1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
61 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
62 otherwise, it will be performed by the shell. This variable will
63 be set by the wgetrc-reading function. */
65 static bool enable_tilde_expansion;
68 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
70 CMD_DECLARE (cmd_boolean);
71 CMD_DECLARE (cmd_bytes);
72 CMD_DECLARE (cmd_bytes_sum);
74 CMD_DECLARE (cmd_cert_type);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_number);
78 CMD_DECLARE (cmd_number_inf);
79 CMD_DECLARE (cmd_string);
80 CMD_DECLARE (cmd_file);
81 CMD_DECLARE (cmd_directory);
82 CMD_DECLARE (cmd_time);
83 CMD_DECLARE (cmd_vector);
85 CMD_DECLARE (cmd_spec_dirstruct);
86 CMD_DECLARE (cmd_spec_header);
87 CMD_DECLARE (cmd_spec_htmlify);
88 CMD_DECLARE (cmd_spec_mirror);
89 CMD_DECLARE (cmd_spec_prefer_family);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_secure_protocol);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
98 CMD_DECLARE (cmd_spec_verbose);
100 /* List of recognized commands, each consisting of name, place and
101 function. When adding a new command, simply add it to the list,
102 but be sure to keep the list sorted alphabetically, as
103 command_by_name's binary search depends on it. Also, be sure to
104 add any entries that allocate memory (e.g. cmd_string and
105 cmd_vector) to the cleanup() function below. */
107 static const struct {
110 bool (*action) (const char *, const char *, void *);
112 /* KEEP THIS LIST ALPHABETICALLY SORTED */
113 { "accept", &opt.accepts, cmd_vector },
114 { "addhostdir", &opt.add_hostdir, cmd_boolean },
115 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
116 { "askpassword", &opt.ask_passwd, cmd_boolean },
117 { "authnochallenge", &opt.auth_without_challenge,
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "contentdisposition", &opt.content_disposition, cmd_boolean },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "defaultpage", &opt.default_page, cmd_string},
144 { "deleteafter", &opt.delete_after, cmd_boolean },
145 { "dirprefix", &opt.dir_prefix, cmd_directory },
146 { "dirstruct", NULL, cmd_spec_dirstruct },
147 { "dnscache", &opt.dns_cache, cmd_boolean },
148 { "dnstimeout", &opt.dns_timeout, cmd_time },
149 { "domains", &opt.domains, cmd_vector },
150 { "dotbytes", &opt.dot_bytes, cmd_bytes },
151 { "dotsinline", &opt.dots_in_line, cmd_number },
152 { "dotspacing", &opt.dot_spacing, cmd_number },
153 { "dotstyle", &opt.dot_style, cmd_string },
155 { "egdfile", &opt.egd_file, cmd_file },
157 { "excludedirectories", &opt.excludes, cmd_directory_vector },
158 { "excludedomains", &opt.exclude_domains, cmd_vector },
159 { "followftp", &opt.follow_ftp, cmd_boolean },
160 { "followtags", &opt.follow_tags, cmd_vector },
161 { "forcehtml", &opt.force_html, cmd_boolean },
162 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
163 { "ftppassword", &opt.ftp_passwd, cmd_string },
164 { "ftpproxy", &opt.ftp_proxy, cmd_string },
165 { "ftpuser", &opt.ftp_user, cmd_string },
166 { "glob", &opt.ftp_glob, cmd_boolean },
167 { "header", NULL, cmd_spec_header },
168 { "htmlextension", &opt.html_extension, cmd_boolean },
169 { "htmlify", NULL, cmd_spec_htmlify },
170 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
171 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
172 { "httppassword", &opt.http_passwd, cmd_string },
173 { "httpproxy", &opt.http_proxy, cmd_string },
174 { "httpsproxy", &opt.https_proxy, cmd_string },
175 { "httpuser", &opt.http_user, cmd_string },
176 { "ignorecase", &opt.ignore_case, cmd_boolean },
177 { "ignorelength", &opt.ignore_length, cmd_boolean },
178 { "ignoretags", &opt.ignore_tags, cmd_vector },
179 { "includedirectories", &opt.includes, cmd_directory_vector },
181 { "inet4only", &opt.ipv4_only, cmd_boolean },
182 { "inet6only", &opt.ipv6_only, cmd_boolean },
184 { "input", &opt.input_filename, cmd_file },
185 { "iri", &opt.enable_iri, cmd_boolean },
186 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
187 { "limitrate", &opt.limit_rate, cmd_bytes },
188 { "loadcookies", &opt.cookies_input, cmd_file },
189 { "locale", &opt.locale, cmd_string },
190 { "logfile", &opt.lfilename, cmd_file },
191 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
192 { "maxredirect", &opt.max_redirect, cmd_number },
193 { "mirror", NULL, cmd_spec_mirror },
194 { "netrc", &opt.netrc, cmd_boolean },
195 { "noclobber", &opt.noclobber, cmd_boolean },
196 { "noparent", &opt.no_parent, cmd_boolean },
197 { "noproxy", &opt.no_proxy, cmd_vector },
198 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
199 { "outputdocument", &opt.output_document, cmd_file },
200 { "pagerequisites", &opt.page_requisites, cmd_boolean },
201 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
202 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
203 { "password", &opt.passwd, cmd_string },
204 { "postdata", &opt.post_data, cmd_string },
205 { "postfile", &opt.post_file_name, cmd_file },
206 { "preferfamily", NULL, cmd_spec_prefer_family },
207 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
209 { "privatekey", &opt.private_key, cmd_file },
210 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
212 { "progress", &opt.progress_type, cmd_spec_progress },
213 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
214 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
215 { "proxypassword", &opt.proxy_passwd, cmd_string },
216 { "proxyuser", &opt.proxy_user, cmd_string },
217 { "quiet", &opt.quiet, cmd_boolean },
218 { "quota", &opt.quota, cmd_bytes_sum },
220 { "randomfile", &opt.random_file, cmd_file },
222 { "randomwait", &opt.random_wait, cmd_boolean },
223 { "readtimeout", &opt.read_timeout, cmd_time },
224 { "reclevel", &opt.reclevel, cmd_number_inf },
225 { "recursive", NULL, cmd_spec_recursive },
226 { "referer", &opt.referer, cmd_string },
227 { "reject", &opt.rejects, cmd_vector },
228 { "relativeonly", &opt.relative_only, cmd_boolean },
229 { "remoteencoding", &opt.encoding_remote, cmd_string },
230 { "removelisting", &opt.remove_listing, cmd_boolean },
231 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
232 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
233 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
234 { "robots", &opt.use_robots, cmd_boolean },
235 { "savecookies", &opt.cookies_output, cmd_file },
236 { "saveheaders", &opt.save_headers, cmd_boolean },
238 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
240 { "serverresponse", &opt.server_response, cmd_boolean },
241 { "spanhosts", &opt.spanhost, cmd_boolean },
242 { "spider", &opt.spider, cmd_boolean },
243 { "strictcomments", &opt.strict_comments, cmd_boolean },
244 { "timeout", NULL, cmd_spec_timeout },
245 { "timestamping", &opt.timestamping, cmd_boolean },
246 { "tries", &opt.ntry, cmd_number_inf },
247 { "useproxy", &opt.use_proxy, cmd_boolean },
248 { "user", &opt.user, cmd_string },
249 { "useragent", NULL, cmd_spec_useragent },
250 { "verbose", NULL, cmd_spec_verbose },
251 { "wait", &opt.wait, cmd_time },
252 { "waitretry", &opt.waitretry, cmd_time },
254 { "wdebug", &opt.wdebug, cmd_boolean },
258 /* Look up CMDNAME in the commands[] and return its position in the
259 array. If CMDNAME is not found, return -1. */
262 command_by_name (const char *cmdname)
264 /* Use binary search for speed. Wget has ~100 commands, which
265 guarantees a worst case performance of 7 string comparisons. */
266 int lo = 0, hi = countof (commands) - 1;
270 int mid = (lo + hi) >> 1;
271 int cmp = strcasecmp (cmdname, commands[mid].name);
282 /* Reset the variables to default values. */
288 /* Most of the default values are 0 (and 0.0, NULL, and false).
289 Just reset everything, and fill in the non-zero values. Note
290 that initializing pointers to NULL this way is technically
291 illegal, but porting Wget to a machine where NULL is not all-zero
292 bit pattern will be the least of the implementors' worries. */
299 opt.add_hostdir = true;
303 opt.http_keep_alive = true;
304 opt.use_proxy = true;
305 tmp = getenv ("no_proxy");
307 opt.no_proxy = sepstring (tmp);
308 opt.prefer_family = prefer_none;
309 opt.allow_cache = true;
311 opt.read_timeout = 900;
312 opt.use_robots = true;
314 opt.remove_listing = true;
316 opt.dot_bytes = 1024;
317 opt.dot_spacing = 10;
318 opt.dots_in_line = 50;
320 opt.dns_cache = true;
324 opt.check_cert = true;
327 /* The default for file name restriction defaults to the OS type. */
328 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
329 opt.restrict_files_os = restrict_windows;
331 opt.restrict_files_os = restrict_unix;
333 opt.restrict_files_ctrl = true;
334 opt.restrict_files_case = restrict_no_case_restriction;
336 opt.max_redirect = 20;
341 opt.enable_iri = true;
343 opt.enable_iri = false;
346 opt.encoding_remote = NULL;
349 /* Return the user's home directory (strdup-ed), or NULL if none is
354 static char buf[PATH_MAX];
359 home = getenv ("HOME");
363 /* Under MSDOS, if $HOME isn't defined, use the directory where
364 `wget.exe' resides. */
365 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
368 strcpy (buf, _w32_get_argv0 ());
369 p = strrchr (buf, '/'); /* djgpp */
371 p = strrchr (buf, '\\'); /* others */
375 #elif !defined(WINDOWS)
376 /* If HOME is not defined, try getting it from the password
378 struct passwd *pwd = getpwuid (getuid ());
379 if (!pwd || !pwd->pw_dir)
381 strcpy (buf, pwd->pw_dir);
384 /* Under Windows, if $HOME isn't defined, use the directory where
385 `wget.exe' resides. */
391 return home ? xstrdup (home) : NULL;
394 /* Check the 'WGETRC' environment variable and return the file name
395 if 'WGETRC' is set and is a valid file.
396 If the `WGETRC' variable exists but the file does not exist, the
397 function will exit(). */
399 wgetrc_env_file_name (void)
401 char *env = getenv ("WGETRC");
404 if (!file_exists_p (env))
406 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
410 return xstrdup (env);
415 /* Check for the existance of '$HOME/.wgetrc' and return it's path
416 if it exists and is set. */
418 wgetrc_user_file_name (void)
420 char *home = home_dir ();
423 file = aprintf ("%s/.wgetrc", home);
427 if (!file_exists_p (file))
435 /* Return the path to the user's .wgetrc. This is either the value of
436 `WGETRC' environment variable, or `$HOME/.wgetrc'.
438 Additionally, for windows, look in the directory where wget.exe
441 wgetrc_file_name (void)
444 char *file = wgetrc_env_file_name ();
448 file = wgetrc_user_file_name ();
451 /* Under Windows, if we still haven't found .wgetrc, look for the file
452 `wget.ini' in the directory where `wget.exe' resides; we do this for
453 backward compatibility with previous versions of Wget.
454 SYSTEM_WGETRC should not be defined under WINDOWS. */
456 if (!file || !file_exists_p (file))
462 file = aprintf ("%s/wget.ini", home);
469 if (!file_exists_p (file))
477 /* Return values of parse_line. */
485 static enum parse_line parse_line (const char *, char **, char **, int *);
486 static bool setval_internal (int, const char *, const char *);
488 /* Initialize variables from a wgetrc file. Returns zero (failure) if
489 there were errors in the file. */
492 run_wgetrc (const char *file)
499 fp = fopen (file, "rb");
502 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
503 file, strerror (errno));
504 return true; /* not a fatal error */
506 enable_tilde_expansion = true;
508 while ((line = read_whole_line (fp)) != NULL)
510 char *com = NULL, *val = NULL;
513 /* Parse the line. */
514 switch (parse_line (line, &com, &val, &comind))
517 /* If everything is OK, set the value. */
518 if (!setval_internal (comind, com, val))
520 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
521 exec_name, file, ln);
525 case line_syntax_error:
526 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
527 exec_name, file, ln);
530 case line_unknown_command:
531 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
532 exec_name, quote (com), file, ln);
545 enable_tilde_expansion = false;
551 /* Initialize the defaults and run the system wgetrc and user's own
559 /* Load the hard-coded defaults. */
562 /* If SYSTEM_WGETRC is defined, use it. */
564 if (file_exists_p (SYSTEM_WGETRC))
565 ok &= run_wgetrc (SYSTEM_WGETRC);
567 /* Override it with your own, if one exists. */
568 file = wgetrc_file_name ();
571 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
572 something like realpath() before comparing them with `strcmp' */
574 if (!strcmp (file, SYSTEM_WGETRC))
576 fprintf (stderr, _("\
577 %s: Warning: Both system and user wgetrc point to %s.\n"),
578 exec_name, quote (file));
582 ok &= run_wgetrc (file);
584 /* If there were errors processing either `.wgetrc', abort. */
592 /* Remove dashes and underscores from S, modifying S in the
598 char *t = s; /* t - tortoise */
599 char *h = s; /* h - hare */
601 if (*h == '_' || *h == '-')
608 /* Parse the line pointed by line, with the syntax:
609 <sp>* command <sp>* = <sp>* value <sp>*
610 Uses malloc to allocate space for command and value.
612 Returns one of line_ok, line_empty, line_syntax_error, or
613 line_unknown_command.
615 In case of line_ok, *COM and *VAL point to freshly allocated
616 strings, and *COMIND points to com's index. In case of error or
617 empty line, their values are unmodified. */
619 static enum parse_line
620 parse_line (const char *line, char **com, char **val, int *comind)
623 const char *end = line + strlen (line);
624 const char *cmdstart, *cmdend;
625 const char *valstart, *valend;
630 /* Skip leading and trailing whitespace. */
631 while (*line && c_isspace (*line))
633 while (end > line && c_isspace (end[-1]))
636 /* Skip empty lines and comments. */
637 if (!*line || *line == '#')
643 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
647 /* Skip '=', as well as any space before or after it. */
648 while (p < end && c_isspace (*p))
650 if (p == end || *p != '=')
651 return line_syntax_error;
653 while (p < end && c_isspace (*p))
659 /* The syntax is valid (even though the command might not be). Fill
660 in the command name and value. */
661 *com = strdupdelim (cmdstart, cmdend);
662 *val = strdupdelim (valstart, valend);
664 /* The line now known to be syntactically correct. Check whether
665 the command is valid. */
666 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
668 ind = command_by_name (cmdcopy);
670 return line_unknown_command;
672 /* Report success to the caller. */
677 /* Run commands[comind].action. */
680 setval_internal (int comind, const char *com, const char *val)
682 assert (0 <= comind && ((size_t) comind) < countof (commands));
683 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
684 return commands[comind].action (com, val, commands[comind].place);
687 /* Run command COM with value VAL. If running the command produces an
688 error, report the error and exit.
690 This is intended to be called from main() to modify Wget's behavior
691 through command-line switches. Since COM is hard-coded in main(),
692 it is not canonicalized, and this aborts when COM is not found.
694 If COMIND's are exported to init.h, this function will be changed
695 to accept COMIND directly. */
698 setoptval (const char *com, const char *val, const char *optname)
700 /* Prepend "--" to OPTNAME. */
701 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
704 strcpy (dd_optname + 2, optname);
706 assert (val != NULL);
707 if (!setval_internal (command_by_name (com), dd_optname, val))
711 /* Parse OPT into command and value and run it. For example,
712 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
713 This is used by the `--execute' flag in main.c. */
716 run_command (const char *opt)
720 switch (parse_line (opt, &com, &val, &comind))
723 if (!setval_internal (comind, com, val))
729 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
730 exec_name, quote (opt));
735 /* Generic helper functions, for use with `commands'. */
737 /* Forward declarations: */
742 static bool decode_string (const char *, const struct decode_item *, int, int *);
743 static bool simple_atoi (const char *, const char *, int *);
744 static bool simple_atof (const char *, const char *, double *);
746 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
748 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
749 && c_tolower((p)[1]) == (c1) \
752 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
753 && c_tolower((p)[1]) == (c1) \
754 && c_tolower((p)[2]) == (c2) \
758 /* Store the boolean value from VAL to PLACE. COM is ignored,
759 except for error messages. */
761 cmd_boolean (const char *com, const char *val, void *place)
765 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
766 /* "on", "yes" and "1" mean true. */
768 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
769 /* "off", "no" and "0" mean false. */
774 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
775 exec_name, com, quote (val));
779 *(bool *) place = value;
783 /* Set the non-negative integer value from VAL to PLACE. With
784 incorrect specification, the number remains unchanged. */
786 cmd_number (const char *com, const char *val, void *place)
788 if (!simple_atoi (val, val + strlen (val), place)
789 || *(int *) place < 0)
791 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
792 exec_name, com, quote (val));
798 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
800 cmd_number_inf (const char *com, const char *val, void *place)
802 if (!strcasecmp (val, "inf"))
807 return cmd_number (com, val, place);
810 /* Copy (strdup) the string at COM to a new location and place a
811 pointer to *PLACE. */
813 cmd_string (const char *com, const char *val, void *place)
815 char **pstring = (char **)place;
817 xfree_null (*pstring);
818 *pstring = xstrdup (val);
822 #if defined(WINDOWS) || defined(MSDOS)
823 # define ISSEP(c) ((c) == '/' || (c) == '\\')
825 # define ISSEP(c) ((c) == '/')
828 /* Like the above, but handles tilde-expansion when reading a user's
829 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
830 gets expanded to the user's home directory. */
832 cmd_file (const char *com, const char *val, void *place)
834 char **pstring = (char **)place;
836 xfree_null (*pstring);
838 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
840 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
843 *pstring = xstrdup (val);
848 char *home = home_dir ();
852 homelen = strlen (home);
853 while (homelen && ISSEP (home[homelen - 1]))
854 home[--homelen] = '\0';
856 /* Skip the leading "~/". */
857 for (++val; ISSEP (*val); val++)
860 *pstring = concat_strings (home, "/", val, (char *) 0);
863 #if defined(WINDOWS) || defined(MSDOS)
864 /* Convert "\" to "/". */
867 for (s = *pstring; *s; s++)
875 /* Like cmd_file, but strips trailing '/' characters. */
877 cmd_directory (const char *com, const char *val, void *place)
881 /* Call cmd_file() for tilde expansion and separator
882 canonicalization (backslash -> slash under Windows). These
883 things should perhaps be in a separate function. */
884 if (!cmd_file (com, val, place))
889 while (t > s && *--t == '/')
895 /* Split VAL by space to a vector of values, and append those values
896 to vector pointed to by the PLACE argument. If VAL is empty, the
897 PLACE vector is cleared instead. */
900 cmd_vector (const char *com, const char *val, void *place)
902 char ***pvec = (char ***)place;
905 *pvec = merge_vecs (*pvec, sepstring (val));
915 cmd_directory_vector (const char *com, const char *val, void *place)
917 char ***pvec = (char ***)place;
921 /* Strip the trailing slashes from directories. */
924 seps = sepstring (val);
925 for (t = seps; t && *t; t++)
927 int len = strlen (*t);
928 /* Skip degenerate case of root directory. */
931 if ((*t)[len - 1] == '/')
932 (*t)[len - 1] = '\0';
935 *pvec = merge_vecs (*pvec, seps);
945 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
946 "100k" or "2.5G" to a floating point number. */
949 parse_bytes_helper (const char *val, double *result)
952 const char *end = val + strlen (val);
954 /* Check for "inf". */
955 if (0 == strcmp (val, "inf"))
961 /* Strip trailing whitespace. */
962 while (val < end && c_isspace (end[-1]))
967 switch (c_tolower (end[-1]))
970 --end, mult = 1024.0;
973 --end, mult = 1048576.0;
976 --end, mult = 1073741824.0;
979 --end, mult = 1099511627776.0;
982 /* Not a recognized suffix: assume it's a digit. (If not,
983 simple_atof will raise an error.) */
987 /* Skip leading and trailing whitespace. */
988 while (val < end && c_isspace (*val))
990 while (val < end && c_isspace (end[-1]))
995 if (!simple_atof (val, end, &number) || number < 0)
998 *result = number * mult;
1002 /* Parse VAL as a number and set its value to PLACE (which should
1005 By default, the value is assumed to be in bytes. If "K", "M", or
1006 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1007 1<<30, respectively. Floating point values are allowed and are
1008 cast to integer before use. The idea is to be able to use things
1009 like 1.5k instead of "1536".
1011 The string "inf" is returned as 0.
1013 In case of error, false is returned and memory pointed to by PLACE
1014 remains unmodified. */
1017 cmd_bytes (const char *com, const char *val, void *place)
1020 if (!parse_bytes_helper (val, &byte_value))
1022 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1023 exec_name, com, quote (val));
1026 *(wgint *)place = (wgint)byte_value;
1030 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1031 SIZE_SUM. It works by converting the string to double, therefore
1032 working with values up to 2^53-1 without loss of precision. This
1033 value (8192 TB) is large enough to serve for a while. */
1036 cmd_bytes_sum (const char *com, const char *val, void *place)
1039 if (!parse_bytes_helper (val, &byte_value))
1041 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1042 exec_name, com, quote (val));
1045 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1049 /* Store the value of VAL to *OUT. The value is a time period, by
1050 default expressed in seconds, but also accepting suffixes "m", "h",
1051 "d", and "w" for minutes, hours, days, and weeks respectively. */
1054 cmd_time (const char *com, const char *val, void *place)
1056 double number, mult;
1057 const char *end = val + strlen (val);
1059 /* Strip trailing whitespace. */
1060 while (val < end && c_isspace (end[-1]))
1066 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1067 exec_name, com, quote (val));
1071 switch (c_tolower (end[-1]))
1074 --end, mult = 1; /* seconds */
1077 --end, mult = 60; /* minutes */
1080 --end, mult = 3600; /* hours */
1083 --end, mult = 86400.0; /* days */
1086 --end, mult = 604800.0; /* weeks */
1089 /* Not a recognized suffix: assume it belongs to the number.
1090 (If not, simple_atof will raise an error.) */
1094 /* Skip leading and trailing whitespace. */
1095 while (val < end && c_isspace (*val))
1097 while (val < end && c_isspace (end[-1]))
1102 if (!simple_atof (val, end, &number))
1105 *(double *)place = number * mult;
1111 cmd_cert_type (const char *com, const char *val, void *place)
1113 static const struct decode_item choices[] = {
1114 { "pem", keyfile_pem },
1115 { "der", keyfile_asn1 },
1116 { "asn1", keyfile_asn1 },
1118 int ok = decode_string (val, choices, countof (choices), place);
1120 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1125 /* Specialized helper functions, used by `commands' to handle some
1126 options specially. */
1128 static bool check_user_specified_header (const char *);
1131 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1133 if (!cmd_boolean (com, val, &opt.dirstruct))
1135 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1136 must be affected inversely. */
1138 opt.no_dirstruct = false;
1140 opt.no_dirstruct = true;
1145 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1147 /* Empty value means reset the list of headers. */
1150 free_vec (opt.user_headers);
1151 opt.user_headers = NULL;
1155 if (!check_user_specified_header (val))
1157 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1158 exec_name, com, quote (val));
1161 opt.user_headers = vec_append (opt.user_headers, val);
1166 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1168 int flag = cmd_boolean (com, val, &opt.htmlify);
1169 if (flag && !opt.htmlify)
1170 opt.remove_listing = false;
1174 /* Set the "mirror" mode. It means: recursive download, timestamping,
1175 no limit on max. recursion depth, and don't remove listings. */
1178 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1182 if (!cmd_boolean (com, val, &mirror))
1186 opt.recursive = true;
1187 if (!opt.no_dirstruct)
1188 opt.dirstruct = true;
1189 opt.timestamping = true;
1190 opt.reclevel = INFINITE_RECURSION;
1191 opt.remove_listing = false;
1196 /* Validate --prefer-family and set the choice. Allowed values are
1197 "IPv4", "IPv6", and "none". */
1200 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1202 static const struct decode_item choices[] = {
1203 { "IPv4", prefer_ipv4 },
1204 { "IPv6", prefer_ipv6 },
1205 { "none", prefer_none },
1207 int prefer_family = prefer_none;
1208 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1210 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1211 opt.prefer_family = prefer_family;
1215 /* Set progress.type to VAL, but verify that it's a valid progress
1216 implementation before that. */
1219 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1221 if (!valid_progress_implementation_p (val))
1223 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1224 exec_name, com, quote (val));
1227 xfree_null (opt.progress_type);
1229 /* Don't call set_progress_implementation here. It will be called
1230 in main() when it becomes clear what the log output is. */
1231 opt.progress_type = xstrdup (val);
1235 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1236 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1240 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1242 if (!cmd_boolean (com, val, &opt.recursive))
1246 if (opt.recursive && !opt.no_dirstruct)
1247 opt.dirstruct = true;
1253 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1255 int restrict_os = opt.restrict_files_os;
1256 int restrict_ctrl = opt.restrict_files_ctrl;
1257 int restrict_case = opt.restrict_files_case;
1261 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1265 end = strchr (val, ',');
1267 end = val + strlen (val);
1269 if (VAL_IS ("unix"))
1270 restrict_os = restrict_unix;
1271 else if (VAL_IS ("windows"))
1272 restrict_os = restrict_windows;
1273 else if (VAL_IS ("lowercase"))
1274 restrict_case = restrict_lowercase;
1275 else if (VAL_IS ("uppercase"))
1276 restrict_case = restrict_uppercase;
1277 else if (VAL_IS ("nocontrol"))
1278 restrict_ctrl = false;
1282 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1283 exec_name, com, quote (val));
1290 while (*val && *end);
1294 opt.restrict_files_os = restrict_os;
1295 opt.restrict_files_ctrl = restrict_ctrl;
1296 opt.restrict_files_case = restrict_case;
1303 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1305 static const struct decode_item choices[] = {
1306 { "auto", secure_protocol_auto },
1307 { "sslv2", secure_protocol_sslv2 },
1308 { "sslv3", secure_protocol_sslv3 },
1309 { "tlsv1", secure_protocol_tlsv1 },
1311 int ok = decode_string (val, choices, countof (choices), place);
1313 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1318 /* Set all three timeout values. */
1321 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1324 if (!cmd_time (com, val, &value))
1326 opt.read_timeout = value;
1327 opt.connect_timeout = value;
1328 opt.dns_timeout = value;
1333 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1335 /* Disallow embedded newlines. */
1336 if (strchr (val, '\n'))
1338 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1339 exec_name, com, quote (val));
1342 xfree_null (opt.useragent);
1343 opt.useragent = xstrdup (val);
1347 /* The "verbose" option cannot be cmd_boolean because the variable is
1348 not bool -- it's of type int (-1 means uninitialized because of
1349 some random hackery for disallowing -q -v). */
1352 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1355 if (cmd_boolean (com, val, &flag))
1363 /* Miscellaneous useful routines. */
1365 /* A very simple atoi clone, more useful than atoi because it works on
1366 delimited strings, and has error reportage. Returns true on success,
1367 false on failure. If successful, stores result to *DEST. */
1370 simple_atoi (const char *beg, const char *end, int *dest)
1373 bool negative = false;
1374 const char *p = beg;
1376 while (p < end && c_isspace (*p))
1378 if (p < end && (*p == '-' || *p == '+'))
1380 negative = (*p == '-');
1386 /* Read negative numbers in a separate loop because the most
1387 negative integer cannot be represented as a positive number. */
1390 for (; p < end && c_isdigit (*p); p++)
1392 int next = (10 * result) + (*p - '0');
1394 return false; /* overflow */
1398 for (; p < end && c_isdigit (*p); p++)
1400 int next = (10 * result) - (*p - '0');
1402 return false; /* underflow */
1413 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1414 doesn't handle exponential notation. Returns true on success,
1415 false on failure. In case of success, stores its result to
1419 simple_atof (const char *beg, const char *end, double *dest)
1423 bool negative = false;
1424 bool seen_dot = false;
1425 bool seen_digit = false;
1428 const char *p = beg;
1430 while (p < end && c_isspace (*p))
1432 if (p < end && (*p == '-' || *p == '+'))
1434 negative = (*p == '-');
1438 for (; p < end; p++)
1444 result = (10 * result) + (ch - '0');
1446 result += (ch - '0') / (divider *= 10);
1468 /* Verify that the user-specified header in S is valid. It must
1469 contain a colon preceded by non-white-space characters and must not
1470 contain newlines. */
1473 check_user_specified_header (const char *s)
1477 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1479 /* The header MUST contain `:' preceded by at least one
1480 non-whitespace character. */
1481 if (*p != ':' || p == s)
1483 /* The header MUST NOT contain newlines. */
1484 if (strchr (s, '\n'))
1489 /* Decode VAL into a number, according to ITEMS. */
1492 decode_string (const char *val, const struct decode_item *items, int itemcount,
1496 for (i = 0; i < itemcount; i++)
1497 if (0 == strcasecmp (val, items[i].name))
1499 *place = items[i].code;
1506 void cleanup_html_url (void);
1509 /* Free the memory allocated by global variables. */
1513 /* Free external resources, close files, etc. */
1516 fclose (output_stream);
1517 /* No need to check for error because Wget flushes its output (and
1518 checks for errors) after any data arrives. */
1520 /* We're exiting anyway so there's no real need to call free()
1521 hundreds of times. Skipping the frees will make Wget exit
1524 However, when detecting leaks, it's crucial to free() everything
1525 because then you can find the real leaks, i.e. the allocated
1526 memory which grows with the size of the program. */
1532 cleanup_html_url ();
1537 extern acc_t *netrc_list;
1538 free_netrc (netrc_list);
1540 xfree_null (opt.lfilename);
1541 xfree_null (opt.dir_prefix);
1542 xfree_null (opt.input_filename);
1543 xfree_null (opt.output_document);
1544 free_vec (opt.accepts);
1545 free_vec (opt.rejects);
1546 free_vec (opt.excludes);
1547 free_vec (opt.includes);
1548 free_vec (opt.domains);
1549 free_vec (opt.follow_tags);
1550 free_vec (opt.ignore_tags);
1551 xfree_null (opt.progress_type);
1552 xfree_null (opt.ftp_user);
1553 xfree_null (opt.ftp_passwd);
1554 xfree_null (opt.ftp_proxy);
1555 xfree_null (opt.https_proxy);
1556 xfree_null (opt.http_proxy);
1557 free_vec (opt.no_proxy);
1558 xfree_null (opt.useragent);
1559 xfree_null (opt.referer);
1560 xfree_null (opt.http_user);
1561 xfree_null (opt.http_passwd);
1562 free_vec (opt.user_headers);
1564 xfree_null (opt.cert_file);
1565 xfree_null (opt.private_key);
1566 xfree_null (opt.ca_directory);
1567 xfree_null (opt.ca_cert);
1568 xfree_null (opt.random_file);
1569 xfree_null (opt.egd_file);
1571 xfree_null (opt.bind_address);
1572 xfree_null (opt.cookies_input);
1573 xfree_null (opt.cookies_output);
1574 xfree_null (opt.user);
1575 xfree_null (opt.passwd);
1576 xfree_null (opt.base_href);
1578 #endif /* DEBUG_MALLOC */
1581 /* Unit testing routines. */
1586 test_commands_sorted()
1588 int prev_idx = 0, next_idx = 1;
1589 int command_count = countof (commands) - 1;
1591 while (next_idx <= command_count)
1593 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1596 mu_assert ("FAILED", false);
1609 test_cmd_spec_restrict_file_names()
1614 int expected_restrict_files_os;
1615 int expected_restrict_files_ctrl;
1616 int expected_restrict_files_case;
1619 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1620 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1621 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1622 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1625 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1630 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1633 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1634 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1635 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1636 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1638 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1639 res == test_array[i].result
1640 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1641 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1642 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1648 #endif /* TESTING */