1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007 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 In addition, as a special exception, the Free Software Foundation
21 gives permission to link the code of its release of Wget with the
22 OpenSSL project's "OpenSSL" library (or with modified versions of it
23 that use the same license as the "OpenSSL" library), and distribute
24 the linked executables. You must obey the GNU General Public License
25 in all respects for all of the code used other than "OpenSSL". If you
26 modify this file, you may extend this exception to your version of the
27 file, but you are not obligated to do so. If you do not wish to do
28 so, delete this exception statement from your version. */
50 #include "recur.h" /* for INFINITE_RECURSION */
51 #include "convert.h" /* for convert_cleanup */
52 #include "res.h" /* for res_cleanup */
53 #include "http.h" /* for http_cleanup */
54 #include "retr.h" /* for output_stream */
60 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
61 otherwise, it will be performed by the shell. This variable will
62 be set by the wgetrc-reading function. */
64 static bool enable_tilde_expansion;
67 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
69 CMD_DECLARE (cmd_boolean);
70 CMD_DECLARE (cmd_bytes);
71 CMD_DECLARE (cmd_bytes_sum);
73 CMD_DECLARE (cmd_cert_type);
75 CMD_DECLARE (cmd_directory_vector);
76 CMD_DECLARE (cmd_number);
77 CMD_DECLARE (cmd_number_inf);
78 CMD_DECLARE (cmd_string);
79 CMD_DECLARE (cmd_file);
80 CMD_DECLARE (cmd_directory);
81 CMD_DECLARE (cmd_time);
82 CMD_DECLARE (cmd_vector);
84 CMD_DECLARE (cmd_spec_dirstruct);
85 CMD_DECLARE (cmd_spec_header);
86 CMD_DECLARE (cmd_spec_htmlify);
87 CMD_DECLARE (cmd_spec_mirror);
88 CMD_DECLARE (cmd_spec_prefer_family);
89 CMD_DECLARE (cmd_spec_progress);
90 CMD_DECLARE (cmd_spec_recursive);
91 CMD_DECLARE (cmd_spec_restrict_file_names);
93 CMD_DECLARE (cmd_spec_secure_protocol);
95 CMD_DECLARE (cmd_spec_timeout);
96 CMD_DECLARE (cmd_spec_useragent);
97 CMD_DECLARE (cmd_spec_verbose);
99 /* List of recognized commands, each consisting of name, place and
100 function. When adding a new command, simply add it to the list,
101 but be sure to keep the list sorted alphabetically, as
102 command_by_name's binary search depends on it. Also, be sure to
103 add any entries that allocate memory (e.g. cmd_string and
104 cmd_vector) to the cleanup() function below. */
106 static const struct {
109 bool (*action) (const char *, const char *, void *);
111 /* KEEP THIS LIST ALPHABETICALLY SORTED */
112 { "accept", &opt.accepts, cmd_vector },
113 { "addhostdir", &opt.add_hostdir, cmd_boolean },
114 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
115 { "background", &opt.background, cmd_boolean },
116 { "backupconverted", &opt.backup_converted, cmd_boolean },
117 { "backups", &opt.backups, cmd_number },
118 { "base", &opt.base_href, cmd_string },
119 { "bindaddress", &opt.bind_address, cmd_string },
121 { "cacertificate", &opt.ca_cert, cmd_file },
123 { "cache", &opt.allow_cache, cmd_boolean },
125 { "cadirectory", &opt.ca_directory, cmd_directory },
126 { "certificate", &opt.cert_file, cmd_file },
127 { "certificatetype", &opt.cert_type, cmd_cert_type },
128 { "checkcertificate", &opt.check_cert, cmd_boolean },
130 { "connecttimeout", &opt.connect_timeout, cmd_time },
131 { "contentdisposition", &opt.content_disposition, cmd_boolean },
132 { "continue", &opt.always_rest, cmd_boolean },
133 { "convertlinks", &opt.convert_links, cmd_boolean },
134 { "cookies", &opt.cookies, cmd_boolean },
135 { "cutdirs", &opt.cut_dirs, cmd_number },
137 { "debug", &opt.debug, cmd_boolean },
139 { "deleteafter", &opt.delete_after, cmd_boolean },
140 { "dirprefix", &opt.dir_prefix, cmd_directory },
141 { "dirstruct", NULL, cmd_spec_dirstruct },
142 { "dnscache", &opt.dns_cache, cmd_boolean },
143 { "dnstimeout", &opt.dns_timeout, cmd_time },
144 { "domains", &opt.domains, cmd_vector },
145 { "dotbytes", &opt.dot_bytes, cmd_bytes },
146 { "dotsinline", &opt.dots_in_line, cmd_number },
147 { "dotspacing", &opt.dot_spacing, cmd_number },
148 { "dotstyle", &opt.dot_style, cmd_string },
150 { "egdfile", &opt.egd_file, cmd_file },
152 { "excludedirectories", &opt.excludes, cmd_directory_vector },
153 { "excludedomains", &opt.exclude_domains, cmd_vector },
154 { "followftp", &opt.follow_ftp, cmd_boolean },
155 { "followtags", &opt.follow_tags, cmd_vector },
156 { "forcehtml", &opt.force_html, cmd_boolean },
157 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
158 { "ftppassword", &opt.ftp_passwd, cmd_string },
159 { "ftpproxy", &opt.ftp_proxy, cmd_string },
160 { "ftpuser", &opt.ftp_user, cmd_string },
161 { "glob", &opt.ftp_glob, cmd_boolean },
162 { "header", NULL, cmd_spec_header },
163 { "htmlextension", &opt.html_extension, cmd_boolean },
164 { "htmlify", NULL, cmd_spec_htmlify },
165 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
166 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
167 { "httppassword", &opt.http_passwd, cmd_string },
168 { "httpproxy", &opt.http_proxy, cmd_string },
169 { "httpsproxy", &opt.https_proxy, cmd_string },
170 { "httpuser", &opt.http_user, cmd_string },
171 { "ignorecase", &opt.ignore_case, cmd_boolean },
172 { "ignorelength", &opt.ignore_length, cmd_boolean },
173 { "ignoretags", &opt.ignore_tags, cmd_vector },
174 { "includedirectories", &opt.includes, cmd_directory_vector },
176 { "inet4only", &opt.ipv4_only, cmd_boolean },
177 { "inet6only", &opt.ipv6_only, cmd_boolean },
179 { "input", &opt.input_filename, cmd_file },
180 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
181 { "limitrate", &opt.limit_rate, cmd_bytes },
182 { "loadcookies", &opt.cookies_input, cmd_file },
183 { "logfile", &opt.lfilename, cmd_file },
184 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
185 { "maxredirect", &opt.max_redirect, cmd_number },
186 { "mirror", NULL, cmd_spec_mirror },
187 { "netrc", &opt.netrc, cmd_boolean },
188 { "noclobber", &opt.noclobber, cmd_boolean },
189 { "noparent", &opt.no_parent, cmd_boolean },
190 { "noproxy", &opt.no_proxy, cmd_vector },
191 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
192 { "outputdocument", &opt.output_document, cmd_file },
193 { "pagerequisites", &opt.page_requisites, cmd_boolean },
194 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
195 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
196 { "password", &opt.passwd, cmd_string },
197 { "postdata", &opt.post_data, cmd_string },
198 { "postfile", &opt.post_file_name, cmd_file },
199 { "preferfamily", NULL, cmd_spec_prefer_family },
200 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
202 { "privatekey", &opt.private_key, cmd_file },
203 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
205 { "progress", &opt.progress_type, cmd_spec_progress },
206 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
207 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
208 { "proxypassword", &opt.proxy_passwd, cmd_string },
209 { "proxyuser", &opt.proxy_user, cmd_string },
210 { "quiet", &opt.quiet, cmd_boolean },
211 { "quota", &opt.quota, cmd_bytes_sum },
213 { "randomfile", &opt.random_file, cmd_file },
215 { "randomwait", &opt.random_wait, cmd_boolean },
216 { "readtimeout", &opt.read_timeout, cmd_time },
217 { "reclevel", &opt.reclevel, cmd_number_inf },
218 { "recursive", NULL, cmd_spec_recursive },
219 { "referer", &opt.referer, cmd_string },
220 { "reject", &opt.rejects, cmd_vector },
221 { "relativeonly", &opt.relative_only, cmd_boolean },
222 { "removelisting", &opt.remove_listing, cmd_boolean },
223 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
224 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
225 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
226 { "robots", &opt.use_robots, cmd_boolean },
227 { "savecookies", &opt.cookies_output, cmd_file },
228 { "saveheaders", &opt.save_headers, cmd_boolean },
230 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
232 { "serverresponse", &opt.server_response, cmd_boolean },
233 { "spanhosts", &opt.spanhost, cmd_boolean },
234 { "spider", &opt.spider, cmd_boolean },
235 { "strictcomments", &opt.strict_comments, cmd_boolean },
236 { "timeout", NULL, cmd_spec_timeout },
237 { "timestamping", &opt.timestamping, cmd_boolean },
238 { "tries", &opt.ntry, cmd_number_inf },
239 { "useproxy", &opt.use_proxy, cmd_boolean },
240 { "user", &opt.user, cmd_string },
241 { "useragent", NULL, cmd_spec_useragent },
242 { "verbose", NULL, cmd_spec_verbose },
243 { "wait", &opt.wait, cmd_time },
244 { "waitretry", &opt.waitretry, cmd_time },
246 { "wdebug", &opt.wdebug, cmd_boolean },
250 /* Look up CMDNAME in the commands[] and return its position in the
251 array. If CMDNAME is not found, return -1. */
254 command_by_name (const char *cmdname)
256 /* Use binary search for speed. Wget has ~100 commands, which
257 guarantees a worst case performance of 7 string comparisons. */
258 int lo = 0, hi = countof (commands) - 1;
262 int mid = (lo + hi) >> 1;
263 int cmp = strcasecmp (cmdname, commands[mid].name);
274 /* Reset the variables to default values. */
280 /* Most of the default values are 0 (and 0.0, NULL, and false).
281 Just reset everything, and fill in the non-zero values. Note
282 that initializing pointers to NULL this way is technically
283 illegal, but porting Wget to a machine where NULL is not all-zero
284 bit pattern will be the least of the implementors' worries. */
291 opt.add_hostdir = true;
295 opt.http_keep_alive = true;
296 opt.use_proxy = true;
297 tmp = getenv ("no_proxy");
299 opt.no_proxy = sepstring (tmp);
300 opt.allow_cache = true;
302 opt.read_timeout = 900;
303 opt.use_robots = true;
305 opt.remove_listing = true;
307 opt.dot_bytes = 1024;
308 opt.dot_spacing = 10;
309 opt.dots_in_line = 50;
311 opt.dns_cache = true;
315 opt.check_cert = true;
318 /* The default for file name restriction defaults to the OS type. */
319 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
320 opt.restrict_files_os = restrict_windows;
322 opt.restrict_files_os = restrict_unix;
324 opt.restrict_files_ctrl = true;
325 opt.restrict_files_case = restrict_no_case_restriction;
327 opt.max_redirect = 20;
330 /* Return the user's home directory (strdup-ed), or NULL if none is
335 char *home = getenv ("HOME");
340 /* Under MSDOS, if $HOME isn't defined, use the directory where
341 `wget.exe' resides. */
342 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
343 char *p, buf[PATH_MAX];
345 strcpy (buf, _w32_get_argv0 ());
346 p = strrchr (buf, '/'); /* djgpp */
348 p = strrchr (buf, '\\'); /* others */
352 #elif !defined(WINDOWS)
353 /* If HOME is not defined, try getting it from the password
355 struct passwd *pwd = getpwuid (getuid ());
356 if (!pwd || !pwd->pw_dir)
360 /* Under Windows, if $HOME isn't defined, use the directory where
361 `wget.exe' resides. */
366 return home ? xstrdup (home) : NULL;
369 /* Return the path to the user's .wgetrc. This is either the value of
370 `WGETRC' environment variable, or `$HOME/.wgetrc'.
372 If the `WGETRC' variable exists but the file does not exist, the
373 function will exit(). */
375 wgetrc_file_name (void)
380 /* Try the environment. */
381 env = getenv ("WGETRC");
384 if (!file_exists_p (env))
386 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
390 return xstrdup (env);
393 /* If that failed, try $HOME/.wgetrc. */
396 file = aprintf ("%s/.wgetrc", home);
400 /* Under Windows, if we still haven't found .wgetrc, look for the file
401 `wget.ini' in the directory where `wget.exe' resides; we do this for
402 backward compatibility with previous versions of Wget.
403 SYSTEM_WGETRC should not be defined under WINDOWS. */
404 if (!file || !file_exists_p (file))
410 file = aprintf ("%s/wget.ini", home);
416 if (!file_exists_p (file))
424 /* Return values of parse_line. */
432 static enum parse_line parse_line (const char *, char **, char **, int *);
433 static bool setval_internal (int, const char *, const char *);
435 /* Initialize variables from a wgetrc file. Returns zero (failure) if
436 there were errors in the file. */
439 run_wgetrc (const char *file)
446 fp = fopen (file, "rb");
449 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
450 file, strerror (errno));
451 return true; /* not a fatal error */
453 enable_tilde_expansion = true;
455 while ((line = read_whole_line (fp)) != NULL)
457 char *com = NULL, *val = NULL;
460 /* Parse the line. */
461 switch (parse_line (line, &com, &val, &comind))
464 /* If everything is OK, set the value. */
465 if (!setval_internal (comind, com, val))
467 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
468 exec_name, file, ln);
472 case line_syntax_error:
473 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
474 exec_name, file, ln);
477 case line_unknown_command:
478 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
479 exec_name, com, file, ln);
492 enable_tilde_expansion = false;
498 /* Initialize the defaults and run the system wgetrc and user's own
506 /* Load the hard-coded defaults. */
509 /* If SYSTEM_WGETRC is defined, use it. */
511 if (file_exists_p (SYSTEM_WGETRC))
512 ok &= run_wgetrc (SYSTEM_WGETRC);
514 /* Override it with your own, if one exists. */
515 file = wgetrc_file_name ();
518 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
519 something like realpath() before comparing them with `strcmp' */
521 if (!strcmp (file, SYSTEM_WGETRC))
523 fprintf (stderr, _("\
524 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
529 ok &= run_wgetrc (file);
531 /* If there were errors processing either `.wgetrc', abort. */
539 /* Remove dashes and underscores from S, modifying S in the
545 char *t = s; /* t - tortoise */
546 char *h = s; /* h - hare */
548 if (*h == '_' || *h == '-')
555 /* Parse the line pointed by line, with the syntax:
556 <sp>* command <sp>* = <sp>* value <sp>*
557 Uses malloc to allocate space for command and value.
559 Returns one of line_ok, line_empty, line_syntax_error, or
560 line_unknown_command.
562 In case of line_ok, *COM and *VAL point to freshly allocated
563 strings, and *COMIND points to com's index. In case of error or
564 empty line, their values are unmodified. */
566 static enum parse_line
567 parse_line (const char *line, char **com, char **val, int *comind)
570 const char *end = line + strlen (line);
571 const char *cmdstart, *cmdend;
572 const char *valstart, *valend;
577 /* Skip leading and trailing whitespace. */
578 while (*line && c_isspace (*line))
580 while (end > line && c_isspace (end[-1]))
583 /* Skip empty lines and comments. */
584 if (!*line || *line == '#')
590 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
594 /* Skip '=', as well as any space before or after it. */
595 while (p < end && c_isspace (*p))
597 if (p == end || *p != '=')
598 return line_syntax_error;
600 while (p < end && c_isspace (*p))
606 /* The syntax is valid (even though the command might not be). Fill
607 in the command name and value. */
608 *com = strdupdelim (cmdstart, cmdend);
609 *val = strdupdelim (valstart, valend);
611 /* The line now known to be syntactically correct. Check whether
612 the command is valid. */
613 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
615 ind = command_by_name (cmdcopy);
617 return line_unknown_command;
619 /* Report success to the caller. */
624 /* Run commands[comind].action. */
627 setval_internal (int comind, const char *com, const char *val)
629 assert (0 <= comind && comind < countof (commands));
630 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
631 return commands[comind].action (com, val, commands[comind].place);
634 /* Run command COM with value VAL. If running the command produces an
635 error, report the error and exit.
637 This is intended to be called from main() to modify Wget's behavior
638 through command-line switches. Since COM is hard-coded in main(),
639 it is not canonicalized, and this aborts when COM is not found.
641 If COMIND's are exported to init.h, this function will be changed
642 to accept COMIND directly. */
645 setoptval (const char *com, const char *val, const char *optname)
647 /* Prepend "--" to OPTNAME. */
648 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
651 strcpy (dd_optname + 2, optname);
653 assert (val != NULL);
654 if (!setval_internal (command_by_name (com), dd_optname, val))
658 /* Parse OPT into command and value and run it. For example,
659 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
660 This is used by the `--execute' flag in main.c. */
663 run_command (const char *opt)
667 switch (parse_line (opt, &com, &val, &comind))
670 if (!setval_internal (comind, com, val))
676 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
682 /* Generic helper functions, for use with `commands'. */
684 /* Forward declarations: */
689 static bool decode_string (const char *, const struct decode_item *, int, int *);
690 static bool simple_atoi (const char *, const char *, int *);
691 static bool simple_atof (const char *, const char *, double *);
693 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
695 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
696 && c_tolower((p)[1]) == (c1) \
699 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
700 && c_tolower((p)[1]) == (c1) \
701 && c_tolower((p)[2]) == (c2) \
705 /* Store the boolean value from VAL to PLACE. COM is ignored,
706 except for error messages. */
708 cmd_boolean (const char *com, const char *val, void *place)
712 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
713 /* "on", "yes" and "1" mean true. */
715 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
716 /* "off", "no" and "0" mean false. */
721 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
722 exec_name, com, val);
726 *(bool *) place = value;
730 /* Set the non-negative integer value from VAL to PLACE. With
731 incorrect specification, the number remains unchanged. */
733 cmd_number (const char *com, const char *val, void *place)
735 if (!simple_atoi (val, val + strlen (val), place)
736 || *(int *) place < 0)
738 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
739 exec_name, com, val);
745 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
747 cmd_number_inf (const char *com, const char *val, void *place)
749 if (!strcasecmp (val, "inf"))
754 return cmd_number (com, val, place);
757 /* Copy (strdup) the string at COM to a new location and place a
758 pointer to *PLACE. */
760 cmd_string (const char *com, const char *val, void *place)
762 char **pstring = (char **)place;
764 xfree_null (*pstring);
765 *pstring = xstrdup (val);
769 #if defined(WINDOWS) || defined(MSDOS)
770 # define ISSEP(c) ((c) == '/' || (c) == '\\')
772 # define ISSEP(c) ((c) == '/')
775 /* Like the above, but handles tilde-expansion when reading a user's
776 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
777 gets expanded to the user's home directory. */
779 cmd_file (const char *com, const char *val, void *place)
781 char **pstring = (char **)place;
783 xfree_null (*pstring);
785 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
787 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
790 *pstring = xstrdup (val);
795 char *home = home_dir ();
799 homelen = strlen (home);
800 while (homelen && ISSEP (home[homelen - 1]))
801 home[--homelen] = '\0';
803 /* Skip the leading "~/". */
804 for (++val; ISSEP (*val); val++)
807 *pstring = concat_strings (home, "/", val, (char *) 0);
810 #if defined(WINDOWS) || defined(MSDOS)
811 /* Convert "\" to "/". */
814 for (s = *pstring; *s; s++)
822 /* Like cmd_file, but strips trailing '/' characters. */
824 cmd_directory (const char *com, const char *val, void *place)
828 /* Call cmd_file() for tilde expansion and separator
829 canonicalization (backslash -> slash under Windows). These
830 things should perhaps be in a separate function. */
831 if (!cmd_file (com, val, place))
836 while (t > s && *--t == '/')
842 /* Split VAL by space to a vector of values, and append those values
843 to vector pointed to by the PLACE argument. If VAL is empty, the
844 PLACE vector is cleared instead. */
847 cmd_vector (const char *com, const char *val, void *place)
849 char ***pvec = (char ***)place;
852 *pvec = merge_vecs (*pvec, sepstring (val));
862 cmd_directory_vector (const char *com, const char *val, void *place)
864 char ***pvec = (char ***)place;
868 /* Strip the trailing slashes from directories. */
871 seps = sepstring (val);
872 for (t = seps; t && *t; t++)
874 int len = strlen (*t);
875 /* Skip degenerate case of root directory. */
878 if ((*t)[len - 1] == '/')
879 (*t)[len - 1] = '\0';
882 *pvec = merge_vecs (*pvec, seps);
892 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
893 "100k" or "2.5G" to a floating point number. */
896 parse_bytes_helper (const char *val, double *result)
899 const char *end = val + strlen (val);
901 /* Check for "inf". */
902 if (0 == strcmp (val, "inf"))
908 /* Strip trailing whitespace. */
909 while (val < end && c_isspace (end[-1]))
914 switch (c_tolower (end[-1]))
917 --end, mult = 1024.0;
920 --end, mult = 1048576.0;
923 --end, mult = 1073741824.0;
926 --end, mult = 1099511627776.0;
929 /* Not a recognized suffix: assume it's a digit. (If not,
930 simple_atof will raise an error.) */
934 /* Skip leading and trailing whitespace. */
935 while (val < end && c_isspace (*val))
937 while (val < end && c_isspace (end[-1]))
942 if (!simple_atof (val, end, &number) || number < 0)
945 *result = number * mult;
949 /* Parse VAL as a number and set its value to PLACE (which should
952 By default, the value is assumed to be in bytes. If "K", "M", or
953 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
954 1<<30, respectively. Floating point values are allowed and are
955 cast to integer before use. The idea is to be able to use things
956 like 1.5k instead of "1536".
958 The string "inf" is returned as 0.
960 In case of error, false is returned and memory pointed to by PLACE
961 remains unmodified. */
964 cmd_bytes (const char *com, const char *val, void *place)
967 if (!parse_bytes_helper (val, &byte_value))
969 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
970 exec_name, com, val);
973 *(wgint *)place = (wgint)byte_value;
977 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
978 SIZE_SUM. It works by converting the string to double, therefore
979 working with values up to 2^53-1 without loss of precision. This
980 value (8192 TB) is large enough to serve for a while. */
983 cmd_bytes_sum (const char *com, const char *val, void *place)
986 if (!parse_bytes_helper (val, &byte_value))
988 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
989 exec_name, com, val);
992 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
996 /* Store the value of VAL to *OUT. The value is a time period, by
997 default expressed in seconds, but also accepting suffixes "m", "h",
998 "d", and "w" for minutes, hours, days, and weeks respectively. */
1001 cmd_time (const char *com, const char *val, void *place)
1003 double number, mult;
1004 const char *end = val + strlen (val);
1006 /* Strip trailing whitespace. */
1007 while (val < end && c_isspace (end[-1]))
1013 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1014 exec_name, com, val);
1018 switch (c_tolower (end[-1]))
1021 --end, mult = 1; /* seconds */
1024 --end, mult = 60; /* minutes */
1027 --end, mult = 3600; /* hours */
1030 --end, mult = 86400.0; /* days */
1033 --end, mult = 604800.0; /* weeks */
1036 /* Not a recognized suffix: assume it belongs to the number.
1037 (If not, simple_atof will raise an error.) */
1041 /* Skip leading and trailing whitespace. */
1042 while (val < end && c_isspace (*val))
1044 while (val < end && c_isspace (end[-1]))
1049 if (!simple_atof (val, end, &number))
1052 *(double *)place = number * mult;
1058 cmd_cert_type (const char *com, const char *val, void *place)
1060 static const struct decode_item choices[] = {
1061 { "pem", keyfile_pem },
1062 { "der", keyfile_asn1 },
1063 { "asn1", keyfile_asn1 },
1065 int ok = decode_string (val, choices, countof (choices), place);
1067 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1072 /* Specialized helper functions, used by `commands' to handle some
1073 options specially. */
1075 static bool check_user_specified_header (const char *);
1078 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1080 if (!cmd_boolean (com, val, &opt.dirstruct))
1082 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1083 must be affected inversely. */
1085 opt.no_dirstruct = false;
1087 opt.no_dirstruct = true;
1092 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1094 /* Empty value means reset the list of headers. */
1097 free_vec (opt.user_headers);
1098 opt.user_headers = NULL;
1102 if (!check_user_specified_header (val))
1104 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1105 exec_name, com, val);
1108 opt.user_headers = vec_append (opt.user_headers, val);
1113 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1115 int flag = cmd_boolean (com, val, &opt.htmlify);
1116 if (flag && !opt.htmlify)
1117 opt.remove_listing = false;
1121 /* Set the "mirror" mode. It means: recursive download, timestamping,
1122 no limit on max. recursion depth, and don't remove listings. */
1125 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1129 if (!cmd_boolean (com, val, &mirror))
1133 opt.recursive = true;
1134 if (!opt.no_dirstruct)
1135 opt.dirstruct = true;
1136 opt.timestamping = true;
1137 opt.reclevel = INFINITE_RECURSION;
1138 opt.remove_listing = false;
1143 /* Validate --prefer-family and set the choice. Allowed values are
1144 "IPv4", "IPv6", and "none". */
1147 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1149 static const struct decode_item choices[] = {
1150 { "IPv4", prefer_ipv4 },
1151 { "IPv6", prefer_ipv6 },
1152 { "none", prefer_none },
1154 int prefer_family = prefer_ipv4;
1155 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1157 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1158 opt.prefer_family = prefer_family;
1162 /* Set progress.type to VAL, but verify that it's a valid progress
1163 implementation before that. */
1166 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1168 if (!valid_progress_implementation_p (val))
1170 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1171 exec_name, com, val);
1174 xfree_null (opt.progress_type);
1176 /* Don't call set_progress_implementation here. It will be called
1177 in main() when it becomes clear what the log output is. */
1178 opt.progress_type = xstrdup (val);
1182 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1183 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1187 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1189 if (!cmd_boolean (com, val, &opt.recursive))
1193 if (opt.recursive && !opt.no_dirstruct)
1194 opt.dirstruct = true;
1200 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1202 int restrict_os = opt.restrict_files_os;
1203 int restrict_ctrl = opt.restrict_files_ctrl;
1204 int restrict_case = opt.restrict_files_case;
1208 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1212 end = strchr (val, ',');
1214 end = val + strlen (val);
1216 if (VAL_IS ("unix"))
1217 restrict_os = restrict_unix;
1218 else if (VAL_IS ("windows"))
1219 restrict_os = restrict_windows;
1220 else if (VAL_IS ("lowercase"))
1221 restrict_case = restrict_lowercase;
1222 else if (VAL_IS ("uppercase"))
1223 restrict_case = restrict_uppercase;
1224 else if (VAL_IS ("nocontrol"))
1225 restrict_ctrl = false;
1229 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1230 exec_name, com, val);
1237 while (*val && *end);
1241 opt.restrict_files_os = restrict_os;
1242 opt.restrict_files_ctrl = restrict_ctrl;
1243 opt.restrict_files_case = restrict_case;
1250 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1252 static const struct decode_item choices[] = {
1253 { "auto", secure_protocol_auto },
1254 { "sslv2", secure_protocol_sslv2 },
1255 { "sslv3", secure_protocol_sslv3 },
1256 { "tlsv1", secure_protocol_tlsv1 },
1258 int ok = decode_string (val, choices, countof (choices), place);
1260 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1265 /* Set all three timeout values. */
1268 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1271 if (!cmd_time (com, val, &value))
1273 opt.read_timeout = value;
1274 opt.connect_timeout = value;
1275 opt.dns_timeout = value;
1280 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1282 /* Disallow embedded newlines. */
1283 if (strchr (val, '\n'))
1285 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1286 exec_name, com, val);
1289 xfree_null (opt.useragent);
1290 opt.useragent = xstrdup (val);
1294 /* The "verbose" option cannot be cmd_boolean because the variable is
1295 not bool -- it's of type int (-1 means uninitialized because of
1296 some random hackery for disallowing -q -v). */
1299 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1302 if (cmd_boolean (com, val, &flag))
1310 /* Miscellaneous useful routines. */
1312 /* A very simple atoi clone, more useful than atoi because it works on
1313 delimited strings, and has error reportage. Returns true on success,
1314 false on failure. If successful, stores result to *DEST. */
1317 simple_atoi (const char *beg, const char *end, int *dest)
1320 bool negative = false;
1321 const char *p = beg;
1323 while (p < end && c_isspace (*p))
1325 if (p < end && (*p == '-' || *p == '+'))
1327 negative = (*p == '-');
1333 /* Read negative numbers in a separate loop because the most
1334 negative integer cannot be represented as a positive number. */
1337 for (; p < end && c_isdigit (*p); p++)
1339 int next = (10 * result) + (*p - '0');
1341 return false; /* overflow */
1345 for (; p < end && c_isdigit (*p); p++)
1347 int next = (10 * result) - (*p - '0');
1349 return false; /* underflow */
1360 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1361 doesn't handle exponential notation. Returns true on success,
1362 false on failure. In case of success, stores its result to
1366 simple_atof (const char *beg, const char *end, double *dest)
1370 bool negative = false;
1371 bool seen_dot = false;
1372 bool seen_digit = false;
1375 const char *p = beg;
1377 while (p < end && c_isspace (*p))
1379 if (p < end && (*p == '-' || *p == '+'))
1381 negative = (*p == '-');
1385 for (; p < end; p++)
1391 result = (10 * result) + (ch - '0');
1393 result += (ch - '0') / (divider *= 10);
1415 /* Verify that the user-specified header in S is valid. It must
1416 contain a colon preceded by non-white-space characters and must not
1417 contain newlines. */
1420 check_user_specified_header (const char *s)
1424 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1426 /* The header MUST contain `:' preceded by at least one
1427 non-whitespace character. */
1428 if (*p != ':' || p == s)
1430 /* The header MUST NOT contain newlines. */
1431 if (strchr (s, '\n'))
1436 /* Decode VAL into a number, according to ITEMS. */
1439 decode_string (const char *val, const struct decode_item *items, int itemcount,
1443 for (i = 0; i < itemcount; i++)
1444 if (0 == strcasecmp (val, items[i].name))
1446 *place = items[i].code;
1453 void cleanup_html_url (void);
1456 /* Free the memory allocated by global variables. */
1460 /* Free external resources, close files, etc. */
1463 fclose (output_stream);
1464 /* No need to check for error because Wget flushes its output (and
1465 checks for errors) after any data arrives. */
1467 /* We're exiting anyway so there's no real need to call free()
1468 hundreds of times. Skipping the frees will make Wget exit
1471 However, when detecting leaks, it's crucial to free() everything
1472 because then you can find the real leaks, i.e. the allocated
1473 memory which grows with the size of the program. */
1479 cleanup_html_url ();
1484 extern acc_t *netrc_list;
1485 free_netrc (netrc_list);
1487 xfree_null (opt.lfilename);
1488 xfree_null (opt.dir_prefix);
1489 xfree_null (opt.input_filename);
1490 xfree_null (opt.output_document);
1491 free_vec (opt.accepts);
1492 free_vec (opt.rejects);
1493 free_vec (opt.excludes);
1494 free_vec (opt.includes);
1495 free_vec (opt.domains);
1496 free_vec (opt.follow_tags);
1497 free_vec (opt.ignore_tags);
1498 xfree_null (opt.progress_type);
1499 xfree_null (opt.ftp_user);
1500 xfree_null (opt.ftp_passwd);
1501 xfree_null (opt.ftp_proxy);
1502 xfree_null (opt.https_proxy);
1503 xfree_null (opt.http_proxy);
1504 free_vec (opt.no_proxy);
1505 xfree_null (opt.useragent);
1506 xfree_null (opt.referer);
1507 xfree_null (opt.http_user);
1508 xfree_null (opt.http_passwd);
1509 free_vec (opt.user_headers);
1511 xfree_null (opt.cert_file);
1512 xfree_null (opt.private_key);
1513 xfree_null (opt.ca_directory);
1514 xfree_null (opt.ca_cert);
1515 xfree_null (opt.random_file);
1516 xfree_null (opt.egd_file);
1518 xfree_null (opt.bind_address);
1519 xfree_null (opt.cookies_input);
1520 xfree_null (opt.cookies_output);
1521 xfree_null (opt.user);
1522 xfree_null (opt.passwd);
1523 #endif /* DEBUG_MALLOC */
1526 /* Unit testing routines. */
1531 test_cmd_spec_restrict_file_names()
1536 int expected_restrict_files_os;
1537 int expected_restrict_files_ctrl;
1538 int expected_restrict_files_case;
1541 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1542 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1543 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1544 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1547 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1552 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1555 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1556 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1557 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1558 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1560 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1561 res == test_array[i].result
1562 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1563 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1564 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1570 #endif /* TESTING */