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 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186 { "limitrate", &opt.limit_rate, cmd_bytes },
187 { "loadcookies", &opt.cookies_input, cmd_file },
188 { "logfile", &opt.lfilename, cmd_file },
189 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
190 { "maxredirect", &opt.max_redirect, cmd_number },
191 { "mirror", NULL, cmd_spec_mirror },
192 { "netrc", &opt.netrc, cmd_boolean },
193 { "noclobber", &opt.noclobber, cmd_boolean },
194 { "noparent", &opt.no_parent, cmd_boolean },
195 { "noproxy", &opt.no_proxy, cmd_vector },
196 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
197 { "outputdocument", &opt.output_document, cmd_file },
198 { "pagerequisites", &opt.page_requisites, cmd_boolean },
199 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
200 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
201 { "password", &opt.passwd, cmd_string },
202 { "postdata", &opt.post_data, cmd_string },
203 { "postfile", &opt.post_file_name, cmd_file },
204 { "preferfamily", NULL, cmd_spec_prefer_family },
205 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
207 { "privatekey", &opt.private_key, cmd_file },
208 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
210 { "progress", &opt.progress_type, cmd_spec_progress },
211 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
212 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
213 { "proxypassword", &opt.proxy_passwd, cmd_string },
214 { "proxyuser", &opt.proxy_user, cmd_string },
215 { "quiet", &opt.quiet, cmd_boolean },
216 { "quota", &opt.quota, cmd_bytes_sum },
218 { "randomfile", &opt.random_file, cmd_file },
220 { "randomwait", &opt.random_wait, cmd_boolean },
221 { "readtimeout", &opt.read_timeout, cmd_time },
222 { "reclevel", &opt.reclevel, cmd_number_inf },
223 { "recursive", NULL, cmd_spec_recursive },
224 { "referer", &opt.referer, cmd_string },
225 { "reject", &opt.rejects, cmd_vector },
226 { "relativeonly", &opt.relative_only, cmd_boolean },
227 { "removelisting", &opt.remove_listing, cmd_boolean },
228 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
229 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
230 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
231 { "robots", &opt.use_robots, cmd_boolean },
232 { "savecookies", &opt.cookies_output, cmd_file },
233 { "saveheaders", &opt.save_headers, cmd_boolean },
235 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
237 { "serverresponse", &opt.server_response, cmd_boolean },
238 { "spanhosts", &opt.spanhost, cmd_boolean },
239 { "spider", &opt.spider, cmd_boolean },
240 { "strictcomments", &opt.strict_comments, cmd_boolean },
241 { "timeout", NULL, cmd_spec_timeout },
242 { "timestamping", &opt.timestamping, cmd_boolean },
243 { "tries", &opt.ntry, cmd_number_inf },
244 { "useproxy", &opt.use_proxy, cmd_boolean },
245 { "user", &opt.user, cmd_string },
246 { "useragent", NULL, cmd_spec_useragent },
247 { "verbose", NULL, cmd_spec_verbose },
248 { "wait", &opt.wait, cmd_time },
249 { "waitretry", &opt.waitretry, cmd_time },
251 { "wdebug", &opt.wdebug, cmd_boolean },
255 /* Look up CMDNAME in the commands[] and return its position in the
256 array. If CMDNAME is not found, return -1. */
259 command_by_name (const char *cmdname)
261 /* Use binary search for speed. Wget has ~100 commands, which
262 guarantees a worst case performance of 7 string comparisons. */
263 int lo = 0, hi = countof (commands) - 1;
267 int mid = (lo + hi) >> 1;
268 int cmp = strcasecmp (cmdname, commands[mid].name);
279 /* Reset the variables to default values. */
285 /* Most of the default values are 0 (and 0.0, NULL, and false).
286 Just reset everything, and fill in the non-zero values. Note
287 that initializing pointers to NULL this way is technically
288 illegal, but porting Wget to a machine where NULL is not all-zero
289 bit pattern will be the least of the implementors' worries. */
296 opt.add_hostdir = true;
300 opt.http_keep_alive = true;
301 opt.use_proxy = true;
302 tmp = getenv ("no_proxy");
304 opt.no_proxy = sepstring (tmp);
305 opt.prefer_family = prefer_none;
306 opt.allow_cache = true;
308 opt.read_timeout = 900;
309 opt.use_robots = true;
311 opt.remove_listing = true;
313 opt.dot_bytes = 1024;
314 opt.dot_spacing = 10;
315 opt.dots_in_line = 50;
317 opt.dns_cache = true;
321 opt.check_cert = true;
324 /* The default for file name restriction defaults to the OS type. */
325 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
326 opt.restrict_files_os = restrict_windows;
328 opt.restrict_files_os = restrict_unix;
330 opt.restrict_files_ctrl = true;
331 opt.restrict_files_case = restrict_no_case_restriction;
333 opt.max_redirect = 20;
336 /* Return the user's home directory (strdup-ed), or NULL if none is
341 static char buf[PATH_MAX];
346 home = getenv ("HOME");
350 /* Under MSDOS, if $HOME isn't defined, use the directory where
351 `wget.exe' resides. */
352 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
355 strcpy (buf, _w32_get_argv0 ());
356 p = strrchr (buf, '/'); /* djgpp */
358 p = strrchr (buf, '\\'); /* others */
362 #elif !defined(WINDOWS)
363 /* If HOME is not defined, try getting it from the password
365 struct passwd *pwd = getpwuid (getuid ());
366 if (!pwd || !pwd->pw_dir)
368 strcpy (buf, pwd->pw_dir);
371 /* Under Windows, if $HOME isn't defined, use the directory where
372 `wget.exe' resides. */
378 return home ? xstrdup (home) : NULL;
381 /* Check the 'WGETRC' environment variable and return the file name
382 if 'WGETRC' is set and is a valid file.
383 If the `WGETRC' variable exists but the file does not exist, the
384 function will exit(). */
386 wgetrc_env_file_name (void)
388 char *env = getenv ("WGETRC");
391 if (!file_exists_p (env))
393 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
397 return xstrdup (env);
402 /* Check for the existance of '$HOME/.wgetrc' and return it's path
403 if it exists and is set. */
405 wgetrc_user_file_name (void)
407 char *home = home_dir ();
410 file = aprintf ("%s/.wgetrc", home);
414 if (!file_exists_p (file))
422 /* Return the path to the user's .wgetrc. This is either the value of
423 `WGETRC' environment variable, or `$HOME/.wgetrc'.
425 Additionally, for windows, look in the directory where wget.exe
428 wgetrc_file_name (void)
431 char *file = wgetrc_env_file_name ();
435 file = wgetrc_user_file_name ();
438 /* Under Windows, if we still haven't found .wgetrc, look for the file
439 `wget.ini' in the directory where `wget.exe' resides; we do this for
440 backward compatibility with previous versions of Wget.
441 SYSTEM_WGETRC should not be defined under WINDOWS. */
443 if (!file || !file_exists_p (file))
449 file = aprintf ("%s/wget.ini", home);
456 if (!file_exists_p (file))
464 /* Return values of parse_line. */
472 static enum parse_line parse_line (const char *, char **, char **, int *);
473 static bool setval_internal (int, const char *, const char *);
475 /* Initialize variables from a wgetrc file. Returns zero (failure) if
476 there were errors in the file. */
479 run_wgetrc (const char *file)
486 fp = fopen (file, "rb");
489 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
490 file, strerror (errno));
491 return true; /* not a fatal error */
493 enable_tilde_expansion = true;
495 while ((line = read_whole_line (fp)) != NULL)
497 char *com = NULL, *val = NULL;
500 /* Parse the line. */
501 switch (parse_line (line, &com, &val, &comind))
504 /* If everything is OK, set the value. */
505 if (!setval_internal (comind, com, val))
507 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
508 exec_name, file, ln);
512 case line_syntax_error:
513 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
514 exec_name, file, ln);
517 case line_unknown_command:
518 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
519 exec_name, quote (com), file, ln);
532 enable_tilde_expansion = false;
538 /* Initialize the defaults and run the system wgetrc and user's own
546 /* Load the hard-coded defaults. */
549 /* If SYSTEM_WGETRC is defined, use it. */
551 if (file_exists_p (SYSTEM_WGETRC))
552 ok &= run_wgetrc (SYSTEM_WGETRC);
554 /* Override it with your own, if one exists. */
555 file = wgetrc_file_name ();
558 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
559 something like realpath() before comparing them with `strcmp' */
561 if (!strcmp (file, SYSTEM_WGETRC))
563 fprintf (stderr, _("\
564 %s: Warning: Both system and user wgetrc point to %s.\n"),
565 exec_name, quote (file));
569 ok &= run_wgetrc (file);
571 /* If there were errors processing either `.wgetrc', abort. */
579 /* Remove dashes and underscores from S, modifying S in the
585 char *t = s; /* t - tortoise */
586 char *h = s; /* h - hare */
588 if (*h == '_' || *h == '-')
595 /* Parse the line pointed by line, with the syntax:
596 <sp>* command <sp>* = <sp>* value <sp>*
597 Uses malloc to allocate space for command and value.
599 Returns one of line_ok, line_empty, line_syntax_error, or
600 line_unknown_command.
602 In case of line_ok, *COM and *VAL point to freshly allocated
603 strings, and *COMIND points to com's index. In case of error or
604 empty line, their values are unmodified. */
606 static enum parse_line
607 parse_line (const char *line, char **com, char **val, int *comind)
610 const char *end = line + strlen (line);
611 const char *cmdstart, *cmdend;
612 const char *valstart, *valend;
617 /* Skip leading and trailing whitespace. */
618 while (*line && c_isspace (*line))
620 while (end > line && c_isspace (end[-1]))
623 /* Skip empty lines and comments. */
624 if (!*line || *line == '#')
630 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
634 /* Skip '=', as well as any space before or after it. */
635 while (p < end && c_isspace (*p))
637 if (p == end || *p != '=')
638 return line_syntax_error;
640 while (p < end && c_isspace (*p))
646 /* The syntax is valid (even though the command might not be). Fill
647 in the command name and value. */
648 *com = strdupdelim (cmdstart, cmdend);
649 *val = strdupdelim (valstart, valend);
651 /* The line now known to be syntactically correct. Check whether
652 the command is valid. */
653 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
655 ind = command_by_name (cmdcopy);
657 return line_unknown_command;
659 /* Report success to the caller. */
664 /* Run commands[comind].action. */
667 setval_internal (int comind, const char *com, const char *val)
669 assert (0 <= comind && ((size_t) comind) < countof (commands));
670 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
671 return commands[comind].action (com, val, commands[comind].place);
674 /* Run command COM with value VAL. If running the command produces an
675 error, report the error and exit.
677 This is intended to be called from main() to modify Wget's behavior
678 through command-line switches. Since COM is hard-coded in main(),
679 it is not canonicalized, and this aborts when COM is not found.
681 If COMIND's are exported to init.h, this function will be changed
682 to accept COMIND directly. */
685 setoptval (const char *com, const char *val, const char *optname)
687 /* Prepend "--" to OPTNAME. */
688 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
691 strcpy (dd_optname + 2, optname);
693 assert (val != NULL);
694 if (!setval_internal (command_by_name (com), dd_optname, val))
698 /* Parse OPT into command and value and run it. For example,
699 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
700 This is used by the `--execute' flag in main.c. */
703 run_command (const char *opt)
707 switch (parse_line (opt, &com, &val, &comind))
710 if (!setval_internal (comind, com, val))
716 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
717 exec_name, quote (opt));
722 /* Generic helper functions, for use with `commands'. */
724 /* Forward declarations: */
729 static bool decode_string (const char *, const struct decode_item *, int, int *);
730 static bool simple_atoi (const char *, const char *, int *);
731 static bool simple_atof (const char *, const char *, double *);
733 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
735 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
736 && c_tolower((p)[1]) == (c1) \
739 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
740 && c_tolower((p)[1]) == (c1) \
741 && c_tolower((p)[2]) == (c2) \
745 /* Store the boolean value from VAL to PLACE. COM is ignored,
746 except for error messages. */
748 cmd_boolean (const char *com, const char *val, void *place)
752 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
753 /* "on", "yes" and "1" mean true. */
755 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
756 /* "off", "no" and "0" mean false. */
761 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
762 exec_name, com, quote (val));
766 *(bool *) place = value;
770 /* Set the non-negative integer value from VAL to PLACE. With
771 incorrect specification, the number remains unchanged. */
773 cmd_number (const char *com, const char *val, void *place)
775 if (!simple_atoi (val, val + strlen (val), place)
776 || *(int *) place < 0)
778 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
779 exec_name, com, quote (val));
785 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
787 cmd_number_inf (const char *com, const char *val, void *place)
789 if (!strcasecmp (val, "inf"))
794 return cmd_number (com, val, place);
797 /* Copy (strdup) the string at COM to a new location and place a
798 pointer to *PLACE. */
800 cmd_string (const char *com, const char *val, void *place)
802 char **pstring = (char **)place;
804 xfree_null (*pstring);
805 *pstring = xstrdup (val);
809 #if defined(WINDOWS) || defined(MSDOS)
810 # define ISSEP(c) ((c) == '/' || (c) == '\\')
812 # define ISSEP(c) ((c) == '/')
815 /* Like the above, but handles tilde-expansion when reading a user's
816 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
817 gets expanded to the user's home directory. */
819 cmd_file (const char *com, const char *val, void *place)
821 char **pstring = (char **)place;
823 xfree_null (*pstring);
825 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
827 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
830 *pstring = xstrdup (val);
835 char *home = home_dir ();
839 homelen = strlen (home);
840 while (homelen && ISSEP (home[homelen - 1]))
841 home[--homelen] = '\0';
843 /* Skip the leading "~/". */
844 for (++val; ISSEP (*val); val++)
847 *pstring = concat_strings (home, "/", val, (char *) 0);
850 #if defined(WINDOWS) || defined(MSDOS)
851 /* Convert "\" to "/". */
854 for (s = *pstring; *s; s++)
862 /* Like cmd_file, but strips trailing '/' characters. */
864 cmd_directory (const char *com, const char *val, void *place)
868 /* Call cmd_file() for tilde expansion and separator
869 canonicalization (backslash -> slash under Windows). These
870 things should perhaps be in a separate function. */
871 if (!cmd_file (com, val, place))
876 while (t > s && *--t == '/')
882 /* Split VAL by space to a vector of values, and append those values
883 to vector pointed to by the PLACE argument. If VAL is empty, the
884 PLACE vector is cleared instead. */
887 cmd_vector (const char *com, const char *val, void *place)
889 char ***pvec = (char ***)place;
892 *pvec = merge_vecs (*pvec, sepstring (val));
902 cmd_directory_vector (const char *com, const char *val, void *place)
904 char ***pvec = (char ***)place;
908 /* Strip the trailing slashes from directories. */
911 seps = sepstring (val);
912 for (t = seps; t && *t; t++)
914 int len = strlen (*t);
915 /* Skip degenerate case of root directory. */
918 if ((*t)[len - 1] == '/')
919 (*t)[len - 1] = '\0';
922 *pvec = merge_vecs (*pvec, seps);
932 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
933 "100k" or "2.5G" to a floating point number. */
936 parse_bytes_helper (const char *val, double *result)
939 const char *end = val + strlen (val);
941 /* Check for "inf". */
942 if (0 == strcmp (val, "inf"))
948 /* Strip trailing whitespace. */
949 while (val < end && c_isspace (end[-1]))
954 switch (c_tolower (end[-1]))
957 --end, mult = 1024.0;
960 --end, mult = 1048576.0;
963 --end, mult = 1073741824.0;
966 --end, mult = 1099511627776.0;
969 /* Not a recognized suffix: assume it's a digit. (If not,
970 simple_atof will raise an error.) */
974 /* Skip leading and trailing whitespace. */
975 while (val < end && c_isspace (*val))
977 while (val < end && c_isspace (end[-1]))
982 if (!simple_atof (val, end, &number) || number < 0)
985 *result = number * mult;
989 /* Parse VAL as a number and set its value to PLACE (which should
992 By default, the value is assumed to be in bytes. If "K", "M", or
993 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
994 1<<30, respectively. Floating point values are allowed and are
995 cast to integer before use. The idea is to be able to use things
996 like 1.5k instead of "1536".
998 The string "inf" is returned as 0.
1000 In case of error, false is returned and memory pointed to by PLACE
1001 remains unmodified. */
1004 cmd_bytes (const char *com, const char *val, void *place)
1007 if (!parse_bytes_helper (val, &byte_value))
1009 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1010 exec_name, com, quote (val));
1013 *(wgint *)place = (wgint)byte_value;
1017 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1018 SIZE_SUM. It works by converting the string to double, therefore
1019 working with values up to 2^53-1 without loss of precision. This
1020 value (8192 TB) is large enough to serve for a while. */
1023 cmd_bytes_sum (const char *com, const char *val, void *place)
1026 if (!parse_bytes_helper (val, &byte_value))
1028 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1029 exec_name, com, quote (val));
1032 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1036 /* Store the value of VAL to *OUT. The value is a time period, by
1037 default expressed in seconds, but also accepting suffixes "m", "h",
1038 "d", and "w" for minutes, hours, days, and weeks respectively. */
1041 cmd_time (const char *com, const char *val, void *place)
1043 double number, mult;
1044 const char *end = val + strlen (val);
1046 /* Strip trailing whitespace. */
1047 while (val < end && c_isspace (end[-1]))
1053 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1054 exec_name, com, quote (val));
1058 switch (c_tolower (end[-1]))
1061 --end, mult = 1; /* seconds */
1064 --end, mult = 60; /* minutes */
1067 --end, mult = 3600; /* hours */
1070 --end, mult = 86400.0; /* days */
1073 --end, mult = 604800.0; /* weeks */
1076 /* Not a recognized suffix: assume it belongs to the number.
1077 (If not, simple_atof will raise an error.) */
1081 /* Skip leading and trailing whitespace. */
1082 while (val < end && c_isspace (*val))
1084 while (val < end && c_isspace (end[-1]))
1089 if (!simple_atof (val, end, &number))
1092 *(double *)place = number * mult;
1098 cmd_cert_type (const char *com, const char *val, void *place)
1100 static const struct decode_item choices[] = {
1101 { "pem", keyfile_pem },
1102 { "der", keyfile_asn1 },
1103 { "asn1", keyfile_asn1 },
1105 int ok = decode_string (val, choices, countof (choices), place);
1107 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1112 /* Specialized helper functions, used by `commands' to handle some
1113 options specially. */
1115 static bool check_user_specified_header (const char *);
1118 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1120 if (!cmd_boolean (com, val, &opt.dirstruct))
1122 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1123 must be affected inversely. */
1125 opt.no_dirstruct = false;
1127 opt.no_dirstruct = true;
1132 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1134 /* Empty value means reset the list of headers. */
1137 free_vec (opt.user_headers);
1138 opt.user_headers = NULL;
1142 if (!check_user_specified_header (val))
1144 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1145 exec_name, com, quote (val));
1148 opt.user_headers = vec_append (opt.user_headers, val);
1153 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1155 int flag = cmd_boolean (com, val, &opt.htmlify);
1156 if (flag && !opt.htmlify)
1157 opt.remove_listing = false;
1161 /* Set the "mirror" mode. It means: recursive download, timestamping,
1162 no limit on max. recursion depth, and don't remove listings. */
1165 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1169 if (!cmd_boolean (com, val, &mirror))
1173 opt.recursive = true;
1174 if (!opt.no_dirstruct)
1175 opt.dirstruct = true;
1176 opt.timestamping = true;
1177 opt.reclevel = INFINITE_RECURSION;
1178 opt.remove_listing = false;
1183 /* Validate --prefer-family and set the choice. Allowed values are
1184 "IPv4", "IPv6", and "none". */
1187 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1189 static const struct decode_item choices[] = {
1190 { "IPv4", prefer_ipv4 },
1191 { "IPv6", prefer_ipv6 },
1192 { "none", prefer_none },
1194 int prefer_family = prefer_none;
1195 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1197 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1198 opt.prefer_family = prefer_family;
1202 /* Set progress.type to VAL, but verify that it's a valid progress
1203 implementation before that. */
1206 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1208 if (!valid_progress_implementation_p (val))
1210 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1211 exec_name, com, quote (val));
1214 xfree_null (opt.progress_type);
1216 /* Don't call set_progress_implementation here. It will be called
1217 in main() when it becomes clear what the log output is. */
1218 opt.progress_type = xstrdup (val);
1222 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1223 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1227 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1229 if (!cmd_boolean (com, val, &opt.recursive))
1233 if (opt.recursive && !opt.no_dirstruct)
1234 opt.dirstruct = true;
1240 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1242 int restrict_os = opt.restrict_files_os;
1243 int restrict_ctrl = opt.restrict_files_ctrl;
1244 int restrict_case = opt.restrict_files_case;
1248 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1252 end = strchr (val, ',');
1254 end = val + strlen (val);
1256 if (VAL_IS ("unix"))
1257 restrict_os = restrict_unix;
1258 else if (VAL_IS ("windows"))
1259 restrict_os = restrict_windows;
1260 else if (VAL_IS ("lowercase"))
1261 restrict_case = restrict_lowercase;
1262 else if (VAL_IS ("uppercase"))
1263 restrict_case = restrict_uppercase;
1264 else if (VAL_IS ("nocontrol"))
1265 restrict_ctrl = false;
1269 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1270 exec_name, com, quote (val));
1277 while (*val && *end);
1281 opt.restrict_files_os = restrict_os;
1282 opt.restrict_files_ctrl = restrict_ctrl;
1283 opt.restrict_files_case = restrict_case;
1290 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1292 static const struct decode_item choices[] = {
1293 { "auto", secure_protocol_auto },
1294 { "sslv2", secure_protocol_sslv2 },
1295 { "sslv3", secure_protocol_sslv3 },
1296 { "tlsv1", secure_protocol_tlsv1 },
1298 int ok = decode_string (val, choices, countof (choices), place);
1300 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1305 /* Set all three timeout values. */
1308 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1311 if (!cmd_time (com, val, &value))
1313 opt.read_timeout = value;
1314 opt.connect_timeout = value;
1315 opt.dns_timeout = value;
1320 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1322 /* Disallow embedded newlines. */
1323 if (strchr (val, '\n'))
1325 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1326 exec_name, com, quote (val));
1329 xfree_null (opt.useragent);
1330 opt.useragent = xstrdup (val);
1334 /* The "verbose" option cannot be cmd_boolean because the variable is
1335 not bool -- it's of type int (-1 means uninitialized because of
1336 some random hackery for disallowing -q -v). */
1339 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1342 if (cmd_boolean (com, val, &flag))
1350 /* Miscellaneous useful routines. */
1352 /* A very simple atoi clone, more useful than atoi because it works on
1353 delimited strings, and has error reportage. Returns true on success,
1354 false on failure. If successful, stores result to *DEST. */
1357 simple_atoi (const char *beg, const char *end, int *dest)
1360 bool negative = false;
1361 const char *p = beg;
1363 while (p < end && c_isspace (*p))
1365 if (p < end && (*p == '-' || *p == '+'))
1367 negative = (*p == '-');
1373 /* Read negative numbers in a separate loop because the most
1374 negative integer cannot be represented as a positive number. */
1377 for (; p < end && c_isdigit (*p); p++)
1379 int next = (10 * result) + (*p - '0');
1381 return false; /* overflow */
1385 for (; p < end && c_isdigit (*p); p++)
1387 int next = (10 * result) - (*p - '0');
1389 return false; /* underflow */
1400 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1401 doesn't handle exponential notation. Returns true on success,
1402 false on failure. In case of success, stores its result to
1406 simple_atof (const char *beg, const char *end, double *dest)
1410 bool negative = false;
1411 bool seen_dot = false;
1412 bool seen_digit = false;
1415 const char *p = beg;
1417 while (p < end && c_isspace (*p))
1419 if (p < end && (*p == '-' || *p == '+'))
1421 negative = (*p == '-');
1425 for (; p < end; p++)
1431 result = (10 * result) + (ch - '0');
1433 result += (ch - '0') / (divider *= 10);
1455 /* Verify that the user-specified header in S is valid. It must
1456 contain a colon preceded by non-white-space characters and must not
1457 contain newlines. */
1460 check_user_specified_header (const char *s)
1464 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1466 /* The header MUST contain `:' preceded by at least one
1467 non-whitespace character. */
1468 if (*p != ':' || p == s)
1470 /* The header MUST NOT contain newlines. */
1471 if (strchr (s, '\n'))
1476 /* Decode VAL into a number, according to ITEMS. */
1479 decode_string (const char *val, const struct decode_item *items, int itemcount,
1483 for (i = 0; i < itemcount; i++)
1484 if (0 == strcasecmp (val, items[i].name))
1486 *place = items[i].code;
1493 void cleanup_html_url (void);
1496 /* Free the memory allocated by global variables. */
1500 /* Free external resources, close files, etc. */
1503 fclose (output_stream);
1504 /* No need to check for error because Wget flushes its output (and
1505 checks for errors) after any data arrives. */
1507 /* We're exiting anyway so there's no real need to call free()
1508 hundreds of times. Skipping the frees will make Wget exit
1511 However, when detecting leaks, it's crucial to free() everything
1512 because then you can find the real leaks, i.e. the allocated
1513 memory which grows with the size of the program. */
1519 cleanup_html_url ();
1524 extern acc_t *netrc_list;
1525 free_netrc (netrc_list);
1527 xfree_null (opt.lfilename);
1528 xfree_null (opt.dir_prefix);
1529 xfree_null (opt.input_filename);
1530 xfree_null (opt.output_document);
1531 free_vec (opt.accepts);
1532 free_vec (opt.rejects);
1533 free_vec (opt.excludes);
1534 free_vec (opt.includes);
1535 free_vec (opt.domains);
1536 free_vec (opt.follow_tags);
1537 free_vec (opt.ignore_tags);
1538 xfree_null (opt.progress_type);
1539 xfree_null (opt.ftp_user);
1540 xfree_null (opt.ftp_passwd);
1541 xfree_null (opt.ftp_proxy);
1542 xfree_null (opt.https_proxy);
1543 xfree_null (opt.http_proxy);
1544 free_vec (opt.no_proxy);
1545 xfree_null (opt.useragent);
1546 xfree_null (opt.referer);
1547 xfree_null (opt.http_user);
1548 xfree_null (opt.http_passwd);
1549 free_vec (opt.user_headers);
1551 xfree_null (opt.cert_file);
1552 xfree_null (opt.private_key);
1553 xfree_null (opt.ca_directory);
1554 xfree_null (opt.ca_cert);
1555 xfree_null (opt.random_file);
1556 xfree_null (opt.egd_file);
1558 xfree_null (opt.bind_address);
1559 xfree_null (opt.cookies_input);
1560 xfree_null (opt.cookies_output);
1561 xfree_null (opt.user);
1562 xfree_null (opt.passwd);
1563 xfree_null (opt.base_href);
1565 #endif /* DEBUG_MALLOC */
1568 /* Unit testing routines. */
1573 test_commands_sorted()
1575 int prev_idx = 0, next_idx = 1;
1576 int command_count = countof (commands) - 1;
1578 while (next_idx <= command_count)
1580 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1583 mu_assert ("FAILED", false);
1596 test_cmd_spec_restrict_file_names()
1601 int expected_restrict_files_os;
1602 int expected_restrict_files_ctrl;
1603 int expected_restrict_files_case;
1606 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1607 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1608 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1609 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1612 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1617 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1620 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1621 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1622 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1623 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1625 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1626 res == test_array[i].result
1627 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1628 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1629 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1635 #endif /* TESTING */