1 /* Reading/parsing the initialization file.
2 Copyright (C) 2005 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
18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, 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 */
55 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
56 otherwise, it will be performed by the shell. This variable will
57 be set by the wgetrc-reading function. */
59 static int enable_tilde_expansion;
62 #define CMD_DECLARE(func) static int func (const char *, const char *, void *)
64 CMD_DECLARE (cmd_boolean);
65 CMD_DECLARE (cmd_bytes);
66 CMD_DECLARE (cmd_bytes_large);
68 CMD_DECLARE (cmd_cert_type);
70 CMD_DECLARE (cmd_directory_vector);
71 CMD_DECLARE (cmd_number);
72 CMD_DECLARE (cmd_number_inf);
73 CMD_DECLARE (cmd_string);
74 CMD_DECLARE (cmd_file);
75 CMD_DECLARE (cmd_directory);
76 CMD_DECLARE (cmd_time);
77 CMD_DECLARE (cmd_vector);
79 CMD_DECLARE (cmd_spec_dirstruct);
80 CMD_DECLARE (cmd_spec_header);
81 CMD_DECLARE (cmd_spec_htmlify);
82 CMD_DECLARE (cmd_spec_mirror);
83 CMD_DECLARE (cmd_spec_prefer_family);
84 CMD_DECLARE (cmd_spec_progress);
85 CMD_DECLARE (cmd_spec_recursive);
86 CMD_DECLARE (cmd_spec_restrict_file_names);
88 CMD_DECLARE (cmd_spec_secure_protocol);
90 CMD_DECLARE (cmd_spec_timeout);
91 CMD_DECLARE (cmd_spec_useragent);
93 /* List of recognized commands, each consisting of name, place and
94 function. When adding a new command, simply add it to the list,
95 but be sure to keep the list sorted alphabetically, as
96 command_by_name depends on it. Also, be sure to add any entries
97 that allocate memory (e.g. cmd_string and cmd_vector) to the
98 cleanup() function below. */
103 int (*action) (const char *, const char *, void *);
105 { "accept", &opt.accepts, cmd_vector },
106 { "addhostdir", &opt.add_hostdir, cmd_boolean },
107 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
108 { "background", &opt.background, cmd_boolean },
109 { "backupconverted", &opt.backup_converted, cmd_boolean },
110 { "backups", &opt.backups, cmd_number },
111 { "base", &opt.base_href, cmd_string },
112 { "bindaddress", &opt.bind_address, cmd_string },
114 { "cacertificate", &opt.ca_cert, cmd_file },
116 { "cache", &opt.allow_cache, cmd_boolean },
118 { "cadirectory", &opt.ca_directory, cmd_directory },
119 { "certificate", &opt.cert_file, cmd_file },
120 { "certificatetype", &opt.cert_type, cmd_cert_type },
121 { "checkcertificate", &opt.check_cert, cmd_boolean },
123 { "connecttimeout", &opt.connect_timeout, cmd_time },
124 { "continue", &opt.always_rest, cmd_boolean },
125 { "convertlinks", &opt.convert_links, cmd_boolean },
126 { "cookies", &opt.cookies, cmd_boolean },
127 { "cutdirs", &opt.cut_dirs, cmd_number },
129 { "debug", &opt.debug, cmd_boolean },
131 { "deleteafter", &opt.delete_after, cmd_boolean },
132 { "dirprefix", &opt.dir_prefix, cmd_directory },
133 { "dirstruct", NULL, cmd_spec_dirstruct },
134 { "dnscache", &opt.dns_cache, cmd_boolean },
135 { "dnstimeout", &opt.dns_timeout, cmd_time },
136 { "domains", &opt.domains, cmd_vector },
137 { "dotbytes", &opt.dot_bytes, cmd_bytes },
138 { "dotsinline", &opt.dots_in_line, cmd_number },
139 { "dotspacing", &opt.dot_spacing, cmd_number },
140 { "dotstyle", &opt.dot_style, cmd_string },
142 { "egdfile", &opt.egd_file, cmd_file },
144 { "excludedirectories", &opt.excludes, cmd_directory_vector },
145 { "excludedomains", &opt.exclude_domains, cmd_vector },
146 { "followftp", &opt.follow_ftp, cmd_boolean },
147 { "followtags", &opt.follow_tags, cmd_vector },
148 { "forcehtml", &opt.force_html, cmd_boolean },
149 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
150 { "ftppassword", &opt.ftp_passwd, cmd_string },
151 { "ftpproxy", &opt.ftp_proxy, cmd_string },
152 { "ftpuser", &opt.ftp_user, cmd_string },
153 { "glob", &opt.ftp_glob, cmd_boolean },
154 { "header", NULL, cmd_spec_header },
155 { "htmlextension", &opt.html_extension, cmd_boolean },
156 { "htmlify", NULL, cmd_spec_htmlify },
157 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
158 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
159 { "httppassword", &opt.http_passwd, cmd_string },
160 { "httpproxy", &opt.http_proxy, cmd_string },
161 { "httpsproxy", &opt.https_proxy, cmd_string },
162 { "httpuser", &opt.http_user, cmd_string },
163 { "ignorelength", &opt.ignore_length, cmd_boolean },
164 { "ignoretags", &opt.ignore_tags, cmd_vector },
165 { "includedirectories", &opt.includes, cmd_directory_vector },
167 { "inet4only", &opt.ipv4_only, cmd_boolean },
168 { "inet6only", &opt.ipv6_only, cmd_boolean },
170 { "input", &opt.input_filename, cmd_file },
171 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
172 { "killlonger", &opt.kill_longer, cmd_boolean },
173 { "limitrate", &opt.limit_rate, cmd_bytes },
174 { "loadcookies", &opt.cookies_input, cmd_file },
175 { "logfile", &opt.lfilename, cmd_file },
176 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
177 { "mirror", NULL, cmd_spec_mirror },
178 { "netrc", &opt.netrc, cmd_boolean },
179 { "noclobber", &opt.noclobber, cmd_boolean },
180 { "noparent", &opt.no_parent, cmd_boolean },
181 { "noproxy", &opt.no_proxy, cmd_vector },
182 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
183 { "outputdocument", &opt.output_document, cmd_file },
184 { "pagerequisites", &opt.page_requisites, cmd_boolean },
185 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
186 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
187 { "password", &opt.passwd, cmd_string },
188 { "postdata", &opt.post_data, cmd_string },
189 { "postfile", &opt.post_file_name, cmd_file },
190 { "preferfamily", NULL, cmd_spec_prefer_family },
191 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
193 { "privatekey", &opt.private_key, cmd_file },
194 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
196 { "progress", &opt.progress_type, cmd_spec_progress },
197 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
198 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
199 { "proxypassword", &opt.proxy_passwd, cmd_string },
200 { "proxyuser", &opt.proxy_user, cmd_string },
201 { "quiet", &opt.quiet, cmd_boolean },
202 { "quota", &opt.quota, cmd_bytes_large },
204 { "randomfile", &opt.random_file, cmd_file },
206 { "randomwait", &opt.random_wait, cmd_boolean },
207 { "readtimeout", &opt.read_timeout, cmd_time },
208 { "reclevel", &opt.reclevel, cmd_number_inf },
209 { "recursive", NULL, cmd_spec_recursive },
210 { "referer", &opt.referer, cmd_string },
211 { "reject", &opt.rejects, cmd_vector },
212 { "relativeonly", &opt.relative_only, cmd_boolean },
213 { "removelisting", &opt.remove_listing, cmd_boolean },
214 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
215 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
216 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
217 { "robots", &opt.use_robots, cmd_boolean },
218 { "savecookies", &opt.cookies_output, cmd_file },
219 { "saveheaders", &opt.save_headers, cmd_boolean },
221 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
223 { "serverresponse", &opt.server_response, cmd_boolean },
224 { "spanhosts", &opt.spanhost, cmd_boolean },
225 { "spider", &opt.spider, cmd_boolean },
226 { "strictcomments", &opt.strict_comments, cmd_boolean },
227 { "timeout", NULL, cmd_spec_timeout },
228 { "timestamping", &opt.timestamping, cmd_boolean },
229 { "tries", &opt.ntry, cmd_number_inf },
230 { "useproxy", &opt.use_proxy, cmd_boolean },
231 { "user", &opt.user, cmd_string },
232 { "useragent", NULL, cmd_spec_useragent },
233 { "verbose", &opt.verbose, cmd_boolean },
234 { "wait", &opt.wait, cmd_time },
235 { "waitretry", &opt.waitretry, cmd_time }
238 /* Look up CMDNAME in the commands[] and return its position in the
239 array. If CMDNAME is not found, return -1. */
242 command_by_name (const char *cmdname)
244 /* Use binary search for speed. Wget has ~100 commands, which
245 guarantees a worst case performance of 7 string comparisons. */
246 int lo = 0, hi = countof (commands) - 1;
250 int mid = (lo + hi) >> 1;
251 int cmp = strcasecmp (cmdname, commands[mid].name);
262 /* Reset the variables to default values. */
268 /* Most of the default values are 0. Just reset everything, and
269 fill in the non-zero values. Note that initializing pointers to
270 NULL this way is technically illegal, but porting Wget to a
271 machine where NULL is not all-zero bit pattern will be the least
272 of the implementors' worries. */
283 opt.http_keep_alive = 1;
285 tmp = getenv ("no_proxy");
287 opt.no_proxy = sepstring (tmp);
290 opt.read_timeout = 900;
293 opt.remove_listing = 1;
295 opt.dot_bytes = 1024;
296 opt.dot_spacing = 10;
297 opt.dots_in_line = 50;
306 /* The default for file name restriction defaults to the OS type. */
307 #if !defined(WINDOWS) && !defined(__CYGWIN__)
308 opt.restrict_files_os = restrict_unix;
310 opt.restrict_files_os = restrict_windows;
312 opt.restrict_files_ctrl = 1;
315 /* Return the user's home directory (strdup-ed), or NULL if none is
320 char *home = getenv ("HOME");
325 /* If HOME is not defined, try getting it from the password
327 struct passwd *pwd = getpwuid (getuid ());
328 if (!pwd || !pwd->pw_dir)
332 /* Under Windows, if $HOME isn't defined, use the directory where
333 `wget.exe' resides. */
338 return home ? xstrdup (home) : NULL;
341 /* Return the path to the user's .wgetrc. This is either the value of
342 `WGETRC' environment variable, or `$HOME/.wgetrc'.
344 If the `WGETRC' variable exists but the file does not exist, the
345 function will exit(). */
347 wgetrc_file_name (void)
352 /* Try the environment. */
353 env = getenv ("WGETRC");
356 if (!file_exists_p (env))
358 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
362 return xstrdup (env);
365 /* If that failed, try $HOME/.wgetrc. */
368 file = aprintf ("%s/.wgetrc", home);
372 /* Under Windows, if we still haven't found .wgetrc, look for the file
373 `wget.ini' in the directory where `wget.exe' resides; we do this for
374 backward compatibility with previous versions of Wget.
375 SYSTEM_WGETRC should not be defined under WINDOWS. */
376 if (!file || !file_exists_p (file))
382 file = aprintf ("%s/wget.ini", home);
388 if (!file_exists_p (file))
396 /* Return values of parse_line. */
404 static enum parse_line parse_line (const char *, char **, char **, int *);
405 static int setval_internal (int, const char *, const char *);
407 /* Initialize variables from a wgetrc file. Returns zero (failure) if
408 there were errors in the file. */
411 run_wgetrc (const char *file)
418 fp = fopen (file, "rb");
421 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
422 file, strerror (errno));
423 return 1; /* not a fatal error */
425 enable_tilde_expansion = 1;
427 while ((line = read_whole_line (fp)) != NULL)
429 char *com = NULL, *val = NULL;
432 /* Parse the line. */
433 switch (parse_line (line, &com, &val, &comind))
436 /* If everything is OK, set the value. */
437 if (!setval_internal (comind, com, val))
439 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
440 exec_name, file, ln);
444 case line_syntax_error:
445 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
446 exec_name, file, ln);
449 case line_unknown_command:
450 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
451 exec_name, com, file, ln);
464 enable_tilde_expansion = 0;
470 /* Initialize the defaults and run the system wgetrc and user's own
478 /* Load the hard-coded defaults. */
481 /* If SYSTEM_WGETRC is defined, use it. */
483 if (file_exists_p (SYSTEM_WGETRC))
484 ok &= run_wgetrc (SYSTEM_WGETRC);
486 /* Override it with your own, if one exists. */
487 file = wgetrc_file_name ();
490 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
491 something like realpath() before comparing them with `strcmp' */
493 if (!strcmp (file, SYSTEM_WGETRC))
495 fprintf (stderr, _("\
496 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
501 ok &= run_wgetrc (file);
503 /* If there were errors processing either `.wgetrc', abort. */
511 /* Remove dashes and underscores from S, modifying S in the
517 char *t = s; /* t - tortoise */
518 char *h = s; /* h - hare */
520 if (*h == '_' || *h == '-')
527 /* Parse the line pointed by line, with the syntax:
528 <sp>* command <sp>* = <sp>* value <sp>*
529 Uses malloc to allocate space for command and value.
530 If the line is invalid, data is freed and 0 is returned.
532 Returns one of line_ok, line_empty, line_syntax_error, or
533 line_unknown_command.
535 In case of line_ok, *COM and *VAL point to freshly allocated
536 strings, and *COMIND points to com's index. In case of error or
537 empty line, their values are unmodified. */
539 static enum parse_line
540 parse_line (const char *line, char **com, char **val, int *comind)
543 const char *end = line + strlen (line);
544 const char *cmdstart, *cmdend;
545 const char *valstart, *valend;
550 /* Skip leading and trailing whitespace. */
551 while (*line && ISSPACE (*line))
553 while (end > line && ISSPACE (end[-1]))
556 /* Skip empty lines and comments. */
557 if (!*line || *line == '#')
563 while (p < end && (ISALPHA (*p) || *p == '_' || *p == '-'))
567 /* Skip '=', as well as any space before or after it. */
568 while (p < end && ISSPACE (*p))
570 if (p == end || *p != '=')
571 return line_syntax_error;
573 while (p < end && ISSPACE (*p))
579 /* The syntax is valid (even though the command might not be). Fill
580 in the command name and value. */
581 *com = strdupdelim (cmdstart, cmdend);
582 *val = strdupdelim (valstart, valend);
584 /* The line now known to be syntactically correct. Check whether
585 the command is valid. */
586 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
588 ind = command_by_name (cmdcopy);
590 return line_unknown_command;
592 /* Report success to the caller. */
597 /* Run commands[comind].action. */
600 setval_internal (int comind, const char *com, const char *val)
602 assert (0 <= comind && comind < countof (commands));
603 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
604 return commands[comind].action (com, val, commands[comind].place);
607 /* Run command COM with value VAL. If running the command produces an
608 error, report the error and exit.
610 This is intended to be called from main() to modify Wget's behavior
611 through command-line switches. Since COM is hard-coded in main(),
612 it is not canonicalized, and this aborts when COM is not found.
614 If COMIND's are exported to init.h, this function will be changed
615 to accept COMIND directly. */
618 setoptval (const char *com, const char *val, const char *optname)
620 /* Prepend "--" to OPTNAME. */
621 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
624 strcpy (dd_optname + 2, optname);
626 assert (val != NULL);
627 if (!setval_internal (command_by_name (com), dd_optname, val))
631 /* Parse OPT into command and value and run it. For example,
632 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
633 This is used by the `--execute' flag in main.c. */
636 run_command (const char *opt)
640 switch (parse_line (opt, &com, &val, &comind))
643 if (!setval_internal (comind, com, val))
649 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
655 /* Generic helper functions, for use with `commands'. */
657 /* Forward declarations: */
662 static int decode_string (const char *, const struct decode_item *, int, int *);
663 static int simple_atoi (const char *, const char *, int *);
664 static int simple_atof (const char *, const char *, double *);
666 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
668 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
669 && TOLOWER((p)[1]) == (c1) \
672 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
673 && TOLOWER((p)[1]) == (c1) \
674 && TOLOWER((p)[2]) == (c2) \
678 /* Store the boolean value from VAL to PLACE. COM is ignored,
679 except for error messages. */
681 cmd_boolean (const char *com, const char *val, void *place)
685 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
686 /* "on", "yes" and "1" mean true. */
688 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
689 /* "off", "no" and "0" mean false. */
694 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
695 exec_name, com, val);
699 *(int *)place = bool_value;
703 /* Set the non-negative integer value from VAL to PLACE. With
704 incorrect specification, the number remains unchanged. */
706 cmd_number (const char *com, const char *val, void *place)
708 if (!simple_atoi (val, val + strlen (val), place)
709 || *(int *) place < 0)
711 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
712 exec_name, com, val);
718 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
720 cmd_number_inf (const char *com, const char *val, void *place)
722 if (!strcasecmp (val, "inf"))
727 return cmd_number (com, val, place);
730 /* Copy (strdup) the string at COM to a new location and place a
731 pointer to *PLACE. */
733 cmd_string (const char *com, const char *val, void *place)
735 char **pstring = (char **)place;
737 xfree_null (*pstring);
738 *pstring = xstrdup (val);
743 # define ISSEP(c) ((c) == '/')
745 # define ISSEP(c) ((c) == '/' || (c) == '\\')
748 /* Like the above, but handles tilde-expansion when reading a user's
749 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
750 gets expanded to the user's home directory. */
752 cmd_file (const char *com, const char *val, void *place)
754 char **pstring = (char **)place;
756 xfree_null (*pstring);
758 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
760 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
763 *pstring = xstrdup (val);
768 char *home = home_dir ();
772 homelen = strlen (home);
773 while (homelen && ISSEP (home[homelen - 1]))
774 home[--homelen] = '\0';
776 /* Skip the leading "~/". */
777 for (++val; ISSEP (*val); val++)
780 *pstring = concat_strings (home, "/", val, (char *) 0);
784 /* Convert "\" to "/". */
787 for (s = *pstring; *s; s++)
795 /* Like cmd_file, but strips trailing '/' characters. */
797 cmd_directory (const char *com, const char *val, void *place)
801 /* Call cmd_file() for tilde expansion and separator
802 canonicalization (backslash -> slash under Windows). These
803 things should perhaps be in a separate function. */
804 if (!cmd_file (com, val, place))
809 while (t > s && *--t == '/')
815 /* Split VAL by space to a vector of values, and append those values
816 to vector pointed to by the PLACE argument. If VAL is empty, the
817 PLACE vector is cleared instead. */
820 cmd_vector (const char *com, const char *val, void *place)
822 char ***pvec = (char ***)place;
825 *pvec = merge_vecs (*pvec, sepstring (val));
835 cmd_directory_vector (const char *com, const char *val, void *place)
837 char ***pvec = (char ***)place;
841 /* Strip the trailing slashes from directories. */
844 seps = sepstring (val);
845 for (t = seps; t && *t; t++)
847 int len = strlen (*t);
848 /* Skip degenerate case of root directory. */
851 if ((*t)[len - 1] == '/')
852 (*t)[len - 1] = '\0';
855 *pvec = merge_vecs (*pvec, seps);
865 /* Engine for cmd_bytes and cmd_bytes_large: converts a string such as
866 "100k" or "2.5G" to a floating point number. */
869 parse_bytes_helper (const char *val, double *result)
872 const char *end = val + strlen (val);
874 /* Check for "inf". */
875 if (0 == strcmp (val, "inf"))
881 /* Strip trailing whitespace. */
882 while (val < end && ISSPACE (end[-1]))
887 switch (TOLOWER (end[-1]))
890 --end, mult = 1024.0;
893 --end, mult = 1048576.0;
896 --end, mult = 1073741824.0;
899 --end, mult = 1099511627776.0;
902 /* Not a recognized suffix: assume it's a digit. (If not,
903 simple_atof will raise an error.) */
907 /* Skip leading and trailing whitespace. */
908 while (val < end && ISSPACE (*val))
910 while (val < end && ISSPACE (end[-1]))
915 if (!simple_atof (val, end, &number) || number < 0)
918 *result = number * mult;
922 /* Parse VAL as a number and set its value to PLACE (which should
925 By default, the value is assumed to be in bytes. If "K", "M", or
926 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
927 1<<30, respectively. Floating point values are allowed and are
928 cast to integer before use. The idea is to be able to use things
929 like 1.5k instead of "1536".
931 The string "inf" is returned as 0.
933 In case of error, 0 is returned and memory pointed to by PLACE
934 remains unmodified. */
937 cmd_bytes (const char *com, const char *val, void *place)
940 if (!parse_bytes_helper (val, &byte_value))
942 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
943 exec_name, com, val);
946 *(wgint *)place = (wgint)byte_value;
950 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
951 LARGE_INT. It works by converting the string to double, therefore
952 working with values up to 2^53-1 without loss of precision. This
953 value (8192 TB) is large enough to serve for a while. */
956 cmd_bytes_large (const char *com, const char *val, void *place)
959 if (!parse_bytes_helper (val, &byte_value))
961 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
962 exec_name, com, val);
965 *(LARGE_INT *)place = (LARGE_INT)byte_value;
969 /* Store the value of VAL to *OUT. The value is a time period, by
970 default expressed in seconds, but also accepting suffixes "m", "h",
971 "d", and "w" for minutes, hours, days, and weeks respectively. */
974 cmd_time (const char *com, const char *val, void *place)
977 const char *end = val + strlen (val);
979 /* Strip trailing whitespace. */
980 while (val < end && ISSPACE (end[-1]))
986 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
987 exec_name, com, val);
991 switch (TOLOWER (end[-1]))
994 --end, mult = 1; /* seconds */
997 --end, mult = 60; /* minutes */
1000 --end, mult = 3600; /* hours */
1003 --end, mult = 86400.0; /* days */
1006 --end, mult = 604800.0; /* weeks */
1009 /* Not a recognized suffix: assume it belongs to the number.
1010 (If not, simple_atof will raise an error.) */
1014 /* Skip leading and trailing whitespace. */
1015 while (val < end && ISSPACE (*val))
1017 while (val < end && ISSPACE (end[-1]))
1022 if (!simple_atof (val, end, &number))
1025 *(double *)place = number * mult;
1031 cmd_cert_type (const char *com, const char *val, void *place)
1033 static const struct decode_item choices[] = {
1034 { "pem", keyfile_pem },
1035 { "der", keyfile_asn1 },
1036 { "asn1", keyfile_asn1 },
1038 int ok = decode_string (val, choices, countof (choices), place);
1040 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1045 /* Specialized helper functions, used by `commands' to handle some
1046 options specially. */
1048 static int check_user_specified_header (const char *);
1051 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1053 if (!cmd_boolean (com, val, &opt.dirstruct))
1055 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1056 must be affected inversely. */
1058 opt.no_dirstruct = 0;
1060 opt.no_dirstruct = 1;
1065 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1067 /* Empty value means reset the list of headers. */
1070 free_vec (opt.user_headers);
1071 opt.user_headers = NULL;
1075 if (!check_user_specified_header (val))
1077 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1078 exec_name, com, val);
1081 opt.user_headers = vec_append (opt.user_headers, val);
1086 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1088 int flag = cmd_boolean (com, val, &opt.htmlify);
1089 if (flag && !opt.htmlify)
1090 opt.remove_listing = 0;
1094 /* Set the "mirror" mode. It means: recursive download, timestamping,
1095 no limit on max. recursion depth, and don't remove listings. */
1098 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1102 if (!cmd_boolean (com, val, &mirror))
1107 if (!opt.no_dirstruct)
1109 opt.timestamping = 1;
1110 opt.reclevel = INFINITE_RECURSION;
1111 opt.remove_listing = 0;
1116 /* Validate --prefer-family and set the choice. Allowed values are
1117 "IPv4", "IPv6", and "none". */
1120 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1122 static const struct decode_item choices[] = {
1123 { "IPv4", prefer_ipv4 },
1124 { "IPv6", prefer_ipv6 },
1125 { "none", prefer_none },
1127 int ok = decode_string (val, choices, countof (choices),
1128 (int *) &opt.prefer_family);
1130 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1134 /* Set progress.type to VAL, but verify that it's a valid progress
1135 implementation before that. */
1138 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1140 if (!valid_progress_implementation_p (val))
1142 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1143 exec_name, com, val);
1146 xfree_null (opt.progress_type);
1148 /* Don't call set_progress_implementation here. It will be called
1149 in main() when it becomes clear what the log output is. */
1150 opt.progress_type = xstrdup (val);
1154 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1155 set to true, also set opt.dirstruct to 1, unless opt.no_dirstruct
1159 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1161 if (!cmd_boolean (com, val, &opt.recursive))
1165 if (opt.recursive && !opt.no_dirstruct)
1172 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1174 int restrict_os = opt.restrict_files_os;
1175 int restrict_ctrl = opt.restrict_files_ctrl;
1177 const char *end = strchr (val, ',');
1179 end = val + strlen (val);
1181 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1183 if (VAL_IS ("unix"))
1184 restrict_os = restrict_unix;
1185 else if (VAL_IS ("windows"))
1186 restrict_os = restrict_windows;
1187 else if (VAL_IS ("nocontrol"))
1193 _("%s: %s: Invalid restriction `%s', use `unix' or `windows'.\n"),
1194 exec_name, com, val);
1202 if (!strcmp (end + 1, "nocontrol"))
1208 opt.restrict_files_os = restrict_os;
1209 opt.restrict_files_ctrl = restrict_ctrl;
1215 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1217 static const struct decode_item choices[] = {
1218 { "auto", secure_protocol_auto },
1219 { "sslv2", secure_protocol_sslv2 },
1220 { "sslv3", secure_protocol_sslv3 },
1221 { "tlsv1", secure_protocol_tlsv1 },
1223 int ok = decode_string (val, choices, countof (choices), place);
1225 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1230 /* Set all three timeout values. */
1233 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1236 if (!cmd_time (com, val, &value))
1238 opt.read_timeout = value;
1239 opt.connect_timeout = value;
1240 opt.dns_timeout = value;
1245 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1247 /* Disallow embedded newlines. */
1248 if (strchr (val, '\n'))
1250 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1251 exec_name, com, val);
1254 xfree_null (opt.useragent);
1255 opt.useragent = xstrdup (val);
1259 /* Miscellaneous useful routines. */
1261 /* A very simple atoi clone, more useful than atoi because it works on
1262 delimited strings, and has error reportage. Returns 1 on success,
1263 0 on failure. If successful, stores result to *DEST. */
1266 simple_atoi (const char *beg, const char *end, int *dest)
1270 const char *p = beg;
1272 while (p < end && ISSPACE (*p))
1274 if (p < end && (*p == '-' || *p == '+'))
1276 negative = (*p == '-');
1282 /* Read negative numbers in a separate loop because the most
1283 negative integer cannot be represented as a positive number. */
1286 for (; p < end && ISDIGIT (*p); p++)
1288 int next = (10 * result) + (*p - '0');
1290 return 0; /* overflow */
1294 for (; p < end && ISDIGIT (*p); p++)
1296 int next = (10 * result) - (*p - '0');
1298 return 0; /* underflow */
1309 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1310 doesn't handle exponential notation. Returns 1 on success, 0 on
1311 failure. In case of success, stores its result to *DEST. */
1314 simple_atof (const char *beg, const char *end, double *dest)
1323 const char *p = beg;
1325 while (p < end && ISSPACE (*p))
1327 if (p < end && (*p == '-' || *p == '+'))
1329 negative = (*p == '-');
1333 for (; p < end; p++)
1339 result = (10 * result) + (ch - '0');
1341 result += (ch - '0') / (divider *= 10);
1363 /* Verify that the user-specified header in S is valid. It must
1364 contain a colon preceded by non-white-space characters and must not
1365 contain newlines. */
1368 check_user_specified_header (const char *s)
1372 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++);
1373 /* The header MUST contain `:' preceded by at least one
1374 non-whitespace character. */
1375 if (*p != ':' || p == s)
1377 /* The header MUST NOT contain newlines. */
1378 if (strchr (s, '\n'))
1383 /* Decode VAL into a number, according to ITEMS. */
1386 decode_string (const char *val, const struct decode_item *items, int itemcount,
1390 for (i = 0; i < itemcount; i++)
1391 if (0 == strcasecmp (val, items[i].name))
1393 *place = items[i].code;
1400 void cleanup_html_url (void);
1401 void http_cleanup (void);
1404 /* Free the memory allocated by global variables. */
1408 /* Free external resources, close files, etc. */
1411 extern FILE *output_stream;
1413 fclose (output_stream);
1414 /* No need to check for error because Wget flushes its output (and
1415 checks for errors) after any data arrives. */
1418 /* We're exiting anyway so there's no real need to call free()
1419 hundreds of times. Skipping the frees will make Wget exit
1422 However, when detecting leaks, it's crucial to free() everything
1423 because then you can find the real leaks, i.e. the allocated
1424 memory which grows with the size of the program. */
1430 cleanup_html_url ();
1435 extern acc_t *netrc_list;
1436 free_netrc (netrc_list);
1438 xfree_null (opt.lfilename);
1439 xfree_null (opt.dir_prefix);
1440 xfree_null (opt.input_filename);
1441 xfree_null (opt.output_document);
1442 free_vec (opt.accepts);
1443 free_vec (opt.rejects);
1444 free_vec (opt.excludes);
1445 free_vec (opt.includes);
1446 free_vec (opt.domains);
1447 free_vec (opt.follow_tags);
1448 free_vec (opt.ignore_tags);
1449 xfree_null (opt.progress_type);
1450 xfree_null (opt.ftp_user);
1451 xfree_null (opt.ftp_passwd);
1452 xfree_null (opt.ftp_proxy);
1453 xfree_null (opt.https_proxy);
1454 xfree_null (opt.http_proxy);
1455 free_vec (opt.no_proxy);
1456 xfree_null (opt.useragent);
1457 xfree_null (opt.referer);
1458 xfree_null (opt.http_user);
1459 xfree_null (opt.http_passwd);
1460 free_vec (opt.user_headers);
1462 xfree_null (opt.cert_file);
1463 xfree_null (opt.private_key);
1464 xfree_null (opt.ca_directory);
1465 xfree_null (opt.ca_cert);
1466 xfree_null (opt.random_file);
1467 xfree_null (opt.egd_file);
1469 xfree_null (opt.bind_address);
1470 xfree_null (opt.cookies_input);
1471 xfree_null (opt.cookies_output);
1472 xfree_null (opt.user);
1473 xfree_null (opt.passwd);
1474 #endif /* DEBUG_MALLOC */