1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
61 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
62 otherwise, it will be performed by the shell. This variable will
63 be set by the wgetrc-reading function. */
65 static bool enable_tilde_expansion;
68 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
70 CMD_DECLARE (cmd_boolean);
71 CMD_DECLARE (cmd_bytes);
72 CMD_DECLARE (cmd_bytes_sum);
74 CMD_DECLARE (cmd_cert_type);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_number);
78 CMD_DECLARE (cmd_number_inf);
79 CMD_DECLARE (cmd_string);
80 CMD_DECLARE (cmd_file);
81 CMD_DECLARE (cmd_directory);
82 CMD_DECLARE (cmd_time);
83 CMD_DECLARE (cmd_vector);
85 CMD_DECLARE (cmd_spec_dirstruct);
86 CMD_DECLARE (cmd_spec_header);
87 CMD_DECLARE (cmd_spec_htmlify);
88 CMD_DECLARE (cmd_spec_mirror);
89 CMD_DECLARE (cmd_spec_prefer_family);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_secure_protocol);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
98 CMD_DECLARE (cmd_spec_verbose);
100 /* List of recognized commands, each consisting of name, place and
101 function. When adding a new command, simply add it to the list,
102 but be sure to keep the list sorted alphabetically, as
103 command_by_name's binary search depends on it. Also, be sure to
104 add any entries that allocate memory (e.g. cmd_string and
105 cmd_vector) to the cleanup() function below. */
107 static const struct {
110 bool (*action) (const char *, const char *, void *);
112 /* KEEP THIS LIST ALPHABETICALLY SORTED */
113 { "accept", &opt.accepts, cmd_vector },
114 { "addhostdir", &opt.add_hostdir, cmd_boolean },
115 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
116 { "askpassword", &opt.ask_passwd, cmd_boolean },
117 { "authnochallenge", &opt.auth_without_challenge,
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "contentdisposition", &opt.content_disposition, cmd_boolean },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "deleteafter", &opt.delete_after, cmd_boolean },
144 { "dirprefix", &opt.dir_prefix, cmd_directory },
145 { "dirstruct", NULL, cmd_spec_dirstruct },
146 { "dnscache", &opt.dns_cache, cmd_boolean },
147 { "dnstimeout", &opt.dns_timeout, cmd_time },
148 { "domains", &opt.domains, cmd_vector },
149 { "dotbytes", &opt.dot_bytes, cmd_bytes },
150 { "dotsinline", &opt.dots_in_line, cmd_number },
151 { "dotspacing", &opt.dot_spacing, cmd_number },
152 { "dotstyle", &opt.dot_style, cmd_string },
154 { "egdfile", &opt.egd_file, cmd_file },
156 { "excludedirectories", &opt.excludes, cmd_directory_vector },
157 { "excludedomains", &opt.exclude_domains, cmd_vector },
158 { "followftp", &opt.follow_ftp, cmd_boolean },
159 { "followtags", &opt.follow_tags, cmd_vector },
160 { "forcehtml", &opt.force_html, cmd_boolean },
161 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
162 { "ftppassword", &opt.ftp_passwd, cmd_string },
163 { "ftpproxy", &opt.ftp_proxy, cmd_string },
164 { "ftpuser", &opt.ftp_user, cmd_string },
165 { "glob", &opt.ftp_glob, cmd_boolean },
166 { "header", NULL, cmd_spec_header },
167 { "htmlextension", &opt.html_extension, cmd_boolean },
168 { "htmlify", NULL, cmd_spec_htmlify },
169 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
170 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
171 { "httppassword", &opt.http_passwd, cmd_string },
172 { "httpproxy", &opt.http_proxy, cmd_string },
173 { "httpsproxy", &opt.https_proxy, cmd_string },
174 { "httpuser", &opt.http_user, cmd_string },
175 { "ignorecase", &opt.ignore_case, cmd_boolean },
176 { "ignorelength", &opt.ignore_length, cmd_boolean },
177 { "ignoretags", &opt.ignore_tags, cmd_vector },
178 { "includedirectories", &opt.includes, cmd_directory_vector },
180 { "inet4only", &opt.ipv4_only, cmd_boolean },
181 { "inet6only", &opt.ipv6_only, cmd_boolean },
183 { "input", &opt.input_filename, cmd_file },
185 { "iri", &opt.enable_iri, cmd_boolean },
187 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
188 { "limitrate", &opt.limit_rate, cmd_bytes },
189 { "loadcookies", &opt.cookies_input, cmd_file },
191 { "locale", &opt.locale, cmd_string },
193 { "logfile", &opt.lfilename, cmd_file },
194 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
195 { "maxredirect", &opt.max_redirect, cmd_number },
196 { "mirror", NULL, cmd_spec_mirror },
197 { "netrc", &opt.netrc, cmd_boolean },
198 { "noclobber", &opt.noclobber, cmd_boolean },
199 { "noparent", &opt.no_parent, cmd_boolean },
200 { "noproxy", &opt.no_proxy, cmd_vector },
201 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
202 { "outputdocument", &opt.output_document, cmd_file },
203 { "pagerequisites", &opt.page_requisites, cmd_boolean },
204 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
205 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
206 { "password", &opt.passwd, cmd_string },
207 { "postdata", &opt.post_data, cmd_string },
208 { "postfile", &opt.post_file_name, cmd_file },
209 { "preferfamily", NULL, cmd_spec_prefer_family },
210 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
212 { "privatekey", &opt.private_key, cmd_file },
213 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
215 { "progress", &opt.progress_type, cmd_spec_progress },
216 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
217 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
218 { "proxypassword", &opt.proxy_passwd, cmd_string },
219 { "proxyuser", &opt.proxy_user, cmd_string },
220 { "quiet", &opt.quiet, cmd_boolean },
221 { "quota", &opt.quota, cmd_bytes_sum },
223 { "randomfile", &opt.random_file, cmd_file },
225 { "randomwait", &opt.random_wait, cmd_boolean },
226 { "readtimeout", &opt.read_timeout, cmd_time },
227 { "reclevel", &opt.reclevel, cmd_number_inf },
228 { "recursive", NULL, cmd_spec_recursive },
229 { "referer", &opt.referer, cmd_string },
230 { "reject", &opt.rejects, cmd_vector },
231 { "relativeonly", &opt.relative_only, cmd_boolean },
233 { "remoteencoding", &opt.encoding_remote, cmd_string },
235 { "removelisting", &opt.remove_listing, cmd_boolean },
236 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
237 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
238 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
239 { "robots", &opt.use_robots, cmd_boolean },
240 { "savecookies", &opt.cookies_output, cmd_file },
241 { "saveheaders", &opt.save_headers, cmd_boolean },
243 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
245 { "serverresponse", &opt.server_response, cmd_boolean },
246 { "spanhosts", &opt.spanhost, cmd_boolean },
247 { "spider", &opt.spider, cmd_boolean },
248 { "strictcomments", &opt.strict_comments, cmd_boolean },
249 { "timeout", NULL, cmd_spec_timeout },
250 { "timestamping", &opt.timestamping, cmd_boolean },
251 { "tries", &opt.ntry, cmd_number_inf },
252 { "useproxy", &opt.use_proxy, cmd_boolean },
253 { "user", &opt.user, cmd_string },
254 { "useragent", NULL, cmd_spec_useragent },
255 { "verbose", NULL, cmd_spec_verbose },
256 { "wait", &opt.wait, cmd_time },
257 { "waitretry", &opt.waitretry, cmd_time },
259 { "wdebug", &opt.wdebug, cmd_boolean },
263 /* Look up CMDNAME in the commands[] and return its position in the
264 array. If CMDNAME is not found, return -1. */
267 command_by_name (const char *cmdname)
269 /* Use binary search for speed. Wget has ~100 commands, which
270 guarantees a worst case performance of 7 string comparisons. */
271 int lo = 0, hi = countof (commands) - 1;
275 int mid = (lo + hi) >> 1;
276 int cmp = strcasecmp (cmdname, commands[mid].name);
287 /* Reset the variables to default values. */
293 /* Most of the default values are 0 (and 0.0, NULL, and false).
294 Just reset everything, and fill in the non-zero values. Note
295 that initializing pointers to NULL this way is technically
296 illegal, but porting Wget to a machine where NULL is not all-zero
297 bit pattern will be the least of the implementors' worries. */
304 opt.add_hostdir = true;
308 opt.http_keep_alive = true;
309 opt.use_proxy = true;
310 tmp = getenv ("no_proxy");
312 opt.no_proxy = sepstring (tmp);
313 opt.prefer_family = prefer_none;
314 opt.allow_cache = true;
316 opt.read_timeout = 900;
317 opt.use_robots = true;
319 opt.remove_listing = true;
321 opt.dot_bytes = 1024;
322 opt.dot_spacing = 10;
323 opt.dots_in_line = 50;
325 opt.dns_cache = true;
329 opt.check_cert = true;
332 /* The default for file name restriction defaults to the OS type. */
333 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
334 opt.restrict_files_os = restrict_windows;
336 opt.restrict_files_os = restrict_unix;
338 opt.restrict_files_ctrl = true;
339 opt.restrict_files_case = restrict_no_case_restriction;
341 opt.max_redirect = 20;
344 /* Return the user's home directory (strdup-ed), or NULL if none is
349 char *home = getenv ("HOME");
354 /* Under MSDOS, if $HOME isn't defined, use the directory where
355 `wget.exe' resides. */
356 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
357 char *p, buf[PATH_MAX];
359 strcpy (buf, _w32_get_argv0 ());
360 p = strrchr (buf, '/'); /* djgpp */
362 p = strrchr (buf, '\\'); /* others */
366 #elif !defined(WINDOWS)
367 /* If HOME is not defined, try getting it from the password
369 struct passwd *pwd = getpwuid (getuid ());
370 if (!pwd || !pwd->pw_dir)
374 /* Under Windows, if $HOME isn't defined, use the directory where
375 `wget.exe' resides. */
380 return home ? xstrdup (home) : NULL;
383 /* Return the path to the user's .wgetrc. This is either the value of
384 `WGETRC' environment variable, or `$HOME/.wgetrc'.
386 If the `WGETRC' variable exists but the file does not exist, the
387 function will exit(). */
389 wgetrc_file_name (void)
394 /* Try the environment. */
395 env = getenv ("WGETRC");
398 if (!file_exists_p (env))
400 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
404 return xstrdup (env);
407 /* If that failed, try $HOME/.wgetrc. */
410 file = aprintf ("%s/.wgetrc", home);
414 /* Under Windows, if we still haven't found .wgetrc, look for the file
415 `wget.ini' in the directory where `wget.exe' resides; we do this for
416 backward compatibility with previous versions of Wget.
417 SYSTEM_WGETRC should not be defined under WINDOWS. */
418 if (!file || !file_exists_p (file))
424 file = aprintf ("%s/wget.ini", home);
430 if (!file_exists_p (file))
438 /* Return values of parse_line. */
446 static enum parse_line parse_line (const char *, char **, char **, int *);
447 static bool setval_internal (int, const char *, const char *);
449 /* Initialize variables from a wgetrc file. Returns zero (failure) if
450 there were errors in the file. */
453 run_wgetrc (const char *file)
460 fp = fopen (file, "rb");
463 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
464 file, strerror (errno));
465 return true; /* not a fatal error */
467 enable_tilde_expansion = true;
469 while ((line = read_whole_line (fp)) != NULL)
471 char *com = NULL, *val = NULL;
474 /* Parse the line. */
475 switch (parse_line (line, &com, &val, &comind))
478 /* If everything is OK, set the value. */
479 if (!setval_internal (comind, com, val))
481 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
482 exec_name, file, ln);
486 case line_syntax_error:
487 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
488 exec_name, file, ln);
491 case line_unknown_command:
492 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
493 exec_name, quote (com), file, ln);
506 enable_tilde_expansion = false;
512 /* Initialize the defaults and run the system wgetrc and user's own
520 /* Load the hard-coded defaults. */
523 /* If SYSTEM_WGETRC is defined, use it. */
525 if (file_exists_p (SYSTEM_WGETRC))
526 ok &= run_wgetrc (SYSTEM_WGETRC);
528 /* Override it with your own, if one exists. */
529 file = wgetrc_file_name ();
532 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
533 something like realpath() before comparing them with `strcmp' */
535 if (!strcmp (file, SYSTEM_WGETRC))
537 fprintf (stderr, _("\
538 %s: Warning: Both system and user wgetrc point to %s.\n"),
539 exec_name, quote (file));
543 ok &= run_wgetrc (file);
545 /* If there were errors processing either `.wgetrc', abort. */
553 /* Remove dashes and underscores from S, modifying S in the
559 char *t = s; /* t - tortoise */
560 char *h = s; /* h - hare */
562 if (*h == '_' || *h == '-')
569 /* Parse the line pointed by line, with the syntax:
570 <sp>* command <sp>* = <sp>* value <sp>*
571 Uses malloc to allocate space for command and value.
573 Returns one of line_ok, line_empty, line_syntax_error, or
574 line_unknown_command.
576 In case of line_ok, *COM and *VAL point to freshly allocated
577 strings, and *COMIND points to com's index. In case of error or
578 empty line, their values are unmodified. */
580 static enum parse_line
581 parse_line (const char *line, char **com, char **val, int *comind)
584 const char *end = line + strlen (line);
585 const char *cmdstart, *cmdend;
586 const char *valstart, *valend;
591 /* Skip leading and trailing whitespace. */
592 while (*line && c_isspace (*line))
594 while (end > line && c_isspace (end[-1]))
597 /* Skip empty lines and comments. */
598 if (!*line || *line == '#')
604 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
608 /* Skip '=', as well as any space before or after it. */
609 while (p < end && c_isspace (*p))
611 if (p == end || *p != '=')
612 return line_syntax_error;
614 while (p < end && c_isspace (*p))
620 /* The syntax is valid (even though the command might not be). Fill
621 in the command name and value. */
622 *com = strdupdelim (cmdstart, cmdend);
623 *val = strdupdelim (valstart, valend);
625 /* The line now known to be syntactically correct. Check whether
626 the command is valid. */
627 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
629 ind = command_by_name (cmdcopy);
631 return line_unknown_command;
633 /* Report success to the caller. */
638 /* Run commands[comind].action. */
641 setval_internal (int comind, const char *com, const char *val)
643 assert (0 <= comind && comind < countof (commands));
644 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
645 return commands[comind].action (com, val, commands[comind].place);
648 /* Run command COM with value VAL. If running the command produces an
649 error, report the error and exit.
651 This is intended to be called from main() to modify Wget's behavior
652 through command-line switches. Since COM is hard-coded in main(),
653 it is not canonicalized, and this aborts when COM is not found.
655 If COMIND's are exported to init.h, this function will be changed
656 to accept COMIND directly. */
659 setoptval (const char *com, const char *val, const char *optname)
661 /* Prepend "--" to OPTNAME. */
662 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
665 strcpy (dd_optname + 2, optname);
667 assert (val != NULL);
668 if (!setval_internal (command_by_name (com), dd_optname, val))
672 /* Parse OPT into command and value and run it. For example,
673 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
674 This is used by the `--execute' flag in main.c. */
677 run_command (const char *opt)
681 switch (parse_line (opt, &com, &val, &comind))
684 if (!setval_internal (comind, com, val))
690 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
691 exec_name, quote (opt));
696 /* Generic helper functions, for use with `commands'. */
698 /* Forward declarations: */
703 static bool decode_string (const char *, const struct decode_item *, int, int *);
704 static bool simple_atoi (const char *, const char *, int *);
705 static bool simple_atof (const char *, const char *, double *);
707 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
709 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
710 && c_tolower((p)[1]) == (c1) \
713 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
714 && c_tolower((p)[1]) == (c1) \
715 && c_tolower((p)[2]) == (c2) \
719 /* Store the boolean value from VAL to PLACE. COM is ignored,
720 except for error messages. */
722 cmd_boolean (const char *com, const char *val, void *place)
726 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
727 /* "on", "yes" and "1" mean true. */
729 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
730 /* "off", "no" and "0" mean false. */
735 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
736 exec_name, com, quote (val));
740 *(bool *) place = value;
744 /* Set the non-negative integer value from VAL to PLACE. With
745 incorrect specification, the number remains unchanged. */
747 cmd_number (const char *com, const char *val, void *place)
749 if (!simple_atoi (val, val + strlen (val), place)
750 || *(int *) place < 0)
752 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
753 exec_name, com, quote (val));
759 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
761 cmd_number_inf (const char *com, const char *val, void *place)
763 if (!strcasecmp (val, "inf"))
768 return cmd_number (com, val, place);
771 /* Copy (strdup) the string at COM to a new location and place a
772 pointer to *PLACE. */
774 cmd_string (const char *com, const char *val, void *place)
776 char **pstring = (char **)place;
778 xfree_null (*pstring);
779 *pstring = xstrdup (val);
783 #if defined(WINDOWS) || defined(MSDOS)
784 # define ISSEP(c) ((c) == '/' || (c) == '\\')
786 # define ISSEP(c) ((c) == '/')
789 /* Like the above, but handles tilde-expansion when reading a user's
790 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
791 gets expanded to the user's home directory. */
793 cmd_file (const char *com, const char *val, void *place)
795 char **pstring = (char **)place;
797 xfree_null (*pstring);
799 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
801 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
804 *pstring = xstrdup (val);
809 char *home = home_dir ();
813 homelen = strlen (home);
814 while (homelen && ISSEP (home[homelen - 1]))
815 home[--homelen] = '\0';
817 /* Skip the leading "~/". */
818 for (++val; ISSEP (*val); val++)
821 *pstring = concat_strings (home, "/", val, (char *) 0);
824 #if defined(WINDOWS) || defined(MSDOS)
825 /* Convert "\" to "/". */
828 for (s = *pstring; *s; s++)
836 /* Like cmd_file, but strips trailing '/' characters. */
838 cmd_directory (const char *com, const char *val, void *place)
842 /* Call cmd_file() for tilde expansion and separator
843 canonicalization (backslash -> slash under Windows). These
844 things should perhaps be in a separate function. */
845 if (!cmd_file (com, val, place))
850 while (t > s && *--t == '/')
856 /* Split VAL by space to a vector of values, and append those values
857 to vector pointed to by the PLACE argument. If VAL is empty, the
858 PLACE vector is cleared instead. */
861 cmd_vector (const char *com, const char *val, void *place)
863 char ***pvec = (char ***)place;
866 *pvec = merge_vecs (*pvec, sepstring (val));
876 cmd_directory_vector (const char *com, const char *val, void *place)
878 char ***pvec = (char ***)place;
882 /* Strip the trailing slashes from directories. */
885 seps = sepstring (val);
886 for (t = seps; t && *t; t++)
888 int len = strlen (*t);
889 /* Skip degenerate case of root directory. */
892 if ((*t)[len - 1] == '/')
893 (*t)[len - 1] = '\0';
896 *pvec = merge_vecs (*pvec, seps);
906 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
907 "100k" or "2.5G" to a floating point number. */
910 parse_bytes_helper (const char *val, double *result)
913 const char *end = val + strlen (val);
915 /* Check for "inf". */
916 if (0 == strcmp (val, "inf"))
922 /* Strip trailing whitespace. */
923 while (val < end && c_isspace (end[-1]))
928 switch (c_tolower (end[-1]))
931 --end, mult = 1024.0;
934 --end, mult = 1048576.0;
937 --end, mult = 1073741824.0;
940 --end, mult = 1099511627776.0;
943 /* Not a recognized suffix: assume it's a digit. (If not,
944 simple_atof will raise an error.) */
948 /* Skip leading and trailing whitespace. */
949 while (val < end && c_isspace (*val))
951 while (val < end && c_isspace (end[-1]))
956 if (!simple_atof (val, end, &number) || number < 0)
959 *result = number * mult;
963 /* Parse VAL as a number and set its value to PLACE (which should
966 By default, the value is assumed to be in bytes. If "K", "M", or
967 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
968 1<<30, respectively. Floating point values are allowed and are
969 cast to integer before use. The idea is to be able to use things
970 like 1.5k instead of "1536".
972 The string "inf" is returned as 0.
974 In case of error, false is returned and memory pointed to by PLACE
975 remains unmodified. */
978 cmd_bytes (const char *com, const char *val, void *place)
981 if (!parse_bytes_helper (val, &byte_value))
983 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
984 exec_name, com, quote (val));
987 *(wgint *)place = (wgint)byte_value;
991 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
992 SIZE_SUM. It works by converting the string to double, therefore
993 working with values up to 2^53-1 without loss of precision. This
994 value (8192 TB) is large enough to serve for a while. */
997 cmd_bytes_sum (const char *com, const char *val, void *place)
1000 if (!parse_bytes_helper (val, &byte_value))
1002 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1003 exec_name, com, quote (val));
1006 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1010 /* Store the value of VAL to *OUT. The value is a time period, by
1011 default expressed in seconds, but also accepting suffixes "m", "h",
1012 "d", and "w" for minutes, hours, days, and weeks respectively. */
1015 cmd_time (const char *com, const char *val, void *place)
1017 double number, mult;
1018 const char *end = val + strlen (val);
1020 /* Strip trailing whitespace. */
1021 while (val < end && c_isspace (end[-1]))
1027 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1028 exec_name, com, quote (val));
1032 switch (c_tolower (end[-1]))
1035 --end, mult = 1; /* seconds */
1038 --end, mult = 60; /* minutes */
1041 --end, mult = 3600; /* hours */
1044 --end, mult = 86400.0; /* days */
1047 --end, mult = 604800.0; /* weeks */
1050 /* Not a recognized suffix: assume it belongs to the number.
1051 (If not, simple_atof will raise an error.) */
1055 /* Skip leading and trailing whitespace. */
1056 while (val < end && c_isspace (*val))
1058 while (val < end && c_isspace (end[-1]))
1063 if (!simple_atof (val, end, &number))
1066 *(double *)place = number * mult;
1072 cmd_cert_type (const char *com, const char *val, void *place)
1074 static const struct decode_item choices[] = {
1075 { "pem", keyfile_pem },
1076 { "der", keyfile_asn1 },
1077 { "asn1", keyfile_asn1 },
1079 int ok = decode_string (val, choices, countof (choices), place);
1081 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1086 /* Specialized helper functions, used by `commands' to handle some
1087 options specially. */
1089 static bool check_user_specified_header (const char *);
1092 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1094 if (!cmd_boolean (com, val, &opt.dirstruct))
1096 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1097 must be affected inversely. */
1099 opt.no_dirstruct = false;
1101 opt.no_dirstruct = true;
1106 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1108 /* Empty value means reset the list of headers. */
1111 free_vec (opt.user_headers);
1112 opt.user_headers = NULL;
1116 if (!check_user_specified_header (val))
1118 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1119 exec_name, com, quote (val));
1122 opt.user_headers = vec_append (opt.user_headers, val);
1127 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1129 int flag = cmd_boolean (com, val, &opt.htmlify);
1130 if (flag && !opt.htmlify)
1131 opt.remove_listing = false;
1135 /* Set the "mirror" mode. It means: recursive download, timestamping,
1136 no limit on max. recursion depth, and don't remove listings. */
1139 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1143 if (!cmd_boolean (com, val, &mirror))
1147 opt.recursive = true;
1148 if (!opt.no_dirstruct)
1149 opt.dirstruct = true;
1150 opt.timestamping = true;
1151 opt.reclevel = INFINITE_RECURSION;
1152 opt.remove_listing = false;
1157 /* Validate --prefer-family and set the choice. Allowed values are
1158 "IPv4", "IPv6", and "none". */
1161 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1163 static const struct decode_item choices[] = {
1164 { "IPv4", prefer_ipv4 },
1165 { "IPv6", prefer_ipv6 },
1166 { "none", prefer_none },
1168 int prefer_family = prefer_none;
1169 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1171 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1172 opt.prefer_family = prefer_family;
1176 /* Set progress.type to VAL, but verify that it's a valid progress
1177 implementation before that. */
1180 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1182 if (!valid_progress_implementation_p (val))
1184 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1185 exec_name, com, quote (val));
1188 xfree_null (opt.progress_type);
1190 /* Don't call set_progress_implementation here. It will be called
1191 in main() when it becomes clear what the log output is. */
1192 opt.progress_type = xstrdup (val);
1196 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1197 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1201 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1203 if (!cmd_boolean (com, val, &opt.recursive))
1207 if (opt.recursive && !opt.no_dirstruct)
1208 opt.dirstruct = true;
1214 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1216 int restrict_os = opt.restrict_files_os;
1217 int restrict_ctrl = opt.restrict_files_ctrl;
1218 int restrict_case = opt.restrict_files_case;
1222 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1226 end = strchr (val, ',');
1228 end = val + strlen (val);
1230 if (VAL_IS ("unix"))
1231 restrict_os = restrict_unix;
1232 else if (VAL_IS ("windows"))
1233 restrict_os = restrict_windows;
1234 else if (VAL_IS ("lowercase"))
1235 restrict_case = restrict_lowercase;
1236 else if (VAL_IS ("uppercase"))
1237 restrict_case = restrict_uppercase;
1238 else if (VAL_IS ("nocontrol"))
1239 restrict_ctrl = false;
1243 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1244 exec_name, com, quote (val));
1251 while (*val && *end);
1255 opt.restrict_files_os = restrict_os;
1256 opt.restrict_files_ctrl = restrict_ctrl;
1257 opt.restrict_files_case = restrict_case;
1264 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1266 static const struct decode_item choices[] = {
1267 { "auto", secure_protocol_auto },
1268 { "sslv2", secure_protocol_sslv2 },
1269 { "sslv3", secure_protocol_sslv3 },
1270 { "tlsv1", secure_protocol_tlsv1 },
1272 int ok = decode_string (val, choices, countof (choices), place);
1274 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1279 /* Set all three timeout values. */
1282 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1285 if (!cmd_time (com, val, &value))
1287 opt.read_timeout = value;
1288 opt.connect_timeout = value;
1289 opt.dns_timeout = value;
1294 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1296 /* Disallow embedded newlines. */
1297 if (strchr (val, '\n'))
1299 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1300 exec_name, com, quote (val));
1303 xfree_null (opt.useragent);
1304 opt.useragent = xstrdup (val);
1308 /* The "verbose" option cannot be cmd_boolean because the variable is
1309 not bool -- it's of type int (-1 means uninitialized because of
1310 some random hackery for disallowing -q -v). */
1313 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1316 if (cmd_boolean (com, val, &flag))
1324 /* Miscellaneous useful routines. */
1326 /* A very simple atoi clone, more useful than atoi because it works on
1327 delimited strings, and has error reportage. Returns true on success,
1328 false on failure. If successful, stores result to *DEST. */
1331 simple_atoi (const char *beg, const char *end, int *dest)
1334 bool negative = false;
1335 const char *p = beg;
1337 while (p < end && c_isspace (*p))
1339 if (p < end && (*p == '-' || *p == '+'))
1341 negative = (*p == '-');
1347 /* Read negative numbers in a separate loop because the most
1348 negative integer cannot be represented as a positive number. */
1351 for (; p < end && c_isdigit (*p); p++)
1353 int next = (10 * result) + (*p - '0');
1355 return false; /* overflow */
1359 for (; p < end && c_isdigit (*p); p++)
1361 int next = (10 * result) - (*p - '0');
1363 return false; /* underflow */
1374 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1375 doesn't handle exponential notation. Returns true on success,
1376 false on failure. In case of success, stores its result to
1380 simple_atof (const char *beg, const char *end, double *dest)
1384 bool negative = false;
1385 bool seen_dot = false;
1386 bool seen_digit = false;
1389 const char *p = beg;
1391 while (p < end && c_isspace (*p))
1393 if (p < end && (*p == '-' || *p == '+'))
1395 negative = (*p == '-');
1399 for (; p < end; p++)
1405 result = (10 * result) + (ch - '0');
1407 result += (ch - '0') / (divider *= 10);
1429 /* Verify that the user-specified header in S is valid. It must
1430 contain a colon preceded by non-white-space characters and must not
1431 contain newlines. */
1434 check_user_specified_header (const char *s)
1438 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1440 /* The header MUST contain `:' preceded by at least one
1441 non-whitespace character. */
1442 if (*p != ':' || p == s)
1444 /* The header MUST NOT contain newlines. */
1445 if (strchr (s, '\n'))
1450 /* Decode VAL into a number, according to ITEMS. */
1453 decode_string (const char *val, const struct decode_item *items, int itemcount,
1457 for (i = 0; i < itemcount; i++)
1458 if (0 == strcasecmp (val, items[i].name))
1460 *place = items[i].code;
1467 void cleanup_html_url (void);
1470 /* Free the memory allocated by global variables. */
1474 /* Free external resources, close files, etc. */
1477 fclose (output_stream);
1478 /* No need to check for error because Wget flushes its output (and
1479 checks for errors) after any data arrives. */
1481 /* We're exiting anyway so there's no real need to call free()
1482 hundreds of times. Skipping the frees will make Wget exit
1485 However, when detecting leaks, it's crucial to free() everything
1486 because then you can find the real leaks, i.e. the allocated
1487 memory which grows with the size of the program. */
1493 cleanup_html_url ();
1498 extern acc_t *netrc_list;
1499 free_netrc (netrc_list);
1501 xfree_null (opt.lfilename);
1502 xfree_null (opt.dir_prefix);
1503 xfree_null (opt.input_filename);
1504 xfree_null (opt.output_document);
1505 free_vec (opt.accepts);
1506 free_vec (opt.rejects);
1507 free_vec (opt.excludes);
1508 free_vec (opt.includes);
1509 free_vec (opt.domains);
1510 free_vec (opt.follow_tags);
1511 free_vec (opt.ignore_tags);
1512 xfree_null (opt.progress_type);
1513 xfree_null (opt.ftp_user);
1514 xfree_null (opt.ftp_passwd);
1515 xfree_null (opt.ftp_proxy);
1516 xfree_null (opt.https_proxy);
1517 xfree_null (opt.http_proxy);
1518 free_vec (opt.no_proxy);
1519 xfree_null (opt.useragent);
1520 xfree_null (opt.referer);
1521 xfree_null (opt.http_user);
1522 xfree_null (opt.http_passwd);
1523 free_vec (opt.user_headers);
1525 xfree_null (opt.cert_file);
1526 xfree_null (opt.private_key);
1527 xfree_null (opt.ca_directory);
1528 xfree_null (opt.ca_cert);
1529 xfree_null (opt.random_file);
1530 xfree_null (opt.egd_file);
1532 xfree_null (opt.bind_address);
1533 xfree_null (opt.cookies_input);
1534 xfree_null (opt.cookies_output);
1535 xfree_null (opt.user);
1536 xfree_null (opt.passwd);
1537 #endif /* DEBUG_MALLOC */
1540 /* Unit testing routines. */
1545 test_commands_sorted()
1547 int prev_idx = 0, next_idx = 1;
1548 int command_count = countof (commands) - 1;
1550 while (next_idx <= command_count)
1552 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1555 mu_assert ("FAILED", false);
1568 test_cmd_spec_restrict_file_names()
1573 int expected_restrict_files_os;
1574 int expected_restrict_files_ctrl;
1575 int expected_restrict_files_case;
1578 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1579 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1580 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1581 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1584 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1589 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1592 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1593 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1594 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1595 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1597 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1598 res == test_array[i].result
1599 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1600 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1601 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1607 #endif /* TESTING */