1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
3 2005, 2006, 2007, 2008, 2009, 2010 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. */
41 /* not all systems provide PATH_MAX in limits.h */
43 # include <sys/param.h>
45 # define PATH_MAX MAXPATHLEN
60 #include "recur.h" /* for INFINITE_RECURSION */
61 #include "convert.h" /* for convert_cleanup */
62 #include "res.h" /* for res_cleanup */
63 #include "http.h" /* for http_cleanup */
64 #include "retr.h" /* for output_stream */
72 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
74 CMD_DECLARE (cmd_boolean);
75 CMD_DECLARE (cmd_bytes);
76 CMD_DECLARE (cmd_bytes_sum);
78 CMD_DECLARE (cmd_cert_type);
80 CMD_DECLARE (cmd_directory_vector);
81 CMD_DECLARE (cmd_number);
82 CMD_DECLARE (cmd_number_inf);
83 CMD_DECLARE (cmd_string);
84 CMD_DECLARE (cmd_file);
85 CMD_DECLARE (cmd_directory);
86 CMD_DECLARE (cmd_time);
87 CMD_DECLARE (cmd_vector);
89 CMD_DECLARE (cmd_spec_dirstruct);
90 CMD_DECLARE (cmd_spec_header);
91 CMD_DECLARE (cmd_spec_htmlify);
92 CMD_DECLARE (cmd_spec_mirror);
93 CMD_DECLARE (cmd_spec_prefer_family);
94 CMD_DECLARE (cmd_spec_progress);
95 CMD_DECLARE (cmd_spec_recursive);
96 CMD_DECLARE (cmd_spec_restrict_file_names);
98 CMD_DECLARE (cmd_spec_secure_protocol);
100 CMD_DECLARE (cmd_spec_timeout);
101 CMD_DECLARE (cmd_spec_useragent);
102 CMD_DECLARE (cmd_spec_verbose);
104 /* List of recognized commands, each consisting of name, place and
105 function. When adding a new command, simply add it to the list,
106 but be sure to keep the list sorted alphabetically, as
107 command_by_name's binary search depends on it. Also, be sure to
108 add any entries that allocate memory (e.g. cmd_string and
109 cmd_vector) to the cleanup() function below. */
111 static const struct {
114 bool (*action) (const char *, const char *, void *);
116 /* KEEP THIS LIST ALPHABETICALLY SORTED */
117 { "accept", &opt.accepts, cmd_vector },
118 { "addhostdir", &opt.add_hostdir, cmd_boolean },
119 { "adjustextension", &opt.adjust_extension, cmd_boolean },
120 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
121 { "askpassword", &opt.ask_passwd, cmd_boolean },
122 { "authnochallenge", &opt.auth_without_challenge,
124 { "background", &opt.background, cmd_boolean },
125 { "backupconverted", &opt.backup_converted, cmd_boolean },
126 { "backups", &opt.backups, cmd_number },
127 { "base", &opt.base_href, cmd_string },
128 { "bindaddress", &opt.bind_address, cmd_string },
130 { "cacertificate", &opt.ca_cert, cmd_file },
132 { "cache", &opt.allow_cache, cmd_boolean },
134 { "cadirectory", &opt.ca_directory, cmd_directory },
135 { "certificate", &opt.cert_file, cmd_file },
136 { "certificatetype", &opt.cert_type, cmd_cert_type },
137 { "checkcertificate", &opt.check_cert, cmd_boolean },
139 { "connecttimeout", &opt.connect_timeout, cmd_time },
140 { "contentdisposition", &opt.content_disposition, cmd_boolean },
141 { "continue", &opt.always_rest, cmd_boolean },
142 { "convertlinks", &opt.convert_links, cmd_boolean },
143 { "cookies", &opt.cookies, cmd_boolean },
144 { "cutdirs", &opt.cut_dirs, cmd_number },
146 { "debug", &opt.debug, cmd_boolean },
148 { "defaultpage", &opt.default_page, cmd_string},
149 { "deleteafter", &opt.delete_after, cmd_boolean },
150 { "dirprefix", &opt.dir_prefix, cmd_directory },
151 { "dirstruct", NULL, cmd_spec_dirstruct },
152 { "dnscache", &opt.dns_cache, cmd_boolean },
153 { "dnstimeout", &opt.dns_timeout, cmd_time },
154 { "domains", &opt.domains, cmd_vector },
155 { "dotbytes", &opt.dot_bytes, cmd_bytes },
156 { "dotsinline", &opt.dots_in_line, cmd_number },
157 { "dotspacing", &opt.dot_spacing, cmd_number },
158 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
160 { "egdfile", &opt.egd_file, cmd_file },
162 { "excludedirectories", &opt.excludes, cmd_directory_vector },
163 { "excludedomains", &opt.exclude_domains, cmd_vector },
164 { "followftp", &opt.follow_ftp, cmd_boolean },
165 { "followtags", &opt.follow_tags, cmd_vector },
166 { "forcehtml", &opt.force_html, cmd_boolean },
167 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
168 { "ftppassword", &opt.ftp_passwd, cmd_string },
169 { "ftpproxy", &opt.ftp_proxy, cmd_string },
171 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
172 #endif /* def __VMS */
173 { "ftpuser", &opt.ftp_user, cmd_string },
174 { "glob", &opt.ftp_glob, cmd_boolean },
175 { "header", NULL, cmd_spec_header },
176 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
177 { "htmlify", NULL, cmd_spec_htmlify },
178 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
179 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
180 { "httppassword", &opt.http_passwd, cmd_string },
181 { "httpproxy", &opt.http_proxy, cmd_string },
182 { "httpsproxy", &opt.https_proxy, cmd_string },
183 { "httpuser", &opt.http_user, cmd_string },
184 { "ignorecase", &opt.ignore_case, cmd_boolean },
185 { "ignorelength", &opt.ignore_length, cmd_boolean },
186 { "ignoretags", &opt.ignore_tags, cmd_vector },
187 { "includedirectories", &opt.includes, cmd_directory_vector },
189 { "inet4only", &opt.ipv4_only, cmd_boolean },
190 { "inet6only", &opt.ipv6_only, cmd_boolean },
192 { "input", &opt.input_filename, cmd_file },
193 { "iri", &opt.enable_iri, cmd_boolean },
194 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
195 { "limitrate", &opt.limit_rate, cmd_bytes },
196 { "loadcookies", &opt.cookies_input, cmd_file },
197 { "localencoding", &opt.locale, cmd_string },
198 { "logfile", &opt.lfilename, cmd_file },
199 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
200 { "maxredirect", &opt.max_redirect, cmd_number },
201 { "mirror", NULL, cmd_spec_mirror },
202 { "netrc", &opt.netrc, cmd_boolean },
203 { "noclobber", &opt.noclobber, cmd_boolean },
204 { "noparent", &opt.no_parent, cmd_boolean },
205 { "noproxy", &opt.no_proxy, cmd_vector },
206 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
207 { "outputdocument", &opt.output_document, cmd_file },
208 { "pagerequisites", &opt.page_requisites, cmd_boolean },
209 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
210 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
211 { "password", &opt.passwd, cmd_string },
212 { "postdata", &opt.post_data, cmd_string },
213 { "postfile", &opt.post_file_name, cmd_file },
214 { "preferfamily", NULL, cmd_spec_prefer_family },
215 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
217 { "privatekey", &opt.private_key, cmd_file },
218 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
220 { "progress", &opt.progress_type, cmd_spec_progress },
221 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
222 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
223 { "proxypassword", &opt.proxy_passwd, cmd_string },
224 { "proxyuser", &opt.proxy_user, cmd_string },
225 { "quiet", &opt.quiet, cmd_boolean },
226 { "quota", &opt.quota, cmd_bytes_sum },
228 { "randomfile", &opt.random_file, cmd_file },
230 { "randomwait", &opt.random_wait, cmd_boolean },
231 { "readtimeout", &opt.read_timeout, cmd_time },
232 { "reclevel", &opt.reclevel, cmd_number_inf },
233 { "recursive", NULL, cmd_spec_recursive },
234 { "referer", &opt.referer, cmd_string },
235 { "reject", &opt.rejects, cmd_vector },
236 { "relativeonly", &opt.relative_only, cmd_boolean },
237 { "remoteencoding", &opt.encoding_remote, cmd_string },
238 { "removelisting", &opt.remove_listing, cmd_boolean },
239 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
240 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
241 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
242 { "robots", &opt.use_robots, cmd_boolean },
243 { "savecookies", &opt.cookies_output, cmd_file },
244 { "saveheaders", &opt.save_headers, cmd_boolean },
246 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
248 { "serverresponse", &opt.server_response, cmd_boolean },
249 { "spanhosts", &opt.spanhost, cmd_boolean },
250 { "spider", &opt.spider, cmd_boolean },
251 { "strictcomments", &opt.strict_comments, cmd_boolean },
252 { "timeout", NULL, cmd_spec_timeout },
253 { "timestamping", &opt.timestamping, cmd_boolean },
254 { "tries", &opt.ntry, cmd_number_inf },
255 { "trustservernames", &opt.trustservernames, cmd_boolean },
256 { "unlink", &opt.unlink, cmd_boolean },
257 { "useproxy", &opt.use_proxy, cmd_boolean },
258 { "user", &opt.user, cmd_string },
259 { "useragent", NULL, cmd_spec_useragent },
260 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
261 { "verbose", NULL, cmd_spec_verbose },
262 { "wait", &opt.wait, cmd_time },
263 { "waitretry", &opt.waitretry, cmd_time },
265 { "wdebug", &opt.wdebug, cmd_boolean },
269 /* Look up CMDNAME in the commands[] and return its position in the
270 array. If CMDNAME is not found, return -1. */
273 command_by_name (const char *cmdname)
275 /* Use binary search for speed. Wget has ~100 commands, which
276 guarantees a worst case performance of 7 string comparisons. */
277 int lo = 0, hi = countof (commands) - 1;
281 int mid = (lo + hi) >> 1;
282 int cmp = strcasecmp (cmdname, commands[mid].name);
293 /* Reset the variables to default values. */
299 /* Most of the default values are 0 (and 0.0, NULL, and false).
300 Just reset everything, and fill in the non-zero values. Note
301 that initializing pointers to NULL this way is technically
302 illegal, but porting Wget to a machine where NULL is not all-zero
303 bit pattern will be the least of the implementors' worries. */
310 opt.add_hostdir = true;
314 opt.http_keep_alive = true;
315 opt.use_proxy = true;
316 tmp = getenv ("no_proxy");
318 opt.no_proxy = sepstring (tmp);
319 opt.prefer_family = prefer_none;
320 opt.allow_cache = true;
322 opt.read_timeout = 900;
323 opt.use_robots = true;
325 opt.remove_listing = true;
327 opt.dot_bytes = 1024;
328 opt.dot_spacing = 10;
329 opt.dots_in_line = 50;
331 opt.dns_cache = true;
335 opt.check_cert = true;
338 /* The default for file name restriction defaults to the OS type. */
339 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
340 opt.restrict_files_os = restrict_windows;
342 opt.restrict_files_os = restrict_unix;
344 opt.restrict_files_ctrl = true;
345 opt.restrict_files_nonascii = false;
346 opt.restrict_files_case = restrict_no_case_restriction;
348 opt.max_redirect = 20;
353 opt.enable_iri = true;
355 opt.enable_iri = false;
358 opt.encoding_remote = NULL;
360 opt.useservertimestamps = true;
363 /* Return the user's home directory (strdup-ed), or NULL if none is
368 static char buf[PATH_MAX];
373 home = getenv ("HOME");
377 /* Under MSDOS, if $HOME isn't defined, use the directory where
378 `wget.exe' resides. */
379 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
382 strcpy (buf, _w32_get_argv0 ());
383 p = strrchr (buf, '/'); /* djgpp */
385 p = strrchr (buf, '\\'); /* others */
389 #elif !defined(WINDOWS)
390 /* If HOME is not defined, try getting it from the password
392 struct passwd *pwd = getpwuid (getuid ());
393 if (!pwd || !pwd->pw_dir)
395 strcpy (buf, pwd->pw_dir);
398 /* Under Windows, if $HOME isn't defined, use the directory where
399 `wget.exe' resides. */
405 return home ? xstrdup (home) : NULL;
408 /* Check the 'WGETRC' environment variable and return the file name
409 if 'WGETRC' is set and is a valid file.
410 If the `WGETRC' variable exists but the file does not exist, the
411 function will exit(). */
413 wgetrc_env_file_name (void)
415 char *env = getenv ("WGETRC");
418 if (!file_exists_p (env))
420 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
424 return xstrdup (env);
429 /* Check for the existance of '$HOME/.wgetrc' and return it's path
430 if it exists and is set. */
432 wgetrc_user_file_name (void)
436 /* If that failed, try $HOME/.wgetrc (or equivalent). */
439 file = "SYS$LOGIN:.wgetrc";
440 #else /* def __VMS */
443 file = aprintf ("%s/.wgetrc", home);
445 #endif /* def __VMS [else] */
449 if (!file_exists_p (file))
457 /* Return the path to the user's .wgetrc. This is either the value of
458 `WGETRC' environment variable, or `$HOME/.wgetrc'.
460 Additionally, for windows, look in the directory where wget.exe
463 wgetrc_file_name (void)
465 char *file = wgetrc_env_file_name ();
469 file = wgetrc_user_file_name ();
472 /* Under Windows, if we still haven't found .wgetrc, look for the file
473 `wget.ini' in the directory where `wget.exe' resides; we do this for
474 backward compatibility with previous versions of Wget.
475 SYSTEM_WGETRC should not be defined under WINDOWS. */
478 char *home = home_dir ();
484 file = aprintf ("%s/wget.ini", home);
485 if (!file_exists_p (file))
498 /* Return values of parse_line. */
506 static enum parse_line parse_line (const char *, char **, char **, int *);
507 static bool setval_internal (int, const char *, const char *);
508 static bool setval_internal_tilde (int, const char *, const char *);
510 /* Initialize variables from a wgetrc file. Returns zero (failure) if
511 there were errors in the file. */
514 run_wgetrc (const char *file)
521 fp = fopen (file, "r");
524 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
525 file, strerror (errno));
526 return true; /* not a fatal error */
529 while ((line = read_whole_line (fp)) != NULL)
531 char *com = NULL, *val = NULL;
534 /* Parse the line. */
535 switch (parse_line (line, &com, &val, &comind))
538 /* If everything is OK, set the value. */
539 if (!setval_internal_tilde (comind, com, val))
541 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
542 exec_name, file, ln);
546 case line_syntax_error:
547 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
548 exec_name, file, ln);
551 case line_unknown_command:
552 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
553 exec_name, quote (com), file, ln);
571 /* Initialize the defaults and run the system wgetrc and user's own
576 char *file, *env_sysrc;
579 /* Load the hard-coded defaults. */
582 /* Run a non-standard system rc file when the according environment
583 variable has been set. For internal testing purposes only! */
584 env_sysrc = getenv ("SYSTEM_WGETRC");
585 if (env_sysrc && file_exists_p (env_sysrc))
586 ok &= run_wgetrc (env_sysrc);
587 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
589 else if (file_exists_p (SYSTEM_WGETRC))
590 ok &= run_wgetrc (SYSTEM_WGETRC);
592 /* Override it with your own, if one exists. */
593 file = wgetrc_file_name ();
596 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
597 something like realpath() before comparing them with `strcmp' */
599 if (!strcmp (file, SYSTEM_WGETRC))
601 fprintf (stderr, _("\
602 %s: Warning: Both system and user wgetrc point to %s.\n"),
603 exec_name, quote (file));
607 ok &= run_wgetrc (file);
609 /* If there were errors processing either `.wgetrc', abort. */
617 /* Remove dashes and underscores from S, modifying S in the
623 char *t = s; /* t - tortoise */
624 char *h = s; /* h - hare */
626 if (*h == '_' || *h == '-')
633 /* Parse the line pointed by line, with the syntax:
634 <sp>* command <sp>* = <sp>* value <sp>*
635 Uses malloc to allocate space for command and value.
637 Returns one of line_ok, line_empty, line_syntax_error, or
638 line_unknown_command.
640 In case of line_ok, *COM and *VAL point to freshly allocated
641 strings, and *COMIND points to com's index. In case of error or
642 empty line, their values are unmodified. */
644 static enum parse_line
645 parse_line (const char *line, char **com, char **val, int *comind)
648 const char *end = line + strlen (line);
649 const char *cmdstart, *cmdend;
650 const char *valstart, *valend;
655 /* Skip leading and trailing whitespace. */
656 while (*line && c_isspace (*line))
658 while (end > line && c_isspace (end[-1]))
661 /* Skip empty lines and comments. */
662 if (!*line || *line == '#')
668 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
672 /* Skip '=', as well as any space before or after it. */
673 while (p < end && c_isspace (*p))
675 if (p == end || *p != '=')
676 return line_syntax_error;
678 while (p < end && c_isspace (*p))
684 /* The syntax is valid (even though the command might not be). Fill
685 in the command name and value. */
686 *com = strdupdelim (cmdstart, cmdend);
687 *val = strdupdelim (valstart, valend);
689 /* The line now known to be syntactically correct. Check whether
690 the command is valid. */
691 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
693 ind = command_by_name (cmdcopy);
695 return line_unknown_command;
697 /* Report success to the caller. */
702 #if defined(WINDOWS) || defined(MSDOS)
703 # define ISSEP(c) ((c) == '/' || (c) == '\\')
705 # define ISSEP(c) ((c) == '/')
708 /* Run commands[comind].action. */
711 setval_internal (int comind, const char *com, const char *val)
713 assert (0 <= comind && ((size_t) comind) < countof (commands));
714 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
715 return commands[comind].action (com, val, commands[comind].place);
719 setval_internal_tilde (int comind, const char *com, const char *val)
725 ret = setval_internal (comind, com, val);
727 /* We make tilde expansion for cmd_file and cmd_directory */
728 if (((commands[comind].action == cmd_file) ||
729 (commands[comind].action == cmd_directory))
730 && ret && (*val == '~' && ISSEP (val[1])))
732 pstring = commands[comind].place;
736 homelen = strlen (home);
737 while (homelen && ISSEP (home[homelen - 1]))
738 home[--homelen] = '\0';
740 /* Skip the leading "~/". */
741 for (++val; ISSEP (*val); val++)
743 *pstring = concat_strings (home, "/", val, (char *)0);
749 /* Run command COM with value VAL. If running the command produces an
750 error, report the error and exit.
752 This is intended to be called from main() to modify Wget's behavior
753 through command-line switches. Since COM is hard-coded in main(),
754 it is not canonicalized, and this aborts when COM is not found.
756 If COMIND's are exported to init.h, this function will be changed
757 to accept COMIND directly. */
760 setoptval (const char *com, const char *val, const char *optname)
762 /* Prepend "--" to OPTNAME. */
763 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
766 strcpy (dd_optname + 2, optname);
768 assert (val != NULL);
769 if (!setval_internal (command_by_name (com), dd_optname, val))
773 /* Parse OPT into command and value and run it. For example,
774 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
775 This is used by the `--execute' flag in main.c. */
778 run_command (const char *opt)
782 switch (parse_line (opt, &com, &val, &comind))
785 if (!setval_internal (comind, com, val))
791 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
792 exec_name, quote (opt));
797 /* Generic helper functions, for use with `commands'. */
799 /* Forward declarations: */
804 static bool decode_string (const char *, const struct decode_item *, int, int *);
805 static bool simple_atoi (const char *, const char *, int *);
806 static bool simple_atof (const char *, const char *, double *);
808 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
810 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
811 && c_tolower((p)[1]) == (c1) \
814 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
815 && c_tolower((p)[1]) == (c1) \
816 && c_tolower((p)[2]) == (c2) \
820 /* Store the boolean value from VAL to PLACE. COM is ignored,
821 except for error messages. */
823 cmd_boolean (const char *com, const char *val, void *place)
827 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
828 /* "on", "yes" and "1" mean true. */
830 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
831 /* "off", "no" and "0" mean false. */
836 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
837 exec_name, com, quote (val));
841 *(bool *) place = value;
845 /* Set the non-negative integer value from VAL to PLACE. With
846 incorrect specification, the number remains unchanged. */
848 cmd_number (const char *com, const char *val, void *place)
850 if (!simple_atoi (val, val + strlen (val), place)
851 || *(int *) place < 0)
853 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
854 exec_name, com, quote (val));
860 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
862 cmd_number_inf (const char *com, const char *val, void *place)
864 if (!strcasecmp (val, "inf"))
869 return cmd_number (com, val, place);
872 /* Copy (strdup) the string at COM to a new location and place a
873 pointer to *PLACE. */
875 cmd_string (const char *com, const char *val, void *place)
877 char **pstring = (char **)place;
879 xfree_null (*pstring);
880 *pstring = xstrdup (val);
885 /* Like the above, but handles tilde-expansion when reading a user's
886 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
887 gets expanded to the user's home directory. */
889 cmd_file (const char *com, const char *val, void *place)
891 char **pstring = (char **)place;
893 xfree_null (*pstring);
895 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
897 *pstring = xstrdup (val);
899 #if defined(WINDOWS) || defined(MSDOS)
900 /* Convert "\" to "/". */
903 for (s = *pstring; *s; s++)
911 /* Like cmd_file, but strips trailing '/' characters. */
913 cmd_directory (const char *com, const char *val, void *place)
917 /* Call cmd_file() for tilde expansion and separator
918 canonicalization (backslash -> slash under Windows). These
919 things should perhaps be in a separate function. */
920 if (!cmd_file (com, val, place))
925 while (t > s && *--t == '/')
931 /* Split VAL by space to a vector of values, and append those values
932 to vector pointed to by the PLACE argument. If VAL is empty, the
933 PLACE vector is cleared instead. */
936 cmd_vector (const char *com, const char *val, void *place)
938 char ***pvec = (char ***)place;
941 *pvec = merge_vecs (*pvec, sepstring (val));
951 cmd_directory_vector (const char *com, const char *val, void *place)
953 char ***pvec = (char ***)place;
957 /* Strip the trailing slashes from directories. */
960 seps = sepstring (val);
961 for (t = seps; t && *t; t++)
963 int len = strlen (*t);
964 /* Skip degenerate case of root directory. */
967 if ((*t)[len - 1] == '/')
968 (*t)[len - 1] = '\0';
971 *pvec = merge_vecs (*pvec, seps);
981 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
982 "100k" or "2.5G" to a floating point number. */
985 parse_bytes_helper (const char *val, double *result)
988 const char *end = val + strlen (val);
990 /* Check for "inf". */
991 if (0 == strcmp (val, "inf"))
997 /* Strip trailing whitespace. */
998 while (val < end && c_isspace (end[-1]))
1003 switch (c_tolower (end[-1]))
1006 --end, mult = 1024.0;
1009 --end, mult = 1048576.0;
1012 --end, mult = 1073741824.0;
1015 --end, mult = 1099511627776.0;
1018 /* Not a recognized suffix: assume it's a digit. (If not,
1019 simple_atof will raise an error.) */
1023 /* Skip leading and trailing whitespace. */
1024 while (val < end && c_isspace (*val))
1026 while (val < end && c_isspace (end[-1]))
1031 if (!simple_atof (val, end, &number) || number < 0)
1034 *result = number * mult;
1038 /* Parse VAL as a number and set its value to PLACE (which should
1041 By default, the value is assumed to be in bytes. If "K", "M", or
1042 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1043 1<<30, respectively. Floating point values are allowed and are
1044 cast to integer before use. The idea is to be able to use things
1045 like 1.5k instead of "1536".
1047 The string "inf" is returned as 0.
1049 In case of error, false is returned and memory pointed to by PLACE
1050 remains unmodified. */
1053 cmd_bytes (const char *com, const char *val, void *place)
1056 if (!parse_bytes_helper (val, &byte_value))
1058 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1059 exec_name, com, quote (val));
1062 *(wgint *)place = (wgint)byte_value;
1066 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1067 SIZE_SUM. It works by converting the string to double, therefore
1068 working with values up to 2^53-1 without loss of precision. This
1069 value (8192 TB) is large enough to serve for a while. */
1072 cmd_bytes_sum (const char *com, const char *val, void *place)
1075 if (!parse_bytes_helper (val, &byte_value))
1077 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1078 exec_name, com, quote (val));
1081 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1085 /* Store the value of VAL to *OUT. The value is a time period, by
1086 default expressed in seconds, but also accepting suffixes "m", "h",
1087 "d", and "w" for minutes, hours, days, and weeks respectively. */
1090 cmd_time (const char *com, const char *val, void *place)
1092 double number, mult;
1093 const char *end = val + strlen (val);
1095 /* Strip trailing whitespace. */
1096 while (val < end && c_isspace (end[-1]))
1102 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1103 exec_name, com, quote (val));
1107 switch (c_tolower (end[-1]))
1110 --end, mult = 1; /* seconds */
1113 --end, mult = 60; /* minutes */
1116 --end, mult = 3600; /* hours */
1119 --end, mult = 86400.0; /* days */
1122 --end, mult = 604800.0; /* weeks */
1125 /* Not a recognized suffix: assume it belongs to the number.
1126 (If not, simple_atof will raise an error.) */
1130 /* Skip leading and trailing whitespace. */
1131 while (val < end && c_isspace (*val))
1133 while (val < end && c_isspace (end[-1]))
1138 if (!simple_atof (val, end, &number))
1141 *(double *)place = number * mult;
1147 cmd_cert_type (const char *com, const char *val, void *place)
1149 static const struct decode_item choices[] = {
1150 { "pem", keyfile_pem },
1151 { "der", keyfile_asn1 },
1152 { "asn1", keyfile_asn1 },
1154 int ok = decode_string (val, choices, countof (choices), place);
1156 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1161 /* Specialized helper functions, used by `commands' to handle some
1162 options specially. */
1164 static bool check_user_specified_header (const char *);
1167 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1169 if (!cmd_boolean (com, val, &opt.dirstruct))
1171 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1172 must be affected inversely. */
1174 opt.no_dirstruct = false;
1176 opt.no_dirstruct = true;
1181 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1183 /* Empty value means reset the list of headers. */
1186 free_vec (opt.user_headers);
1187 opt.user_headers = NULL;
1191 if (!check_user_specified_header (val))
1193 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1194 exec_name, com, quote (val));
1197 opt.user_headers = vec_append (opt.user_headers, val);
1202 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1204 int flag = cmd_boolean (com, val, &opt.htmlify);
1205 if (flag && !opt.htmlify)
1206 opt.remove_listing = false;
1210 /* Set the "mirror" mode. It means: recursive download, timestamping,
1211 no limit on max. recursion depth, and don't remove listings. */
1214 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1218 if (!cmd_boolean (com, val, &mirror))
1222 opt.recursive = true;
1223 if (!opt.no_dirstruct)
1224 opt.dirstruct = true;
1225 opt.timestamping = true;
1226 opt.reclevel = INFINITE_RECURSION;
1227 opt.remove_listing = false;
1232 /* Validate --prefer-family and set the choice. Allowed values are
1233 "IPv4", "IPv6", and "none". */
1236 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1238 static const struct decode_item choices[] = {
1239 { "IPv4", prefer_ipv4 },
1240 { "IPv6", prefer_ipv6 },
1241 { "none", prefer_none },
1243 int prefer_family = prefer_none;
1244 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1246 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1247 opt.prefer_family = prefer_family;
1251 /* Set progress.type to VAL, but verify that it's a valid progress
1252 implementation before that. */
1255 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1257 if (!valid_progress_implementation_p (val))
1259 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1260 exec_name, com, quote (val));
1263 xfree_null (opt.progress_type);
1265 /* Don't call set_progress_implementation here. It will be called
1266 in main() when it becomes clear what the log output is. */
1267 opt.progress_type = xstrdup (val);
1271 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1272 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1276 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1278 if (!cmd_boolean (com, val, &opt.recursive))
1282 if (opt.recursive && !opt.no_dirstruct)
1283 opt.dirstruct = true;
1289 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1291 int restrict_os = opt.restrict_files_os;
1292 int restrict_ctrl = opt.restrict_files_ctrl;
1293 int restrict_case = opt.restrict_files_case;
1294 int restrict_nonascii = opt.restrict_files_nonascii;
1298 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1302 end = strchr (val, ',');
1304 end = val + strlen (val);
1306 if (VAL_IS ("unix"))
1307 restrict_os = restrict_unix;
1308 else if (VAL_IS ("windows"))
1309 restrict_os = restrict_windows;
1310 else if (VAL_IS ("lowercase"))
1311 restrict_case = restrict_lowercase;
1312 else if (VAL_IS ("uppercase"))
1313 restrict_case = restrict_uppercase;
1314 else if (VAL_IS ("nocontrol"))
1315 restrict_ctrl = false;
1316 else if (VAL_IS ("ascii"))
1317 restrict_nonascii = true;
1320 fprintf (stderr, _("\
1321 %s: %s: Invalid restriction %s,\n\
1322 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1323 exec_name, com, quote (val));
1330 while (*val && *end);
1334 opt.restrict_files_os = restrict_os;
1335 opt.restrict_files_ctrl = restrict_ctrl;
1336 opt.restrict_files_case = restrict_case;
1337 opt.restrict_files_nonascii = restrict_nonascii;
1344 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1346 static const struct decode_item choices[] = {
1347 { "auto", secure_protocol_auto },
1348 { "sslv2", secure_protocol_sslv2 },
1349 { "sslv3", secure_protocol_sslv3 },
1350 { "tlsv1", secure_protocol_tlsv1 },
1352 int ok = decode_string (val, choices, countof (choices), place);
1354 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1359 /* Set all three timeout values. */
1362 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1365 if (!cmd_time (com, val, &value))
1367 opt.read_timeout = value;
1368 opt.connect_timeout = value;
1369 opt.dns_timeout = value;
1374 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1376 /* Disallow embedded newlines. */
1377 if (strchr (val, '\n'))
1379 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1380 exec_name, com, quote (val));
1383 xfree_null (opt.useragent);
1384 opt.useragent = xstrdup (val);
1388 /* The "verbose" option cannot be cmd_boolean because the variable is
1389 not bool -- it's of type int (-1 means uninitialized because of
1390 some random hackery for disallowing -q -v). */
1393 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1396 if (cmd_boolean (com, val, &flag))
1404 /* Miscellaneous useful routines. */
1406 /* A very simple atoi clone, more useful than atoi because it works on
1407 delimited strings, and has error reportage. Returns true on success,
1408 false on failure. If successful, stores result to *DEST. */
1411 simple_atoi (const char *beg, const char *end, int *dest)
1414 bool negative = false;
1415 const char *p = beg;
1417 while (p < end && c_isspace (*p))
1419 if (p < end && (*p == '-' || *p == '+'))
1421 negative = (*p == '-');
1427 /* Read negative numbers in a separate loop because the most
1428 negative integer cannot be represented as a positive number. */
1431 for (; p < end && c_isdigit (*p); p++)
1433 int next = (10 * result) + (*p - '0');
1435 return false; /* overflow */
1439 for (; p < end && c_isdigit (*p); p++)
1441 int next = (10 * result) - (*p - '0');
1443 return false; /* underflow */
1454 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1455 doesn't handle exponential notation. Returns true on success,
1456 false on failure. In case of success, stores its result to
1460 simple_atof (const char *beg, const char *end, double *dest)
1464 bool negative = false;
1465 bool seen_dot = false;
1466 bool seen_digit = false;
1469 const char *p = beg;
1471 while (p < end && c_isspace (*p))
1473 if (p < end && (*p == '-' || *p == '+'))
1475 negative = (*p == '-');
1479 for (; p < end; p++)
1485 result = (10 * result) + (ch - '0');
1487 result += (ch - '0') / (divider *= 10);
1509 /* Verify that the user-specified header in S is valid. It must
1510 contain a colon preceded by non-white-space characters and must not
1511 contain newlines. */
1514 check_user_specified_header (const char *s)
1518 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1520 /* The header MUST contain `:' preceded by at least one
1521 non-whitespace character. */
1522 if (*p != ':' || p == s)
1524 /* The header MUST NOT contain newlines. */
1525 if (strchr (s, '\n'))
1530 /* Decode VAL into a number, according to ITEMS. */
1533 decode_string (const char *val, const struct decode_item *items, int itemcount,
1537 for (i = 0; i < itemcount; i++)
1538 if (0 == strcasecmp (val, items[i].name))
1540 *place = items[i].code;
1547 void cleanup_html_url (void);
1550 /* Free the memory allocated by global variables. */
1554 /* Free external resources, close files, etc. */
1557 fclose (output_stream);
1558 /* No need to check for error because Wget flushes its output (and
1559 checks for errors) after any data arrives. */
1561 /* We're exiting anyway so there's no real need to call free()
1562 hundreds of times. Skipping the frees will make Wget exit
1565 However, when detecting leaks, it's crucial to free() everything
1566 because then you can find the real leaks, i.e. the allocated
1567 memory which grows with the size of the program. */
1573 cleanup_html_url ();
1578 extern acc_t *netrc_list;
1579 free_netrc (netrc_list);
1581 xfree_null (opt.lfilename);
1582 xfree_null (opt.dir_prefix);
1583 xfree_null (opt.input_filename);
1584 xfree_null (opt.output_document);
1585 free_vec (opt.accepts);
1586 free_vec (opt.rejects);
1587 free_vec (opt.excludes);
1588 free_vec (opt.includes);
1589 free_vec (opt.domains);
1590 free_vec (opt.follow_tags);
1591 free_vec (opt.ignore_tags);
1592 xfree_null (opt.progress_type);
1593 xfree_null (opt.ftp_user);
1594 xfree_null (opt.ftp_passwd);
1595 xfree_null (opt.ftp_proxy);
1596 xfree_null (opt.https_proxy);
1597 xfree_null (opt.http_proxy);
1598 free_vec (opt.no_proxy);
1599 xfree_null (opt.useragent);
1600 xfree_null (opt.referer);
1601 xfree_null (opt.http_user);
1602 xfree_null (opt.http_passwd);
1603 free_vec (opt.user_headers);
1605 xfree_null (opt.cert_file);
1606 xfree_null (opt.private_key);
1607 xfree_null (opt.ca_directory);
1608 xfree_null (opt.ca_cert);
1609 xfree_null (opt.random_file);
1610 xfree_null (opt.egd_file);
1612 xfree_null (opt.bind_address);
1613 xfree_null (opt.cookies_input);
1614 xfree_null (opt.cookies_output);
1615 xfree_null (opt.user);
1616 xfree_null (opt.passwd);
1617 xfree_null (opt.base_href);
1619 #endif /* DEBUG_MALLOC */
1622 /* Unit testing routines. */
1627 test_commands_sorted()
1629 int prev_idx = 0, next_idx = 1;
1630 int command_count = countof (commands) - 1;
1632 while (next_idx <= command_count)
1634 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1637 mu_assert ("FAILED", false);
1650 test_cmd_spec_restrict_file_names()
1655 int expected_restrict_files_os;
1656 int expected_restrict_files_ctrl;
1657 int expected_restrict_files_case;
1660 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1661 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1662 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1663 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1666 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1671 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1674 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1675 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1676 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1677 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1679 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1680 res == test_array[i].result
1681 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1682 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1683 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1689 #endif /* TESTING */