1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
61 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
62 otherwise, it will be performed by the shell. This variable will
63 be set by the wgetrc-reading function. */
65 static bool enable_tilde_expansion;
68 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
70 CMD_DECLARE (cmd_boolean);
71 CMD_DECLARE (cmd_bytes);
72 CMD_DECLARE (cmd_bytes_sum);
74 CMD_DECLARE (cmd_cert_type);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_number);
78 CMD_DECLARE (cmd_number_inf);
79 CMD_DECLARE (cmd_string);
80 CMD_DECLARE (cmd_file);
81 CMD_DECLARE (cmd_directory);
82 CMD_DECLARE (cmd_time);
83 CMD_DECLARE (cmd_vector);
85 CMD_DECLARE (cmd_spec_dirstruct);
86 CMD_DECLARE (cmd_spec_header);
87 CMD_DECLARE (cmd_spec_htmlify);
88 CMD_DECLARE (cmd_spec_mirror);
89 CMD_DECLARE (cmd_spec_prefer_family);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_secure_protocol);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
98 CMD_DECLARE (cmd_spec_verbose);
100 /* List of recognized commands, each consisting of name, place and
101 function. When adding a new command, simply add it to the list,
102 but be sure to keep the list sorted alphabetically, as
103 command_by_name's binary search depends on it. Also, be sure to
104 add any entries that allocate memory (e.g. cmd_string and
105 cmd_vector) to the cleanup() function below. */
107 static const struct {
110 bool (*action) (const char *, const char *, void *);
112 /* KEEP THIS LIST ALPHABETICALLY SORTED */
113 { "accept", &opt.accepts, cmd_vector },
114 { "addhostdir", &opt.add_hostdir, cmd_boolean },
115 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
116 { "askpassword", &opt.ask_passwd, cmd_boolean },
117 { "authnochallenge", &opt.auth_without_challenge,
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "contentdisposition", &opt.content_disposition, cmd_boolean },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "defaultpage", &opt.default_page, cmd_string},
144 { "deleteafter", &opt.delete_after, cmd_boolean },
145 { "dirprefix", &opt.dir_prefix, cmd_directory },
146 { "dirstruct", NULL, cmd_spec_dirstruct },
147 { "dnscache", &opt.dns_cache, cmd_boolean },
148 { "dnstimeout", &opt.dns_timeout, cmd_time },
149 { "domains", &opt.domains, cmd_vector },
150 { "dotbytes", &opt.dot_bytes, cmd_bytes },
151 { "dotsinline", &opt.dots_in_line, cmd_number },
152 { "dotspacing", &opt.dot_spacing, cmd_number },
153 { "dotstyle", &opt.dot_style, cmd_string },
155 { "egdfile", &opt.egd_file, cmd_file },
157 { "excludedirectories", &opt.excludes, cmd_directory_vector },
158 { "excludedomains", &opt.exclude_domains, cmd_vector },
159 { "followftp", &opt.follow_ftp, cmd_boolean },
160 { "followtags", &opt.follow_tags, cmd_vector },
161 { "forcehtml", &opt.force_html, cmd_boolean },
162 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
163 { "ftppassword", &opt.ftp_passwd, cmd_string },
164 { "ftpproxy", &opt.ftp_proxy, cmd_string },
165 { "ftpuser", &opt.ftp_user, cmd_string },
166 { "glob", &opt.ftp_glob, cmd_boolean },
167 { "header", NULL, cmd_spec_header },
168 { "htmlextension", &opt.html_extension, cmd_boolean },
169 { "htmlify", NULL, cmd_spec_htmlify },
170 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
171 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
172 { "httppassword", &opt.http_passwd, cmd_string },
173 { "httpproxy", &opt.http_proxy, cmd_string },
174 { "httpsproxy", &opt.https_proxy, cmd_string },
175 { "httpuser", &opt.http_user, cmd_string },
176 { "ignorecase", &opt.ignore_case, cmd_boolean },
177 { "ignorelength", &opt.ignore_length, cmd_boolean },
178 { "ignoretags", &opt.ignore_tags, cmd_vector },
179 { "includedirectories", &opt.includes, cmd_directory_vector },
181 { "inet4only", &opt.ipv4_only, cmd_boolean },
182 { "inet6only", &opt.ipv6_only, cmd_boolean },
184 { "input", &opt.input_filename, cmd_file },
185 { "iri", &opt.enable_iri, cmd_boolean },
186 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
187 { "limitrate", &opt.limit_rate, cmd_bytes },
188 { "loadcookies", &opt.cookies_input, cmd_file },
189 { "locale", &opt.locale, cmd_string },
190 { "logfile", &opt.lfilename, cmd_file },
191 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
192 { "maxredirect", &opt.max_redirect, cmd_number },
193 { "mirror", NULL, cmd_spec_mirror },
194 { "netrc", &opt.netrc, cmd_boolean },
195 { "noclobber", &opt.noclobber, cmd_boolean },
196 { "noparent", &opt.no_parent, cmd_boolean },
197 { "noproxy", &opt.no_proxy, cmd_vector },
198 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
199 { "outputdocument", &opt.output_document, cmd_file },
200 { "pagerequisites", &opt.page_requisites, cmd_boolean },
201 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
202 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
203 { "password", &opt.passwd, cmd_string },
204 { "postdata", &opt.post_data, cmd_string },
205 { "postfile", &opt.post_file_name, cmd_file },
206 { "preferfamily", NULL, cmd_spec_prefer_family },
207 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
209 { "privatekey", &opt.private_key, cmd_file },
210 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
212 { "progress", &opt.progress_type, cmd_spec_progress },
213 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
214 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
215 { "proxypassword", &opt.proxy_passwd, cmd_string },
216 { "proxyuser", &opt.proxy_user, cmd_string },
217 { "quiet", &opt.quiet, cmd_boolean },
218 { "quota", &opt.quota, cmd_bytes_sum },
220 { "randomfile", &opt.random_file, cmd_file },
222 { "randomwait", &opt.random_wait, cmd_boolean },
223 { "readtimeout", &opt.read_timeout, cmd_time },
224 { "reclevel", &opt.reclevel, cmd_number_inf },
225 { "recursive", NULL, cmd_spec_recursive },
226 { "referer", &opt.referer, cmd_string },
227 { "reject", &opt.rejects, cmd_vector },
228 { "relativeonly", &opt.relative_only, cmd_boolean },
229 { "remoteencoding", &opt.encoding_remote, cmd_string },
230 { "removelisting", &opt.remove_listing, cmd_boolean },
231 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
232 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
233 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
234 { "robots", &opt.use_robots, cmd_boolean },
235 { "savecookies", &opt.cookies_output, cmd_file },
236 { "saveheaders", &opt.save_headers, cmd_boolean },
238 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
240 { "serverresponse", &opt.server_response, cmd_boolean },
241 { "spanhosts", &opt.spanhost, cmd_boolean },
242 { "spider", &opt.spider, cmd_boolean },
243 { "strictcomments", &opt.strict_comments, cmd_boolean },
244 { "timeout", NULL, cmd_spec_timeout },
245 { "timestamping", &opt.timestamping, cmd_boolean },
246 { "tries", &opt.ntry, cmd_number_inf },
247 { "useproxy", &opt.use_proxy, cmd_boolean },
248 { "user", &opt.user, cmd_string },
249 { "useragent", NULL, cmd_spec_useragent },
250 { "verbose", NULL, cmd_spec_verbose },
251 { "wait", &opt.wait, cmd_time },
252 { "waitretry", &opt.waitretry, cmd_time },
254 { "wdebug", &opt.wdebug, cmd_boolean },
258 /* Look up CMDNAME in the commands[] and return its position in the
259 array. If CMDNAME is not found, return -1. */
262 command_by_name (const char *cmdname)
264 /* Use binary search for speed. Wget has ~100 commands, which
265 guarantees a worst case performance of 7 string comparisons. */
266 int lo = 0, hi = countof (commands) - 1;
270 int mid = (lo + hi) >> 1;
271 int cmp = strcasecmp (cmdname, commands[mid].name);
282 /* Reset the variables to default values. */
288 /* Most of the default values are 0 (and 0.0, NULL, and false).
289 Just reset everything, and fill in the non-zero values. Note
290 that initializing pointers to NULL this way is technically
291 illegal, but porting Wget to a machine where NULL is not all-zero
292 bit pattern will be the least of the implementors' worries. */
299 opt.add_hostdir = true;
303 opt.http_keep_alive = true;
304 opt.use_proxy = true;
305 tmp = getenv ("no_proxy");
307 opt.no_proxy = sepstring (tmp);
308 opt.prefer_family = prefer_none;
309 opt.allow_cache = true;
311 opt.read_timeout = 900;
312 opt.use_robots = true;
314 opt.remove_listing = true;
316 opt.dot_bytes = 1024;
317 opt.dot_spacing = 10;
318 opt.dots_in_line = 50;
320 opt.dns_cache = true;
324 opt.check_cert = true;
327 /* The default for file name restriction defaults to the OS type. */
328 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
329 opt.restrict_files_os = restrict_windows;
331 opt.restrict_files_os = restrict_unix;
333 opt.restrict_files_ctrl = true;
334 opt.restrict_files_case = restrict_no_case_restriction;
336 opt.max_redirect = 20;
339 opt.enable_iri = true;
341 opt.enable_iri = false;
344 opt.encoding_remote = NULL;
347 /* Return the user's home directory (strdup-ed), or NULL if none is
352 static char buf[PATH_MAX];
357 home = getenv ("HOME");
361 /* Under MSDOS, if $HOME isn't defined, use the directory where
362 `wget.exe' resides. */
363 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
366 strcpy (buf, _w32_get_argv0 ());
367 p = strrchr (buf, '/'); /* djgpp */
369 p = strrchr (buf, '\\'); /* others */
373 #elif !defined(WINDOWS)
374 /* If HOME is not defined, try getting it from the password
376 struct passwd *pwd = getpwuid (getuid ());
377 if (!pwd || !pwd->pw_dir)
379 strcpy (buf, pwd->pw_dir);
382 /* Under Windows, if $HOME isn't defined, use the directory where
383 `wget.exe' resides. */
389 return home ? xstrdup (home) : NULL;
392 /* Check the 'WGETRC' environment variable and return the file name
393 if 'WGETRC' is set and is a valid file.
394 If the `WGETRC' variable exists but the file does not exist, the
395 function will exit(). */
397 wgetrc_env_file_name (void)
399 char *env = getenv ("WGETRC");
402 if (!file_exists_p (env))
404 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
408 return xstrdup (env);
413 /* Check for the existance of '$HOME/.wgetrc' and return it's path
414 if it exists and is set. */
416 wgetrc_user_file_name (void)
418 char *home = home_dir ();
421 file = aprintf ("%s/.wgetrc", home);
425 if (!file_exists_p (file))
433 /* Return the path to the user's .wgetrc. This is either the value of
434 `WGETRC' environment variable, or `$HOME/.wgetrc'.
436 Additionally, for windows, look in the directory where wget.exe
439 wgetrc_file_name (void)
442 char *file = wgetrc_env_file_name ();
446 file = wgetrc_user_file_name ();
449 /* Under Windows, if we still haven't found .wgetrc, look for the file
450 `wget.ini' in the directory where `wget.exe' resides; we do this for
451 backward compatibility with previous versions of Wget.
452 SYSTEM_WGETRC should not be defined under WINDOWS. */
454 if (!file || !file_exists_p (file))
460 file = aprintf ("%s/wget.ini", home);
467 if (!file_exists_p (file))
475 /* Return values of parse_line. */
483 static enum parse_line parse_line (const char *, char **, char **, int *);
484 static bool setval_internal (int, const char *, const char *);
486 /* Initialize variables from a wgetrc file. Returns zero (failure) if
487 there were errors in the file. */
490 run_wgetrc (const char *file)
497 fp = fopen (file, "rb");
500 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
501 file, strerror (errno));
502 return true; /* not a fatal error */
504 enable_tilde_expansion = true;
506 while ((line = read_whole_line (fp)) != NULL)
508 char *com = NULL, *val = NULL;
511 /* Parse the line. */
512 switch (parse_line (line, &com, &val, &comind))
515 /* If everything is OK, set the value. */
516 if (!setval_internal (comind, com, val))
518 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
519 exec_name, file, ln);
523 case line_syntax_error:
524 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
525 exec_name, file, ln);
528 case line_unknown_command:
529 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
530 exec_name, quote (com), file, ln);
543 enable_tilde_expansion = false;
549 /* Initialize the defaults and run the system wgetrc and user's own
557 /* Load the hard-coded defaults. */
560 /* If SYSTEM_WGETRC is defined, use it. */
562 if (file_exists_p (SYSTEM_WGETRC))
563 ok &= run_wgetrc (SYSTEM_WGETRC);
565 /* Override it with your own, if one exists. */
566 file = wgetrc_file_name ();
569 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
570 something like realpath() before comparing them with `strcmp' */
572 if (!strcmp (file, SYSTEM_WGETRC))
574 fprintf (stderr, _("\
575 %s: Warning: Both system and user wgetrc point to %s.\n"),
576 exec_name, quote (file));
580 ok &= run_wgetrc (file);
582 /* If there were errors processing either `.wgetrc', abort. */
590 /* Remove dashes and underscores from S, modifying S in the
596 char *t = s; /* t - tortoise */
597 char *h = s; /* h - hare */
599 if (*h == '_' || *h == '-')
606 /* Parse the line pointed by line, with the syntax:
607 <sp>* command <sp>* = <sp>* value <sp>*
608 Uses malloc to allocate space for command and value.
610 Returns one of line_ok, line_empty, line_syntax_error, or
611 line_unknown_command.
613 In case of line_ok, *COM and *VAL point to freshly allocated
614 strings, and *COMIND points to com's index. In case of error or
615 empty line, their values are unmodified. */
617 static enum parse_line
618 parse_line (const char *line, char **com, char **val, int *comind)
621 const char *end = line + strlen (line);
622 const char *cmdstart, *cmdend;
623 const char *valstart, *valend;
628 /* Skip leading and trailing whitespace. */
629 while (*line && c_isspace (*line))
631 while (end > line && c_isspace (end[-1]))
634 /* Skip empty lines and comments. */
635 if (!*line || *line == '#')
641 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
645 /* Skip '=', as well as any space before or after it. */
646 while (p < end && c_isspace (*p))
648 if (p == end || *p != '=')
649 return line_syntax_error;
651 while (p < end && c_isspace (*p))
657 /* The syntax is valid (even though the command might not be). Fill
658 in the command name and value. */
659 *com = strdupdelim (cmdstart, cmdend);
660 *val = strdupdelim (valstart, valend);
662 /* The line now known to be syntactically correct. Check whether
663 the command is valid. */
664 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
666 ind = command_by_name (cmdcopy);
668 return line_unknown_command;
670 /* Report success to the caller. */
675 /* Run commands[comind].action. */
678 setval_internal (int comind, const char *com, const char *val)
680 assert (0 <= comind && ((size_t) comind) < countof (commands));
681 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
682 return commands[comind].action (com, val, commands[comind].place);
685 /* Run command COM with value VAL. If running the command produces an
686 error, report the error and exit.
688 This is intended to be called from main() to modify Wget's behavior
689 through command-line switches. Since COM is hard-coded in main(),
690 it is not canonicalized, and this aborts when COM is not found.
692 If COMIND's are exported to init.h, this function will be changed
693 to accept COMIND directly. */
696 setoptval (const char *com, const char *val, const char *optname)
698 /* Prepend "--" to OPTNAME. */
699 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
702 strcpy (dd_optname + 2, optname);
704 assert (val != NULL);
705 if (!setval_internal (command_by_name (com), dd_optname, val))
709 /* Parse OPT into command and value and run it. For example,
710 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
711 This is used by the `--execute' flag in main.c. */
714 run_command (const char *opt)
718 switch (parse_line (opt, &com, &val, &comind))
721 if (!setval_internal (comind, com, val))
727 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
728 exec_name, quote (opt));
733 /* Generic helper functions, for use with `commands'. */
735 /* Forward declarations: */
740 static bool decode_string (const char *, const struct decode_item *, int, int *);
741 static bool simple_atoi (const char *, const char *, int *);
742 static bool simple_atof (const char *, const char *, double *);
744 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
746 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
747 && c_tolower((p)[1]) == (c1) \
750 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
751 && c_tolower((p)[1]) == (c1) \
752 && c_tolower((p)[2]) == (c2) \
756 /* Store the boolean value from VAL to PLACE. COM is ignored,
757 except for error messages. */
759 cmd_boolean (const char *com, const char *val, void *place)
763 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
764 /* "on", "yes" and "1" mean true. */
766 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
767 /* "off", "no" and "0" mean false. */
772 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
773 exec_name, com, quote (val));
777 *(bool *) place = value;
781 /* Set the non-negative integer value from VAL to PLACE. With
782 incorrect specification, the number remains unchanged. */
784 cmd_number (const char *com, const char *val, void *place)
786 if (!simple_atoi (val, val + strlen (val), place)
787 || *(int *) place < 0)
789 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
790 exec_name, com, quote (val));
796 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
798 cmd_number_inf (const char *com, const char *val, void *place)
800 if (!strcasecmp (val, "inf"))
805 return cmd_number (com, val, place);
808 /* Copy (strdup) the string at COM to a new location and place a
809 pointer to *PLACE. */
811 cmd_string (const char *com, const char *val, void *place)
813 char **pstring = (char **)place;
815 xfree_null (*pstring);
816 *pstring = xstrdup (val);
820 #if defined(WINDOWS) || defined(MSDOS)
821 # define ISSEP(c) ((c) == '/' || (c) == '\\')
823 # define ISSEP(c) ((c) == '/')
826 /* Like the above, but handles tilde-expansion when reading a user's
827 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
828 gets expanded to the user's home directory. */
830 cmd_file (const char *com, const char *val, void *place)
832 char **pstring = (char **)place;
834 xfree_null (*pstring);
836 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
838 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
841 *pstring = xstrdup (val);
846 char *home = home_dir ();
850 homelen = strlen (home);
851 while (homelen && ISSEP (home[homelen - 1]))
852 home[--homelen] = '\0';
854 /* Skip the leading "~/". */
855 for (++val; ISSEP (*val); val++)
858 *pstring = concat_strings (home, "/", val, (char *) 0);
861 #if defined(WINDOWS) || defined(MSDOS)
862 /* Convert "\" to "/". */
865 for (s = *pstring; *s; s++)
873 /* Like cmd_file, but strips trailing '/' characters. */
875 cmd_directory (const char *com, const char *val, void *place)
879 /* Call cmd_file() for tilde expansion and separator
880 canonicalization (backslash -> slash under Windows). These
881 things should perhaps be in a separate function. */
882 if (!cmd_file (com, val, place))
887 while (t > s && *--t == '/')
893 /* Split VAL by space to a vector of values, and append those values
894 to vector pointed to by the PLACE argument. If VAL is empty, the
895 PLACE vector is cleared instead. */
898 cmd_vector (const char *com, const char *val, void *place)
900 char ***pvec = (char ***)place;
903 *pvec = merge_vecs (*pvec, sepstring (val));
913 cmd_directory_vector (const char *com, const char *val, void *place)
915 char ***pvec = (char ***)place;
919 /* Strip the trailing slashes from directories. */
922 seps = sepstring (val);
923 for (t = seps; t && *t; t++)
925 int len = strlen (*t);
926 /* Skip degenerate case of root directory. */
929 if ((*t)[len - 1] == '/')
930 (*t)[len - 1] = '\0';
933 *pvec = merge_vecs (*pvec, seps);
943 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
944 "100k" or "2.5G" to a floating point number. */
947 parse_bytes_helper (const char *val, double *result)
950 const char *end = val + strlen (val);
952 /* Check for "inf". */
953 if (0 == strcmp (val, "inf"))
959 /* Strip trailing whitespace. */
960 while (val < end && c_isspace (end[-1]))
965 switch (c_tolower (end[-1]))
968 --end, mult = 1024.0;
971 --end, mult = 1048576.0;
974 --end, mult = 1073741824.0;
977 --end, mult = 1099511627776.0;
980 /* Not a recognized suffix: assume it's a digit. (If not,
981 simple_atof will raise an error.) */
985 /* Skip leading and trailing whitespace. */
986 while (val < end && c_isspace (*val))
988 while (val < end && c_isspace (end[-1]))
993 if (!simple_atof (val, end, &number) || number < 0)
996 *result = number * mult;
1000 /* Parse VAL as a number and set its value to PLACE (which should
1003 By default, the value is assumed to be in bytes. If "K", "M", or
1004 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1005 1<<30, respectively. Floating point values are allowed and are
1006 cast to integer before use. The idea is to be able to use things
1007 like 1.5k instead of "1536".
1009 The string "inf" is returned as 0.
1011 In case of error, false is returned and memory pointed to by PLACE
1012 remains unmodified. */
1015 cmd_bytes (const char *com, const char *val, void *place)
1018 if (!parse_bytes_helper (val, &byte_value))
1020 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1021 exec_name, com, quote (val));
1024 *(wgint *)place = (wgint)byte_value;
1028 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1029 SIZE_SUM. It works by converting the string to double, therefore
1030 working with values up to 2^53-1 without loss of precision. This
1031 value (8192 TB) is large enough to serve for a while. */
1034 cmd_bytes_sum (const char *com, const char *val, void *place)
1037 if (!parse_bytes_helper (val, &byte_value))
1039 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1040 exec_name, com, quote (val));
1043 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1047 /* Store the value of VAL to *OUT. The value is a time period, by
1048 default expressed in seconds, but also accepting suffixes "m", "h",
1049 "d", and "w" for minutes, hours, days, and weeks respectively. */
1052 cmd_time (const char *com, const char *val, void *place)
1054 double number, mult;
1055 const char *end = val + strlen (val);
1057 /* Strip trailing whitespace. */
1058 while (val < end && c_isspace (end[-1]))
1064 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1065 exec_name, com, quote (val));
1069 switch (c_tolower (end[-1]))
1072 --end, mult = 1; /* seconds */
1075 --end, mult = 60; /* minutes */
1078 --end, mult = 3600; /* hours */
1081 --end, mult = 86400.0; /* days */
1084 --end, mult = 604800.0; /* weeks */
1087 /* Not a recognized suffix: assume it belongs to the number.
1088 (If not, simple_atof will raise an error.) */
1092 /* Skip leading and trailing whitespace. */
1093 while (val < end && c_isspace (*val))
1095 while (val < end && c_isspace (end[-1]))
1100 if (!simple_atof (val, end, &number))
1103 *(double *)place = number * mult;
1109 cmd_cert_type (const char *com, const char *val, void *place)
1111 static const struct decode_item choices[] = {
1112 { "pem", keyfile_pem },
1113 { "der", keyfile_asn1 },
1114 { "asn1", keyfile_asn1 },
1116 int ok = decode_string (val, choices, countof (choices), place);
1118 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1123 /* Specialized helper functions, used by `commands' to handle some
1124 options specially. */
1126 static bool check_user_specified_header (const char *);
1129 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1131 if (!cmd_boolean (com, val, &opt.dirstruct))
1133 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1134 must be affected inversely. */
1136 opt.no_dirstruct = false;
1138 opt.no_dirstruct = true;
1143 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1145 /* Empty value means reset the list of headers. */
1148 free_vec (opt.user_headers);
1149 opt.user_headers = NULL;
1153 if (!check_user_specified_header (val))
1155 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1156 exec_name, com, quote (val));
1159 opt.user_headers = vec_append (opt.user_headers, val);
1164 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1166 int flag = cmd_boolean (com, val, &opt.htmlify);
1167 if (flag && !opt.htmlify)
1168 opt.remove_listing = false;
1172 /* Set the "mirror" mode. It means: recursive download, timestamping,
1173 no limit on max. recursion depth, and don't remove listings. */
1176 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1180 if (!cmd_boolean (com, val, &mirror))
1184 opt.recursive = true;
1185 if (!opt.no_dirstruct)
1186 opt.dirstruct = true;
1187 opt.timestamping = true;
1188 opt.reclevel = INFINITE_RECURSION;
1189 opt.remove_listing = false;
1194 /* Validate --prefer-family and set the choice. Allowed values are
1195 "IPv4", "IPv6", and "none". */
1198 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1200 static const struct decode_item choices[] = {
1201 { "IPv4", prefer_ipv4 },
1202 { "IPv6", prefer_ipv6 },
1203 { "none", prefer_none },
1205 int prefer_family = prefer_none;
1206 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1208 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1209 opt.prefer_family = prefer_family;
1213 /* Set progress.type to VAL, but verify that it's a valid progress
1214 implementation before that. */
1217 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1219 if (!valid_progress_implementation_p (val))
1221 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1222 exec_name, com, quote (val));
1225 xfree_null (opt.progress_type);
1227 /* Don't call set_progress_implementation here. It will be called
1228 in main() when it becomes clear what the log output is. */
1229 opt.progress_type = xstrdup (val);
1233 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1234 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1238 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1240 if (!cmd_boolean (com, val, &opt.recursive))
1244 if (opt.recursive && !opt.no_dirstruct)
1245 opt.dirstruct = true;
1251 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1253 int restrict_os = opt.restrict_files_os;
1254 int restrict_ctrl = opt.restrict_files_ctrl;
1255 int restrict_case = opt.restrict_files_case;
1259 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1263 end = strchr (val, ',');
1265 end = val + strlen (val);
1267 if (VAL_IS ("unix"))
1268 restrict_os = restrict_unix;
1269 else if (VAL_IS ("windows"))
1270 restrict_os = restrict_windows;
1271 else if (VAL_IS ("lowercase"))
1272 restrict_case = restrict_lowercase;
1273 else if (VAL_IS ("uppercase"))
1274 restrict_case = restrict_uppercase;
1275 else if (VAL_IS ("nocontrol"))
1276 restrict_ctrl = false;
1280 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1281 exec_name, com, quote (val));
1288 while (*val && *end);
1292 opt.restrict_files_os = restrict_os;
1293 opt.restrict_files_ctrl = restrict_ctrl;
1294 opt.restrict_files_case = restrict_case;
1301 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1303 static const struct decode_item choices[] = {
1304 { "auto", secure_protocol_auto },
1305 { "sslv2", secure_protocol_sslv2 },
1306 { "sslv3", secure_protocol_sslv3 },
1307 { "tlsv1", secure_protocol_tlsv1 },
1309 int ok = decode_string (val, choices, countof (choices), place);
1311 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1316 /* Set all three timeout values. */
1319 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1322 if (!cmd_time (com, val, &value))
1324 opt.read_timeout = value;
1325 opt.connect_timeout = value;
1326 opt.dns_timeout = value;
1331 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1333 /* Disallow embedded newlines. */
1334 if (strchr (val, '\n'))
1336 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1337 exec_name, com, quote (val));
1340 xfree_null (opt.useragent);
1341 opt.useragent = xstrdup (val);
1345 /* The "verbose" option cannot be cmd_boolean because the variable is
1346 not bool -- it's of type int (-1 means uninitialized because of
1347 some random hackery for disallowing -q -v). */
1350 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1353 if (cmd_boolean (com, val, &flag))
1361 /* Miscellaneous useful routines. */
1363 /* A very simple atoi clone, more useful than atoi because it works on
1364 delimited strings, and has error reportage. Returns true on success,
1365 false on failure. If successful, stores result to *DEST. */
1368 simple_atoi (const char *beg, const char *end, int *dest)
1371 bool negative = false;
1372 const char *p = beg;
1374 while (p < end && c_isspace (*p))
1376 if (p < end && (*p == '-' || *p == '+'))
1378 negative = (*p == '-');
1384 /* Read negative numbers in a separate loop because the most
1385 negative integer cannot be represented as a positive number. */
1388 for (; p < end && c_isdigit (*p); p++)
1390 int next = (10 * result) + (*p - '0');
1392 return false; /* overflow */
1396 for (; p < end && c_isdigit (*p); p++)
1398 int next = (10 * result) - (*p - '0');
1400 return false; /* underflow */
1411 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1412 doesn't handle exponential notation. Returns true on success,
1413 false on failure. In case of success, stores its result to
1417 simple_atof (const char *beg, const char *end, double *dest)
1421 bool negative = false;
1422 bool seen_dot = false;
1423 bool seen_digit = false;
1426 const char *p = beg;
1428 while (p < end && c_isspace (*p))
1430 if (p < end && (*p == '-' || *p == '+'))
1432 negative = (*p == '-');
1436 for (; p < end; p++)
1442 result = (10 * result) + (ch - '0');
1444 result += (ch - '0') / (divider *= 10);
1466 /* Verify that the user-specified header in S is valid. It must
1467 contain a colon preceded by non-white-space characters and must not
1468 contain newlines. */
1471 check_user_specified_header (const char *s)
1475 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1477 /* The header MUST contain `:' preceded by at least one
1478 non-whitespace character. */
1479 if (*p != ':' || p == s)
1481 /* The header MUST NOT contain newlines. */
1482 if (strchr (s, '\n'))
1487 /* Decode VAL into a number, according to ITEMS. */
1490 decode_string (const char *val, const struct decode_item *items, int itemcount,
1494 for (i = 0; i < itemcount; i++)
1495 if (0 == strcasecmp (val, items[i].name))
1497 *place = items[i].code;
1504 void cleanup_html_url (void);
1507 /* Free the memory allocated by global variables. */
1511 /* Free external resources, close files, etc. */
1514 fclose (output_stream);
1515 /* No need to check for error because Wget flushes its output (and
1516 checks for errors) after any data arrives. */
1518 /* We're exiting anyway so there's no real need to call free()
1519 hundreds of times. Skipping the frees will make Wget exit
1522 However, when detecting leaks, it's crucial to free() everything
1523 because then you can find the real leaks, i.e. the allocated
1524 memory which grows with the size of the program. */
1530 cleanup_html_url ();
1535 extern acc_t *netrc_list;
1536 free_netrc (netrc_list);
1538 xfree_null (opt.lfilename);
1539 xfree_null (opt.dir_prefix);
1540 xfree_null (opt.input_filename);
1541 xfree_null (opt.output_document);
1542 free_vec (opt.accepts);
1543 free_vec (opt.rejects);
1544 free_vec (opt.excludes);
1545 free_vec (opt.includes);
1546 free_vec (opt.domains);
1547 free_vec (opt.follow_tags);
1548 free_vec (opt.ignore_tags);
1549 xfree_null (opt.progress_type);
1550 xfree_null (opt.ftp_user);
1551 xfree_null (opt.ftp_passwd);
1552 xfree_null (opt.ftp_proxy);
1553 xfree_null (opt.https_proxy);
1554 xfree_null (opt.http_proxy);
1555 free_vec (opt.no_proxy);
1556 xfree_null (opt.useragent);
1557 xfree_null (opt.referer);
1558 xfree_null (opt.http_user);
1559 xfree_null (opt.http_passwd);
1560 free_vec (opt.user_headers);
1562 xfree_null (opt.cert_file);
1563 xfree_null (opt.private_key);
1564 xfree_null (opt.ca_directory);
1565 xfree_null (opt.ca_cert);
1566 xfree_null (opt.random_file);
1567 xfree_null (opt.egd_file);
1569 xfree_null (opt.bind_address);
1570 xfree_null (opt.cookies_input);
1571 xfree_null (opt.cookies_output);
1572 xfree_null (opt.user);
1573 xfree_null (opt.passwd);
1574 xfree_null (opt.base_href);
1576 #endif /* DEBUG_MALLOC */
1579 /* Unit testing routines. */
1584 test_commands_sorted()
1586 int prev_idx = 0, next_idx = 1;
1587 int command_count = countof (commands) - 1;
1589 while (next_idx <= command_count)
1591 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1594 mu_assert ("FAILED", false);
1607 test_cmd_spec_restrict_file_names()
1612 int expected_restrict_files_os;
1613 int expected_restrict_files_ctrl;
1614 int expected_restrict_files_case;
1617 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1618 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1619 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1620 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1623 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1628 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1631 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1632 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1633 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1634 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1636 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1637 res == test_array[i].result
1638 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1639 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1640 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1646 #endif /* TESTING */