1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
5 This file is part of GNU Wget.
7 GNU Wget is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3 of the License, or
10 (at your option) any later version.
12 GNU Wget is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with Wget. If not, see <http://www.gnu.org/licenses/>.
20 Additional permission under GNU GPL version 3 section 7
22 If you modify this program, or any covered work, by linking or
23 combining it with the OpenSSL project's OpenSSL library (or a
24 modified version of that library), containing parts covered by the
25 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
26 grants you additional permission to convey the resulting work.
27 Corresponding Source for a non-source form of such a combination
28 shall include the source code for the parts of OpenSSL used as well
29 as that of the covered work. */
41 /* not all systems provide PATH_MAX in limits.h */
43 # include <sys/param.h>
45 # define PATH_MAX MAXPATHLEN
60 #include "recur.h" /* for INFINITE_RECURSION */
61 #include "convert.h" /* for convert_cleanup */
62 #include "res.h" /* for res_cleanup */
63 #include "http.h" /* for http_cleanup */
64 #include "retr.h" /* for output_stream */
72 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
74 CMD_DECLARE (cmd_boolean);
75 CMD_DECLARE (cmd_bytes);
76 CMD_DECLARE (cmd_bytes_sum);
78 CMD_DECLARE (cmd_cert_type);
80 CMD_DECLARE (cmd_directory_vector);
81 CMD_DECLARE (cmd_number);
82 CMD_DECLARE (cmd_number_inf);
83 CMD_DECLARE (cmd_string);
84 CMD_DECLARE (cmd_file);
85 CMD_DECLARE (cmd_directory);
86 CMD_DECLARE (cmd_time);
87 CMD_DECLARE (cmd_vector);
89 CMD_DECLARE (cmd_spec_dirstruct);
90 CMD_DECLARE (cmd_spec_header);
91 CMD_DECLARE (cmd_spec_htmlify);
92 CMD_DECLARE (cmd_spec_mirror);
93 CMD_DECLARE (cmd_spec_prefer_family);
94 CMD_DECLARE (cmd_spec_progress);
95 CMD_DECLARE (cmd_spec_recursive);
96 CMD_DECLARE (cmd_spec_restrict_file_names);
98 CMD_DECLARE (cmd_spec_secure_protocol);
100 CMD_DECLARE (cmd_spec_timeout);
101 CMD_DECLARE (cmd_spec_useragent);
102 CMD_DECLARE (cmd_spec_verbose);
104 /* List of recognized commands, each consisting of name, place and
105 function. When adding a new command, simply add it to the list,
106 but be sure to keep the list sorted alphabetically, as
107 command_by_name's binary search depends on it. Also, be sure to
108 add any entries that allocate memory (e.g. cmd_string and
109 cmd_vector) to the cleanup() function below. */
111 static const struct {
114 bool (*action) (const char *, const char *, void *);
116 /* KEEP THIS LIST ALPHABETICALLY SORTED */
117 { "accept", &opt.accepts, cmd_vector },
118 { "addhostdir", &opt.add_hostdir, cmd_boolean },
119 { "adjustextension", &opt.adjust_extension, cmd_boolean },
120 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
121 { "askpassword", &opt.ask_passwd, cmd_boolean },
122 { "authnochallenge", &opt.auth_without_challenge,
124 { "background", &opt.background, cmd_boolean },
125 { "backupconverted", &opt.backup_converted, cmd_boolean },
126 { "backups", &opt.backups, cmd_number },
127 { "base", &opt.base_href, cmd_string },
128 { "bindaddress", &opt.bind_address, cmd_string },
130 { "cacertificate", &opt.ca_cert, cmd_file },
132 { "cache", &opt.allow_cache, cmd_boolean },
134 { "cadirectory", &opt.ca_directory, cmd_directory },
135 { "certificate", &opt.cert_file, cmd_file },
136 { "certificatetype", &opt.cert_type, cmd_cert_type },
137 { "checkcertificate", &opt.check_cert, cmd_boolean },
139 { "connecttimeout", &opt.connect_timeout, cmd_time },
140 { "contentdisposition", &opt.content_disposition, cmd_boolean },
141 { "continue", &opt.always_rest, cmd_boolean },
142 { "convertlinks", &opt.convert_links, cmd_boolean },
143 { "cookies", &opt.cookies, cmd_boolean },
144 { "cutdirs", &opt.cut_dirs, cmd_number },
146 { "debug", &opt.debug, cmd_boolean },
148 { "defaultpage", &opt.default_page, cmd_string},
149 { "deleteafter", &opt.delete_after, cmd_boolean },
150 { "dirprefix", &opt.dir_prefix, cmd_directory },
151 { "dirstruct", NULL, cmd_spec_dirstruct },
152 { "dnscache", &opt.dns_cache, cmd_boolean },
153 { "dnstimeout", &opt.dns_timeout, cmd_time },
154 { "domains", &opt.domains, cmd_vector },
155 { "dotbytes", &opt.dot_bytes, cmd_bytes },
156 { "dotsinline", &opt.dots_in_line, cmd_number },
157 { "dotspacing", &opt.dot_spacing, cmd_number },
158 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
160 { "egdfile", &opt.egd_file, cmd_file },
162 { "excludedirectories", &opt.excludes, cmd_directory_vector },
163 { "excludedomains", &opt.exclude_domains, cmd_vector },
164 { "followftp", &opt.follow_ftp, cmd_boolean },
165 { "followtags", &opt.follow_tags, cmd_vector },
166 { "forcehtml", &opt.force_html, cmd_boolean },
167 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
168 { "ftppassword", &opt.ftp_passwd, cmd_string },
169 { "ftpproxy", &opt.ftp_proxy, cmd_string },
171 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
172 #endif /* def __VMS */
173 { "ftpuser", &opt.ftp_user, cmd_string },
174 { "glob", &opt.ftp_glob, cmd_boolean },
175 { "header", NULL, cmd_spec_header },
176 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
177 { "htmlify", NULL, cmd_spec_htmlify },
178 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
179 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
180 { "httppassword", &opt.http_passwd, cmd_string },
181 { "httpproxy", &opt.http_proxy, cmd_string },
182 { "httpsproxy", &opt.https_proxy, cmd_string },
183 { "httpuser", &opt.http_user, cmd_string },
184 { "ignorecase", &opt.ignore_case, cmd_boolean },
185 { "ignorelength", &opt.ignore_length, cmd_boolean },
186 { "ignoretags", &opt.ignore_tags, cmd_vector },
187 { "includedirectories", &opt.includes, cmd_directory_vector },
189 { "inet4only", &opt.ipv4_only, cmd_boolean },
190 { "inet6only", &opt.ipv6_only, cmd_boolean },
192 { "input", &opt.input_filename, cmd_file },
193 { "iri", &opt.enable_iri, cmd_boolean },
194 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
195 { "limitrate", &opt.limit_rate, cmd_bytes },
196 { "loadcookies", &opt.cookies_input, cmd_file },
197 { "localencoding", &opt.locale, cmd_string },
198 { "logfile", &opt.lfilename, cmd_file },
199 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
200 { "maxredirect", &opt.max_redirect, cmd_number },
201 { "mirror", NULL, cmd_spec_mirror },
202 { "netrc", &opt.netrc, cmd_boolean },
203 { "noclobber", &opt.noclobber, cmd_boolean },
204 { "noparent", &opt.no_parent, cmd_boolean },
205 { "noproxy", &opt.no_proxy, cmd_vector },
206 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
207 { "outputdocument", &opt.output_document, cmd_file },
208 { "pagerequisites", &opt.page_requisites, cmd_boolean },
209 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
210 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
211 { "password", &opt.passwd, cmd_string },
212 { "postdata", &opt.post_data, cmd_string },
213 { "postfile", &opt.post_file_name, cmd_file },
214 { "preferfamily", NULL, cmd_spec_prefer_family },
215 { "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
217 { "privatekey", &opt.private_key, cmd_file },
218 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
220 { "progress", &opt.progress_type, cmd_spec_progress },
221 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
222 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
223 { "proxypassword", &opt.proxy_passwd, cmd_string },
224 { "proxyuser", &opt.proxy_user, cmd_string },
225 { "quiet", &opt.quiet, cmd_boolean },
226 { "quota", &opt.quota, cmd_bytes_sum },
228 { "randomfile", &opt.random_file, cmd_file },
230 { "randomwait", &opt.random_wait, cmd_boolean },
231 { "readtimeout", &opt.read_timeout, cmd_time },
232 { "reclevel", &opt.reclevel, cmd_number_inf },
233 { "recursive", NULL, cmd_spec_recursive },
234 { "referer", &opt.referer, cmd_string },
235 { "reject", &opt.rejects, cmd_vector },
236 { "relativeonly", &opt.relative_only, cmd_boolean },
237 { "remoteencoding", &opt.encoding_remote, cmd_string },
238 { "removelisting", &opt.remove_listing, cmd_boolean },
239 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
240 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
241 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
242 { "robots", &opt.use_robots, cmd_boolean },
243 { "savecookies", &opt.cookies_output, cmd_file },
244 { "saveheaders", &opt.save_headers, cmd_boolean },
246 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
248 { "serverresponse", &opt.server_response, cmd_boolean },
249 { "spanhosts", &opt.spanhost, cmd_boolean },
250 { "spider", &opt.spider, cmd_boolean },
251 { "strictcomments", &opt.strict_comments, cmd_boolean },
252 { "timeout", NULL, cmd_spec_timeout },
253 { "timestamping", &opt.timestamping, cmd_boolean },
254 { "tries", &opt.ntry, cmd_number_inf },
255 { "useproxy", &opt.use_proxy, cmd_boolean },
256 { "user", &opt.user, cmd_string },
257 { "useragent", NULL, cmd_spec_useragent },
258 { "verbose", NULL, cmd_spec_verbose },
259 { "wait", &opt.wait, cmd_time },
260 { "waitretry", &opt.waitretry, cmd_time },
262 { "wdebug", &opt.wdebug, cmd_boolean },
266 /* Look up CMDNAME in the commands[] and return its position in the
267 array. If CMDNAME is not found, return -1. */
270 command_by_name (const char *cmdname)
272 /* Use binary search for speed. Wget has ~100 commands, which
273 guarantees a worst case performance of 7 string comparisons. */
274 int lo = 0, hi = countof (commands) - 1;
278 int mid = (lo + hi) >> 1;
279 int cmp = strcasecmp (cmdname, commands[mid].name);
290 /* Reset the variables to default values. */
296 /* Most of the default values are 0 (and 0.0, NULL, and false).
297 Just reset everything, and fill in the non-zero values. Note
298 that initializing pointers to NULL this way is technically
299 illegal, but porting Wget to a machine where NULL is not all-zero
300 bit pattern will be the least of the implementors' worries. */
307 opt.add_hostdir = true;
311 opt.http_keep_alive = true;
312 opt.use_proxy = true;
313 tmp = getenv ("no_proxy");
315 opt.no_proxy = sepstring (tmp);
316 opt.prefer_family = prefer_none;
317 opt.allow_cache = true;
319 opt.read_timeout = 900;
320 opt.use_robots = true;
322 opt.remove_listing = true;
324 opt.dot_bytes = 1024;
325 opt.dot_spacing = 10;
326 opt.dots_in_line = 50;
328 opt.dns_cache = true;
332 opt.check_cert = true;
335 /* The default for file name restriction defaults to the OS type. */
336 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
337 opt.restrict_files_os = restrict_windows;
339 opt.restrict_files_os = restrict_unix;
341 opt.restrict_files_ctrl = true;
342 opt.restrict_files_nonascii = false;
343 opt.restrict_files_case = restrict_no_case_restriction;
345 opt.max_redirect = 20;
350 opt.enable_iri = true;
352 opt.enable_iri = false;
355 opt.encoding_remote = NULL;
358 /* Return the user's home directory (strdup-ed), or NULL if none is
363 static char buf[PATH_MAX];
368 home = getenv ("HOME");
372 /* Under MSDOS, if $HOME isn't defined, use the directory where
373 `wget.exe' resides. */
374 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
377 strcpy (buf, _w32_get_argv0 ());
378 p = strrchr (buf, '/'); /* djgpp */
380 p = strrchr (buf, '\\'); /* others */
384 #elif !defined(WINDOWS)
385 /* If HOME is not defined, try getting it from the password
387 struct passwd *pwd = getpwuid (getuid ());
388 if (!pwd || !pwd->pw_dir)
390 strcpy (buf, pwd->pw_dir);
393 /* Under Windows, if $HOME isn't defined, use the directory where
394 `wget.exe' resides. */
400 return home ? xstrdup (home) : NULL;
403 /* Check the 'WGETRC' environment variable and return the file name
404 if 'WGETRC' is set and is a valid file.
405 If the `WGETRC' variable exists but the file does not exist, the
406 function will exit(). */
408 wgetrc_env_file_name (void)
410 char *env = getenv ("WGETRC");
413 if (!file_exists_p (env))
415 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
419 return xstrdup (env);
424 /* Check for the existance of '$HOME/.wgetrc' and return it's path
425 if it exists and is set. */
427 wgetrc_user_file_name (void)
429 char *home = home_dir ();
431 /* If that failed, try $HOME/.wgetrc (or equivalent). */
434 file = "SYS$LOGIN:.wgetrc";
435 #else /* def __VMS */
438 file = aprintf ("%s/.wgetrc", home);
440 #endif /* def __VMS [else] */
444 if (!file_exists_p (file))
452 /* Return the path to the user's .wgetrc. This is either the value of
453 `WGETRC' environment variable, or `$HOME/.wgetrc'.
455 Additionally, for windows, look in the directory where wget.exe
458 wgetrc_file_name (void)
460 char *file = wgetrc_env_file_name ();
464 file = wgetrc_user_file_name ();
467 /* Under Windows, if we still haven't found .wgetrc, look for the file
468 `wget.ini' in the directory where `wget.exe' resides; we do this for
469 backward compatibility with previous versions of Wget.
470 SYSTEM_WGETRC should not be defined under WINDOWS. */
473 char *home = home_dir ();
479 file = aprintf ("%s/wget.ini", home);
480 if (!file_exists_p (file))
493 /* Return values of parse_line. */
501 static enum parse_line parse_line (const char *, char **, char **, int *);
502 static bool setval_internal (int, const char *, const char *);
503 static bool setval_internal_tilde (int, const char *, const char *);
505 /* Initialize variables from a wgetrc file. Returns zero (failure) if
506 there were errors in the file. */
509 run_wgetrc (const char *file)
516 fp = fopen (file, "r");
519 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
520 file, strerror (errno));
521 return true; /* not a fatal error */
524 while ((line = read_whole_line (fp)) != NULL)
526 char *com = NULL, *val = NULL;
529 /* Parse the line. */
530 switch (parse_line (line, &com, &val, &comind))
533 /* If everything is OK, set the value. */
534 if (!setval_internal_tilde (comind, com, val))
536 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
537 exec_name, file, ln);
541 case line_syntax_error:
542 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
543 exec_name, file, ln);
546 case line_unknown_command:
547 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
548 exec_name, quote (com), file, ln);
566 /* Initialize the defaults and run the system wgetrc and user's own
571 char *file, *env_sysrc;
574 /* Load the hard-coded defaults. */
577 /* Run a non-standard system rc file when the according environment
578 variable has been set. For internal testing purposes only! */
579 env_sysrc = getenv ("SYSTEM_WGETRC");
580 if (env_sysrc && file_exists_p (env_sysrc))
581 ok &= run_wgetrc (env_sysrc);
582 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
584 else if (file_exists_p (SYSTEM_WGETRC))
585 ok &= run_wgetrc (SYSTEM_WGETRC);
587 /* Override it with your own, if one exists. */
588 file = wgetrc_file_name ();
591 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
592 something like realpath() before comparing them with `strcmp' */
594 if (!strcmp (file, SYSTEM_WGETRC))
596 fprintf (stderr, _("\
597 %s: Warning: Both system and user wgetrc point to %s.\n"),
598 exec_name, quote (file));
602 ok &= run_wgetrc (file);
604 /* If there were errors processing either `.wgetrc', abort. */
612 /* Remove dashes and underscores from S, modifying S in the
618 char *t = s; /* t - tortoise */
619 char *h = s; /* h - hare */
621 if (*h == '_' || *h == '-')
628 /* Parse the line pointed by line, with the syntax:
629 <sp>* command <sp>* = <sp>* value <sp>*
630 Uses malloc to allocate space for command and value.
632 Returns one of line_ok, line_empty, line_syntax_error, or
633 line_unknown_command.
635 In case of line_ok, *COM and *VAL point to freshly allocated
636 strings, and *COMIND points to com's index. In case of error or
637 empty line, their values are unmodified. */
639 static enum parse_line
640 parse_line (const char *line, char **com, char **val, int *comind)
643 const char *end = line + strlen (line);
644 const char *cmdstart, *cmdend;
645 const char *valstart, *valend;
650 /* Skip leading and trailing whitespace. */
651 while (*line && c_isspace (*line))
653 while (end > line && c_isspace (end[-1]))
656 /* Skip empty lines and comments. */
657 if (!*line || *line == '#')
663 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
667 /* Skip '=', as well as any space before or after it. */
668 while (p < end && c_isspace (*p))
670 if (p == end || *p != '=')
671 return line_syntax_error;
673 while (p < end && c_isspace (*p))
679 /* The syntax is valid (even though the command might not be). Fill
680 in the command name and value. */
681 *com = strdupdelim (cmdstart, cmdend);
682 *val = strdupdelim (valstart, valend);
684 /* The line now known to be syntactically correct. Check whether
685 the command is valid. */
686 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
688 ind = command_by_name (cmdcopy);
690 return line_unknown_command;
692 /* Report success to the caller. */
697 #if defined(WINDOWS) || defined(MSDOS)
698 # define ISSEP(c) ((c) == '/' || (c) == '\\')
700 # define ISSEP(c) ((c) == '/')
703 /* Run commands[comind].action. */
706 setval_internal (int comind, const char *com, const char *val)
708 assert (0 <= comind && ((size_t) comind) < countof (commands));
709 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
710 return commands[comind].action (com, val, commands[comind].place);
714 setval_internal_tilde (int comind, const char *com, const char *val)
720 ret = setval_internal (comind, com, val);
722 /* We make tilde expansion for cmd_file and cmd_directory */
723 if (((commands[comind].action == cmd_file) ||
724 (commands[comind].action == cmd_directory))
725 && ret && (*val == '~' && ISSEP (val[1])))
727 pstring = commands[comind].place;
731 homelen = strlen (home);
732 while (homelen && ISSEP (home[homelen - 1]))
733 home[--homelen] = '\0';
735 /* Skip the leading "~/". */
736 for (++val; ISSEP (*val); val++)
738 *pstring = concat_strings (home, "/", val, (char *)0);
744 /* Run command COM with value VAL. If running the command produces an
745 error, report the error and exit.
747 This is intended to be called from main() to modify Wget's behavior
748 through command-line switches. Since COM is hard-coded in main(),
749 it is not canonicalized, and this aborts when COM is not found.
751 If COMIND's are exported to init.h, this function will be changed
752 to accept COMIND directly. */
755 setoptval (const char *com, const char *val, const char *optname)
757 /* Prepend "--" to OPTNAME. */
758 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
761 strcpy (dd_optname + 2, optname);
763 assert (val != NULL);
764 if (!setval_internal (command_by_name (com), dd_optname, val))
768 /* Parse OPT into command and value and run it. For example,
769 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
770 This is used by the `--execute' flag in main.c. */
773 run_command (const char *opt)
777 switch (parse_line (opt, &com, &val, &comind))
780 if (!setval_internal (comind, com, val))
786 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
787 exec_name, quote (opt));
792 /* Generic helper functions, for use with `commands'. */
794 /* Forward declarations: */
799 static bool decode_string (const char *, const struct decode_item *, int, int *);
800 static bool simple_atoi (const char *, const char *, int *);
801 static bool simple_atof (const char *, const char *, double *);
803 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
805 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
806 && c_tolower((p)[1]) == (c1) \
809 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
810 && c_tolower((p)[1]) == (c1) \
811 && c_tolower((p)[2]) == (c2) \
815 /* Store the boolean value from VAL to PLACE. COM is ignored,
816 except for error messages. */
818 cmd_boolean (const char *com, const char *val, void *place)
822 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
823 /* "on", "yes" and "1" mean true. */
825 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
826 /* "off", "no" and "0" mean false. */
831 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
832 exec_name, com, quote (val));
836 *(bool *) place = value;
840 /* Set the non-negative integer value from VAL to PLACE. With
841 incorrect specification, the number remains unchanged. */
843 cmd_number (const char *com, const char *val, void *place)
845 if (!simple_atoi (val, val + strlen (val), place)
846 || *(int *) place < 0)
848 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
849 exec_name, com, quote (val));
855 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
857 cmd_number_inf (const char *com, const char *val, void *place)
859 if (!strcasecmp (val, "inf"))
864 return cmd_number (com, val, place);
867 /* Copy (strdup) the string at COM to a new location and place a
868 pointer to *PLACE. */
870 cmd_string (const char *com, const char *val, void *place)
872 char **pstring = (char **)place;
874 xfree_null (*pstring);
875 *pstring = xstrdup (val);
880 /* Like the above, but handles tilde-expansion when reading a user's
881 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
882 gets expanded to the user's home directory. */
884 cmd_file (const char *com, const char *val, void *place)
886 char **pstring = (char **)place;
888 xfree_null (*pstring);
890 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
892 *pstring = xstrdup (val);
894 #if defined(WINDOWS) || defined(MSDOS)
895 /* Convert "\" to "/". */
898 for (s = *pstring; *s; s++)
906 /* Like cmd_file, but strips trailing '/' characters. */
908 cmd_directory (const char *com, const char *val, void *place)
912 /* Call cmd_file() for tilde expansion and separator
913 canonicalization (backslash -> slash under Windows). These
914 things should perhaps be in a separate function. */
915 if (!cmd_file (com, val, place))
920 while (t > s && *--t == '/')
926 /* Split VAL by space to a vector of values, and append those values
927 to vector pointed to by the PLACE argument. If VAL is empty, the
928 PLACE vector is cleared instead. */
931 cmd_vector (const char *com, const char *val, void *place)
933 char ***pvec = (char ***)place;
936 *pvec = merge_vecs (*pvec, sepstring (val));
946 cmd_directory_vector (const char *com, const char *val, void *place)
948 char ***pvec = (char ***)place;
952 /* Strip the trailing slashes from directories. */
955 seps = sepstring (val);
956 for (t = seps; t && *t; t++)
958 int len = strlen (*t);
959 /* Skip degenerate case of root directory. */
962 if ((*t)[len - 1] == '/')
963 (*t)[len - 1] = '\0';
966 *pvec = merge_vecs (*pvec, seps);
976 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
977 "100k" or "2.5G" to a floating point number. */
980 parse_bytes_helper (const char *val, double *result)
983 const char *end = val + strlen (val);
985 /* Check for "inf". */
986 if (0 == strcmp (val, "inf"))
992 /* Strip trailing whitespace. */
993 while (val < end && c_isspace (end[-1]))
998 switch (c_tolower (end[-1]))
1001 --end, mult = 1024.0;
1004 --end, mult = 1048576.0;
1007 --end, mult = 1073741824.0;
1010 --end, mult = 1099511627776.0;
1013 /* Not a recognized suffix: assume it's a digit. (If not,
1014 simple_atof will raise an error.) */
1018 /* Skip leading and trailing whitespace. */
1019 while (val < end && c_isspace (*val))
1021 while (val < end && c_isspace (end[-1]))
1026 if (!simple_atof (val, end, &number) || number < 0)
1029 *result = number * mult;
1033 /* Parse VAL as a number and set its value to PLACE (which should
1036 By default, the value is assumed to be in bytes. If "K", "M", or
1037 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1038 1<<30, respectively. Floating point values are allowed and are
1039 cast to integer before use. The idea is to be able to use things
1040 like 1.5k instead of "1536".
1042 The string "inf" is returned as 0.
1044 In case of error, false is returned and memory pointed to by PLACE
1045 remains unmodified. */
1048 cmd_bytes (const char *com, const char *val, void *place)
1051 if (!parse_bytes_helper (val, &byte_value))
1053 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1054 exec_name, com, quote (val));
1057 *(wgint *)place = (wgint)byte_value;
1061 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1062 SIZE_SUM. It works by converting the string to double, therefore
1063 working with values up to 2^53-1 without loss of precision. This
1064 value (8192 TB) is large enough to serve for a while. */
1067 cmd_bytes_sum (const char *com, const char *val, void *place)
1070 if (!parse_bytes_helper (val, &byte_value))
1072 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1073 exec_name, com, quote (val));
1076 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1080 /* Store the value of VAL to *OUT. The value is a time period, by
1081 default expressed in seconds, but also accepting suffixes "m", "h",
1082 "d", and "w" for minutes, hours, days, and weeks respectively. */
1085 cmd_time (const char *com, const char *val, void *place)
1087 double number, mult;
1088 const char *end = val + strlen (val);
1090 /* Strip trailing whitespace. */
1091 while (val < end && c_isspace (end[-1]))
1097 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1098 exec_name, com, quote (val));
1102 switch (c_tolower (end[-1]))
1105 --end, mult = 1; /* seconds */
1108 --end, mult = 60; /* minutes */
1111 --end, mult = 3600; /* hours */
1114 --end, mult = 86400.0; /* days */
1117 --end, mult = 604800.0; /* weeks */
1120 /* Not a recognized suffix: assume it belongs to the number.
1121 (If not, simple_atof will raise an error.) */
1125 /* Skip leading and trailing whitespace. */
1126 while (val < end && c_isspace (*val))
1128 while (val < end && c_isspace (end[-1]))
1133 if (!simple_atof (val, end, &number))
1136 *(double *)place = number * mult;
1142 cmd_cert_type (const char *com, const char *val, void *place)
1144 static const struct decode_item choices[] = {
1145 { "pem", keyfile_pem },
1146 { "der", keyfile_asn1 },
1147 { "asn1", keyfile_asn1 },
1149 int ok = decode_string (val, choices, countof (choices), place);
1151 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1156 /* Specialized helper functions, used by `commands' to handle some
1157 options specially. */
1159 static bool check_user_specified_header (const char *);
1162 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1164 if (!cmd_boolean (com, val, &opt.dirstruct))
1166 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1167 must be affected inversely. */
1169 opt.no_dirstruct = false;
1171 opt.no_dirstruct = true;
1176 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1178 /* Empty value means reset the list of headers. */
1181 free_vec (opt.user_headers);
1182 opt.user_headers = NULL;
1186 if (!check_user_specified_header (val))
1188 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1189 exec_name, com, quote (val));
1192 opt.user_headers = vec_append (opt.user_headers, val);
1197 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1199 int flag = cmd_boolean (com, val, &opt.htmlify);
1200 if (flag && !opt.htmlify)
1201 opt.remove_listing = false;
1205 /* Set the "mirror" mode. It means: recursive download, timestamping,
1206 no limit on max. recursion depth, and don't remove listings. */
1209 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1213 if (!cmd_boolean (com, val, &mirror))
1217 opt.recursive = true;
1218 if (!opt.no_dirstruct)
1219 opt.dirstruct = true;
1220 opt.timestamping = true;
1221 opt.reclevel = INFINITE_RECURSION;
1222 opt.remove_listing = false;
1227 /* Validate --prefer-family and set the choice. Allowed values are
1228 "IPv4", "IPv6", and "none". */
1231 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1233 static const struct decode_item choices[] = {
1234 { "IPv4", prefer_ipv4 },
1235 { "IPv6", prefer_ipv6 },
1236 { "none", prefer_none },
1238 int prefer_family = prefer_none;
1239 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1241 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1242 opt.prefer_family = prefer_family;
1246 /* Set progress.type to VAL, but verify that it's a valid progress
1247 implementation before that. */
1250 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1252 if (!valid_progress_implementation_p (val))
1254 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1255 exec_name, com, quote (val));
1258 xfree_null (opt.progress_type);
1260 /* Don't call set_progress_implementation here. It will be called
1261 in main() when it becomes clear what the log output is. */
1262 opt.progress_type = xstrdup (val);
1266 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1267 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1271 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1273 if (!cmd_boolean (com, val, &opt.recursive))
1277 if (opt.recursive && !opt.no_dirstruct)
1278 opt.dirstruct = true;
1284 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1286 int restrict_os = opt.restrict_files_os;
1287 int restrict_ctrl = opt.restrict_files_ctrl;
1288 int restrict_case = opt.restrict_files_case;
1289 int restrict_nonascii = opt.restrict_files_nonascii;
1293 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1297 end = strchr (val, ',');
1299 end = val + strlen (val);
1301 if (VAL_IS ("unix"))
1302 restrict_os = restrict_unix;
1303 else if (VAL_IS ("windows"))
1304 restrict_os = restrict_windows;
1305 else if (VAL_IS ("lowercase"))
1306 restrict_case = restrict_lowercase;
1307 else if (VAL_IS ("uppercase"))
1308 restrict_case = restrict_uppercase;
1309 else if (VAL_IS ("nocontrol"))
1310 restrict_ctrl = false;
1311 else if (VAL_IS ("ascii"))
1312 restrict_nonascii = true;
1315 fprintf (stderr, _("\
1316 %s: %s: Invalid restriction %s,\n\
1317 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1318 exec_name, com, quote (val));
1325 while (*val && *end);
1329 opt.restrict_files_os = restrict_os;
1330 opt.restrict_files_ctrl = restrict_ctrl;
1331 opt.restrict_files_case = restrict_case;
1332 opt.restrict_files_nonascii = restrict_nonascii;
1339 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1341 static const struct decode_item choices[] = {
1342 { "auto", secure_protocol_auto },
1343 { "sslv2", secure_protocol_sslv2 },
1344 { "sslv3", secure_protocol_sslv3 },
1345 { "tlsv1", secure_protocol_tlsv1 },
1347 int ok = decode_string (val, choices, countof (choices), place);
1349 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1354 /* Set all three timeout values. */
1357 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1360 if (!cmd_time (com, val, &value))
1362 opt.read_timeout = value;
1363 opt.connect_timeout = value;
1364 opt.dns_timeout = value;
1369 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1371 /* Disallow embedded newlines. */
1372 if (strchr (val, '\n'))
1374 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1375 exec_name, com, quote (val));
1378 xfree_null (opt.useragent);
1379 opt.useragent = xstrdup (val);
1383 /* The "verbose" option cannot be cmd_boolean because the variable is
1384 not bool -- it's of type int (-1 means uninitialized because of
1385 some random hackery for disallowing -q -v). */
1388 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1391 if (cmd_boolean (com, val, &flag))
1399 /* Miscellaneous useful routines. */
1401 /* A very simple atoi clone, more useful than atoi because it works on
1402 delimited strings, and has error reportage. Returns true on success,
1403 false on failure. If successful, stores result to *DEST. */
1406 simple_atoi (const char *beg, const char *end, int *dest)
1409 bool negative = false;
1410 const char *p = beg;
1412 while (p < end && c_isspace (*p))
1414 if (p < end && (*p == '-' || *p == '+'))
1416 negative = (*p == '-');
1422 /* Read negative numbers in a separate loop because the most
1423 negative integer cannot be represented as a positive number. */
1426 for (; p < end && c_isdigit (*p); p++)
1428 int next = (10 * result) + (*p - '0');
1430 return false; /* overflow */
1434 for (; p < end && c_isdigit (*p); p++)
1436 int next = (10 * result) - (*p - '0');
1438 return false; /* underflow */
1449 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1450 doesn't handle exponential notation. Returns true on success,
1451 false on failure. In case of success, stores its result to
1455 simple_atof (const char *beg, const char *end, double *dest)
1459 bool negative = false;
1460 bool seen_dot = false;
1461 bool seen_digit = false;
1464 const char *p = beg;
1466 while (p < end && c_isspace (*p))
1468 if (p < end && (*p == '-' || *p == '+'))
1470 negative = (*p == '-');
1474 for (; p < end; p++)
1480 result = (10 * result) + (ch - '0');
1482 result += (ch - '0') / (divider *= 10);
1504 /* Verify that the user-specified header in S is valid. It must
1505 contain a colon preceded by non-white-space characters and must not
1506 contain newlines. */
1509 check_user_specified_header (const char *s)
1513 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1515 /* The header MUST contain `:' preceded by at least one
1516 non-whitespace character. */
1517 if (*p != ':' || p == s)
1519 /* The header MUST NOT contain newlines. */
1520 if (strchr (s, '\n'))
1525 /* Decode VAL into a number, according to ITEMS. */
1528 decode_string (const char *val, const struct decode_item *items, int itemcount,
1532 for (i = 0; i < itemcount; i++)
1533 if (0 == strcasecmp (val, items[i].name))
1535 *place = items[i].code;
1542 void cleanup_html_url (void);
1545 /* Free the memory allocated by global variables. */
1549 /* Free external resources, close files, etc. */
1552 fclose (output_stream);
1553 /* No need to check for error because Wget flushes its output (and
1554 checks for errors) after any data arrives. */
1556 /* We're exiting anyway so there's no real need to call free()
1557 hundreds of times. Skipping the frees will make Wget exit
1560 However, when detecting leaks, it's crucial to free() everything
1561 because then you can find the real leaks, i.e. the allocated
1562 memory which grows with the size of the program. */
1568 cleanup_html_url ();
1573 extern acc_t *netrc_list;
1574 free_netrc (netrc_list);
1576 xfree_null (opt.lfilename);
1577 xfree_null (opt.dir_prefix);
1578 xfree_null (opt.input_filename);
1579 xfree_null (opt.output_document);
1580 free_vec (opt.accepts);
1581 free_vec (opt.rejects);
1582 free_vec (opt.excludes);
1583 free_vec (opt.includes);
1584 free_vec (opt.domains);
1585 free_vec (opt.follow_tags);
1586 free_vec (opt.ignore_tags);
1587 xfree_null (opt.progress_type);
1588 xfree_null (opt.ftp_user);
1589 xfree_null (opt.ftp_passwd);
1590 xfree_null (opt.ftp_proxy);
1591 xfree_null (opt.https_proxy);
1592 xfree_null (opt.http_proxy);
1593 free_vec (opt.no_proxy);
1594 xfree_null (opt.useragent);
1595 xfree_null (opt.referer);
1596 xfree_null (opt.http_user);
1597 xfree_null (opt.http_passwd);
1598 free_vec (opt.user_headers);
1600 xfree_null (opt.cert_file);
1601 xfree_null (opt.private_key);
1602 xfree_null (opt.ca_directory);
1603 xfree_null (opt.ca_cert);
1604 xfree_null (opt.random_file);
1605 xfree_null (opt.egd_file);
1607 xfree_null (opt.bind_address);
1608 xfree_null (opt.cookies_input);
1609 xfree_null (opt.cookies_output);
1610 xfree_null (opt.user);
1611 xfree_null (opt.passwd);
1612 xfree_null (opt.base_href);
1614 #endif /* DEBUG_MALLOC */
1617 /* Unit testing routines. */
1622 test_commands_sorted()
1624 int prev_idx = 0, next_idx = 1;
1625 int command_count = countof (commands) - 1;
1627 while (next_idx <= command_count)
1629 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1632 mu_assert ("FAILED", false);
1645 test_cmd_spec_restrict_file_names()
1650 int expected_restrict_files_os;
1651 int expected_restrict_files_ctrl;
1652 int expected_restrict_files_case;
1655 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1656 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1657 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1658 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1661 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1666 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1669 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1670 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1671 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1672 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1674 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1675 res == test_array[i].result
1676 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1677 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1678 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1684 #endif /* TESTING */