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 },
185 { "iri", &opt.enable_iri, cmd_boolean },
187 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
188 { "limitrate", &opt.limit_rate, cmd_bytes },
189 { "loadcookies", &opt.cookies_input, cmd_file },
191 { "locale", &opt.locale, cmd_string },
193 { "logfile", &opt.lfilename, cmd_file },
194 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
195 { "maxredirect", &opt.max_redirect, cmd_number },
196 { "mirror", NULL, cmd_spec_mirror },
197 { "netrc", &opt.netrc, cmd_boolean },
198 { "noclobber", &opt.noclobber, cmd_boolean },
199 { "noparent", &opt.no_parent, cmd_boolean },
200 { "noproxy", &opt.no_proxy, cmd_vector },
201 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
202 { "outputdocument", &opt.output_document, cmd_file },
203 { "pagerequisites", &opt.page_requisites, cmd_boolean },
204 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
205 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
206 { "password", &opt.passwd, cmd_string },
207 { "postdata", &opt.post_data, cmd_string },
208 { "postfile", &opt.post_file_name, cmd_file },
209 { "preferfamily", NULL, cmd_spec_prefer_family },
210 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
212 { "privatekey", &opt.private_key, cmd_file },
213 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
215 { "progress", &opt.progress_type, cmd_spec_progress },
216 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
217 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
218 { "proxypassword", &opt.proxy_passwd, cmd_string },
219 { "proxyuser", &opt.proxy_user, cmd_string },
220 { "quiet", &opt.quiet, cmd_boolean },
221 { "quota", &opt.quota, cmd_bytes_sum },
223 { "randomfile", &opt.random_file, cmd_file },
225 { "randomwait", &opt.random_wait, cmd_boolean },
226 { "readtimeout", &opt.read_timeout, cmd_time },
227 { "reclevel", &opt.reclevel, cmd_number_inf },
228 { "recursive", NULL, cmd_spec_recursive },
229 { "referer", &opt.referer, cmd_string },
230 { "reject", &opt.rejects, cmd_vector },
231 { "relativeonly", &opt.relative_only, cmd_boolean },
233 { "remoteencoding", &opt.encoding_remote, cmd_string },
235 { "removelisting", &opt.remove_listing, cmd_boolean },
236 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
237 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
238 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
239 { "robots", &opt.use_robots, cmd_boolean },
240 { "savecookies", &opt.cookies_output, cmd_file },
241 { "saveheaders", &opt.save_headers, cmd_boolean },
243 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
245 { "serverresponse", &opt.server_response, cmd_boolean },
246 { "spanhosts", &opt.spanhost, cmd_boolean },
247 { "spider", &opt.spider, cmd_boolean },
248 { "strictcomments", &opt.strict_comments, cmd_boolean },
249 { "timeout", NULL, cmd_spec_timeout },
250 { "timestamping", &opt.timestamping, cmd_boolean },
251 { "tries", &opt.ntry, cmd_number_inf },
252 { "useproxy", &opt.use_proxy, cmd_boolean },
253 { "user", &opt.user, cmd_string },
254 { "useragent", NULL, cmd_spec_useragent },
255 { "verbose", NULL, cmd_spec_verbose },
256 { "wait", &opt.wait, cmd_time },
257 { "waitretry", &opt.waitretry, cmd_time },
259 { "wdebug", &opt.wdebug, cmd_boolean },
263 /* Look up CMDNAME in the commands[] and return its position in the
264 array. If CMDNAME is not found, return -1. */
267 command_by_name (const char *cmdname)
269 /* Use binary search for speed. Wget has ~100 commands, which
270 guarantees a worst case performance of 7 string comparisons. */
271 int lo = 0, hi = countof (commands) - 1;
275 int mid = (lo + hi) >> 1;
276 int cmp = strcasecmp (cmdname, commands[mid].name);
287 /* Reset the variables to default values. */
293 /* Most of the default values are 0 (and 0.0, NULL, and false).
294 Just reset everything, and fill in the non-zero values. Note
295 that initializing pointers to NULL this way is technically
296 illegal, but porting Wget to a machine where NULL is not all-zero
297 bit pattern will be the least of the implementors' worries. */
304 opt.add_hostdir = true;
308 opt.http_keep_alive = true;
309 opt.use_proxy = true;
310 tmp = getenv ("no_proxy");
312 opt.no_proxy = sepstring (tmp);
313 opt.prefer_family = prefer_none;
314 opt.allow_cache = true;
316 opt.read_timeout = 900;
317 opt.use_robots = true;
319 opt.remove_listing = true;
321 opt.dot_bytes = 1024;
322 opt.dot_spacing = 10;
323 opt.dots_in_line = 50;
325 opt.dns_cache = true;
329 opt.check_cert = true;
332 /* The default for file name restriction defaults to the OS type. */
333 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
334 opt.restrict_files_os = restrict_windows;
336 opt.restrict_files_os = restrict_unix;
338 opt.restrict_files_ctrl = true;
339 opt.restrict_files_case = restrict_no_case_restriction;
341 opt.max_redirect = 20;
344 /* Return the user's home directory (strdup-ed), or NULL if none is
349 char *home = getenv ("HOME");
354 /* Under MSDOS, if $HOME isn't defined, use the directory where
355 `wget.exe' resides. */
356 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
357 char *p, buf[PATH_MAX];
359 strcpy (buf, _w32_get_argv0 ());
360 p = strrchr (buf, '/'); /* djgpp */
362 p = strrchr (buf, '\\'); /* others */
366 #elif !defined(WINDOWS)
367 /* If HOME is not defined, try getting it from the password
369 struct passwd *pwd = getpwuid (getuid ());
370 if (!pwd || !pwd->pw_dir)
374 /* Under Windows, if $HOME isn't defined, use the directory where
375 `wget.exe' resides. */
380 return home ? xstrdup (home) : NULL;
383 /* Check the 'WGETRC' environment variable and return the file name
384 if 'WGETRC' is set and is a valid file.
385 If the `WGETRC' variable exists but the file does not exist, the
386 function will exit(). */
388 wgetrc_env_file_name (void)
390 char *env = getenv ("WGETRC");
393 if (!file_exists_p (env))
395 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
399 return xstrdup (env);
403 /* Check for the existance of '$HOME/.wgetrc' and return it's path
404 if it exists and is set. */
406 wgetrc_user_file_name (void)
408 char *home = home_dir();
411 file = aprintf ("%s/.wgetrc", home);
415 if (!file_exists_p (file))
422 /* Return the path to the user's .wgetrc. This is either the value of
423 `WGETRC' environment variable, or `$HOME/.wgetrc'.
425 Additionally, for windows, look in the directory where wget.exe
428 wgetrc_file_name (void)
430 char *file = wgetrc_env_file_name ();
434 file = wgetrc_user_file_name ();
437 /* Under Windows, if we still haven't found .wgetrc, look for the file
438 `wget.ini' in the directory where `wget.exe' resides; we do this for
439 backward compatibility with previous versions of Wget.
440 SYSTEM_WGETRC should not be defined under WINDOWS. */
441 if (!file || !file_exists_p (file))
447 file = aprintf ("%s/wget.ini", home);
453 if (!file_exists_p (file))
461 /* Return values of parse_line. */
469 static enum parse_line parse_line (const char *, char **, char **, int *);
470 static bool setval_internal (int, const char *, const char *);
472 /* Initialize variables from a wgetrc file. Returns zero (failure) if
473 there were errors in the file. */
476 run_wgetrc (const char *file)
483 fp = fopen (file, "rb");
486 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
487 file, strerror (errno));
488 return true; /* not a fatal error */
490 enable_tilde_expansion = true;
492 while ((line = read_whole_line (fp)) != NULL)
494 char *com = NULL, *val = NULL;
497 /* Parse the line. */
498 switch (parse_line (line, &com, &val, &comind))
501 /* If everything is OK, set the value. */
502 if (!setval_internal (comind, com, val))
504 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
505 exec_name, file, ln);
509 case line_syntax_error:
510 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
511 exec_name, file, ln);
514 case line_unknown_command:
515 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
516 exec_name, quote (com), file, ln);
529 enable_tilde_expansion = false;
535 /* Initialize the defaults and run the system wgetrc and user's own
543 /* Load the hard-coded defaults. */
546 /* If SYSTEM_WGETRC is defined, use it. */
548 if (file_exists_p (SYSTEM_WGETRC))
549 ok &= run_wgetrc (SYSTEM_WGETRC);
551 /* Override it with your own, if one exists. */
552 file = wgetrc_file_name ();
555 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
556 something like realpath() before comparing them with `strcmp' */
558 if (!strcmp (file, SYSTEM_WGETRC))
560 fprintf (stderr, _("\
561 %s: Warning: Both system and user wgetrc point to %s.\n"),
562 exec_name, quote (file));
566 ok &= run_wgetrc (file);
568 /* If there were errors processing either `.wgetrc', abort. */
576 /* Remove dashes and underscores from S, modifying S in the
582 char *t = s; /* t - tortoise */
583 char *h = s; /* h - hare */
585 if (*h == '_' || *h == '-')
592 /* Parse the line pointed by line, with the syntax:
593 <sp>* command <sp>* = <sp>* value <sp>*
594 Uses malloc to allocate space for command and value.
596 Returns one of line_ok, line_empty, line_syntax_error, or
597 line_unknown_command.
599 In case of line_ok, *COM and *VAL point to freshly allocated
600 strings, and *COMIND points to com's index. In case of error or
601 empty line, their values are unmodified. */
603 static enum parse_line
604 parse_line (const char *line, char **com, char **val, int *comind)
607 const char *end = line + strlen (line);
608 const char *cmdstart, *cmdend;
609 const char *valstart, *valend;
614 /* Skip leading and trailing whitespace. */
615 while (*line && c_isspace (*line))
617 while (end > line && c_isspace (end[-1]))
620 /* Skip empty lines and comments. */
621 if (!*line || *line == '#')
627 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
631 /* Skip '=', as well as any space before or after it. */
632 while (p < end && c_isspace (*p))
634 if (p == end || *p != '=')
635 return line_syntax_error;
637 while (p < end && c_isspace (*p))
643 /* The syntax is valid (even though the command might not be). Fill
644 in the command name and value. */
645 *com = strdupdelim (cmdstart, cmdend);
646 *val = strdupdelim (valstart, valend);
648 /* The line now known to be syntactically correct. Check whether
649 the command is valid. */
650 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
652 ind = command_by_name (cmdcopy);
654 return line_unknown_command;
656 /* Report success to the caller. */
661 /* Run commands[comind].action. */
664 setval_internal (int comind, const char *com, const char *val)
666 assert (0 <= comind && ((size_t) comind) < countof (commands));
667 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
668 return commands[comind].action (com, val, commands[comind].place);
671 /* Run command COM with value VAL. If running the command produces an
672 error, report the error and exit.
674 This is intended to be called from main() to modify Wget's behavior
675 through command-line switches. Since COM is hard-coded in main(),
676 it is not canonicalized, and this aborts when COM is not found.
678 If COMIND's are exported to init.h, this function will be changed
679 to accept COMIND directly. */
682 setoptval (const char *com, const char *val, const char *optname)
684 /* Prepend "--" to OPTNAME. */
685 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
688 strcpy (dd_optname + 2, optname);
690 assert (val != NULL);
691 if (!setval_internal (command_by_name (com), dd_optname, val))
695 /* Parse OPT into command and value and run it. For example,
696 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
697 This is used by the `--execute' flag in main.c. */
700 run_command (const char *opt)
704 switch (parse_line (opt, &com, &val, &comind))
707 if (!setval_internal (comind, com, val))
713 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
714 exec_name, quote (opt));
719 /* Generic helper functions, for use with `commands'. */
721 /* Forward declarations: */
726 static bool decode_string (const char *, const struct decode_item *, int, int *);
727 static bool simple_atoi (const char *, const char *, int *);
728 static bool simple_atof (const char *, const char *, double *);
730 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
732 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
733 && c_tolower((p)[1]) == (c1) \
736 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
737 && c_tolower((p)[1]) == (c1) \
738 && c_tolower((p)[2]) == (c2) \
742 /* Store the boolean value from VAL to PLACE. COM is ignored,
743 except for error messages. */
745 cmd_boolean (const char *com, const char *val, void *place)
749 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
750 /* "on", "yes" and "1" mean true. */
752 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
753 /* "off", "no" and "0" mean false. */
758 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
759 exec_name, com, quote (val));
763 *(bool *) place = value;
767 /* Set the non-negative integer value from VAL to PLACE. With
768 incorrect specification, the number remains unchanged. */
770 cmd_number (const char *com, const char *val, void *place)
772 if (!simple_atoi (val, val + strlen (val), place)
773 || *(int *) place < 0)
775 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
776 exec_name, com, quote (val));
782 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
784 cmd_number_inf (const char *com, const char *val, void *place)
786 if (!strcasecmp (val, "inf"))
791 return cmd_number (com, val, place);
794 /* Copy (strdup) the string at COM to a new location and place a
795 pointer to *PLACE. */
797 cmd_string (const char *com, const char *val, void *place)
799 char **pstring = (char **)place;
801 xfree_null (*pstring);
802 *pstring = xstrdup (val);
806 #if defined(WINDOWS) || defined(MSDOS)
807 # define ISSEP(c) ((c) == '/' || (c) == '\\')
809 # define ISSEP(c) ((c) == '/')
812 /* Like the above, but handles tilde-expansion when reading a user's
813 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
814 gets expanded to the user's home directory. */
816 cmd_file (const char *com, const char *val, void *place)
818 char **pstring = (char **)place;
820 xfree_null (*pstring);
822 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
824 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
827 *pstring = xstrdup (val);
832 char *home = home_dir ();
836 homelen = strlen (home);
837 while (homelen && ISSEP (home[homelen - 1]))
838 home[--homelen] = '\0';
840 /* Skip the leading "~/". */
841 for (++val; ISSEP (*val); val++)
844 *pstring = concat_strings (home, "/", val, (char *) 0);
847 #if defined(WINDOWS) || defined(MSDOS)
848 /* Convert "\" to "/". */
851 for (s = *pstring; *s; s++)
859 /* Like cmd_file, but strips trailing '/' characters. */
861 cmd_directory (const char *com, const char *val, void *place)
865 /* Call cmd_file() for tilde expansion and separator
866 canonicalization (backslash -> slash under Windows). These
867 things should perhaps be in a separate function. */
868 if (!cmd_file (com, val, place))
873 while (t > s && *--t == '/')
879 /* Split VAL by space to a vector of values, and append those values
880 to vector pointed to by the PLACE argument. If VAL is empty, the
881 PLACE vector is cleared instead. */
884 cmd_vector (const char *com, const char *val, void *place)
886 char ***pvec = (char ***)place;
889 *pvec = merge_vecs (*pvec, sepstring (val));
899 cmd_directory_vector (const char *com, const char *val, void *place)
901 char ***pvec = (char ***)place;
905 /* Strip the trailing slashes from directories. */
908 seps = sepstring (val);
909 for (t = seps; t && *t; t++)
911 int len = strlen (*t);
912 /* Skip degenerate case of root directory. */
915 if ((*t)[len - 1] == '/')
916 (*t)[len - 1] = '\0';
919 *pvec = merge_vecs (*pvec, seps);
929 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
930 "100k" or "2.5G" to a floating point number. */
933 parse_bytes_helper (const char *val, double *result)
936 const char *end = val + strlen (val);
938 /* Check for "inf". */
939 if (0 == strcmp (val, "inf"))
945 /* Strip trailing whitespace. */
946 while (val < end && c_isspace (end[-1]))
951 switch (c_tolower (end[-1]))
954 --end, mult = 1024.0;
957 --end, mult = 1048576.0;
960 --end, mult = 1073741824.0;
963 --end, mult = 1099511627776.0;
966 /* Not a recognized suffix: assume it's a digit. (If not,
967 simple_atof will raise an error.) */
971 /* Skip leading and trailing whitespace. */
972 while (val < end && c_isspace (*val))
974 while (val < end && c_isspace (end[-1]))
979 if (!simple_atof (val, end, &number) || number < 0)
982 *result = number * mult;
986 /* Parse VAL as a number and set its value to PLACE (which should
989 By default, the value is assumed to be in bytes. If "K", "M", or
990 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
991 1<<30, respectively. Floating point values are allowed and are
992 cast to integer before use. The idea is to be able to use things
993 like 1.5k instead of "1536".
995 The string "inf" is returned as 0.
997 In case of error, false is returned and memory pointed to by PLACE
998 remains unmodified. */
1001 cmd_bytes (const char *com, const char *val, void *place)
1004 if (!parse_bytes_helper (val, &byte_value))
1006 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1007 exec_name, com, quote (val));
1010 *(wgint *)place = (wgint)byte_value;
1014 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1015 SIZE_SUM. It works by converting the string to double, therefore
1016 working with values up to 2^53-1 without loss of precision. This
1017 value (8192 TB) is large enough to serve for a while. */
1020 cmd_bytes_sum (const char *com, const char *val, void *place)
1023 if (!parse_bytes_helper (val, &byte_value))
1025 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1026 exec_name, com, quote (val));
1029 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1033 /* Store the value of VAL to *OUT. The value is a time period, by
1034 default expressed in seconds, but also accepting suffixes "m", "h",
1035 "d", and "w" for minutes, hours, days, and weeks respectively. */
1038 cmd_time (const char *com, const char *val, void *place)
1040 double number, mult;
1041 const char *end = val + strlen (val);
1043 /* Strip trailing whitespace. */
1044 while (val < end && c_isspace (end[-1]))
1050 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1051 exec_name, com, quote (val));
1055 switch (c_tolower (end[-1]))
1058 --end, mult = 1; /* seconds */
1061 --end, mult = 60; /* minutes */
1064 --end, mult = 3600; /* hours */
1067 --end, mult = 86400.0; /* days */
1070 --end, mult = 604800.0; /* weeks */
1073 /* Not a recognized suffix: assume it belongs to the number.
1074 (If not, simple_atof will raise an error.) */
1078 /* Skip leading and trailing whitespace. */
1079 while (val < end && c_isspace (*val))
1081 while (val < end && c_isspace (end[-1]))
1086 if (!simple_atof (val, end, &number))
1089 *(double *)place = number * mult;
1095 cmd_cert_type (const char *com, const char *val, void *place)
1097 static const struct decode_item choices[] = {
1098 { "pem", keyfile_pem },
1099 { "der", keyfile_asn1 },
1100 { "asn1", keyfile_asn1 },
1102 int ok = decode_string (val, choices, countof (choices), place);
1104 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1109 /* Specialized helper functions, used by `commands' to handle some
1110 options specially. */
1112 static bool check_user_specified_header (const char *);
1115 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1117 if (!cmd_boolean (com, val, &opt.dirstruct))
1119 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1120 must be affected inversely. */
1122 opt.no_dirstruct = false;
1124 opt.no_dirstruct = true;
1129 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1131 /* Empty value means reset the list of headers. */
1134 free_vec (opt.user_headers);
1135 opt.user_headers = NULL;
1139 if (!check_user_specified_header (val))
1141 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1142 exec_name, com, quote (val));
1145 opt.user_headers = vec_append (opt.user_headers, val);
1150 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1152 int flag = cmd_boolean (com, val, &opt.htmlify);
1153 if (flag && !opt.htmlify)
1154 opt.remove_listing = false;
1158 /* Set the "mirror" mode. It means: recursive download, timestamping,
1159 no limit on max. recursion depth, and don't remove listings. */
1162 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1166 if (!cmd_boolean (com, val, &mirror))
1170 opt.recursive = true;
1171 if (!opt.no_dirstruct)
1172 opt.dirstruct = true;
1173 opt.timestamping = true;
1174 opt.reclevel = INFINITE_RECURSION;
1175 opt.remove_listing = false;
1180 /* Validate --prefer-family and set the choice. Allowed values are
1181 "IPv4", "IPv6", and "none". */
1184 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1186 static const struct decode_item choices[] = {
1187 { "IPv4", prefer_ipv4 },
1188 { "IPv6", prefer_ipv6 },
1189 { "none", prefer_none },
1191 int prefer_family = prefer_none;
1192 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1194 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1195 opt.prefer_family = prefer_family;
1199 /* Set progress.type to VAL, but verify that it's a valid progress
1200 implementation before that. */
1203 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1205 if (!valid_progress_implementation_p (val))
1207 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1208 exec_name, com, quote (val));
1211 xfree_null (opt.progress_type);
1213 /* Don't call set_progress_implementation here. It will be called
1214 in main() when it becomes clear what the log output is. */
1215 opt.progress_type = xstrdup (val);
1219 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1220 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1224 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1226 if (!cmd_boolean (com, val, &opt.recursive))
1230 if (opt.recursive && !opt.no_dirstruct)
1231 opt.dirstruct = true;
1237 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1239 int restrict_os = opt.restrict_files_os;
1240 int restrict_ctrl = opt.restrict_files_ctrl;
1241 int restrict_case = opt.restrict_files_case;
1245 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1249 end = strchr (val, ',');
1251 end = val + strlen (val);
1253 if (VAL_IS ("unix"))
1254 restrict_os = restrict_unix;
1255 else if (VAL_IS ("windows"))
1256 restrict_os = restrict_windows;
1257 else if (VAL_IS ("lowercase"))
1258 restrict_case = restrict_lowercase;
1259 else if (VAL_IS ("uppercase"))
1260 restrict_case = restrict_uppercase;
1261 else if (VAL_IS ("nocontrol"))
1262 restrict_ctrl = false;
1266 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1267 exec_name, com, quote (val));
1274 while (*val && *end);
1278 opt.restrict_files_os = restrict_os;
1279 opt.restrict_files_ctrl = restrict_ctrl;
1280 opt.restrict_files_case = restrict_case;
1287 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1289 static const struct decode_item choices[] = {
1290 { "auto", secure_protocol_auto },
1291 { "sslv2", secure_protocol_sslv2 },
1292 { "sslv3", secure_protocol_sslv3 },
1293 { "tlsv1", secure_protocol_tlsv1 },
1295 int ok = decode_string (val, choices, countof (choices), place);
1297 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1302 /* Set all three timeout values. */
1305 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1308 if (!cmd_time (com, val, &value))
1310 opt.read_timeout = value;
1311 opt.connect_timeout = value;
1312 opt.dns_timeout = value;
1317 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1319 /* Disallow embedded newlines. */
1320 if (strchr (val, '\n'))
1322 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1323 exec_name, com, quote (val));
1326 xfree_null (opt.useragent);
1327 opt.useragent = xstrdup (val);
1331 /* The "verbose" option cannot be cmd_boolean because the variable is
1332 not bool -- it's of type int (-1 means uninitialized because of
1333 some random hackery for disallowing -q -v). */
1336 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1339 if (cmd_boolean (com, val, &flag))
1347 /* Miscellaneous useful routines. */
1349 /* A very simple atoi clone, more useful than atoi because it works on
1350 delimited strings, and has error reportage. Returns true on success,
1351 false on failure. If successful, stores result to *DEST. */
1354 simple_atoi (const char *beg, const char *end, int *dest)
1357 bool negative = false;
1358 const char *p = beg;
1360 while (p < end && c_isspace (*p))
1362 if (p < end && (*p == '-' || *p == '+'))
1364 negative = (*p == '-');
1370 /* Read negative numbers in a separate loop because the most
1371 negative integer cannot be represented as a positive number. */
1374 for (; p < end && c_isdigit (*p); p++)
1376 int next = (10 * result) + (*p - '0');
1378 return false; /* overflow */
1382 for (; p < end && c_isdigit (*p); p++)
1384 int next = (10 * result) - (*p - '0');
1386 return false; /* underflow */
1397 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1398 doesn't handle exponential notation. Returns true on success,
1399 false on failure. In case of success, stores its result to
1403 simple_atof (const char *beg, const char *end, double *dest)
1407 bool negative = false;
1408 bool seen_dot = false;
1409 bool seen_digit = false;
1412 const char *p = beg;
1414 while (p < end && c_isspace (*p))
1416 if (p < end && (*p == '-' || *p == '+'))
1418 negative = (*p == '-');
1422 for (; p < end; p++)
1428 result = (10 * result) + (ch - '0');
1430 result += (ch - '0') / (divider *= 10);
1452 /* Verify that the user-specified header in S is valid. It must
1453 contain a colon preceded by non-white-space characters and must not
1454 contain newlines. */
1457 check_user_specified_header (const char *s)
1461 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1463 /* The header MUST contain `:' preceded by at least one
1464 non-whitespace character. */
1465 if (*p != ':' || p == s)
1467 /* The header MUST NOT contain newlines. */
1468 if (strchr (s, '\n'))
1473 /* Decode VAL into a number, according to ITEMS. */
1476 decode_string (const char *val, const struct decode_item *items, int itemcount,
1480 for (i = 0; i < itemcount; i++)
1481 if (0 == strcasecmp (val, items[i].name))
1483 *place = items[i].code;
1490 void cleanup_html_url (void);
1493 /* Free the memory allocated by global variables. */
1497 /* Free external resources, close files, etc. */
1500 fclose (output_stream);
1501 /* No need to check for error because Wget flushes its output (and
1502 checks for errors) after any data arrives. */
1504 /* We're exiting anyway so there's no real need to call free()
1505 hundreds of times. Skipping the frees will make Wget exit
1508 However, when detecting leaks, it's crucial to free() everything
1509 because then you can find the real leaks, i.e. the allocated
1510 memory which grows with the size of the program. */
1516 cleanup_html_url ();
1521 extern acc_t *netrc_list;
1522 free_netrc (netrc_list);
1524 xfree_null (opt.lfilename);
1525 xfree_null (opt.dir_prefix);
1526 xfree_null (opt.input_filename);
1527 xfree_null (opt.output_document);
1528 free_vec (opt.accepts);
1529 free_vec (opt.rejects);
1530 free_vec (opt.excludes);
1531 free_vec (opt.includes);
1532 free_vec (opt.domains);
1533 free_vec (opt.follow_tags);
1534 free_vec (opt.ignore_tags);
1535 xfree_null (opt.progress_type);
1536 xfree_null (opt.ftp_user);
1537 xfree_null (opt.ftp_passwd);
1538 xfree_null (opt.ftp_proxy);
1539 xfree_null (opt.https_proxy);
1540 xfree_null (opt.http_proxy);
1541 free_vec (opt.no_proxy);
1542 xfree_null (opt.useragent);
1543 xfree_null (opt.referer);
1544 xfree_null (opt.http_user);
1545 xfree_null (opt.http_passwd);
1546 free_vec (opt.user_headers);
1548 xfree_null (opt.cert_file);
1549 xfree_null (opt.private_key);
1550 xfree_null (opt.ca_directory);
1551 xfree_null (opt.ca_cert);
1552 xfree_null (opt.random_file);
1553 xfree_null (opt.egd_file);
1555 xfree_null (opt.bind_address);
1556 xfree_null (opt.cookies_input);
1557 xfree_null (opt.cookies_output);
1558 xfree_null (opt.user);
1559 xfree_null (opt.passwd);
1560 #endif /* DEBUG_MALLOC */
1563 /* Unit testing routines. */
1568 test_commands_sorted()
1570 int prev_idx = 0, next_idx = 1;
1571 int command_count = countof (commands) - 1;
1573 while (next_idx <= command_count)
1575 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1578 mu_assert ("FAILED", false);
1591 test_cmd_spec_restrict_file_names()
1596 int expected_restrict_files_os;
1597 int expected_restrict_files_ctrl;
1598 int expected_restrict_files_case;
1601 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1602 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1603 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1604 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1607 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1612 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1615 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1616 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1617 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1618 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1620 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1621 res == test_array[i].result
1622 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1623 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1624 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1630 #endif /* TESTING */