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 { "authnochallenge", &opt.auth_without_challenge,
118 { "background", &opt.background, cmd_boolean },
119 { "backupconverted", &opt.backup_converted, cmd_boolean },
120 { "backups", &opt.backups, cmd_number },
121 { "base", &opt.base_href, cmd_string },
122 { "bindaddress", &opt.bind_address, cmd_string },
124 { "cacertificate", &opt.ca_cert, cmd_file },
126 { "cache", &opt.allow_cache, cmd_boolean },
128 { "cadirectory", &opt.ca_directory, cmd_directory },
129 { "certificate", &opt.cert_file, cmd_file },
130 { "certificatetype", &opt.cert_type, cmd_cert_type },
131 { "checkcertificate", &opt.check_cert, cmd_boolean },
133 { "connecttimeout", &opt.connect_timeout, cmd_time },
134 { "contentdisposition", &opt.content_disposition, cmd_boolean },
135 { "continue", &opt.always_rest, cmd_boolean },
136 { "convertlinks", &opt.convert_links, cmd_boolean },
137 { "cookies", &opt.cookies, cmd_boolean },
138 { "cutdirs", &opt.cut_dirs, cmd_number },
140 { "debug", &opt.debug, cmd_boolean },
142 { "deleteafter", &opt.delete_after, cmd_boolean },
143 { "dirprefix", &opt.dir_prefix, cmd_directory },
144 { "dirstruct", NULL, cmd_spec_dirstruct },
145 { "dnscache", &opt.dns_cache, cmd_boolean },
146 { "dnstimeout", &opt.dns_timeout, cmd_time },
147 { "domains", &opt.domains, cmd_vector },
148 { "dotbytes", &opt.dot_bytes, cmd_bytes },
149 { "dotsinline", &opt.dots_in_line, cmd_number },
150 { "dotspacing", &opt.dot_spacing, cmd_number },
151 { "dotstyle", &opt.dot_style, cmd_string },
153 { "egdfile", &opt.egd_file, cmd_file },
155 { "excludedirectories", &opt.excludes, cmd_directory_vector },
156 { "excludedomains", &opt.exclude_domains, cmd_vector },
157 { "followftp", &opt.follow_ftp, cmd_boolean },
158 { "followtags", &opt.follow_tags, cmd_vector },
159 { "forcehtml", &opt.force_html, cmd_boolean },
160 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
161 { "ftppassword", &opt.ftp_passwd, cmd_string },
162 { "ftpproxy", &opt.ftp_proxy, cmd_string },
163 { "ftpuser", &opt.ftp_user, cmd_string },
164 { "glob", &opt.ftp_glob, cmd_boolean },
165 { "header", NULL, cmd_spec_header },
166 { "htmlextension", &opt.html_extension, cmd_boolean },
167 { "htmlify", NULL, cmd_spec_htmlify },
168 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
169 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
170 { "httppassword", &opt.http_passwd, cmd_string },
171 { "httpproxy", &opt.http_proxy, cmd_string },
172 { "httpsproxy", &opt.https_proxy, cmd_string },
173 { "httpuser", &opt.http_user, cmd_string },
174 { "ignorecase", &opt.ignore_case, cmd_boolean },
175 { "ignorelength", &opt.ignore_length, cmd_boolean },
176 { "ignoretags", &opt.ignore_tags, cmd_vector },
177 { "includedirectories", &opt.includes, cmd_directory_vector },
179 { "inet4only", &opt.ipv4_only, cmd_boolean },
180 { "inet6only", &opt.ipv6_only, cmd_boolean },
182 { "input", &opt.input_filename, cmd_file },
183 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
184 { "limitrate", &opt.limit_rate, cmd_bytes },
185 { "loadcookies", &opt.cookies_input, cmd_file },
186 { "logfile", &opt.lfilename, cmd_file },
187 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
188 { "maxredirect", &opt.max_redirect, cmd_number },
189 { "mirror", NULL, cmd_spec_mirror },
190 { "netrc", &opt.netrc, cmd_boolean },
191 { "noclobber", &opt.noclobber, cmd_boolean },
192 { "noparent", &opt.no_parent, cmd_boolean },
193 { "noproxy", &opt.no_proxy, cmd_vector },
194 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
195 { "outputdocument", &opt.output_document, cmd_file },
196 { "pagerequisites", &opt.page_requisites, cmd_boolean },
197 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
198 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
199 { "password", &opt.passwd, cmd_string },
200 { "postdata", &opt.post_data, cmd_string },
201 { "postfile", &opt.post_file_name, cmd_file },
202 { "preferfamily", NULL, cmd_spec_prefer_family },
203 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
205 { "privatekey", &opt.private_key, cmd_file },
206 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
208 { "progress", &opt.progress_type, cmd_spec_progress },
209 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
210 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
211 { "proxypassword", &opt.proxy_passwd, cmd_string },
212 { "proxyuser", &opt.proxy_user, cmd_string },
213 { "quiet", &opt.quiet, cmd_boolean },
214 { "quota", &opt.quota, cmd_bytes_sum },
216 { "randomfile", &opt.random_file, cmd_file },
218 { "randomwait", &opt.random_wait, cmd_boolean },
219 { "readtimeout", &opt.read_timeout, cmd_time },
220 { "reclevel", &opt.reclevel, cmd_number_inf },
221 { "recursive", NULL, cmd_spec_recursive },
222 { "referer", &opt.referer, cmd_string },
223 { "reject", &opt.rejects, cmd_vector },
224 { "relativeonly", &opt.relative_only, cmd_boolean },
225 { "removelisting", &opt.remove_listing, cmd_boolean },
226 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
227 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
228 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
229 { "robots", &opt.use_robots, cmd_boolean },
230 { "savecookies", &opt.cookies_output, cmd_file },
231 { "saveheaders", &opt.save_headers, cmd_boolean },
233 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
235 { "serverresponse", &opt.server_response, cmd_boolean },
236 { "spanhosts", &opt.spanhost, cmd_boolean },
237 { "spider", &opt.spider, cmd_boolean },
238 { "strictcomments", &opt.strict_comments, cmd_boolean },
239 { "timeout", NULL, cmd_spec_timeout },
240 { "timestamping", &opt.timestamping, cmd_boolean },
241 { "tries", &opt.ntry, cmd_number_inf },
242 { "useproxy", &opt.use_proxy, cmd_boolean },
243 { "user", &opt.user, cmd_string },
244 { "useragent", NULL, cmd_spec_useragent },
245 { "verbose", NULL, cmd_spec_verbose },
246 { "wait", &opt.wait, cmd_time },
247 { "waitretry", &opt.waitretry, cmd_time },
249 { "wdebug", &opt.wdebug, cmd_boolean },
253 /* Look up CMDNAME in the commands[] and return its position in the
254 array. If CMDNAME is not found, return -1. */
257 command_by_name (const char *cmdname)
259 /* Use binary search for speed. Wget has ~100 commands, which
260 guarantees a worst case performance of 7 string comparisons. */
261 int lo = 0, hi = countof (commands) - 1;
265 int mid = (lo + hi) >> 1;
266 int cmp = strcasecmp (cmdname, commands[mid].name);
277 /* Reset the variables to default values. */
283 /* Most of the default values are 0 (and 0.0, NULL, and false).
284 Just reset everything, and fill in the non-zero values. Note
285 that initializing pointers to NULL this way is technically
286 illegal, but porting Wget to a machine where NULL is not all-zero
287 bit pattern will be the least of the implementors' worries. */
294 opt.add_hostdir = true;
298 opt.http_keep_alive = true;
299 opt.use_proxy = true;
300 tmp = getenv ("no_proxy");
302 opt.no_proxy = sepstring (tmp);
303 opt.allow_cache = true;
305 opt.read_timeout = 900;
306 opt.use_robots = true;
308 opt.remove_listing = true;
310 opt.dot_bytes = 1024;
311 opt.dot_spacing = 10;
312 opt.dots_in_line = 50;
314 opt.dns_cache = true;
318 opt.check_cert = true;
321 /* The default for file name restriction defaults to the OS type. */
322 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
323 opt.restrict_files_os = restrict_windows;
325 opt.restrict_files_os = restrict_unix;
327 opt.restrict_files_ctrl = true;
328 opt.restrict_files_case = restrict_no_case_restriction;
330 opt.max_redirect = 20;
333 /* Return the user's home directory (strdup-ed), or NULL if none is
338 char *home = getenv ("HOME");
343 /* Under MSDOS, if $HOME isn't defined, use the directory where
344 `wget.exe' resides. */
345 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
346 char *p, buf[PATH_MAX];
348 strcpy (buf, _w32_get_argv0 ());
349 p = strrchr (buf, '/'); /* djgpp */
351 p = strrchr (buf, '\\'); /* others */
355 #elif !defined(WINDOWS)
356 /* If HOME is not defined, try getting it from the password
358 struct passwd *pwd = getpwuid (getuid ());
359 if (!pwd || !pwd->pw_dir)
363 /* Under Windows, if $HOME isn't defined, use the directory where
364 `wget.exe' resides. */
369 return home ? xstrdup (home) : NULL;
372 /* Check the 'WGETRC' environment variable and return the file name
373 if 'WGETRC' is set and is a valid file.
374 If the `WGETRC' variable exists but the file does not exist, the
375 function will exit(). */
377 wgetrc_env_file_name (void)
379 char *env = getenv ("WGETRC");
382 if (!file_exists_p (env))
384 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
388 return xstrdup (env);
392 /* Check for the existance of '$HOME/.wgetrc' and return it's path
393 if it exists and is set. */
395 wgetrc_user_file_name (void)
397 char *home = home_dir();
400 file = aprintf ("%s/.wgetrc", home);
404 if (!file_exists_p (file))
411 /* Return the path to the user's .wgetrc. This is either the value of
412 `WGETRC' environment variable, or `$HOME/.wgetrc'.
414 Additionally, for windows, look in the directory where wget.exe
417 wgetrc_file_name (void)
419 char *file = wgetrc_env_file_name ();
423 file = wgetrc_user_file_name ();
426 /* Under Windows, if we still haven't found .wgetrc, look for the file
427 `wget.ini' in the directory where `wget.exe' resides; we do this for
428 backward compatibility with previous versions of Wget.
429 SYSTEM_WGETRC should not be defined under WINDOWS. */
430 if (!file || !file_exists_p (file))
436 file = aprintf ("%s/wget.ini", home);
442 if (!file_exists_p (file))
450 /* Return values of parse_line. */
458 static enum parse_line parse_line (const char *, char **, char **, int *);
459 static bool setval_internal (int, const char *, const char *);
461 /* Initialize variables from a wgetrc file. Returns zero (failure) if
462 there were errors in the file. */
465 run_wgetrc (const char *file)
472 fp = fopen (file, "rb");
475 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
476 file, strerror (errno));
477 return true; /* not a fatal error */
479 enable_tilde_expansion = true;
481 while ((line = read_whole_line (fp)) != NULL)
483 char *com = NULL, *val = NULL;
486 /* Parse the line. */
487 switch (parse_line (line, &com, &val, &comind))
490 /* If everything is OK, set the value. */
491 if (!setval_internal (comind, com, val))
493 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
494 exec_name, file, ln);
498 case line_syntax_error:
499 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
500 exec_name, file, ln);
503 case line_unknown_command:
504 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
505 exec_name, com, file, ln);
518 enable_tilde_expansion = false;
524 /* Initialize the defaults and run the system wgetrc and user's own
532 /* Load the hard-coded defaults. */
535 /* If SYSTEM_WGETRC is defined, use it. */
537 if (file_exists_p (SYSTEM_WGETRC))
538 ok &= run_wgetrc (SYSTEM_WGETRC);
540 /* Override it with your own, if one exists. */
541 file = wgetrc_file_name ();
544 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
545 something like realpath() before comparing them with `strcmp' */
547 if (!strcmp (file, SYSTEM_WGETRC))
549 fprintf (stderr, _("\
550 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
555 ok &= run_wgetrc (file);
557 /* If there were errors processing either `.wgetrc', abort. */
565 /* Remove dashes and underscores from S, modifying S in the
571 char *t = s; /* t - tortoise */
572 char *h = s; /* h - hare */
574 if (*h == '_' || *h == '-')
581 /* Parse the line pointed by line, with the syntax:
582 <sp>* command <sp>* = <sp>* value <sp>*
583 Uses malloc to allocate space for command and value.
585 Returns one of line_ok, line_empty, line_syntax_error, or
586 line_unknown_command.
588 In case of line_ok, *COM and *VAL point to freshly allocated
589 strings, and *COMIND points to com's index. In case of error or
590 empty line, their values are unmodified. */
592 static enum parse_line
593 parse_line (const char *line, char **com, char **val, int *comind)
596 const char *end = line + strlen (line);
597 const char *cmdstart, *cmdend;
598 const char *valstart, *valend;
603 /* Skip leading and trailing whitespace. */
604 while (*line && c_isspace (*line))
606 while (end > line && c_isspace (end[-1]))
609 /* Skip empty lines and comments. */
610 if (!*line || *line == '#')
616 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
620 /* Skip '=', as well as any space before or after it. */
621 while (p < end && c_isspace (*p))
623 if (p == end || *p != '=')
624 return line_syntax_error;
626 while (p < end && c_isspace (*p))
632 /* The syntax is valid (even though the command might not be). Fill
633 in the command name and value. */
634 *com = strdupdelim (cmdstart, cmdend);
635 *val = strdupdelim (valstart, valend);
637 /* The line now known to be syntactically correct. Check whether
638 the command is valid. */
639 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
641 ind = command_by_name (cmdcopy);
643 return line_unknown_command;
645 /* Report success to the caller. */
650 /* Run commands[comind].action. */
653 setval_internal (int comind, const char *com, const char *val)
655 assert (0 <= comind && comind < countof (commands));
656 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
657 return commands[comind].action (com, val, commands[comind].place);
660 /* Run command COM with value VAL. If running the command produces an
661 error, report the error and exit.
663 This is intended to be called from main() to modify Wget's behavior
664 through command-line switches. Since COM is hard-coded in main(),
665 it is not canonicalized, and this aborts when COM is not found.
667 If COMIND's are exported to init.h, this function will be changed
668 to accept COMIND directly. */
671 setoptval (const char *com, const char *val, const char *optname)
673 /* Prepend "--" to OPTNAME. */
674 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
677 strcpy (dd_optname + 2, optname);
679 assert (val != NULL);
680 if (!setval_internal (command_by_name (com), dd_optname, val))
684 /* Parse OPT into command and value and run it. For example,
685 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
686 This is used by the `--execute' flag in main.c. */
689 run_command (const char *opt)
693 switch (parse_line (opt, &com, &val, &comind))
696 if (!setval_internal (comind, com, val))
702 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
708 /* Generic helper functions, for use with `commands'. */
710 /* Forward declarations: */
715 static bool decode_string (const char *, const struct decode_item *, int, int *);
716 static bool simple_atoi (const char *, const char *, int *);
717 static bool simple_atof (const char *, const char *, double *);
719 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
721 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
722 && c_tolower((p)[1]) == (c1) \
725 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
726 && c_tolower((p)[1]) == (c1) \
727 && c_tolower((p)[2]) == (c2) \
731 /* Store the boolean value from VAL to PLACE. COM is ignored,
732 except for error messages. */
734 cmd_boolean (const char *com, const char *val, void *place)
738 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
739 /* "on", "yes" and "1" mean true. */
741 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
742 /* "off", "no" and "0" mean false. */
747 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
748 exec_name, com, val);
752 *(bool *) place = value;
756 /* Set the non-negative integer value from VAL to PLACE. With
757 incorrect specification, the number remains unchanged. */
759 cmd_number (const char *com, const char *val, void *place)
761 if (!simple_atoi (val, val + strlen (val), place)
762 || *(int *) place < 0)
764 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
765 exec_name, com, val);
771 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
773 cmd_number_inf (const char *com, const char *val, void *place)
775 if (!strcasecmp (val, "inf"))
780 return cmd_number (com, val, place);
783 /* Copy (strdup) the string at COM to a new location and place a
784 pointer to *PLACE. */
786 cmd_string (const char *com, const char *val, void *place)
788 char **pstring = (char **)place;
790 xfree_null (*pstring);
791 *pstring = xstrdup (val);
795 #if defined(WINDOWS) || defined(MSDOS)
796 # define ISSEP(c) ((c) == '/' || (c) == '\\')
798 # define ISSEP(c) ((c) == '/')
801 /* Like the above, but handles tilde-expansion when reading a user's
802 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
803 gets expanded to the user's home directory. */
805 cmd_file (const char *com, const char *val, void *place)
807 char **pstring = (char **)place;
809 xfree_null (*pstring);
811 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
813 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
816 *pstring = xstrdup (val);
821 char *home = home_dir ();
825 homelen = strlen (home);
826 while (homelen && ISSEP (home[homelen - 1]))
827 home[--homelen] = '\0';
829 /* Skip the leading "~/". */
830 for (++val; ISSEP (*val); val++)
833 *pstring = concat_strings (home, "/", val, (char *) 0);
836 #if defined(WINDOWS) || defined(MSDOS)
837 /* Convert "\" to "/". */
840 for (s = *pstring; *s; s++)
848 /* Like cmd_file, but strips trailing '/' characters. */
850 cmd_directory (const char *com, const char *val, void *place)
854 /* Call cmd_file() for tilde expansion and separator
855 canonicalization (backslash -> slash under Windows). These
856 things should perhaps be in a separate function. */
857 if (!cmd_file (com, val, place))
862 while (t > s && *--t == '/')
868 /* Split VAL by space to a vector of values, and append those values
869 to vector pointed to by the PLACE argument. If VAL is empty, the
870 PLACE vector is cleared instead. */
873 cmd_vector (const char *com, const char *val, void *place)
875 char ***pvec = (char ***)place;
878 *pvec = merge_vecs (*pvec, sepstring (val));
888 cmd_directory_vector (const char *com, const char *val, void *place)
890 char ***pvec = (char ***)place;
894 /* Strip the trailing slashes from directories. */
897 seps = sepstring (val);
898 for (t = seps; t && *t; t++)
900 int len = strlen (*t);
901 /* Skip degenerate case of root directory. */
904 if ((*t)[len - 1] == '/')
905 (*t)[len - 1] = '\0';
908 *pvec = merge_vecs (*pvec, seps);
918 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
919 "100k" or "2.5G" to a floating point number. */
922 parse_bytes_helper (const char *val, double *result)
925 const char *end = val + strlen (val);
927 /* Check for "inf". */
928 if (0 == strcmp (val, "inf"))
934 /* Strip trailing whitespace. */
935 while (val < end && c_isspace (end[-1]))
940 switch (c_tolower (end[-1]))
943 --end, mult = 1024.0;
946 --end, mult = 1048576.0;
949 --end, mult = 1073741824.0;
952 --end, mult = 1099511627776.0;
955 /* Not a recognized suffix: assume it's a digit. (If not,
956 simple_atof will raise an error.) */
960 /* Skip leading and trailing whitespace. */
961 while (val < end && c_isspace (*val))
963 while (val < end && c_isspace (end[-1]))
968 if (!simple_atof (val, end, &number) || number < 0)
971 *result = number * mult;
975 /* Parse VAL as a number and set its value to PLACE (which should
978 By default, the value is assumed to be in bytes. If "K", "M", or
979 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
980 1<<30, respectively. Floating point values are allowed and are
981 cast to integer before use. The idea is to be able to use things
982 like 1.5k instead of "1536".
984 The string "inf" is returned as 0.
986 In case of error, false is returned and memory pointed to by PLACE
987 remains unmodified. */
990 cmd_bytes (const char *com, const char *val, void *place)
993 if (!parse_bytes_helper (val, &byte_value))
995 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
996 exec_name, com, val);
999 *(wgint *)place = (wgint)byte_value;
1003 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1004 SIZE_SUM. It works by converting the string to double, therefore
1005 working with values up to 2^53-1 without loss of precision. This
1006 value (8192 TB) is large enough to serve for a while. */
1009 cmd_bytes_sum (const char *com, const char *val, void *place)
1012 if (!parse_bytes_helper (val, &byte_value))
1014 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
1015 exec_name, com, val);
1018 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1022 /* Store the value of VAL to *OUT. The value is a time period, by
1023 default expressed in seconds, but also accepting suffixes "m", "h",
1024 "d", and "w" for minutes, hours, days, and weeks respectively. */
1027 cmd_time (const char *com, const char *val, void *place)
1029 double number, mult;
1030 const char *end = val + strlen (val);
1032 /* Strip trailing whitespace. */
1033 while (val < end && c_isspace (end[-1]))
1039 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1040 exec_name, com, val);
1044 switch (c_tolower (end[-1]))
1047 --end, mult = 1; /* seconds */
1050 --end, mult = 60; /* minutes */
1053 --end, mult = 3600; /* hours */
1056 --end, mult = 86400.0; /* days */
1059 --end, mult = 604800.0; /* weeks */
1062 /* Not a recognized suffix: assume it belongs to the number.
1063 (If not, simple_atof will raise an error.) */
1067 /* Skip leading and trailing whitespace. */
1068 while (val < end && c_isspace (*val))
1070 while (val < end && c_isspace (end[-1]))
1075 if (!simple_atof (val, end, &number))
1078 *(double *)place = number * mult;
1084 cmd_cert_type (const char *com, const char *val, void *place)
1086 static const struct decode_item choices[] = {
1087 { "pem", keyfile_pem },
1088 { "der", keyfile_asn1 },
1089 { "asn1", keyfile_asn1 },
1091 int ok = decode_string (val, choices, countof (choices), place);
1093 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1098 /* Specialized helper functions, used by `commands' to handle some
1099 options specially. */
1101 static bool check_user_specified_header (const char *);
1104 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1106 if (!cmd_boolean (com, val, &opt.dirstruct))
1108 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1109 must be affected inversely. */
1111 opt.no_dirstruct = false;
1113 opt.no_dirstruct = true;
1118 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1120 /* Empty value means reset the list of headers. */
1123 free_vec (opt.user_headers);
1124 opt.user_headers = NULL;
1128 if (!check_user_specified_header (val))
1130 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1131 exec_name, com, val);
1134 opt.user_headers = vec_append (opt.user_headers, val);
1139 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1141 int flag = cmd_boolean (com, val, &opt.htmlify);
1142 if (flag && !opt.htmlify)
1143 opt.remove_listing = false;
1147 /* Set the "mirror" mode. It means: recursive download, timestamping,
1148 no limit on max. recursion depth, and don't remove listings. */
1151 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1155 if (!cmd_boolean (com, val, &mirror))
1159 opt.recursive = true;
1160 if (!opt.no_dirstruct)
1161 opt.dirstruct = true;
1162 opt.timestamping = true;
1163 opt.reclevel = INFINITE_RECURSION;
1164 opt.remove_listing = false;
1169 /* Validate --prefer-family and set the choice. Allowed values are
1170 "IPv4", "IPv6", and "none". */
1173 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1175 static const struct decode_item choices[] = {
1176 { "IPv4", prefer_ipv4 },
1177 { "IPv6", prefer_ipv6 },
1178 { "none", prefer_none },
1180 int prefer_family = prefer_ipv4;
1181 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1183 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1184 opt.prefer_family = prefer_family;
1188 /* Set progress.type to VAL, but verify that it's a valid progress
1189 implementation before that. */
1192 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1194 if (!valid_progress_implementation_p (val))
1196 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1197 exec_name, com, val);
1200 xfree_null (opt.progress_type);
1202 /* Don't call set_progress_implementation here. It will be called
1203 in main() when it becomes clear what the log output is. */
1204 opt.progress_type = xstrdup (val);
1208 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1209 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1213 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1215 if (!cmd_boolean (com, val, &opt.recursive))
1219 if (opt.recursive && !opt.no_dirstruct)
1220 opt.dirstruct = true;
1226 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1228 int restrict_os = opt.restrict_files_os;
1229 int restrict_ctrl = opt.restrict_files_ctrl;
1230 int restrict_case = opt.restrict_files_case;
1234 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1238 end = strchr (val, ',');
1240 end = val + strlen (val);
1242 if (VAL_IS ("unix"))
1243 restrict_os = restrict_unix;
1244 else if (VAL_IS ("windows"))
1245 restrict_os = restrict_windows;
1246 else if (VAL_IS ("lowercase"))
1247 restrict_case = restrict_lowercase;
1248 else if (VAL_IS ("uppercase"))
1249 restrict_case = restrict_uppercase;
1250 else if (VAL_IS ("nocontrol"))
1251 restrict_ctrl = false;
1255 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1256 exec_name, com, val);
1263 while (*val && *end);
1267 opt.restrict_files_os = restrict_os;
1268 opt.restrict_files_ctrl = restrict_ctrl;
1269 opt.restrict_files_case = restrict_case;
1276 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1278 static const struct decode_item choices[] = {
1279 { "auto", secure_protocol_auto },
1280 { "sslv2", secure_protocol_sslv2 },
1281 { "sslv3", secure_protocol_sslv3 },
1282 { "tlsv1", secure_protocol_tlsv1 },
1284 int ok = decode_string (val, choices, countof (choices), place);
1286 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1291 /* Set all three timeout values. */
1294 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1297 if (!cmd_time (com, val, &value))
1299 opt.read_timeout = value;
1300 opt.connect_timeout = value;
1301 opt.dns_timeout = value;
1306 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1308 /* Disallow embedded newlines. */
1309 if (strchr (val, '\n'))
1311 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1312 exec_name, com, val);
1315 xfree_null (opt.useragent);
1316 opt.useragent = xstrdup (val);
1320 /* The "verbose" option cannot be cmd_boolean because the variable is
1321 not bool -- it's of type int (-1 means uninitialized because of
1322 some random hackery for disallowing -q -v). */
1325 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1328 if (cmd_boolean (com, val, &flag))
1336 /* Miscellaneous useful routines. */
1338 /* A very simple atoi clone, more useful than atoi because it works on
1339 delimited strings, and has error reportage. Returns true on success,
1340 false on failure. If successful, stores result to *DEST. */
1343 simple_atoi (const char *beg, const char *end, int *dest)
1346 bool negative = false;
1347 const char *p = beg;
1349 while (p < end && c_isspace (*p))
1351 if (p < end && (*p == '-' || *p == '+'))
1353 negative = (*p == '-');
1359 /* Read negative numbers in a separate loop because the most
1360 negative integer cannot be represented as a positive number. */
1363 for (; p < end && c_isdigit (*p); p++)
1365 int next = (10 * result) + (*p - '0');
1367 return false; /* overflow */
1371 for (; p < end && c_isdigit (*p); p++)
1373 int next = (10 * result) - (*p - '0');
1375 return false; /* underflow */
1386 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1387 doesn't handle exponential notation. Returns true on success,
1388 false on failure. In case of success, stores its result to
1392 simple_atof (const char *beg, const char *end, double *dest)
1396 bool negative = false;
1397 bool seen_dot = false;
1398 bool seen_digit = false;
1401 const char *p = beg;
1403 while (p < end && c_isspace (*p))
1405 if (p < end && (*p == '-' || *p == '+'))
1407 negative = (*p == '-');
1411 for (; p < end; p++)
1417 result = (10 * result) + (ch - '0');
1419 result += (ch - '0') / (divider *= 10);
1441 /* Verify that the user-specified header in S is valid. It must
1442 contain a colon preceded by non-white-space characters and must not
1443 contain newlines. */
1446 check_user_specified_header (const char *s)
1450 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1452 /* The header MUST contain `:' preceded by at least one
1453 non-whitespace character. */
1454 if (*p != ':' || p == s)
1456 /* The header MUST NOT contain newlines. */
1457 if (strchr (s, '\n'))
1462 /* Decode VAL into a number, according to ITEMS. */
1465 decode_string (const char *val, const struct decode_item *items, int itemcount,
1469 for (i = 0; i < itemcount; i++)
1470 if (0 == strcasecmp (val, items[i].name))
1472 *place = items[i].code;
1479 void cleanup_html_url (void);
1482 /* Free the memory allocated by global variables. */
1486 /* Free external resources, close files, etc. */
1489 fclose (output_stream);
1490 /* No need to check for error because Wget flushes its output (and
1491 checks for errors) after any data arrives. */
1493 /* We're exiting anyway so there's no real need to call free()
1494 hundreds of times. Skipping the frees will make Wget exit
1497 However, when detecting leaks, it's crucial to free() everything
1498 because then you can find the real leaks, i.e. the allocated
1499 memory which grows with the size of the program. */
1505 cleanup_html_url ();
1510 extern acc_t *netrc_list;
1511 free_netrc (netrc_list);
1513 xfree_null (opt.lfilename);
1514 xfree_null (opt.dir_prefix);
1515 xfree_null (opt.input_filename);
1516 xfree_null (opt.output_document);
1517 free_vec (opt.accepts);
1518 free_vec (opt.rejects);
1519 free_vec (opt.excludes);
1520 free_vec (opt.includes);
1521 free_vec (opt.domains);
1522 free_vec (opt.follow_tags);
1523 free_vec (opt.ignore_tags);
1524 xfree_null (opt.progress_type);
1525 xfree_null (opt.ftp_user);
1526 xfree_null (opt.ftp_passwd);
1527 xfree_null (opt.ftp_proxy);
1528 xfree_null (opt.https_proxy);
1529 xfree_null (opt.http_proxy);
1530 free_vec (opt.no_proxy);
1531 xfree_null (opt.useragent);
1532 xfree_null (opt.referer);
1533 xfree_null (opt.http_user);
1534 xfree_null (opt.http_passwd);
1535 free_vec (opt.user_headers);
1537 xfree_null (opt.cert_file);
1538 xfree_null (opt.private_key);
1539 xfree_null (opt.ca_directory);
1540 xfree_null (opt.ca_cert);
1541 xfree_null (opt.random_file);
1542 xfree_null (opt.egd_file);
1544 xfree_null (opt.bind_address);
1545 xfree_null (opt.cookies_input);
1546 xfree_null (opt.cookies_output);
1547 xfree_null (opt.user);
1548 xfree_null (opt.passwd);
1549 #endif /* DEBUG_MALLOC */
1552 /* Unit testing routines. */
1557 test_commands_sorted()
1559 int prev_idx = 0, next_idx = 1;
1560 int command_count = countof (commands) - 1;
1562 while (next_idx <= command_count)
1564 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1567 mu_assert ("FAILED", false);
1580 test_cmd_spec_restrict_file_names()
1585 int expected_restrict_files_os;
1586 int expected_restrict_files_ctrl;
1587 int expected_restrict_files_case;
1590 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1591 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1592 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1593 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1596 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1601 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1604 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1605 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1606 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1607 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1609 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1610 res == test_array[i].result
1611 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1612 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1613 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1619 #endif /* TESTING */