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 2 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, write to the Free Software Foundation, Inc.,
18 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
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. */
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 depends on it. Also, be sure to add any entries
104 that allocate memory (e.g. cmd_string and cmd_vector) to the
105 cleanup() function below. */
110 bool (*action) (const char *, const char *, void *);
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 { "continue", &opt.always_rest, cmd_boolean },
132 { "convertlinks", &opt.convert_links, cmd_boolean },
133 { "contentdisposition", &opt.content_disposition, 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 { "mirror", NULL, cmd_spec_mirror },
186 { "netrc", &opt.netrc, cmd_boolean },
187 { "noclobber", &opt.noclobber, cmd_boolean },
188 { "noparent", &opt.no_parent, cmd_boolean },
189 { "noproxy", &opt.no_proxy, cmd_vector },
190 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
191 { "outputdocument", &opt.output_document, cmd_file },
192 { "pagerequisites", &opt.page_requisites, cmd_boolean },
193 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
194 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
195 { "password", &opt.passwd, cmd_string },
196 { "postdata", &opt.post_data, cmd_string },
197 { "postfile", &opt.post_file_name, cmd_file },
198 { "preferfamily", NULL, cmd_spec_prefer_family },
199 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
201 { "privatekey", &opt.private_key, cmd_file },
202 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
204 { "progress", &opt.progress_type, cmd_spec_progress },
205 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
206 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
207 { "proxypassword", &opt.proxy_passwd, cmd_string },
208 { "proxyuser", &opt.proxy_user, cmd_string },
209 { "quiet", &opt.quiet, cmd_boolean },
210 { "quota", &opt.quota, cmd_bytes_sum },
212 { "randomfile", &opt.random_file, cmd_file },
214 { "randomwait", &opt.random_wait, cmd_boolean },
215 { "readtimeout", &opt.read_timeout, cmd_time },
216 { "reclevel", &opt.reclevel, cmd_number_inf },
217 { "recursive", NULL, cmd_spec_recursive },
218 { "referer", &opt.referer, cmd_string },
219 { "reject", &opt.rejects, cmd_vector },
220 { "relativeonly", &opt.relative_only, cmd_boolean },
221 { "removelisting", &opt.remove_listing, cmd_boolean },
222 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
223 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
224 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
225 { "robots", &opt.use_robots, cmd_boolean },
226 { "savecookies", &opt.cookies_output, cmd_file },
227 { "saveheaders", &opt.save_headers, cmd_boolean },
229 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
231 { "serverresponse", &opt.server_response, cmd_boolean },
232 { "spanhosts", &opt.spanhost, cmd_boolean },
233 { "spider", &opt.spider, cmd_boolean },
234 { "strictcomments", &opt.strict_comments, cmd_boolean },
235 { "timeout", NULL, cmd_spec_timeout },
236 { "timestamping", &opt.timestamping, cmd_boolean },
237 { "tries", &opt.ntry, cmd_number_inf },
238 { "useproxy", &opt.use_proxy, cmd_boolean },
239 { "user", &opt.user, cmd_string },
240 { "useragent", NULL, cmd_spec_useragent },
241 { "verbose", NULL, cmd_spec_verbose },
242 { "wait", &opt.wait, cmd_time },
243 { "waitretry", &opt.waitretry, cmd_time }
246 /* Look up CMDNAME in the commands[] and return its position in the
247 array. If CMDNAME is not found, return -1. */
250 command_by_name (const char *cmdname)
252 /* Use binary search for speed. Wget has ~100 commands, which
253 guarantees a worst case performance of 7 string comparisons. */
254 int lo = 0, hi = countof (commands) - 1;
258 int mid = (lo + hi) >> 1;
259 int cmp = strcasecmp (cmdname, commands[mid].name);
270 /* Reset the variables to default values. */
276 /* Most of the default values are 0 (and 0.0, NULL, and false).
277 Just reset everything, and fill in the non-zero values. Note
278 that initializing pointers to NULL this way is technically
279 illegal, but porting Wget to a machine where NULL is not all-zero
280 bit pattern will be the least of the implementors' worries. */
287 opt.add_hostdir = true;
291 opt.http_keep_alive = true;
292 opt.use_proxy = true;
293 tmp = getenv ("no_proxy");
295 opt.no_proxy = sepstring (tmp);
296 opt.allow_cache = true;
298 opt.read_timeout = 900;
299 opt.use_robots = true;
301 opt.remove_listing = true;
303 opt.dot_bytes = 1024;
304 opt.dot_spacing = 10;
305 opt.dots_in_line = 50;
307 opt.dns_cache = true;
311 opt.check_cert = true;
314 /* The default for file name restriction defaults to the OS type. */
315 #if !defined(WINDOWS) && !defined(__CYGWIN__)
316 opt.restrict_files_os = restrict_unix;
318 opt.restrict_files_os = restrict_windows;
320 opt.restrict_files_ctrl = true;
321 opt.restrict_files_case = restrict_no_case_restriction;
323 opt.content_disposition = true;
326 /* Return the user's home directory (strdup-ed), or NULL if none is
331 char *home = getenv ("HOME");
336 /* If HOME is not defined, try getting it from the password
338 struct passwd *pwd = getpwuid (getuid ());
339 if (!pwd || !pwd->pw_dir)
343 /* Under Windows, if $HOME isn't defined, use the directory where
344 `wget.exe' resides. */
349 return home ? xstrdup (home) : NULL;
352 /* Return the path to the user's .wgetrc. This is either the value of
353 `WGETRC' environment variable, or `$HOME/.wgetrc'.
355 If the `WGETRC' variable exists but the file does not exist, the
356 function will exit(). */
358 wgetrc_file_name (void)
363 /* Try the environment. */
364 env = getenv ("WGETRC");
367 if (!file_exists_p (env))
369 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
373 return xstrdup (env);
376 /* If that failed, try $HOME/.wgetrc. */
379 file = aprintf ("%s/.wgetrc", home);
383 /* Under Windows, if we still haven't found .wgetrc, look for the file
384 `wget.ini' in the directory where `wget.exe' resides; we do this for
385 backward compatibility with previous versions of Wget.
386 SYSTEM_WGETRC should not be defined under WINDOWS. */
387 if (!file || !file_exists_p (file))
393 file = aprintf ("%s/wget.ini", home);
399 if (!file_exists_p (file))
407 /* Return values of parse_line. */
415 static enum parse_line parse_line (const char *, char **, char **, int *);
416 static bool setval_internal (int, const char *, const char *);
418 /* Initialize variables from a wgetrc file. Returns zero (failure) if
419 there were errors in the file. */
422 run_wgetrc (const char *file)
429 fp = fopen (file, "rb");
432 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
433 file, strerror (errno));
434 return true; /* not a fatal error */
436 enable_tilde_expansion = true;
438 while ((line = read_whole_line (fp)) != NULL)
440 char *com = NULL, *val = NULL;
443 /* Parse the line. */
444 switch (parse_line (line, &com, &val, &comind))
447 /* If everything is OK, set the value. */
448 if (!setval_internal (comind, com, val))
450 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
451 exec_name, file, ln);
455 case line_syntax_error:
456 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
457 exec_name, file, ln);
460 case line_unknown_command:
461 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
462 exec_name, com, file, ln);
475 enable_tilde_expansion = false;
481 /* Initialize the defaults and run the system wgetrc and user's own
489 /* Load the hard-coded defaults. */
492 /* If SYSTEM_WGETRC is defined, use it. */
494 if (file_exists_p (SYSTEM_WGETRC))
495 ok &= run_wgetrc (SYSTEM_WGETRC);
497 /* Override it with your own, if one exists. */
498 file = wgetrc_file_name ();
501 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
502 something like realpath() before comparing them with `strcmp' */
504 if (!strcmp (file, SYSTEM_WGETRC))
506 fprintf (stderr, _("\
507 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
512 ok &= run_wgetrc (file);
514 /* If there were errors processing either `.wgetrc', abort. */
522 /* Remove dashes and underscores from S, modifying S in the
528 char *t = s; /* t - tortoise */
529 char *h = s; /* h - hare */
531 if (*h == '_' || *h == '-')
538 /* Parse the line pointed by line, with the syntax:
539 <sp>* command <sp>* = <sp>* value <sp>*
540 Uses malloc to allocate space for command and value.
542 Returns one of line_ok, line_empty, line_syntax_error, or
543 line_unknown_command.
545 In case of line_ok, *COM and *VAL point to freshly allocated
546 strings, and *COMIND points to com's index. In case of error or
547 empty line, their values are unmodified. */
549 static enum parse_line
550 parse_line (const char *line, char **com, char **val, int *comind)
553 const char *end = line + strlen (line);
554 const char *cmdstart, *cmdend;
555 const char *valstart, *valend;
560 /* Skip leading and trailing whitespace. */
561 while (*line && ISSPACE (*line))
563 while (end > line && ISSPACE (end[-1]))
566 /* Skip empty lines and comments. */
567 if (!*line || *line == '#')
573 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
577 /* Skip '=', as well as any space before or after it. */
578 while (p < end && ISSPACE (*p))
580 if (p == end || *p != '=')
581 return line_syntax_error;
583 while (p < end && ISSPACE (*p))
589 /* The syntax is valid (even though the command might not be). Fill
590 in the command name and value. */
591 *com = strdupdelim (cmdstart, cmdend);
592 *val = strdupdelim (valstart, valend);
594 /* The line now known to be syntactically correct. Check whether
595 the command is valid. */
596 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
598 ind = command_by_name (cmdcopy);
600 return line_unknown_command;
602 /* Report success to the caller. */
607 /* Run commands[comind].action. */
610 setval_internal (int comind, const char *com, const char *val)
612 assert (0 <= comind && comind < countof (commands));
613 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
614 return commands[comind].action (com, val, commands[comind].place);
617 /* Run command COM with value VAL. If running the command produces an
618 error, report the error and exit.
620 This is intended to be called from main() to modify Wget's behavior
621 through command-line switches. Since COM is hard-coded in main(),
622 it is not canonicalized, and this aborts when COM is not found.
624 If COMIND's are exported to init.h, this function will be changed
625 to accept COMIND directly. */
628 setoptval (const char *com, const char *val, const char *optname)
630 /* Prepend "--" to OPTNAME. */
631 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
634 strcpy (dd_optname + 2, optname);
636 assert (val != NULL);
637 if (!setval_internal (command_by_name (com), dd_optname, val))
641 /* Parse OPT into command and value and run it. For example,
642 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
643 This is used by the `--execute' flag in main.c. */
646 run_command (const char *opt)
650 switch (parse_line (opt, &com, &val, &comind))
653 if (!setval_internal (comind, com, val))
659 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
665 /* Generic helper functions, for use with `commands'. */
667 /* Forward declarations: */
672 static bool decode_string (const char *, const struct decode_item *, int, int *);
673 static bool simple_atoi (const char *, const char *, int *);
674 static bool simple_atof (const char *, const char *, double *);
676 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
678 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
679 && TOLOWER((p)[1]) == (c1) \
682 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
683 && TOLOWER((p)[1]) == (c1) \
684 && TOLOWER((p)[2]) == (c2) \
688 /* Store the boolean value from VAL to PLACE. COM is ignored,
689 except for error messages. */
691 cmd_boolean (const char *com, const char *val, void *place)
695 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
696 /* "on", "yes" and "1" mean true. */
698 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
699 /* "off", "no" and "0" mean false. */
704 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
705 exec_name, com, val);
709 *(bool *) place = value;
713 /* Set the non-negative integer value from VAL to PLACE. With
714 incorrect specification, the number remains unchanged. */
716 cmd_number (const char *com, const char *val, void *place)
718 if (!simple_atoi (val, val + strlen (val), place)
719 || *(int *) place < 0)
721 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
722 exec_name, com, val);
728 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
730 cmd_number_inf (const char *com, const char *val, void *place)
732 if (!strcasecmp (val, "inf"))
737 return cmd_number (com, val, place);
740 /* Copy (strdup) the string at COM to a new location and place a
741 pointer to *PLACE. */
743 cmd_string (const char *com, const char *val, void *place)
745 char **pstring = (char **)place;
747 xfree_null (*pstring);
748 *pstring = xstrdup (val);
753 # define ISSEP(c) ((c) == '/')
755 # define ISSEP(c) ((c) == '/' || (c) == '\\')
758 /* Like the above, but handles tilde-expansion when reading a user's
759 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
760 gets expanded to the user's home directory. */
762 cmd_file (const char *com, const char *val, void *place)
764 char **pstring = (char **)place;
766 xfree_null (*pstring);
768 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
770 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
773 *pstring = xstrdup (val);
778 char *home = home_dir ();
782 homelen = strlen (home);
783 while (homelen && ISSEP (home[homelen - 1]))
784 home[--homelen] = '\0';
786 /* Skip the leading "~/". */
787 for (++val; ISSEP (*val); val++)
790 *pstring = concat_strings (home, "/", val, (char *) 0);
794 /* Convert "\" to "/". */
797 for (s = *pstring; *s; s++)
805 /* Like cmd_file, but strips trailing '/' characters. */
807 cmd_directory (const char *com, const char *val, void *place)
811 /* Call cmd_file() for tilde expansion and separator
812 canonicalization (backslash -> slash under Windows). These
813 things should perhaps be in a separate function. */
814 if (!cmd_file (com, val, place))
819 while (t > s && *--t == '/')
825 /* Split VAL by space to a vector of values, and append those values
826 to vector pointed to by the PLACE argument. If VAL is empty, the
827 PLACE vector is cleared instead. */
830 cmd_vector (const char *com, const char *val, void *place)
832 char ***pvec = (char ***)place;
835 *pvec = merge_vecs (*pvec, sepstring (val));
845 cmd_directory_vector (const char *com, const char *val, void *place)
847 char ***pvec = (char ***)place;
851 /* Strip the trailing slashes from directories. */
854 seps = sepstring (val);
855 for (t = seps; t && *t; t++)
857 int len = strlen (*t);
858 /* Skip degenerate case of root directory. */
861 if ((*t)[len - 1] == '/')
862 (*t)[len - 1] = '\0';
865 *pvec = merge_vecs (*pvec, seps);
875 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
876 "100k" or "2.5G" to a floating point number. */
879 parse_bytes_helper (const char *val, double *result)
882 const char *end = val + strlen (val);
884 /* Check for "inf". */
885 if (0 == strcmp (val, "inf"))
891 /* Strip trailing whitespace. */
892 while (val < end && ISSPACE (end[-1]))
897 switch (TOLOWER (end[-1]))
900 --end, mult = 1024.0;
903 --end, mult = 1048576.0;
906 --end, mult = 1073741824.0;
909 --end, mult = 1099511627776.0;
912 /* Not a recognized suffix: assume it's a digit. (If not,
913 simple_atof will raise an error.) */
917 /* Skip leading and trailing whitespace. */
918 while (val < end && ISSPACE (*val))
920 while (val < end && ISSPACE (end[-1]))
925 if (!simple_atof (val, end, &number) || number < 0)
928 *result = number * mult;
932 /* Parse VAL as a number and set its value to PLACE (which should
935 By default, the value is assumed to be in bytes. If "K", "M", or
936 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
937 1<<30, respectively. Floating point values are allowed and are
938 cast to integer before use. The idea is to be able to use things
939 like 1.5k instead of "1536".
941 The string "inf" is returned as 0.
943 In case of error, false is returned and memory pointed to by PLACE
944 remains unmodified. */
947 cmd_bytes (const char *com, const char *val, void *place)
950 if (!parse_bytes_helper (val, &byte_value))
952 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
953 exec_name, com, val);
956 *(wgint *)place = (wgint)byte_value;
960 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
961 SIZE_SUM. It works by converting the string to double, therefore
962 working with values up to 2^53-1 without loss of precision. This
963 value (8192 TB) is large enough to serve for a while. */
966 cmd_bytes_sum (const char *com, const char *val, void *place)
969 if (!parse_bytes_helper (val, &byte_value))
971 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
972 exec_name, com, val);
975 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
979 /* Store the value of VAL to *OUT. The value is a time period, by
980 default expressed in seconds, but also accepting suffixes "m", "h",
981 "d", and "w" for minutes, hours, days, and weeks respectively. */
984 cmd_time (const char *com, const char *val, void *place)
987 const char *end = val + strlen (val);
989 /* Strip trailing whitespace. */
990 while (val < end && ISSPACE (end[-1]))
996 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
997 exec_name, com, val);
1001 switch (TOLOWER (end[-1]))
1004 --end, mult = 1; /* seconds */
1007 --end, mult = 60; /* minutes */
1010 --end, mult = 3600; /* hours */
1013 --end, mult = 86400.0; /* days */
1016 --end, mult = 604800.0; /* weeks */
1019 /* Not a recognized suffix: assume it belongs to the number.
1020 (If not, simple_atof will raise an error.) */
1024 /* Skip leading and trailing whitespace. */
1025 while (val < end && ISSPACE (*val))
1027 while (val < end && ISSPACE (end[-1]))
1032 if (!simple_atof (val, end, &number))
1035 *(double *)place = number * mult;
1041 cmd_cert_type (const char *com, const char *val, void *place)
1043 static const struct decode_item choices[] = {
1044 { "pem", keyfile_pem },
1045 { "der", keyfile_asn1 },
1046 { "asn1", keyfile_asn1 },
1048 int ok = decode_string (val, choices, countof (choices), place);
1050 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1055 /* Specialized helper functions, used by `commands' to handle some
1056 options specially. */
1058 static bool check_user_specified_header (const char *);
1061 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1063 if (!cmd_boolean (com, val, &opt.dirstruct))
1065 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1066 must be affected inversely. */
1068 opt.no_dirstruct = false;
1070 opt.no_dirstruct = true;
1075 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1077 /* Empty value means reset the list of headers. */
1080 free_vec (opt.user_headers);
1081 opt.user_headers = NULL;
1085 if (!check_user_specified_header (val))
1087 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1088 exec_name, com, val);
1091 opt.user_headers = vec_append (opt.user_headers, val);
1096 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1098 int flag = cmd_boolean (com, val, &opt.htmlify);
1099 if (flag && !opt.htmlify)
1100 opt.remove_listing = false;
1104 /* Set the "mirror" mode. It means: recursive download, timestamping,
1105 no limit on max. recursion depth, and don't remove listings. */
1108 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1112 if (!cmd_boolean (com, val, &mirror))
1116 opt.recursive = true;
1117 if (!opt.no_dirstruct)
1118 opt.dirstruct = true;
1119 opt.timestamping = true;
1120 opt.reclevel = INFINITE_RECURSION;
1121 opt.remove_listing = false;
1126 /* Validate --prefer-family and set the choice. Allowed values are
1127 "IPv4", "IPv6", and "none". */
1130 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1132 static const struct decode_item choices[] = {
1133 { "IPv4", prefer_ipv4 },
1134 { "IPv6", prefer_ipv6 },
1135 { "none", prefer_none },
1137 int ok = decode_string (val, choices, countof (choices),
1138 (int *) &opt.prefer_family);
1140 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1144 /* Set progress.type to VAL, but verify that it's a valid progress
1145 implementation before that. */
1148 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1150 if (!valid_progress_implementation_p (val))
1152 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1153 exec_name, com, val);
1156 xfree_null (opt.progress_type);
1158 /* Don't call set_progress_implementation here. It will be called
1159 in main() when it becomes clear what the log output is. */
1160 opt.progress_type = xstrdup (val);
1164 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1165 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1169 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1171 if (!cmd_boolean (com, val, &opt.recursive))
1175 if (opt.recursive && !opt.no_dirstruct)
1176 opt.dirstruct = true;
1182 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1184 int restrict_os = opt.restrict_files_os;
1185 int restrict_ctrl = opt.restrict_files_ctrl;
1186 int restrict_case = opt.restrict_files_case;
1190 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1194 end = strchr (val, ',');
1196 end = val + strlen (val);
1198 if (VAL_IS ("unix"))
1199 restrict_os = restrict_unix;
1200 else if (VAL_IS ("windows"))
1201 restrict_os = restrict_windows;
1202 else if (VAL_IS ("lowercase"))
1203 restrict_case = restrict_lowercase;
1204 else if (VAL_IS ("uppercase"))
1205 restrict_case = restrict_uppercase;
1206 else if (VAL_IS ("nocontrol"))
1207 restrict_ctrl = false;
1211 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1212 exec_name, com, val);
1219 while (*val && *end);
1223 opt.restrict_files_os = restrict_os;
1224 opt.restrict_files_ctrl = restrict_ctrl;
1225 opt.restrict_files_case = restrict_case;
1232 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1234 static const struct decode_item choices[] = {
1235 { "auto", secure_protocol_auto },
1236 { "sslv2", secure_protocol_sslv2 },
1237 { "sslv3", secure_protocol_sslv3 },
1238 { "tlsv1", secure_protocol_tlsv1 },
1240 int ok = decode_string (val, choices, countof (choices), place);
1242 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1247 /* Set all three timeout values. */
1250 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1253 if (!cmd_time (com, val, &value))
1255 opt.read_timeout = value;
1256 opt.connect_timeout = value;
1257 opt.dns_timeout = value;
1262 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1264 /* Disallow embedded newlines. */
1265 if (strchr (val, '\n'))
1267 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1268 exec_name, com, val);
1271 xfree_null (opt.useragent);
1272 opt.useragent = xstrdup (val);
1276 /* The "verbose" option cannot be cmd_boolean because the variable is
1277 not bool -- it's of type int (-1 means uninitialized because of
1278 some random hackery for disallowing -q -v). */
1281 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1284 if (cmd_boolean (com, val, &flag))
1292 /* Miscellaneous useful routines. */
1294 /* A very simple atoi clone, more useful than atoi because it works on
1295 delimited strings, and has error reportage. Returns true on success,
1296 false on failure. If successful, stores result to *DEST. */
1299 simple_atoi (const char *beg, const char *end, int *dest)
1302 bool negative = false;
1303 const char *p = beg;
1305 while (p < end && ISSPACE (*p))
1307 if (p < end && (*p == '-' || *p == '+'))
1309 negative = (*p == '-');
1315 /* Read negative numbers in a separate loop because the most
1316 negative integer cannot be represented as a positive number. */
1319 for (; p < end && ISDIGIT (*p); p++)
1321 int next = (10 * result) + (*p - '0');
1323 return false; /* overflow */
1327 for (; p < end && ISDIGIT (*p); p++)
1329 int next = (10 * result) - (*p - '0');
1331 return false; /* underflow */
1342 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1343 doesn't handle exponential notation. Returns true on success,
1344 false on failure. In case of success, stores its result to
1348 simple_atof (const char *beg, const char *end, double *dest)
1352 bool negative = false;
1353 bool seen_dot = false;
1354 bool seen_digit = false;
1357 const char *p = beg;
1359 while (p < end && ISSPACE (*p))
1361 if (p < end && (*p == '-' || *p == '+'))
1363 negative = (*p == '-');
1367 for (; p < end; p++)
1373 result = (10 * result) + (ch - '0');
1375 result += (ch - '0') / (divider *= 10);
1397 /* Verify that the user-specified header in S is valid. It must
1398 contain a colon preceded by non-white-space characters and must not
1399 contain newlines. */
1402 check_user_specified_header (const char *s)
1406 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++)
1408 /* The header MUST contain `:' preceded by at least one
1409 non-whitespace character. */
1410 if (*p != ':' || p == s)
1412 /* The header MUST NOT contain newlines. */
1413 if (strchr (s, '\n'))
1418 /* Decode VAL into a number, according to ITEMS. */
1421 decode_string (const char *val, const struct decode_item *items, int itemcount,
1425 for (i = 0; i < itemcount; i++)
1426 if (0 == strcasecmp (val, items[i].name))
1428 *place = items[i].code;
1435 void cleanup_html_url (void);
1438 /* Free the memory allocated by global variables. */
1442 /* Free external resources, close files, etc. */
1445 fclose (output_stream);
1446 /* No need to check for error because Wget flushes its output (and
1447 checks for errors) after any data arrives. */
1449 /* We're exiting anyway so there's no real need to call free()
1450 hundreds of times. Skipping the frees will make Wget exit
1453 However, when detecting leaks, it's crucial to free() everything
1454 because then you can find the real leaks, i.e. the allocated
1455 memory which grows with the size of the program. */
1461 cleanup_html_url ();
1466 extern acc_t *netrc_list;
1467 free_netrc (netrc_list);
1469 xfree_null (opt.lfilename);
1470 xfree_null (opt.dir_prefix);
1471 xfree_null (opt.input_filename);
1472 xfree_null (opt.output_document);
1473 free_vec (opt.accepts);
1474 free_vec (opt.rejects);
1475 free_vec (opt.excludes);
1476 free_vec (opt.includes);
1477 free_vec (opt.domains);
1478 free_vec (opt.follow_tags);
1479 free_vec (opt.ignore_tags);
1480 xfree_null (opt.progress_type);
1481 xfree_null (opt.ftp_user);
1482 xfree_null (opt.ftp_passwd);
1483 xfree_null (opt.ftp_proxy);
1484 xfree_null (opt.https_proxy);
1485 xfree_null (opt.http_proxy);
1486 free_vec (opt.no_proxy);
1487 xfree_null (opt.useragent);
1488 xfree_null (opt.referer);
1489 xfree_null (opt.http_user);
1490 xfree_null (opt.http_passwd);
1491 free_vec (opt.user_headers);
1493 xfree_null (opt.cert_file);
1494 xfree_null (opt.private_key);
1495 xfree_null (opt.ca_directory);
1496 xfree_null (opt.ca_cert);
1497 xfree_null (opt.random_file);
1498 xfree_null (opt.egd_file);
1500 xfree_null (opt.bind_address);
1501 xfree_null (opt.cookies_input);
1502 xfree_null (opt.cookies_output);
1503 xfree_null (opt.user);
1504 xfree_null (opt.passwd);
1505 #endif /* DEBUG_MALLOC */
1508 /* Unit testing routines. */
1513 test_cmd_spec_restrict_file_names()
1518 int expected_restrict_files_os;
1519 int expected_restrict_files_ctrl;
1520 int expected_restrict_files_case;
1523 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1524 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1525 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1526 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1529 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1534 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1537 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1538 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1539 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1540 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1542 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1543 res == test_array[i].result
1544 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1545 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1546 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1552 #endif /* TESTING */