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 { "limitrate", &opt.limit_rate, cmd_bytes },
177 { "loadcookies", &opt.cookies_input, cmd_file },
178 { "logfile", &opt.lfilename, cmd_file },
179 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
180 { "mirror", NULL, cmd_spec_mirror },
181 { "netrc", &opt.netrc, cmd_boolean },
182 { "noclobber", &opt.noclobber, cmd_boolean },
183 { "noparent", &opt.no_parent, cmd_boolean },
184 { "noproxy", &opt.no_proxy, cmd_vector },
185 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
186 { "outputdocument", &opt.output_document, cmd_file },
187 { "pagerequisites", &opt.page_requisites, cmd_boolean },
188 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
189 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
190 { "password", &opt.passwd, cmd_string },
191 { "postdata", &opt.post_data, cmd_string },
192 { "postfile", &opt.post_file_name, cmd_file },
193 { "preferfamily", NULL, cmd_spec_prefer_family },
194 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
196 { "privatekey", &opt.private_key, cmd_file },
197 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
199 { "progress", &opt.progress_type, cmd_spec_progress },
200 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
201 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
202 { "proxypassword", &opt.proxy_passwd, cmd_string },
203 { "proxyuser", &opt.proxy_user, cmd_string },
204 { "quiet", &opt.quiet, cmd_boolean },
205 { "quota", &opt.quota, cmd_bytes_sum },
207 { "randomfile", &opt.random_file, cmd_file },
209 { "randomwait", &opt.random_wait, cmd_boolean },
210 { "readtimeout", &opt.read_timeout, cmd_time },
211 { "reclevel", &opt.reclevel, cmd_number_inf },
212 { "recursive", NULL, cmd_spec_recursive },
213 { "referer", &opt.referer, cmd_string },
214 { "reject", &opt.rejects, cmd_vector },
215 { "relativeonly", &opt.relative_only, cmd_boolean },
216 { "removelisting", &opt.remove_listing, cmd_boolean },
217 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
218 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
219 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
220 { "robots", &opt.use_robots, cmd_boolean },
221 { "savecookies", &opt.cookies_output, cmd_file },
222 { "saveheaders", &opt.save_headers, cmd_boolean },
224 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
226 { "serverresponse", &opt.server_response, cmd_boolean },
227 { "spanhosts", &opt.spanhost, cmd_boolean },
228 { "spider", &opt.spider, cmd_boolean },
229 { "strictcomments", &opt.strict_comments, cmd_boolean },
230 { "timeout", NULL, cmd_spec_timeout },
231 { "timestamping", &opt.timestamping, cmd_boolean },
232 { "tries", &opt.ntry, cmd_number_inf },
233 { "useproxy", &opt.use_proxy, cmd_boolean },
234 { "user", &opt.user, cmd_string },
235 { "useragent", NULL, cmd_spec_useragent },
236 { "verbose", NULL, cmd_spec_verbose },
237 { "wait", &opt.wait, cmd_time },
238 { "waitretry", &opt.waitretry, cmd_time }
241 /* Look up CMDNAME in the commands[] and return its position in the
242 array. If CMDNAME is not found, return -1. */
245 command_by_name (const char *cmdname)
247 /* Use binary search for speed. Wget has ~100 commands, which
248 guarantees a worst case performance of 7 string comparisons. */
249 int lo = 0, hi = countof (commands) - 1;
253 int mid = (lo + hi) >> 1;
254 int cmp = strcasecmp (cmdname, commands[mid].name);
265 /* Reset the variables to default values. */
271 /* Most of the default values are 0 (and 0.0, NULL, and false).
272 Just reset everything, and fill in the non-zero values. Note
273 that initializing pointers to NULL this way is technically
274 illegal, but porting Wget to a machine where NULL is not all-zero
275 bit pattern will be the least of the implementors' worries. */
282 opt.add_hostdir = true;
286 opt.http_keep_alive = true;
287 opt.use_proxy = true;
288 tmp = getenv ("no_proxy");
290 opt.no_proxy = sepstring (tmp);
291 opt.allow_cache = true;
293 opt.read_timeout = 900;
294 opt.use_robots = true;
296 opt.remove_listing = true;
298 opt.dot_bytes = 1024;
299 opt.dot_spacing = 10;
300 opt.dots_in_line = 50;
302 opt.dns_cache = true;
306 opt.check_cert = true;
309 /* The default for file name restriction defaults to the OS type. */
310 #if !defined(WINDOWS) && !defined(__CYGWIN__)
311 opt.restrict_files_os = restrict_unix;
313 opt.restrict_files_os = restrict_windows;
315 opt.restrict_files_ctrl = true;
318 /* Return the user's home directory (strdup-ed), or NULL if none is
323 char *home = getenv ("HOME");
328 /* If HOME is not defined, try getting it from the password
330 struct passwd *pwd = getpwuid (getuid ());
331 if (!pwd || !pwd->pw_dir)
335 /* Under Windows, if $HOME isn't defined, use the directory where
336 `wget.exe' resides. */
341 return home ? xstrdup (home) : NULL;
344 /* Return the path to the user's .wgetrc. This is either the value of
345 `WGETRC' environment variable, or `$HOME/.wgetrc'.
347 If the `WGETRC' variable exists but the file does not exist, the
348 function will exit(). */
350 wgetrc_file_name (void)
355 /* Try the environment. */
356 env = getenv ("WGETRC");
359 if (!file_exists_p (env))
361 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
365 return xstrdup (env);
368 /* If that failed, try $HOME/.wgetrc. */
371 file = aprintf ("%s/.wgetrc", home);
375 /* Under Windows, if we still haven't found .wgetrc, look for the file
376 `wget.ini' in the directory where `wget.exe' resides; we do this for
377 backward compatibility with previous versions of Wget.
378 SYSTEM_WGETRC should not be defined under WINDOWS. */
379 if (!file || !file_exists_p (file))
385 file = aprintf ("%s/wget.ini", home);
391 if (!file_exists_p (file))
399 /* Return values of parse_line. */
407 static enum parse_line parse_line (const char *, char **, char **, int *);
408 static bool setval_internal (int, const char *, const char *);
410 /* Initialize variables from a wgetrc file. Returns zero (failure) if
411 there were errors in the file. */
414 run_wgetrc (const char *file)
421 fp = fopen (file, "rb");
424 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
425 file, strerror (errno));
426 return true; /* not a fatal error */
428 enable_tilde_expansion = true;
430 while ((line = read_whole_line (fp)) != NULL)
432 char *com = NULL, *val = NULL;
435 /* Parse the line. */
436 switch (parse_line (line, &com, &val, &comind))
439 /* If everything is OK, set the value. */
440 if (!setval_internal (comind, com, val))
442 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
443 exec_name, file, ln);
447 case line_syntax_error:
448 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
449 exec_name, file, ln);
452 case line_unknown_command:
453 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
454 exec_name, com, file, ln);
467 enable_tilde_expansion = false;
473 /* Initialize the defaults and run the system wgetrc and user's own
481 /* Load the hard-coded defaults. */
484 /* If SYSTEM_WGETRC is defined, use it. */
486 if (file_exists_p (SYSTEM_WGETRC))
487 ok &= run_wgetrc (SYSTEM_WGETRC);
489 /* Override it with your own, if one exists. */
490 file = wgetrc_file_name ();
493 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
494 something like realpath() before comparing them with `strcmp' */
496 if (!strcmp (file, SYSTEM_WGETRC))
498 fprintf (stderr, _("\
499 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
504 ok &= run_wgetrc (file);
506 /* If there were errors processing either `.wgetrc', abort. */
514 /* Remove dashes and underscores from S, modifying S in the
520 char *t = s; /* t - tortoise */
521 char *h = s; /* h - hare */
523 if (*h == '_' || *h == '-')
530 /* Parse the line pointed by line, with the syntax:
531 <sp>* command <sp>* = <sp>* value <sp>*
532 Uses malloc to allocate space for command and value.
534 Returns one of line_ok, line_empty, line_syntax_error, or
535 line_unknown_command.
537 In case of line_ok, *COM and *VAL point to freshly allocated
538 strings, and *COMIND points to com's index. In case of error or
539 empty line, their values are unmodified. */
541 static enum parse_line
542 parse_line (const char *line, char **com, char **val, int *comind)
545 const char *end = line + strlen (line);
546 const char *cmdstart, *cmdend;
547 const char *valstart, *valend;
552 /* Skip leading and trailing whitespace. */
553 while (*line && ISSPACE (*line))
555 while (end > line && ISSPACE (end[-1]))
558 /* Skip empty lines and comments. */
559 if (!*line || *line == '#')
565 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
569 /* Skip '=', as well as any space before or after it. */
570 while (p < end && ISSPACE (*p))
572 if (p == end || *p != '=')
573 return line_syntax_error;
575 while (p < end && ISSPACE (*p))
581 /* The syntax is valid (even though the command might not be). Fill
582 in the command name and value. */
583 *com = strdupdelim (cmdstart, cmdend);
584 *val = strdupdelim (valstart, valend);
586 /* The line now known to be syntactically correct. Check whether
587 the command is valid. */
588 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
590 ind = command_by_name (cmdcopy);
592 return line_unknown_command;
594 /* Report success to the caller. */
599 /* Run commands[comind].action. */
602 setval_internal (int comind, const char *com, const char *val)
604 assert (0 <= comind && comind < countof (commands));
605 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
606 return commands[comind].action (com, val, commands[comind].place);
609 /* Run command COM with value VAL. If running the command produces an
610 error, report the error and exit.
612 This is intended to be called from main() to modify Wget's behavior
613 through command-line switches. Since COM is hard-coded in main(),
614 it is not canonicalized, and this aborts when COM is not found.
616 If COMIND's are exported to init.h, this function will be changed
617 to accept COMIND directly. */
620 setoptval (const char *com, const char *val, const char *optname)
622 /* Prepend "--" to OPTNAME. */
623 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
626 strcpy (dd_optname + 2, optname);
628 assert (val != NULL);
629 if (!setval_internal (command_by_name (com), dd_optname, val))
633 /* Parse OPT into command and value and run it. For example,
634 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
635 This is used by the `--execute' flag in main.c. */
638 run_command (const char *opt)
642 switch (parse_line (opt, &com, &val, &comind))
645 if (!setval_internal (comind, com, val))
651 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
657 /* Generic helper functions, for use with `commands'. */
659 /* Forward declarations: */
664 static bool decode_string (const char *, const struct decode_item *, int, int *);
665 static bool simple_atoi (const char *, const char *, int *);
666 static bool simple_atof (const char *, const char *, double *);
668 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
670 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
671 && TOLOWER((p)[1]) == (c1) \
674 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
675 && TOLOWER((p)[1]) == (c1) \
676 && TOLOWER((p)[2]) == (c2) \
680 /* Store the boolean value from VAL to PLACE. COM is ignored,
681 except for error messages. */
683 cmd_boolean (const char *com, const char *val, void *place)
687 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
688 /* "on", "yes" and "1" mean true. */
690 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
691 /* "off", "no" and "0" mean false. */
696 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
697 exec_name, com, val);
701 *(bool *) place = value;
705 /* Set the non-negative integer value from VAL to PLACE. With
706 incorrect specification, the number remains unchanged. */
708 cmd_number (const char *com, const char *val, void *place)
710 if (!simple_atoi (val, val + strlen (val), place)
711 || *(int *) place < 0)
713 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
714 exec_name, com, val);
720 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
722 cmd_number_inf (const char *com, const char *val, void *place)
724 if (!strcasecmp (val, "inf"))
729 return cmd_number (com, val, place);
732 /* Copy (strdup) the string at COM to a new location and place a
733 pointer to *PLACE. */
735 cmd_string (const char *com, const char *val, void *place)
737 char **pstring = (char **)place;
739 xfree_null (*pstring);
740 *pstring = xstrdup (val);
745 # define ISSEP(c) ((c) == '/')
747 # define ISSEP(c) ((c) == '/' || (c) == '\\')
750 /* Like the above, but handles tilde-expansion when reading a user's
751 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
752 gets expanded to the user's home directory. */
754 cmd_file (const char *com, const char *val, void *place)
756 char **pstring = (char **)place;
758 xfree_null (*pstring);
760 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
762 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
765 *pstring = xstrdup (val);
770 char *home = home_dir ();
774 homelen = strlen (home);
775 while (homelen && ISSEP (home[homelen - 1]))
776 home[--homelen] = '\0';
778 /* Skip the leading "~/". */
779 for (++val; ISSEP (*val); val++)
782 *pstring = concat_strings (home, "/", val, (char *) 0);
786 /* Convert "\" to "/". */
789 for (s = *pstring; *s; s++)
797 /* Like cmd_file, but strips trailing '/' characters. */
799 cmd_directory (const char *com, const char *val, void *place)
803 /* Call cmd_file() for tilde expansion and separator
804 canonicalization (backslash -> slash under Windows). These
805 things should perhaps be in a separate function. */
806 if (!cmd_file (com, val, place))
811 while (t > s && *--t == '/')
817 /* Split VAL by space to a vector of values, and append those values
818 to vector pointed to by the PLACE argument. If VAL is empty, the
819 PLACE vector is cleared instead. */
822 cmd_vector (const char *com, const char *val, void *place)
824 char ***pvec = (char ***)place;
827 *pvec = merge_vecs (*pvec, sepstring (val));
837 cmd_directory_vector (const char *com, const char *val, void *place)
839 char ***pvec = (char ***)place;
843 /* Strip the trailing slashes from directories. */
846 seps = sepstring (val);
847 for (t = seps; t && *t; t++)
849 int len = strlen (*t);
850 /* Skip degenerate case of root directory. */
853 if ((*t)[len - 1] == '/')
854 (*t)[len - 1] = '\0';
857 *pvec = merge_vecs (*pvec, seps);
867 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
868 "100k" or "2.5G" to a floating point number. */
871 parse_bytes_helper (const char *val, double *result)
874 const char *end = val + strlen (val);
876 /* Check for "inf". */
877 if (0 == strcmp (val, "inf"))
883 /* Strip trailing whitespace. */
884 while (val < end && ISSPACE (end[-1]))
889 switch (TOLOWER (end[-1]))
892 --end, mult = 1024.0;
895 --end, mult = 1048576.0;
898 --end, mult = 1073741824.0;
901 --end, mult = 1099511627776.0;
904 /* Not a recognized suffix: assume it's a digit. (If not,
905 simple_atof will raise an error.) */
909 /* Skip leading and trailing whitespace. */
910 while (val < end && ISSPACE (*val))
912 while (val < end && ISSPACE (end[-1]))
917 if (!simple_atof (val, end, &number) || number < 0)
920 *result = number * mult;
924 /* Parse VAL as a number and set its value to PLACE (which should
927 By default, the value is assumed to be in bytes. If "K", "M", or
928 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
929 1<<30, respectively. Floating point values are allowed and are
930 cast to integer before use. The idea is to be able to use things
931 like 1.5k instead of "1536".
933 The string "inf" is returned as 0.
935 In case of error, false is returned and memory pointed to by PLACE
936 remains unmodified. */
939 cmd_bytes (const char *com, const char *val, void *place)
942 if (!parse_bytes_helper (val, &byte_value))
944 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
945 exec_name, com, val);
948 *(wgint *)place = (wgint)byte_value;
952 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
953 SIZE_SUM. It works by converting the string to double, therefore
954 working with values up to 2^53-1 without loss of precision. This
955 value (8192 TB) is large enough to serve for a while. */
958 cmd_bytes_sum (const char *com, const char *val, void *place)
961 if (!parse_bytes_helper (val, &byte_value))
963 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
964 exec_name, com, val);
967 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
971 /* Store the value of VAL to *OUT. The value is a time period, by
972 default expressed in seconds, but also accepting suffixes "m", "h",
973 "d", and "w" for minutes, hours, days, and weeks respectively. */
976 cmd_time (const char *com, const char *val, void *place)
979 const char *end = val + strlen (val);
981 /* Strip trailing whitespace. */
982 while (val < end && ISSPACE (end[-1]))
988 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
989 exec_name, com, val);
993 switch (TOLOWER (end[-1]))
996 --end, mult = 1; /* seconds */
999 --end, mult = 60; /* minutes */
1002 --end, mult = 3600; /* hours */
1005 --end, mult = 86400.0; /* days */
1008 --end, mult = 604800.0; /* weeks */
1011 /* Not a recognized suffix: assume it belongs to the number.
1012 (If not, simple_atof will raise an error.) */
1016 /* Skip leading and trailing whitespace. */
1017 while (val < end && ISSPACE (*val))
1019 while (val < end && ISSPACE (end[-1]))
1024 if (!simple_atof (val, end, &number))
1027 *(double *)place = number * mult;
1033 cmd_cert_type (const char *com, const char *val, void *place)
1035 static const struct decode_item choices[] = {
1036 { "pem", keyfile_pem },
1037 { "der", keyfile_asn1 },
1038 { "asn1", keyfile_asn1 },
1040 int ok = decode_string (val, choices, countof (choices), place);
1042 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1047 /* Specialized helper functions, used by `commands' to handle some
1048 options specially. */
1050 static bool check_user_specified_header (const char *);
1053 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1055 if (!cmd_boolean (com, val, &opt.dirstruct))
1057 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1058 must be affected inversely. */
1060 opt.no_dirstruct = false;
1062 opt.no_dirstruct = true;
1067 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1069 /* Empty value means reset the list of headers. */
1072 free_vec (opt.user_headers);
1073 opt.user_headers = NULL;
1077 if (!check_user_specified_header (val))
1079 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1080 exec_name, com, val);
1083 opt.user_headers = vec_append (opt.user_headers, val);
1088 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1090 int flag = cmd_boolean (com, val, &opt.htmlify);
1091 if (flag && !opt.htmlify)
1092 opt.remove_listing = false;
1096 /* Set the "mirror" mode. It means: recursive download, timestamping,
1097 no limit on max. recursion depth, and don't remove listings. */
1100 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1104 if (!cmd_boolean (com, val, &mirror))
1108 opt.recursive = true;
1109 if (!opt.no_dirstruct)
1110 opt.dirstruct = true;
1111 opt.timestamping = true;
1112 opt.reclevel = INFINITE_RECURSION;
1113 opt.remove_listing = false;
1118 /* Validate --prefer-family and set the choice. Allowed values are
1119 "IPv4", "IPv6", and "none". */
1122 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1124 static const struct decode_item choices[] = {
1125 { "IPv4", prefer_ipv4 },
1126 { "IPv6", prefer_ipv6 },
1127 { "none", prefer_none },
1129 int ok = decode_string (val, choices, countof (choices),
1130 (int *) &opt.prefer_family);
1132 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1136 /* Set progress.type to VAL, but verify that it's a valid progress
1137 implementation before that. */
1140 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1142 if (!valid_progress_implementation_p (val))
1144 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1145 exec_name, com, val);
1148 xfree_null (opt.progress_type);
1150 /* Don't call set_progress_implementation here. It will be called
1151 in main() when it becomes clear what the log output is. */
1152 opt.progress_type = xstrdup (val);
1156 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1157 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1161 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1163 if (!cmd_boolean (com, val, &opt.recursive))
1167 if (opt.recursive && !opt.no_dirstruct)
1168 opt.dirstruct = true;
1174 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1176 int restrict_os = opt.restrict_files_os;
1177 int restrict_ctrl = opt.restrict_files_ctrl;
1179 const char *end = strchr (val, ',');
1181 end = val + strlen (val);
1183 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1185 if (VAL_IS ("unix"))
1186 restrict_os = restrict_unix;
1187 else if (VAL_IS ("windows"))
1188 restrict_os = restrict_windows;
1189 else if (VAL_IS ("nocontrol"))
1195 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1196 exec_name, com, val);
1204 if (!strcmp (end + 1, "nocontrol"))
1205 restrict_ctrl = false;
1210 opt.restrict_files_os = restrict_os;
1211 opt.restrict_files_ctrl = restrict_ctrl;
1217 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1219 static const struct decode_item choices[] = {
1220 { "auto", secure_protocol_auto },
1221 { "sslv2", secure_protocol_sslv2 },
1222 { "sslv3", secure_protocol_sslv3 },
1223 { "tlsv1", secure_protocol_tlsv1 },
1225 int ok = decode_string (val, choices, countof (choices), place);
1227 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1232 /* Set all three timeout values. */
1235 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1238 if (!cmd_time (com, val, &value))
1240 opt.read_timeout = value;
1241 opt.connect_timeout = value;
1242 opt.dns_timeout = value;
1247 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1249 /* Disallow embedded newlines. */
1250 if (strchr (val, '\n'))
1252 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1253 exec_name, com, val);
1256 xfree_null (opt.useragent);
1257 opt.useragent = xstrdup (val);
1261 /* The "verbose" option cannot be cmd_boolean because the variable is
1262 not bool -- it's of type int (-1 means uninitialized because of
1263 some random hackery for disallowing -q -v). */
1266 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1269 if (cmd_boolean (com, val, &flag))
1277 /* Miscellaneous useful routines. */
1279 /* A very simple atoi clone, more useful than atoi because it works on
1280 delimited strings, and has error reportage. Returns true on success,
1281 false on failure. If successful, stores result to *DEST. */
1284 simple_atoi (const char *beg, const char *end, int *dest)
1287 bool negative = false;
1288 const char *p = beg;
1290 while (p < end && ISSPACE (*p))
1292 if (p < end && (*p == '-' || *p == '+'))
1294 negative = (*p == '-');
1300 /* Read negative numbers in a separate loop because the most
1301 negative integer cannot be represented as a positive number. */
1304 for (; p < end && ISDIGIT (*p); p++)
1306 int next = (10 * result) + (*p - '0');
1308 return false; /* overflow */
1312 for (; p < end && ISDIGIT (*p); p++)
1314 int next = (10 * result) - (*p - '0');
1316 return false; /* underflow */
1327 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1328 doesn't handle exponential notation. Returns true on success,
1329 false on failure. In case of success, stores its result to
1333 simple_atof (const char *beg, const char *end, double *dest)
1337 bool negative = false;
1338 bool seen_dot = false;
1339 bool seen_digit = false;
1342 const char *p = beg;
1344 while (p < end && ISSPACE (*p))
1346 if (p < end && (*p == '-' || *p == '+'))
1348 negative = (*p == '-');
1352 for (; p < end; p++)
1358 result = (10 * result) + (ch - '0');
1360 result += (ch - '0') / (divider *= 10);
1382 /* Verify that the user-specified header in S is valid. It must
1383 contain a colon preceded by non-white-space characters and must not
1384 contain newlines. */
1387 check_user_specified_header (const char *s)
1391 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 */