is
1265 not bool -- it's of type int (-1 means uninitialized because of
1266 some random hackery for disallowing -q -v). */
1269 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1272 if (cmd_boolean (com, val, &flag))
1280 /* Miscellaneous useful routines. */
1282 /* A very simple atoi clone, more useful than atoi because it works on
1283 delimited strings, and has error reportage. Returns true on success,
1284 false on failure. If successful, stores result to *DEST. */
1287 simple_atoi (const char *beg, const char *end, int *dest)
1290 bool negative = false;
1291 const char *p = beg;
1293 while (p < end && ISSPACE (*p))
1295 if (p < end && (*p == '-' || *p == '+'))
1297 negative = (*p == '-');
1303 /* Read negative numbers in a separate loop because the most
1304 negative integer cannot be represented as a positive number. */
1307 for (; p < end && ISDIGIT (*p); p++)
1309 int next = (10 * result) + (*p - '0');
1311 return false; /* overflow */
1315 for (; p < end && ISDIGIT (*p); p++)
1317 int next = (10 * result) - (*p - '0');
1319 return false; /* underflow */
1330 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1331 doesn't handle exponential notation. Returns true on success,
1332 false on failure. In case of success, stores its result to
1336 simple_atof (const char *beg, const char *end, double *dest)
1340 bool negative = false;
1341 bool seen_dot = false;
1342 bool seen_digit = false;
1345 const char *p = beg;
1347 while (p < end && ISSPACE (*p))
1349 if (p < end && (*p == '-' || *p == '+'))
1351 negative = (*p == '-');
1355 for (; p < end; p++)
1361 result = (10 * result) + (ch - '0');
1363 result += (ch - '0') / (divider *= 10);
1385 /* Verify that the user-specified header in S is valid. It must
1386 contain a colon preceded by non-white-space characters and must not
1387 contain newlines. */
1390 check_user_specified_header (const char *s)
1394 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++)
1396 /* The header MUST contain `:' preceded by at least one
1397 non-whitespace character. */
1398 if (*p != ':' || p == s)
1400 /* The header MUST NOT contain newlines. */
1401 if (strchr (s, '\n'))
1406 /* Decode VAL into a number, according to ITEMS. */
1409 decode_string (const char *val, const struct decode_item *items, int itemcount,
1413 for (i = 0; i < itemcount; i++)
1414 if (0 == strcasecmp (val, items[i].name))
1416 *place = items[i].code;
1423 void cleanup_html_url (void);
1426 /* Free the memory allocated by global variables. */
1430 /* Free external resources, close files, etc. */
1433 fclose (output_stream);
1434 /* No need to check for error because Wget flushes its output (and
1435 checks for errors) after any data arrives. */
1437 /* We're exiting anyway so there's no real need to call free()
1438 hundreds of times. Skipping the frees will make Wget exit
1441 However, when detecting leaks, it's crucial to free() everything
1442 because then you can find the real leaks, i.e. the allocated
1443 memory which grows with the size of the program. */
1449 cleanup_html_url ();
1454 extern acc_t *netrc_list;
1455 free_netrc (netrc_list);
1457 xfree_null (opt.lfilename);
1458 xfree_null (opt.dir_prefix);
1459 xfree_null (opt.input_filename);
1460 xfree_null (opt.output_document);
1461 free_vec (opt.accepts);
1462 free_vec (opt.rejects);
1463 free_vec (opt.excludes);
1464 free_vec (opt.includes);
1465 free_vec (opt.domains);
1466 free_vec (opt.follow_tags);
1467 free_vec (opt.ignore_tags);
1468 xfree_null (opt.progress_type);
1469 xfree_null (opt.ftp_user);
1470 xfree_null (opt.ftp_passwd);
1471 xfree_null (opt.ftp_proxy);
1472 xfree_null (opt.https_proxy);
1473 xfree_null (opt.http_proxy);
1474 free_vec (opt.no_proxy);
1475 xfree_null (opt.useragent);
1476 xfree_null (opt.referer);
1477 xfree_null (opt.http_user);
1478 xfree_null (opt.http_passwd);
1479 free_vec (opt.user_headers);
1481 xfree_null (opt.cert_file);
1482 xfree_null (opt.private_key);
1483 xfree_null (opt.ca_directory);
1484 xfree_null (opt.ca_cert);
1485 xfree_null (opt.random_file);
1486 xfree_null (opt.egd_file);
1488 xfree_null (opt.bind_address);
1489 xfree_null (opt.cookies_input);
1490 xfree_null (opt.cookies_output);
1491 xfree_null (opt.user);
1492 xfree_null (opt.passwd);
1493 #endif /* DEBUG_MALLOC */