1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
61 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
62 otherwise, it will be performed by the shell. This variable will
63 be set by the wgetrc-reading function. */
65 static bool enable_tilde_expansion;
68 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
70 CMD_DECLARE (cmd_boolean);
71 CMD_DECLARE (cmd_bytes);
72 CMD_DECLARE (cmd_bytes_sum);
74 CMD_DECLARE (cmd_cert_type);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_number);
78 CMD_DECLARE (cmd_number_inf);
79 CMD_DECLARE (cmd_string);
80 CMD_DECLARE (cmd_file);
81 CMD_DECLARE (cmd_directory);
82 CMD_DECLARE (cmd_time);
83 CMD_DECLARE (cmd_vector);
85 CMD_DECLARE (cmd_spec_dirstruct);
86 CMD_DECLARE (cmd_spec_header);
87 CMD_DECLARE (cmd_spec_htmlify);
88 CMD_DECLARE (cmd_spec_mirror);
89 CMD_DECLARE (cmd_spec_prefer_family);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_secure_protocol);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
98 CMD_DECLARE (cmd_spec_verbose);
100 /* List of recognized commands, each consisting of name, place and
101 function. When adding a new command, simply add it to the list,
102 but be sure to keep the list sorted alphabetically, as
103 command_by_name's binary search depends on it. Also, be sure to
104 add any entries that allocate memory (e.g. cmd_string and
105 cmd_vector) to the cleanup() function below. */
107 static const struct {
110 bool (*action) (const char *, const char *, void *);
112 /* KEEP THIS LIST ALPHABETICALLY SORTED */
113 { "accept", &opt.accepts, cmd_vector },
114 { "addhostdir", &opt.add_hostdir, cmd_boolean },
115 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
116 { "askpassword", &opt.ask_passwd, cmd_boolean },
117 { "authnochallenge", &opt.auth_without_challenge,
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "contentdisposition", &opt.content_disposition, cmd_boolean },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "deleteafter", &opt.delete_after, cmd_boolean },
144 { "dirprefix", &opt.dir_prefix, cmd_directory },
145 { "dirstruct", NULL, cmd_spec_dirstruct },
146 { "dnscache", &opt.dns_cache, cmd_boolean },
147 { "dnstimeout", &opt.dns_timeout, cmd_time },
148 { "domains", &opt.domains, cmd_vector },
149 { "dotbytes", &opt.dot_bytes, cmd_bytes },
150 { "dotsinline", &opt.dots_in_line, cmd_number },
151 { "dotspacing", &opt.dot_spacing, cmd_number },
152 { "dotstyle", &opt.dot_style, cmd_string },
154 { "egdfile", &opt.egd_file, cmd_file },
156 { "excludedirectories", &opt.excludes, cmd_directory_vector },
157 { "excludedomains", &opt.exclude_domains, cmd_vector },
158 { "followftp", &opt.follow_ftp, cmd_boolean },
159 { "followtags", &opt.follow_tags, cmd_vector },
160 { "forcehtml", &opt.force_html, cmd_boolean },
161 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
162 { "ftppassword", &opt.ftp_passwd, cmd_string },
163 { "ftpproxy", &opt.ftp_proxy, cmd_string },
164 { "ftpuser", &opt.ftp_user, cmd_string },
165 { "glob", &opt.ftp_glob, cmd_boolean },
166 { "header", NULL, cmd_spec_header },
167 { "htmlextension", &opt.html_extension, cmd_boolean },
168 { "htmlify", NULL, cmd_spec_htmlify },
169 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
170 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
171 { "httppassword", &opt.http_passwd, cmd_string },
172 { "httpproxy", &opt.http_proxy, cmd_string },
173 { "httpsproxy", &opt.https_proxy, cmd_string },
174 { "httpuser", &opt.http_user, cmd_string },
175 { "ignorecase", &opt.ignore_case, cmd_boolean },
176 { "ignorelength", &opt.ignore_length, cmd_boolean },
177 { "ignoretags", &opt.ignore_tags, cmd_vector },
178 { "includedirectories", &opt.includes, cmd_directory_vector },
180 { "inet4only", &opt.ipv4_only, cmd_boolean },
181 { "inet6only", &opt.ipv6_only, cmd_boolean },
183 { "input", &opt.input_filename, cmd_file },
184 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
185 { "limitrate", &opt.limit_rate, cmd_bytes },
186 { "loadcookies", &opt.cookies_input, cmd_file },
187 { "logfile", &opt.lfilename, cmd_file },
188 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
189 { "maxredirect", &opt.max_redirect, cmd_number },
190 { "mirror", NULL, cmd_spec_mirror },
191 { "netrc", &opt.netrc, cmd_boolean },
192 { "noclobber", &opt.noclobber, cmd_boolean },
193 { "noparent", &opt.no_parent, cmd_boolean },
194 { "noproxy", &opt.no_proxy, cmd_vector },
195 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
196 { "outputdocument", &opt.output_document, cmd_file },
197 { "pagerequisites", &opt.page_requisites, cmd_boolean },
198 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
199 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
200 { "password", &opt.passwd, cmd_string },
201 { "postdata", &opt.post_data, cmd_string },
202 { "postfile", &opt.post_file_name, cmd_file },
203 { "preferfamily", NULL, cmd_spec_prefer_family },
204 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
206 { "privatekey", &opt.private_key, cmd_file },
207 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
209 { "progress", &opt.progress_type, cmd_spec_progress },
210 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
211 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
212 { "proxypassword", &opt.proxy_passwd, cmd_string },
213 { "proxyuser", &opt.proxy_user, cmd_string },
214 { "quiet", &opt.quiet, cmd_boolean },
215 { "quota", &opt.quota, cmd_bytes_sum },
217 { "randomfile", &opt.random_file, cmd_file },
219 { "randomwait", &opt.random_wait, cmd_boolean },
220 { "readtimeout", &opt.read_timeout, cmd_time },
221 { "reclevel", &opt.reclevel, cmd_number_inf },
222 { "recursive", NULL, cmd_spec_recursive },
223 { "referer", &opt.referer, cmd_string },
224 { "reject", &opt.rejects, cmd_vector },
225 { "relativeonly", &opt.relative_only, cmd_boolean },
226 { "removelisting", &opt.remove_listing, cmd_boolean },
227 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
228 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
229 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
230 { "robots", &opt.use_robots, cmd_boolean },
231 { "savecookies", &opt.cookies_output, cmd_file },
232 { "saveheaders", &opt.save_headers, cmd_boolean },
234 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
236 { "serverresponse", &opt.server_response, cmd_boolean },
237 { "spanhosts", &opt.spanhost, cmd_boolean },
238 { "spider", &opt.spider, cmd_boolean },
239 { "strictcomments", &opt.strict_comments, cmd_boolean },
240 { "timeout", NULL, cmd_spec_timeout },
241 { "timestamping", &opt.timestamping, cmd_boolean },
242 { "tries", &opt.ntry, cmd_number_inf },
243 { "useproxy", &opt.use_proxy, cmd_boolean },
244 { "user", &opt.user, cmd_string },
245 { "useragent", NULL, cmd_spec_useragent },
246 { "verbose", NULL, cmd_spec_verbose },
247 { "wait", &opt.wait, cmd_time },
248 { "waitretry", &opt.waitretry, cmd_time },
250 { "wdebug", &opt.wdebug, cmd_boolean },
254 /* Look up CMDNAME in the commands[] and return its position in the
255 array. If CMDNAME is not found, return -1. */
258 command_by_name (const char *cmdname)
260 /* Use binary search for speed. Wget has ~100 commands, which
261 guarantees a worst case performance of 7 string comparisons. */
262 int lo = 0, hi = countof (commands) - 1;
266 int mid = (lo + hi) >> 1;
267 int cmp = strcasecmp (cmdname, commands[mid].name);
278 /* Reset the variables to default values. */
284 /* Most of the default values are 0 (and 0.0, NULL, and false).
285 Just reset everything, and fill in the non-zero values. Note
286 that initializing pointers to NULL this way is technically
287 illegal, but porting Wget to a machine where NULL is not all-zero
288 bit pattern will be the least of the implementors' worries. */
295 opt.add_hostdir = true;
299 opt.http_keep_alive = true;
300 opt.use_proxy = true;
301 tmp = getenv ("no_proxy");
303 opt.no_proxy = sepstring (tmp);
304 opt.allow_cache = true;
306 opt.read_timeout = 900;
307 opt.use_robots = true;
309 opt.remove_listing = true;
311 opt.dot_bytes = 1024;
312 opt.dot_spacing = 10;
313 opt.dots_in_line = 50;
315 opt.dns_cache = true;
319 opt.check_cert = true;
322 /* The default for file name restriction defaults to the OS type. */
323 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
324 opt.restrict_files_os = restrict_windows;
326 opt.restrict_files_os = restrict_unix;
328 opt.restrict_files_ctrl = true;
329 opt.restrict_files_case = restrict_no_case_restriction;
331 opt.max_redirect = 20;
334 /* Return the user's home directory (strdup-ed), or NULL if none is
339 char *home = getenv ("HOME");
344 /* Under MSDOS, if $HOME isn't defined, use the directory where
345 `wget.exe' resides. */
346 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
347 char *p, buf[PATH_MAX];
349 strcpy (buf, _w32_get_argv0 ());
350 p = strrchr (buf, '/'); /* djgpp */
352 p = strrchr (buf, '\\'); /* others */
356 #elif !defined(WINDOWS)
357 /* If HOME is not defined, try getting it from the password
359 struct passwd *pwd = getpwuid (getuid ());
360 if (!pwd || !pwd->pw_dir)
364 /* Under Windows, if $HOME isn't defined, use the directory where
365 `wget.exe' resides. */
370 return home ? xstrdup (home) : NULL;
373 /* Return the path to the user's .wgetrc. This is either the value of
374 `WGETRC' environment variable, or `$HOME/.wgetrc'.
376 If the `WGETRC' variable exists but the file does not exist, the
377 function will exit(). */
379 wgetrc_file_name (void)
384 /* Try the environment. */
385 env = getenv ("WGETRC");
388 if (!file_exists_p (env))
390 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
394 return xstrdup (env);
397 /* If that failed, try $HOME/.wgetrc. */
400 file = aprintf ("%s/.wgetrc", home);
404 /* Under Windows, if we still haven't found .wgetrc, look for the file
405 `wget.ini' in the directory where `wget.exe' resides; we do this for
406 backward compatibility with previous versions of Wget.
407 SYSTEM_WGETRC should not be defined under WINDOWS. */
408 if (!file || !file_exists_p (file))
414 file = aprintf ("%s/wget.ini", home);
420 if (!file_exists_p (file))
428 /* Return values of parse_line. */
436 static enum parse_line parse_line (const char *, char **, char **, int *);
437 static bool setval_internal (int, const char *, const char *);
439 /* Initialize variables from a wgetrc file. Returns zero (failure) if
440 there were errors in the file. */
443 run_wgetrc (const char *file)
450 fp = fopen (file, "rb");
453 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
454 file, strerror (errno));
455 return true; /* not a fatal error */
457 enable_tilde_expansion = true;
459 while ((line = read_whole_line (fp)) != NULL)
461 char *com = NULL, *val = NULL;
464 /* Parse the line. */
465 switch (parse_line (line, &com, &val, &comind))
468 /* If everything is OK, set the value. */
469 if (!setval_internal (comind, com, val))
471 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
472 exec_name, file, ln);
476 case line_syntax_error:
477 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
478 exec_name, file, ln);
481 case line_unknown_command:
482 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
483 exec_name, quote (com), file, ln);
496 enable_tilde_expansion = false;
502 /* Initialize the defaults and run the system wgetrc and user's own
510 /* Load the hard-coded defaults. */
513 /* If SYSTEM_WGETRC is defined, use it. */
515 if (file_exists_p (SYSTEM_WGETRC))
516 ok &= run_wgetrc (SYSTEM_WGETRC);
518 /* Override it with your own, if one exists. */
519 file = wgetrc_file_name ();
522 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
523 something like realpath() before comparing them with `strcmp' */
525 if (!strcmp (file, SYSTEM_WGETRC))
527 fprintf (stderr, _("\
528 %s: Warning: Both system and user wgetrc point to %s.\n"),
529 exec_name, quote (file));
533 ok &= run_wgetrc (file);
535 /* If there were errors processing either `.wgetrc', abort. */
543 /* Remove dashes and underscores from S, modifying S in the
549 char *t = s; /* t - tortoise */
550 char *h = s; /* h - hare */
552 if (*h == '_' || *h == '-')
559 /* Parse the line pointed by line, with the syntax:
560 <sp>* command <sp>* = <sp>* value <sp>*
561 Uses malloc to allocate space for command and value.
563 Returns one of line_ok, line_empty, line_syntax_error, or
564 line_unknown_command.
566 In case of line_ok, *COM and *VAL point to freshly allocated
567 strings, and *COMIND points to com's index. In case of error or
568 empty line, their values are unmodified. */
570 static enum parse_line
571 parse_line (const char *line, char **com, char **val, int *comind)
574 const char *end = line + strlen (line);
575 const char *cmdstart, *cmdend;
576 const char *valstart, *valend;
581 /* Skip leading and trailing whitespace. */
582 while (*line && c_isspace (*line))
584 while (end > line && c_isspace (end[-1]))
587 /* Skip empty lines and comments. */
588 if (!*line || *line == '#')
594 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
598 /* Skip '=', as well as any space before or after it. */
599 while (p < end && c_isspace (*p))
601 if (p == end || *p != '=')
602 return line_syntax_error;
604 while (p < end && c_isspace (*p))
610 /* The syntax is valid (even though the command might not be). Fill
611 in the command name and value. */
612 *com = strdupdelim (cmdstart, cmdend);
613 *val = strdupdelim (valstart, valend);
615 /* The line now known to be syntactically correct. Check whether
616 the command is valid. */
617 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
619 ind = command_by_name (cmdcopy);
621 return line_unknown_command;
623 /* Report success to the caller. */
628 /* Run commands[comind].action. */
631 setval_internal (int comind, const char *com, const char *val)
633 assert (0 <= comind && comind < countof (commands));
634 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
635 return commands[comind].action (com, val, commands[comind].place);
638 /* Run command COM with value VAL. If running the command produces an
639 error, report the error and exit.
641 This is intended to be called from main() to modify Wget's behavior
642 through command-line switches. Since COM is hard-coded in main(),
643 it is not canonicalized, and this aborts when COM is not found.
645 If COMIND's are exported to init.h, this function will be changed
646 to accept COMIND directly. */
649 setoptval (const char *com, const char *val, const char *optname)
651 /* Prepend "--" to OPTNAME. */
652 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
655 strcpy (dd_optname + 2, optname);
657 assert (val != NULL);
658 if (!setval_internal (command_by_name (com), dd_optname, val))
662 /* Parse OPT into command and value and run it. For example,
663 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
664 This is used by the `--execute' flag in main.c. */
667 run_command (const char *opt)
671 switch (parse_line (opt, &com, &val, &comind))
674 if (!setval_internal (comind, com, val))
680 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
681 exec_name, quote (opt));
686 /* Generic helper functions, for use with `commands'. */
688 /* Forward declarations: */
693 static bool decode_string (const char *, const struct decode_item *, int, int *);
694 static bool simple_atoi (const char *, const char *, int *);
695 static bool simple_atof (const char *, const char *, double *);
697 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
699 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
700 && c_tolower((p)[1]) == (c1) \
703 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
704 && c_tolower((p)[1]) == (c1) \
705 && c_tolower((p)[2]) == (c2) \
709 /* Store the boolean value from VAL to PLACE. COM is ignored,
710 except for error messages. */
712 cmd_boolean (const char *com, const char *val, void *place)
716 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
717 /* "on", "yes" and "1" mean true. */
719 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
720 /* "off", "no" and "0" mean false. */
725 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
726 exec_name, com, quote (val));
730 *(bool *) place = value;
734 /* Set the non-negative integer value from VAL to PLACE. With
735 incorrect specification, the number remains unchanged. */
737 cmd_number (const char *com, const char *val, void *place)
739 if (!simple_atoi (val, val + strlen (val), place)
740 || *(int *) place < 0)
742 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
743 exec_name, com, quote (val));
749 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
751 cmd_number_inf (const char *com, const char *val, void *place)
753 if (!strcasecmp (val, "inf"))
758 return cmd_number (com, val, place);
761 /* Copy (strdup) the string at COM to a new location and place a
762 pointer to *PLACE. */
764 cmd_string (const char *com, const char *val, void *place)
766 char **pstring = (char **)place;
768 xfree_null (*pstring);
769 *pstring = xstrdup (val);
773 #if defined(WINDOWS) || defined(MSDOS)
774 # define ISSEP(c) ((c) == '/' || (c) == '\\')
776 # define ISSEP(c) ((c) == '/')
779 /* Like the above, but handles tilde-expansion when reading a user's
780 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
781 gets expanded to the user's home directory. */
783 cmd_file (const char *com, const char *val, void *place)
785 char **pstring = (char **)place;
787 xfree_null (*pstring);
789 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
791 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
794 *pstring = xstrdup (val);
799 char *home = home_dir ();
803 homelen = strlen (home);
804 while (homelen && ISSEP (home[homelen - 1]))
805 home[--homelen] = '\0';
807 /* Skip the leading "~/". */
808 for (++val; ISSEP (*val); val++)
811 *pstring = concat_strings (home, "/", val, (char *) 0);
814 #if defined(WINDOWS) || defined(MSDOS)
815 /* Convert "\" to "/". */
818 for (s = *pstring; *s; s++)
826 /* Like cmd_file, but strips trailing '/' characters. */
828 cmd_directory (const char *com, const char *val, void *place)
832 /* Call cmd_file() for tilde expansion and separator
833 canonicalization (backslash -> slash under Windows). These
834 things should perhaps be in a separate function. */
835 if (!cmd_file (com, val, place))
840 while (t > s && *--t == '/')
846 /* Split VAL by space to a vector of values, and append those values
847 to vector pointed to by the PLACE argument. If VAL is empty, the
848 PLACE vector is cleared instead. */
851 cmd_vector (const char *com, const char *val, void *place)
853 char ***pvec = (char ***)place;
856 *pvec = merge_vecs (*pvec, sepstring (val));
866 cmd_directory_vector (const char *com, const char *val, void *place)
868 char ***pvec = (char ***)place;
872 /* Strip the trailing slashes from directories. */
875 seps = sepstring (val);
876 for (t = seps; t && *t; t++)
878 int len = strlen (*t);
879 /* Skip degenerate case of root directory. */
882 if ((*t)[len - 1] == '/')
883 (*t)[len - 1] = '\0';
886 *pvec = merge_vecs (*pvec, seps);
896 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
897 "100k" or "2.5G" to a floating point number. */
900 parse_bytes_helper (const char *val, double *result)
903 const char *end = val + strlen (val);
905 /* Check for "inf". */
906 if (0 == strcmp (val, "inf"))
912 /* Strip trailing whitespace. */
913 while (val < end && c_isspace (end[-1]))
918 switch (c_tolower (end[-1]))
921 --end, mult = 1024.0;
924 --end, mult = 1048576.0;
927 --end, mult = 1073741824.0;
930 --end, mult = 1099511627776.0;
933 /* Not a recognized suffix: assume it's a digit. (If not,
934 simple_atof will raise an error.) */
938 /* Skip leading and trailing whitespace. */
939 while (val < end && c_isspace (*val))
941 while (val < end && c_isspace (end[-1]))
946 if (!simple_atof (val, end, &number) || number < 0)
949 *result = number * mult;
953 /* Parse VAL as a number and set its value to PLACE (which should
956 By default, the value is assumed to be in bytes. If "K", "M", or
957 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
958 1<<30, respectively. Floating point values are allowed and are
959 cast to integer before use. The idea is to be able to use things
960 like 1.5k instead of "1536".
962 The string "inf" is returned as 0.
964 In case of error, false is returned and memory pointed to by PLACE
965 remains unmodified. */
968 cmd_bytes (const char *com, const char *val, void *place)
971 if (!parse_bytes_helper (val, &byte_value))
973 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
974 exec_name, com, quote (val));
977 *(wgint *)place = (wgint)byte_value;
981 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
982 SIZE_SUM. It works by converting the string to double, therefore
983 working with values up to 2^53-1 without loss of precision. This
984 value (8192 TB) is large enough to serve for a while. */
987 cmd_bytes_sum (const char *com, const char *val, void *place)
990 if (!parse_bytes_helper (val, &byte_value))
992 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
993 exec_name, com, quote (val));
996 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1000 /* Store the value of VAL to *OUT. The value is a time period, by
1001 default expressed in seconds, but also accepting suffixes "m", "h",
1002 "d", and "w" for minutes, hours, days, and weeks respectively. */
1005 cmd_time (const char *com, const char *val, void *place)
1007 double number, mult;
1008 const char *end = val + strlen (val);
1010 /* Strip trailing whitespace. */
1011 while (val < end && c_isspace (end[-1]))
1017 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1018 exec_name, com, quote (val));
1022 switch (c_tolower (end[-1]))
1025 --end, mult = 1; /* seconds */
1028 --end, mult = 60; /* minutes */
1031 --end, mult = 3600; /* hours */
1034 --end, mult = 86400.0; /* days */
1037 --end, mult = 604800.0; /* weeks */
1040 /* Not a recognized suffix: assume it belongs to the number.
1041 (If not, simple_atof will raise an error.) */
1045 /* Skip leading and trailing whitespace. */
1046 while (val < end && c_isspace (*val))
1048 while (val < end && c_isspace (end[-1]))
1053 if (!simple_atof (val, end, &number))
1056 *(double *)place = number * mult;
1062 cmd_cert_type (const char *com, const char *val, void *place)
1064 static const struct decode_item choices[] = {
1065 { "pem", keyfile_pem },
1066 { "der", keyfile_asn1 },
1067 { "asn1", keyfile_asn1 },
1069 int ok = decode_string (val, choices, countof (choices), place);
1071 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1076 /* Specialized helper functions, used by `commands' to handle some
1077 options specially. */
1079 static bool check_user_specified_header (const char *);
1082 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1084 if (!cmd_boolean (com, val, &opt.dirstruct))
1086 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1087 must be affected inversely. */
1089 opt.no_dirstruct = false;
1091 opt.no_dirstruct = true;
1096 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1098 /* Empty value means reset the list of headers. */
1101 free_vec (opt.user_headers);
1102 opt.user_headers = NULL;
1106 if (!check_user_specified_header (val))
1108 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1109 exec_name, com, quote (val));
1112 opt.user_headers = vec_append (opt.user_headers, val);
1117 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1119 int flag = cmd_boolean (com, val, &opt.htmlify);
1120 if (flag && !opt.htmlify)
1121 opt.remove_listing = false;
1125 /* Set the "mirror" mode. It means: recursive download, timestamping,
1126 no limit on max. recursion depth, and don't remove listings. */
1129 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1133 if (!cmd_boolean (com, val, &mirror))
1137 opt.recursive = true;
1138 if (!opt.no_dirstruct)
1139 opt.dirstruct = true;
1140 opt.timestamping = true;
1141 opt.reclevel = INFINITE_RECURSION;
1142 opt.remove_listing = false;
1147 /* Validate --prefer-family and set the choice. Allowed values are
1148 "IPv4", "IPv6", and "none". */
1151 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1153 static const struct decode_item choices[] = {
1154 { "IPv4", prefer_ipv4 },
1155 { "IPv6", prefer_ipv6 },
1156 { "none", prefer_none },
1158 int prefer_family = prefer_ipv4;
1159 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1161 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1162 opt.prefer_family = prefer_family;
1166 /* Set progress.type to VAL, but verify that it's a valid progress
1167 implementation before that. */
1170 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1172 if (!valid_progress_implementation_p (val))
1174 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1175 exec_name, com, quote (val));
1178 xfree_null (opt.progress_type);
1180 /* Don't call set_progress_implementation here. It will be called
1181 in main() when it becomes clear what the log output is. */
1182 opt.progress_type = xstrdup (val);
1186 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1187 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1191 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1193 if (!cmd_boolean (com, val, &opt.recursive))
1197 if (opt.recursive && !opt.no_dirstruct)
1198 opt.dirstruct = true;
1204 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1206 int restrict_os = opt.restrict_files_os;
1207 int restrict_ctrl = opt.restrict_files_ctrl;
1208 int restrict_case = opt.restrict_files_case;
1212 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1216 end = strchr (val, ',');
1218 end = val + strlen (val);
1220 if (VAL_IS ("unix"))
1221 restrict_os = restrict_unix;
1222 else if (VAL_IS ("windows"))
1223 restrict_os = restrict_windows;
1224 else if (VAL_IS ("lowercase"))
1225 restrict_case = restrict_lowercase;
1226 else if (VAL_IS ("uppercase"))
1227 restrict_case = restrict_uppercase;
1228 else if (VAL_IS ("nocontrol"))
1229 restrict_ctrl = false;
1233 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1234 exec_name, com, quote (val));
1241 while (*val && *end);
1245 opt.restrict_files_os = restrict_os;
1246 opt.restrict_files_ctrl = restrict_ctrl;
1247 opt.restrict_files_case = restrict_case;
1254 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1256 static const struct decode_item choices[] = {
1257 { "auto", secure_protocol_auto },
1258 { "sslv2", secure_protocol_sslv2 },
1259 { "sslv3", secure_protocol_sslv3 },
1260 { "tlsv1", secure_protocol_tlsv1 },
1262 int ok = decode_string (val, choices, countof (choices), place);
1264 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1269 /* Set all three timeout values. */
1272 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1275 if (!cmd_time (com, val, &value))
1277 opt.read_timeout = value;
1278 opt.connect_timeout = value;
1279 opt.dns_timeout = value;
1284 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1286 /* Disallow embedded newlines. */
1287 if (strchr (val, '\n'))
1289 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1290 exec_name, com, quote (val));
1293 xfree_null (opt.useragent);
1294 opt.useragent = xstrdup (val);
1298 /* The "verbose" option cannot be cmd_boolean because the variable is
1299 not bool -- it's of type int (-1 means uninitialized because of
1300 some random hackery for disallowing -q -v). */
1303 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1306 if (cmd_boolean (com, val, &flag))
1314 /* Miscellaneous useful routines. */
1316 /* A very simple atoi clone, more useful than atoi because it works on
1317 delimited strings, and has error reportage. Returns true on success,
1318 false on failure. If successful, stores result to *DEST. */
1321 simple_atoi (const char *beg, const char *end, int *dest)
1324 bool negative = false;
1325 const char *p = beg;
1327 while (p < end && c_isspace (*p))
1329 if (p < end && (*p == '-' || *p == '+'))
1331 negative = (*p == '-');
1337 /* Read negative numbers in a separate loop because the most
1338 negative integer cannot be represented as a positive number. */
1341 for (; p < end && c_isdigit (*p); p++)
1343 int next = (10 * result) + (*p - '0');
1345 return false; /* overflow */
1349 for (; p < end && c_isdigit (*p); p++)
1351 int next = (10 * result) - (*p - '0');
1353 return false; /* underflow */
1364 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1365 doesn't handle exponential notation. Returns true on success,
1366 false on failure. In case of success, stores its result to
1370 simple_atof (const char *beg, const char *end, double *dest)
1374 bool negative = false;
1375 bool seen_dot = false;
1376 bool seen_digit = false;
1379 const char *p = beg;
1381 while (p < end && c_isspace (*p))
1383 if (p < end && (*p == '-' || *p == '+'))
1385 negative = (*p == '-');
1389 for (; p < end; p++)
1395 result = (10 * result) + (ch - '0');
1397 result += (ch - '0') / (divider *= 10);
1419 /* Verify that the user-specified header in S is valid. It must
1420 contain a colon preceded by non-white-space characters and must not
1421 contain newlines. */
1424 check_user_specified_header (const char *s)
1428 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1430 /* The header MUST contain `:' preceded by at least one
1431 non-whitespace character. */
1432 if (*p != ':' || p == s)
1434 /* The header MUST NOT contain newlines. */
1435 if (strchr (s, '\n'))
1440 /* Decode VAL into a number, according to ITEMS. */
1443 decode_string (const char *val, const struct decode_item *items, int itemcount,
1447 for (i = 0; i < itemcount; i++)
1448 if (0 == strcasecmp (val, items[i].name))
1450 *place = items[i].code;
1457 void cleanup_html_url (void);
1460 /* Free the memory allocated by global variables. */
1464 /* Free external resources, close files, etc. */
1467 fclose (output_stream);
1468 /* No need to check for error because Wget flushes its output (and
1469 checks for errors) after any data arrives. */
1471 /* We're exiting anyway so there's no real need to call free()
1472 hundreds of times. Skipping the frees will make Wget exit
1475 However, when detecting leaks, it's crucial to free() everything
1476 because then you can find the real leaks, i.e. the allocated
1477 memory which grows with the size of the program. */
1483 cleanup_html_url ();
1488 extern acc_t *netrc_list;
1489 free_netrc (netrc_list);
1491 xfree_null (opt.lfilename);
1492 xfree_null (opt.dir_prefix);
1493 xfree_null (opt.input_filename);
1494 xfree_null (opt.output_document);
1495 free_vec (opt.accepts);
1496 free_vec (opt.rejects);
1497 free_vec (opt.excludes);
1498 free_vec (opt.includes);
1499 free_vec (opt.domains);
1500 free_vec (opt.follow_tags);
1501 free_vec (opt.ignore_tags);
1502 xfree_null (opt.progress_type);
1503 xfree_null (opt.ftp_user);
1504 xfree_null (opt.ftp_passwd);
1505 xfree_null (opt.ftp_proxy);
1506 xfree_null (opt.https_proxy);
1507 xfree_null (opt.http_proxy);
1508 free_vec (opt.no_proxy);
1509 xfree_null (opt.useragent);
1510 xfree_null (opt.referer);
1511 xfree_null (opt.http_user);
1512 xfree_null (opt.http_passwd);
1513 free_vec (opt.user_headers);
1515 xfree_null (opt.cert_file);
1516 xfree_null (opt.private_key);
1517 xfree_null (opt.ca_directory);
1518 xfree_null (opt.ca_cert);
1519 xfree_null (opt.random_file);
1520 xfree_null (opt.egd_file);
1522 xfree_null (opt.bind_address);
1523 xfree_null (opt.cookies_input);
1524 xfree_null (opt.cookies_output);
1525 xfree_null (opt.user);
1526 xfree_null (opt.passwd);
1527 #endif /* DEBUG_MALLOC */
1530 /* Unit testing routines. */
1535 test_commands_sorted()
1537 int prev_idx = 0, next_idx = 1;
1538 int command_count = countof (commands) - 1;
1540 while (next_idx <= command_count)
1542 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1545 mu_assert ("FAILED", false);
1558 test_cmd_spec_restrict_file_names()
1563 int expected_restrict_files_os;
1564 int expected_restrict_files_ctrl;
1565 int expected_restrict_files_case;
1568 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1569 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1570 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1571 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1574 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1579 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1582 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1583 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1584 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1585 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1587 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1588 res == test_array[i].result
1589 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1590 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1591 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1597 #endif /* TESTING */