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 { "deleteafter", &opt.delete_after, cmd_boolean },
144 { "dirprefix", &opt.dir_prefix, cmd_directory },
145 { "dirstruct", NULL, cmd_spec_dirstruct },
146 { "dnscache", &opt.dns_cache, cmd_boolean },
147 { "dnstimeout", &opt.dns_timeout, cmd_time },
148 { "domains", &opt.domains, cmd_vector },
149 { "dotbytes", &opt.dot_bytes, cmd_bytes },
150 { "dotsinline", &opt.dots_in_line, cmd_number },
151 { "dotspacing", &opt.dot_spacing, cmd_number },
152 { "dotstyle", &opt.dot_style, cmd_string },
154 { "egdfile", &opt.egd_file, cmd_file },
156 { "excludedirectories", &opt.excludes, cmd_directory_vector },
157 { "excludedomains", &opt.exclude_domains, cmd_vector },
158 { "followftp", &opt.follow_ftp, cmd_boolean },
159 { "followtags", &opt.follow_tags, cmd_vector },
160 { "forcehtml", &opt.force_html, cmd_boolean },
161 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
162 { "ftppassword", &opt.ftp_passwd, cmd_string },
163 { "ftpproxy", &opt.ftp_proxy, cmd_string },
164 { "ftpuser", &opt.ftp_user, cmd_string },
165 { "glob", &opt.ftp_glob, cmd_boolean },
166 { "header", NULL, cmd_spec_header },
167 { "htmlextension", &opt.html_extension, cmd_boolean },
168 { "htmlify", NULL, cmd_spec_htmlify },
169 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
170 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
171 { "httppassword", &opt.http_passwd, cmd_string },
172 { "httpproxy", &opt.http_proxy, cmd_string },
173 { "httpsproxy", &opt.https_proxy, cmd_string },
174 { "httpuser", &opt.http_user, cmd_string },
175 { "ignorecase", &opt.ignore_case, cmd_boolean },
176 { "ignorelength", &opt.ignore_length, cmd_boolean },
177 { "ignoretags", &opt.ignore_tags, cmd_vector },
178 { "includedirectories", &opt.includes, cmd_directory_vector },
180 { "inet4only", &opt.ipv4_only, cmd_boolean },
181 { "inet6only", &opt.ipv6_only, cmd_boolean },
183 { "input", &opt.input_filename, cmd_file },
184 { "iri", &opt.enable_iri, cmd_boolean },
185 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186 { "limitrate", &opt.limit_rate, cmd_bytes },
187 { "loadcookies", &opt.cookies_input, cmd_file },
188 { "locale", &opt.locale, cmd_string },
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 { "remoteencoding", &opt.encoding_remote, cmd_string },
229 { "removelisting", &opt.remove_listing, cmd_boolean },
230 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
231 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
232 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
233 { "robots", &opt.use_robots, cmd_boolean },
234 { "savecookies", &opt.cookies_output, cmd_file },
235 { "saveheaders", &opt.save_headers, cmd_boolean },
237 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
239 { "serverresponse", &opt.server_response, cmd_boolean },
240 { "spanhosts", &opt.spanhost, cmd_boolean },
241 { "spider", &opt.spider, cmd_boolean },
242 { "strictcomments", &opt.strict_comments, cmd_boolean },
243 { "timeout", NULL, cmd_spec_timeout },
244 { "timestamping", &opt.timestamping, cmd_boolean },
245 { "tries", &opt.ntry, cmd_number_inf },
246 { "useproxy", &opt.use_proxy, cmd_boolean },
247 { "user", &opt.user, cmd_string },
248 { "useragent", NULL, cmd_spec_useragent },
249 { "verbose", NULL, cmd_spec_verbose },
250 { "wait", &opt.wait, cmd_time },
251 { "waitretry", &opt.waitretry, cmd_time },
253 { "wdebug", &opt.wdebug, cmd_boolean },
257 /* Look up CMDNAME in the commands[] and return its position in the
258 array. If CMDNAME is not found, return -1. */
261 command_by_name (const char *cmdname)
263 /* Use binary search for speed. Wget has ~100 commands, which
264 guarantees a worst case performance of 7 string comparisons. */
265 int lo = 0, hi = countof (commands) - 1;
269 int mid = (lo + hi) >> 1;
270 int cmp = strcasecmp (cmdname, commands[mid].name);
281 /* Reset the variables to default values. */
287 /* Most of the default values are 0 (and 0.0, NULL, and false).
288 Just reset everything, and fill in the non-zero values. Note
289 that initializing pointers to NULL this way is technically
290 illegal, but porting Wget to a machine where NULL is not all-zero
291 bit pattern will be the least of the implementors' worries. */
298 opt.add_hostdir = true;
302 opt.http_keep_alive = true;
303 opt.use_proxy = true;
304 tmp = getenv ("no_proxy");
306 opt.no_proxy = sepstring (tmp);
307 opt.prefer_family = prefer_none;
308 opt.allow_cache = true;
310 opt.read_timeout = 900;
311 opt.use_robots = true;
313 opt.remove_listing = true;
315 opt.dot_bytes = 1024;
316 opt.dot_spacing = 10;
317 opt.dots_in_line = 50;
319 opt.dns_cache = true;
323 opt.check_cert = true;
326 /* The default for file name restriction defaults to the OS type. */
327 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
328 opt.restrict_files_os = restrict_windows;
330 opt.restrict_files_os = restrict_unix;
332 opt.restrict_files_ctrl = true;
333 opt.restrict_files_case = restrict_no_case_restriction;
335 opt.max_redirect = 20;
338 opt.enable_iri = true;
340 opt.enable_iri = false;
343 opt.encoding_remote = NULL;
346 /* Return the user's home directory (strdup-ed), or NULL if none is
351 char *home = getenv ("HOME");
356 /* Under MSDOS, if $HOME isn't defined, use the directory where
357 `wget.exe' resides. */
358 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
359 char *p, buf[PATH_MAX];
361 strcpy (buf, _w32_get_argv0 ());
362 p = strrchr (buf, '/'); /* djgpp */
364 p = strrchr (buf, '\\'); /* others */
368 #elif !defined(WINDOWS)
369 /* If HOME is not defined, try getting it from the password
371 struct passwd *pwd = getpwuid (getuid ());
372 if (!pwd || !pwd->pw_dir)
376 /* Under Windows, if $HOME isn't defined, use the directory where
377 `wget.exe' resides. */
382 return home ? xstrdup (home) : NULL;
385 /* Check the 'WGETRC' environment variable and return the file name
386 if 'WGETRC' is set and is a valid file.
387 If the `WGETRC' variable exists but the file does not exist, the
388 function will exit(). */
390 wgetrc_env_file_name (void)
392 char *env = getenv ("WGETRC");
395 if (!file_exists_p (env))
397 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
401 return xstrdup (env);
405 /* Check for the existance of '$HOME/.wgetrc' and return it's path
406 if it exists and is set. */
408 wgetrc_user_file_name (void)
410 char *home = home_dir();
413 file = aprintf ("%s/.wgetrc", home);
417 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)
432 char *file = wgetrc_env_file_name ();
436 file = wgetrc_user_file_name ();
439 /* Under Windows, if we still haven't found .wgetrc, look for the file
440 `wget.ini' in the directory where `wget.exe' resides; we do this for
441 backward compatibility with previous versions of Wget.
442 SYSTEM_WGETRC should not be defined under WINDOWS. */
443 if (!file || !file_exists_p (file))
449 file = aprintf ("%s/wget.ini", home);
455 if (!file_exists_p (file))
463 /* Return values of parse_line. */
471 static enum parse_line parse_line (const char *, char **, char **, int *);
472 static bool setval_internal (int, const char *, const char *);
474 /* Initialize variables from a wgetrc file. Returns zero (failure) if
475 there were errors in the file. */
478 run_wgetrc (const char *file)
485 fp = fopen (file, "rb");
488 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
489 file, strerror (errno));
490 return true; /* not a fatal error */
492 enable_tilde_expansion = true;
494 while ((line = read_whole_line (fp)) != NULL)
496 char *com = NULL, *val = NULL;
499 /* Parse the line. */
500 switch (parse_line (line, &com, &val, &comind))
503 /* If everything is OK, set the value. */
504 if (!setval_internal (comind, com, val))
506 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
507 exec_name, file, ln);
511 case line_syntax_error:
512 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
513 exec_name, file, ln);
516 case line_unknown_command:
517 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
518 exec_name, quote (com), file, ln);
531 enable_tilde_expansion = false;
537 /* Initialize the defaults and run the system wgetrc and user's own
545 /* Load the hard-coded defaults. */
548 /* If SYSTEM_WGETRC is defined, use it. */
550 if (file_exists_p (SYSTEM_WGETRC))
551 ok &= run_wgetrc (SYSTEM_WGETRC);
553 /* Override it with your own, if one exists. */
554 file = wgetrc_file_name ();
557 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
558 something like realpath() before comparing them with `strcmp' */
560 if (!strcmp (file, SYSTEM_WGETRC))
562 fprintf (stderr, _("\
563 %s: Warning: Both system and user wgetrc point to %s.\n"),
564 exec_name, quote (file));
568 ok &= run_wgetrc (file);
570 /* If there were errors processing either `.wgetrc', abort. */
578 /* Remove dashes and underscores from S, modifying S in the
584 char *t = s; /* t - tortoise */
585 char *h = s; /* h - hare */
587 if (*h == '_' || *h == '-')
594 /* Parse the line pointed by line, with the syntax:
595 <sp>* command <sp>* = <sp>* value <sp>*
596 Uses malloc to allocate space for command and value.
598 Returns one of line_ok, line_empty, line_syntax_error, or
599 line_unknown_command.
601 In case of line_ok, *COM and *VAL point to freshly allocated
602 strings, and *COMIND points to com's index. In case of error or
603 empty line, their values are unmodified. */
605 static enum parse_line
606 parse_line (const char *line, char **com, char **val, int *comind)
609 const char *end = line + strlen (line);
610 const char *cmdstart, *cmdend;
611 const char *valstart, *valend;
616 /* Skip leading and trailing whitespace. */
617 while (*line && c_isspace (*line))
619 while (end > line && c_isspace (end[-1]))
622 /* Skip empty lines and comments. */
623 if (!*line || *line == '#')
629 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
633 /* Skip '=', as well as any space before or after it. */
634 while (p < end && c_isspace (*p))
636 if (p == end || *p != '=')
637 return line_syntax_error;
639 while (p < end && c_isspace (*p))
645 /* The syntax is valid (even though the command might not be). Fill
646 in the command name and value. */
647 *com = strdupdelim (cmdstart, cmdend);
648 *val = strdupdelim (valstart, valend);
650 /* The line now known to be syntactically correct. Check whether
651 the command is valid. */
652 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
654 ind = command_by_name (cmdcopy);
656 return line_unknown_command;
658 /* Report success to the caller. */
663 /* Run commands[comind].action. */
666 setval_internal (int comind, const char *com, const char *val)
668 assert (0 <= comind && ((size_t) comind) < countof (commands));
669 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
670 return commands[comind].action (com, val, commands[comind].place);
673 /* Run command COM with value VAL. If running the command produces an
674 error, report the error and exit.
676 This is intended to be called from main() to modify Wget's behavior
677 through command-line switches. Since COM is hard-coded in main(),
678 it is not canonicalized, and this aborts when COM is not found.
680 If COMIND's are exported to init.h, this function will be changed
681 to accept COMIND directly. */
684 setoptval (const char *com, const char *val, const char *optname)
686 /* Prepend "--" to OPTNAME. */
687 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
690 strcpy (dd_optname + 2, optname);
692 assert (val != NULL);
693 if (!setval_internal (command_by_name (com), dd_optname, val))
697 /* Parse OPT into command and value and run it. For example,
698 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
699 This is used by the `--execute' flag in main.c. */
702 run_command (const char *opt)
706 switch (parse_line (opt, &com, &val, &comind))
709 if (!setval_internal (comind, com, val))
715 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
716 exec_name, quote (opt));
721 /* Generic helper functions, for use with `commands'. */
723 /* Forward declarations: */
728 static bool decode_string (const char *, const struct decode_item *, int, int *);
729 static bool simple_atoi (const char *, const char *, int *);
730 static bool simple_atof (const char *, const char *, double *);
732 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
734 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
735 && c_tolower((p)[1]) == (c1) \
738 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
739 && c_tolower((p)[1]) == (c1) \
740 && c_tolower((p)[2]) == (c2) \
744 /* Store the boolean value from VAL to PLACE. COM is ignored,
745 except for error messages. */
747 cmd_boolean (const char *com, const char *val, void *place)
751 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
752 /* "on", "yes" and "1" mean true. */
754 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
755 /* "off", "no" and "0" mean false. */
760 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
761 exec_name, com, quote (val));
765 *(bool *) place = value;
769 /* Set the non-negative integer value from VAL to PLACE. With
770 incorrect specification, the number remains unchanged. */
772 cmd_number (const char *com, const char *val, void *place)
774 if (!simple_atoi (val, val + strlen (val), place)
775 || *(int *) place < 0)
777 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
778 exec_name, com, quote (val));
784 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
786 cmd_number_inf (const char *com, const char *val, void *place)
788 if (!strcasecmp (val, "inf"))
793 return cmd_number (com, val, place);
796 /* Copy (strdup) the string at COM to a new location and place a
797 pointer to *PLACE. */
799 cmd_string (const char *com, const char *val, void *place)
801 char **pstring = (char **)place;
803 xfree_null (*pstring);
804 *pstring = xstrdup (val);
808 #if defined(WINDOWS) || defined(MSDOS)
809 # define ISSEP(c) ((c) == '/' || (c) == '\\')
811 # define ISSEP(c) ((c) == '/')
814 /* Like the above, but handles tilde-expansion when reading a user's
815 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
816 gets expanded to the user's home directory. */
818 cmd_file (const char *com, const char *val, void *place)
820 char **pstring = (char **)place;
822 xfree_null (*pstring);
824 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
826 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
829 *pstring = xstrdup (val);
834 char *home = home_dir ();
838 homelen = strlen (home);
839 while (homelen && ISSEP (home[homelen - 1]))
840 home[--homelen] = '\0';
842 /* Skip the leading "~/". */
843 for (++val; ISSEP (*val); val++)
846 *pstring = concat_strings (home, "/", val, (char *) 0);
849 #if defined(WINDOWS) || defined(MSDOS)
850 /* Convert "\" to "/". */
853 for (s = *pstring; *s; s++)
861 /* Like cmd_file, but strips trailing '/' characters. */
863 cmd_directory (const char *com, const char *val, void *place)
867 /* Call cmd_file() for tilde expansion and separator
868 canonicalization (backslash -> slash under Windows). These
869 things should perhaps be in a separate function. */
870 if (!cmd_file (com, val, place))
875 while (t > s && *--t == '/')
881 /* Split VAL by space to a vector of values, and append those values
882 to vector pointed to by the PLACE argument. If VAL is empty, the
883 PLACE vector is cleared instead. */
886 cmd_vector (const char *com, const char *val, void *place)
888 char ***pvec = (char ***)place;
891 *pvec = merge_vecs (*pvec, sepstring (val));
901 cmd_directory_vector (const char *com, const char *val, void *place)
903 char ***pvec = (char ***)place;
907 /* Strip the trailing slashes from directories. */
910 seps = sepstring (val);
911 for (t = seps; t && *t; t++)
913 int len = strlen (*t);
914 /* Skip degenerate case of root directory. */
917 if ((*t)[len - 1] == '/')
918 (*t)[len - 1] = '\0';
921 *pvec = merge_vecs (*pvec, seps);
931 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
932 "100k" or "2.5G" to a floating point number. */
935 parse_bytes_helper (const char *val, double *result)
938 const char *end = val + strlen (val);
940 /* Check for "inf". */
941 if (0 == strcmp (val, "inf"))
947 /* Strip trailing whitespace. */
948 while (val < end && c_isspace (end[-1]))
953 switch (c_tolower (end[-1]))
956 --end, mult = 1024.0;
959 --end, mult = 1048576.0;
962 --end, mult = 1073741824.0;
965 --end, mult = 1099511627776.0;
968 /* Not a recognized suffix: assume it's a digit. (If not,
969 simple_atof will raise an error.) */
973 /* Skip leading and trailing whitespace. */
974 while (val < end && c_isspace (*val))
976 while (val < end && c_isspace (end[-1]))
981 if (!simple_atof (val, end, &number) || number < 0)
984 *result = number * mult;
988 /* Parse VAL as a number and set its value to PLACE (which should
991 By default, the value is assumed to be in bytes. If "K", "M", or
992 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
993 1<<30, respectively. Floating point values are allowed and are
994 cast to integer before use. The idea is to be able to use things
995 like 1.5k instead of "1536".
997 The string "inf" is returned as 0.
999 In case of error, false is returned and memory pointed to by PLACE
1000 remains unmodified. */
1003 cmd_bytes (const char *com, const char *val, void *place)
1006 if (!parse_bytes_helper (val, &byte_value))
1008 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1009 exec_name, com, quote (val));
1012 *(wgint *)place = (wgint)byte_value;
1016 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
1017 SIZE_SUM. It works by converting the string to double, therefore
1018 working with values up to 2^53-1 without loss of precision. This
1019 value (8192 TB) is large enough to serve for a while. */
1022 cmd_bytes_sum (const char *com, const char *val, void *place)
1025 if (!parse_bytes_helper (val, &byte_value))
1027 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1028 exec_name, com, quote (val));
1031 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1035 /* Store the value of VAL to *OUT. The value is a time period, by
1036 default expressed in seconds, but also accepting suffixes "m", "h",
1037 "d", and "w" for minutes, hours, days, and weeks respectively. */
1040 cmd_time (const char *com, const char *val, void *place)
1042 double number, mult;
1043 const char *end = val + strlen (val);
1045 /* Strip trailing whitespace. */
1046 while (val < end && c_isspace (end[-1]))
1052 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1053 exec_name, com, quote (val));
1057 switch (c_tolower (end[-1]))
1060 --end, mult = 1; /* seconds */
1063 --end, mult = 60; /* minutes */
1066 --end, mult = 3600; /* hours */
1069 --end, mult = 86400.0; /* days */
1072 --end, mult = 604800.0; /* weeks */
1075 /* Not a recognized suffix: assume it belongs to the number.
1076 (If not, simple_atof will raise an error.) */
1080 /* Skip leading and trailing whitespace. */
1081 while (val < end && c_isspace (*val))
1083 while (val < end && c_isspace (end[-1]))
1088 if (!simple_atof (val, end, &number))
1091 *(double *)place = number * mult;
1097 cmd_cert_type (const char *com, const char *val, void *place)
1099 static const struct decode_item choices[] = {
1100 { "pem", keyfile_pem },
1101 { "der", keyfile_asn1 },
1102 { "asn1", keyfile_asn1 },
1104 int ok = decode_string (val, choices, countof (choices), place);
1106 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1111 /* Specialized helper functions, used by `commands' to handle some
1112 options specially. */
1114 static bool check_user_specified_header (const char *);
1117 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1119 if (!cmd_boolean (com, val, &opt.dirstruct))
1121 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1122 must be affected inversely. */
1124 opt.no_dirstruct = false;
1126 opt.no_dirstruct = true;
1131 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1133 /* Empty value means reset the list of headers. */
1136 free_vec (opt.user_headers);
1137 opt.user_headers = NULL;
1141 if (!check_user_specified_header (val))
1143 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1144 exec_name, com, quote (val));
1147 opt.user_headers = vec_append (opt.user_headers, val);
1152 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1154 int flag = cmd_boolean (com, val, &opt.htmlify);
1155 if (flag && !opt.htmlify)
1156 opt.remove_listing = false;
1160 /* Set the "mirror" mode. It means: recursive download, timestamping,
1161 no limit on max. recursion depth, and don't remove listings. */
1164 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1168 if (!cmd_boolean (com, val, &mirror))
1172 opt.recursive = true;
1173 if (!opt.no_dirstruct)
1174 opt.dirstruct = true;
1175 opt.timestamping = true;
1176 opt.reclevel = INFINITE_RECURSION;
1177 opt.remove_listing = false;
1182 /* Validate --prefer-family and set the choice. Allowed values are
1183 "IPv4", "IPv6", and "none". */
1186 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1188 static const struct decode_item choices[] = {
1189 { "IPv4", prefer_ipv4 },
1190 { "IPv6", prefer_ipv6 },
1191 { "none", prefer_none },
1193 int prefer_family = prefer_none;
1194 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1196 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1197 opt.prefer_family = prefer_family;
1201 /* Set progress.type to VAL, but verify that it's a valid progress
1202 implementation before that. */
1205 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1207 if (!valid_progress_implementation_p (val))
1209 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1210 exec_name, com, quote (val));
1213 xfree_null (opt.progress_type);
1215 /* Don't call set_progress_implementation here. It will be called
1216 in main() when it becomes clear what the log output is. */
1217 opt.progress_type = xstrdup (val);
1221 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1222 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1226 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1228 if (!cmd_boolean (com, val, &opt.recursive))
1232 if (opt.recursive && !opt.no_dirstruct)
1233 opt.dirstruct = true;
1239 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1241 int restrict_os = opt.restrict_files_os;
1242 int restrict_ctrl = opt.restrict_files_ctrl;
1243 int restrict_case = opt.restrict_files_case;
1247 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1251 end = strchr (val, ',');
1253 end = val + strlen (val);
1255 if (VAL_IS ("unix"))
1256 restrict_os = restrict_unix;
1257 else if (VAL_IS ("windows"))
1258 restrict_os = restrict_windows;
1259 else if (VAL_IS ("lowercase"))
1260 restrict_case = restrict_lowercase;
1261 else if (VAL_IS ("uppercase"))
1262 restrict_case = restrict_uppercase;
1263 else if (VAL_IS ("nocontrol"))
1264 restrict_ctrl = false;
1268 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1269 exec_name, com, quote (val));
1276 while (*val && *end);
1280 opt.restrict_files_os = restrict_os;
1281 opt.restrict_files_ctrl = restrict_ctrl;
1282 opt.restrict_files_case = restrict_case;
1289 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1291 static const struct decode_item choices[] = {
1292 { "auto", secure_protocol_auto },
1293 { "sslv2", secure_protocol_sslv2 },
1294 { "sslv3", secure_protocol_sslv3 },
1295 { "tlsv1", secure_protocol_tlsv1 },
1297 int ok = decode_string (val, choices, countof (choices), place);
1299 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1304 /* Set all three timeout values. */
1307 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1310 if (!cmd_time (com, val, &value))
1312 opt.read_timeout = value;
1313 opt.connect_timeout = value;
1314 opt.dns_timeout = value;
1319 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1321 /* Disallow embedded newlines. */
1322 if (strchr (val, '\n'))
1324 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1325 exec_name, com, quote (val));
1328 xfree_null (opt.useragent);
1329 opt.useragent = xstrdup (val);
1333 /* The "verbose" option cannot be cmd_boolean because the variable is
1334 not bool -- it's of type int (-1 means uninitialized because of
1335 some random hackery for disallowing -q -v). */
1338 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1341 if (cmd_boolean (com, val, &flag))
1349 /* Miscellaneous useful routines. */
1351 /* A very simple atoi clone, more useful than atoi because it works on
1352 delimited strings, and has error reportage. Returns true on success,
1353 false on failure. If successful, stores result to *DEST. */
1356 simple_atoi (const char *beg, const char *end, int *dest)
1359 bool negative = false;
1360 const char *p = beg;
1362 while (p < end && c_isspace (*p))
1364 if (p < end && (*p == '-' || *p == '+'))
1366 negative = (*p == '-');
1372 /* Read negative numbers in a separate loop because the most
1373 negative integer cannot be represented as a positive number. */
1376 for (; p < end && c_isdigit (*p); p++)
1378 int next = (10 * result) + (*p - '0');
1380 return false; /* overflow */
1384 for (; p < end && c_isdigit (*p); p++)
1386 int next = (10 * result) - (*p - '0');
1388 return false; /* underflow */
1399 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1400 doesn't handle exponential notation. Returns true on success,
1401 false on failure. In case of success, stores its result to
1405 simple_atof (const char *beg, const char *end, double *dest)
1409 bool negative = false;
1410 bool seen_dot = false;
1411 bool seen_digit = false;
1414 const char *p = beg;
1416 while (p < end && c_isspace (*p))
1418 if (p < end && (*p == '-' || *p == '+'))
1420 negative = (*p == '-');
1424 for (; p < end; p++)
1430 result = (10 * result) + (ch - '0');
1432 result += (ch - '0') / (divider *= 10);
1454 /* Verify that the user-specified header in S is valid. It must
1455 contain a colon preceded by non-white-space characters and must not
1456 contain newlines. */
1459 check_user_specified_header (const char *s)
1463 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1465 /* The header MUST contain `:' preceded by at least one
1466 non-whitespace character. */
1467 if (*p != ':' || p == s)
1469 /* The header MUST NOT contain newlines. */
1470 if (strchr (s, '\n'))
1475 /* Decode VAL into a number, according to ITEMS. */
1478 decode_string (const char *val, const struct decode_item *items, int itemcount,
1482 for (i = 0; i < itemcount; i++)
1483 if (0 == strcasecmp (val, items[i].name))
1485 *place = items[i].code;
1492 void cleanup_html_url (void);
1495 /* Free the memory allocated by global variables. */
1499 /* Free external resources, close files, etc. */
1502 fclose (output_stream);
1503 /* No need to check for error because Wget flushes its output (and
1504 checks for errors) after any data arrives. */
1506 /* We're exiting anyway so there's no real need to call free()
1507 hundreds of times. Skipping the frees will make Wget exit
1510 However, when detecting leaks, it's crucial to free() everything
1511 because then you can find the real leaks, i.e. the allocated
1512 memory which grows with the size of the program. */
1518 cleanup_html_url ();
1523 extern acc_t *netrc_list;
1524 free_netrc (netrc_list);
1526 xfree_null (opt.lfilename);
1527 xfree_null (opt.dir_prefix);
1528 xfree_null (opt.input_filename);
1529 xfree_null (opt.output_document);
1530 free_vec (opt.accepts);
1531 free_vec (opt.rejects);
1532 free_vec (opt.excludes);
1533 free_vec (opt.includes);
1534 free_vec (opt.domains);
1535 free_vec (opt.follow_tags);
1536 free_vec (opt.ignore_tags);
1537 xfree_null (opt.progress_type);
1538 xfree_null (opt.ftp_user);
1539 xfree_null (opt.ftp_passwd);
1540 xfree_null (opt.ftp_proxy);
1541 xfree_null (opt.https_proxy);
1542 xfree_null (opt.http_proxy);
1543 free_vec (opt.no_proxy);
1544 xfree_null (opt.useragent);
1545 xfree_null (opt.referer);
1546 xfree_null (opt.http_user);
1547 xfree_null (opt.http_passwd);
1548 free_vec (opt.user_headers);
1550 xfree_null (opt.cert_file);
1551 xfree_null (opt.private_key);
1552 xfree_null (opt.ca_directory);
1553 xfree_null (opt.ca_cert);
1554 xfree_null (opt.random_file);
1555 xfree_null (opt.egd_file);
1557 xfree_null (opt.bind_address);
1558 xfree_null (opt.cookies_input);
1559 xfree_null (opt.cookies_output);
1560 xfree_null (opt.user);
1561 xfree_null (opt.passwd);
1562 #endif /* DEBUG_MALLOC */
1565 /* Unit testing routines. */
1570 test_commands_sorted()
1572 int prev_idx = 0, next_idx = 1;
1573 int command_count = countof (commands) - 1;
1575 while (next_idx <= command_count)
1577 cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1580 mu_assert ("FAILED", false);
1593 test_cmd_spec_restrict_file_names()
1598 int expected_restrict_files_os;
1599 int expected_restrict_files_ctrl;
1600 int expected_restrict_files_case;
1603 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1604 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1605 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1606 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1609 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1614 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1617 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1618 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1619 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1620 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1622 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1623 res == test_array[i].result
1624 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1625 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1626 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1632 #endif /* TESTING */