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;
338 /* Return the user's home directory (strdup-ed), or NULL if none is
343 static char buf[PATH_MAX];
348 home = getenv ("HOME");
352 /* Under MSDOS, if $HOME isn't defined, use the directory where
353 `wget.exe' resides. */
354 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
357 strcpy (buf, _w32_get_argv0 ());
358 p = strrchr (buf, '/'); /* djgpp */
360 p = strrchr (buf, '\\'); /* others */
364 #elif !defined(WINDOWS)
365 /* If HOME is not defined, try getting it from the password
367 struct passwd *pwd = getpwuid (getuid ());
368 if (!pwd || !pwd->pw_dir)
370 strcpy (buf, pwd->pw_dir);
373 /* Under Windows, if $HOME isn't defined, use the directory where
374 `wget.exe' resides. */
380 return home ? xstrdup (home) : NULL;
383 /* Check the 'WGETRC' environment variable and return the file name
384 if 'WGETRC' is set and is a valid file.
385 If the `WGETRC' variable exists but the file does not exist, the
386 function will exit(). */
388 wgetrc_env_file_name (void)
390 char *env = getenv ("WGETRC");
393 if (!file_exists_p (env))
395 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
399 return xstrdup (env);
404 /* Check for the existance of '$HOME/.wgetrc' and return it's path
405 if it exists and is set. */
407 wgetrc_user_file_name (void)
409 char *home = home_dir ();
412 file = aprintf ("%s/.wgetrc", home);
416 if (!file_exists_p (file))
424 /* Return the path to the user's .wgetrc. This is either the value of
425 `WGETRC' environment variable, or `$HOME/.wgetrc'.
427 Additionally, for windows, look in the directory where wget.exe
430 wgetrc_file_name (void)
433 char *file = wgetrc_env_file_name ();
437 file = wgetrc_user_file_name ();
440 /* Under Windows, if we still haven't found .wgetrc, look for the file
441 `wget.ini' in the directory where `wget.exe' resides; we do this for
442 backward compatibility with previous versions of Wget.
443 SYSTEM_WGETRC should not be defined under WINDOWS. */
445 if (!file || !file_exists_p (file))
451 file = aprintf ("%s/wget.ini", home);
458 if (!file_exists_p (file))
466 /* Return values of parse_line. */
474 static enum parse_line parse_line (const char *, char **, char **, int *);
475 static bool setval_internal (int, const char *, const char *);
477 /* Initialize variables from a wgetrc file. Returns zero (failure) if
478 there were errors in the file. */
481 run_wgetrc (const char *file)
488 fp = fopen (file, "rb");
491 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
492 file, strerror (errno));
493 return true; /* not a fatal error */
495 enable_tilde_expansion = true;
497 while ((line = read_whole_line (fp)) != NULL)
499 char *com = NULL, *val = NULL;
502 /* Parse the line. */
503 switch (parse_line (line, &com, &val, &comind))
506 /* If everything is OK, set the value. */
507 if (!setval_internal (comind, com, val))
509 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
510 exec_name, file, ln);
514 case line_syntax_error:
515 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
516 exec_name, file, ln);
519 case line_unknown_command:
520 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
521 exec_name, quote (com), file, ln);
534 enable_tilde_expansion = false;
540 /* Initialize the defaults and run the system wgetrc and user's own
548 /* Load the hard-coded defaults. */
551 /* If SYSTEM_WGETRC is defined, use it. */
553 if (file_exists_p (SYSTEM_WGETRC))
554 ok &= run_wgetrc (SYSTEM_WGETRC);
556 /* Override it with your own, if one exists. */
557 file = wgetrc_file_name ();
560 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
561 something like realpath() before comparing them with `strcmp' */
563 if (!strcmp (file, SYSTEM_WGETRC))
565 fprintf (stderr, _("\
566 %s: Warning: Both system and user wgetrc point to %s.\n"),
567 exec_name, quote (file));
571 ok &= run_wgetrc (file);
573 /* If there were errors processing either `.wgetrc', abort. */
581 /* Remove dashes and underscores from S, modifying S in the
587 char *t = s; /* t - tortoise */
588 char *h = s; /* h - hare */
590 if (*h == '_' || *h == '-')
597 /* Parse the line pointed by line, with the syntax:
598 <sp>* command <sp>* = <sp>* value <sp>*
599 Uses malloc to allocate space for command and value.
601 Returns one of line_ok, line_empty, line_syntax_error, or
602 line_unknown_command.
604 In case of line_ok, *COM and *VAL point to freshly allocated
605 strings, and *COMIND points to com's index. In case of error or
606 empty line, their values are unmodified. */
608 static enum parse_line
609 parse_line (const char *line, char **com, char **val, int *comind)
612 const char *end = line + strlen (line);
613 const char *cmdstart, *cmdend;
614 const char *valstart, *valend;
619 /* Skip leading and trailing whitespace. */
620 while (*line && c_isspace (*line))
622 while (end > line && c_isspace (end[-1]))
625 /* Skip empty lines and comments. */
626 if (!*line || *line == '#')
632 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
636 /* Skip '=', as well as any space before or after it. */
637 while (p < end && c_isspace (*p))
639 if (p == end || *p != '=')
640 return line_syntax_error;
642 while (p < end && c_isspace (*p))
648 /* The syntax is valid (even though the command might not be). Fill
649 in the command name and value. */
650 *com = strdupdelim (cmdstart, cmdend);
651 *val = strdupdelim (valstart, valend);
653 /* The line now known to be syntactically correct. Check whether
654 the command is valid. */
655 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
657 ind = command_by_name (cmdcopy);
659 return line_unknown_command;
661 /* Report success to the caller. */
666 /* Run commands[comind].action. */
669 setval_internal (int comind, const char *com, const char *val)
671 assert (0 <= comind && ((size_t) comind) < countof (commands));
672 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
673 return commands[comind].action (com, val, commands[comind].place);
676 /* Run command COM with value VAL. If running the command produces an
677 error, report the error and exit.
679 This is intended to be called from main() to modify Wget's behavior
680 through command-line switches. Since COM is hard-coded in main(),
681 it is not canonicalized, and this aborts when COM is not found.
683 If COMIND's are exported to init.h, this function will be changed
684 to accept COMIND directly. */
687 setoptval (const char *com, const char *val, const char *optname)
689 /* Prepend "--" to OPTNAME. */
690 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
693 strcpy (dd_optname + 2, optname);
695 assert (val != NULL);
696 if (!setval_internal (command_by_name (com), dd_optname, val))
700 /* Parse OPT into command and value and run it. For example,
701 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
702 This is used by the `--execute' flag in main.c. */
705 run_command (const char *opt)
709 switch (parse_line (opt, &com, &val, &comind))
712 if (!setval_internal (comind, com, val))
718 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
719 exec_name, quote (opt));
724 /* Generic helper functions, for use with `commands'. */
726 /* Forward declarations: */
731 static bool decode_string (const char *, const struct decode_item *, int, int *);
732 static bool simple_atoi (const char *, const char *, int *);
733 static bool simple_atof (const char *, const char *, double *);
735 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
737 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
738 && c_tolower((p)[1]) == (c1) \
741 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
742 && c_tolower((p)[1]) == (c1) \
743 && c_tolower((p)[2]) == (c2) \
747 /* Store the boolean value from VAL to PLACE. COM is ignored,
748 except for error messages. */
750 cmd_boolean (const char *com, const char *val, void *place)
754 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
755 /* "on", "yes" and "1" mean true. */
757 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
758 /* "off", "no" and "0" mean false. */
763 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
764 exec_name, com, quote (val));
768 *(bool *) place = value;
772 /* Set the non-negative integer value from VAL to PLACE. With
773 incorrect specification, the number remains unchanged. */
775 cmd_number (const char *com, const char *val, void *place)
777 if (!simple_atoi (val, val + strlen (val), place)
778 || *(int *) place < 0)
780 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
781 exec_name, com, quote (val));
787 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
789 cmd_number_inf (const char *com, const char *val, void *place)
791 if (!strcasecmp (val, "inf"))
796 return cmd_number (com, val, place);
799 /* Copy (strdup) the string at COM to a new location and place a
800 pointer to *PLACE. */
802 cmd_string (const char *com, const char *val, void *place)
804 char **pstring = (char **)place;
806 xfree_null (*pstring);
807 *pstring = xstrdup (val);
811 #if defined(WINDOWS) || defined(MSDOS)
812 # define ISSEP(c) ((c) == '/' || (c) == '\\')
814 # define ISSEP(c) ((c) == '/')
817 /* Like the above, but handles tilde-expansion when reading a user's
818 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
819 gets expanded to the user's home directory. */
821 cmd_file (const char *com, const char *val, void *place)
823 char **pstring = (char **)place;
825 xfree_null (*pstring);
827 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
829 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
832 *pstring = xstrdup (val);
837 char *home = home_dir ();
841 homelen = strlen (home);
842 while (homelen && ISSEP (home[homelen - 1]))
843 home[--homelen] = '\0';
845 /* Skip the leading "~/". */
846 for (++val; ISSEP (*val); val++)
849 *pstring = concat_strings (home, "/", val, (char *) 0);
852 #if defined(WINDOWS) || defined(MSDOS)
853 /* Convert "\" to "/". */
856 for (s = *pstring; *s; s++)
864 /* Like cmd_file, but strips trailing '/' characters. */
866 cmd_directory (const char *com, const char *val, void *place)
870 /* Call cmd_file() for tilde expansion and separator
871 canonicalization (backslash -> slash under Windows). These
872 things should perhaps be in a separate function. */
873 if (!cmd_file (com, val, place))
878 while (t > s && *--t == '/')
884 /* Split VAL by space to a vector of values, and append those values
885 to vector pointed to by the PLACE argument. If VAL is empty, the
886 PLACE vector is cleared instead. */
889 cmd_vector (const char *com, const char *val, void *place)
891 char ***pvec = (char ***)place;
894 *pvec = merge_vecs (*pvec, sepstring (val));
904 cmd_directory_vector (const char *com, const char *val, void *place)
906 char ***pvec = (char ***)place;
910 /* Strip the trailing slashes from directories. */
913 seps = sepstring (val);
914 for (t = seps; t && *t; t++)
916 int len = strlen (*t);
917 /* Skip degenerate case of root directory. */
920 if ((*t)[len - 1] == '/')
921 (*t)[len - 1] = '\0';
924 *pvec = merge_vecs (*pvec, seps);
934 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
935 "100k" or "2.5G" to a floating point number. */
938 parse_bytes_helper (const char *val, double *result)
941 const char *end = val + strlen (val);
943 /* Check for "inf". */
944 if (0 == strcmp (val, "inf"))
950 /* Strip trailing whitespace. */
951 while (val < end && c_isspace (end[-1]))
956 switch (c_tolower (end[-1]))
959 --end, mult = 1024.0;
962 --end, mult = 1048576.0;
965 --end, mult = 1073741824.0;
968 --end, mult = 1099511627776.0;
971 /* Not a recognized suffix: assume it's a digit. (If not,
972 simple_atof will raise an error.) */
976 /* Skip leading and trailing whitespace. */
977 while (val < end && c_isspace (*val))
979 while (val < end && c_isspace (end[-1]))
984 if (!simple_atof (val, end, &number) || number < 0)
987 *result = number * mult;
991 /* Parse VAL as a number and set its value to PLACE (which should
994 By default, the value is assumed to be in bytes. If "K", "M", or
995 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
996 1<<30, respectively. Floating point values are allowed and are
997 cast to integer before use. The idea is to be able to use things
998 like 1.5k instead of "1536".
1000 The string "inf" is returned as 0.
1002 In case of error, false is returned and memory pointed to by PLACE
1003 remains unmodified. */
1006 cmd_bytes (const char *com, const char *val, void *place)
1009 if (!parse_bytes_helper (val, &byte_value))
1011 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1012 exec_name, com, quote (val));
1015 *(wgint *)place = (wgint)byte_value;
1019 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1020 SIZE_SUM. It works by converting the string to double, therefore
1021 working with values up to 2^53-1 without loss of precision. This
1022 value (8192 TB) is large enough to serve for a while. */
1025 cmd_bytes_sum (const char *com, const char *val, void *place)
1028 if (!parse_bytes_helper (val, &byte_value))
1030 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1031 exec_name, com, quote (val));
1034 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1038 /* Store the value of VAL to *OUT. The value is a time period, by
1039 default expressed in seconds, but also accepting suffixes "m", "h",
1040 "d", and "w" for minutes, hours, days, and weeks respectively. */
1043 cmd_time (const char *com, const char *val, void *place)
1045 double number, mult;
1046 const char *end = val + strlen (val);
1048 /* Strip trailing whitespace. */
1049 while (val < end && c_isspace (end[-1]))
1055 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1056 exec_name, com, quote (val));
1060 switch (c_tolower (end[-1]))
1063 --end, mult = 1; /* seconds */
1066 --end, mult = 60; /* minutes */
1069 --end, mult = 3600; /* hours */
1072 --end, mult = 86400.0; /* days */
1075 --end, mult = 604800.0; /* weeks */
1078 /* Not a recognized suffix: assume it belongs to the number.
1079 (If not, simple_atof will raise an error.) */
1083 /* Skip leading and trailing whitespace. */
1084 while (val < end && c_isspace (*val))
1086 while (val < end && c_isspace (end[-1]))
1091 if (!simple_atof (val, end, &number))
1094 *(double *)place = number * mult;
1100 cmd_cert_type (const char *com, const char *val, void *place)
1102 static const struct decode_item choices[] = {
1103 { "pem", keyfile_pem },
1104 { "der", keyfile_asn1 },
1105 { "asn1", keyfile_asn1 },
1107 int ok = decode_string (val, choices, countof (choices), place);
1109 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1114 /* Specialized helper functions, used by `commands' to handle some
1115 options specially. */
1117 static bool check_user_specified_header (const char *);
1120 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1122 if (!cmd_boolean (com, val, &opt.dirstruct))
1124 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1125 must be affected inversely. */
1127 opt.no_dirstruct = false;
1129 opt.no_dirstruct = true;
1134 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1136 /* Empty value means reset the list of headers. */
1139 free_vec (opt.user_headers);
1140 opt.user_headers = NULL;
1144 if (!check_user_specified_header (val))
1146 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1147 exec_name, com, quote (val));
1150 opt.user_headers = vec_append (opt.user_headers, val);
1155 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1157 int flag = cmd_boolean (com, val, &opt.htmlify);
1158 if (flag && !opt.htmlify)
1159 opt.remove_listing = false;
1163 /* Set the "mirror" mode. It means: recursive download, timestamping,
1164 no limit on max. recursion depth, and don't remove listings. */
1167 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1171 if (!cmd_boolean (com, val, &mirror))
1175 opt.recursive = true;
1176 if (!opt.no_dirstruct)
1177 opt.dirstruct = true;
1178 opt.timestamping = true;
1179 opt.reclevel = INFINITE_RECURSION;
1180 opt.remove_listing = false;
1185 /* Validate --prefer-family and set the choice. Allowed values are
1186 "IPv4", "IPv6", and "none". */
1189 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1191 static const struct decode_item choices[] = {
1192 { "IPv4", prefer_ipv4 },
1193 { "IPv6", prefer_ipv6 },
1194 { "none", prefer_none },
1196 int prefer_family = prefer_none;
1197 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1199 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1200 opt.prefer_family = prefer_family;
1204 /* Set progress.type to VAL, but verify that it's a valid progress
1205 implementation before that. */
1208 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1210 if (!valid_progress_implementation_p (val))
1212 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1213 exec_name, com, quote (val));
1216 xfree_null (opt.progress_type);
1218 /* Don't call set_progress_implementation here. It will be called
1219 in main() when it becomes clear what the log output is. */
1220 opt.progress_type = xstrdup (val);
1224 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1225 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1229 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1231 if (!cmd_boolean (com, val, &opt.recursive))
1235 if (opt.recursive && !opt.no_dirstruct)
1236 opt.dirstruct = true;
1242 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1244 int restrict_os = opt.restrict_files_os;
1245 int restrict_ctrl = opt.restrict_files_ctrl;
1246 int restrict_case = opt.restrict_files_case;
1250 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1254 end = strchr (val, ',');
1256 end = val + strlen (val);
1258 if (VAL_IS ("unix"))
1259 restrict_os = restrict_unix;
1260 else if (VAL_IS ("windows"))
1261 restrict_os = restrict_windows;
1262 else if (VAL_IS ("lowercase"))
1263 restrict_case = restrict_lowercase;
1264 else if (VAL_IS ("uppercase"))
1265 restrict_case = restrict_uppercase;
1266 else if (VAL_IS ("nocontrol"))
1267 restrict_ctrl = false;
1271 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1272 exec_name, com, quote (val));
1279 while (*val && *end);
1283 opt.restrict_files_os = restrict_os;
1284 opt.restrict_files_ctrl = restrict_ctrl;
1285 opt.restrict_files_case = restrict_case;
1292 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1294 static const struct decode_item choices[] = {
1295 { "auto", secure_protocol_auto },
1296 { "sslv2", secure_protocol_sslv2 },
1297 { "sslv3", secure_protocol_sslv3 },
1298 { "tlsv1", secure_protocol_tlsv1 },
1300 int ok = decode_string (val, choices, countof (choices), place);
1302 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1307 /* Set all three timeout values. */
1310 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1313 if (!cmd_time (com, val, &value))
1315 opt.read_timeout = value;
1316 opt.connect_timeout = value;
1317 opt.dns_timeout = value;
1322 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1324 /* Disallow embedded newlines. */
1325 if (strchr (val, '\n'))
1327 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1328 exec_name, com, quote (val));
1331 xfree_null (opt.useragent);
1332 opt.useragent = xstrdup (val);
1336 /* The "verbose" option cannot be cmd_boolean because the variable is
1337 not bool -- it's of type int (-1 means uninitialized because of
1338 some random hackery for disallowing -q -v). */
1341 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1344 if (cmd_boolean (com, val, &flag))
1352 /* Miscellaneous useful routines. */
1354 /* A very simple atoi clone, more useful than atoi because it works on
1355 delimited strings, and has error reportage. Returns true on success,
1356 false on failure. If successful, stores result to *DEST. */
1359 simple_atoi (const char *beg, const char *end, int *dest)
1362 bool negative = false;
1363 const char *p = beg;
1365 while (p < end && c_isspace (*p))
1367 if (p < end && (*p == '-' || *p == '+'))
1369 negative = (*p == '-');
1375 /* Read negative numbers in a separate loop because the most
1376 negative integer cannot be represented as a positive number. */
1379 for (; p < end && c_isdigit (*p); p++)
1381 int next = (10 * result) + (*p - '0');
1383 return false; /* overflow */
1387 for (; p < end && c_isdigit (*p); p++)
1389 int next = (10 * result) - (*p - '0');
1391 return false; /* underflow */
1402 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1403 doesn't handle exponential notation. Returns true on success,
1404 false on failure. In case of success, stores its result to
1408 simple_atof (const char *beg, const char *end, double *dest)
1412 bool negative = false;
1413 bool seen_dot = false;
1414 bool seen_digit = false;
1417 const char *p = beg;
1419 while (p < end && c_isspace (*p))
1421 if (p < end && (*p == '-' || *p == '+'))
1423 negative = (*p == '-');
1427 for (; p < end; p++)
1433 result = (10 * result) + (ch - '0');
1435 result += (ch - '0') / (divider *= 10);
1457 /* Verify that the user-specified header in S is valid. It must
1458 contain a colon preceded by non-white-space characters and must not
1459 contain newlines. */
1462 check_user_specified_header (const char *s)
1466 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1468 /* The header MUST contain `:' preceded by at least one
1469 non-whitespace character. */
1470 if (*p != ':' || p == s)
1472 /* The header MUST NOT contain newlines. */
1473 if (strchr (s, '\n'))
1478 /* Decode VAL into a number, according to ITEMS. */
1481 decode_string (const char *val, const struct decode_item *items, int itemcount,
1485 for (i = 0; i < itemcount; i++)
1486 if (0 == strcasecmp (val, items[i].name))
1488 *place = items[i].code;
1495 void cleanup_html_url (void);
1498 /* Free the memory allocated by global variables. */
1502 /* Free external resources, close files, etc. */
1505 fclose (output_stream);
1506 /* No need to check for error because Wget flushes its output (and
1507 checks for errors) after any data arrives. */
1509 /* We're exiting anyway so there's no real need to call free()
1510 hundreds of times. Skipping the frees will make Wget exit
1513 However, when detecting leaks, it's crucial to free() everything
1514 because then you can find the real leaks, i.e. the allocated
1515 memory which grows with the size of the program. */
1521 cleanup_html_url ();
1526 extern acc_t *netrc_list;
1527 free_netrc (netrc_list);
1529 xfree_null (opt.lfilename);
1530 xfree_null (opt.dir_prefix);
1531 xfree_null (opt.input_filename);
1532 xfree_null (opt.output_document);
1533 free_vec (opt.accepts);
1534 free_vec (opt.rejects);
1535 free_vec (opt.excludes);
1536 free_vec (opt.includes);
1537 free_vec (opt.domains);
1538 free_vec (opt.follow_tags);
1539 free_vec (opt.ignore_tags);
1540 xfree_null (opt.progress_type);
1541 xfree_null (opt.ftp_user);
1542 xfree_null (opt.ftp_passwd);
1543 xfree_null (opt.ftp_proxy);
1544 xfree_null (opt.https_proxy);
1545 xfree_null (opt.http_proxy);
1546 free_vec (opt.no_proxy);
1547 xfree_null (opt.useragent);
1548 xfree_null (opt.referer);
1549 xfree_null (opt.http_user);
1550 xfree_null (opt.http_passwd);
1551 free_vec (opt.user_headers);
1553 xfree_null (opt.cert_file);
1554 xfree_null (opt.private_key);
1555 xfree_null (opt.ca_directory);
1556 xfree_null (opt.ca_cert);
1557 xfree_null (opt.random_file);
1558 xfree_null (opt.egd_file);
1560 xfree_null (opt.bind_address);
1561 xfree_null (opt.cookies_input);
1562 xfree_null (opt.cookies_output);
1563 xfree_null (opt.user);
1564 xfree_null (opt.passwd);
1565 xfree_null (opt.base_href);
1567 #endif /* DEBUG_MALLOC */
1570 /* Unit testing routines. */
1575 test_commands_sorted()
1577 int prev_idx = 0, next_idx = 1;
1578 int command_count = countof (commands) - 1;
1580 while (next_idx <= command_count)
1582 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1585 mu_assert ("FAILED", false);
1598 test_cmd_spec_restrict_file_names()
1603 int expected_restrict_files_os;
1604 int expected_restrict_files_ctrl;
1605 int expected_restrict_files_case;
1608 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1609 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1610 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1611 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1614 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1619 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1622 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1623 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1624 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1625 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1627 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1628 res == test_array[i].result
1629 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1630 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1631 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1637 #endif /* TESTING */