1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008, 2009 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 */
63 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
65 CMD_DECLARE (cmd_boolean);
66 CMD_DECLARE (cmd_bytes);
67 CMD_DECLARE (cmd_bytes_sum);
69 CMD_DECLARE (cmd_cert_type);
71 CMD_DECLARE (cmd_directory_vector);
72 CMD_DECLARE (cmd_number);
73 CMD_DECLARE (cmd_number_inf);
74 CMD_DECLARE (cmd_string);
75 CMD_DECLARE (cmd_file);
76 CMD_DECLARE (cmd_directory);
77 CMD_DECLARE (cmd_time);
78 CMD_DECLARE (cmd_vector);
80 CMD_DECLARE (cmd_spec_dirstruct);
81 CMD_DECLARE (cmd_spec_header);
82 CMD_DECLARE (cmd_spec_htmlify);
83 CMD_DECLARE (cmd_spec_mirror);
84 CMD_DECLARE (cmd_spec_prefer_family);
85 CMD_DECLARE (cmd_spec_progress);
86 CMD_DECLARE (cmd_spec_recursive);
87 CMD_DECLARE (cmd_spec_restrict_file_names);
89 CMD_DECLARE (cmd_spec_secure_protocol);
91 CMD_DECLARE (cmd_spec_timeout);
92 CMD_DECLARE (cmd_spec_useragent);
93 CMD_DECLARE (cmd_spec_verbose);
95 /* List of recognized commands, each consisting of name, place and
96 function. When adding a new command, simply add it to the list,
97 but be sure to keep the list sorted alphabetically, as
98 command_by_name's binary search depends on it. Also, be sure to
99 add any entries that allocate memory (e.g. cmd_string and
100 cmd_vector) to the cleanup() function below. */
102 static const struct {
105 bool (*action) (const char *, const char *, void *);
107 /* KEEP THIS LIST ALPHABETICALLY SORTED */
108 { "accept", &opt.accepts, cmd_vector },
109 { "addhostdir", &opt.add_hostdir, cmd_boolean },
110 { "adjustextension", &opt.adjust_extension, cmd_boolean },
111 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
112 { "askpassword", &opt.ask_passwd, cmd_boolean },
113 { "authnochallenge", &opt.auth_without_challenge,
115 { "background", &opt.background, cmd_boolean },
116 { "backupconverted", &opt.backup_converted, cmd_boolean },
117 { "backups", &opt.backups, cmd_number },
118 { "base", &opt.base_href, cmd_string },
119 { "bindaddress", &opt.bind_address, cmd_string },
121 { "cacertificate", &opt.ca_cert, cmd_file },
123 { "cache", &opt.allow_cache, cmd_boolean },
125 { "cadirectory", &opt.ca_directory, cmd_directory },
126 { "certificate", &opt.cert_file, cmd_file },
127 { "certificatetype", &opt.cert_type, cmd_cert_type },
128 { "checkcertificate", &opt.check_cert, cmd_boolean },
130 { "connecttimeout", &opt.connect_timeout, cmd_time },
131 { "contentdisposition", &opt.content_disposition, cmd_boolean },
132 { "continue", &opt.always_rest, cmd_boolean },
133 { "convertlinks", &opt.convert_links, cmd_boolean },
134 { "cookies", &opt.cookies, cmd_boolean },
135 { "cutdirs", &opt.cut_dirs, cmd_number },
137 { "debug", &opt.debug, cmd_boolean },
139 { "defaultpage", &opt.default_page, cmd_string},
140 { "deleteafter", &opt.delete_after, cmd_boolean },
141 { "dirprefix", &opt.dir_prefix, cmd_directory },
142 { "dirstruct", NULL, cmd_spec_dirstruct },
143 { "dnscache", &opt.dns_cache, cmd_boolean },
144 { "dnstimeout", &opt.dns_timeout, cmd_time },
145 { "domains", &opt.domains, cmd_vector },
146 { "dotbytes", &opt.dot_bytes, cmd_bytes },
147 { "dotsinline", &opt.dots_in_line, cmd_number },
148 { "dotspacing", &opt.dot_spacing, cmd_number },
149 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
151 { "egdfile", &opt.egd_file, cmd_file },
153 { "excludedirectories", &opt.excludes, cmd_directory_vector },
154 { "excludedomains", &opt.exclude_domains, cmd_vector },
155 { "followftp", &opt.follow_ftp, cmd_boolean },
156 { "followtags", &opt.follow_tags, cmd_vector },
157 { "forcehtml", &opt.force_html, cmd_boolean },
158 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
159 { "ftppassword", &opt.ftp_passwd, cmd_string },
160 { "ftpproxy", &opt.ftp_proxy, cmd_string },
162 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
163 #endif /* def __VMS */
164 { "ftpuser", &opt.ftp_user, cmd_string },
165 { "glob", &opt.ftp_glob, cmd_boolean },
166 { "header", NULL, cmd_spec_header },
167 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
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 { "iri", &opt.enable_iri, cmd_boolean },
185 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186 { "limitrate", &opt.limit_rate, cmd_bytes },
187 { "loadcookies", &opt.cookies_input, cmd_file },
188 { "localencoding", &opt.locale, cmd_string },
189 { "logfile", &opt.lfilename, cmd_file },
190 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
191 { "maxredirect", &opt.max_redirect, cmd_number },
192 { "mirror", NULL, cmd_spec_mirror },
193 { "netrc", &opt.netrc, cmd_boolean },
194 { "noclobber", &opt.noclobber, cmd_boolean },
195 { "noparent", &opt.no_parent, cmd_boolean },
196 { "noproxy", &opt.no_proxy, cmd_vector },
197 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
198 { "outputdocument", &opt.output_document, cmd_file },
199 { "pagerequisites", &opt.page_requisites, cmd_boolean },
200 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
201 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
202 { "password", &opt.passwd, cmd_string },
203 { "postdata", &opt.post_data, cmd_string },
204 { "postfile", &opt.post_file_name, cmd_file },
205 { "preferfamily", NULL, cmd_spec_prefer_family },
206 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
208 { "privatekey", &opt.private_key, cmd_file },
209 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
211 { "progress", &opt.progress_type, cmd_spec_progress },
212 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
213 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
214 { "proxypassword", &opt.proxy_passwd, cmd_string },
215 { "proxyuser", &opt.proxy_user, cmd_string },
216 { "quiet", &opt.quiet, cmd_boolean },
217 { "quota", &opt.quota, cmd_bytes_sum },
219 { "randomfile", &opt.random_file, cmd_file },
221 { "randomwait", &opt.random_wait, cmd_boolean },
222 { "readtimeout", &opt.read_timeout, cmd_time },
223 { "reclevel", &opt.reclevel, cmd_number_inf },
224 { "recursive", NULL, cmd_spec_recursive },
225 { "referer", &opt.referer, cmd_string },
226 { "reject", &opt.rejects, cmd_vector },
227 { "relativeonly", &opt.relative_only, cmd_boolean },
228 { "remoteencoding", &opt.encoding_remote, cmd_string },
229 { "removelisting", &opt.remove_listing, cmd_boolean },
230 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
231 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
232 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
233 { "robots", &opt.use_robots, cmd_boolean },
234 { "savecookies", &opt.cookies_output, cmd_file },
235 { "saveheaders", &opt.save_headers, cmd_boolean },
237 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
239 { "serverresponse", &opt.server_response, cmd_boolean },
240 { "spanhosts", &opt.spanhost, cmd_boolean },
241 { "spider", &opt.spider, cmd_boolean },
242 { "strictcomments", &opt.strict_comments, cmd_boolean },
243 { "timeout", NULL, cmd_spec_timeout },
244 { "timestamping", &opt.timestamping, cmd_boolean },
245 { "tries", &opt.ntry, cmd_number_inf },
246 { "useproxy", &opt.use_proxy, cmd_boolean },
247 { "user", &opt.user, cmd_string },
248 { "useragent", NULL, cmd_spec_useragent },
249 { "verbose", NULL, cmd_spec_verbose },
250 { "wait", &opt.wait, cmd_time },
251 { "waitretry", &opt.waitretry, cmd_time },
253 { "wdebug", &opt.wdebug, cmd_boolean },
257 /* Look up CMDNAME in the commands[] and return its position in the
258 array. If CMDNAME is not found, return -1. */
261 command_by_name (const char *cmdname)
263 /* Use binary search for speed. Wget has ~100 commands, which
264 guarantees a worst case performance of 7 string comparisons. */
265 int lo = 0, hi = countof (commands) - 1;
269 int mid = (lo + hi) >> 1;
270 int cmp = strcasecmp (cmdname, commands[mid].name);
281 /* Reset the variables to default values. */
287 /* Most of the default values are 0 (and 0.0, NULL, and false).
288 Just reset everything, and fill in the non-zero values. Note
289 that initializing pointers to NULL this way is technically
290 illegal, but porting Wget to a machine where NULL is not all-zero
291 bit pattern will be the least of the implementors' worries. */
298 opt.add_hostdir = true;
302 opt.http_keep_alive = true;
303 opt.use_proxy = true;
304 tmp = getenv ("no_proxy");
306 opt.no_proxy = sepstring (tmp);
307 opt.prefer_family = prefer_none;
308 opt.allow_cache = true;
310 opt.read_timeout = 900;
311 opt.use_robots = true;
313 opt.remove_listing = true;
315 opt.dot_bytes = 1024;
316 opt.dot_spacing = 10;
317 opt.dots_in_line = 50;
319 opt.dns_cache = true;
323 opt.check_cert = true;
326 /* The default for file name restriction defaults to the OS type. */
327 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
328 opt.restrict_files_os = restrict_windows;
330 opt.restrict_files_os = restrict_unix;
332 opt.restrict_files_ctrl = true;
333 opt.restrict_files_nonascii = false;
334 opt.restrict_files_case = restrict_no_case_restriction;
336 opt.max_redirect = 20;
341 opt.enable_iri = true;
343 opt.enable_iri = false;
346 opt.encoding_remote = NULL;
349 /* Return the user's home directory (strdup-ed), or NULL if none is
354 static char buf[PATH_MAX];
359 home = getenv ("HOME");
363 /* Under MSDOS, if $HOME isn't defined, use the directory where
364 `wget.exe' resides. */
365 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
368 strcpy (buf, _w32_get_argv0 ());
369 p = strrchr (buf, '/'); /* djgpp */
371 p = strrchr (buf, '\\'); /* others */
375 #elif !defined(WINDOWS)
376 /* If HOME is not defined, try getting it from the password
378 struct passwd *pwd = getpwuid (getuid ());
379 if (!pwd || !pwd->pw_dir)
381 strcpy (buf, pwd->pw_dir);
384 /* Under Windows, if $HOME isn't defined, use the directory where
385 `wget.exe' resides. */
391 return home ? xstrdup (home) : NULL;
394 /* Check the 'WGETRC' environment variable and return the file name
395 if 'WGETRC' is set and is a valid file.
396 If the `WGETRC' variable exists but the file does not exist, the
397 function will exit(). */
399 wgetrc_env_file_name (void)
401 char *env = getenv ("WGETRC");
404 if (!file_exists_p (env))
406 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
410 return xstrdup (env);
415 /* Check for the existance of '$HOME/.wgetrc' and return it's path
416 if it exists and is set. */
418 wgetrc_user_file_name (void)
420 char *home = home_dir ();
422 /* If that failed, try $HOME/.wgetrc (or equivalent). */
425 file = "SYS$LOGIN:.wgetrc";
426 #else /* def __VMS */
429 file = aprintf ("%s/.wgetrc", home);
431 #endif /* def __VMS [else] */
435 if (!file_exists_p (file))
443 /* Return the path to the user's .wgetrc. This is either the value of
444 `WGETRC' environment variable, or `$HOME/.wgetrc'.
446 Additionally, for windows, look in the directory where wget.exe
449 wgetrc_file_name (void)
451 char *file = wgetrc_env_file_name ();
455 file = wgetrc_user_file_name ();
458 /* Under Windows, if we still haven't found .wgetrc, look for the file
459 `wget.ini' in the directory where `wget.exe' resides; we do this for
460 backward compatibility with previous versions of Wget.
461 SYSTEM_WGETRC should not be defined under WINDOWS. */
464 char *home = home_dir ();
470 file = aprintf ("%s/wget.ini", home);
471 if (!file_exists_p (file))
484 /* Return values of parse_line. */
492 static enum parse_line parse_line (const char *, char **, char **, int *);
493 static bool setval_internal (int, const char *, const char *);
494 static bool setval_internal_tilde (int, const char *, const char *);
496 /* Initialize variables from a wgetrc file. Returns zero (failure) if
497 there were errors in the file. */
500 run_wgetrc (const char *file)
507 fp = fopen (file, "r");
510 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
511 file, strerror (errno));
512 return true; /* not a fatal error */
515 while ((line = read_whole_line (fp)) != NULL)
517 char *com = NULL, *val = NULL;
520 /* Parse the line. */
521 switch (parse_line (line, &com, &val, &comind))
524 /* If everything is OK, set the value. */
525 if (!setval_internal_tilde (comind, com, val))
527 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
528 exec_name, file, ln);
532 case line_syntax_error:
533 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
534 exec_name, file, ln);
537 case line_unknown_command:
538 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
539 exec_name, quote (com), file, ln);
557 /* Initialize the defaults and run the system wgetrc and user's own
562 char *file, *env_sysrc;
565 /* Load the hard-coded defaults. */
568 /* Run a non-standard system rc file when the according environment
569 variable has been set. For internal testing purposes only! */
570 env_sysrc = getenv ("SYSTEM_WGETRC");
571 if (env_sysrc && file_exists_p (env_sysrc))
572 ok &= run_wgetrc (env_sysrc);
573 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
575 else if (file_exists_p (SYSTEM_WGETRC))
576 ok &= run_wgetrc (SYSTEM_WGETRC);
578 /* Override it with your own, if one exists. */
579 file = wgetrc_file_name ();
582 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
583 something like realpath() before comparing them with `strcmp' */
585 if (!strcmp (file, SYSTEM_WGETRC))
587 fprintf (stderr, _("\
588 %s: Warning: Both system and user wgetrc point to %s.\n"),
589 exec_name, quote (file));
593 ok &= run_wgetrc (file);
595 /* If there were errors processing either `.wgetrc', abort. */
603 /* Remove dashes and underscores from S, modifying S in the
609 char *t = s; /* t - tortoise */
610 char *h = s; /* h - hare */
612 if (*h == '_' || *h == '-')
619 /* Parse the line pointed by line, with the syntax:
620 <sp>* command <sp>* = <sp>* value <sp>*
621 Uses malloc to allocate space for command and value.
623 Returns one of line_ok, line_empty, line_syntax_error, or
624 line_unknown_command.
626 In case of line_ok, *COM and *VAL point to freshly allocated
627 strings, and *COMIND points to com's index. In case of error or
628 empty line, their values are unmodified. */
630 static enum parse_line
631 parse_line (const char *line, char **com, char **val, int *comind)
634 const char *end = line + strlen (line);
635 const char *cmdstart, *cmdend;
636 const char *valstart, *valend;
641 /* Skip leading and trailing whitespace. */
642 while (*line && c_isspace (*line))
644 while (end > line && c_isspace (end[-1]))
647 /* Skip empty lines and comments. */
648 if (!*line || *line == '#')
654 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
658 /* Skip '=', as well as any space before or after it. */
659 while (p < end && c_isspace (*p))
661 if (p == end || *p != '=')
662 return line_syntax_error;
664 while (p < end && c_isspace (*p))
670 /* The syntax is valid (even though the command might not be). Fill
671 in the command name and value. */
672 *com = strdupdelim (cmdstart, cmdend);
673 *val = strdupdelim (valstart, valend);
675 /* The line now known to be syntactically correct. Check whether
676 the command is valid. */
677 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
679 ind = command_by_name (cmdcopy);
681 return line_unknown_command;
683 /* Report success to the caller. */
688 #if defined(WINDOWS) || defined(MSDOS)
689 # define ISSEP(c) ((c) == '/' || (c) == '\\')
691 # define ISSEP(c) ((c) == '/')
694 /* Run commands[comind].action. */
697 setval_internal (int comind, const char *com, const char *val)
699 assert (0 <= comind && ((size_t) comind) < countof (commands));
700 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
701 return commands[comind].action (com, val, commands[comind].place);
705 setval_internal_tilde (int comind, const char *com, const char *val)
711 ret = setval_internal (comind, com, val);
713 /* We make tilde expansion for cmd_file and cmd_directory */
714 if (((commands[comind].action == cmd_file) ||
715 (commands[comind].action == cmd_directory))
716 && ret && (*val == '~' && ISSEP (val[1])))
718 pstring = commands[comind].place;
722 homelen = strlen (home);
723 while (homelen && ISSEP (home[homelen - 1]))
724 home[--homelen] = '\0';
726 /* Skip the leading "~/". */
727 for (++val; ISSEP (*val); val++)
729 *pstring = concat_strings (home, "/", val, (char *)0);
735 /* Run command COM with value VAL. If running the command produces an
736 error, report the error and exit.
738 This is intended to be called from main() to modify Wget's behavior
739 through command-line switches. Since COM is hard-coded in main(),
740 it is not canonicalized, and this aborts when COM is not found.
742 If COMIND's are exported to init.h, this function will be changed
743 to accept COMIND directly. */
746 setoptval (const char *com, const char *val, const char *optname)
748 /* Prepend "--" to OPTNAME. */
749 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
752 strcpy (dd_optname + 2, optname);
754 assert (val != NULL);
755 if (!setval_internal (command_by_name (com), dd_optname, val))
759 /* Parse OPT into command and value and run it. For example,
760 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
761 This is used by the `--execute' flag in main.c. */
764 run_command (const char *opt)
768 switch (parse_line (opt, &com, &val, &comind))
771 if (!setval_internal (comind, com, val))
777 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
778 exec_name, quote (opt));
783 /* Generic helper functions, for use with `commands'. */
785 /* Forward declarations: */
790 static bool decode_string (const char *, const struct decode_item *, int, int *);
791 static bool simple_atoi (const char *, const char *, int *);
792 static bool simple_atof (const char *, const char *, double *);
794 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
796 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
797 && c_tolower((p)[1]) == (c1) \
800 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
801 && c_tolower((p)[1]) == (c1) \
802 && c_tolower((p)[2]) == (c2) \
806 /* Store the boolean value from VAL to PLACE. COM is ignored,
807 except for error messages. */
809 cmd_boolean (const char *com, const char *val, void *place)
813 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
814 /* "on", "yes" and "1" mean true. */
816 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
817 /* "off", "no" and "0" mean false. */
822 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
823 exec_name, com, quote (val));
827 *(bool *) place = value;
831 /* Set the non-negative integer value from VAL to PLACE. With
832 incorrect specification, the number remains unchanged. */
834 cmd_number (const char *com, const char *val, void *place)
836 if (!simple_atoi (val, val + strlen (val), place)
837 || *(int *) place < 0)
839 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
840 exec_name, com, quote (val));
846 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
848 cmd_number_inf (const char *com, const char *val, void *place)
850 if (!strcasecmp (val, "inf"))
855 return cmd_number (com, val, place);
858 /* Copy (strdup) the string at COM to a new location and place a
859 pointer to *PLACE. */
861 cmd_string (const char *com, const char *val, void *place)
863 char **pstring = (char **)place;
865 xfree_null (*pstring);
866 *pstring = xstrdup (val);
871 /* Like the above, but handles tilde-expansion when reading a user's
872 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
873 gets expanded to the user's home directory. */
875 cmd_file (const char *com, const char *val, void *place)
877 char **pstring = (char **)place;
879 xfree_null (*pstring);
881 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
883 *pstring = xstrdup (val);
885 #if defined(WINDOWS) || defined(MSDOS)
886 /* Convert "\" to "/". */
889 for (s = *pstring; *s; s++)
897 /* Like cmd_file, but strips trailing '/' characters. */
899 cmd_directory (const char *com, const char *val, void *place)
903 /* Call cmd_file() for tilde expansion and separator
904 canonicalization (backslash -> slash under Windows). These
905 things should perhaps be in a separate function. */
906 if (!cmd_file (com, val, place))
911 while (t > s && *--t == '/')
917 /* Split VAL by space to a vector of values, and append those values
918 to vector pointed to by the PLACE argument. If VAL is empty, the
919 PLACE vector is cleared instead. */
922 cmd_vector (const char *com, const char *val, void *place)
924 char ***pvec = (char ***)place;
927 *pvec = merge_vecs (*pvec, sepstring (val));
937 cmd_directory_vector (const char *com, const char *val, void *place)
939 char ***pvec = (char ***)place;
943 /* Strip the trailing slashes from directories. */
946 seps = sepstring (val);
947 for (t = seps; t && *t; t++)
949 int len = strlen (*t);
950 /* Skip degenerate case of root directory. */
953 if ((*t)[len - 1] == '/')
954 (*t)[len - 1] = '\0';
957 *pvec = merge_vecs (*pvec, seps);
967 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
968 "100k" or "2.5G" to a floating point number. */
971 parse_bytes_helper (const char *val, double *result)
974 const char *end = val + strlen (val);
976 /* Check for "inf". */
977 if (0 == strcmp (val, "inf"))
983 /* Strip trailing whitespace. */
984 while (val < end && c_isspace (end[-1]))
989 switch (c_tolower (end[-1]))
992 --end, mult = 1024.0;
995 --end, mult = 1048576.0;
998 --end, mult = 1073741824.0;
1001 --end, mult = 1099511627776.0;
1004 /* Not a recognized suffix: assume it's a digit. (If not,
1005 simple_atof will raise an error.) */
1009 /* Skip leading and trailing whitespace. */
1010 while (val < end && c_isspace (*val))
1012 while (val < end && c_isspace (end[-1]))
1017 if (!simple_atof (val, end, &number) || number < 0)
1020 *result = number * mult;
1024 /* Parse VAL as a number and set its value to PLACE (which should
1027 By default, the value is assumed to be in bytes. If "K", "M", or
1028 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1029 1<<30, respectively. Floating point values are allowed and are
1030 cast to integer before use. The idea is to be able to use things
1031 like 1.5k instead of "1536".
1033 The string "inf" is returned as 0.
1035 In case of error, false is returned and memory pointed to by PLACE
1036 remains unmodified. */
1039 cmd_bytes (const char *com, const char *val, void *place)
1042 if (!parse_bytes_helper (val, &byte_value))
1044 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1045 exec_name, com, quote (val));
1048 *(wgint *)place = (wgint)byte_value;
1052 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1053 SIZE_SUM. It works by converting the string to double, therefore
1054 working with values up to 2^53-1 without loss of precision. This
1055 value (8192 TB) is large enough to serve for a while. */
1058 cmd_bytes_sum (const char *com, const char *val, void *place)
1061 if (!parse_bytes_helper (val, &byte_value))
1063 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1064 exec_name, com, quote (val));
1067 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1071 /* Store the value of VAL to *OUT. The value is a time period, by
1072 default expressed in seconds, but also accepting suffixes "m", "h",
1073 "d", and "w" for minutes, hours, days, and weeks respectively. */
1076 cmd_time (const char *com, const char *val, void *place)
1078 double number, mult;
1079 const char *end = val + strlen (val);
1081 /* Strip trailing whitespace. */
1082 while (val < end && c_isspace (end[-1]))
1088 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1089 exec_name, com, quote (val));
1093 switch (c_tolower (end[-1]))
1096 --end, mult = 1; /* seconds */
1099 --end, mult = 60; /* minutes */
1102 --end, mult = 3600; /* hours */
1105 --end, mult = 86400.0; /* days */
1108 --end, mult = 604800.0; /* weeks */
1111 /* Not a recognized suffix: assume it belongs to the number.
1112 (If not, simple_atof will raise an error.) */
1116 /* Skip leading and trailing whitespace. */
1117 while (val < end && c_isspace (*val))
1119 while (val < end && c_isspace (end[-1]))
1124 if (!simple_atof (val, end, &number))
1127 *(double *)place = number * mult;
1133 cmd_cert_type (const char *com, const char *val, void *place)
1135 static const struct decode_item choices[] = {
1136 { "pem", keyfile_pem },
1137 { "der", keyfile_asn1 },
1138 { "asn1", keyfile_asn1 },
1140 int ok = decode_string (val, choices, countof (choices), place);
1142 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1147 /* Specialized helper functions, used by `commands' to handle some
1148 options specially. */
1150 static bool check_user_specified_header (const char *);
1153 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1155 if (!cmd_boolean (com, val, &opt.dirstruct))
1157 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1158 must be affected inversely. */
1160 opt.no_dirstruct = false;
1162 opt.no_dirstruct = true;
1167 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1169 /* Empty value means reset the list of headers. */
1172 free_vec (opt.user_headers);
1173 opt.user_headers = NULL;
1177 if (!check_user_specified_header (val))
1179 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1180 exec_name, com, quote (val));
1183 opt.user_headers = vec_append (opt.user_headers, val);
1188 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1190 int flag = cmd_boolean (com, val, &opt.htmlify);
1191 if (flag && !opt.htmlify)
1192 opt.remove_listing = false;
1196 /* Set the "mirror" mode. It means: recursive download, timestamping,
1197 no limit on max. recursion depth, and don't remove listings. */
1200 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1204 if (!cmd_boolean (com, val, &mirror))
1208 opt.recursive = true;
1209 if (!opt.no_dirstruct)
1210 opt.dirstruct = true;
1211 opt.timestamping = true;
1212 opt.reclevel = INFINITE_RECURSION;
1213 opt.remove_listing = false;
1218 /* Validate --prefer-family and set the choice. Allowed values are
1219 "IPv4", "IPv6", and "none". */
1222 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1224 static const struct decode_item choices[] = {
1225 { "IPv4", prefer_ipv4 },
1226 { "IPv6", prefer_ipv6 },
1227 { "none", prefer_none },
1229 int prefer_family = prefer_none;
1230 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1232 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1233 opt.prefer_family = prefer_family;
1237 /* Set progress.type to VAL, but verify that it's a valid progress
1238 implementation before that. */
1241 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1243 if (!valid_progress_implementation_p (val))
1245 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1246 exec_name, com, quote (val));
1249 xfree_null (opt.progress_type);
1251 /* Don't call set_progress_implementation here. It will be called
1252 in main() when it becomes clear what the log output is. */
1253 opt.progress_type = xstrdup (val);
1257 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1258 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1262 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1264 if (!cmd_boolean (com, val, &opt.recursive))
1268 if (opt.recursive && !opt.no_dirstruct)
1269 opt.dirstruct = true;
1275 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1277 int restrict_os = opt.restrict_files_os;
1278 int restrict_ctrl = opt.restrict_files_ctrl;
1279 int restrict_case = opt.restrict_files_case;
1280 int restrict_nonascii = opt.restrict_files_nonascii;
1284 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1288 end = strchr (val, ',');
1290 end = val + strlen (val);
1292 if (VAL_IS ("unix"))
1293 restrict_os = restrict_unix;
1294 else if (VAL_IS ("windows"))
1295 restrict_os = restrict_windows;
1296 else if (VAL_IS ("lowercase"))
1297 restrict_case = restrict_lowercase;
1298 else if (VAL_IS ("uppercase"))
1299 restrict_case = restrict_uppercase;
1300 else if (VAL_IS ("nocontrol"))
1301 restrict_ctrl = false;
1302 else if (VAL_IS ("ascii"))
1303 restrict_nonascii = true;
1306 fprintf (stderr, _("\
1307 %s: %s: Invalid restriction %s,\n\
1308 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1309 exec_name, com, quote (val));
1316 while (*val && *end);
1320 opt.restrict_files_os = restrict_os;
1321 opt.restrict_files_ctrl = restrict_ctrl;
1322 opt.restrict_files_case = restrict_case;
1323 opt.restrict_files_nonascii = restrict_nonascii;
1330 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1332 static const struct decode_item choices[] = {
1333 { "auto", secure_protocol_auto },
1334 { "sslv2", secure_protocol_sslv2 },
1335 { "sslv3", secure_protocol_sslv3 },
1336 { "tlsv1", secure_protocol_tlsv1 },
1338 int ok = decode_string (val, choices, countof (choices), place);
1340 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1345 /* Set all three timeout values. */
1348 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1351 if (!cmd_time (com, val, &value))
1353 opt.read_timeout = value;
1354 opt.connect_timeout = value;
1355 opt.dns_timeout = value;
1360 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1362 /* Disallow embedded newlines. */
1363 if (strchr (val, '\n'))
1365 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1366 exec_name, com, quote (val));
1369 xfree_null (opt.useragent);
1370 opt.useragent = xstrdup (val);
1374 /* The "verbose" option cannot be cmd_boolean because the variable is
1375 not bool -- it's of type int (-1 means uninitialized because of
1376 some random hackery for disallowing -q -v). */
1379 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1382 if (cmd_boolean (com, val, &flag))
1390 /* Miscellaneous useful routines. */
1392 /* A very simple atoi clone, more useful than atoi because it works on
1393 delimited strings, and has error reportage. Returns true on success,
1394 false on failure. If successful, stores result to *DEST. */
1397 simple_atoi (const char *beg, const char *end, int *dest)
1400 bool negative = false;
1401 const char *p = beg;
1403 while (p < end && c_isspace (*p))
1405 if (p < end && (*p == '-' || *p == '+'))
1407 negative = (*p == '-');
1413 /* Read negative numbers in a separate loop because the most
1414 negative integer cannot be represented as a positive number. */
1417 for (; p < end && c_isdigit (*p); p++)
1419 int next = (10 * result) + (*p - '0');
1421 return false; /* overflow */
1425 for (; p < end && c_isdigit (*p); p++)
1427 int next = (10 * result) - (*p - '0');
1429 return false; /* underflow */
1440 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1441 doesn't handle exponential notation. Returns true on success,
1442 false on failure. In case of success, stores its result to
1446 simple_atof (const char *beg, const char *end, double *dest)
1450 bool negative = false;
1451 bool seen_dot = false;
1452 bool seen_digit = false;
1455 const char *p = beg;
1457 while (p < end && c_isspace (*p))
1459 if (p < end && (*p == '-' || *p == '+'))
1461 negative = (*p == '-');
1465 for (; p < end; p++)
1471 result = (10 * result) + (ch - '0');
1473 result += (ch - '0') / (divider *= 10);
1495 /* Verify that the user-specified header in S is valid. It must
1496 contain a colon preceded by non-white-space characters and must not
1497 contain newlines. */
1500 check_user_specified_header (const char *s)
1504 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1506 /* The header MUST contain `:' preceded by at least one
1507 non-whitespace character. */
1508 if (*p != ':' || p == s)
1510 /* The header MUST NOT contain newlines. */
1511 if (strchr (s, '\n'))
1516 /* Decode VAL into a number, according to ITEMS. */
1519 decode_string (const char *val, const struct decode_item *items, int itemcount,
1523 for (i = 0; i < itemcount; i++)
1524 if (0 == strcasecmp (val, items[i].name))
1526 *place = items[i].code;
1533 void cleanup_html_url (void);
1536 /* Free the memory allocated by global variables. */
1540 /* Free external resources, close files, etc. */
1543 fclose (output_stream);
1544 /* No need to check for error because Wget flushes its output (and
1545 checks for errors) after any data arrives. */
1547 /* We're exiting anyway so there's no real need to call free()
1548 hundreds of times. Skipping the frees will make Wget exit
1551 However, when detecting leaks, it's crucial to free() everything
1552 because then you can find the real leaks, i.e. the allocated
1553 memory which grows with the size of the program. */
1559 cleanup_html_url ();
1564 extern acc_t *netrc_list;
1565 free_netrc (netrc_list);
1567 xfree_null (opt.lfilename);
1568 xfree_null (opt.dir_prefix);
1569 xfree_null (opt.input_filename);
1570 xfree_null (opt.output_document);
1571 free_vec (opt.accepts);
1572 free_vec (opt.rejects);
1573 free_vec (opt.excludes);
1574 free_vec (opt.includes);
1575 free_vec (opt.domains);
1576 free_vec (opt.follow_tags);
1577 free_vec (opt.ignore_tags);
1578 xfree_null (opt.progress_type);
1579 xfree_null (opt.ftp_user);
1580 xfree_null (opt.ftp_passwd);
1581 xfree_null (opt.ftp_proxy);
1582 xfree_null (opt.https_proxy);
1583 xfree_null (opt.http_proxy);
1584 free_vec (opt.no_proxy);
1585 xfree_null (opt.useragent);
1586 xfree_null (opt.referer);
1587 xfree_null (opt.http_user);
1588 xfree_null (opt.http_passwd);
1589 free_vec (opt.user_headers);
1591 xfree_null (opt.cert_file);
1592 xfree_null (opt.private_key);
1593 xfree_null (opt.ca_directory);
1594 xfree_null (opt.ca_cert);
1595 xfree_null (opt.random_file);
1596 xfree_null (opt.egd_file);
1598 xfree_null (opt.bind_address);
1599 xfree_null (opt.cookies_input);
1600 xfree_null (opt.cookies_output);
1601 xfree_null (opt.user);
1602 xfree_null (opt.passwd);
1603 xfree_null (opt.base_href);
1605 #endif /* DEBUG_MALLOC */
1608 /* Unit testing routines. */
1613 test_commands_sorted()
1615 int prev_idx = 0, next_idx = 1;
1616 int command_count = countof (commands) - 1;
1618 while (next_idx <= command_count)
1620 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1623 mu_assert ("FAILED", false);
1636 test_cmd_spec_restrict_file_names()
1641 int expected_restrict_files_os;
1642 int expected_restrict_files_ctrl;
1643 int expected_restrict_files_case;
1646 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1647 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1648 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1649 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1652 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1657 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1660 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1661 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1662 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1663 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1665 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1666 res == test_array[i].result
1667 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1668 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1669 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1675 #endif /* TESTING */