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 },
184 { "iri", &opt.enable_iri, cmd_boolean },
185 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186 { "limitrate", &opt.limit_rate, cmd_bytes },
187 { "loadcookies", &opt.cookies_input, cmd_file },
188 { "locale", &opt.locale, cmd_string },
189 { "logfile", &opt.lfilename, cmd_file },
190 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
191 { "maxredirect", &opt.max_redirect, cmd_number },
192 { "mirror", NULL, cmd_spec_mirror },
193 { "netrc", &opt.netrc, cmd_boolean },
194 { "noclobber", &opt.noclobber, cmd_boolean },
195 { "noparent", &opt.no_parent, cmd_boolean },
196 { "noproxy", &opt.no_proxy, cmd_vector },
197 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
198 { "outputdocument", &opt.output_document, cmd_file },
199 { "pagerequisites", &opt.page_requisites, cmd_boolean },
200 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
201 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
202 { "password", &opt.passwd, cmd_string },
203 { "postdata", &opt.post_data, cmd_string },
204 { "postfile", &opt.post_file_name, cmd_file },
205 { "preferfamily", NULL, cmd_spec_prefer_family },
206 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
208 { "privatekey", &opt.private_key, cmd_file },
209 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
211 { "progress", &opt.progress_type, cmd_spec_progress },
212 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
213 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
214 { "proxypassword", &opt.proxy_passwd, cmd_string },
215 { "proxyuser", &opt.proxy_user, cmd_string },
216 { "quiet", &opt.quiet, cmd_boolean },
217 { "quota", &opt.quota, cmd_bytes_sum },
219 { "randomfile", &opt.random_file, cmd_file },
221 { "randomwait", &opt.random_wait, cmd_boolean },
222 { "readtimeout", &opt.read_timeout, cmd_time },
223 { "reclevel", &opt.reclevel, cmd_number_inf },
224 { "recursive", NULL, cmd_spec_recursive },
225 { "referer", &opt.referer, cmd_string },
226 { "reject", &opt.rejects, cmd_vector },
227 { "relativeonly", &opt.relative_only, cmd_boolean },
228 { "remoteencoding", &opt.encoding_remote, cmd_string },
229 { "removelisting", &opt.remove_listing, cmd_boolean },
230 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
231 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
232 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
233 { "robots", &opt.use_robots, cmd_boolean },
234 { "savecookies", &opt.cookies_output, cmd_file },
235 { "saveheaders", &opt.save_headers, cmd_boolean },
237 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
239 { "serverresponse", &opt.server_response, cmd_boolean },
240 { "spanhosts", &opt.spanhost, cmd_boolean },
241 { "spider", &opt.spider, cmd_boolean },
242 { "strictcomments", &opt.strict_comments, cmd_boolean },
243 { "timeout", NULL, cmd_spec_timeout },
244 { "timestamping", &opt.timestamping, cmd_boolean },
245 { "tries", &opt.ntry, cmd_number_inf },
246 { "useproxy", &opt.use_proxy, cmd_boolean },
247 { "user", &opt.user, cmd_string },
248 { "useragent", NULL, cmd_spec_useragent },
249 { "verbose", NULL, cmd_spec_verbose },
250 { "wait", &opt.wait, cmd_time },
251 { "waitretry", &opt.waitretry, cmd_time },
253 { "wdebug", &opt.wdebug, cmd_boolean },
257 /* Look up CMDNAME in the commands[] and return its position in the
258 array. If CMDNAME is not found, return -1. */
261 command_by_name (const char *cmdname)
263 /* Use binary search for speed. Wget has ~100 commands, which
264 guarantees a worst case performance of 7 string comparisons. */
265 int lo = 0, hi = countof (commands) - 1;
269 int mid = (lo + hi) >> 1;
270 int cmp = strcasecmp (cmdname, commands[mid].name);
281 /* Reset the variables to default values. */
287 /* Most of the default values are 0 (and 0.0, NULL, and false).
288 Just reset everything, and fill in the non-zero values. Note
289 that initializing pointers to NULL this way is technically
290 illegal, but porting Wget to a machine where NULL is not all-zero
291 bit pattern will be the least of the implementors' worries. */
298 opt.add_hostdir = true;
302 opt.http_keep_alive = true;
303 opt.use_proxy = true;
304 tmp = getenv ("no_proxy");
306 opt.no_proxy = sepstring (tmp);
307 opt.prefer_family = prefer_none;
308 opt.allow_cache = true;
310 opt.read_timeout = 900;
311 opt.use_robots = true;
313 opt.remove_listing = true;
315 opt.dot_bytes = 1024;
316 opt.dot_spacing = 10;
317 opt.dots_in_line = 50;
319 opt.dns_cache = true;
323 opt.check_cert = true;
326 /* The default for file name restriction defaults to the OS type. */
327 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
328 opt.restrict_files_os = restrict_windows;
330 opt.restrict_files_os = restrict_unix;
332 opt.restrict_files_ctrl = true;
333 opt.restrict_files_case = restrict_no_case_restriction;
335 opt.max_redirect = 20;
338 /* Return the user's home directory (strdup-ed), or NULL if none is
343 char *home = getenv ("HOME");
348 /* Under MSDOS, if $HOME isn't defined, use the directory where
349 `wget.exe' resides. */
350 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
351 char *p, buf[PATH_MAX];
353 strcpy (buf, _w32_get_argv0 ());
354 p = strrchr (buf, '/'); /* djgpp */
356 p = strrchr (buf, '\\'); /* others */
360 #elif !defined(WINDOWS)
361 /* If HOME is not defined, try getting it from the password
363 struct passwd *pwd = getpwuid (getuid ());
364 if (!pwd || !pwd->pw_dir)
368 /* Under Windows, if $HOME isn't defined, use the directory where
369 `wget.exe' resides. */
374 return home ? xstrdup (home) : NULL;
377 /* Check the 'WGETRC' environment variable and return the file name
378 if 'WGETRC' is set and is a valid file.
379 If the `WGETRC' variable exists but the file does not exist, the
380 function will exit(). */
382 wgetrc_env_file_name (void)
384 char *env = getenv ("WGETRC");
387 if (!file_exists_p (env))
389 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
393 return xstrdup (env);
397 /* Check for the existance of '$HOME/.wgetrc' and return it's path
398 if it exists and is set. */
400 wgetrc_user_file_name (void)
402 char *home = home_dir();
405 file = aprintf ("%s/.wgetrc", home);
409 if (!file_exists_p (file))
416 /* Return the path to the user's .wgetrc. This is either the value of
417 `WGETRC' environment variable, or `$HOME/.wgetrc'.
419 Additionally, for windows, look in the directory where wget.exe
422 wgetrc_file_name (void)
424 char *file = wgetrc_env_file_name ();
428 file = wgetrc_user_file_name ();
431 /* Under Windows, if we still haven't found .wgetrc, look for the file
432 `wget.ini' in the directory where `wget.exe' resides; we do this for
433 backward compatibility with previous versions of Wget.
434 SYSTEM_WGETRC should not be defined under WINDOWS. */
435 if (!file || !file_exists_p (file))
441 file = aprintf ("%s/wget.ini", home);
447 if (!file_exists_p (file))
455 /* Return values of parse_line. */
463 static enum parse_line parse_line (const char *, char **, char **, int *);
464 static bool setval_internal (int, const char *, const char *);
466 /* Initialize variables from a wgetrc file. Returns zero (failure) if
467 there were errors in the file. */
470 run_wgetrc (const char *file)
477 fp = fopen (file, "rb");
480 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
481 file, strerror (errno));
482 return true; /* not a fatal error */
484 enable_tilde_expansion = true;
486 while ((line = read_whole_line (fp)) != NULL)
488 char *com = NULL, *val = NULL;
491 /* Parse the line. */
492 switch (parse_line (line, &com, &val, &comind))
495 /* If everything is OK, set the value. */
496 if (!setval_internal (comind, com, val))
498 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
499 exec_name, file, ln);
503 case line_syntax_error:
504 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
505 exec_name, file, ln);
508 case line_unknown_command:
509 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
510 exec_name, quote (com), file, ln);
523 enable_tilde_expansion = false;
529 /* Initialize the defaults and run the system wgetrc and user's own
537 /* Load the hard-coded defaults. */
540 /* If SYSTEM_WGETRC is defined, use it. */
542 if (file_exists_p (SYSTEM_WGETRC))
543 ok &= run_wgetrc (SYSTEM_WGETRC);
545 /* Override it with your own, if one exists. */
546 file = wgetrc_file_name ();
549 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
550 something like realpath() before comparing them with `strcmp' */
552 if (!strcmp (file, SYSTEM_WGETRC))
554 fprintf (stderr, _("\
555 %s: Warning: Both system and user wgetrc point to %s.\n"),
556 exec_name, quote (file));
560 ok &= run_wgetrc (file);
562 /* If there were errors processing either `.wgetrc', abort. */
570 /* Remove dashes and underscores from S, modifying S in the
576 char *t = s; /* t - tortoise */
577 char *h = s; /* h - hare */
579 if (*h == '_' || *h == '-')
586 /* Parse the line pointed by line, with the syntax:
587 <sp>* command <sp>* = <sp>* value <sp>*
588 Uses malloc to allocate space for command and value.
590 Returns one of line_ok, line_empty, line_syntax_error, or
591 line_unknown_command.
593 In case of line_ok, *COM and *VAL point to freshly allocated
594 strings, and *COMIND points to com's index. In case of error or
595 empty line, their values are unmodified. */
597 static enum parse_line
598 parse_line (const char *line, char **com, char **val, int *comind)
601 const char *end = line + strlen (line);
602 const char *cmdstart, *cmdend;
603 const char *valstart, *valend;
608 /* Skip leading and trailing whitespace. */
609 while (*line && c_isspace (*line))
611 while (end > line && c_isspace (end[-1]))
614 /* Skip empty lines and comments. */
615 if (!*line || *line == '#')
621 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
625 /* Skip '=', as well as any space before or after it. */
626 while (p < end && c_isspace (*p))
628 if (p == end || *p != '=')
629 return line_syntax_error;
631 while (p < end && c_isspace (*p))
637 /* The syntax is valid (even though the command might not be). Fill
638 in the command name and value. */
639 *com = strdupdelim (cmdstart, cmdend);
640 *val = strdupdelim (valstart, valend);
642 /* The line now known to be syntactically correct. Check whether
643 the command is valid. */
644 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
646 ind = command_by_name (cmdcopy);
648 return line_unknown_command;
650 /* Report success to the caller. */
655 /* Run commands[comind].action. */
658 setval_internal (int comind, const char *com, const char *val)
660 assert (0 <= comind && ((size_t) comind) < countof (commands));
661 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
662 return commands[comind].action (com, val, commands[comind].place);
665 /* Run command COM with value VAL. If running the command produces an
666 error, report the error and exit.
668 This is intended to be called from main() to modify Wget's behavior
669 through command-line switches. Since COM is hard-coded in main(),
670 it is not canonicalized, and this aborts when COM is not found.
672 If COMIND's are exported to init.h, this function will be changed
673 to accept COMIND directly. */
676 setoptval (const char *com, const char *val, const char *optname)
678 /* Prepend "--" to OPTNAME. */
679 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
682 strcpy (dd_optname + 2, optname);
684 assert (val != NULL);
685 if (!setval_internal (command_by_name (com), dd_optname, val))
689 /* Parse OPT into command and value and run it. For example,
690 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
691 This is used by the `--execute' flag in main.c. */
694 run_command (const char *opt)
698 switch (parse_line (opt, &com, &val, &comind))
701 if (!setval_internal (comind, com, val))
707 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
708 exec_name, quote (opt));
713 /* Generic helper functions, for use with `commands'. */
715 /* Forward declarations: */
720 static bool decode_string (const char *, const struct decode_item *, int, int *);
721 static bool simple_atoi (const char *, const char *, int *);
722 static bool simple_atof (const char *, const char *, double *);
724 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
726 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
727 && c_tolower((p)[1]) == (c1) \
730 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
731 && c_tolower((p)[1]) == (c1) \
732 && c_tolower((p)[2]) == (c2) \
736 /* Store the boolean value from VAL to PLACE. COM is ignored,
737 except for error messages. */
739 cmd_boolean (const char *com, const char *val, void *place)
743 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
744 /* "on", "yes" and "1" mean true. */
746 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
747 /* "off", "no" and "0" mean false. */
752 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
753 exec_name, com, quote (val));
757 *(bool *) place = value;
761 /* Set the non-negative integer value from VAL to PLACE. With
762 incorrect specification, the number remains unchanged. */
764 cmd_number (const char *com, const char *val, void *place)
766 if (!simple_atoi (val, val + strlen (val), place)
767 || *(int *) place < 0)
769 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
770 exec_name, com, quote (val));
776 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
778 cmd_number_inf (const char *com, const char *val, void *place)
780 if (!strcasecmp (val, "inf"))
785 return cmd_number (com, val, place);
788 /* Copy (strdup) the string at COM to a new location and place a
789 pointer to *PLACE. */
791 cmd_string (const char *com, const char *val, void *place)
793 char **pstring = (char **)place;
795 xfree_null (*pstring);
796 *pstring = xstrdup (val);
800 #if defined(WINDOWS) || defined(MSDOS)
801 # define ISSEP(c) ((c) == '/' || (c) == '\\')
803 # define ISSEP(c) ((c) == '/')
806 /* Like the above, but handles tilde-expansion when reading a user's
807 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
808 gets expanded to the user's home directory. */
810 cmd_file (const char *com, const char *val, void *place)
812 char **pstring = (char **)place;
814 xfree_null (*pstring);
816 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
818 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
821 *pstring = xstrdup (val);
826 char *home = home_dir ();
830 homelen = strlen (home);
831 while (homelen && ISSEP (home[homelen - 1]))
832 home[--homelen] = '\0';
834 /* Skip the leading "~/". */
835 for (++val; ISSEP (*val); val++)
838 *pstring = concat_strings (home, "/", val, (char *) 0);
841 #if defined(WINDOWS) || defined(MSDOS)
842 /* Convert "\" to "/". */
845 for (s = *pstring; *s; s++)
853 /* Like cmd_file, but strips trailing '/' characters. */
855 cmd_directory (const char *com, const char *val, void *place)
859 /* Call cmd_file() for tilde expansion and separator
860 canonicalization (backslash -> slash under Windows). These
861 things should perhaps be in a separate function. */
862 if (!cmd_file (com, val, place))
867 while (t > s && *--t == '/')
873 /* Split VAL by space to a vector of values, and append those values
874 to vector pointed to by the PLACE argument. If VAL is empty, the
875 PLACE vector is cleared instead. */
878 cmd_vector (const char *com, const char *val, void *place)
880 char ***pvec = (char ***)place;
883 *pvec = merge_vecs (*pvec, sepstring (val));
893 cmd_directory_vector (const char *com, const char *val, void *place)
895 char ***pvec = (char ***)place;
899 /* Strip the trailing slashes from directories. */
902 seps = sepstring (val);
903 for (t = seps; t && *t; t++)
905 int len = strlen (*t);
906 /* Skip degenerate case of root directory. */
909 if ((*t)[len - 1] == '/')
910 (*t)[len - 1] = '\0';
913 *pvec = merge_vecs (*pvec, seps);
923 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
924 "100k" or "2.5G" to a floating point number. */
927 parse_bytes_helper (const char *val, double *result)
930 const char *end = val + strlen (val);
932 /* Check for "inf". */
933 if (0 == strcmp (val, "inf"))
939 /* Strip trailing whitespace. */
940 while (val < end && c_isspace (end[-1]))
945 switch (c_tolower (end[-1]))
948 --end, mult = 1024.0;
951 --end, mult = 1048576.0;
954 --end, mult = 1073741824.0;
957 --end, mult = 1099511627776.0;
960 /* Not a recognized suffix: assume it's a digit. (If not,
961 simple_atof will raise an error.) */
965 /* Skip leading and trailing whitespace. */
966 while (val < end && c_isspace (*val))
968 while (val < end && c_isspace (end[-1]))
973 if (!simple_atof (val, end, &number) || number < 0)
976 *result = number * mult;
980 /* Parse VAL as a number and set its value to PLACE (which should
983 By default, the value is assumed to be in bytes. If "K", "M", or
984 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
985 1<<30, respectively. Floating point values are allowed and are
986 cast to integer before use. The idea is to be able to use things
987 like 1.5k instead of "1536".
989 The string "inf" is returned as 0.
991 In case of error, false is returned and memory pointed to by PLACE
992 remains unmodified. */
995 cmd_bytes (const char *com, const char *val, void *place)
998 if (!parse_bytes_helper (val, &byte_value))
1000 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1001 exec_name, com, quote (val));
1004 *(wgint *)place = (wgint)byte_value;
1008 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1009 SIZE_SUM. It works by converting the string to double, therefore
1010 working with values up to 2^53-1 without loss of precision. This
1011 value (8192 TB) is large enough to serve for a while. */
1014 cmd_bytes_sum (const char *com, const char *val, void *place)
1017 if (!parse_bytes_helper (val, &byte_value))
1019 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1020 exec_name, com, quote (val));
1023 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1027 /* Store the value of VAL to *OUT. The value is a time period, by
1028 default expressed in seconds, but also accepting suffixes "m", "h",
1029 "d", and "w" for minutes, hours, days, and weeks respectively. */
1032 cmd_time (const char *com, const char *val, void *place)
1034 double number, mult;
1035 const char *end = val + strlen (val);
1037 /* Strip trailing whitespace. */
1038 while (val < end && c_isspace (end[-1]))
1044 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1045 exec_name, com, quote (val));
1049 switch (c_tolower (end[-1]))
1052 --end, mult = 1; /* seconds */
1055 --end, mult = 60; /* minutes */
1058 --end, mult = 3600; /* hours */
1061 --end, mult = 86400.0; /* days */
1064 --end, mult = 604800.0; /* weeks */
1067 /* Not a recognized suffix: assume it belongs to the number.
1068 (If not, simple_atof will raise an error.) */
1072 /* Skip leading and trailing whitespace. */
1073 while (val < end && c_isspace (*val))
1075 while (val < end && c_isspace (end[-1]))
1080 if (!simple_atof (val, end, &number))
1083 *(double *)place = number * mult;
1089 cmd_cert_type (const char *com, const char *val, void *place)
1091 static const struct decode_item choices[] = {
1092 { "pem", keyfile_pem },
1093 { "der", keyfile_asn1 },
1094 { "asn1", keyfile_asn1 },
1096 int ok = decode_string (val, choices, countof (choices), place);
1098 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1103 /* Specialized helper functions, used by `commands' to handle some
1104 options specially. */
1106 static bool check_user_specified_header (const char *);
1109 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1111 if (!cmd_boolean (com, val, &opt.dirstruct))
1113 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1114 must be affected inversely. */
1116 opt.no_dirstruct = false;
1118 opt.no_dirstruct = true;
1123 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1125 /* Empty value means reset the list of headers. */
1128 free_vec (opt.user_headers);
1129 opt.user_headers = NULL;
1133 if (!check_user_specified_header (val))
1135 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1136 exec_name, com, quote (val));
1139 opt.user_headers = vec_append (opt.user_headers, val);
1144 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1146 int flag = cmd_boolean (com, val, &opt.htmlify);
1147 if (flag && !opt.htmlify)
1148 opt.remove_listing = false;
1152 /* Set the "mirror" mode. It means: recursive download, timestamping,
1153 no limit on max. recursion depth, and don't remove listings. */
1156 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1160 if (!cmd_boolean (com, val, &mirror))
1164 opt.recursive = true;
1165 if (!opt.no_dirstruct)
1166 opt.dirstruct = true;
1167 opt.timestamping = true;
1168 opt.reclevel = INFINITE_RECURSION;
1169 opt.remove_listing = false;
1174 /* Validate --prefer-family and set the choice. Allowed values are
1175 "IPv4", "IPv6", and "none". */
1178 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1180 static const struct decode_item choices[] = {
1181 { "IPv4", prefer_ipv4 },
1182 { "IPv6", prefer_ipv6 },
1183 { "none", prefer_none },
1185 int prefer_family = prefer_none;
1186 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1188 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1189 opt.prefer_family = prefer_family;
1193 /* Set progress.type to VAL, but verify that it's a valid progress
1194 implementation before that. */
1197 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1199 if (!valid_progress_implementation_p (val))
1201 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1202 exec_name, com, quote (val));
1205 xfree_null (opt.progress_type);
1207 /* Don't call set_progress_implementation here. It will be called
1208 in main() when it becomes clear what the log output is. */
1209 opt.progress_type = xstrdup (val);
1213 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1214 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1218 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1220 if (!cmd_boolean (com, val, &opt.recursive))
1224 if (opt.recursive && !opt.no_dirstruct)
1225 opt.dirstruct = true;
1231 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1233 int restrict_os = opt.restrict_files_os;
1234 int restrict_ctrl = opt.restrict_files_ctrl;
1235 int restrict_case = opt.restrict_files_case;
1239 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1243 end = strchr (val, ',');
1245 end = val + strlen (val);
1247 if (VAL_IS ("unix"))
1248 restrict_os = restrict_unix;
1249 else if (VAL_IS ("windows"))
1250 restrict_os = restrict_windows;
1251 else if (VAL_IS ("lowercase"))
1252 restrict_case = restrict_lowercase;
1253 else if (VAL_IS ("uppercase"))
1254 restrict_case = restrict_uppercase;
1255 else if (VAL_IS ("nocontrol"))
1256 restrict_ctrl = false;
1260 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1261 exec_name, com, quote (val));
1268 while (*val && *end);
1272 opt.restrict_files_os = restrict_os;
1273 opt.restrict_files_ctrl = restrict_ctrl;
1274 opt.restrict_files_case = restrict_case;
1281 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1283 static const struct decode_item choices[] = {
1284 { "auto", secure_protocol_auto },
1285 { "sslv2", secure_protocol_sslv2 },
1286 { "sslv3", secure_protocol_sslv3 },
1287 { "tlsv1", secure_protocol_tlsv1 },
1289 int ok = decode_string (val, choices, countof (choices), place);
1291 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1296 /* Set all three timeout values. */
1299 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1302 if (!cmd_time (com, val, &value))
1304 opt.read_timeout = value;
1305 opt.connect_timeout = value;
1306 opt.dns_timeout = value;
1311 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1313 /* Disallow embedded newlines. */
1314 if (strchr (val, '\n'))
1316 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1317 exec_name, com, quote (val));
1320 xfree_null (opt.useragent);
1321 opt.useragent = xstrdup (val);
1325 /* The "verbose" option cannot be cmd_boolean because the variable is
1326 not bool -- it's of type int (-1 means uninitialized because of
1327 some random hackery for disallowing -q -v). */
1330 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1333 if (cmd_boolean (com, val, &flag))
1341 /* Miscellaneous useful routines. */
1343 /* A very simple atoi clone, more useful than atoi because it works on
1344 delimited strings, and has error reportage. Returns true on success,
1345 false on failure. If successful, stores result to *DEST. */
1348 simple_atoi (const char *beg, const char *end, int *dest)
1351 bool negative = false;
1352 const char *p = beg;
1354 while (p < end && c_isspace (*p))
1356 if (p < end && (*p == '-' || *p == '+'))
1358 negative = (*p == '-');
1364 /* Read negative numbers in a separate loop because the most
1365 negative integer cannot be represented as a positive number. */
1368 for (; p < end && c_isdigit (*p); p++)
1370 int next = (10 * result) + (*p - '0');
1372 return false; /* overflow */
1376 for (; p < end && c_isdigit (*p); p++)
1378 int next = (10 * result) - (*p - '0');
1380 return false; /* underflow */
1391 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1392 doesn't handle exponential notation. Returns true on success,
1393 false on failure. In case of success, stores its result to
1397 simple_atof (const char *beg, const char *end, double *dest)
1401 bool negative = false;
1402 bool seen_dot = false;
1403 bool seen_digit = false;
1406 const char *p = beg;
1408 while (p < end && c_isspace (*p))
1410 if (p < end && (*p == '-' || *p == '+'))
1412 negative = (*p == '-');
1416 for (; p < end; p++)
1422 result = (10 * result) + (ch - '0');
1424 result += (ch - '0') / (divider *= 10);
1446 /* Verify that the user-specified header in S is valid. It must
1447 contain a colon preceded by non-white-space characters and must not
1448 contain newlines. */
1451 check_user_specified_header (const char *s)
1455 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1457 /* The header MUST contain `:' preceded by at least one
1458 non-whitespace character. */
1459 if (*p != ':' || p == s)
1461 /* The header MUST NOT contain newlines. */
1462 if (strchr (s, '\n'))
1467 /* Decode VAL into a number, according to ITEMS. */
1470 decode_string (const char *val, const struct decode_item *items, int itemcount,
1474 for (i = 0; i < itemcount; i++)
1475 if (0 == strcasecmp (val, items[i].name))
1477 *place = items[i].code;
1484 void cleanup_html_url (void);
1487 /* Free the memory allocated by global variables. */
1491 /* Free external resources, close files, etc. */
1494 fclose (output_stream);
1495 /* No need to check for error because Wget flushes its output (and
1496 checks for errors) after any data arrives. */
1498 /* We're exiting anyway so there's no real need to call free()
1499 hundreds of times. Skipping the frees will make Wget exit
1502 However, when detecting leaks, it's crucial to free() everything
1503 because then you can find the real leaks, i.e. the allocated
1504 memory which grows with the size of the program. */
1510 cleanup_html_url ();
1515 extern acc_t *netrc_list;
1516 free_netrc (netrc_list);
1518 xfree_null (opt.lfilename);
1519 xfree_null (opt.dir_prefix);
1520 xfree_null (opt.input_filename);
1521 xfree_null (opt.output_document);
1522 free_vec (opt.accepts);
1523 free_vec (opt.rejects);
1524 free_vec (opt.excludes);
1525 free_vec (opt.includes);
1526 free_vec (opt.domains);
1527 free_vec (opt.follow_tags);
1528 free_vec (opt.ignore_tags);
1529 xfree_null (opt.progress_type);
1530 xfree_null (opt.ftp_user);
1531 xfree_null (opt.ftp_passwd);
1532 xfree_null (opt.ftp_proxy);
1533 xfree_null (opt.https_proxy);
1534 xfree_null (opt.http_proxy);
1535 free_vec (opt.no_proxy);
1536 xfree_null (opt.useragent);
1537 xfree_null (opt.referer);
1538 xfree_null (opt.http_user);
1539 xfree_null (opt.http_passwd);
1540 free_vec (opt.user_headers);
1542 xfree_null (opt.cert_file);
1543 xfree_null (opt.private_key);
1544 xfree_null (opt.ca_directory);
1545 xfree_null (opt.ca_cert);
1546 xfree_null (opt.random_file);
1547 xfree_null (opt.egd_file);
1549 xfree_null (opt.bind_address);
1550 xfree_null (opt.cookies_input);
1551 xfree_null (opt.cookies_output);
1552 xfree_null (opt.user);
1553 xfree_null (opt.passwd);
1554 #endif /* DEBUG_MALLOC */
1557 /* Unit testing routines. */
1562 test_commands_sorted()
1564 int prev_idx = 0, next_idx = 1;
1565 int command_count = countof (commands) - 1;
1567 while (next_idx <= command_count)
1569 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1572 mu_assert ("FAILED", false);
1585 test_cmd_spec_restrict_file_names()
1590 int expected_restrict_files_os;
1591 int expected_restrict_files_ctrl;
1592 int expected_restrict_files_case;
1595 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1596 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1597 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1598 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1601 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1606 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1609 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1610 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1611 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1612 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1614 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1615 res == test_array[i].result
1616 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1617 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1618 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1624 #endif /* TESTING */