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 { "askpassword", &opt.ask_passwd, cmd_boolean },
117 { "authnochallenge", &opt.auth_without_challenge,
119 { "background", &opt.background, cmd_boolean },
120 { "backupconverted", &opt.backup_converted, cmd_boolean },
121 { "backups", &opt.backups, cmd_number },
122 { "base", &opt.base_href, cmd_string },
123 { "bindaddress", &opt.bind_address, cmd_string },
125 { "cacertificate", &opt.ca_cert, cmd_file },
127 { "cache", &opt.allow_cache, cmd_boolean },
129 { "cadirectory", &opt.ca_directory, cmd_directory },
130 { "certificate", &opt.cert_file, cmd_file },
131 { "certificatetype", &opt.cert_type, cmd_cert_type },
132 { "checkcertificate", &opt.check_cert, cmd_boolean },
134 { "connecttimeout", &opt.connect_timeout, cmd_time },
135 { "contentdisposition", &opt.content_disposition, cmd_boolean },
136 { "continue", &opt.always_rest, cmd_boolean },
137 { "convertlinks", &opt.convert_links, cmd_boolean },
138 { "cookies", &opt.cookies, cmd_boolean },
139 { "cutdirs", &opt.cut_dirs, cmd_number },
141 { "debug", &opt.debug, cmd_boolean },
143 { "defaultpage", &opt.default_page, cmd_string},
144 { "deleteafter", &opt.delete_after, cmd_boolean },
145 { "dirprefix", &opt.dir_prefix, cmd_directory },
146 { "dirstruct", NULL, cmd_spec_dirstruct },
147 { "dnscache", &opt.dns_cache, cmd_boolean },
148 { "dnstimeout", &opt.dns_timeout, cmd_time },
149 { "domains", &opt.domains, cmd_vector },
150 { "dotbytes", &opt.dot_bytes, cmd_bytes },
151 { "dotsinline", &opt.dots_in_line, cmd_number },
152 { "dotspacing", &opt.dot_spacing, cmd_number },
153 { "dotstyle", &opt.dot_style, cmd_string },
155 { "egdfile", &opt.egd_file, cmd_file },
157 { "excludedirectories", &opt.excludes, cmd_directory_vector },
158 { "excludedomains", &opt.exclude_domains, cmd_vector },
159 { "followftp", &opt.follow_ftp, cmd_boolean },
160 { "followtags", &opt.follow_tags, cmd_vector },
161 { "forcehtml", &opt.force_html, cmd_boolean },
162 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
163 { "ftppassword", &opt.ftp_passwd, cmd_string },
164 { "ftpproxy", &opt.ftp_proxy, cmd_string },
165 { "ftpuser", &opt.ftp_user, cmd_string },
166 { "glob", &opt.ftp_glob, cmd_boolean },
167 { "header", NULL, cmd_spec_header },
168 { "htmlextension", &opt.html_extension, cmd_boolean },
169 { "htmlify", NULL, cmd_spec_htmlify },
170 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
171 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
172 { "httppassword", &opt.http_passwd, cmd_string },
173 { "httpproxy", &opt.http_proxy, cmd_string },
174 { "httpsproxy", &opt.https_proxy, cmd_string },
175 { "httpuser", &opt.http_user, cmd_string },
176 { "ignorecase", &opt.ignore_case, cmd_boolean },
177 { "ignorelength", &opt.ignore_length, cmd_boolean },
178 { "ignoretags", &opt.ignore_tags, cmd_vector },
179 { "includedirectories", &opt.includes, cmd_directory_vector },
181 { "inet4only", &opt.ipv4_only, cmd_boolean },
182 { "inet6only", &opt.ipv6_only, cmd_boolean },
184 { "input", &opt.input_filename, cmd_file },
185 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186 { "limitrate", &opt.limit_rate, cmd_bytes },
187 { "loadcookies", &opt.cookies_input, cmd_file },
188 { "logfile", &opt.lfilename, cmd_file },
189 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
190 { "maxredirect", &opt.max_redirect, cmd_number },
191 { "mirror", NULL, cmd_spec_mirror },
192 { "netrc", &opt.netrc, cmd_boolean },
193 { "noclobber", &opt.noclobber, cmd_boolean },
194 { "noparent", &opt.no_parent, cmd_boolean },
195 { "noproxy", &opt.no_proxy, cmd_vector },
196 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
197 { "outputdocument", &opt.output_document, cmd_file },
198 { "pagerequisites", &opt.page_requisites, cmd_boolean },
199 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
200 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
201 { "password", &opt.passwd, cmd_string },
202 { "postdata", &opt.post_data, cmd_string },
203 { "postfile", &opt.post_file_name, cmd_file },
204 { "preferfamily", NULL, cmd_spec_prefer_family },
205 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
207 { "privatekey", &opt.private_key, cmd_file },
208 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
210 { "progress", &opt.progress_type, cmd_spec_progress },
211 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
212 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
213 { "proxypassword", &opt.proxy_passwd, cmd_string },
214 { "proxyuser", &opt.proxy_user, cmd_string },
215 { "quiet", &opt.quiet, cmd_boolean },
216 { "quota", &opt.quota, cmd_bytes_sum },
218 { "randomfile", &opt.random_file, cmd_file },
220 { "randomwait", &opt.random_wait, cmd_boolean },
221 { "readtimeout", &opt.read_timeout, cmd_time },
222 { "reclevel", &opt.reclevel, cmd_number_inf },
223 { "recursive", NULL, cmd_spec_recursive },
224 { "referer", &opt.referer, cmd_string },
225 { "reject", &opt.rejects, cmd_vector },
226 { "relativeonly", &opt.relative_only, cmd_boolean },
227 { "removelisting", &opt.remove_listing, cmd_boolean },
228 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
229 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
230 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
231 { "robots", &opt.use_robots, cmd_boolean },
232 { "savecookies", &opt.cookies_output, cmd_file },
233 { "saveheaders", &opt.save_headers, cmd_boolean },
235 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
237 { "serverresponse", &opt.server_response, cmd_boolean },
238 { "spanhosts", &opt.spanhost, cmd_boolean },
239 { "spider", &opt.spider, cmd_boolean },
240 { "strictcomments", &opt.strict_comments, cmd_boolean },
241 { "timeout", NULL, cmd_spec_timeout },
242 { "timestamping", &opt.timestamping, cmd_boolean },
243 { "tries", &opt.ntry, cmd_number_inf },
244 { "useproxy", &opt.use_proxy, cmd_boolean },
245 { "user", &opt.user, cmd_string },
246 { "useragent", NULL, cmd_spec_useragent },
247 { "verbose", NULL, cmd_spec_verbose },
248 { "wait", &opt.wait, cmd_time },
249 { "waitretry", &opt.waitretry, cmd_time },
251 { "wdebug", &opt.wdebug, cmd_boolean },
255 /* Look up CMDNAME in the commands[] and return its position in the
256 array. If CMDNAME is not found, return -1. */
259 command_by_name (const char *cmdname)
261 /* Use binary search for speed. Wget has ~100 commands, which
262 guarantees a worst case performance of 7 string comparisons. */
263 int lo = 0, hi = countof (commands) - 1;
267 int mid = (lo + hi) >> 1;
268 int cmp = strcasecmp (cmdname, commands[mid].name);
279 /* Reset the variables to default values. */
285 /* Most of the default values are 0 (and 0.0, NULL, and false).
286 Just reset everything, and fill in the non-zero values. Note
287 that initializing pointers to NULL this way is technically
288 illegal, but porting Wget to a machine where NULL is not all-zero
289 bit pattern will be the least of the implementors' worries. */
296 opt.add_hostdir = true;
300 opt.http_keep_alive = true;
301 opt.use_proxy = true;
302 tmp = getenv ("no_proxy");
304 opt.no_proxy = sepstring (tmp);
305 opt.prefer_family = prefer_none;
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 $HOME isn't defined, use the directory where
367 `wget.exe' resides. */
372 return home ? xstrdup (home) : NULL;
375 /* Check the 'WGETRC' environment variable and return the file name
376 if 'WGETRC' is set and is a valid file.
377 If the `WGETRC' variable exists but the file does not exist, the
378 function will exit(). */
380 wgetrc_env_file_name (void)
382 char *env = getenv ("WGETRC");
385 if (!file_exists_p (env))
387 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
391 return xstrdup (env);
395 /* Check for the existance of '$HOME/.wgetrc' and return it's path
396 if it exists and is set. */
398 wgetrc_user_file_name (void)
400 char *home = home_dir();
403 file = aprintf ("%s/.wgetrc", home);
407 if (!file_exists_p (file))
414 /* Return the path to the user's .wgetrc. This is either the value of
415 `WGETRC' environment variable, or `$HOME/.wgetrc'.
417 Additionally, for windows, look in the directory where wget.exe
420 wgetrc_file_name (void)
422 char *file = wgetrc_env_file_name ();
426 file = wgetrc_user_file_name ();
429 /* Under Windows, if we still haven't found .wgetrc, look for the file
430 `wget.ini' in the directory where `wget.exe' resides; we do this for
431 backward compatibility with previous versions of Wget.
432 SYSTEM_WGETRC should not be defined under WINDOWS. */
433 if (!file || !file_exists_p (file))
439 file = aprintf ("%s/wget.ini", home);
445 if (!file_exists_p (file))
453 /* Return values of parse_line. */
461 static enum parse_line parse_line (const char *, char **, char **, int *);
462 static bool setval_internal (int, const char *, const char *);
464 /* Initialize variables from a wgetrc file. Returns zero (failure) if
465 there were errors in the file. */
468 run_wgetrc (const char *file)
475 fp = fopen (file, "rb");
478 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
479 file, strerror (errno));
480 return true; /* not a fatal error */
482 enable_tilde_expansion = true;
484 while ((line = read_whole_line (fp)) != NULL)
486 char *com = NULL, *val = NULL;
489 /* Parse the line. */
490 switch (parse_line (line, &com, &val, &comind))
493 /* If everything is OK, set the value. */
494 if (!setval_internal (comind, com, val))
496 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
497 exec_name, file, ln);
501 case line_syntax_error:
502 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
503 exec_name, file, ln);
506 case line_unknown_command:
507 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
508 exec_name, quote (com), file, ln);
521 enable_tilde_expansion = false;
527 /* Initialize the defaults and run the system wgetrc and user's own
535 /* Load the hard-coded defaults. */
538 /* If SYSTEM_WGETRC is defined, use it. */
540 if (file_exists_p (SYSTEM_WGETRC))
541 ok &= run_wgetrc (SYSTEM_WGETRC);
543 /* Override it with your own, if one exists. */
544 file = wgetrc_file_name ();
547 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
548 something like realpath() before comparing them with `strcmp' */
550 if (!strcmp (file, SYSTEM_WGETRC))
552 fprintf (stderr, _("\
553 %s: Warning: Both system and user wgetrc point to %s.\n"),
554 exec_name, quote (file));
558 ok &= run_wgetrc (file);
560 /* If there were errors processing either `.wgetrc', abort. */
568 /* Remove dashes and underscores from S, modifying S in the
574 char *t = s; /* t - tortoise */
575 char *h = s; /* h - hare */
577 if (*h == '_' || *h == '-')
584 /* Parse the line pointed by line, with the syntax:
585 <sp>* command <sp>* = <sp>* value <sp>*
586 Uses malloc to allocate space for command and value.
588 Returns one of line_ok, line_empty, line_syntax_error, or
589 line_unknown_command.
591 In case of line_ok, *COM and *VAL point to freshly allocated
592 strings, and *COMIND points to com's index. In case of error or
593 empty line, their values are unmodified. */
595 static enum parse_line
596 parse_line (const char *line, char **com, char **val, int *comind)
599 const char *end = line + strlen (line);
600 const char *cmdstart, *cmdend;
601 const char *valstart, *valend;
606 /* Skip leading and trailing whitespace. */
607 while (*line && c_isspace (*line))
609 while (end > line && c_isspace (end[-1]))
612 /* Skip empty lines and comments. */
613 if (!*line || *line == '#')
619 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
623 /* Skip '=', as well as any space before or after it. */
624 while (p < end && c_isspace (*p))
626 if (p == end || *p != '=')
627 return line_syntax_error;
629 while (p < end && c_isspace (*p))
635 /* The syntax is valid (even though the command might not be). Fill
636 in the command name and value. */
637 *com = strdupdelim (cmdstart, cmdend);
638 *val = strdupdelim (valstart, valend);
640 /* The line now known to be syntactically correct. Check whether
641 the command is valid. */
642 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
644 ind = command_by_name (cmdcopy);
646 return line_unknown_command;
648 /* Report success to the caller. */
653 /* Run commands[comind].action. */
656 setval_internal (int comind, const char *com, const char *val)
658 assert (0 <= comind && ((size_t) comind) < countof (commands));
659 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
660 return commands[comind].action (com, val, commands[comind].place);
663 /* Run command COM with value VAL. If running the command produces an
664 error, report the error and exit.
666 This is intended to be called from main() to modify Wget's behavior
667 through command-line switches. Since COM is hard-coded in main(),
668 it is not canonicalized, and this aborts when COM is not found.
670 If COMIND's are exported to init.h, this function will be changed
671 to accept COMIND directly. */
674 setoptval (const char *com, const char *val, const char *optname)
676 /* Prepend "--" to OPTNAME. */
677 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
680 strcpy (dd_optname + 2, optname);
682 assert (val != NULL);
683 if (!setval_internal (command_by_name (com), dd_optname, val))
687 /* Parse OPT into command and value and run it. For example,
688 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
689 This is used by the `--execute' flag in main.c. */
692 run_command (const char *opt)
696 switch (parse_line (opt, &com, &val, &comind))
699 if (!setval_internal (comind, com, val))
705 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
706 exec_name, quote (opt));
711 /* Generic helper functions, for use with `commands'. */
713 /* Forward declarations: */
718 static bool decode_string (const char *, const struct decode_item *, int, int *);
719 static bool simple_atoi (const char *, const char *, int *);
720 static bool simple_atof (const char *, const char *, double *);
722 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
724 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
725 && c_tolower((p)[1]) == (c1) \
728 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
729 && c_tolower((p)[1]) == (c1) \
730 && c_tolower((p)[2]) == (c2) \
734 /* Store the boolean value from VAL to PLACE. COM is ignored,
735 except for error messages. */
737 cmd_boolean (const char *com, const char *val, void *place)
741 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
742 /* "on", "yes" and "1" mean true. */
744 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
745 /* "off", "no" and "0" mean false. */
750 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
751 exec_name, com, quote (val));
755 *(bool *) place = value;
759 /* Set the non-negative integer value from VAL to PLACE. With
760 incorrect specification, the number remains unchanged. */
762 cmd_number (const char *com, const char *val, void *place)
764 if (!simple_atoi (val, val + strlen (val), place)
765 || *(int *) place < 0)
767 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
768 exec_name, com, quote (val));
774 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
776 cmd_number_inf (const char *com, const char *val, void *place)
778 if (!strcasecmp (val, "inf"))
783 return cmd_number (com, val, place);
786 /* Copy (strdup) the string at COM to a new location and place a
787 pointer to *PLACE. */
789 cmd_string (const char *com, const char *val, void *place)
791 char **pstring = (char **)place;
793 xfree_null (*pstring);
794 *pstring = xstrdup (val);
798 #if defined(WINDOWS) || defined(MSDOS)
799 # define ISSEP(c) ((c) == '/' || (c) == '\\')
801 # define ISSEP(c) ((c) == '/')
804 /* Like the above, but handles tilde-expansion when reading a user's
805 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
806 gets expanded to the user's home directory. */
808 cmd_file (const char *com, const char *val, void *place)
810 char **pstring = (char **)place;
812 xfree_null (*pstring);
814 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
816 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
819 *pstring = xstrdup (val);
824 char *home = home_dir ();
828 homelen = strlen (home);
829 while (homelen && ISSEP (home[homelen - 1]))
830 home[--homelen] = '\0';
832 /* Skip the leading "~/". */
833 for (++val; ISSEP (*val); val++)
836 *pstring = concat_strings (home, "/", val, (char *) 0);
839 #if defined(WINDOWS) || defined(MSDOS)
840 /* Convert "\" to "/". */
843 for (s = *pstring; *s; s++)
851 /* Like cmd_file, but strips trailing '/' characters. */
853 cmd_directory (const char *com, const char *val, void *place)
857 /* Call cmd_file() for tilde expansion and separator
858 canonicalization (backslash -> slash under Windows). These
859 things should perhaps be in a separate function. */
860 if (!cmd_file (com, val, place))
865 while (t > s && *--t == '/')
871 /* Split VAL by space to a vector of values, and append those values
872 to vector pointed to by the PLACE argument. If VAL is empty, the
873 PLACE vector is cleared instead. */
876 cmd_vector (const char *com, const char *val, void *place)
878 char ***pvec = (char ***)place;
881 *pvec = merge_vecs (*pvec, sepstring (val));
891 cmd_directory_vector (const char *com, const char *val, void *place)
893 char ***pvec = (char ***)place;
897 /* Strip the trailing slashes from directories. */
900 seps = sepstring (val);
901 for (t = seps; t && *t; t++)
903 int len = strlen (*t);
904 /* Skip degenerate case of root directory. */
907 if ((*t)[len - 1] == '/')
908 (*t)[len - 1] = '\0';
911 *pvec = merge_vecs (*pvec, seps);
921 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
922 "100k" or "2.5G" to a floating point number. */
925 parse_bytes_helper (const char *val, double *result)
928 const char *end = val + strlen (val);
930 /* Check for "inf". */
931 if (0 == strcmp (val, "inf"))
937 /* Strip trailing whitespace. */
938 while (val < end && c_isspace (end[-1]))
943 switch (c_tolower (end[-1]))
946 --end, mult = 1024.0;
949 --end, mult = 1048576.0;
952 --end, mult = 1073741824.0;
955 --end, mult = 1099511627776.0;
958 /* Not a recognized suffix: assume it's a digit. (If not,
959 simple_atof will raise an error.) */
963 /* Skip leading and trailing whitespace. */
964 while (val < end && c_isspace (*val))
966 while (val < end && c_isspace (end[-1]))
971 if (!simple_atof (val, end, &number) || number < 0)
974 *result = number * mult;
978 /* Parse VAL as a number and set its value to PLACE (which should
981 By default, the value is assumed to be in bytes. If "K", "M", or
982 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
983 1<<30, respectively. Floating point values are allowed and are
984 cast to integer before use. The idea is to be able to use things
985 like 1.5k instead of "1536".
987 The string "inf" is returned as 0.
989 In case of error, false is returned and memory pointed to by PLACE
990 remains unmodified. */
993 cmd_bytes (const char *com, const char *val, void *place)
996 if (!parse_bytes_helper (val, &byte_value))
998 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
999 exec_name, com, quote (val));
1002 *(wgint *)place = (wgint)byte_value;
1006 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1007 SIZE_SUM. It works by converting the string to double, therefore
1008 working with values up to 2^53-1 without loss of precision. This
1009 value (8192 TB) is large enough to serve for a while. */
1012 cmd_bytes_sum (const char *com, const char *val, void *place)
1015 if (!parse_bytes_helper (val, &byte_value))
1017 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1018 exec_name, com, quote (val));
1021 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1025 /* Store the value of VAL to *OUT. The value is a time period, by
1026 default expressed in seconds, but also accepting suffixes "m", "h",
1027 "d", and "w" for minutes, hours, days, and weeks respectively. */
1030 cmd_time (const char *com, const char *val, void *place)
1032 double number, mult;
1033 const char *end = val + strlen (val);
1035 /* Strip trailing whitespace. */
1036 while (val < end && c_isspace (end[-1]))
1042 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1043 exec_name, com, quote (val));
1047 switch (c_tolower (end[-1]))
1050 --end, mult = 1; /* seconds */
1053 --end, mult = 60; /* minutes */
1056 --end, mult = 3600; /* hours */
1059 --end, mult = 86400.0; /* days */
1062 --end, mult = 604800.0; /* weeks */
1065 /* Not a recognized suffix: assume it belongs to the number.
1066 (If not, simple_atof will raise an error.) */
1070 /* Skip leading and trailing whitespace. */
1071 while (val < end && c_isspace (*val))
1073 while (val < end && c_isspace (end[-1]))
1078 if (!simple_atof (val, end, &number))
1081 *(double *)place = number * mult;
1087 cmd_cert_type (const char *com, const char *val, void *place)
1089 static const struct decode_item choices[] = {
1090 { "pem", keyfile_pem },
1091 { "der", keyfile_asn1 },
1092 { "asn1", keyfile_asn1 },
1094 int ok = decode_string (val, choices, countof (choices), place);
1096 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1101 /* Specialized helper functions, used by `commands' to handle some
1102 options specially. */
1104 static bool check_user_specified_header (const char *);
1107 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1109 if (!cmd_boolean (com, val, &opt.dirstruct))
1111 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1112 must be affected inversely. */
1114 opt.no_dirstruct = false;
1116 opt.no_dirstruct = true;
1121 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1123 /* Empty value means reset the list of headers. */
1126 free_vec (opt.user_headers);
1127 opt.user_headers = NULL;
1131 if (!check_user_specified_header (val))
1133 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1134 exec_name, com, quote (val));
1137 opt.user_headers = vec_append (opt.user_headers, val);
1142 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1144 int flag = cmd_boolean (com, val, &opt.htmlify);
1145 if (flag && !opt.htmlify)
1146 opt.remove_listing = false;
1150 /* Set the "mirror" mode. It means: recursive download, timestamping,
1151 no limit on max. recursion depth, and don't remove listings. */
1154 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1158 if (!cmd_boolean (com, val, &mirror))
1162 opt.recursive = true;
1163 if (!opt.no_dirstruct)
1164 opt.dirstruct = true;
1165 opt.timestamping = true;
1166 opt.reclevel = INFINITE_RECURSION;
1167 opt.remove_listing = false;
1172 /* Validate --prefer-family and set the choice. Allowed values are
1173 "IPv4", "IPv6", and "none". */
1176 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1178 static const struct decode_item choices[] = {
1179 { "IPv4", prefer_ipv4 },
1180 { "IPv6", prefer_ipv6 },
1181 { "none", prefer_none },
1183 int prefer_family = prefer_none;
1184 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1186 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1187 opt.prefer_family = prefer_family;
1191 /* Set progress.type to VAL, but verify that it's a valid progress
1192 implementation before that. */
1195 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1197 if (!valid_progress_implementation_p (val))
1199 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1200 exec_name, com, quote (val));
1203 xfree_null (opt.progress_type);
1205 /* Don't call set_progress_implementation here. It will be called
1206 in main() when it becomes clear what the log output is. */
1207 opt.progress_type = xstrdup (val);
1211 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1212 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1216 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1218 if (!cmd_boolean (com, val, &opt.recursive))
1222 if (opt.recursive && !opt.no_dirstruct)
1223 opt.dirstruct = true;
1229 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1231 int restrict_os = opt.restrict_files_os;
1232 int restrict_ctrl = opt.restrict_files_ctrl;
1233 int restrict_case = opt.restrict_files_case;
1237 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1241 end = strchr (val, ',');
1243 end = val + strlen (val);
1245 if (VAL_IS ("unix"))
1246 restrict_os = restrict_unix;
1247 else if (VAL_IS ("windows"))
1248 restrict_os = restrict_windows;
1249 else if (VAL_IS ("lowercase"))
1250 restrict_case = restrict_lowercase;
1251 else if (VAL_IS ("uppercase"))
1252 restrict_case = restrict_uppercase;
1253 else if (VAL_IS ("nocontrol"))
1254 restrict_ctrl = false;
1258 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1259 exec_name, com, quote (val));
1266 while (*val && *end);
1270 opt.restrict_files_os = restrict_os;
1271 opt.restrict_files_ctrl = restrict_ctrl;
1272 opt.restrict_files_case = restrict_case;
1279 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1281 static const struct decode_item choices[] = {
1282 { "auto", secure_protocol_auto },
1283 { "sslv2", secure_protocol_sslv2 },
1284 { "sslv3", secure_protocol_sslv3 },
1285 { "tlsv1", secure_protocol_tlsv1 },
1287 int ok = decode_string (val, choices, countof (choices), place);
1289 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1294 /* Set all three timeout values. */
1297 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1300 if (!cmd_time (com, val, &value))
1302 opt.read_timeout = value;
1303 opt.connect_timeout = value;
1304 opt.dns_timeout = value;
1309 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1311 /* Disallow embedded newlines. */
1312 if (strchr (val, '\n'))
1314 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1315 exec_name, com, quote (val));
1318 xfree_null (opt.useragent);
1319 opt.useragent = xstrdup (val);
1323 /* The "verbose" option cannot be cmd_boolean because the variable is
1324 not bool -- it's of type int (-1 means uninitialized because of
1325 some random hackery for disallowing -q -v). */
1328 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1331 if (cmd_boolean (com, val, &flag))
1339 /* Miscellaneous useful routines. */
1341 /* A very simple atoi clone, more useful than atoi because it works on
1342 delimited strings, and has error reportage. Returns true on success,
1343 false on failure. If successful, stores result to *DEST. */
1346 simple_atoi (const char *beg, const char *end, int *dest)
1349 bool negative = false;
1350 const char *p = beg;
1352 while (p < end && c_isspace (*p))
1354 if (p < end && (*p == '-' || *p == '+'))
1356 negative = (*p == '-');
1362 /* Read negative numbers in a separate loop because the most
1363 negative integer cannot be represented as a positive number. */
1366 for (; p < end && c_isdigit (*p); p++)
1368 int next = (10 * result) + (*p - '0');
1370 return false; /* overflow */
1374 for (; p < end && c_isdigit (*p); p++)
1376 int next = (10 * result) - (*p - '0');
1378 return false; /* underflow */
1389 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1390 doesn't handle exponential notation. Returns true on success,
1391 false on failure. In case of success, stores its result to
1395 simple_atof (const char *beg, const char *end, double *dest)
1399 bool negative = false;
1400 bool seen_dot = false;
1401 bool seen_digit = false;
1404 const char *p = beg;
1406 while (p < end && c_isspace (*p))
1408 if (p < end && (*p == '-' || *p == '+'))
1410 negative = (*p == '-');
1414 for (; p < end; p++)
1420 result = (10 * result) + (ch - '0');
1422 result += (ch - '0') / (divider *= 10);
1444 /* Verify that the user-specified header in S is valid. It must
1445 contain a colon preceded by non-white-space characters and must not
1446 contain newlines. */
1449 check_user_specified_header (const char *s)
1453 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1455 /* The header MUST contain `:' preceded by at least one
1456 non-whitespace character. */
1457 if (*p != ':' || p == s)
1459 /* The header MUST NOT contain newlines. */
1460 if (strchr (s, '\n'))
1465 /* Decode VAL into a number, according to ITEMS. */
1468 decode_string (const char *val, const struct decode_item *items, int itemcount,
1472 for (i = 0; i < itemcount; i++)
1473 if (0 == strcasecmp (val, items[i].name))
1475 *place = items[i].code;
1482 void cleanup_html_url (void);
1485 /* Free the memory allocated by global variables. */
1489 /* Free external resources, close files, etc. */
1492 fclose (output_stream);
1493 /* No need to check for error because Wget flushes its output (and
1494 checks for errors) after any data arrives. */
1496 /* We're exiting anyway so there's no real need to call free()
1497 hundreds of times. Skipping the frees will make Wget exit
1500 However, when detecting leaks, it's crucial to free() everything
1501 because then you can find the real leaks, i.e. the allocated
1502 memory which grows with the size of the program. */
1508 cleanup_html_url ();
1513 extern acc_t *netrc_list;
1514 free_netrc (netrc_list);
1516 xfree_null (opt.lfilename);
1517 xfree_null (opt.dir_prefix);
1518 xfree_null (opt.input_filename);
1519 xfree_null (opt.output_document);
1520 free_vec (opt.accepts);
1521 free_vec (opt.rejects);
1522 free_vec (opt.excludes);
1523 free_vec (opt.includes);
1524 free_vec (opt.domains);
1525 free_vec (opt.follow_tags);
1526 free_vec (opt.ignore_tags);
1527 xfree_null (opt.progress_type);
1528 xfree_null (opt.ftp_user);
1529 xfree_null (opt.ftp_passwd);
1530 xfree_null (opt.ftp_proxy);
1531 xfree_null (opt.https_proxy);
1532 xfree_null (opt.http_proxy);
1533 free_vec (opt.no_proxy);
1534 xfree_null (opt.useragent);
1535 xfree_null (opt.referer);
1536 xfree_null (opt.http_user);
1537 xfree_null (opt.http_passwd);
1538 free_vec (opt.user_headers);
1540 xfree_null (opt.cert_file);
1541 xfree_null (opt.private_key);
1542 xfree_null (opt.ca_directory);
1543 xfree_null (opt.ca_cert);
1544 xfree_null (opt.random_file);
1545 xfree_null (opt.egd_file);
1547 xfree_null (opt.bind_address);
1548 xfree_null (opt.cookies_input);
1549 xfree_null (opt.cookies_output);
1550 xfree_null (opt.user);
1551 xfree_null (opt.passwd);
1552 xfree_null (opt.base_href);
1554 #endif /* DEBUG_MALLOC */
1557 /* Unit testing routines. */
1562 test_commands_sorted()
1564 int prev_idx = 0, next_idx = 1;
1565 int command_count = countof (commands) - 1;
1567 while (next_idx <= command_count)
1569 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1572 mu_assert ("FAILED", false);
1585 test_cmd_spec_restrict_file_names()
1590 int expected_restrict_files_os;
1591 int expected_restrict_files_ctrl;
1592 int expected_restrict_files_case;
1595 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1596 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1597 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1598 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1601 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1606 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1609 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1610 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1611 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1612 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1614 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1615 res == test_array[i].result
1616 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1617 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1618 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1624 #endif /* TESTING */