1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007, 2008 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. */
51 #include "recur.h" /* for INFINITE_RECURSION */
52 #include "convert.h" /* for convert_cleanup */
53 #include "res.h" /* for res_cleanup */
54 #include "http.h" /* for http_cleanup */
55 #include "retr.h" /* for output_stream */
61 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
62 otherwise, it will be performed by the shell. This variable will
63 be set by the wgetrc-reading function. */
65 static bool enable_tilde_expansion;
68 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
70 CMD_DECLARE (cmd_boolean);
71 CMD_DECLARE (cmd_bytes);
72 CMD_DECLARE (cmd_bytes_sum);
74 CMD_DECLARE (cmd_cert_type);
76 CMD_DECLARE (cmd_directory_vector);
77 CMD_DECLARE (cmd_number);
78 CMD_DECLARE (cmd_number_inf);
79 CMD_DECLARE (cmd_string);
80 CMD_DECLARE (cmd_file);
81 CMD_DECLARE (cmd_directory);
82 CMD_DECLARE (cmd_time);
83 CMD_DECLARE (cmd_vector);
85 CMD_DECLARE (cmd_spec_dirstruct);
86 CMD_DECLARE (cmd_spec_header);
87 CMD_DECLARE (cmd_spec_htmlify);
88 CMD_DECLARE (cmd_spec_mirror);
89 CMD_DECLARE (cmd_spec_prefer_family);
90 CMD_DECLARE (cmd_spec_progress);
91 CMD_DECLARE (cmd_spec_recursive);
92 CMD_DECLARE (cmd_spec_restrict_file_names);
94 CMD_DECLARE (cmd_spec_secure_protocol);
96 CMD_DECLARE (cmd_spec_timeout);
97 CMD_DECLARE (cmd_spec_useragent);
98 CMD_DECLARE (cmd_spec_verbose);
100 /* List of recognized commands, each consisting of name, place and
101 function. When adding a new command, simply add it to the list,
102 but be sure to keep the list sorted alphabetically, as
103 command_by_name's binary search depends on it. Also, be sure to
104 add any entries that allocate memory (e.g. cmd_string and
105 cmd_vector) to the cleanup() function below. */
107 static const struct {
110 bool (*action) (const char *, const char *, void *);
112 /* KEEP THIS LIST ALPHABETICALLY SORTED */
113 { "accept", &opt.accepts, cmd_vector },
114 { "addhostdir", &opt.add_hostdir, cmd_boolean },
115 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
116 { "authnochallenge", &opt.auth_without_challenge,
118 { "background", &opt.background, cmd_boolean },
119 { "backupconverted", &opt.backup_converted, cmd_boolean },
120 { "backups", &opt.backups, cmd_number },
121 { "base", &opt.base_href, cmd_string },
122 { "bindaddress", &opt.bind_address, cmd_string },
124 { "cacertificate", &opt.ca_cert, cmd_file },
126 { "cache", &opt.allow_cache, cmd_boolean },
128 { "cadirectory", &opt.ca_directory, cmd_directory },
129 { "certificate", &opt.cert_file, cmd_file },
130 { "certificatetype", &opt.cert_type, cmd_cert_type },
131 { "checkcertificate", &opt.check_cert, cmd_boolean },
133 { "connecttimeout", &opt.connect_timeout, cmd_time },
134 { "contentdisposition", &opt.content_disposition, cmd_boolean },
135 { "continue", &opt.always_rest, cmd_boolean },
136 { "convertlinks", &opt.convert_links, cmd_boolean },
137 { "cookies", &opt.cookies, cmd_boolean },
138 { "cutdirs", &opt.cut_dirs, cmd_number },
140 { "debug", &opt.debug, cmd_boolean },
142 { "deleteafter", &opt.delete_after, cmd_boolean },
143 { "dirprefix", &opt.dir_prefix, cmd_directory },
144 { "dirstruct", NULL, cmd_spec_dirstruct },
145 { "dnscache", &opt.dns_cache, cmd_boolean },
146 { "dnstimeout", &opt.dns_timeout, cmd_time },
147 { "domains", &opt.domains, cmd_vector },
148 { "dotbytes", &opt.dot_bytes, cmd_bytes },
149 { "dotsinline", &opt.dots_in_line, cmd_number },
150 { "dotspacing", &opt.dot_spacing, cmd_number },
151 { "dotstyle", &opt.dot_style, cmd_string },
153 { "egdfile", &opt.egd_file, cmd_file },
155 { "excludedirectories", &opt.excludes, cmd_directory_vector },
156 { "excludedomains", &opt.exclude_domains, cmd_vector },
157 { "followftp", &opt.follow_ftp, cmd_boolean },
158 { "followtags", &opt.follow_tags, cmd_vector },
159 { "forcehtml", &opt.force_html, cmd_boolean },
160 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
161 { "ftppassword", &opt.ftp_passwd, cmd_string },
162 { "ftpproxy", &opt.ftp_proxy, cmd_string },
164 { "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
165 #endif /* def __VMS */
166 { "ftpuser", &opt.ftp_user, cmd_string },
167 { "glob", &opt.ftp_glob, cmd_boolean },
168 { "header", NULL, cmd_spec_header },
169 { "htmlextension", &opt.html_extension, cmd_boolean },
170 { "htmlify", NULL, cmd_spec_htmlify },
171 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
172 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
173 { "httppassword", &opt.http_passwd, cmd_string },
174 { "httpproxy", &opt.http_proxy, cmd_string },
175 { "httpsproxy", &opt.https_proxy, cmd_string },
176 { "httpuser", &opt.http_user, cmd_string },
177 { "ignorecase", &opt.ignore_case, cmd_boolean },
178 { "ignorelength", &opt.ignore_length, cmd_boolean },
179 { "ignoretags", &opt.ignore_tags, cmd_vector },
180 { "includedirectories", &opt.includes, cmd_directory_vector },
182 { "inet4only", &opt.ipv4_only, cmd_boolean },
183 { "inet6only", &opt.ipv6_only, cmd_boolean },
185 { "input", &opt.input_filename, cmd_file },
186 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
187 { "limitrate", &opt.limit_rate, cmd_bytes },
188 { "loadcookies", &opt.cookies_input, cmd_file },
189 { "logfile", &opt.lfilename, cmd_file },
190 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
191 { "maxredirect", &opt.max_redirect, cmd_number },
192 { "mirror", NULL, cmd_spec_mirror },
193 { "netrc", &opt.netrc, cmd_boolean },
194 { "noclobber", &opt.noclobber, cmd_boolean },
195 { "noparent", &opt.no_parent, cmd_boolean },
196 { "noproxy", &opt.no_proxy, cmd_vector },
197 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
198 { "outputdocument", &opt.output_document, cmd_file },
199 { "pagerequisites", &opt.page_requisites, cmd_boolean },
200 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
201 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
202 { "password", &opt.passwd, cmd_string },
203 { "postdata", &opt.post_data, cmd_string },
204 { "postfile", &opt.post_file_name, cmd_file },
205 { "preferfamily", NULL, cmd_spec_prefer_family },
206 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
208 { "privatekey", &opt.private_key, cmd_file },
209 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
211 { "progress", &opt.progress_type, cmd_spec_progress },
212 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
213 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
214 { "proxypassword", &opt.proxy_passwd, cmd_string },
215 { "proxyuser", &opt.proxy_user, cmd_string },
216 { "quiet", &opt.quiet, cmd_boolean },
217 { "quota", &opt.quota, cmd_bytes_sum },
219 { "randomfile", &opt.random_file, cmd_file },
221 { "randomwait", &opt.random_wait, cmd_boolean },
222 { "readtimeout", &opt.read_timeout, cmd_time },
223 { "reclevel", &opt.reclevel, cmd_number_inf },
224 { "recursive", NULL, cmd_spec_recursive },
225 { "referer", &opt.referer, cmd_string },
226 { "reject", &opt.rejects, cmd_vector },
227 { "relativeonly", &opt.relative_only, cmd_boolean },
228 { "removelisting", &opt.remove_listing, cmd_boolean },
229 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
230 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
231 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
232 { "robots", &opt.use_robots, cmd_boolean },
233 { "savecookies", &opt.cookies_output, cmd_file },
234 { "saveheaders", &opt.save_headers, cmd_boolean },
236 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
238 { "serverresponse", &opt.server_response, cmd_boolean },
239 { "spanhosts", &opt.spanhost, cmd_boolean },
240 { "spider", &opt.spider, cmd_boolean },
241 { "strictcomments", &opt.strict_comments, cmd_boolean },
242 { "timeout", NULL, cmd_spec_timeout },
243 { "timestamping", &opt.timestamping, cmd_boolean },
244 { "tries", &opt.ntry, cmd_number_inf },
245 { "useproxy", &opt.use_proxy, cmd_boolean },
246 { "user", &opt.user, cmd_string },
247 { "useragent", NULL, cmd_spec_useragent },
248 { "verbose", NULL, cmd_spec_verbose },
249 { "wait", &opt.wait, cmd_time },
250 { "waitretry", &opt.waitretry, cmd_time },
252 { "wdebug", &opt.wdebug, cmd_boolean },
256 /* Look up CMDNAME in the commands[] and return its position in the
257 array. If CMDNAME is not found, return -1. */
260 command_by_name (const char *cmdname)
262 /* Use binary search for speed. Wget has ~100 commands, which
263 guarantees a worst case performance of 7 string comparisons. */
264 int lo = 0, hi = countof (commands) - 1;
268 int mid = (lo + hi) >> 1;
269 int cmp = strcasecmp (cmdname, commands[mid].name);
280 /* Reset the variables to default values. */
286 /* Most of the default values are 0 (and 0.0, NULL, and false).
287 Just reset everything, and fill in the non-zero values. Note
288 that initializing pointers to NULL this way is technically
289 illegal, but porting Wget to a machine where NULL is not all-zero
290 bit pattern will be the least of the implementors' worries. */
297 opt.add_hostdir = true;
301 opt.http_keep_alive = true;
302 opt.use_proxy = true;
303 tmp = getenv ("no_proxy");
305 opt.no_proxy = sepstring (tmp);
306 opt.allow_cache = true;
308 opt.read_timeout = 900;
309 opt.use_robots = true;
311 opt.remove_listing = true;
313 opt.dot_bytes = 1024;
314 opt.dot_spacing = 10;
315 opt.dots_in_line = 50;
317 opt.dns_cache = true;
321 opt.check_cert = true;
324 /* The default for file name restriction defaults to the OS type. */
325 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
326 opt.restrict_files_os = restrict_windows;
328 opt.restrict_files_os = restrict_unix;
330 opt.restrict_files_ctrl = true;
331 opt.restrict_files_case = restrict_no_case_restriction;
333 opt.max_redirect = 20;
336 /* Return the user's home directory (strdup-ed), or NULL if none is
341 char *home = getenv ("HOME");
346 /* Under MSDOS, if $HOME isn't defined, use the directory where
347 `wget.exe' resides. */
348 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
349 char *p, buf[PATH_MAX];
351 strcpy (buf, _w32_get_argv0 ());
352 p = strrchr (buf, '/'); /* djgpp */
354 p = strrchr (buf, '\\'); /* others */
358 #elif !defined(WINDOWS)
359 /* If HOME is not defined, try getting it from the password
361 struct passwd *pwd = getpwuid (getuid ());
362 if (!pwd || !pwd->pw_dir)
366 /* Under Windows, if $HOME isn't defined, use the directory where
367 `wget.exe' resides. */
372 return home ? xstrdup (home) : NULL;
375 /* Return the path to the user's .wgetrc. This is either the value of
376 `WGETRC' environment variable, or `$HOME/.wgetrc'.
378 If the `WGETRC' variable exists but the file does not exist, the
379 function will exit(). */
381 wgetrc_file_name (void)
386 /* Try the environment. */
387 env = getenv ("WGETRC");
390 if (!file_exists_p (env))
392 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
396 return xstrdup (env);
399 /* If that failed, try $HOME/.wgetrc (or equivalent). */
402 file = "SYS$LOGIN:.wgetrc";
403 #else /* def __VMS */
406 file = aprintf ("%s/.wgetrc", home);
408 #endif /* def __VMS [else] */
411 /* Under Windows, if we still haven't found .wgetrc, look for the file
412 `wget.ini' in the directory where `wget.exe' resides; we do this for
413 backward compatibility with previous versions of Wget.
414 SYSTEM_WGETRC should not be defined under WINDOWS. */
415 if (!file || !file_exists_p (file))
421 file = aprintf ("%s/wget.ini", home);
427 if (!file_exists_p (file))
435 /* Return values of parse_line. */
443 static enum parse_line parse_line (const char *, char **, char **, int *);
444 static bool setval_internal (int, const char *, const char *);
446 /* Initialize variables from a wgetrc file. Returns zero (failure) if
447 there were errors in the file. */
450 run_wgetrc (const char *file)
457 fp = fopen (file, "r");
460 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
461 file, strerror (errno));
462 return true; /* not a fatal error */
464 enable_tilde_expansion = true;
466 while ((line = read_whole_line (fp)) != NULL)
468 char *com = NULL, *val = NULL;
471 /* Parse the line. */
472 switch (parse_line (line, &com, &val, &comind))
475 /* If everything is OK, set the value. */
476 if (!setval_internal (comind, com, val))
478 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
479 exec_name, file, ln);
483 case line_syntax_error:
484 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
485 exec_name, file, ln);
488 case line_unknown_command:
489 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
490 exec_name, com, file, ln);
503 enable_tilde_expansion = false;
509 /* Initialize the defaults and run the system wgetrc and user's own
517 /* Load the hard-coded defaults. */
520 /* If SYSTEM_WGETRC is defined, use it. */
522 if (file_exists_p (SYSTEM_WGETRC))
523 ok &= run_wgetrc (SYSTEM_WGETRC);
525 /* Override it with your own, if one exists. */
526 file = wgetrc_file_name ();
529 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
530 something like realpath() before comparing them with `strcmp' */
532 if (!strcmp (file, SYSTEM_WGETRC))
534 fprintf (stderr, _("\
535 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
540 ok &= run_wgetrc (file);
542 /* If there were errors processing either `.wgetrc', abort. */
550 /* Remove dashes and underscores from S, modifying S in the
556 char *t = s; /* t - tortoise */
557 char *h = s; /* h - hare */
559 if (*h == '_' || *h == '-')
566 /* Parse the line pointed by line, with the syntax:
567 <sp>* command <sp>* = <sp>* value <sp>*
568 Uses malloc to allocate space for command and value.
570 Returns one of line_ok, line_empty, line_syntax_error, or
571 line_unknown_command.
573 In case of line_ok, *COM and *VAL point to freshly allocated
574 strings, and *COMIND points to com's index. In case of error or
575 empty line, their values are unmodified. */
577 static enum parse_line
578 parse_line (const char *line, char **com, char **val, int *comind)
581 const char *end = line + strlen (line);
582 const char *cmdstart, *cmdend;
583 const char *valstart, *valend;
588 /* Skip leading and trailing whitespace. */
589 while (*line && c_isspace (*line))
591 while (end > line && c_isspace (end[-1]))
594 /* Skip empty lines and comments. */
595 if (!*line || *line == '#')
601 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
605 /* Skip '=', as well as any space before or after it. */
606 while (p < end && c_isspace (*p))
608 if (p == end || *p != '=')
609 return line_syntax_error;
611 while (p < end && c_isspace (*p))
617 /* The syntax is valid (even though the command might not be). Fill
618 in the command name and value. */
619 *com = strdupdelim (cmdstart, cmdend);
620 *val = strdupdelim (valstart, valend);
622 /* The line now known to be syntactically correct. Check whether
623 the command is valid. */
624 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
626 ind = command_by_name (cmdcopy);
628 return line_unknown_command;
630 /* Report success to the caller. */
635 /* Run commands[comind].action. */
638 setval_internal (int comind, const char *com, const char *val)
640 assert (0 <= comind && comind < countof (commands));
641 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
642 return commands[comind].action (com, val, commands[comind].place);
645 /* Run command COM with value VAL. If running the command produces an
646 error, report the error and exit.
648 This is intended to be called from main() to modify Wget's behavior
649 through command-line switches. Since COM is hard-coded in main(),
650 it is not canonicalized, and this aborts when COM is not found.
652 If COMIND's are exported to init.h, this function will be changed
653 to accept COMIND directly. */
656 setoptval (const char *com, const char *val, const char *optname)
658 /* Prepend "--" to OPTNAME. */
659 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
662 strcpy (dd_optname + 2, optname);
664 assert (val != NULL);
665 if (!setval_internal (command_by_name (com), dd_optname, val))
669 /* Parse OPT into command and value and run it. For example,
670 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
671 This is used by the `--execute' flag in main.c. */
674 run_command (const char *opt)
678 switch (parse_line (opt, &com, &val, &comind))
681 if (!setval_internal (comind, com, val))
687 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
693 /* Generic helper functions, for use with `commands'. */
695 /* Forward declarations: */
700 static bool decode_string (const char *, const struct decode_item *, int, int *);
701 static bool simple_atoi (const char *, const char *, int *);
702 static bool simple_atof (const char *, const char *, double *);
704 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
706 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
707 && c_tolower((p)[1]) == (c1) \
710 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
711 && c_tolower((p)[1]) == (c1) \
712 && c_tolower((p)[2]) == (c2) \
716 /* Store the boolean value from VAL to PLACE. COM is ignored,
717 except for error messages. */
719 cmd_boolean (const char *com, const char *val, void *place)
723 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
724 /* "on", "yes" and "1" mean true. */
726 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
727 /* "off", "no" and "0" mean false. */
732 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
733 exec_name, com, val);
737 *(bool *) place = value;
741 /* Set the non-negative integer value from VAL to PLACE. With
742 incorrect specification, the number remains unchanged. */
744 cmd_number (const char *com, const char *val, void *place)
746 if (!simple_atoi (val, val + strlen (val), place)
747 || *(int *) place < 0)
749 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
750 exec_name, com, val);
756 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
758 cmd_number_inf (const char *com, const char *val, void *place)
760 if (!strcasecmp (val, "inf"))
765 return cmd_number (com, val, place);
768 /* Copy (strdup) the string at COM to a new location and place a
769 pointer to *PLACE. */
771 cmd_string (const char *com, const char *val, void *place)
773 char **pstring = (char **)place;
775 xfree_null (*pstring);
776 *pstring = xstrdup (val);
780 #if defined(WINDOWS) || defined(MSDOS)
781 # define ISSEP(c) ((c) == '/' || (c) == '\\')
783 # define ISSEP(c) ((c) == '/')
786 /* Like the above, but handles tilde-expansion when reading a user's
787 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
788 gets expanded to the user's home directory. */
790 cmd_file (const char *com, const char *val, void *place)
792 char **pstring = (char **)place;
794 xfree_null (*pstring);
796 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
798 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
801 *pstring = xstrdup (val);
806 char *home = home_dir ();
810 homelen = strlen (home);
811 while (homelen && ISSEP (home[homelen - 1]))
812 home[--homelen] = '\0';
814 /* Skip the leading "~/". */
815 for (++val; ISSEP (*val); val++)
818 *pstring = concat_strings (home, "/", val, (char *) 0);
821 #if defined(WINDOWS) || defined(MSDOS)
822 /* Convert "\" to "/". */
825 for (s = *pstring; *s; s++)
833 /* Like cmd_file, but strips trailing '/' characters. */
835 cmd_directory (const char *com, const char *val, void *place)
839 /* Call cmd_file() for tilde expansion and separator
840 canonicalization (backslash -> slash under Windows). These
841 things should perhaps be in a separate function. */
842 if (!cmd_file (com, val, place))
847 while (t > s && *--t == '/')
853 /* Split VAL by space to a vector of values, and append those values
854 to vector pointed to by the PLACE argument. If VAL is empty, the
855 PLACE vector is cleared instead. */
858 cmd_vector (const char *com, const char *val, void *place)
860 char ***pvec = (char ***)place;
863 *pvec = merge_vecs (*pvec, sepstring (val));
873 cmd_directory_vector (const char *com, const char *val, void *place)
875 char ***pvec = (char ***)place;
879 /* Strip the trailing slashes from directories. */
882 seps = sepstring (val);
883 for (t = seps; t && *t; t++)
885 int len = strlen (*t);
886 /* Skip degenerate case of root directory. */
889 if ((*t)[len - 1] == '/')
890 (*t)[len - 1] = '\0';
893 *pvec = merge_vecs (*pvec, seps);
903 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
904 "100k" or "2.5G" to a floating point number. */
907 parse_bytes_helper (const char *val, double *result)
910 const char *end = val + strlen (val);
912 /* Check for "inf". */
913 if (0 == strcmp (val, "inf"))
919 /* Strip trailing whitespace. */
920 while (val < end && c_isspace (end[-1]))
925 switch (c_tolower (end[-1]))
928 --end, mult = 1024.0;
931 --end, mult = 1048576.0;
934 --end, mult = 1073741824.0;
937 --end, mult = 1099511627776.0;
940 /* Not a recognized suffix: assume it's a digit. (If not,
941 simple_atof will raise an error.) */
945 /* Skip leading and trailing whitespace. */
946 while (val < end && c_isspace (*val))
948 while (val < end && c_isspace (end[-1]))
953 if (!simple_atof (val, end, &number) || number < 0)
956 *result = number * mult;
960 /* Parse VAL as a number and set its value to PLACE (which should
963 By default, the value is assumed to be in bytes. If "K", "M", or
964 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
965 1<<30, respectively. Floating point values are allowed and are
966 cast to integer before use. The idea is to be able to use things
967 like 1.5k instead of "1536".
969 The string "inf" is returned as 0.
971 In case of error, false is returned and memory pointed to by PLACE
972 remains unmodified. */
975 cmd_bytes (const char *com, const char *val, void *place)
978 if (!parse_bytes_helper (val, &byte_value))
980 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
981 exec_name, com, val);
984 *(wgint *)place = (wgint)byte_value;
988 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
989 SIZE_SUM. It works by converting the string to double, therefore
990 working with values up to 2^53-1 without loss of precision. This
991 value (8192 TB) is large enough to serve for a while. */
994 cmd_bytes_sum (const char *com, const char *val, void *place)
997 if (!parse_bytes_helper (val, &byte_value))
999 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
1000 exec_name, com, val);
1003 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1007 /* Store the value of VAL to *OUT. The value is a time period, by
1008 default expressed in seconds, but also accepting suffixes "m", "h",
1009 "d", and "w" for minutes, hours, days, and weeks respectively. */
1012 cmd_time (const char *com, const char *val, void *place)
1014 double number, mult;
1015 const char *end = val + strlen (val);
1017 /* Strip trailing whitespace. */
1018 while (val < end && c_isspace (end[-1]))
1024 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1025 exec_name, com, val);
1029 switch (c_tolower (end[-1]))
1032 --end, mult = 1; /* seconds */
1035 --end, mult = 60; /* minutes */
1038 --end, mult = 3600; /* hours */
1041 --end, mult = 86400.0; /* days */
1044 --end, mult = 604800.0; /* weeks */
1047 /* Not a recognized suffix: assume it belongs to the number.
1048 (If not, simple_atof will raise an error.) */
1052 /* Skip leading and trailing whitespace. */
1053 while (val < end && c_isspace (*val))
1055 while (val < end && c_isspace (end[-1]))
1060 if (!simple_atof (val, end, &number))
1063 *(double *)place = number * mult;
1069 cmd_cert_type (const char *com, const char *val, void *place)
1071 static const struct decode_item choices[] = {
1072 { "pem", keyfile_pem },
1073 { "der", keyfile_asn1 },
1074 { "asn1", keyfile_asn1 },
1076 int ok = decode_string (val, choices, countof (choices), place);
1078 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1083 /* Specialized helper functions, used by `commands' to handle some
1084 options specially. */
1086 static bool check_user_specified_header (const char *);
1089 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1091 if (!cmd_boolean (com, val, &opt.dirstruct))
1093 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1094 must be affected inversely. */
1096 opt.no_dirstruct = false;
1098 opt.no_dirstruct = true;
1103 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1105 /* Empty value means reset the list of headers. */
1108 free_vec (opt.user_headers);
1109 opt.user_headers = NULL;
1113 if (!check_user_specified_header (val))
1115 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1116 exec_name, com, val);
1119 opt.user_headers = vec_append (opt.user_headers, val);
1124 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1126 int flag = cmd_boolean (com, val, &opt.htmlify);
1127 if (flag && !opt.htmlify)
1128 opt.remove_listing = false;
1132 /* Set the "mirror" mode. It means: recursive download, timestamping,
1133 no limit on max. recursion depth, and don't remove listings. */
1136 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1140 if (!cmd_boolean (com, val, &mirror))
1144 opt.recursive = true;
1145 if (!opt.no_dirstruct)
1146 opt.dirstruct = true;
1147 opt.timestamping = true;
1148 opt.reclevel = INFINITE_RECURSION;
1149 opt.remove_listing = false;
1154 /* Validate --prefer-family and set the choice. Allowed values are
1155 "IPv4", "IPv6", and "none". */
1158 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1160 static const struct decode_item choices[] = {
1161 { "IPv4", prefer_ipv4 },
1162 { "IPv6", prefer_ipv6 },
1163 { "none", prefer_none },
1165 int prefer_family = prefer_ipv4;
1166 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1168 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1169 opt.prefer_family = prefer_family;
1173 /* Set progress.type to VAL, but verify that it's a valid progress
1174 implementation before that. */
1177 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1179 if (!valid_progress_implementation_p (val))
1181 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1182 exec_name, com, val);
1185 xfree_null (opt.progress_type);
1187 /* Don't call set_progress_implementation here. It will be called
1188 in main() when it becomes clear what the log output is. */
1189 opt.progress_type = xstrdup (val);
1193 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1194 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1198 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1200 if (!cmd_boolean (com, val, &opt.recursive))
1204 if (opt.recursive && !opt.no_dirstruct)
1205 opt.dirstruct = true;
1211 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1213 int restrict_os = opt.restrict_files_os;
1214 int restrict_ctrl = opt.restrict_files_ctrl;
1215 int restrict_case = opt.restrict_files_case;
1219 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1223 end = strchr (val, ',');
1225 end = val + strlen (val);
1227 if (VAL_IS ("unix"))
1228 restrict_os = restrict_unix;
1229 else if (VAL_IS ("windows"))
1230 restrict_os = restrict_windows;
1231 else if (VAL_IS ("lowercase"))
1232 restrict_case = restrict_lowercase;
1233 else if (VAL_IS ("uppercase"))
1234 restrict_case = restrict_uppercase;
1235 else if (VAL_IS ("nocontrol"))
1236 restrict_ctrl = false;
1240 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1241 exec_name, com, val);
1248 while (*val && *end);
1252 opt.restrict_files_os = restrict_os;
1253 opt.restrict_files_ctrl = restrict_ctrl;
1254 opt.restrict_files_case = restrict_case;
1261 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1263 static const struct decode_item choices[] = {
1264 { "auto", secure_protocol_auto },
1265 { "sslv2", secure_protocol_sslv2 },
1266 { "sslv3", secure_protocol_sslv3 },
1267 { "tlsv1", secure_protocol_tlsv1 },
1269 int ok = decode_string (val, choices, countof (choices), place);
1271 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1276 /* Set all three timeout values. */
1279 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1282 if (!cmd_time (com, val, &value))
1284 opt.read_timeout = value;
1285 opt.connect_timeout = value;
1286 opt.dns_timeout = value;
1291 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1293 /* Disallow embedded newlines. */
1294 if (strchr (val, '\n'))
1296 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1297 exec_name, com, val);
1300 xfree_null (opt.useragent);
1301 opt.useragent = xstrdup (val);
1305 /* The "verbose" option cannot be cmd_boolean because the variable is
1306 not bool -- it's of type int (-1 means uninitialized because of
1307 some random hackery for disallowing -q -v). */
1310 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1313 if (cmd_boolean (com, val, &flag))
1321 /* Miscellaneous useful routines. */
1323 /* A very simple atoi clone, more useful than atoi because it works on
1324 delimited strings, and has error reportage. Returns true on success,
1325 false on failure. If successful, stores result to *DEST. */
1328 simple_atoi (const char *beg, const char *end, int *dest)
1331 bool negative = false;
1332 const char *p = beg;
1334 while (p < end && c_isspace (*p))
1336 if (p < end && (*p == '-' || *p == '+'))
1338 negative = (*p == '-');
1344 /* Read negative numbers in a separate loop because the most
1345 negative integer cannot be represented as a positive number. */
1348 for (; p < end && c_isdigit (*p); p++)
1350 int next = (10 * result) + (*p - '0');
1352 return false; /* overflow */
1356 for (; p < end && c_isdigit (*p); p++)
1358 int next = (10 * result) - (*p - '0');
1360 return false; /* underflow */
1371 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1372 doesn't handle exponential notation. Returns true on success,
1373 false on failure. In case of success, stores its result to
1377 simple_atof (const char *beg, const char *end, double *dest)
1381 bool negative = false;
1382 bool seen_dot = false;
1383 bool seen_digit = false;
1386 const char *p = beg;
1388 while (p < end && c_isspace (*p))
1390 if (p < end && (*p == '-' || *p == '+'))
1392 negative = (*p == '-');
1396 for (; p < end; p++)
1402 result = (10 * result) + (ch - '0');
1404 result += (ch - '0') / (divider *= 10);
1426 /* Verify that the user-specified header in S is valid. It must
1427 contain a colon preceded by non-white-space characters and must not
1428 contain newlines. */
1431 check_user_specified_header (const char *s)
1435 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1437 /* The header MUST contain `:' preceded by at least one
1438 non-whitespace character. */
1439 if (*p != ':' || p == s)
1441 /* The header MUST NOT contain newlines. */
1442 if (strchr (s, '\n'))
1447 /* Decode VAL into a number, according to ITEMS. */
1450 decode_string (const char *val, const struct decode_item *items, int itemcount,
1454 for (i = 0; i < itemcount; i++)
1455 if (0 == strcasecmp (val, items[i].name))
1457 *place = items[i].code;
1464 void cleanup_html_url (void);
1467 /* Free the memory allocated by global variables. */
1471 /* Free external resources, close files, etc. */
1474 fclose (output_stream);
1475 /* No need to check for error because Wget flushes its output (and
1476 checks for errors) after any data arrives. */
1478 /* We're exiting anyway so there's no real need to call free()
1479 hundreds of times. Skipping the frees will make Wget exit
1482 However, when detecting leaks, it's crucial to free() everything
1483 because then you can find the real leaks, i.e. the allocated
1484 memory which grows with the size of the program. */
1490 cleanup_html_url ();
1495 extern acc_t *netrc_list;
1496 free_netrc (netrc_list);
1498 xfree_null (opt.lfilename);
1499 xfree_null (opt.dir_prefix);
1500 xfree_null (opt.input_filename);
1501 xfree_null (opt.output_document);
1502 free_vec (opt.accepts);
1503 free_vec (opt.rejects);
1504 free_vec (opt.excludes);
1505 free_vec (opt.includes);
1506 free_vec (opt.domains);
1507 free_vec (opt.follow_tags);
1508 free_vec (opt.ignore_tags);
1509 xfree_null (opt.progress_type);
1510 xfree_null (opt.ftp_user);
1511 xfree_null (opt.ftp_passwd);
1512 xfree_null (opt.ftp_proxy);
1513 xfree_null (opt.https_proxy);
1514 xfree_null (opt.http_proxy);
1515 free_vec (opt.no_proxy);
1516 xfree_null (opt.useragent);
1517 xfree_null (opt.referer);
1518 xfree_null (opt.http_user);
1519 xfree_null (opt.http_passwd);
1520 free_vec (opt.user_headers);
1522 xfree_null (opt.cert_file);
1523 xfree_null (opt.private_key);
1524 xfree_null (opt.ca_directory);
1525 xfree_null (opt.ca_cert);
1526 xfree_null (opt.random_file);
1527 xfree_null (opt.egd_file);
1529 xfree_null (opt.bind_address);
1530 xfree_null (opt.cookies_input);
1531 xfree_null (opt.cookies_output);
1532 xfree_null (opt.user);
1533 xfree_null (opt.passwd);
1534 #endif /* DEBUG_MALLOC */
1537 /* Unit testing routines. */
1542 test_cmd_spec_restrict_file_names()
1547 int expected_restrict_files_os;
1548 int expected_restrict_files_ctrl;
1549 int expected_restrict_files_case;
1552 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1553 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1554 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1555 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1558 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1563 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1566 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1567 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1568 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1569 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1571 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1572 res == test_array[i].result
1573 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1574 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1575 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1581 #endif /* TESTING */