1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996-2006 Free Software Foundation, Inc.
4 This file is part of GNU Wget.
6 GNU Wget is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 3 of the License, or
9 (at your option) any later version.
11 GNU Wget is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with Wget. If not, see <http://www.gnu.org/licenses/>.
19 In addition, as a special exception, the Free Software Foundation
20 gives permission to link the code of its release of Wget with the
21 OpenSSL project's "OpenSSL" library (or with modified versions of it
22 that use the same license as the "OpenSSL" library), and distribute
23 the linked executables. You must obey the GNU General Public License
24 in all respects for all of the code used other than "OpenSSL". If you
25 modify this file, you may extend this exception to your version of the
26 file, but you are not obligated to do so. If you do not wish to do
27 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. */
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 }
247 /* Look up CMDNAME in the commands[] and return its position in the
248 array. If CMDNAME is not found, return -1. */
251 command_by_name (const char *cmdname)
253 /* Use binary search for speed. Wget has ~100 commands, which
254 guarantees a worst case performance of 7 string comparisons. */
255 int lo = 0, hi = countof (commands) - 1;
259 int mid = (lo + hi) >> 1;
260 int cmp = strcasecmp (cmdname, commands[mid].name);
271 /* Reset the variables to default values. */
277 /* Most of the default values are 0 (and 0.0, NULL, and false).
278 Just reset everything, and fill in the non-zero values. Note
279 that initializing pointers to NULL this way is technically
280 illegal, but porting Wget to a machine where NULL is not all-zero
281 bit pattern will be the least of the implementors' worries. */
288 opt.add_hostdir = true;
292 opt.http_keep_alive = true;
293 opt.use_proxy = true;
294 tmp = getenv ("no_proxy");
296 opt.no_proxy = sepstring (tmp);
297 opt.allow_cache = true;
299 opt.read_timeout = 900;
300 opt.use_robots = true;
302 opt.remove_listing = true;
304 opt.dot_bytes = 1024;
305 opt.dot_spacing = 10;
306 opt.dots_in_line = 50;
308 opt.dns_cache = true;
312 opt.check_cert = true;
315 /* The default for file name restriction defaults to the OS type. */
316 #if !defined(WINDOWS) && !defined(__CYGWIN__)
317 opt.restrict_files_os = restrict_unix;
319 opt.restrict_files_os = restrict_windows;
321 opt.restrict_files_ctrl = true;
322 opt.restrict_files_case = restrict_no_case_restriction;
324 opt.content_disposition = true;
325 opt.max_redirect = 20;
328 /* Return the user's home directory (strdup-ed), or NULL if none is
333 char *home = getenv ("HOME");
338 /* If HOME is not defined, try getting it from the password
340 struct passwd *pwd = getpwuid (getuid ());
341 if (!pwd || !pwd->pw_dir)
345 /* Under Windows, if $HOME isn't defined, use the directory where
346 `wget.exe' resides. */
351 return home ? xstrdup (home) : NULL;
354 /* Return the path to the user's .wgetrc. This is either the value of
355 `WGETRC' environment variable, or `$HOME/.wgetrc'.
357 If the `WGETRC' variable exists but the file does not exist, the
358 function will exit(). */
360 wgetrc_file_name (void)
365 /* Try the environment. */
366 env = getenv ("WGETRC");
369 if (!file_exists_p (env))
371 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
375 return xstrdup (env);
378 /* If that failed, try $HOME/.wgetrc. */
381 file = aprintf ("%s/.wgetrc", home);
385 /* Under Windows, if we still haven't found .wgetrc, look for the file
386 `wget.ini' in the directory where `wget.exe' resides; we do this for
387 backward compatibility with previous versions of Wget.
388 SYSTEM_WGETRC should not be defined under WINDOWS. */
389 if (!file || !file_exists_p (file))
395 file = aprintf ("%s/wget.ini", home);
401 if (!file_exists_p (file))
409 /* Return values of parse_line. */
417 static enum parse_line parse_line (const char *, char **, char **, int *);
418 static bool setval_internal (int, const char *, const char *);
420 /* Initialize variables from a wgetrc file. Returns zero (failure) if
421 there were errors in the file. */
424 run_wgetrc (const char *file)
431 fp = fopen (file, "rb");
434 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
435 file, strerror (errno));
436 return true; /* not a fatal error */
438 enable_tilde_expansion = true;
440 while ((line = read_whole_line (fp)) != NULL)
442 char *com = NULL, *val = NULL;
445 /* Parse the line. */
446 switch (parse_line (line, &com, &val, &comind))
449 /* If everything is OK, set the value. */
450 if (!setval_internal (comind, com, val))
452 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
453 exec_name, file, ln);
457 case line_syntax_error:
458 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
459 exec_name, file, ln);
462 case line_unknown_command:
463 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
464 exec_name, com, file, ln);
477 enable_tilde_expansion = false;
483 /* Initialize the defaults and run the system wgetrc and user's own
491 /* Load the hard-coded defaults. */
494 /* If SYSTEM_WGETRC is defined, use it. */
496 if (file_exists_p (SYSTEM_WGETRC))
497 ok &= run_wgetrc (SYSTEM_WGETRC);
499 /* Override it with your own, if one exists. */
500 file = wgetrc_file_name ();
503 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
504 something like realpath() before comparing them with `strcmp' */
506 if (!strcmp (file, SYSTEM_WGETRC))
508 fprintf (stderr, _("\
509 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
514 ok &= run_wgetrc (file);
516 /* If there were errors processing either `.wgetrc', abort. */
524 /* Remove dashes and underscores from S, modifying S in the
530 char *t = s; /* t - tortoise */
531 char *h = s; /* h - hare */
533 if (*h == '_' || *h == '-')
540 /* Parse the line pointed by line, with the syntax:
541 <sp>* command <sp>* = <sp>* value <sp>*
542 Uses malloc to allocate space for command and value.
544 Returns one of line_ok, line_empty, line_syntax_error, or
545 line_unknown_command.
547 In case of line_ok, *COM and *VAL point to freshly allocated
548 strings, and *COMIND points to com's index. In case of error or
549 empty line, their values are unmodified. */
551 static enum parse_line
552 parse_line (const char *line, char **com, char **val, int *comind)
555 const char *end = line + strlen (line);
556 const char *cmdstart, *cmdend;
557 const char *valstart, *valend;
562 /* Skip leading and trailing whitespace. */
563 while (*line && ISSPACE (*line))
565 while (end > line && ISSPACE (end[-1]))
568 /* Skip empty lines and comments. */
569 if (!*line || *line == '#')
575 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
579 /* Skip '=', as well as any space before or after it. */
580 while (p < end && ISSPACE (*p))
582 if (p == end || *p != '=')
583 return line_syntax_error;
585 while (p < end && ISSPACE (*p))
591 /* The syntax is valid (even though the command might not be). Fill
592 in the command name and value. */
593 *com = strdupdelim (cmdstart, cmdend);
594 *val = strdupdelim (valstart, valend);
596 /* The line now known to be syntactically correct. Check whether
597 the command is valid. */
598 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
600 ind = command_by_name (cmdcopy);
602 return line_unknown_command;
604 /* Report success to the caller. */
609 /* Run commands[comind].action. */
612 setval_internal (int comind, const char *com, const char *val)
614 assert (0 <= comind && comind < countof (commands));
615 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
616 return commands[comind].action (com, val, commands[comind].place);
619 /* Run command COM with value VAL. If running the command produces an
620 error, report the error and exit.
622 This is intended to be called from main() to modify Wget's behavior
623 through command-line switches. Since COM is hard-coded in main(),
624 it is not canonicalized, and this aborts when COM is not found.
626 If COMIND's are exported to init.h, this function will be changed
627 to accept COMIND directly. */
630 setoptval (const char *com, const char *val, const char *optname)
632 /* Prepend "--" to OPTNAME. */
633 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
636 strcpy (dd_optname + 2, optname);
638 assert (val != NULL);
639 if (!setval_internal (command_by_name (com), dd_optname, val))
643 /* Parse OPT into command and value and run it. For example,
644 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
645 This is used by the `--execute' flag in main.c. */
648 run_command (const char *opt)
652 switch (parse_line (opt, &com, &val, &comind))
655 if (!setval_internal (comind, com, val))
661 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
667 /* Generic helper functions, for use with `commands'. */
669 /* Forward declarations: */
674 static bool decode_string (const char *, const struct decode_item *, int, int *);
675 static bool simple_atoi (const char *, const char *, int *);
676 static bool simple_atof (const char *, const char *, double *);
678 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
680 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
681 && TOLOWER((p)[1]) == (c1) \
684 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
685 && TOLOWER((p)[1]) == (c1) \
686 && TOLOWER((p)[2]) == (c2) \
690 /* Store the boolean value from VAL to PLACE. COM is ignored,
691 except for error messages. */
693 cmd_boolean (const char *com, const char *val, void *place)
697 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
698 /* "on", "yes" and "1" mean true. */
700 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
701 /* "off", "no" and "0" mean false. */
706 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
707 exec_name, com, val);
711 *(bool *) place = value;
715 /* Set the non-negative integer value from VAL to PLACE. With
716 incorrect specification, the number remains unchanged. */
718 cmd_number (const char *com, const char *val, void *place)
720 if (!simple_atoi (val, val + strlen (val), place)
721 || *(int *) place < 0)
723 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
724 exec_name, com, val);
730 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
732 cmd_number_inf (const char *com, const char *val, void *place)
734 if (!strcasecmp (val, "inf"))
739 return cmd_number (com, val, place);
742 /* Copy (strdup) the string at COM to a new location and place a
743 pointer to *PLACE. */
745 cmd_string (const char *com, const char *val, void *place)
747 char **pstring = (char **)place;
749 xfree_null (*pstring);
750 *pstring = xstrdup (val);
755 # define ISSEP(c) ((c) == '/')
757 # define ISSEP(c) ((c) == '/' || (c) == '\\')
760 /* Like the above, but handles tilde-expansion when reading a user's
761 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
762 gets expanded to the user's home directory. */
764 cmd_file (const char *com, const char *val, void *place)
766 char **pstring = (char **)place;
768 xfree_null (*pstring);
770 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
772 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
775 *pstring = xstrdup (val);
780 char *home = home_dir ();
784 homelen = strlen (home);
785 while (homelen && ISSEP (home[homelen - 1]))
786 home[--homelen] = '\0';
788 /* Skip the leading "~/". */
789 for (++val; ISSEP (*val); val++)
792 *pstring = concat_strings (home, "/", val, (char *) 0);
796 /* Convert "\" to "/". */
799 for (s = *pstring; *s; s++)
807 /* Like cmd_file, but strips trailing '/' characters. */
809 cmd_directory (const char *com, const char *val, void *place)
813 /* Call cmd_file() for tilde expansion and separator
814 canonicalization (backslash -> slash under Windows). These
815 things should perhaps be in a separate function. */
816 if (!cmd_file (com, val, place))
821 while (t > s && *--t == '/')
827 /* Split VAL by space to a vector of values, and append those values
828 to vector pointed to by the PLACE argument. If VAL is empty, the
829 PLACE vector is cleared instead. */
832 cmd_vector (const char *com, const char *val, void *place)
834 char ***pvec = (char ***)place;
837 *pvec = merge_vecs (*pvec, sepstring (val));
847 cmd_directory_vector (const char *com, const char *val, void *place)
849 char ***pvec = (char ***)place;
853 /* Strip the trailing slashes from directories. */
856 seps = sepstring (val);
857 for (t = seps; t && *t; t++)
859 int len = strlen (*t);
860 /* Skip degenerate case of root directory. */
863 if ((*t)[len - 1] == '/')
864 (*t)[len - 1] = '\0';
867 *pvec = merge_vecs (*pvec, seps);
877 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
878 "100k" or "2.5G" to a floating point number. */
881 parse_bytes_helper (const char *val, double *result)
884 const char *end = val + strlen (val);
886 /* Check for "inf". */
887 if (0 == strcmp (val, "inf"))
893 /* Strip trailing whitespace. */
894 while (val < end && ISSPACE (end[-1]))
899 switch (TOLOWER (end[-1]))
902 --end, mult = 1024.0;
905 --end, mult = 1048576.0;
908 --end, mult = 1073741824.0;
911 --end, mult = 1099511627776.0;
914 /* Not a recognized suffix: assume it's a digit. (If not,
915 simple_atof will raise an error.) */
919 /* Skip leading and trailing whitespace. */
920 while (val < end && ISSPACE (*val))
922 while (val < end && ISSPACE (end[-1]))
927 if (!simple_atof (val, end, &number) || number < 0)
930 *result = number * mult;
934 /* Parse VAL as a number and set its value to PLACE (which should
937 By default, the value is assumed to be in bytes. If "K", "M", or
938 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
939 1<<30, respectively. Floating point values are allowed and are
940 cast to integer before use. The idea is to be able to use things
941 like 1.5k instead of "1536".
943 The string "inf" is returned as 0.
945 In case of error, false is returned and memory pointed to by PLACE
946 remains unmodified. */
949 cmd_bytes (const char *com, const char *val, void *place)
952 if (!parse_bytes_helper (val, &byte_value))
954 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
955 exec_name, com, val);
958 *(wgint *)place = (wgint)byte_value;
962 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
963 SIZE_SUM. It works by converting the string to double, therefore
964 working with values up to 2^53-1 without loss of precision. This
965 value (8192 TB) is large enough to serve for a while. */
968 cmd_bytes_sum (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, val);
977 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
981 /* Store the value of VAL to *OUT. The value is a time period, by
982 default expressed in seconds, but also accepting suffixes "m", "h",
983 "d", and "w" for minutes, hours, days, and weeks respectively. */
986 cmd_time (const char *com, const char *val, void *place)
989 const char *end = val + strlen (val);
991 /* Strip trailing whitespace. */
992 while (val < end && ISSPACE (end[-1]))
998 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
999 exec_name, com, val);
1003 switch (TOLOWER (end[-1]))
1006 --end, mult = 1; /* seconds */
1009 --end, mult = 60; /* minutes */
1012 --end, mult = 3600; /* hours */
1015 --end, mult = 86400.0; /* days */
1018 --end, mult = 604800.0; /* weeks */
1021 /* Not a recognized suffix: assume it belongs to the number.
1022 (If not, simple_atof will raise an error.) */
1026 /* Skip leading and trailing whitespace. */
1027 while (val < end && ISSPACE (*val))
1029 while (val < end && ISSPACE (end[-1]))
1034 if (!simple_atof (val, end, &number))
1037 *(double *)place = number * mult;
1043 cmd_cert_type (const char *com, const char *val, void *place)
1045 static const struct decode_item choices[] = {
1046 { "pem", keyfile_pem },
1047 { "der", keyfile_asn1 },
1048 { "asn1", keyfile_asn1 },
1050 int ok = decode_string (val, choices, countof (choices), place);
1052 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1057 /* Specialized helper functions, used by `commands' to handle some
1058 options specially. */
1060 static bool check_user_specified_header (const char *);
1063 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1065 if (!cmd_boolean (com, val, &opt.dirstruct))
1067 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1068 must be affected inversely. */
1070 opt.no_dirstruct = false;
1072 opt.no_dirstruct = true;
1077 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1079 /* Empty value means reset the list of headers. */
1082 free_vec (opt.user_headers);
1083 opt.user_headers = NULL;
1087 if (!check_user_specified_header (val))
1089 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1090 exec_name, com, val);
1093 opt.user_headers = vec_append (opt.user_headers, val);
1098 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1100 int flag = cmd_boolean (com, val, &opt.htmlify);
1101 if (flag && !opt.htmlify)
1102 opt.remove_listing = false;
1106 /* Set the "mirror" mode. It means: recursive download, timestamping,
1107 no limit on max. recursion depth, and don't remove listings. */
1110 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1114 if (!cmd_boolean (com, val, &mirror))
1118 opt.recursive = true;
1119 if (!opt.no_dirstruct)
1120 opt.dirstruct = true;
1121 opt.timestamping = true;
1122 opt.reclevel = INFINITE_RECURSION;
1123 opt.remove_listing = false;
1128 /* Validate --prefer-family and set the choice. Allowed values are
1129 "IPv4", "IPv6", and "none". */
1132 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1134 static const struct decode_item choices[] = {
1135 { "IPv4", prefer_ipv4 },
1136 { "IPv6", prefer_ipv6 },
1137 { "none", prefer_none },
1139 int prefer_family = prefer_ipv4;
1140 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1142 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1143 opt.prefer_family = prefer_family;
1147 /* Set progress.type to VAL, but verify that it's a valid progress
1148 implementation before that. */
1151 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1153 if (!valid_progress_implementation_p (val))
1155 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1156 exec_name, com, val);
1159 xfree_null (opt.progress_type);
1161 /* Don't call set_progress_implementation here. It will be called
1162 in main() when it becomes clear what the log output is. */
1163 opt.progress_type = xstrdup (val);
1167 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1168 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1172 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1174 if (!cmd_boolean (com, val, &opt.recursive))
1178 if (opt.recursive && !opt.no_dirstruct)
1179 opt.dirstruct = true;
1185 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1187 int restrict_os = opt.restrict_files_os;
1188 int restrict_ctrl = opt.restrict_files_ctrl;
1189 int restrict_case = opt.restrict_files_case;
1193 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1197 end = strchr (val, ',');
1199 end = val + strlen (val);
1201 if (VAL_IS ("unix"))
1202 restrict_os = restrict_unix;
1203 else if (VAL_IS ("windows"))
1204 restrict_os = restrict_windows;
1205 else if (VAL_IS ("lowercase"))
1206 restrict_case = restrict_lowercase;
1207 else if (VAL_IS ("uppercase"))
1208 restrict_case = restrict_uppercase;
1209 else if (VAL_IS ("nocontrol"))
1210 restrict_ctrl = false;
1214 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1215 exec_name, com, val);
1222 while (*val && *end);
1226 opt.restrict_files_os = restrict_os;
1227 opt.restrict_files_ctrl = restrict_ctrl;
1228 opt.restrict_files_case = restrict_case;
1235 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1237 static const struct decode_item choices[] = {
1238 { "auto", secure_protocol_auto },
1239 { "sslv2", secure_protocol_sslv2 },
1240 { "sslv3", secure_protocol_sslv3 },
1241 { "tlsv1", secure_protocol_tlsv1 },
1243 int ok = decode_string (val, choices, countof (choices), place);
1245 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1250 /* Set all three timeout values. */
1253 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1256 if (!cmd_time (com, val, &value))
1258 opt.read_timeout = value;
1259 opt.connect_timeout = value;
1260 opt.dns_timeout = value;
1265 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1267 /* Disallow embedded newlines. */
1268 if (strchr (val, '\n'))
1270 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1271 exec_name, com, val);
1274 xfree_null (opt.useragent);
1275 opt.useragent = xstrdup (val);
1279 /* The "verbose" option cannot be cmd_boolean because the variable is
1280 not bool -- it's of type int (-1 means uninitialized because of
1281 some random hackery for disallowing -q -v). */
1284 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1287 if (cmd_boolean (com, val, &flag))
1295 /* Miscellaneous useful routines. */
1297 /* A very simple atoi clone, more useful than atoi because it works on
1298 delimited strings, and has error reportage. Returns true on success,
1299 false on failure. If successful, stores result to *DEST. */
1302 simple_atoi (const char *beg, const char *end, int *dest)
1305 bool negative = false;
1306 const char *p = beg;
1308 while (p < end && ISSPACE (*p))
1310 if (p < end && (*p == '-' || *p == '+'))
1312 negative = (*p == '-');
1318 /* Read negative numbers in a separate loop because the most
1319 negative integer cannot be represented as a positive number. */
1322 for (; p < end && ISDIGIT (*p); p++)
1324 int next = (10 * result) + (*p - '0');
1326 return false; /* overflow */
1330 for (; p < end && ISDIGIT (*p); p++)
1332 int next = (10 * result) - (*p - '0');
1334 return false; /* underflow */
1345 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1346 doesn't handle exponential notation. Returns true on success,
1347 false on failure. In case of success, stores its result to
1351 simple_atof (const char *beg, const char *end, double *dest)
1355 bool negative = false;
1356 bool seen_dot = false;
1357 bool seen_digit = false;
1360 const char *p = beg;
1362 while (p < end && ISSPACE (*p))
1364 if (p < end && (*p == '-' || *p == '+'))
1366 negative = (*p == '-');
1370 for (; p < end; p++)
1376 result = (10 * result) + (ch - '0');
1378 result += (ch - '0') / (divider *= 10);
1400 /* Verify that the user-specified header in S is valid. It must
1401 contain a colon preceded by non-white-space characters and must not
1402 contain newlines. */
1405 check_user_specified_header (const char *s)
1409 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++)
1411 /* The header MUST contain `:' preceded by at least one
1412 non-whitespace character. */
1413 if (*p != ':' || p == s)
1415 /* The header MUST NOT contain newlines. */
1416 if (strchr (s, '\n'))
1421 /* Decode VAL into a number, according to ITEMS. */
1424 decode_string (const char *val, const struct decode_item *items, int itemcount,
1428 for (i = 0; i < itemcount; i++)
1429 if (0 == strcasecmp (val, items[i].name))
1431 *place = items[i].code;
1438 void cleanup_html_url (void);
1441 /* Free the memory allocated by global variables. */
1445 /* Free external resources, close files, etc. */
1448 fclose (output_stream);
1449 /* No need to check for error because Wget flushes its output (and
1450 checks for errors) after any data arrives. */
1452 /* We're exiting anyway so there's no real need to call free()
1453 hundreds of times. Skipping the frees will make Wget exit
1456 However, when detecting leaks, it's crucial to free() everything
1457 because then you can find the real leaks, i.e. the allocated
1458 memory which grows with the size of the program. */
1464 cleanup_html_url ();
1469 extern acc_t *netrc_list;
1470 free_netrc (netrc_list);
1472 xfree_null (opt.lfilename);
1473 xfree_null (opt.dir_prefix);
1474 xfree_null (opt.input_filename);
1475 xfree_null (opt.output_document);
1476 free_vec (opt.accepts);
1477 free_vec (opt.rejects);
1478 free_vec (opt.excludes);
1479 free_vec (opt.includes);
1480 free_vec (opt.domains);
1481 free_vec (opt.follow_tags);
1482 free_vec (opt.ignore_tags);
1483 xfree_null (opt.progress_type);
1484 xfree_null (opt.ftp_user);
1485 xfree_null (opt.ftp_passwd);
1486 xfree_null (opt.ftp_proxy);
1487 xfree_null (opt.https_proxy);
1488 xfree_null (opt.http_proxy);
1489 free_vec (opt.no_proxy);
1490 xfree_null (opt.useragent);
1491 xfree_null (opt.referer);
1492 xfree_null (opt.http_user);
1493 xfree_null (opt.http_passwd);
1494 free_vec (opt.user_headers);
1496 xfree_null (opt.cert_file);
1497 xfree_null (opt.private_key);
1498 xfree_null (opt.ca_directory);
1499 xfree_null (opt.ca_cert);
1500 xfree_null (opt.random_file);
1501 xfree_null (opt.egd_file);
1503 xfree_null (opt.bind_address);
1504 xfree_null (opt.cookies_input);
1505 xfree_null (opt.cookies_output);
1506 xfree_null (opt.user);
1507 xfree_null (opt.passwd);
1508 #endif /* DEBUG_MALLOC */
1511 /* Unit testing routines. */
1516 test_cmd_spec_restrict_file_names()
1521 int expected_restrict_files_os;
1522 int expected_restrict_files_ctrl;
1523 int expected_restrict_files_case;
1526 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1527 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1528 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1529 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1532 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1537 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1540 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1541 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1542 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1543 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1545 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1546 res == test_array[i].result
1547 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1548 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1549 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1555 #endif /* TESTING */