1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
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 */
61 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
62 otherwise, it will be performed by the shell. This variable will
63 be set by the wgetrc-reading function. */
65 static bool enable_tilde_expansion;
68 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
70 CMD_DECLARE (cmd_boolean);
71 CMD_DECLARE (cmd_bytes);
72 CMD_DECLARE (cmd_bytes_sum);
74 CMD_DECLARE (cmd_cert_type);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_number);
78 CMD_DECLARE (cmd_number_inf);
79 CMD_DECLARE (cmd_string);
80 CMD_DECLARE (cmd_file);
81 CMD_DECLARE (cmd_directory);
82 CMD_DECLARE (cmd_time);
83 CMD_DECLARE (cmd_vector);
85 CMD_DECLARE (cmd_spec_dirstruct);
86 CMD_DECLARE (cmd_spec_header);
87 CMD_DECLARE (cmd_spec_htmlify);
88 CMD_DECLARE (cmd_spec_mirror);
89 CMD_DECLARE (cmd_spec_prefer_family);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_secure_protocol);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
98 CMD_DECLARE (cmd_spec_verbose);
100 /* List of recognized commands, each consisting of name, place and
101 function. When adding a new command, simply add it to the list,
102 but be sure to keep the list sorted alphabetically, as
103 command_by_name's binary search depends on it. Also, be sure to
104 add any entries that allocate memory (e.g. cmd_string and
105 cmd_vector) to the cleanup() function below. */
107 static const struct {
110 bool (*action) (const char *, const char *, void *);
112 /* KEEP THIS LIST ALPHABETICALLY SORTED */
113 { "accept", &opt.accepts, cmd_vector },
114 { "addhostdir", &opt.add_hostdir, cmd_boolean },
115 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
116 { "askpassword", &opt.ask_passwd, cmd_boolean },
117 { "authnochallenge", &opt.auth_without_challenge,
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "contentdisposition", &opt.content_disposition, cmd_boolean },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "deleteafter", &opt.delete_after, cmd_boolean },
144 { "dirprefix", &opt.dir_prefix, cmd_directory },
145 { "dirstruct", NULL, cmd_spec_dirstruct },
146 { "dnscache", &opt.dns_cache, cmd_boolean },
147 { "dnstimeout", &opt.dns_timeout, cmd_time },
148 { "domains", &opt.domains, cmd_vector },
149 { "dotbytes", &opt.dot_bytes, cmd_bytes },
150 { "dotsinline", &opt.dots_in_line, cmd_number },
151 { "dotspacing", &opt.dot_spacing, cmd_number },
152 { "dotstyle", &opt.dot_style, cmd_string },
154 { "egdfile", &opt.egd_file, cmd_file },
156 { "excludedirectories", &opt.excludes, cmd_directory_vector },
157 { "excludedomains", &opt.exclude_domains, cmd_vector },
158 { "followftp", &opt.follow_ftp, cmd_boolean },
159 { "followtags", &opt.follow_tags, cmd_vector },
160 { "forcehtml", &opt.force_html, cmd_boolean },
161 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
162 { "ftppassword", &opt.ftp_passwd, cmd_string },
163 { "ftpproxy", &opt.ftp_proxy, cmd_string },
164 { "ftpuser", &opt.ftp_user, cmd_string },
165 { "glob", &opt.ftp_glob, cmd_boolean },
166 { "header", NULL, cmd_spec_header },
167 { "htmlextension", &opt.html_extension, cmd_boolean },
168 { "htmlify", NULL, cmd_spec_htmlify },
169 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
170 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
171 { "httppassword", &opt.http_passwd, cmd_string },
172 { "httpproxy", &opt.http_proxy, cmd_string },
173 { "httpsproxy", &opt.https_proxy, cmd_string },
174 { "httpuser", &opt.http_user, cmd_string },
175 { "ignorecase", &opt.ignore_case, cmd_boolean },
176 { "ignorelength", &opt.ignore_length, cmd_boolean },
177 { "ignoretags", &opt.ignore_tags, cmd_vector },
178 { "includedirectories", &opt.includes, cmd_directory_vector },
180 { "inet4only", &opt.ipv4_only, cmd_boolean },
181 { "inet6only", &opt.ipv6_only, cmd_boolean },
183 { "input", &opt.input_filename, cmd_file },
184 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
185 { "limitrate", &opt.limit_rate, cmd_bytes },
186 { "loadcookies", &opt.cookies_input, cmd_file },
187 { "logfile", &opt.lfilename, cmd_file },
188 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
189 { "maxredirect", &opt.max_redirect, cmd_number },
190 { "mirror", NULL, cmd_spec_mirror },
191 { "netrc", &opt.netrc, cmd_boolean },
192 { "noclobber", &opt.noclobber, cmd_boolean },
193 { "noparent", &opt.no_parent, cmd_boolean },
194 { "noproxy", &opt.no_proxy, cmd_vector },
195 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
196 { "outputdocument", &opt.output_document, cmd_file },
197 { "pagerequisites", &opt.page_requisites, cmd_boolean },
198 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
199 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
200 { "password", &opt.passwd, cmd_string },
201 { "postdata", &opt.post_data, cmd_string },
202 { "postfile", &opt.post_file_name, cmd_file },
203 { "preferfamily", NULL, cmd_spec_prefer_family },
204 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
206 { "privatekey", &opt.private_key, cmd_file },
207 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
209 { "progress", &opt.progress_type, cmd_spec_progress },
210 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
211 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
212 { "proxypassword", &opt.proxy_passwd, cmd_string },
213 { "proxyuser", &opt.proxy_user, cmd_string },
214 { "quiet", &opt.quiet, cmd_boolean },
215 { "quota", &opt.quota, cmd_bytes_sum },
217 { "randomfile", &opt.random_file, cmd_file },
219 { "randomwait", &opt.random_wait, cmd_boolean },
220 { "readtimeout", &opt.read_timeout, cmd_time },
221 { "reclevel", &opt.reclevel, cmd_number_inf },
222 { "recursive", NULL, cmd_spec_recursive },
223 { "referer", &opt.referer, cmd_string },
224 { "reject", &opt.rejects, cmd_vector },
225 { "relativeonly", &opt.relative_only, cmd_boolean },
226 { "removelisting", &opt.remove_listing, cmd_boolean },
227 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
228 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
229 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
230 { "robots", &opt.use_robots, cmd_boolean },
231 { "savecookies", &opt.cookies_output, cmd_file },
232 { "saveheaders", &opt.save_headers, cmd_boolean },
234 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
236 { "serverresponse", &opt.server_response, cmd_boolean },
237 { "spanhosts", &opt.spanhost, cmd_boolean },
238 { "spider", &opt.spider, cmd_boolean },
239 { "strictcomments", &opt.strict_comments, cmd_boolean },
240 { "timeout", NULL, cmd_spec_timeout },
241 { "timestamping", &opt.timestamping, cmd_boolean },
242 { "tries", &opt.ntry, cmd_number_inf },
243 { "useproxy", &opt.use_proxy, cmd_boolean },
244 { "user", &opt.user, cmd_string },
245 { "useragent", NULL, cmd_spec_useragent },
246 { "verbose", NULL, cmd_spec_verbose },
247 { "wait", &opt.wait, cmd_time },
248 { "waitretry", &opt.waitretry, cmd_time },
250 { "wdebug", &opt.wdebug, cmd_boolean },
254 /* Look up CMDNAME in the commands[] and return its position in the
255 array. If CMDNAME is not found, return -1. */
258 command_by_name (const char *cmdname)
260 /* Use binary search for speed. Wget has ~100 commands, which
261 guarantees a worst case performance of 7 string comparisons. */
262 int lo = 0, hi = countof (commands) - 1;
266 int mid = (lo + hi) >> 1;
267 int cmp = strcasecmp (cmdname, commands[mid].name);
278 /* Reset the variables to default values. */
284 /* Most of the default values are 0 (and 0.0, NULL, and false).
285 Just reset everything, and fill in the non-zero values. Note
286 that initializing pointers to NULL this way is technically
287 illegal, but porting Wget to a machine where NULL is not all-zero
288 bit pattern will be the least of the implementors' worries. */
295 opt.add_hostdir = true;
299 opt.http_keep_alive = true;
300 opt.use_proxy = true;
301 tmp = getenv ("no_proxy");
303 opt.no_proxy = sepstring (tmp);
304 opt.prefer_family = prefer_none;
305 opt.allow_cache = true;
307 opt.read_timeout = 900;
308 opt.use_robots = true;
310 opt.remove_listing = true;
312 opt.dot_bytes = 1024;
313 opt.dot_spacing = 10;
314 opt.dots_in_line = 50;
316 opt.dns_cache = true;
320 opt.check_cert = true;
323 /* The default for file name restriction defaults to the OS type. */
324 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
325 opt.restrict_files_os = restrict_windows;
327 opt.restrict_files_os = restrict_unix;
329 opt.restrict_files_ctrl = true;
330 opt.restrict_files_case = restrict_no_case_restriction;
332 opt.max_redirect = 20;
335 /* Return the user's home directory (strdup-ed), or NULL if none is
340 char *home = getenv ("HOME");
345 /* Under MSDOS, if $HOME isn't defined, use the directory where
346 `wget.exe' resides. */
347 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
348 char *p, buf[PATH_MAX];
350 strcpy (buf, _w32_get_argv0 ());
351 p = strrchr (buf, '/'); /* djgpp */
353 p = strrchr (buf, '\\'); /* others */
357 #elif !defined(WINDOWS)
358 /* If HOME is not defined, try getting it from the password
360 struct passwd *pwd = getpwuid (getuid ());
361 if (!pwd || !pwd->pw_dir)
365 /* Under Windows, if $HOME isn't defined, use the directory where
366 `wget.exe' resides. */
371 return home ? xstrdup (home) : NULL;
374 /* Check the 'WGETRC' environment variable and return the file name
375 if 'WGETRC' is set and is a valid file.
376 If the `WGETRC' variable exists but the file does not exist, the
377 function will exit(). */
379 wgetrc_env_file_name (void)
381 char *env = getenv ("WGETRC");
384 if (!file_exists_p (env))
386 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
390 return xstrdup (env);
394 /* Check for the existance of '$HOME/.wgetrc' and return it's path
395 if it exists and is set. */
397 wgetrc_user_file_name (void)
399 char *home = home_dir();
402 file = aprintf ("%s/.wgetrc", home);
406 if (!file_exists_p (file))
413 /* Return the path to the user's .wgetrc. This is either the value of
414 `WGETRC' environment variable, or `$HOME/.wgetrc'.
416 Additionally, for windows, look in the directory where wget.exe
419 wgetrc_file_name (void)
421 char *file = wgetrc_env_file_name ();
425 file = wgetrc_user_file_name ();
428 /* Under Windows, if we still haven't found .wgetrc, look for the file
429 `wget.ini' in the directory where `wget.exe' resides; we do this for
430 backward compatibility with previous versions of Wget.
431 SYSTEM_WGETRC should not be defined under WINDOWS. */
432 if (!file || !file_exists_p (file))
438 file = aprintf ("%s/wget.ini", home);
444 if (!file_exists_p (file))
452 /* Return values of parse_line. */
460 static enum parse_line parse_line (const char *, char **, char **, int *);
461 static bool setval_internal (int, const char *, const char *);
463 /* Initialize variables from a wgetrc file. Returns zero (failure) if
464 there were errors in the file. */
467 run_wgetrc (const char *file)
474 fp = fopen (file, "rb");
477 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
478 file, strerror (errno));
479 return true; /* not a fatal error */
481 enable_tilde_expansion = true;
483 while ((line = read_whole_line (fp)) != NULL)
485 char *com = NULL, *val = NULL;
488 /* Parse the line. */
489 switch (parse_line (line, &com, &val, &comind))
492 /* If everything is OK, set the value. */
493 if (!setval_internal (comind, com, val))
495 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
496 exec_name, file, ln);
500 case line_syntax_error:
501 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
502 exec_name, file, ln);
505 case line_unknown_command:
506 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
507 exec_name, quote (com), file, ln);
520 enable_tilde_expansion = false;
526 /* Initialize the defaults and run the system wgetrc and user's own
534 /* Load the hard-coded defaults. */
537 /* If SYSTEM_WGETRC is defined, use it. */
539 if (file_exists_p (SYSTEM_WGETRC))
540 ok &= run_wgetrc (SYSTEM_WGETRC);
542 /* Override it with your own, if one exists. */
543 file = wgetrc_file_name ();
546 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
547 something like realpath() before comparing them with `strcmp' */
549 if (!strcmp (file, SYSTEM_WGETRC))
551 fprintf (stderr, _("\
552 %s: Warning: Both system and user wgetrc point to %s.\n"),
553 exec_name, quote (file));
557 ok &= run_wgetrc (file);
559 /* If there were errors processing either `.wgetrc', abort. */
567 /* Remove dashes and underscores from S, modifying S in the
573 char *t = s; /* t - tortoise */
574 char *h = s; /* h - hare */
576 if (*h == '_' || *h == '-')
583 /* Parse the line pointed by line, with the syntax:
584 <sp>* command <sp>* = <sp>* value <sp>*
585 Uses malloc to allocate space for command and value.
587 Returns one of line_ok, line_empty, line_syntax_error, or
588 line_unknown_command.
590 In case of line_ok, *COM and *VAL point to freshly allocated
591 strings, and *COMIND points to com's index. In case of error or
592 empty line, their values are unmodified. */
594 static enum parse_line
595 parse_line (const char *line, char **com, char **val, int *comind)
598 const char *end = line + strlen (line);
599 const char *cmdstart, *cmdend;
600 const char *valstart, *valend;
605 /* Skip leading and trailing whitespace. */
606 while (*line && c_isspace (*line))
608 while (end > line && c_isspace (end[-1]))
611 /* Skip empty lines and comments. */
612 if (!*line || *line == '#')
618 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
622 /* Skip '=', as well as any space before or after it. */
623 while (p < end && c_isspace (*p))
625 if (p == end || *p != '=')
626 return line_syntax_error;
628 while (p < end && c_isspace (*p))
634 /* The syntax is valid (even though the command might not be). Fill
635 in the command name and value. */
636 *com = strdupdelim (cmdstart, cmdend);
637 *val = strdupdelim (valstart, valend);
639 /* The line now known to be syntactically correct. Check whether
640 the command is valid. */
641 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
643 ind = command_by_name (cmdcopy);
645 return line_unknown_command;
647 /* Report success to the caller. */
652 /* Run commands[comind].action. */
655 setval_internal (int comind, const char *com, const char *val)
657 assert (0 <= comind && ((size_t) comind) < countof (commands));
658 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
659 return commands[comind].action (com, val, commands[comind].place);
662 /* Run command COM with value VAL. If running the command produces an
663 error, report the error and exit.
665 This is intended to be called from main() to modify Wget's behavior
666 through command-line switches. Since COM is hard-coded in main(),
667 it is not canonicalized, and this aborts when COM is not found.
669 If COMIND's are exported to init.h, this function will be changed
670 to accept COMIND directly. */
673 setoptval (const char *com, const char *val, const char *optname)
675 /* Prepend "--" to OPTNAME. */
676 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
679 strcpy (dd_optname + 2, optname);
681 assert (val != NULL);
682 if (!setval_internal (command_by_name (com), dd_optname, val))
686 /* Parse OPT into command and value and run it. For example,
687 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
688 This is used by the `--execute' flag in main.c. */
691 run_command (const char *opt)
695 switch (parse_line (opt, &com, &val, &comind))
698 if (!setval_internal (comind, com, val))
704 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
705 exec_name, quote (opt));
710 /* Generic helper functions, for use with `commands'. */
712 /* Forward declarations: */
717 static bool decode_string (const char *, const struct decode_item *, int, int *);
718 static bool simple_atoi (const char *, const char *, int *);
719 static bool simple_atof (const char *, const char *, double *);
721 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
723 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
724 && c_tolower((p)[1]) == (c1) \
727 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
728 && c_tolower((p)[1]) == (c1) \
729 && c_tolower((p)[2]) == (c2) \
733 /* Store the boolean value from VAL to PLACE. COM is ignored,
734 except for error messages. */
736 cmd_boolean (const char *com, const char *val, void *place)
740 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
741 /* "on", "yes" and "1" mean true. */
743 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
744 /* "off", "no" and "0" mean false. */
749 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
750 exec_name, com, quote (val));
754 *(bool *) place = value;
758 /* Set the non-negative integer value from VAL to PLACE. With
759 incorrect specification, the number remains unchanged. */
761 cmd_number (const char *com, const char *val, void *place)
763 if (!simple_atoi (val, val + strlen (val), place)
764 || *(int *) place < 0)
766 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
767 exec_name, com, quote (val));
773 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
775 cmd_number_inf (const char *com, const char *val, void *place)
777 if (!strcasecmp (val, "inf"))
782 return cmd_number (com, val, place);
785 /* Copy (strdup) the string at COM to a new location and place a
786 pointer to *PLACE. */
788 cmd_string (const char *com, const char *val, void *place)
790 char **pstring = (char **)place;
792 xfree_null (*pstring);
793 *pstring = xstrdup (val);
797 #if defined(WINDOWS) || defined(MSDOS)
798 # define ISSEP(c) ((c) == '/' || (c) == '\\')
800 # define ISSEP(c) ((c) == '/')
803 /* Like the above, but handles tilde-expansion when reading a user's
804 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
805 gets expanded to the user's home directory. */
807 cmd_file (const char *com, const char *val, void *place)
809 char **pstring = (char **)place;
811 xfree_null (*pstring);
813 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
815 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
818 *pstring = xstrdup (val);
823 char *home = home_dir ();
827 homelen = strlen (home);
828 while (homelen && ISSEP (home[homelen - 1]))
829 home[--homelen] = '\0';
831 /* Skip the leading "~/". */
832 for (++val; ISSEP (*val); val++)
835 *pstring = concat_strings (home, "/", val, (char *) 0);
838 #if defined(WINDOWS) || defined(MSDOS)
839 /* Convert "\" to "/". */
842 for (s = *pstring; *s; s++)
850 /* Like cmd_file, but strips trailing '/' characters. */
852 cmd_directory (const char *com, const char *val, void *place)
856 /* Call cmd_file() for tilde expansion and separator
857 canonicalization (backslash -> slash under Windows). These
858 things should perhaps be in a separate function. */
859 if (!cmd_file (com, val, place))
864 while (t > s && *--t == '/')
870 /* Split VAL by space to a vector of values, and append those values
871 to vector pointed to by the PLACE argument. If VAL is empty, the
872 PLACE vector is cleared instead. */
875 cmd_vector (const char *com, const char *val, void *place)
877 char ***pvec = (char ***)place;
880 *pvec = merge_vecs (*pvec, sepstring (val));
890 cmd_directory_vector (const char *com, const char *val, void *place)
892 char ***pvec = (char ***)place;
896 /* Strip the trailing slashes from directories. */
899 seps = sepstring (val);
900 for (t = seps; t && *t; t++)
902 int len = strlen (*t);
903 /* Skip degenerate case of root directory. */
906 if ((*t)[len - 1] == '/')
907 (*t)[len - 1] = '\0';
910 *pvec = merge_vecs (*pvec, seps);
920 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
921 "100k" or "2.5G" to a floating point number. */
924 parse_bytes_helper (const char *val, double *result)
927 const char *end = val + strlen (val);
929 /* Check for "inf". */
930 if (0 == strcmp (val, "inf"))
936 /* Strip trailing whitespace. */
937 while (val < end && c_isspace (end[-1]))
942 switch (c_tolower (end[-1]))
945 --end, mult = 1024.0;
948 --end, mult = 1048576.0;
951 --end, mult = 1073741824.0;
954 --end, mult = 1099511627776.0;
957 /* Not a recognized suffix: assume it's a digit. (If not,
958 simple_atof will raise an error.) */
962 /* Skip leading and trailing whitespace. */
963 while (val < end && c_isspace (*val))
965 while (val < end && c_isspace (end[-1]))
970 if (!simple_atof (val, end, &number) || number < 0)
973 *result = number * mult;
977 /* Parse VAL as a number and set its value to PLACE (which should
980 By default, the value is assumed to be in bytes. If "K", "M", or
981 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
982 1<<30, respectively. Floating point values are allowed and are
983 cast to integer before use. The idea is to be able to use things
984 like 1.5k instead of "1536".
986 The string "inf" is returned as 0.
988 In case of error, false is returned and memory pointed to by PLACE
989 remains unmodified. */
992 cmd_bytes (const char *com, const char *val, void *place)
995 if (!parse_bytes_helper (val, &byte_value))
997 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
998 exec_name, com, quote (val));
1001 *(wgint *)place = (wgint)byte_value;
1005 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1006 SIZE_SUM. It works by converting the string to double, therefore
1007 working with values up to 2^53-1 without loss of precision. This
1008 value (8192 TB) is large enough to serve for a while. */
1011 cmd_bytes_sum (const char *com, const char *val, void *place)
1014 if (!parse_bytes_helper (val, &byte_value))
1016 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1017 exec_name, com, quote (val));
1020 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1024 /* Store the value of VAL to *OUT. The value is a time period, by
1025 default expressed in seconds, but also accepting suffixes "m", "h",
1026 "d", and "w" for minutes, hours, days, and weeks respectively. */
1029 cmd_time (const char *com, const char *val, void *place)
1031 double number, mult;
1032 const char *end = val + strlen (val);
1034 /* Strip trailing whitespace. */
1035 while (val < end && c_isspace (end[-1]))
1041 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1042 exec_name, com, quote (val));
1046 switch (c_tolower (end[-1]))
1049 --end, mult = 1; /* seconds */
1052 --end, mult = 60; /* minutes */
1055 --end, mult = 3600; /* hours */
1058 --end, mult = 86400.0; /* days */
1061 --end, mult = 604800.0; /* weeks */
1064 /* Not a recognized suffix: assume it belongs to the number.
1065 (If not, simple_atof will raise an error.) */
1069 /* Skip leading and trailing whitespace. */
1070 while (val < end && c_isspace (*val))
1072 while (val < end && c_isspace (end[-1]))
1077 if (!simple_atof (val, end, &number))
1080 *(double *)place = number * mult;
1086 cmd_cert_type (const char *com, const char *val, void *place)
1088 static const struct decode_item choices[] = {
1089 { "pem", keyfile_pem },
1090 { "der", keyfile_asn1 },
1091 { "asn1", keyfile_asn1 },
1093 int ok = decode_string (val, choices, countof (choices), place);
1095 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1100 /* Specialized helper functions, used by `commands' to handle some
1101 options specially. */
1103 static bool check_user_specified_header (const char *);
1106 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1108 if (!cmd_boolean (com, val, &opt.dirstruct))
1110 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1111 must be affected inversely. */
1113 opt.no_dirstruct = false;
1115 opt.no_dirstruct = true;
1120 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1122 /* Empty value means reset the list of headers. */
1125 free_vec (opt.user_headers);
1126 opt.user_headers = NULL;
1130 if (!check_user_specified_header (val))
1132 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1133 exec_name, com, quote (val));
1136 opt.user_headers = vec_append (opt.user_headers, val);
1141 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1143 int flag = cmd_boolean (com, val, &opt.htmlify);
1144 if (flag && !opt.htmlify)
1145 opt.remove_listing = false;
1149 /* Set the "mirror" mode. It means: recursive download, timestamping,
1150 no limit on max. recursion depth, and don't remove listings. */
1153 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1157 if (!cmd_boolean (com, val, &mirror))
1161 opt.recursive = true;
1162 if (!opt.no_dirstruct)
1163 opt.dirstruct = true;
1164 opt.timestamping = true;
1165 opt.reclevel = INFINITE_RECURSION;
1166 opt.remove_listing = false;
1171 /* Validate --prefer-family and set the choice. Allowed values are
1172 "IPv4", "IPv6", and "none". */
1175 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1177 static const struct decode_item choices[] = {
1178 { "IPv4", prefer_ipv4 },
1179 { "IPv6", prefer_ipv6 },
1180 { "none", prefer_none },
1182 int prefer_family = prefer_none;
1183 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1185 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1186 opt.prefer_family = prefer_family;
1190 /* Set progress.type to VAL, but verify that it's a valid progress
1191 implementation before that. */
1194 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1196 if (!valid_progress_implementation_p (val))
1198 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1199 exec_name, com, quote (val));
1202 xfree_null (opt.progress_type);
1204 /* Don't call set_progress_implementation here. It will be called
1205 in main() when it becomes clear what the log output is. */
1206 opt.progress_type = xstrdup (val);
1210 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1211 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1215 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1217 if (!cmd_boolean (com, val, &opt.recursive))
1221 if (opt.recursive && !opt.no_dirstruct)
1222 opt.dirstruct = true;
1228 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1230 int restrict_os = opt.restrict_files_os;
1231 int restrict_ctrl = opt.restrict_files_ctrl;
1232 int restrict_case = opt.restrict_files_case;
1236 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1240 end = strchr (val, ',');
1242 end = val + strlen (val);
1244 if (VAL_IS ("unix"))
1245 restrict_os = restrict_unix;
1246 else if (VAL_IS ("windows"))
1247 restrict_os = restrict_windows;
1248 else if (VAL_IS ("lowercase"))
1249 restrict_case = restrict_lowercase;
1250 else if (VAL_IS ("uppercase"))
1251 restrict_case = restrict_uppercase;
1252 else if (VAL_IS ("nocontrol"))
1253 restrict_ctrl = false;
1257 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1258 exec_name, com, quote (val));
1265 while (*val && *end);
1269 opt.restrict_files_os = restrict_os;
1270 opt.restrict_files_ctrl = restrict_ctrl;
1271 opt.restrict_files_case = restrict_case;
1278 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1280 static const struct decode_item choices[] = {
1281 { "auto", secure_protocol_auto },
1282 { "sslv2", secure_protocol_sslv2 },
1283 { "sslv3", secure_protocol_sslv3 },
1284 { "tlsv1", secure_protocol_tlsv1 },
1286 int ok = decode_string (val, choices, countof (choices), place);
1288 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1293 /* Set all three timeout values. */
1296 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1299 if (!cmd_time (com, val, &value))
1301 opt.read_timeout = value;
1302 opt.connect_timeout = value;
1303 opt.dns_timeout = value;
1308 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1310 /* Disallow embedded newlines. */
1311 if (strchr (val, '\n'))
1313 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1314 exec_name, com, quote (val));
1317 xfree_null (opt.useragent);
1318 opt.useragent = xstrdup (val);
1322 /* The "verbose" option cannot be cmd_boolean because the variable is
1323 not bool -- it's of type int (-1 means uninitialized because of
1324 some random hackery for disallowing -q -v). */
1327 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1330 if (cmd_boolean (com, val, &flag))
1338 /* Miscellaneous useful routines. */
1340 /* A very simple atoi clone, more useful than atoi because it works on
1341 delimited strings, and has error reportage. Returns true on success,
1342 false on failure. If successful, stores result to *DEST. */
1345 simple_atoi (const char *beg, const char *end, int *dest)
1348 bool negative = false;
1349 const char *p = beg;
1351 while (p < end && c_isspace (*p))
1353 if (p < end && (*p == '-' || *p == '+'))
1355 negative = (*p == '-');
1361 /* Read negative numbers in a separate loop because the most
1362 negative integer cannot be represented as a positive number. */
1365 for (; p < end && c_isdigit (*p); p++)
1367 int next = (10 * result) + (*p - '0');
1369 return false; /* overflow */
1373 for (; p < end && c_isdigit (*p); p++)
1375 int next = (10 * result) - (*p - '0');
1377 return false; /* underflow */
1388 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1389 doesn't handle exponential notation. Returns true on success,
1390 false on failure. In case of success, stores its result to
1394 simple_atof (const char *beg, const char *end, double *dest)
1398 bool negative = false;
1399 bool seen_dot = false;
1400 bool seen_digit = false;
1403 const char *p = beg;
1405 while (p < end && c_isspace (*p))
1407 if (p < end && (*p == '-' || *p == '+'))
1409 negative = (*p == '-');
1413 for (; p < end; p++)
1419 result = (10 * result) + (ch - '0');
1421 result += (ch - '0') / (divider *= 10);
1443 /* Verify that the user-specified header in S is valid. It must
1444 contain a colon preceded by non-white-space characters and must not
1445 contain newlines. */
1448 check_user_specified_header (const char *s)
1452 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1454 /* The header MUST contain `:' preceded by at least one
1455 non-whitespace character. */
1456 if (*p != ':' || p == s)
1458 /* The header MUST NOT contain newlines. */
1459 if (strchr (s, '\n'))
1464 /* Decode VAL into a number, according to ITEMS. */
1467 decode_string (const char *val, const struct decode_item *items, int itemcount,
1471 for (i = 0; i < itemcount; i++)
1472 if (0 == strcasecmp (val, items[i].name))
1474 *place = items[i].code;
1481 void cleanup_html_url (void);
1484 /* Free the memory allocated by global variables. */
1488 /* Free external resources, close files, etc. */
1491 fclose (output_stream);
1492 /* No need to check for error because Wget flushes its output (and
1493 checks for errors) after any data arrives. */
1495 /* We're exiting anyway so there's no real need to call free()
1496 hundreds of times. Skipping the frees will make Wget exit
1499 However, when detecting leaks, it's crucial to free() everything
1500 because then you can find the real leaks, i.e. the allocated
1501 memory which grows with the size of the program. */
1507 cleanup_html_url ();
1512 extern acc_t *netrc_list;
1513 free_netrc (netrc_list);
1515 xfree_null (opt.lfilename);
1516 xfree_null (opt.dir_prefix);
1517 xfree_null (opt.input_filename);
1518 xfree_null (opt.output_document);
1519 free_vec (opt.accepts);
1520 free_vec (opt.rejects);
1521 free_vec (opt.excludes);
1522 free_vec (opt.includes);
1523 free_vec (opt.domains);
1524 free_vec (opt.follow_tags);
1525 free_vec (opt.ignore_tags);
1526 xfree_null (opt.progress_type);
1527 xfree_null (opt.ftp_user);
1528 xfree_null (opt.ftp_passwd);
1529 xfree_null (opt.ftp_proxy);
1530 xfree_null (opt.https_proxy);
1531 xfree_null (opt.http_proxy);
1532 free_vec (opt.no_proxy);
1533 xfree_null (opt.useragent);
1534 xfree_null (opt.referer);
1535 xfree_null (opt.http_user);
1536 xfree_null (opt.http_passwd);
1537 free_vec (opt.user_headers);
1539 xfree_null (opt.cert_file);
1540 xfree_null (opt.private_key);
1541 xfree_null (opt.ca_directory);
1542 xfree_null (opt.ca_cert);
1543 xfree_null (opt.random_file);
1544 xfree_null (opt.egd_file);
1546 xfree_null (opt.bind_address);
1547 xfree_null (opt.cookies_input);
1548 xfree_null (opt.cookies_output);
1549 xfree_null (opt.user);
1550 xfree_null (opt.passwd);
1551 xfree_null (opt.base_href);
1553 #endif /* DEBUG_MALLOC */
1556 /* Unit testing routines. */
1561 test_commands_sorted()
1563 int prev_idx = 0, next_idx = 1;
1564 int command_count = countof (commands) - 1;
1566 while (next_idx <= command_count)
1568 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1571 mu_assert ("FAILED", false);
1584 test_cmd_spec_restrict_file_names()
1589 int expected_restrict_files_os;
1590 int expected_restrict_files_ctrl;
1591 int expected_restrict_files_case;
1594 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1595 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1596 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1597 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1600 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1605 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1608 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1609 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1610 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1611 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1613 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1614 res == test_array[i].result
1615 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1616 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1617 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1623 #endif /* TESTING */