1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996-2006 Free Software Foundation, Inc.
4 This file is part of GNU Wget.
6 GNU Wget is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
11 GNU Wget is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with Wget; if not, write to the Free Software Foundation, Inc.,
18 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
20 In addition, as a special exception, the Free Software Foundation
21 gives permission to link the code of its release of Wget with the
22 OpenSSL project's "OpenSSL" library (or with modified versions of it
23 that use the same license as the "OpenSSL" library), and distribute
24 the linked executables. You must obey the GNU General Public License
25 in all respects for all of the code used other than "OpenSSL". If you
26 modify this file, you may extend this exception to your version of the
27 file, but you are not obligated to do so. If you do not wish to do
28 so, delete this exception statement from your version. */
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. */
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 { "background", &opt.background, cmd_boolean },
117 { "backupconverted", &opt.backup_converted, cmd_boolean },
118 { "backups", &opt.backups, cmd_number },
119 { "base", &opt.base_href, cmd_string },
120 { "bindaddress", &opt.bind_address, cmd_string },
122 { "cacertificate", &opt.ca_cert, cmd_file },
124 { "cache", &opt.allow_cache, cmd_boolean },
126 { "cadirectory", &opt.ca_directory, cmd_directory },
127 { "certificate", &opt.cert_file, cmd_file },
128 { "certificatetype", &opt.cert_type, cmd_cert_type },
129 { "checkcertificate", &opt.check_cert, cmd_boolean },
131 { "connecttimeout", &opt.connect_timeout, cmd_time },
132 { "contentdisposition", &opt.content_disposition, cmd_boolean },
133 { "continue", &opt.always_rest, cmd_boolean },
134 { "convertlinks", &opt.convert_links, cmd_boolean },
135 { "cookies", &opt.cookies, cmd_boolean },
136 { "cutdirs", &opt.cut_dirs, cmd_number },
138 { "debug", &opt.debug, cmd_boolean },
140 { "deleteafter", &opt.delete_after, cmd_boolean },
141 { "dirprefix", &opt.dir_prefix, cmd_directory },
142 { "dirstruct", NULL, cmd_spec_dirstruct },
143 { "dnscache", &opt.dns_cache, cmd_boolean },
144 { "dnstimeout", &opt.dns_timeout, cmd_time },
145 { "domains", &opt.domains, cmd_vector },
146 { "dotbytes", &opt.dot_bytes, cmd_bytes },
147 { "dotsinline", &opt.dots_in_line, cmd_number },
148 { "dotspacing", &opt.dot_spacing, cmd_number },
149 { "dotstyle", &opt.dot_style, cmd_string },
151 { "egdfile", &opt.egd_file, cmd_file },
153 { "excludedirectories", &opt.excludes, cmd_directory_vector },
154 { "excludedomains", &opt.exclude_domains, cmd_vector },
155 { "followftp", &opt.follow_ftp, cmd_boolean },
156 { "followtags", &opt.follow_tags, cmd_vector },
157 { "forcehtml", &opt.force_html, cmd_boolean },
158 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
159 { "ftppassword", &opt.ftp_passwd, cmd_string },
160 { "ftpproxy", &opt.ftp_proxy, cmd_string },
161 { "ftpuser", &opt.ftp_user, cmd_string },
162 { "glob", &opt.ftp_glob, cmd_boolean },
163 { "header", NULL, cmd_spec_header },
164 { "htmlextension", &opt.html_extension, cmd_boolean },
165 { "htmlify", NULL, cmd_spec_htmlify },
166 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
167 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
168 { "httppassword", &opt.http_passwd, cmd_string },
169 { "httpproxy", &opt.http_proxy, cmd_string },
170 { "httpsproxy", &opt.https_proxy, cmd_string },
171 { "httpuser", &opt.http_user, cmd_string },
172 { "ignorecase", &opt.ignore_case, cmd_boolean },
173 { "ignorelength", &opt.ignore_length, cmd_boolean },
174 { "ignoretags", &opt.ignore_tags, cmd_vector },
175 { "includedirectories", &opt.includes, cmd_directory_vector },
177 { "inet4only", &opt.ipv4_only, cmd_boolean },
178 { "inet6only", &opt.ipv6_only, cmd_boolean },
180 { "input", &opt.input_filename, cmd_file },
181 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
182 { "limitrate", &opt.limit_rate, cmd_bytes },
183 { "loadcookies", &opt.cookies_input, cmd_file },
184 { "logfile", &opt.lfilename, cmd_file },
185 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
186 { "mirror", NULL, cmd_spec_mirror },
187 { "netrc", &opt.netrc, cmd_boolean },
188 { "noclobber", &opt.noclobber, cmd_boolean },
189 { "noparent", &opt.no_parent, cmd_boolean },
190 { "noproxy", &opt.no_proxy, cmd_vector },
191 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
192 { "outputdocument", &opt.output_document, cmd_file },
193 { "pagerequisites", &opt.page_requisites, cmd_boolean },
194 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
195 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
196 { "password", &opt.passwd, cmd_string },
197 { "postdata", &opt.post_data, cmd_string },
198 { "postfile", &opt.post_file_name, cmd_file },
199 { "preferfamily", NULL, cmd_spec_prefer_family },
200 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
202 { "privatekey", &opt.private_key, cmd_file },
203 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
205 { "progress", &opt.progress_type, cmd_spec_progress },
206 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
207 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
208 { "proxypassword", &opt.proxy_passwd, cmd_string },
209 { "proxyuser", &opt.proxy_user, cmd_string },
210 { "quiet", &opt.quiet, cmd_boolean },
211 { "quota", &opt.quota, cmd_bytes_sum },
213 { "randomfile", &opt.random_file, cmd_file },
215 { "randomwait", &opt.random_wait, cmd_boolean },
216 { "readtimeout", &opt.read_timeout, cmd_time },
217 { "reclevel", &opt.reclevel, cmd_number_inf },
218 { "recursive", NULL, cmd_spec_recursive },
219 { "referer", &opt.referer, cmd_string },
220 { "reject", &opt.rejects, cmd_vector },
221 { "relativeonly", &opt.relative_only, cmd_boolean },
222 { "removelisting", &opt.remove_listing, cmd_boolean },
223 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
224 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
225 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
226 { "robots", &opt.use_robots, cmd_boolean },
227 { "savecookies", &opt.cookies_output, cmd_file },
228 { "saveheaders", &opt.save_headers, cmd_boolean },
230 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
232 { "serverresponse", &opt.server_response, cmd_boolean },
233 { "spanhosts", &opt.spanhost, cmd_boolean },
234 { "spider", &opt.spider, cmd_boolean },
235 { "strictcomments", &opt.strict_comments, cmd_boolean },
236 { "timeout", NULL, cmd_spec_timeout },
237 { "timestamping", &opt.timestamping, cmd_boolean },
238 { "tries", &opt.ntry, cmd_number_inf },
239 { "useproxy", &opt.use_proxy, cmd_boolean },
240 { "user", &opt.user, cmd_string },
241 { "useragent", NULL, cmd_spec_useragent },
242 { "verbose", NULL, cmd_spec_verbose },
243 { "wait", &opt.wait, cmd_time },
244 { "waitretry", &opt.waitretry, cmd_time }
247 /* Look up CMDNAME in the commands[] and return its position in the
248 array. If CMDNAME is not found, return -1. */
251 command_by_name (const char *cmdname)
253 /* Use binary search for speed. Wget has ~100 commands, which
254 guarantees a worst case performance of 7 string comparisons. */
255 int lo = 0, hi = countof (commands) - 1;
259 int mid = (lo + hi) >> 1;
260 int cmp = strcasecmp (cmdname, commands[mid].name);
271 /* Reset the variables to default values. */
277 /* Most of the default values are 0 (and 0.0, NULL, and false).
278 Just reset everything, and fill in the non-zero values. Note
279 that initializing pointers to NULL this way is technically
280 illegal, but porting Wget to a machine where NULL is not all-zero
281 bit pattern will be the least of the implementors' worries. */
288 opt.add_hostdir = true;
292 opt.http_keep_alive = true;
293 opt.use_proxy = true;
294 tmp = getenv ("no_proxy");
296 opt.no_proxy = sepstring (tmp);
297 opt.allow_cache = true;
299 opt.read_timeout = 900;
300 opt.use_robots = true;
302 opt.remove_listing = true;
304 opt.dot_bytes = 1024;
305 opt.dot_spacing = 10;
306 opt.dots_in_line = 50;
308 opt.dns_cache = true;
312 opt.check_cert = true;
315 /* The default for file name restriction defaults to the OS type. */
316 #if !defined(WINDOWS) && !defined(__CYGWIN__)
317 opt.restrict_files_os = restrict_unix;
319 opt.restrict_files_os = restrict_windows;
321 opt.restrict_files_ctrl = true;
322 opt.restrict_files_case = restrict_no_case_restriction;
324 opt.content_disposition = true;
327 /* Return the user's home directory (strdup-ed), or NULL if none is
332 char *home = getenv ("HOME");
337 /* If HOME is not defined, try getting it from the password
339 struct passwd *pwd = getpwuid (getuid ());
340 if (!pwd || !pwd->pw_dir)
344 /* Under Windows, if $HOME isn't defined, use the directory where
345 `wget.exe' resides. */
350 return home ? xstrdup (home) : NULL;
353 /* Return the path to the user's .wgetrc. This is either the value of
354 `WGETRC' environment variable, or `$HOME/.wgetrc'.
356 If the `WGETRC' variable exists but the file does not exist, the
357 function will exit(). */
359 wgetrc_file_name (void)
364 /* Try the environment. */
365 env = getenv ("WGETRC");
368 if (!file_exists_p (env))
370 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
374 return xstrdup (env);
377 /* If that failed, try $HOME/.wgetrc. */
380 file = aprintf ("%s/.wgetrc", home);
384 /* Under Windows, if we still haven't found .wgetrc, look for the file
385 `wget.ini' in the directory where `wget.exe' resides; we do this for
386 backward compatibility with previous versions of Wget.
387 SYSTEM_WGETRC should not be defined under WINDOWS. */
388 if (!file || !file_exists_p (file))
394 file = aprintf ("%s/wget.ini", home);
400 if (!file_exists_p (file))
408 /* Return values of parse_line. */
416 static enum parse_line parse_line (const char *, char **, char **, int *);
417 static bool setval_internal (int, const char *, const char *);
419 /* Initialize variables from a wgetrc file. Returns zero (failure) if
420 there were errors in the file. */
423 run_wgetrc (const char *file)
430 fp = fopen (file, "rb");
433 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
434 file, strerror (errno));
435 return true; /* not a fatal error */
437 enable_tilde_expansion = true;
439 while ((line = read_whole_line (fp)) != NULL)
441 char *com = NULL, *val = NULL;
444 /* Parse the line. */
445 switch (parse_line (line, &com, &val, &comind))
448 /* If everything is OK, set the value. */
449 if (!setval_internal (comind, com, val))
451 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
452 exec_name, file, ln);
456 case line_syntax_error:
457 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
458 exec_name, file, ln);
461 case line_unknown_command:
462 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
463 exec_name, com, file, ln);
476 enable_tilde_expansion = false;
482 /* Initialize the defaults and run the system wgetrc and user's own
490 /* Load the hard-coded defaults. */
493 /* If SYSTEM_WGETRC is defined, use it. */
495 if (file_exists_p (SYSTEM_WGETRC))
496 ok &= run_wgetrc (SYSTEM_WGETRC);
498 /* Override it with your own, if one exists. */
499 file = wgetrc_file_name ();
502 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
503 something like realpath() before comparing them with `strcmp' */
505 if (!strcmp (file, SYSTEM_WGETRC))
507 fprintf (stderr, _("\
508 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
513 ok &= run_wgetrc (file);
515 /* If there were errors processing either `.wgetrc', abort. */
523 /* Remove dashes and underscores from S, modifying S in the
529 char *t = s; /* t - tortoise */
530 char *h = s; /* h - hare */
532 if (*h == '_' || *h == '-')
539 /* Parse the line pointed by line, with the syntax:
540 <sp>* command <sp>* = <sp>* value <sp>*
541 Uses malloc to allocate space for command and value.
543 Returns one of line_ok, line_empty, line_syntax_error, or
544 line_unknown_command.
546 In case of line_ok, *COM and *VAL point to freshly allocated
547 strings, and *COMIND points to com's index. In case of error or
548 empty line, their values are unmodified. */
550 static enum parse_line
551 parse_line (const char *line, char **com, char **val, int *comind)
554 const char *end = line + strlen (line);
555 const char *cmdstart, *cmdend;
556 const char *valstart, *valend;
561 /* Skip leading and trailing whitespace. */
562 while (*line && ISSPACE (*line))
564 while (end > line && ISSPACE (end[-1]))
567 /* Skip empty lines and comments. */
568 if (!*line || *line == '#')
574 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
578 /* Skip '=', as well as any space before or after it. */
579 while (p < end && ISSPACE (*p))
581 if (p == end || *p != '=')
582 return line_syntax_error;
584 while (p < end && ISSPACE (*p))
590 /* The syntax is valid (even though the command might not be). Fill
591 in the command name and value. */
592 *com = strdupdelim (cmdstart, cmdend);
593 *val = strdupdelim (valstart, valend);
595 /* The line now known to be syntactically correct. Check whether
596 the command is valid. */
597 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
599 ind = command_by_name (cmdcopy);
601 return line_unknown_command;
603 /* Report success to the caller. */
608 /* Run commands[comind].action. */
611 setval_internal (int comind, const char *com, const char *val)
613 assert (0 <= comind && comind < countof (commands));
614 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
615 return commands[comind].action (com, val, commands[comind].place);
618 /* Run command COM with value VAL. If running the command produces an
619 error, report the error and exit.
621 This is intended to be called from main() to modify Wget's behavior
622 through command-line switches. Since COM is hard-coded in main(),
623 it is not canonicalized, and this aborts when COM is not found.
625 If COMIND's are exported to init.h, this function will be changed
626 to accept COMIND directly. */
629 setoptval (const char *com, const char *val, const char *optname)
631 /* Prepend "--" to OPTNAME. */
632 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
635 strcpy (dd_optname + 2, optname);
637 assert (val != NULL);
638 if (!setval_internal (command_by_name (com), dd_optname, val))
642 /* Parse OPT into command and value and run it. For example,
643 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
644 This is used by the `--execute' flag in main.c. */
647 run_command (const char *opt)
651 switch (parse_line (opt, &com, &val, &comind))
654 if (!setval_internal (comind, com, val))
660 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
666 /* Generic helper functions, for use with `commands'. */
668 /* Forward declarations: */
673 static bool decode_string (const char *, const struct decode_item *, int, int *);
674 static bool simple_atoi (const char *, const char *, int *);
675 static bool simple_atof (const char *, const char *, double *);
677 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
679 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
680 && TOLOWER((p)[1]) == (c1) \
683 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
684 && TOLOWER((p)[1]) == (c1) \
685 && TOLOWER((p)[2]) == (c2) \
689 /* Store the boolean value from VAL to PLACE. COM is ignored,
690 except for error messages. */
692 cmd_boolean (const char *com, const char *val, void *place)
696 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
697 /* "on", "yes" and "1" mean true. */
699 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
700 /* "off", "no" and "0" mean false. */
705 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
706 exec_name, com, val);
710 *(bool *) place = value;
714 /* Set the non-negative integer value from VAL to PLACE. With
715 incorrect specification, the number remains unchanged. */
717 cmd_number (const char *com, const char *val, void *place)
719 if (!simple_atoi (val, val + strlen (val), place)
720 || *(int *) place < 0)
722 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
723 exec_name, com, val);
729 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
731 cmd_number_inf (const char *com, const char *val, void *place)
733 if (!strcasecmp (val, "inf"))
738 return cmd_number (com, val, place);
741 /* Copy (strdup) the string at COM to a new location and place a
742 pointer to *PLACE. */
744 cmd_string (const char *com, const char *val, void *place)
746 char **pstring = (char **)place;
748 xfree_null (*pstring);
749 *pstring = xstrdup (val);
754 # define ISSEP(c) ((c) == '/')
756 # define ISSEP(c) ((c) == '/' || (c) == '\\')
759 /* Like the above, but handles tilde-expansion when reading a user's
760 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
761 gets expanded to the user's home directory. */
763 cmd_file (const char *com, const char *val, void *place)
765 char **pstring = (char **)place;
767 xfree_null (*pstring);
769 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
771 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
774 *pstring = xstrdup (val);
779 char *home = home_dir ();
783 homelen = strlen (home);
784 while (homelen && ISSEP (home[homelen - 1]))
785 home[--homelen] = '\0';
787 /* Skip the leading "~/". */
788 for (++val; ISSEP (*val); val++)
791 *pstring = concat_strings (home, "/", val, (char *) 0);
795 /* Convert "\" to "/". */
798 for (s = *pstring; *s; s++)
806 /* Like cmd_file, but strips trailing '/' characters. */
808 cmd_directory (const char *com, const char *val, void *place)
812 /* Call cmd_file() for tilde expansion and separator
813 canonicalization (backslash -> slash under Windows). These
814 things should perhaps be in a separate function. */
815 if (!cmd_file (com, val, place))
820 while (t > s && *--t == '/')
826 /* Split VAL by space to a vector of values, and append those values
827 to vector pointed to by the PLACE argument. If VAL is empty, the
828 PLACE vector is cleared instead. */
831 cmd_vector (const char *com, const char *val, void *place)
833 char ***pvec = (char ***)place;
836 *pvec = merge_vecs (*pvec, sepstring (val));
846 cmd_directory_vector (const char *com, const char *val, void *place)
848 char ***pvec = (char ***)place;
852 /* Strip the trailing slashes from directories. */
855 seps = sepstring (val);
856 for (t = seps; t && *t; t++)
858 int len = strlen (*t);
859 /* Skip degenerate case of root directory. */
862 if ((*t)[len - 1] == '/')
863 (*t)[len - 1] = '\0';
866 *pvec = merge_vecs (*pvec, seps);
876 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
877 "100k" or "2.5G" to a floating point number. */
880 parse_bytes_helper (const char *val, double *result)
883 const char *end = val + strlen (val);
885 /* Check for "inf". */
886 if (0 == strcmp (val, "inf"))
892 /* Strip trailing whitespace. */
893 while (val < end && ISSPACE (end[-1]))
898 switch (TOLOWER (end[-1]))
901 --end, mult = 1024.0;
904 --end, mult = 1048576.0;
907 --end, mult = 1073741824.0;
910 --end, mult = 1099511627776.0;
913 /* Not a recognized suffix: assume it's a digit. (If not,
914 simple_atof will raise an error.) */
918 /* Skip leading and trailing whitespace. */
919 while (val < end && ISSPACE (*val))
921 while (val < end && ISSPACE (end[-1]))
926 if (!simple_atof (val, end, &number) || number < 0)
929 *result = number * mult;
933 /* Parse VAL as a number and set its value to PLACE (which should
936 By default, the value is assumed to be in bytes. If "K", "M", or
937 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
938 1<<30, respectively. Floating point values are allowed and are
939 cast to integer before use. The idea is to be able to use things
940 like 1.5k instead of "1536".
942 The string "inf" is returned as 0.
944 In case of error, false is returned and memory pointed to by PLACE
945 remains unmodified. */
948 cmd_bytes (const char *com, const char *val, void *place)
951 if (!parse_bytes_helper (val, &byte_value))
953 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
954 exec_name, com, val);
957 *(wgint *)place = (wgint)byte_value;
961 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
962 SIZE_SUM. It works by converting the string to double, therefore
963 working with values up to 2^53-1 without loss of precision. This
964 value (8192 TB) is large enough to serve for a while. */
967 cmd_bytes_sum (const char *com, const char *val, void *place)
970 if (!parse_bytes_helper (val, &byte_value))
972 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
973 exec_name, com, val);
976 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
980 /* Store the value of VAL to *OUT. The value is a time period, by
981 default expressed in seconds, but also accepting suffixes "m", "h",
982 "d", and "w" for minutes, hours, days, and weeks respectively. */
985 cmd_time (const char *com, const char *val, void *place)
988 const char *end = val + strlen (val);
990 /* Strip trailing whitespace. */
991 while (val < end && ISSPACE (end[-1]))
997 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
998 exec_name, com, val);
1002 switch (TOLOWER (end[-1]))
1005 --end, mult = 1; /* seconds */
1008 --end, mult = 60; /* minutes */
1011 --end, mult = 3600; /* hours */
1014 --end, mult = 86400.0; /* days */
1017 --end, mult = 604800.0; /* weeks */
1020 /* Not a recognized suffix: assume it belongs to the number.
1021 (If not, simple_atof will raise an error.) */
1025 /* Skip leading and trailing whitespace. */
1026 while (val < end && ISSPACE (*val))
1028 while (val < end && ISSPACE (end[-1]))
1033 if (!simple_atof (val, end, &number))
1036 *(double *)place = number * mult;
1042 cmd_cert_type (const char *com, const char *val, void *place)
1044 static const struct decode_item choices[] = {
1045 { "pem", keyfile_pem },
1046 { "der", keyfile_asn1 },
1047 { "asn1", keyfile_asn1 },
1049 int ok = decode_string (val, choices, countof (choices), place);
1051 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1056 /* Specialized helper functions, used by `commands' to handle some
1057 options specially. */
1059 static bool check_user_specified_header (const char *);
1062 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1064 if (!cmd_boolean (com, val, &opt.dirstruct))
1066 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1067 must be affected inversely. */
1069 opt.no_dirstruct = false;
1071 opt.no_dirstruct = true;
1076 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1078 /* Empty value means reset the list of headers. */
1081 free_vec (opt.user_headers);
1082 opt.user_headers = NULL;
1086 if (!check_user_specified_header (val))
1088 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1089 exec_name, com, val);
1092 opt.user_headers = vec_append (opt.user_headers, val);
1097 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1099 int flag = cmd_boolean (com, val, &opt.htmlify);
1100 if (flag && !opt.htmlify)
1101 opt.remove_listing = false;
1105 /* Set the "mirror" mode. It means: recursive download, timestamping,
1106 no limit on max. recursion depth, and don't remove listings. */
1109 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1113 if (!cmd_boolean (com, val, &mirror))
1117 opt.recursive = true;
1118 if (!opt.no_dirstruct)
1119 opt.dirstruct = true;
1120 opt.timestamping = true;
1121 opt.reclevel = INFINITE_RECURSION;
1122 opt.remove_listing = false;
1127 /* Validate --prefer-family and set the choice. Allowed values are
1128 "IPv4", "IPv6", and "none". */
1131 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1133 static const struct decode_item choices[] = {
1134 { "IPv4", prefer_ipv4 },
1135 { "IPv6", prefer_ipv6 },
1136 { "none", prefer_none },
1138 int prefer_family = prefer_ipv4;
1139 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1141 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1142 opt.prefer_family = prefer_family;
1146 /* Set progress.type to VAL, but verify that it's a valid progress
1147 implementation before that. */
1150 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1152 if (!valid_progress_implementation_p (val))
1154 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1155 exec_name, com, val);
1158 xfree_null (opt.progress_type);
1160 /* Don't call set_progress_implementation here. It will be called
1161 in main() when it becomes clear what the log output is. */
1162 opt.progress_type = xstrdup (val);
1166 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1167 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1171 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1173 if (!cmd_boolean (com, val, &opt.recursive))
1177 if (opt.recursive && !opt.no_dirstruct)
1178 opt.dirstruct = true;
1184 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1186 int restrict_os = opt.restrict_files_os;
1187 int restrict_ctrl = opt.restrict_files_ctrl;
1188 int restrict_case = opt.restrict_files_case;
1192 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1196 end = strchr (val, ',');
1198 end = val + strlen (val);
1200 if (VAL_IS ("unix"))
1201 restrict_os = restrict_unix;
1202 else if (VAL_IS ("windows"))
1203 restrict_os = restrict_windows;
1204 else if (VAL_IS ("lowercase"))
1205 restrict_case = restrict_lowercase;
1206 else if (VAL_IS ("uppercase"))
1207 restrict_case = restrict_uppercase;
1208 else if (VAL_IS ("nocontrol"))
1209 restrict_ctrl = false;
1213 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1214 exec_name, com, val);
1221 while (*val && *end);
1225 opt.restrict_files_os = restrict_os;
1226 opt.restrict_files_ctrl = restrict_ctrl;
1227 opt.restrict_files_case = restrict_case;
1234 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1236 static const struct decode_item choices[] = {
1237 { "auto", secure_protocol_auto },
1238 { "sslv2", secure_protocol_sslv2 },
1239 { "sslv3", secure_protocol_sslv3 },
1240 { "tlsv1", secure_protocol_tlsv1 },
1242 int ok = decode_string (val, choices, countof (choices), place);
1244 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1249 /* Set all three timeout values. */
1252 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1255 if (!cmd_time (com, val, &value))
1257 opt.read_timeout = value;
1258 opt.connect_timeout = value;
1259 opt.dns_timeout = value;
1264 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1266 /* Disallow embedded newlines. */
1267 if (strchr (val, '\n'))
1269 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1270 exec_name, com, val);
1273 xfree_null (opt.useragent);
1274 opt.useragent = xstrdup (val);
1278 /* The "verbose" option cannot be cmd_boolean because the variable is
1279 not bool -- it's of type int (-1 means uninitialized because of
1280 some random hackery for disallowing -q -v). */
1283 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1286 if (cmd_boolean (com, val, &flag))
1294 /* Miscellaneous useful routines. */
1296 /* A very simple atoi clone, more useful than atoi because it works on
1297 delimited strings, and has error reportage. Returns true on success,
1298 false on failure. If successful, stores result to *DEST. */
1301 simple_atoi (const char *beg, const char *end, int *dest)
1304 bool negative = false;
1305 const char *p = beg;
1307 while (p < end && ISSPACE (*p))
1309 if (p < end && (*p == '-' || *p == '+'))
1311 negative = (*p == '-');
1317 /* Read negative numbers in a separate loop because the most
1318 negative integer cannot be represented as a positive number. */
1321 for (; p < end && ISDIGIT (*p); p++)
1323 int next = (10 * result) + (*p - '0');
1325 return false; /* overflow */
1329 for (; p < end && ISDIGIT (*p); p++)
1331 int next = (10 * result) - (*p - '0');
1333 return false; /* underflow */
1344 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1345 doesn't handle exponential notation. Returns true on success,
1346 false on failure. In case of success, stores its result to
1350 simple_atof (const char *beg, const char *end, double *dest)
1354 bool negative = false;
1355 bool seen_dot = false;
1356 bool seen_digit = false;
1359 const char *p = beg;
1361 while (p < end && ISSPACE (*p))
1363 if (p < end && (*p == '-' || *p == '+'))
1365 negative = (*p == '-');
1369 for (; p < end; p++)
1375 result = (10 * result) + (ch - '0');
1377 result += (ch - '0') / (divider *= 10);
1399 /* Verify that the user-specified header in S is valid. It must
1400 contain a colon preceded by non-white-space characters and must not
1401 contain newlines. */
1404 check_user_specified_header (const char *s)
1408 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++)
1410 /* The header MUST contain `:' preceded by at least one
1411 non-whitespace character. */
1412 if (*p != ':' || p == s)
1414 /* The header MUST NOT contain newlines. */
1415 if (strchr (s, '\n'))
1420 /* Decode VAL into a number, according to ITEMS. */
1423 decode_string (const char *val, const struct decode_item *items, int itemcount,
1427 for (i = 0; i < itemcount; i++)
1428 if (0 == strcasecmp (val, items[i].name))
1430 *place = items[i].code;
1437 void cleanup_html_url (void);
1440 /* Free the memory allocated by global variables. */
1444 /* Free external resources, close files, etc. */
1447 fclose (output_stream);
1448 /* No need to check for error because Wget flushes its output (and
1449 checks for errors) after any data arrives. */
1451 /* We're exiting anyway so there's no real need to call free()
1452 hundreds of times. Skipping the frees will make Wget exit
1455 However, when detecting leaks, it's crucial to free() everything
1456 because then you can find the real leaks, i.e. the allocated
1457 memory which grows with the size of the program. */
1463 cleanup_html_url ();
1468 extern acc_t *netrc_list;
1469 free_netrc (netrc_list);
1471 xfree_null (opt.lfilename);
1472 xfree_null (opt.dir_prefix);
1473 xfree_null (opt.input_filename);
1474 xfree_null (opt.output_document);
1475 free_vec (opt.accepts);
1476 free_vec (opt.rejects);
1477 free_vec (opt.excludes);
1478 free_vec (opt.includes);
1479 free_vec (opt.domains);
1480 free_vec (opt.follow_tags);
1481 free_vec (opt.ignore_tags);
1482 xfree_null (opt.progress_type);
1483 xfree_null (opt.ftp_user);
1484 xfree_null (opt.ftp_passwd);
1485 xfree_null (opt.ftp_proxy);
1486 xfree_null (opt.https_proxy);
1487 xfree_null (opt.http_proxy);
1488 free_vec (opt.no_proxy);
1489 xfree_null (opt.useragent);
1490 xfree_null (opt.referer);
1491 xfree_null (opt.http_user);
1492 xfree_null (opt.http_passwd);
1493 free_vec (opt.user_headers);
1495 xfree_null (opt.cert_file);
1496 xfree_null (opt.private_key);
1497 xfree_null (opt.ca_directory);
1498 xfree_null (opt.ca_cert);
1499 xfree_null (opt.random_file);
1500 xfree_null (opt.egd_file);
1502 xfree_null (opt.bind_address);
1503 xfree_null (opt.cookies_input);
1504 xfree_null (opt.cookies_output);
1505 xfree_null (opt.user);
1506 xfree_null (opt.passwd);
1507 #endif /* DEBUG_MALLOC */
1510 /* Unit testing routines. */
1515 test_cmd_spec_restrict_file_names()
1520 int expected_restrict_files_os;
1521 int expected_restrict_files_ctrl;
1522 int expected_restrict_files_case;
1525 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1526 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1527 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1528 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1531 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1536 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1539 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1540 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1541 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1542 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1544 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1545 res == test_array[i].result
1546 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1547 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1548 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1554 #endif /* TESTING */