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 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
111 { "askpassword", &opt.ask_passwd, cmd_boolean },
112 { "authnochallenge", &opt.auth_without_challenge,
114 { "background", &opt.background, cmd_boolean },
115 { "backupconverted", &opt.backup_converted, cmd_boolean },
116 { "backups", &opt.backups, cmd_number },
117 { "base", &opt.base_href, cmd_string },
118 { "bindaddress", &opt.bind_address, cmd_string },
120 { "cacertificate", &opt.ca_cert, cmd_file },
122 { "cache", &opt.allow_cache, cmd_boolean },
124 { "cadirectory", &opt.ca_directory, cmd_directory },
125 { "certificate", &opt.cert_file, cmd_file },
126 { "certificatetype", &opt.cert_type, cmd_cert_type },
127 { "checkcertificate", &opt.check_cert, cmd_boolean },
129 { "connecttimeout", &opt.connect_timeout, cmd_time },
130 { "contentdisposition", &opt.content_disposition, cmd_boolean },
131 { "continue", &opt.always_rest, cmd_boolean },
132 { "convertlinks", &opt.convert_links, cmd_boolean },
133 { "cookies", &opt.cookies, cmd_boolean },
134 { "cutdirs", &opt.cut_dirs, cmd_number },
136 { "debug", &opt.debug, cmd_boolean },
138 { "defaultpage", &opt.default_page, cmd_string},
139 { "deleteafter", &opt.delete_after, cmd_boolean },
140 { "dirprefix", &opt.dir_prefix, cmd_directory },
141 { "dirstruct", NULL, cmd_spec_dirstruct },
142 { "dnscache", &opt.dns_cache, cmd_boolean },
143 { "dnstimeout", &opt.dns_timeout, cmd_time },
144 { "domains", &opt.domains, cmd_vector },
145 { "dotbytes", &opt.dot_bytes, cmd_bytes },
146 { "dotsinline", &opt.dots_in_line, cmd_number },
147 { "dotspacing", &opt.dot_spacing, cmd_number },
148 { "dotstyle", &opt.dot_style, cmd_string },
150 { "egdfile", &opt.egd_file, cmd_file },
152 { "excludedirectories", &opt.excludes, cmd_directory_vector },
153 { "excludedomains", &opt.exclude_domains, cmd_vector },
154 { "followftp", &opt.follow_ftp, cmd_boolean },
155 { "followtags", &opt.follow_tags, cmd_vector },
156 { "forcehtml", &opt.force_html, cmd_boolean },
157 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
158 { "ftppassword", &opt.ftp_passwd, cmd_string },
159 { "ftpproxy", &opt.ftp_proxy, cmd_string },
160 { "ftpuser", &opt.ftp_user, cmd_string },
161 { "glob", &opt.ftp_glob, cmd_boolean },
162 { "header", NULL, cmd_spec_header },
163 { "htmlextension", &opt.html_extension, cmd_boolean },
164 { "htmlify", NULL, cmd_spec_htmlify },
165 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
166 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
167 { "httppassword", &opt.http_passwd, cmd_string },
168 { "httpproxy", &opt.http_proxy, cmd_string },
169 { "httpsproxy", &opt.https_proxy, cmd_string },
170 { "httpuser", &opt.http_user, cmd_string },
171 { "ignorecase", &opt.ignore_case, cmd_boolean },
172 { "ignorelength", &opt.ignore_length, cmd_boolean },
173 { "ignoretags", &opt.ignore_tags, cmd_vector },
174 { "includedirectories", &opt.includes, cmd_directory_vector },
176 { "inet4only", &opt.ipv4_only, cmd_boolean },
177 { "inet6only", &opt.ipv6_only, cmd_boolean },
179 { "input", &opt.input_filename, cmd_file },
180 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
181 { "limitrate", &opt.limit_rate, cmd_bytes },
182 { "loadcookies", &opt.cookies_input, cmd_file },
183 { "logfile", &opt.lfilename, cmd_file },
184 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
185 { "maxredirect", &opt.max_redirect, cmd_number },
186 { "mirror", NULL, cmd_spec_mirror },
187 { "netrc", &opt.netrc, cmd_boolean },
188 { "noclobber", &opt.noclobber, cmd_boolean },
189 { "noparent", &opt.no_parent, cmd_boolean },
190 { "noproxy", &opt.no_proxy, cmd_vector },
191 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
192 { "outputdocument", &opt.output_document, cmd_file },
193 { "pagerequisites", &opt.page_requisites, cmd_boolean },
194 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
195 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
196 { "password", &opt.passwd, cmd_string },
197 { "postdata", &opt.post_data, cmd_string },
198 { "postfile", &opt.post_file_name, cmd_file },
199 { "preferfamily", NULL, cmd_spec_prefer_family },
200 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
202 { "privatekey", &opt.private_key, cmd_file },
203 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
205 { "progress", &opt.progress_type, cmd_spec_progress },
206 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
207 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
208 { "proxypassword", &opt.proxy_passwd, cmd_string },
209 { "proxyuser", &opt.proxy_user, cmd_string },
210 { "quiet", &opt.quiet, cmd_boolean },
211 { "quota", &opt.quota, cmd_bytes_sum },
213 { "randomfile", &opt.random_file, cmd_file },
215 { "randomwait", &opt.random_wait, cmd_boolean },
216 { "readtimeout", &opt.read_timeout, cmd_time },
217 { "reclevel", &opt.reclevel, cmd_number_inf },
218 { "recursive", NULL, cmd_spec_recursive },
219 { "referer", &opt.referer, cmd_string },
220 { "reject", &opt.rejects, cmd_vector },
221 { "relativeonly", &opt.relative_only, cmd_boolean },
222 { "removelisting", &opt.remove_listing, cmd_boolean },
223 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
224 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
225 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
226 { "robots", &opt.use_robots, cmd_boolean },
227 { "savecookies", &opt.cookies_output, cmd_file },
228 { "saveheaders", &opt.save_headers, cmd_boolean },
230 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
232 { "serverresponse", &opt.server_response, cmd_boolean },
233 { "spanhosts", &opt.spanhost, cmd_boolean },
234 { "spider", &opt.spider, cmd_boolean },
235 { "strictcomments", &opt.strict_comments, cmd_boolean },
236 { "timeout", NULL, cmd_spec_timeout },
237 { "timestamping", &opt.timestamping, cmd_boolean },
238 { "tries", &opt.ntry, cmd_number_inf },
239 { "useproxy", &opt.use_proxy, cmd_boolean },
240 { "user", &opt.user, cmd_string },
241 { "useragent", NULL, cmd_spec_useragent },
242 { "verbose", NULL, cmd_spec_verbose },
243 { "wait", &opt.wait, cmd_time },
244 { "waitretry", &opt.waitretry, cmd_time },
246 { "wdebug", &opt.wdebug, cmd_boolean },
250 /* Look up CMDNAME in the commands[] and return its position in the
251 array. If CMDNAME is not found, return -1. */
254 command_by_name (const char *cmdname)
256 /* Use binary search for speed. Wget has ~100 commands, which
257 guarantees a worst case performance of 7 string comparisons. */
258 int lo = 0, hi = countof (commands) - 1;
262 int mid = (lo + hi) >> 1;
263 int cmp = strcasecmp (cmdname, commands[mid].name);
274 /* Reset the variables to default values. */
280 /* Most of the default values are 0 (and 0.0, NULL, and false).
281 Just reset everything, and fill in the non-zero values. Note
282 that initializing pointers to NULL this way is technically
283 illegal, but porting Wget to a machine where NULL is not all-zero
284 bit pattern will be the least of the implementors' worries. */
291 opt.add_hostdir = true;
295 opt.http_keep_alive = true;
296 opt.use_proxy = true;
297 tmp = getenv ("no_proxy");
299 opt.no_proxy = sepstring (tmp);
300 opt.prefer_family = prefer_none;
301 opt.allow_cache = true;
303 opt.read_timeout = 900;
304 opt.use_robots = true;
306 opt.remove_listing = true;
308 opt.dot_bytes = 1024;
309 opt.dot_spacing = 10;
310 opt.dots_in_line = 50;
312 opt.dns_cache = true;
316 opt.check_cert = true;
319 /* The default for file name restriction defaults to the OS type. */
320 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
321 opt.restrict_files_os = restrict_windows;
323 opt.restrict_files_os = restrict_unix;
325 opt.restrict_files_ctrl = true;
326 opt.restrict_files_case = restrict_no_case_restriction;
328 opt.max_redirect = 20;
333 /* Return the user's home directory (strdup-ed), or NULL if none is
338 static char buf[PATH_MAX];
343 home = getenv ("HOME");
347 /* Under MSDOS, if $HOME isn't defined, use the directory where
348 `wget.exe' resides. */
349 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
352 strcpy (buf, _w32_get_argv0 ());
353 p = strrchr (buf, '/'); /* djgpp */
355 p = strrchr (buf, '\\'); /* others */
359 #elif !defined(WINDOWS)
360 /* If HOME is not defined, try getting it from the password
362 struct passwd *pwd = getpwuid (getuid ());
363 if (!pwd || !pwd->pw_dir)
365 strcpy (buf, pwd->pw_dir);
368 /* Under Windows, if $HOME isn't defined, use the directory where
369 `wget.exe' resides. */
375 return home ? xstrdup (home) : NULL;
378 /* Check the 'WGETRC' environment variable and return the file name
379 if 'WGETRC' is set and is a valid file.
380 If the `WGETRC' variable exists but the file does not exist, the
381 function will exit(). */
383 wgetrc_env_file_name (void)
385 char *env = getenv ("WGETRC");
388 if (!file_exists_p (env))
390 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
394 return xstrdup (env);
399 /* Check for the existance of '$HOME/.wgetrc' and return it's path
400 if it exists and is set. */
402 wgetrc_user_file_name (void)
404 char *home = home_dir ();
407 file = aprintf ("%s/.wgetrc", home);
411 if (!file_exists_p (file))
419 /* Return the path to the user's .wgetrc. This is either the value of
420 `WGETRC' environment variable, or `$HOME/.wgetrc'.
422 Additionally, for windows, look in the directory where wget.exe
425 wgetrc_file_name (void)
428 char *file = wgetrc_env_file_name ();
432 file = wgetrc_user_file_name ();
435 /* Under Windows, if we still haven't found .wgetrc, look for the file
436 `wget.ini' in the directory where `wget.exe' resides; we do this for
437 backward compatibility with previous versions of Wget.
438 SYSTEM_WGETRC should not be defined under WINDOWS. */
440 if (!file || !file_exists_p (file))
446 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 *);
471 static bool setval_internal_tilde (int, const char *, const char *);
473 /* Initialize variables from a wgetrc file. Returns zero (failure) if
474 there were errors in the file. */
477 run_wgetrc (const char *file)
484 fp = fopen (file, "rb");
487 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
488 file, strerror (errno));
489 return true; /* not a fatal error */
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_tilde (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);
534 /* Initialize the defaults and run the system wgetrc and user's own
539 char *file, *env_sysrc;
542 /* Load the hard-coded defaults. */
545 /* Run a non-standard system rc file when the according environment
546 variable has been set. For internal testing purposes only! */
547 env_sysrc = getenv ("SYSTEM_WGETRC");
548 if (env_sysrc && file_exists_p (env_sysrc))
549 ok &= run_wgetrc (env_sysrc);
550 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
552 else if (file_exists_p (SYSTEM_WGETRC))
553 ok &= run_wgetrc (SYSTEM_WGETRC);
555 /* Override it with your own, if one exists. */
556 file = wgetrc_file_name ();
559 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
560 something like realpath() before comparing them with `strcmp' */
562 if (!strcmp (file, SYSTEM_WGETRC))
564 fprintf (stderr, _("\
565 %s: Warning: Both system and user wgetrc point to %s.\n"),
566 exec_name, quote (file));
570 ok &= run_wgetrc (file);
572 /* If there were errors processing either `.wgetrc', abort. */
580 /* Remove dashes and underscores from S, modifying S in the
586 char *t = s; /* t - tortoise */
587 char *h = s; /* h - hare */
589 if (*h == '_' || *h == '-')
596 /* Parse the line pointed by line, with the syntax:
597 <sp>* command <sp>* = <sp>* value <sp>*
598 Uses malloc to allocate space for command and value.
600 Returns one of line_ok, line_empty, line_syntax_error, or
601 line_unknown_command.
603 In case of line_ok, *COM and *VAL point to freshly allocated
604 strings, and *COMIND points to com's index. In case of error or
605 empty line, their values are unmodified. */
607 static enum parse_line
608 parse_line (const char *line, char **com, char **val, int *comind)
611 const char *end = line + strlen (line);
612 const char *cmdstart, *cmdend;
613 const char *valstart, *valend;
618 /* Skip leading and trailing whitespace. */
619 while (*line && c_isspace (*line))
621 while (end > line && c_isspace (end[-1]))
624 /* Skip empty lines and comments. */
625 if (!*line || *line == '#')
631 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
635 /* Skip '=', as well as any space before or after it. */
636 while (p < end && c_isspace (*p))
638 if (p == end || *p != '=')
639 return line_syntax_error;
641 while (p < end && c_isspace (*p))
647 /* The syntax is valid (even though the command might not be). Fill
648 in the command name and value. */
649 *com = strdupdelim (cmdstart, cmdend);
650 *val = strdupdelim (valstart, valend);
652 /* The line now known to be syntactically correct. Check whether
653 the command is valid. */
654 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
656 ind = command_by_name (cmdcopy);
658 return line_unknown_command;
660 /* Report success to the caller. */
665 #if defined(WINDOWS) || defined(MSDOS)
666 # define ISSEP(c) ((c) == '/' || (c) == '\\')
668 # define ISSEP(c) ((c) == '/')
671 /* Run commands[comind].action. */
674 setval_internal (int comind, const char *com, const char *val)
676 assert (0 <= comind && ((size_t) comind) < countof (commands));
677 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
678 return commands[comind].action (com, val, commands[comind].place);
682 setval_internal_tilde (int comind, const char *com, const char *val)
688 ret = setval_internal (comind, com, val);
690 /* We make tilde expansion for cmd_file and cmd_directory */
691 if (((commands[comind].action == cmd_file) ||
692 (commands[comind].action == cmd_directory))
693 && ret && (*val == '~' && ISSEP (val[1])))
695 pstring = commands[comind].place;
699 homelen = strlen (home);
700 while (homelen && ISSEP (home[homelen - 1]))
701 home[--homelen] = '\0';
703 /* Skip the leading "~/". */
704 for (++val; ISSEP (*val); val++)
706 *pstring = concat_strings (home, "/", val, (char *)0);
712 /* Run command COM with value VAL. If running the command produces an
713 error, report the error and exit.
715 This is intended to be called from main() to modify Wget's behavior
716 through command-line switches. Since COM is hard-coded in main(),
717 it is not canonicalized, and this aborts when COM is not found.
719 If COMIND's are exported to init.h, this function will be changed
720 to accept COMIND directly. */
723 setoptval (const char *com, const char *val, const char *optname)
725 /* Prepend "--" to OPTNAME. */
726 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
729 strcpy (dd_optname + 2, optname);
731 assert (val != NULL);
732 if (!setval_internal (command_by_name (com), dd_optname, val))
736 /* Parse OPT into command and value and run it. For example,
737 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
738 This is used by the `--execute' flag in main.c. */
741 run_command (const char *opt)
745 switch (parse_line (opt, &com, &val, &comind))
748 if (!setval_internal (comind, com, val))
754 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
755 exec_name, quote (opt));
760 /* Generic helper functions, for use with `commands'. */
762 /* Forward declarations: */
767 static bool decode_string (const char *, const struct decode_item *, int, int *);
768 static bool simple_atoi (const char *, const char *, int *);
769 static bool simple_atof (const char *, const char *, double *);
771 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
773 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
774 && c_tolower((p)[1]) == (c1) \
777 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
778 && c_tolower((p)[1]) == (c1) \
779 && c_tolower((p)[2]) == (c2) \
783 /* Store the boolean value from VAL to PLACE. COM is ignored,
784 except for error messages. */
786 cmd_boolean (const char *com, const char *val, void *place)
790 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
791 /* "on", "yes" and "1" mean true. */
793 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
794 /* "off", "no" and "0" mean false. */
799 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
800 exec_name, com, quote (val));
804 *(bool *) place = value;
808 /* Set the non-negative integer value from VAL to PLACE. With
809 incorrect specification, the number remains unchanged. */
811 cmd_number (const char *com, const char *val, void *place)
813 if (!simple_atoi (val, val + strlen (val), place)
814 || *(int *) place < 0)
816 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
817 exec_name, com, quote (val));
823 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
825 cmd_number_inf (const char *com, const char *val, void *place)
827 if (!strcasecmp (val, "inf"))
832 return cmd_number (com, val, place);
835 /* Copy (strdup) the string at COM to a new location and place a
836 pointer to *PLACE. */
838 cmd_string (const char *com, const char *val, void *place)
840 char **pstring = (char **)place;
842 xfree_null (*pstring);
843 *pstring = xstrdup (val);
848 /* Like the above, but handles tilde-expansion when reading a user's
849 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
850 gets expanded to the user's home directory. */
852 cmd_file (const char *com, const char *val, void *place)
854 char **pstring = (char **)place;
856 xfree_null (*pstring);
858 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
860 *pstring = xstrdup (val);
862 #if defined(WINDOWS) || defined(MSDOS)
863 /* Convert "\" to "/". */
866 for (s = *pstring; *s; s++)
874 /* Like cmd_file, but strips trailing '/' characters. */
876 cmd_directory (const char *com, const char *val, void *place)
880 /* Call cmd_file() for tilde expansion and separator
881 canonicalization (backslash -> slash under Windows). These
882 things should perhaps be in a separate function. */
883 if (!cmd_file (com, val, place))
888 while (t > s && *--t == '/')
894 /* Split VAL by space to a vector of values, and append those values
895 to vector pointed to by the PLACE argument. If VAL is empty, the
896 PLACE vector is cleared instead. */
899 cmd_vector (const char *com, const char *val, void *place)
901 char ***pvec = (char ***)place;
904 *pvec = merge_vecs (*pvec, sepstring (val));
914 cmd_directory_vector (const char *com, const char *val, void *place)
916 char ***pvec = (char ***)place;
920 /* Strip the trailing slashes from directories. */
923 seps = sepstring (val);
924 for (t = seps; t && *t; t++)
926 int len = strlen (*t);
927 /* Skip degenerate case of root directory. */
930 if ((*t)[len - 1] == '/')
931 (*t)[len - 1] = '\0';
934 *pvec = merge_vecs (*pvec, seps);
944 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
945 "100k" or "2.5G" to a floating point number. */
948 parse_bytes_helper (const char *val, double *result)
951 const char *end = val + strlen (val);
953 /* Check for "inf". */
954 if (0 == strcmp (val, "inf"))
960 /* Strip trailing whitespace. */
961 while (val < end && c_isspace (end[-1]))
966 switch (c_tolower (end[-1]))
969 --end, mult = 1024.0;
972 --end, mult = 1048576.0;
975 --end, mult = 1073741824.0;
978 --end, mult = 1099511627776.0;
981 /* Not a recognized suffix: assume it's a digit. (If not,
982 simple_atof will raise an error.) */
986 /* Skip leading and trailing whitespace. */
987 while (val < end && c_isspace (*val))
989 while (val < end && c_isspace (end[-1]))
994 if (!simple_atof (val, end, &number) || number < 0)
997 *result = number * mult;
1001 /* Parse VAL as a number and set its value to PLACE (which should
1004 By default, the value is assumed to be in bytes. If "K", "M", or
1005 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1006 1<<30, respectively. Floating point values are allowed and are
1007 cast to integer before use. The idea is to be able to use things
1008 like 1.5k instead of "1536".
1010 The string "inf" is returned as 0.
1012 In case of error, false is returned and memory pointed to by PLACE
1013 remains unmodified. */
1016 cmd_bytes (const char *com, const char *val, void *place)
1019 if (!parse_bytes_helper (val, &byte_value))
1021 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1022 exec_name, com, quote (val));
1025 *(wgint *)place = (wgint)byte_value;
1029 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1030 SIZE_SUM. It works by converting the string to double, therefore
1031 working with values up to 2^53-1 without loss of precision. This
1032 value (8192 TB) is large enough to serve for a while. */
1035 cmd_bytes_sum (const char *com, const char *val, void *place)
1038 if (!parse_bytes_helper (val, &byte_value))
1040 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1041 exec_name, com, quote (val));
1044 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1048 /* Store the value of VAL to *OUT. The value is a time period, by
1049 default expressed in seconds, but also accepting suffixes "m", "h",
1050 "d", and "w" for minutes, hours, days, and weeks respectively. */
1053 cmd_time (const char *com, const char *val, void *place)
1055 double number, mult;
1056 const char *end = val + strlen (val);
1058 /* Strip trailing whitespace. */
1059 while (val < end && c_isspace (end[-1]))
1065 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1066 exec_name, com, quote (val));
1070 switch (c_tolower (end[-1]))
1073 --end, mult = 1; /* seconds */
1076 --end, mult = 60; /* minutes */
1079 --end, mult = 3600; /* hours */
1082 --end, mult = 86400.0; /* days */
1085 --end, mult = 604800.0; /* weeks */
1088 /* Not a recognized suffix: assume it belongs to the number.
1089 (If not, simple_atof will raise an error.) */
1093 /* Skip leading and trailing whitespace. */
1094 while (val < end && c_isspace (*val))
1096 while (val < end && c_isspace (end[-1]))
1101 if (!simple_atof (val, end, &number))
1104 *(double *)place = number * mult;
1110 cmd_cert_type (const char *com, const char *val, void *place)
1112 static const struct decode_item choices[] = {
1113 { "pem", keyfile_pem },
1114 { "der", keyfile_asn1 },
1115 { "asn1", keyfile_asn1 },
1117 int ok = decode_string (val, choices, countof (choices), place);
1119 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1124 /* Specialized helper functions, used by `commands' to handle some
1125 options specially. */
1127 static bool check_user_specified_header (const char *);
1130 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1132 if (!cmd_boolean (com, val, &opt.dirstruct))
1134 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1135 must be affected inversely. */
1137 opt.no_dirstruct = false;
1139 opt.no_dirstruct = true;
1144 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1146 /* Empty value means reset the list of headers. */
1149 free_vec (opt.user_headers);
1150 opt.user_headers = NULL;
1154 if (!check_user_specified_header (val))
1156 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1157 exec_name, com, quote (val));
1160 opt.user_headers = vec_append (opt.user_headers, val);
1165 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1167 int flag = cmd_boolean (com, val, &opt.htmlify);
1168 if (flag && !opt.htmlify)
1169 opt.remove_listing = false;
1173 /* Set the "mirror" mode. It means: recursive download, timestamping,
1174 no limit on max. recursion depth, and don't remove listings. */
1177 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1181 if (!cmd_boolean (com, val, &mirror))
1185 opt.recursive = true;
1186 if (!opt.no_dirstruct)
1187 opt.dirstruct = true;
1188 opt.timestamping = true;
1189 opt.reclevel = INFINITE_RECURSION;
1190 opt.remove_listing = false;
1195 /* Validate --prefer-family and set the choice. Allowed values are
1196 "IPv4", "IPv6", and "none". */
1199 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1201 static const struct decode_item choices[] = {
1202 { "IPv4", prefer_ipv4 },
1203 { "IPv6", prefer_ipv6 },
1204 { "none", prefer_none },
1206 int prefer_family = prefer_none;
1207 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1209 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1210 opt.prefer_family = prefer_family;
1214 /* Set progress.type to VAL, but verify that it's a valid progress
1215 implementation before that. */
1218 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1220 if (!valid_progress_implementation_p (val))
1222 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1223 exec_name, com, quote (val));
1226 xfree_null (opt.progress_type);
1228 /* Don't call set_progress_implementation here. It will be called
1229 in main() when it becomes clear what the log output is. */
1230 opt.progress_type = xstrdup (val);
1234 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1235 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1239 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1241 if (!cmd_boolean (com, val, &opt.recursive))
1245 if (opt.recursive && !opt.no_dirstruct)
1246 opt.dirstruct = true;
1252 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1254 int restrict_os = opt.restrict_files_os;
1255 int restrict_ctrl = opt.restrict_files_ctrl;
1256 int restrict_case = opt.restrict_files_case;
1260 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1264 end = strchr (val, ',');
1266 end = val + strlen (val);
1268 if (VAL_IS ("unix"))
1269 restrict_os = restrict_unix;
1270 else if (VAL_IS ("windows"))
1271 restrict_os = restrict_windows;
1272 else if (VAL_IS ("lowercase"))
1273 restrict_case = restrict_lowercase;
1274 else if (VAL_IS ("uppercase"))
1275 restrict_case = restrict_uppercase;
1276 else if (VAL_IS ("nocontrol"))
1277 restrict_ctrl = false;
1281 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1282 exec_name, com, quote (val));
1289 while (*val && *end);
1293 opt.restrict_files_os = restrict_os;
1294 opt.restrict_files_ctrl = restrict_ctrl;
1295 opt.restrict_files_case = restrict_case;
1302 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1304 static const struct decode_item choices[] = {
1305 { "auto", secure_protocol_auto },
1306 { "sslv2", secure_protocol_sslv2 },
1307 { "sslv3", secure_protocol_sslv3 },
1308 { "tlsv1", secure_protocol_tlsv1 },
1310 int ok = decode_string (val, choices, countof (choices), place);
1312 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1317 /* Set all three timeout values. */
1320 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1323 if (!cmd_time (com, val, &value))
1325 opt.read_timeout = value;
1326 opt.connect_timeout = value;
1327 opt.dns_timeout = value;
1332 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1334 /* Disallow embedded newlines. */
1335 if (strchr (val, '\n'))
1337 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1338 exec_name, com, quote (val));
1341 xfree_null (opt.useragent);
1342 opt.useragent = xstrdup (val);
1346 /* The "verbose" option cannot be cmd_boolean because the variable is
1347 not bool -- it's of type int (-1 means uninitialized because of
1348 some random hackery for disallowing -q -v). */
1351 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1354 if (cmd_boolean (com, val, &flag))
1362 /* Miscellaneous useful routines. */
1364 /* A very simple atoi clone, more useful than atoi because it works on
1365 delimited strings, and has error reportage. Returns true on success,
1366 false on failure. If successful, stores result to *DEST. */
1369 simple_atoi (const char *beg, const char *end, int *dest)
1372 bool negative = false;
1373 const char *p = beg;
1375 while (p < end && c_isspace (*p))
1377 if (p < end && (*p == '-' || *p == '+'))
1379 negative = (*p == '-');
1385 /* Read negative numbers in a separate loop because the most
1386 negative integer cannot be represented as a positive number. */
1389 for (; p < end && c_isdigit (*p); p++)
1391 int next = (10 * result) + (*p - '0');
1393 return false; /* overflow */
1397 for (; p < end && c_isdigit (*p); p++)
1399 int next = (10 * result) - (*p - '0');
1401 return false; /* underflow */
1412 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1413 doesn't handle exponential notation. Returns true on success,
1414 false on failure. In case of success, stores its result to
1418 simple_atof (const char *beg, const char *end, double *dest)
1422 bool negative = false;
1423 bool seen_dot = false;
1424 bool seen_digit = false;
1427 const char *p = beg;
1429 while (p < end && c_isspace (*p))
1431 if (p < end && (*p == '-' || *p == '+'))
1433 negative = (*p == '-');
1437 for (; p < end; p++)
1443 result = (10 * result) + (ch - '0');
1445 result += (ch - '0') / (divider *= 10);
1467 /* Verify that the user-specified header in S is valid. It must
1468 contain a colon preceded by non-white-space characters and must not
1469 contain newlines. */
1472 check_user_specified_header (const char *s)
1476 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1478 /* The header MUST contain `:' preceded by at least one
1479 non-whitespace character. */
1480 if (*p != ':' || p == s)
1482 /* The header MUST NOT contain newlines. */
1483 if (strchr (s, '\n'))
1488 /* Decode VAL into a number, according to ITEMS. */
1491 decode_string (const char *val, const struct decode_item *items, int itemcount,
1495 for (i = 0; i < itemcount; i++)
1496 if (0 == strcasecmp (val, items[i].name))
1498 *place = items[i].code;
1505 void cleanup_html_url (void);
1508 /* Free the memory allocated by global variables. */
1512 /* Free external resources, close files, etc. */
1515 fclose (output_stream);
1516 /* No need to check for error because Wget flushes its output (and
1517 checks for errors) after any data arrives. */
1519 /* We're exiting anyway so there's no real need to call free()
1520 hundreds of times. Skipping the frees will make Wget exit
1523 However, when detecting leaks, it's crucial to free() everything
1524 because then you can find the real leaks, i.e. the allocated
1525 memory which grows with the size of the program. */
1531 cleanup_html_url ();
1536 extern acc_t *netrc_list;
1537 free_netrc (netrc_list);
1539 xfree_null (opt.lfilename);
1540 xfree_null (opt.dir_prefix);
1541 xfree_null (opt.input_filename);
1542 xfree_null (opt.output_document);
1543 free_vec (opt.accepts);
1544 free_vec (opt.rejects);
1545 free_vec (opt.excludes);
1546 free_vec (opt.includes);
1547 free_vec (opt.domains);
1548 free_vec (opt.follow_tags);
1549 free_vec (opt.ignore_tags);
1550 xfree_null (opt.progress_type);
1551 xfree_null (opt.ftp_user);
1552 xfree_null (opt.ftp_passwd);
1553 xfree_null (opt.ftp_proxy);
1554 xfree_null (opt.https_proxy);
1555 xfree_null (opt.http_proxy);
1556 free_vec (opt.no_proxy);
1557 xfree_null (opt.useragent);
1558 xfree_null (opt.referer);
1559 xfree_null (opt.http_user);
1560 xfree_null (opt.http_passwd);
1561 free_vec (opt.user_headers);
1563 xfree_null (opt.cert_file);
1564 xfree_null (opt.private_key);
1565 xfree_null (opt.ca_directory);
1566 xfree_null (opt.ca_cert);
1567 xfree_null (opt.random_file);
1568 xfree_null (opt.egd_file);
1570 xfree_null (opt.bind_address);
1571 xfree_null (opt.cookies_input);
1572 xfree_null (opt.cookies_output);
1573 xfree_null (opt.user);
1574 xfree_null (opt.passwd);
1575 xfree_null (opt.base_href);
1577 #endif /* DEBUG_MALLOC */
1580 /* Unit testing routines. */
1585 test_commands_sorted()
1587 int prev_idx = 0, next_idx = 1;
1588 int command_count = countof (commands) - 1;
1590 while (next_idx <= command_count)
1592 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1595 mu_assert ("FAILED", false);
1608 test_cmd_spec_restrict_file_names()
1613 int expected_restrict_files_os;
1614 int expected_restrict_files_ctrl;
1615 int expected_restrict_files_case;
1618 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1619 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1620 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1621 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1624 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1629 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1632 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1633 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1634 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1635 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1637 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1638 res == test_array[i].result
1639 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1640 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1641 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1647 #endif /* TESTING */