1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
3 2005, 2006, 2007, 2008, 2009, 2010, 2011 Free Software Foundation,
6 This file is part of GNU Wget.
8 GNU Wget is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 3 of the License, or
11 (at your option) any later version.
13 GNU Wget is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
18 You should have received a copy of the GNU General Public License
19 along with Wget. If not, see <http://www.gnu.org/licenses/>.
21 Additional permission under GNU GPL version 3 section 7
23 If you modify this program, or any covered work, by linking or
24 combining it with the OpenSSL project's OpenSSL library (or a
25 modified version of that library), containing parts covered by the
26 terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
27 grants you additional permission to convey the resulting work.
28 Corresponding Source for a non-source form of such a combination
29 shall include the source code for the parts of OpenSSL used as well
30 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_warc_header);
92 CMD_DECLARE (cmd_spec_htmlify);
93 CMD_DECLARE (cmd_spec_mirror);
94 CMD_DECLARE (cmd_spec_prefer_family);
95 CMD_DECLARE (cmd_spec_progress);
96 CMD_DECLARE (cmd_spec_recursive);
97 CMD_DECLARE (cmd_spec_restrict_file_names);
99 CMD_DECLARE (cmd_spec_secure_protocol);
101 CMD_DECLARE (cmd_spec_timeout);
102 CMD_DECLARE (cmd_spec_useragent);
103 CMD_DECLARE (cmd_spec_verbose);
105 /* List of recognized commands, each consisting of name, place and
106 function. When adding a new command, simply add it to the list,
107 but be sure to keep the list sorted alphabetically, as
108 command_by_name's binary search depends on it. Also, be sure to
109 add any entries that allocate memory (e.g. cmd_string and
110 cmd_vector) to the cleanup() function below. */
112 static const struct {
115 bool (*action) (const char *, const char *, void *);
117 /* KEEP THIS LIST ALPHABETICALLY SORTED */
118 { "accept", &opt.accepts, cmd_vector },
119 { "addhostdir", &opt.add_hostdir, cmd_boolean },
120 { "adjustextension", &opt.adjust_extension, cmd_boolean },
121 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
122 { "askpassword", &opt.ask_passwd, cmd_boolean },
123 { "authnochallenge", &opt.auth_without_challenge,
125 { "background", &opt.background, cmd_boolean },
126 { "backupconverted", &opt.backup_converted, cmd_boolean },
127 { "backups", &opt.backups, cmd_number },
128 { "base", &opt.base_href, cmd_string },
129 { "bindaddress", &opt.bind_address, cmd_string },
131 { "cacertificate", &opt.ca_cert, cmd_file },
133 { "cache", &opt.allow_cache, cmd_boolean },
135 { "cadirectory", &opt.ca_directory, cmd_directory },
136 { "certificate", &opt.cert_file, cmd_file },
137 { "certificatetype", &opt.cert_type, cmd_cert_type },
138 { "checkcertificate", &opt.check_cert, cmd_boolean },
140 { "chooseconfig", &opt.choose_config, cmd_file },
141 { "connecttimeout", &opt.connect_timeout, cmd_time },
142 { "contentdisposition", &opt.content_disposition, cmd_boolean },
143 { "contentonerror", &opt.content_on_error, cmd_boolean },
144 { "continue", &opt.always_rest, cmd_boolean },
145 { "convertlinks", &opt.convert_links, cmd_boolean },
146 { "cookies", &opt.cookies, cmd_boolean },
147 { "cutdirs", &opt.cut_dirs, cmd_number },
149 { "debug", &opt.debug, cmd_boolean },
151 { "defaultpage", &opt.default_page, cmd_string},
152 { "deleteafter", &opt.delete_after, cmd_boolean },
153 { "dirprefix", &opt.dir_prefix, cmd_directory },
154 { "dirstruct", NULL, cmd_spec_dirstruct },
155 { "dnscache", &opt.dns_cache, cmd_boolean },
156 { "dnstimeout", &opt.dns_timeout, cmd_time },
157 { "domains", &opt.domains, cmd_vector },
158 { "dotbytes", &opt.dot_bytes, cmd_bytes },
159 { "dotsinline", &opt.dots_in_line, cmd_number },
160 { "dotspacing", &opt.dot_spacing, cmd_number },
161 { "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
163 { "egdfile", &opt.egd_file, cmd_file },
165 { "excludedirectories", &opt.excludes, cmd_directory_vector },
166 { "excludedomains", &opt.exclude_domains, cmd_vector },
167 { "followftp", &opt.follow_ftp, cmd_boolean },
168 { "followtags", &opt.follow_tags, cmd_vector },
169 { "forcehtml", &opt.force_html, cmd_boolean },
170 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
171 { "ftppassword", &opt.ftp_passwd, cmd_string },
172 { "ftpproxy", &opt.ftp_proxy, cmd_string },
174 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
175 #endif /* def __VMS */
176 { "ftpuser", &opt.ftp_user, cmd_string },
177 { "glob", &opt.ftp_glob, cmd_boolean },
178 { "header", NULL, cmd_spec_header },
179 { "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
180 { "htmlify", NULL, cmd_spec_htmlify },
181 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
182 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
183 { "httppassword", &opt.http_passwd, cmd_string },
184 { "httpproxy", &opt.http_proxy, cmd_string },
185 { "httpsproxy", &opt.https_proxy, cmd_string },
186 { "httpuser", &opt.http_user, cmd_string },
187 { "ignorecase", &opt.ignore_case, cmd_boolean },
188 { "ignorelength", &opt.ignore_length, cmd_boolean },
189 { "ignoretags", &opt.ignore_tags, cmd_vector },
190 { "includedirectories", &opt.includes, cmd_directory_vector },
192 { "inet4only", &opt.ipv4_only, cmd_boolean },
193 { "inet6only", &opt.ipv6_only, cmd_boolean },
195 { "input", &opt.input_filename, cmd_file },
196 { "iri", &opt.enable_iri, cmd_boolean },
197 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
198 { "limitrate", &opt.limit_rate, cmd_bytes },
199 { "loadcookies", &opt.cookies_input, cmd_file },
200 { "localencoding", &opt.locale, cmd_string },
201 { "logfile", &opt.lfilename, cmd_file },
202 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
203 { "maxredirect", &opt.max_redirect, cmd_number },
204 { "mirror", NULL, cmd_spec_mirror },
205 { "netrc", &opt.netrc, cmd_boolean },
206 { "noclobber", &opt.noclobber, cmd_boolean },
207 { "noparent", &opt.no_parent, cmd_boolean },
208 { "noproxy", &opt.no_proxy, cmd_vector },
209 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
210 { "outputdocument", &opt.output_document, cmd_file },
211 { "pagerequisites", &opt.page_requisites, cmd_boolean },
212 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
213 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
214 { "password", &opt.passwd, cmd_string },
215 { "postdata", &opt.post_data, cmd_string },
216 { "postfile", &opt.post_file_name, cmd_file },
217 { "preferfamily", NULL, cmd_spec_prefer_family },
218 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
220 { "privatekey", &opt.private_key, cmd_file },
221 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
223 { "progress", &opt.progress_type, cmd_spec_progress },
224 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
225 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
226 { "proxypassword", &opt.proxy_passwd, cmd_string },
227 { "proxyuser", &opt.proxy_user, cmd_string },
228 { "quiet", &opt.quiet, cmd_boolean },
229 { "quota", &opt.quota, cmd_bytes_sum },
231 { "randomfile", &opt.random_file, cmd_file },
233 { "randomwait", &opt.random_wait, cmd_boolean },
234 { "readtimeout", &opt.read_timeout, cmd_time },
235 { "reclevel", &opt.reclevel, cmd_number_inf },
236 { "recursive", NULL, cmd_spec_recursive },
237 { "referer", &opt.referer, cmd_string },
238 { "reject", &opt.rejects, cmd_vector },
239 { "relativeonly", &opt.relative_only, cmd_boolean },
240 { "remoteencoding", &opt.encoding_remote, cmd_string },
241 { "removelisting", &opt.remove_listing, cmd_boolean },
242 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
243 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
244 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
245 { "robots", &opt.use_robots, cmd_boolean },
246 { "savecookies", &opt.cookies_output, cmd_file },
247 { "saveheaders", &opt.save_headers, cmd_boolean },
249 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
251 { "serverresponse", &opt.server_response, cmd_boolean },
252 { "showalldnsentries", &opt.show_all_dns_entries, cmd_boolean },
253 { "spanhosts", &opt.spanhost, cmd_boolean },
254 { "spider", &opt.spider, cmd_boolean },
255 { "strictcomments", &opt.strict_comments, cmd_boolean },
256 { "timeout", NULL, cmd_spec_timeout },
257 { "timestamping", &opt.timestamping, cmd_boolean },
258 { "tries", &opt.ntry, cmd_number_inf },
259 { "trustservernames", &opt.trustservernames, cmd_boolean },
260 { "unlink", &opt.unlink, cmd_boolean },
261 { "useproxy", &opt.use_proxy, cmd_boolean },
262 { "user", &opt.user, cmd_string },
263 { "useragent", NULL, cmd_spec_useragent },
264 { "useservertimestamps", &opt.useservertimestamps, cmd_boolean },
265 { "verbose", NULL, cmd_spec_verbose },
266 { "wait", &opt.wait, cmd_time },
267 { "waitretry", &opt.waitretry, cmd_time },
268 { "warccdx", &opt.warc_cdx_enabled, cmd_boolean },
269 { "warccdxdedup", &opt.warc_cdx_dedup_filename, cmd_file },
270 { "warccompression", &opt.warc_compression_enabled, cmd_boolean },
271 { "warcdigests", &opt.warc_digests_enabled, cmd_boolean },
272 { "warcfile", &opt.warc_filename, cmd_file },
273 { "warcheader", NULL, cmd_spec_warc_header },
274 { "warckeeplog", &opt.warc_keep_log, cmd_boolean },
275 { "warcmaxsize", &opt.warc_maxsize, cmd_bytes },
276 { "warctempdir", &opt.warc_tempdir, cmd_directory },
278 { "wdebug", &opt.wdebug, cmd_boolean },
282 /* Look up CMDNAME in the commands[] and return its position in the
283 array. If CMDNAME is not found, return -1. */
286 command_by_name (const char *cmdname)
288 /* Use binary search for speed. Wget has ~100 commands, which
289 guarantees a worst case performance of 7 string comparisons. */
290 int lo = 0, hi = countof (commands) - 1;
294 int mid = (lo + hi) >> 1;
295 int cmp = strcasecmp (cmdname, commands[mid].name);
306 /* Reset the variables to default values. */
312 /* Most of the default values are 0 (and 0.0, NULL, and false).
313 Just reset everything, and fill in the non-zero values. Note
314 that initializing pointers to NULL this way is technically
315 illegal, but porting Wget to a machine where NULL is not all-zero
316 bit pattern will be the least of the implementors' worries. */
323 opt.add_hostdir = true;
327 opt.http_keep_alive = true;
328 opt.use_proxy = true;
329 tmp = getenv ("no_proxy");
331 opt.no_proxy = sepstring (tmp);
332 opt.prefer_family = prefer_none;
333 opt.allow_cache = true;
335 opt.read_timeout = 900;
336 opt.use_robots = true;
338 opt.remove_listing = true;
340 opt.dot_bytes = 1024;
341 opt.dot_spacing = 10;
342 opt.dots_in_line = 50;
344 opt.dns_cache = true;
348 opt.check_cert = true;
351 /* The default for file name restriction defaults to the OS type. */
352 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
353 opt.restrict_files_os = restrict_windows;
355 opt.restrict_files_os = restrict_unix;
357 opt.restrict_files_ctrl = true;
358 opt.restrict_files_nonascii = false;
359 opt.restrict_files_case = restrict_no_case_restriction;
361 opt.max_redirect = 20;
366 opt.enable_iri = true;
368 opt.enable_iri = false;
371 opt.encoding_remote = NULL;
373 opt.useservertimestamps = true;
374 opt.show_all_dns_entries = false;
376 opt.warc_maxsize = 0; /* 1024 * 1024 * 1024; */
377 opt.warc_compression_enabled = true;
378 opt.warc_digests_enabled = true;
379 opt.warc_cdx_enabled = false;
380 opt.warc_cdx_dedup_filename = NULL;
381 opt.warc_tempdir = NULL;
382 opt.warc_keep_log = true;
385 /* Return the user's home directory (strdup-ed), or NULL if none is
390 static char *buf = NULL;
391 static char *home, *ret;
395 home = getenv ("HOME");
401 /* Under MSDOS, if $HOME isn't defined, use the directory where
402 `wget.exe' resides. */
403 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
406 buff = _w32_get_argv0 ();
408 p = strrchr (buf, '/'); /* djgpp */
410 p = strrchr (buf, '\\'); /* others */
414 buff = malloc (len + 1);
418 strncpy (buff, _w32_get_argv0 (), len);
422 #elif !defined(WINDOWS)
423 /* If HOME is not defined, try getting it from the password
425 struct passwd *pwd = getpwuid (getuid ());
426 if (!pwd || !pwd->pw_dir)
430 /* Under Windows, if $HOME isn't defined, use the directory where
431 `wget.exe' resides. */
437 ret = home ? xstrdup (home) : NULL;
444 /* Check the 'WGETRC' environment variable and return the file name
445 if 'WGETRC' is set and is a valid file.
446 If the `WGETRC' variable exists but the file does not exist, the
447 function will exit(). */
449 wgetrc_env_file_name (void)
451 char *env = getenv ("WGETRC");
454 if (!file_exists_p (env))
456 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
460 return xstrdup (env);
465 /* Check for the existance of '$HOME/.wgetrc' and return it's path
466 if it exists and is set. */
468 wgetrc_user_file_name (void)
472 /* If that failed, try $HOME/.wgetrc (or equivalent). */
475 file = "SYS$LOGIN:.wgetrc";
476 #else /* def __VMS */
479 file = aprintf ("%s/.wgetrc", home);
481 #endif /* def __VMS [else] */
485 if (!file_exists_p (file))
493 /* Return the path to the user's .wgetrc. This is either the value of
494 `WGETRC' environment variable, or `$HOME/.wgetrc'.
496 Additionally, for windows, look in the directory where wget.exe
499 wgetrc_file_name (void)
501 char *file = wgetrc_env_file_name ();
505 file = wgetrc_user_file_name ();
508 /* Under Windows, if we still haven't found .wgetrc, look for the file
509 `wget.ini' in the directory where `wget.exe' resides; we do this for
510 backward compatibility with previous versions of Wget.
511 SYSTEM_WGETRC should not be defined under WINDOWS. */
514 char *home = home_dir ();
520 file = aprintf ("%s/wget.ini", home);
521 if (!file_exists_p (file))
534 /* Return values of parse_line. */
542 static enum parse_line parse_line (const char *, char **, char **, int *);
543 static bool setval_internal (int, const char *, const char *);
544 static bool setval_internal_tilde (int, const char *, const char *);
546 /* Initialize variables from a wgetrc file. Returns zero (failure) if
547 there were errors in the file. */
550 run_wgetrc (const char *file)
557 fp = fopen (file, "r");
560 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
561 file, strerror (errno));
562 return true; /* not a fatal error */
565 while ((line = read_whole_line (fp)) != NULL)
567 char *com = NULL, *val = NULL;
570 /* Parse the line. */
571 switch (parse_line (line, &com, &val, &comind))
574 /* If everything is OK, set the value. */
575 if (!setval_internal_tilde (comind, com, val))
577 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
578 exec_name, file, ln);
582 case line_syntax_error:
583 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
584 exec_name, file, ln);
587 case line_unknown_command:
588 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
589 exec_name, quote (com), file, ln);
607 /* Initialize the defaults and run the system wgetrc and user's own
612 char *file, *env_sysrc;
615 /* Run a non-standard system rc file when the according environment
616 variable has been set. For internal testing purposes only! */
617 env_sysrc = getenv ("SYSTEM_WGETRC");
618 if (env_sysrc && file_exists_p (env_sysrc))
620 ok &= run_wgetrc (env_sysrc);
621 /* If there are any problems parsing the system wgetrc file, tell
625 fprintf (stderr, _("\
626 Parsing system wgetrc file (env SYSTEM_WGETRC) failed. Please check\n\
628 or specify a different file using --config.\n"), env_sysrc);
632 /* Otherwise, if SYSTEM_WGETRC is defined, use it. */
634 else if (file_exists_p (SYSTEM_WGETRC))
635 ok &= run_wgetrc (SYSTEM_WGETRC);
636 /* If there are any problems parsing the system wgetrc file, tell
640 fprintf (stderr, _("\
641 Parsing system wgetrc file failed. Please check\n\
643 or specify a different file using --config.\n"), SYSTEM_WGETRC);
647 /* Override it with your own, if one exists. */
648 file = wgetrc_file_name ();
651 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
652 something like realpath() before comparing them with `strcmp' */
654 if (!strcmp (file, SYSTEM_WGETRC))
656 fprintf (stderr, _("\
657 %s: Warning: Both system and user wgetrc point to %s.\n"),
658 exec_name, quote (file));
662 ok &= run_wgetrc (file);
664 /* If there were errors processing either `.wgetrc', abort. */
672 /* Remove dashes and underscores from S, modifying S in the
678 char *t = s; /* t - tortoise */
679 char *h = s; /* h - hare */
681 if (*h == '_' || *h == '-')
688 /* Parse the line pointed by line, with the syntax:
689 <sp>* command <sp>* = <sp>* value <sp>*
690 Uses malloc to allocate space for command and value.
692 Returns one of line_ok, line_empty, line_syntax_error, or
693 line_unknown_command.
695 In case of line_ok, *COM and *VAL point to freshly allocated
696 strings, and *COMIND points to com's index. In case of error or
697 empty line, their values are unmodified. */
699 static enum parse_line
700 parse_line (const char *line, char **com, char **val, int *comind)
703 const char *end = line + strlen (line);
704 const char *cmdstart, *cmdend;
705 const char *valstart, *valend;
710 /* Skip leading and trailing whitespace. */
711 while (*line && c_isspace (*line))
713 while (end > line && c_isspace (end[-1]))
716 /* Skip empty lines and comments. */
717 if (!*line || *line == '#')
723 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
727 /* Skip '=', as well as any space before or after it. */
728 while (p < end && c_isspace (*p))
730 if (p == end || *p != '=')
731 return line_syntax_error;
733 while (p < end && c_isspace (*p))
739 /* The syntax is valid (even though the command might not be). Fill
740 in the command name and value. */
741 *com = strdupdelim (cmdstart, cmdend);
742 *val = strdupdelim (valstart, valend);
744 /* The line now known to be syntactically correct. Check whether
745 the command is valid. */
746 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
748 ind = command_by_name (cmdcopy);
750 return line_unknown_command;
752 /* Report success to the caller. */
757 #if defined(WINDOWS) || defined(MSDOS)
758 # define ISSEP(c) ((c) == '/' || (c) == '\\')
760 # define ISSEP(c) ((c) == '/')
763 /* Run commands[comind].action. */
766 setval_internal (int comind, const char *com, const char *val)
768 assert (0 <= comind && ((size_t) comind) < countof (commands));
769 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
770 return commands[comind].action (com, val, commands[comind].place);
774 setval_internal_tilde (int comind, const char *com, const char *val)
780 ret = setval_internal (comind, com, val);
782 /* We make tilde expansion for cmd_file and cmd_directory */
783 if (((commands[comind].action == cmd_file) ||
784 (commands[comind].action == cmd_directory))
785 && ret && (*val == '~' && ISSEP (val[1])))
787 pstring = commands[comind].place;
791 homelen = strlen (home);
792 while (homelen && ISSEP (home[homelen - 1]))
793 home[--homelen] = '\0';
795 /* Skip the leading "~/". */
796 for (++val; ISSEP (*val); val++)
798 *pstring = concat_strings (home, "/", val, (char *)0);
804 /* Run command COM with value VAL. If running the command produces an
805 error, report the error and exit.
807 This is intended to be called from main() to modify Wget's behavior
808 through command-line switches. Since COM is hard-coded in main(),
809 it is not canonicalized, and this aborts when COM is not found.
811 If COMIND's are exported to init.h, this function will be changed
812 to accept COMIND directly. */
815 setoptval (const char *com, const char *val, const char *optname)
817 /* Prepend "--" to OPTNAME. */
818 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
821 strcpy (dd_optname + 2, optname);
823 assert (val != NULL);
824 if (!setval_internal (command_by_name (com), dd_optname, val))
828 /* Parse OPT into command and value and run it. For example,
829 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
830 This is used by the `--execute' flag in main.c. */
833 run_command (const char *opt)
837 switch (parse_line (opt, &com, &val, &comind))
840 if (!setval_internal (comind, com, val))
846 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
847 exec_name, quote (opt));
852 /* Generic helper functions, for use with `commands'. */
854 /* Forward declarations: */
859 static bool decode_string (const char *, const struct decode_item *, int, int *);
860 static bool simple_atoi (const char *, const char *, int *);
861 static bool simple_atof (const char *, const char *, double *);
863 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
865 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
866 && c_tolower((p)[1]) == (c1) \
869 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
870 && c_tolower((p)[1]) == (c1) \
871 && c_tolower((p)[2]) == (c2) \
875 /* Store the boolean value from VAL to PLACE. COM is ignored,
876 except for error messages. */
878 cmd_boolean (const char *com, const char *val, void *place)
882 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
883 /* "on", "yes" and "1" mean true. */
885 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
886 /* "off", "no" and "0" mean false. */
891 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
892 exec_name, com, quote (val));
896 *(bool *) place = value;
900 /* Set the non-negative integer value from VAL to PLACE. With
901 incorrect specification, the number remains unchanged. */
903 cmd_number (const char *com, const char *val, void *place)
905 if (!simple_atoi (val, val + strlen (val), place)
906 || *(int *) place < 0)
908 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
909 exec_name, com, quote (val));
915 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
917 cmd_number_inf (const char *com, const char *val, void *place)
919 if (!strcasecmp (val, "inf"))
924 return cmd_number (com, val, place);
927 /* Copy (strdup) the string at COM to a new location and place a
928 pointer to *PLACE. */
930 cmd_string (const char *com, const char *val, void *place)
932 char **pstring = (char **)place;
934 xfree_null (*pstring);
935 *pstring = xstrdup (val);
940 /* Like the above, but handles tilde-expansion when reading a user's
941 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
942 gets expanded to the user's home directory. */
944 cmd_file (const char *com, const char *val, void *place)
946 char **pstring = (char **)place;
948 xfree_null (*pstring);
950 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
952 *pstring = xstrdup (val);
954 #if defined(WINDOWS) || defined(MSDOS)
955 /* Convert "\" to "/". */
958 for (s = *pstring; *s; s++)
966 /* Like cmd_file, but strips trailing '/' characters. */
968 cmd_directory (const char *com, const char *val, void *place)
972 /* Call cmd_file() for tilde expansion and separator
973 canonicalization (backslash -> slash under Windows). These
974 things should perhaps be in a separate function. */
975 if (!cmd_file (com, val, place))
980 while (t > s && *--t == '/')
986 /* Split VAL by space to a vector of values, and append those values
987 to vector pointed to by the PLACE argument. If VAL is empty, the
988 PLACE vector is cleared instead. */
991 cmd_vector (const char *com, const char *val, void *place)
993 char ***pvec = (char ***)place;
996 *pvec = merge_vecs (*pvec, sepstring (val));
1006 cmd_directory_vector (const char *com, const char *val, void *place)
1008 char ***pvec = (char ***)place;
1012 /* Strip the trailing slashes from directories. */
1015 seps = sepstring (val);
1016 for (t = seps; t && *t; t++)
1018 int len = strlen (*t);
1019 /* Skip degenerate case of root directory. */
1022 if ((*t)[len - 1] == '/')
1023 (*t)[len - 1] = '\0';
1026 *pvec = merge_vecs (*pvec, seps);
1036 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
1037 "100k" or "2.5G" to a floating point number. */
1040 parse_bytes_helper (const char *val, double *result)
1042 double number, mult;
1043 const char *end = val + strlen (val);
1045 /* Check for "inf". */
1046 if (0 == strcmp (val, "inf"))
1052 /* Strip trailing whitespace. */
1053 while (val < end && c_isspace (end[-1]))
1058 switch (c_tolower (end[-1]))
1061 --end, mult = 1024.0;
1064 --end, mult = 1048576.0;
1067 --end, mult = 1073741824.0;
1070 --end, mult = 1099511627776.0;
1073 /* Not a recognized suffix: assume it's a digit. (If not,
1074 simple_atof will raise an error.) */
1078 /* Skip leading and trailing whitespace. */
1079 while (val < end && c_isspace (*val))
1081 while (val < end && c_isspace (end[-1]))
1086 if (!simple_atof (val, end, &number) || number < 0)
1089 *result = number * mult;
1093 /* Parse VAL as a number and set its value to PLACE (which should
1096 By default, the value is assumed to be in bytes. If "K", "M", or
1097 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
1098 1<<30, respectively. Floating point values are allowed and are
1099 cast to integer before use. The idea is to be able to use things
1100 like 1.5k instead of "1536".
1102 The string "inf" is returned as 0.
1104 In case of error, false is returned and memory pointed to by PLACE
1105 remains unmodified. */
1108 cmd_bytes (const char *com, const char *val, void *place)
1111 if (!parse_bytes_helper (val, &byte_value))
1113 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1114 exec_name, com, quote (val));
1117 *(wgint *)place = (wgint)byte_value;
1121 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1122 SIZE_SUM. It works by converting the string to double, therefore
1123 working with values up to 2^53-1 without loss of precision. This
1124 value (8192 TB) is large enough to serve for a while. */
1127 cmd_bytes_sum (const char *com, const char *val, void *place)
1130 if (!parse_bytes_helper (val, &byte_value))
1132 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1133 exec_name, com, quote (val));
1136 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1140 /* Store the value of VAL to *OUT. The value is a time period, by
1141 default expressed in seconds, but also accepting suffixes "m", "h",
1142 "d", and "w" for minutes, hours, days, and weeks respectively. */
1145 cmd_time (const char *com, const char *val, void *place)
1147 double number, mult;
1148 const char *end = val + strlen (val);
1150 /* Strip trailing whitespace. */
1151 while (val < end && c_isspace (end[-1]))
1157 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1158 exec_name, com, quote (val));
1162 switch (c_tolower (end[-1]))
1165 --end, mult = 1; /* seconds */
1168 --end, mult = 60; /* minutes */
1171 --end, mult = 3600; /* hours */
1174 --end, mult = 86400.0; /* days */
1177 --end, mult = 604800.0; /* weeks */
1180 /* Not a recognized suffix: assume it belongs to the number.
1181 (If not, simple_atof will raise an error.) */
1185 /* Skip leading and trailing whitespace. */
1186 while (val < end && c_isspace (*val))
1188 while (val < end && c_isspace (end[-1]))
1193 if (!simple_atof (val, end, &number))
1196 *(double *)place = number * mult;
1202 cmd_cert_type (const char *com, const char *val, void *place)
1204 static const struct decode_item choices[] = {
1205 { "pem", keyfile_pem },
1206 { "der", keyfile_asn1 },
1207 { "asn1", keyfile_asn1 },
1209 int ok = decode_string (val, choices, countof (choices), place);
1211 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1216 /* Specialized helper functions, used by `commands' to handle some
1217 options specially. */
1219 static bool check_user_specified_header (const char *);
1222 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1224 if (!cmd_boolean (com, val, &opt.dirstruct))
1226 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1227 must be affected inversely. */
1229 opt.no_dirstruct = false;
1231 opt.no_dirstruct = true;
1236 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1238 /* Empty value means reset the list of headers. */
1241 free_vec (opt.user_headers);
1242 opt.user_headers = NULL;
1246 if (!check_user_specified_header (val))
1248 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1249 exec_name, com, quote (val));
1252 opt.user_headers = vec_append (opt.user_headers, val);
1257 cmd_spec_warc_header (const char *com, const char *val, void *place_ignored)
1259 /* Empty value means reset the list of headers. */
1262 free_vec (opt.warc_user_headers);
1263 opt.warc_user_headers = NULL;
1267 if (!check_user_specified_header (val))
1269 fprintf (stderr, _("%s: %s: Invalid WARC header %s.\n"),
1270 exec_name, com, quote (val));
1273 opt.warc_user_headers = vec_append (opt.warc_user_headers, val);
1278 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1280 int flag = cmd_boolean (com, val, &opt.htmlify);
1281 if (flag && !opt.htmlify)
1282 opt.remove_listing = false;
1286 /* Set the "mirror" mode. It means: recursive download, timestamping,
1287 no limit on max. recursion depth, and don't remove listings. */
1290 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1294 if (!cmd_boolean (com, val, &mirror))
1298 opt.recursive = true;
1299 if (!opt.no_dirstruct)
1300 opt.dirstruct = true;
1301 opt.timestamping = true;
1302 opt.reclevel = INFINITE_RECURSION;
1303 opt.remove_listing = false;
1308 /* Validate --prefer-family and set the choice. Allowed values are
1309 "IPv4", "IPv6", and "none". */
1312 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1314 static const struct decode_item choices[] = {
1315 { "IPv4", prefer_ipv4 },
1316 { "IPv6", prefer_ipv6 },
1317 { "none", prefer_none },
1319 int prefer_family = prefer_none;
1320 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1322 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1323 opt.prefer_family = prefer_family;
1327 /* Set progress.type to VAL, but verify that it's a valid progress
1328 implementation before that. */
1331 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1333 if (!valid_progress_implementation_p (val))
1335 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1336 exec_name, com, quote (val));
1339 xfree_null (opt.progress_type);
1341 /* Don't call set_progress_implementation here. It will be called
1342 in main() when it becomes clear what the log output is. */
1343 opt.progress_type = xstrdup (val);
1347 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1348 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1352 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1354 if (!cmd_boolean (com, val, &opt.recursive))
1358 if (opt.recursive && !opt.no_dirstruct)
1359 opt.dirstruct = true;
1365 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1367 int restrict_os = opt.restrict_files_os;
1368 int restrict_ctrl = opt.restrict_files_ctrl;
1369 int restrict_case = opt.restrict_files_case;
1370 int restrict_nonascii = opt.restrict_files_nonascii;
1374 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1378 end = strchr (val, ',');
1380 end = val + strlen (val);
1382 if (VAL_IS ("unix"))
1383 restrict_os = restrict_unix;
1384 else if (VAL_IS ("windows"))
1385 restrict_os = restrict_windows;
1386 else if (VAL_IS ("lowercase"))
1387 restrict_case = restrict_lowercase;
1388 else if (VAL_IS ("uppercase"))
1389 restrict_case = restrict_uppercase;
1390 else if (VAL_IS ("nocontrol"))
1391 restrict_ctrl = false;
1392 else if (VAL_IS ("ascii"))
1393 restrict_nonascii = true;
1396 fprintf (stderr, _("\
1397 %s: %s: Invalid restriction %s,\n\
1398 use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1399 exec_name, com, quote (val));
1406 while (*val && *end);
1410 opt.restrict_files_os = restrict_os;
1411 opt.restrict_files_ctrl = restrict_ctrl;
1412 opt.restrict_files_case = restrict_case;
1413 opt.restrict_files_nonascii = restrict_nonascii;
1420 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1422 static const struct decode_item choices[] = {
1423 { "auto", secure_protocol_auto },
1424 { "sslv2", secure_protocol_sslv2 },
1425 { "sslv3", secure_protocol_sslv3 },
1426 { "tlsv1", secure_protocol_tlsv1 },
1428 int ok = decode_string (val, choices, countof (choices), place);
1430 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1435 /* Set all three timeout values. */
1438 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1441 if (!cmd_time (com, val, &value))
1443 opt.read_timeout = value;
1444 opt.connect_timeout = value;
1445 opt.dns_timeout = value;
1450 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1452 /* Disallow embedded newlines. */
1453 if (strchr (val, '\n'))
1455 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1456 exec_name, com, quote (val));
1459 xfree_null (opt.useragent);
1460 opt.useragent = xstrdup (val);
1464 /* The "verbose" option cannot be cmd_boolean because the variable is
1465 not bool -- it's of type int (-1 means uninitialized because of
1466 some random hackery for disallowing -q -v). */
1469 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1472 if (cmd_boolean (com, val, &flag))
1480 /* Miscellaneous useful routines. */
1482 /* A very simple atoi clone, more useful than atoi because it works on
1483 delimited strings, and has error reportage. Returns true on success,
1484 false on failure. If successful, stores result to *DEST. */
1487 simple_atoi (const char *beg, const char *end, int *dest)
1490 bool negative = false;
1491 const char *p = beg;
1493 while (p < end && c_isspace (*p))
1495 if (p < end && (*p == '-' || *p == '+'))
1497 negative = (*p == '-');
1503 /* Read negative numbers in a separate loop because the most
1504 negative integer cannot be represented as a positive number. */
1507 for (; p < end && c_isdigit (*p); p++)
1509 int next = (10 * result) + (*p - '0');
1511 return false; /* overflow */
1515 for (; p < end && c_isdigit (*p); p++)
1517 int next = (10 * result) - (*p - '0');
1519 return false; /* underflow */
1530 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1531 doesn't handle exponential notation. Returns true on success,
1532 false on failure. In case of success, stores its result to
1536 simple_atof (const char *beg, const char *end, double *dest)
1540 bool negative = false;
1541 bool seen_dot = false;
1542 bool seen_digit = false;
1545 const char *p = beg;
1547 while (p < end && c_isspace (*p))
1549 if (p < end && (*p == '-' || *p == '+'))
1551 negative = (*p == '-');
1555 for (; p < end; p++)
1561 result = (10 * result) + (ch - '0');
1563 result += (ch - '0') / (divider *= 10);
1585 /* Verify that the user-specified header in S is valid. It must
1586 contain a colon preceded by non-white-space characters and must not
1587 contain newlines. */
1590 check_user_specified_header (const char *s)
1594 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1596 /* The header MUST contain `:' preceded by at least one
1597 non-whitespace character. */
1598 if (*p != ':' || p == s)
1600 /* The header MUST NOT contain newlines. */
1601 if (strchr (s, '\n'))
1606 /* Decode VAL into a number, according to ITEMS. */
1609 decode_string (const char *val, const struct decode_item *items, int itemcount,
1613 for (i = 0; i < itemcount; i++)
1614 if (0 == strcasecmp (val, items[i].name))
1616 *place = items[i].code;
1623 void cleanup_html_url (void);
1626 /* Free the memory allocated by global variables. */
1630 /* Free external resources, close files, etc. */
1633 fclose (output_stream);
1634 /* No need to check for error because Wget flushes its output (and
1635 checks for errors) after any data arrives. */
1637 /* We're exiting anyway so there's no real need to call free()
1638 hundreds of times. Skipping the frees will make Wget exit
1641 However, when detecting leaks, it's crucial to free() everything
1642 because then you can find the real leaks, i.e. the allocated
1643 memory which grows with the size of the program. */
1649 cleanup_html_url ();
1654 extern acc_t *netrc_list;
1655 free_netrc (netrc_list);
1657 xfree_null (opt.choose_config);
1658 xfree_null (opt.lfilename);
1659 xfree_null (opt.dir_prefix);
1660 xfree_null (opt.input_filename);
1661 xfree_null (opt.output_document);
1662 free_vec (opt.accepts);
1663 free_vec (opt.rejects);
1664 free_vec (opt.excludes);
1665 free_vec (opt.includes);
1666 free_vec (opt.domains);
1667 free_vec (opt.follow_tags);
1668 free_vec (opt.ignore_tags);
1669 xfree_null (opt.progress_type);
1670 xfree_null (opt.ftp_user);
1671 xfree_null (opt.ftp_passwd);
1672 xfree_null (opt.ftp_proxy);
1673 xfree_null (opt.https_proxy);
1674 xfree_null (opt.http_proxy);
1675 free_vec (opt.no_proxy);
1676 xfree_null (opt.useragent);
1677 xfree_null (opt.referer);
1678 xfree_null (opt.http_user);
1679 xfree_null (opt.http_passwd);
1680 free_vec (opt.user_headers);
1681 free_vec (opt.warc_user_headers);
1683 xfree_null (opt.cert_file);
1684 xfree_null (opt.private_key);
1685 xfree_null (opt.ca_directory);
1686 xfree_null (opt.ca_cert);
1687 xfree_null (opt.random_file);
1688 xfree_null (opt.egd_file);
1690 xfree_null (opt.bind_address);
1691 xfree_null (opt.cookies_input);
1692 xfree_null (opt.cookies_output);
1693 xfree_null (opt.user);
1694 xfree_null (opt.passwd);
1695 xfree_null (opt.base_href);
1697 #endif /* DEBUG_MALLOC */
1700 /* Unit testing routines. */
1705 test_commands_sorted()
1707 int prev_idx = 0, next_idx = 1;
1708 int command_count = countof (commands) - 1;
1710 while (next_idx <= command_count)
1712 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1715 mu_assert ("FAILED", false);
1728 test_cmd_spec_restrict_file_names()
1733 int expected_restrict_files_os;
1734 int expected_restrict_files_ctrl;
1735 int expected_restrict_files_case;
1738 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1739 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1740 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1741 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1744 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1749 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1752 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1753 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1754 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1755 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1757 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1758 res == test_array[i].result
1759 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1760 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1761 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1767 #endif /* TESTING */