1 /* Reading/parsing the initialization file.
2 Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3 2004, 2005, 2006, 2007 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. */
52 #include "recur.h" /* for INFINITE_RECURSION */
53 #include "convert.h" /* for convert_cleanup */
54 #include "res.h" /* for res_cleanup */
55 #include "http.h" /* for http_cleanup */
56 #include "retr.h" /* for output_stream */
62 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
63 otherwise, it will be performed by the shell. This variable will
64 be set by the wgetrc-reading function. */
66 static bool enable_tilde_expansion;
69 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
71 CMD_DECLARE (cmd_boolean);
72 CMD_DECLARE (cmd_bytes);
73 CMD_DECLARE (cmd_bytes_sum);
75 CMD_DECLARE (cmd_cert_type);
77 CMD_DECLARE (cmd_directory_vector);
78 CMD_DECLARE (cmd_number);
79 CMD_DECLARE (cmd_number_inf);
80 CMD_DECLARE (cmd_string);
81 CMD_DECLARE (cmd_file);
82 CMD_DECLARE (cmd_directory);
83 CMD_DECLARE (cmd_time);
84 CMD_DECLARE (cmd_vector);
86 CMD_DECLARE (cmd_spec_dirstruct);
87 CMD_DECLARE (cmd_spec_header);
88 CMD_DECLARE (cmd_spec_htmlify);
89 CMD_DECLARE (cmd_spec_mirror);
90 CMD_DECLARE (cmd_spec_prefer_family);
91 CMD_DECLARE (cmd_spec_progress);
92 CMD_DECLARE (cmd_spec_recursive);
93 CMD_DECLARE (cmd_spec_restrict_file_names);
95 CMD_DECLARE (cmd_spec_secure_protocol);
97 CMD_DECLARE (cmd_spec_timeout);
98 CMD_DECLARE (cmd_spec_useragent);
99 CMD_DECLARE (cmd_spec_verbose);
101 /* List of recognized commands, each consisting of name, place and
102 function. When adding a new command, simply add it to the list,
103 but be sure to keep the list sorted alphabetically, as
104 command_by_name's binary search depends on it. Also, be sure to
105 add any entries that allocate memory (e.g. cmd_string and
106 cmd_vector) to the cleanup() function below. */
108 static const struct {
111 bool (*action) (const char *, const char *, void *);
113 /* KEEP THIS LIST ALPHABETICALLY SORTED */
114 { "accept", &opt.accepts, cmd_vector },
115 { "addhostdir", &opt.add_hostdir, cmd_boolean },
116 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
117 { "background", &opt.background, cmd_boolean },
118 { "backupconverted", &opt.backup_converted, cmd_boolean },
119 { "backups", &opt.backups, cmd_number },
120 { "base", &opt.base_href, cmd_string },
121 { "bindaddress", &opt.bind_address, cmd_string },
123 { "cacertificate", &opt.ca_cert, cmd_file },
125 { "cache", &opt.allow_cache, cmd_boolean },
127 { "cadirectory", &opt.ca_directory, cmd_directory },
128 { "certificate", &opt.cert_file, cmd_file },
129 { "certificatetype", &opt.cert_type, cmd_cert_type },
130 { "checkcertificate", &opt.check_cert, cmd_boolean },
132 { "connecttimeout", &opt.connect_timeout, cmd_time },
133 { "contentdisposition", &opt.content_disposition, cmd_boolean },
134 { "continue", &opt.always_rest, cmd_boolean },
135 { "convertlinks", &opt.convert_links, cmd_boolean },
136 { "cookies", &opt.cookies, cmd_boolean },
137 { "cutdirs", &opt.cut_dirs, cmd_number },
139 { "debug", &opt.debug, cmd_boolean },
141 { "deleteafter", &opt.delete_after, cmd_boolean },
142 { "dirprefix", &opt.dir_prefix, cmd_directory },
143 { "dirstruct", NULL, cmd_spec_dirstruct },
144 { "dnscache", &opt.dns_cache, cmd_boolean },
145 { "dnstimeout", &opt.dns_timeout, cmd_time },
146 { "domains", &opt.domains, cmd_vector },
147 { "dotbytes", &opt.dot_bytes, cmd_bytes },
148 { "dotsinline", &opt.dots_in_line, cmd_number },
149 { "dotspacing", &opt.dot_spacing, cmd_number },
150 { "dotstyle", &opt.dot_style, cmd_string },
152 { "egdfile", &opt.egd_file, cmd_file },
154 { "excludedirectories", &opt.excludes, cmd_directory_vector },
155 { "excludedomains", &opt.exclude_domains, cmd_vector },
156 { "followftp", &opt.follow_ftp, cmd_boolean },
157 { "followtags", &opt.follow_tags, cmd_vector },
158 { "forcehtml", &opt.force_html, cmd_boolean },
159 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
160 { "ftppassword", &opt.ftp_passwd, cmd_string },
161 { "ftpproxy", &opt.ftp_proxy, cmd_string },
162 { "ftpuser", &opt.ftp_user, cmd_string },
163 { "glob", &opt.ftp_glob, cmd_boolean },
164 { "header", NULL, cmd_spec_header },
165 { "htmlextension", &opt.html_extension, cmd_boolean },
166 { "htmlify", NULL, cmd_spec_htmlify },
167 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
168 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
169 { "httppassword", &opt.http_passwd, cmd_string },
170 { "httpproxy", &opt.http_proxy, cmd_string },
171 { "httpsproxy", &opt.https_proxy, cmd_string },
172 { "httpuser", &opt.http_user, cmd_string },
173 { "ignorecase", &opt.ignore_case, cmd_boolean },
174 { "ignorelength", &opt.ignore_length, cmd_boolean },
175 { "ignoretags", &opt.ignore_tags, cmd_vector },
176 { "includedirectories", &opt.includes, cmd_directory_vector },
178 { "inet4only", &opt.ipv4_only, cmd_boolean },
179 { "inet6only", &opt.ipv6_only, cmd_boolean },
181 { "input", &opt.input_filename, cmd_file },
182 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
183 { "limitrate", &opt.limit_rate, cmd_bytes },
184 { "loadcookies", &opt.cookies_input, cmd_file },
185 { "logfile", &opt.lfilename, cmd_file },
186 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
187 { "maxredirect", &opt.max_redirect, cmd_number },
188 { "mirror", NULL, cmd_spec_mirror },
189 { "netrc", &opt.netrc, cmd_boolean },
190 { "noclobber", &opt.noclobber, cmd_boolean },
191 { "noparent", &opt.no_parent, cmd_boolean },
192 { "noproxy", &opt.no_proxy, cmd_vector },
193 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
194 { "outputdocument", &opt.output_document, cmd_file },
195 { "pagerequisites", &opt.page_requisites, cmd_boolean },
196 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
197 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
198 { "password", &opt.passwd, cmd_string },
199 { "postdata", &opt.post_data, cmd_string },
200 { "postfile", &opt.post_file_name, cmd_file },
201 { "preferfamily", NULL, cmd_spec_prefer_family },
202 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
204 { "privatekey", &opt.private_key, cmd_file },
205 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
207 { "progress", &opt.progress_type, cmd_spec_progress },
208 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
209 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
210 { "proxypassword", &opt.proxy_passwd, cmd_string },
211 { "proxyuser", &opt.proxy_user, cmd_string },
212 { "quiet", &opt.quiet, cmd_boolean },
213 { "quota", &opt.quota, cmd_bytes_sum },
215 { "randomfile", &opt.random_file, cmd_file },
217 { "randomwait", &opt.random_wait, cmd_boolean },
218 { "readtimeout", &opt.read_timeout, cmd_time },
219 { "reclevel", &opt.reclevel, cmd_number_inf },
220 { "recursive", NULL, cmd_spec_recursive },
221 { "referer", &opt.referer, cmd_string },
222 { "reject", &opt.rejects, cmd_vector },
223 { "relativeonly", &opt.relative_only, cmd_boolean },
224 { "removelisting", &opt.remove_listing, cmd_boolean },
225 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
226 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
227 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
228 { "robots", &opt.use_robots, cmd_boolean },
229 { "savecookies", &opt.cookies_output, cmd_file },
230 { "saveheaders", &opt.save_headers, cmd_boolean },
232 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
234 { "serverresponse", &opt.server_response, cmd_boolean },
235 { "spanhosts", &opt.spanhost, cmd_boolean },
236 { "spider", &opt.spider, cmd_boolean },
237 { "strictcomments", &opt.strict_comments, cmd_boolean },
238 { "timeout", NULL, cmd_spec_timeout },
239 { "timestamping", &opt.timestamping, cmd_boolean },
240 { "tries", &opt.ntry, cmd_number_inf },
241 { "useproxy", &opt.use_proxy, cmd_boolean },
242 { "user", &opt.user, cmd_string },
243 { "useragent", NULL, cmd_spec_useragent },
244 { "verbose", NULL, cmd_spec_verbose },
245 { "wait", &opt.wait, cmd_time },
246 { "waitretry", &opt.waitretry, cmd_time },
248 { "wdebug", &opt.wdebug, cmd_boolean },
252 /* Look up CMDNAME in the commands[] and return its position in the
253 array. If CMDNAME is not found, return -1. */
256 command_by_name (const char *cmdname)
258 /* Use binary search for speed. Wget has ~100 commands, which
259 guarantees a worst case performance of 7 string comparisons. */
260 int lo = 0, hi = countof (commands) - 1;
264 int mid = (lo + hi) >> 1;
265 int cmp = strcasecmp (cmdname, commands[mid].name);
276 /* Reset the variables to default values. */
282 /* Most of the default values are 0 (and 0.0, NULL, and false).
283 Just reset everything, and fill in the non-zero values. Note
284 that initializing pointers to NULL this way is technically
285 illegal, but porting Wget to a machine where NULL is not all-zero
286 bit pattern will be the least of the implementors' worries. */
293 opt.add_hostdir = true;
297 opt.http_keep_alive = true;
298 opt.use_proxy = true;
299 tmp = getenv ("no_proxy");
301 opt.no_proxy = sepstring (tmp);
302 opt.allow_cache = true;
304 opt.read_timeout = 900;
305 opt.use_robots = true;
307 opt.remove_listing = true;
309 opt.dot_bytes = 1024;
310 opt.dot_spacing = 10;
311 opt.dots_in_line = 50;
313 opt.dns_cache = true;
317 opt.check_cert = true;
320 /* The default for file name restriction defaults to the OS type. */
321 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
322 opt.restrict_files_os = restrict_windows;
324 opt.restrict_files_os = restrict_unix;
326 opt.restrict_files_ctrl = true;
327 opt.restrict_files_case = restrict_no_case_restriction;
329 opt.max_redirect = 20;
332 /* Return the user's home directory (strdup-ed), or NULL if none is
337 char *home = getenv ("HOME");
342 /* Under MSDOS, if $HOME isn't defined, use the directory where
343 `wget.exe' resides. */
344 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
345 char *p, buf[PATH_MAX];
347 strcpy (buf, _w32_get_argv0 ());
348 p = strrchr (buf, '/'); /* djgpp */
350 p = strrchr (buf, '\\'); /* others */
354 #elif !defined(WINDOWS)
355 /* If HOME is not defined, try getting it from the password
357 struct passwd *pwd = getpwuid (getuid ());
358 if (!pwd || !pwd->pw_dir)
362 /* Under Windows, if $HOME isn't defined, use the directory where
363 `wget.exe' resides. */
368 return home ? xstrdup (home) : NULL;
371 /* Return the path to the user's .wgetrc. This is either the value of
372 `WGETRC' environment variable, or `$HOME/.wgetrc'.
374 If the `WGETRC' variable exists but the file does not exist, the
375 function will exit(). */
377 wgetrc_file_name (void)
382 /* Try the environment. */
383 env = getenv ("WGETRC");
386 if (!file_exists_p (env))
388 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
392 return xstrdup (env);
395 /* If that failed, try $HOME/.wgetrc. */
398 file = aprintf ("%s/.wgetrc", home);
402 /* Under Windows, if we still haven't found .wgetrc, look for the file
403 `wget.ini' in the directory where `wget.exe' resides; we do this for
404 backward compatibility with previous versions of Wget.
405 SYSTEM_WGETRC should not be defined under WINDOWS. */
406 if (!file || !file_exists_p (file))
412 file = aprintf ("%s/wget.ini", home);
418 if (!file_exists_p (file))
426 /* Return values of parse_line. */
434 static enum parse_line parse_line (const char *, char **, char **, int *);
435 static bool setval_internal (int, const char *, const char *);
437 /* Initialize variables from a wgetrc file. Returns zero (failure) if
438 there were errors in the file. */
441 run_wgetrc (const char *file)
448 fp = fopen (file, "rb");
451 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
452 file, strerror (errno));
453 return true; /* not a fatal error */
455 enable_tilde_expansion = true;
457 while ((line = read_whole_line (fp)) != NULL)
459 char *com = NULL, *val = NULL;
462 /* Parse the line. */
463 switch (parse_line (line, &com, &val, &comind))
466 /* If everything is OK, set the value. */
467 if (!setval_internal (comind, com, val))
469 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
470 exec_name, file, ln);
474 case line_syntax_error:
475 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
476 exec_name, file, ln);
479 case line_unknown_command:
480 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
481 exec_name, com, file, ln);
494 enable_tilde_expansion = false;
500 /* Initialize the defaults and run the system wgetrc and user's own
508 /* Load the hard-coded defaults. */
511 /* If SYSTEM_WGETRC is defined, use it. */
513 if (file_exists_p (SYSTEM_WGETRC))
514 ok &= run_wgetrc (SYSTEM_WGETRC);
516 /* Override it with your own, if one exists. */
517 file = wgetrc_file_name ();
520 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
521 something like realpath() before comparing them with `strcmp' */
523 if (!strcmp (file, SYSTEM_WGETRC))
525 fprintf (stderr, _("\
526 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
531 ok &= run_wgetrc (file);
533 /* If there were errors processing either `.wgetrc', abort. */
541 /* Remove dashes and underscores from S, modifying S in the
547 char *t = s; /* t - tortoise */
548 char *h = s; /* h - hare */
550 if (*h == '_' || *h == '-')
557 /* Parse the line pointed by line, with the syntax:
558 <sp>* command <sp>* = <sp>* value <sp>*
559 Uses malloc to allocate space for command and value.
561 Returns one of line_ok, line_empty, line_syntax_error, or
562 line_unknown_command.
564 In case of line_ok, *COM and *VAL point to freshly allocated
565 strings, and *COMIND points to com's index. In case of error or
566 empty line, their values are unmodified. */
568 static enum parse_line
569 parse_line (const char *line, char **com, char **val, int *comind)
572 const char *end = line + strlen (line);
573 const char *cmdstart, *cmdend;
574 const char *valstart, *valend;
579 /* Skip leading and trailing whitespace. */
580 while (*line && ISSPACE (*line))
582 while (end > line && ISSPACE (end[-1]))
585 /* Skip empty lines and comments. */
586 if (!*line || *line == '#')
592 while (p < end && (ISALNUM (*p) || *p == '_' || *p == '-'))
596 /* Skip '=', as well as any space before or after it. */
597 while (p < end && ISSPACE (*p))
599 if (p == end || *p != '=')
600 return line_syntax_error;
602 while (p < end && ISSPACE (*p))
608 /* The syntax is valid (even though the command might not be). Fill
609 in the command name and value. */
610 *com = strdupdelim (cmdstart, cmdend);
611 *val = strdupdelim (valstart, valend);
613 /* The line now known to be syntactically correct. Check whether
614 the command is valid. */
615 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
617 ind = command_by_name (cmdcopy);
619 return line_unknown_command;
621 /* Report success to the caller. */
626 /* Run commands[comind].action. */
629 setval_internal (int comind, const char *com, const char *val)
631 assert (0 <= comind && comind < countof (commands));
632 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
633 return commands[comind].action (com, val, commands[comind].place);
636 /* Run command COM with value VAL. If running the command produces an
637 error, report the error and exit.
639 This is intended to be called from main() to modify Wget's behavior
640 through command-line switches. Since COM is hard-coded in main(),
641 it is not canonicalized, and this aborts when COM is not found.
643 If COMIND's are exported to init.h, this function will be changed
644 to accept COMIND directly. */
647 setoptval (const char *com, const char *val, const char *optname)
649 /* Prepend "--" to OPTNAME. */
650 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
653 strcpy (dd_optname + 2, optname);
655 assert (val != NULL);
656 if (!setval_internal (command_by_name (com), dd_optname, val))
660 /* Parse OPT into command and value and run it. For example,
661 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
662 This is used by the `--execute' flag in main.c. */
665 run_command (const char *opt)
669 switch (parse_line (opt, &com, &val, &comind))
672 if (!setval_internal (comind, com, val))
678 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
684 /* Generic helper functions, for use with `commands'. */
686 /* Forward declarations: */
691 static bool decode_string (const char *, const struct decode_item *, int, int *);
692 static bool simple_atoi (const char *, const char *, int *);
693 static bool simple_atof (const char *, const char *, double *);
695 #define CMP1(p, c0) (TOLOWER((p)[0]) == (c0) && (p)[1] == '\0')
697 #define CMP2(p, c0, c1) (TOLOWER((p)[0]) == (c0) \
698 && TOLOWER((p)[1]) == (c1) \
701 #define CMP3(p, c0, c1, c2) (TOLOWER((p)[0]) == (c0) \
702 && TOLOWER((p)[1]) == (c1) \
703 && TOLOWER((p)[2]) == (c2) \
707 /* Store the boolean value from VAL to PLACE. COM is ignored,
708 except for error messages. */
710 cmd_boolean (const char *com, const char *val, void *place)
714 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
715 /* "on", "yes" and "1" mean true. */
717 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
718 /* "off", "no" and "0" mean false. */
723 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
724 exec_name, com, val);
728 *(bool *) place = value;
732 /* Set the non-negative integer value from VAL to PLACE. With
733 incorrect specification, the number remains unchanged. */
735 cmd_number (const char *com, const char *val, void *place)
737 if (!simple_atoi (val, val + strlen (val), place)
738 || *(int *) place < 0)
740 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
741 exec_name, com, val);
747 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
749 cmd_number_inf (const char *com, const char *val, void *place)
751 if (!strcasecmp (val, "inf"))
756 return cmd_number (com, val, place);
759 /* Copy (strdup) the string at COM to a new location and place a
760 pointer to *PLACE. */
762 cmd_string (const char *com, const char *val, void *place)
764 char **pstring = (char **)place;
766 xfree_null (*pstring);
767 *pstring = xstrdup (val);
771 #if defined(WINDOWS) || defined(MSDOS)
772 # define ISSEP(c) ((c) == '/' || (c) == '\\')
774 # define ISSEP(c) ((c) == '/')
777 /* Like the above, but handles tilde-expansion when reading a user's
778 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
779 gets expanded to the user's home directory. */
781 cmd_file (const char *com, const char *val, void *place)
783 char **pstring = (char **)place;
785 xfree_null (*pstring);
787 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
789 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
792 *pstring = xstrdup (val);
797 char *home = home_dir ();
801 homelen = strlen (home);
802 while (homelen && ISSEP (home[homelen - 1]))
803 home[--homelen] = '\0';
805 /* Skip the leading "~/". */
806 for (++val; ISSEP (*val); val++)
809 *pstring = concat_strings (home, "/", val, (char *) 0);
812 #if defined(WINDOWS) || defined(MSDOS)
813 /* Convert "\" to "/". */
816 for (s = *pstring; *s; s++)
824 /* Like cmd_file, but strips trailing '/' characters. */
826 cmd_directory (const char *com, const char *val, void *place)
830 /* Call cmd_file() for tilde expansion and separator
831 canonicalization (backslash -> slash under Windows). These
832 things should perhaps be in a separate function. */
833 if (!cmd_file (com, val, place))
838 while (t > s && *--t == '/')
844 /* Split VAL by space to a vector of values, and append those values
845 to vector pointed to by the PLACE argument. If VAL is empty, the
846 PLACE vector is cleared instead. */
849 cmd_vector (const char *com, const char *val, void *place)
851 char ***pvec = (char ***)place;
854 *pvec = merge_vecs (*pvec, sepstring (val));
864 cmd_directory_vector (const char *com, const char *val, void *place)
866 char ***pvec = (char ***)place;
870 /* Strip the trailing slashes from directories. */
873 seps = sepstring (val);
874 for (t = seps; t && *t; t++)
876 int len = strlen (*t);
877 /* Skip degenerate case of root directory. */
880 if ((*t)[len - 1] == '/')
881 (*t)[len - 1] = '\0';
884 *pvec = merge_vecs (*pvec, seps);
894 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
895 "100k" or "2.5G" to a floating point number. */
898 parse_bytes_helper (const char *val, double *result)
901 const char *end = val + strlen (val);
903 /* Check for "inf". */
904 if (0 == strcmp (val, "inf"))
910 /* Strip trailing whitespace. */
911 while (val < end && ISSPACE (end[-1]))
916 switch (TOLOWER (end[-1]))
919 --end, mult = 1024.0;
922 --end, mult = 1048576.0;
925 --end, mult = 1073741824.0;
928 --end, mult = 1099511627776.0;
931 /* Not a recognized suffix: assume it's a digit. (If not,
932 simple_atof will raise an error.) */
936 /* Skip leading and trailing whitespace. */
937 while (val < end && ISSPACE (*val))
939 while (val < end && ISSPACE (end[-1]))
944 if (!simple_atof (val, end, &number) || number < 0)
947 *result = number * mult;
951 /* Parse VAL as a number and set its value to PLACE (which should
954 By default, the value is assumed to be in bytes. If "K", "M", or
955 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
956 1<<30, respectively. Floating point values are allowed and are
957 cast to integer before use. The idea is to be able to use things
958 like 1.5k instead of "1536".
960 The string "inf" is returned as 0.
962 In case of error, false is returned and memory pointed to by PLACE
963 remains unmodified. */
966 cmd_bytes (const char *com, const char *val, void *place)
969 if (!parse_bytes_helper (val, &byte_value))
971 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
972 exec_name, com, val);
975 *(wgint *)place = (wgint)byte_value;
979 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
980 SIZE_SUM. It works by converting the string to double, therefore
981 working with values up to 2^53-1 without loss of precision. This
982 value (8192 TB) is large enough to serve for a while. */
985 cmd_bytes_sum (const char *com, const char *val, void *place)
988 if (!parse_bytes_helper (val, &byte_value))
990 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
991 exec_name, com, val);
994 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
998 /* Store the value of VAL to *OUT. The value is a time period, by
999 default expressed in seconds, but also accepting suffixes "m", "h",
1000 "d", and "w" for minutes, hours, days, and weeks respectively. */
1003 cmd_time (const char *com, const char *val, void *place)
1005 double number, mult;
1006 const char *end = val + strlen (val);
1008 /* Strip trailing whitespace. */
1009 while (val < end && ISSPACE (end[-1]))
1015 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1016 exec_name, com, val);
1020 switch (TOLOWER (end[-1]))
1023 --end, mult = 1; /* seconds */
1026 --end, mult = 60; /* minutes */
1029 --end, mult = 3600; /* hours */
1032 --end, mult = 86400.0; /* days */
1035 --end, mult = 604800.0; /* weeks */
1038 /* Not a recognized suffix: assume it belongs to the number.
1039 (If not, simple_atof will raise an error.) */
1043 /* Skip leading and trailing whitespace. */
1044 while (val < end && ISSPACE (*val))
1046 while (val < end && ISSPACE (end[-1]))
1051 if (!simple_atof (val, end, &number))
1054 *(double *)place = number * mult;
1060 cmd_cert_type (const char *com, const char *val, void *place)
1062 static const struct decode_item choices[] = {
1063 { "pem", keyfile_pem },
1064 { "der", keyfile_asn1 },
1065 { "asn1", keyfile_asn1 },
1067 int ok = decode_string (val, choices, countof (choices), place);
1069 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1074 /* Specialized helper functions, used by `commands' to handle some
1075 options specially. */
1077 static bool check_user_specified_header (const char *);
1080 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1082 if (!cmd_boolean (com, val, &opt.dirstruct))
1084 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1085 must be affected inversely. */
1087 opt.no_dirstruct = false;
1089 opt.no_dirstruct = true;
1094 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1096 /* Empty value means reset the list of headers. */
1099 free_vec (opt.user_headers);
1100 opt.user_headers = NULL;
1104 if (!check_user_specified_header (val))
1106 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1107 exec_name, com, val);
1110 opt.user_headers = vec_append (opt.user_headers, val);
1115 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1117 int flag = cmd_boolean (com, val, &opt.htmlify);
1118 if (flag && !opt.htmlify)
1119 opt.remove_listing = false;
1123 /* Set the "mirror" mode. It means: recursive download, timestamping,
1124 no limit on max. recursion depth, and don't remove listings. */
1127 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1131 if (!cmd_boolean (com, val, &mirror))
1135 opt.recursive = true;
1136 if (!opt.no_dirstruct)
1137 opt.dirstruct = true;
1138 opt.timestamping = true;
1139 opt.reclevel = INFINITE_RECURSION;
1140 opt.remove_listing = false;
1145 /* Validate --prefer-family and set the choice. Allowed values are
1146 "IPv4", "IPv6", and "none". */
1149 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1151 static const struct decode_item choices[] = {
1152 { "IPv4", prefer_ipv4 },
1153 { "IPv6", prefer_ipv6 },
1154 { "none", prefer_none },
1156 int prefer_family = prefer_ipv4;
1157 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1159 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1160 opt.prefer_family = prefer_family;
1164 /* Set progress.type to VAL, but verify that it's a valid progress
1165 implementation before that. */
1168 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1170 if (!valid_progress_implementation_p (val))
1172 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1173 exec_name, com, val);
1176 xfree_null (opt.progress_type);
1178 /* Don't call set_progress_implementation here. It will be called
1179 in main() when it becomes clear what the log output is. */
1180 opt.progress_type = xstrdup (val);
1184 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1185 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1189 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1191 if (!cmd_boolean (com, val, &opt.recursive))
1195 if (opt.recursive && !opt.no_dirstruct)
1196 opt.dirstruct = true;
1202 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1204 int restrict_os = opt.restrict_files_os;
1205 int restrict_ctrl = opt.restrict_files_ctrl;
1206 int restrict_case = opt.restrict_files_case;
1210 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1214 end = strchr (val, ',');
1216 end = val + strlen (val);
1218 if (VAL_IS ("unix"))
1219 restrict_os = restrict_unix;
1220 else if (VAL_IS ("windows"))
1221 restrict_os = restrict_windows;
1222 else if (VAL_IS ("lowercase"))
1223 restrict_case = restrict_lowercase;
1224 else if (VAL_IS ("uppercase"))
1225 restrict_case = restrict_uppercase;
1226 else if (VAL_IS ("nocontrol"))
1227 restrict_ctrl = false;
1231 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1232 exec_name, com, val);
1239 while (*val && *end);
1243 opt.restrict_files_os = restrict_os;
1244 opt.restrict_files_ctrl = restrict_ctrl;
1245 opt.restrict_files_case = restrict_case;
1252 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1254 static const struct decode_item choices[] = {
1255 { "auto", secure_protocol_auto },
1256 { "sslv2", secure_protocol_sslv2 },
1257 { "sslv3", secure_protocol_sslv3 },
1258 { "tlsv1", secure_protocol_tlsv1 },
1260 int ok = decode_string (val, choices, countof (choices), place);
1262 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1267 /* Set all three timeout values. */
1270 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1273 if (!cmd_time (com, val, &value))
1275 opt.read_timeout = value;
1276 opt.connect_timeout = value;
1277 opt.dns_timeout = value;
1282 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1284 /* Disallow embedded newlines. */
1285 if (strchr (val, '\n'))
1287 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1288 exec_name, com, val);
1291 xfree_null (opt.useragent);
1292 opt.useragent = xstrdup (val);
1296 /* The "verbose" option cannot be cmd_boolean because the variable is
1297 not bool -- it's of type int (-1 means uninitialized because of
1298 some random hackery for disallowing -q -v). */
1301 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1304 if (cmd_boolean (com, val, &flag))
1312 /* Miscellaneous useful routines. */
1314 /* A very simple atoi clone, more useful than atoi because it works on
1315 delimited strings, and has error reportage. Returns true on success,
1316 false on failure. If successful, stores result to *DEST. */
1319 simple_atoi (const char *beg, const char *end, int *dest)
1322 bool negative = false;
1323 const char *p = beg;
1325 while (p < end && ISSPACE (*p))
1327 if (p < end && (*p == '-' || *p == '+'))
1329 negative = (*p == '-');
1335 /* Read negative numbers in a separate loop because the most
1336 negative integer cannot be represented as a positive number. */
1339 for (; p < end && ISDIGIT (*p); p++)
1341 int next = (10 * result) + (*p - '0');
1343 return false; /* overflow */
1347 for (; p < end && ISDIGIT (*p); p++)
1349 int next = (10 * result) - (*p - '0');
1351 return false; /* underflow */
1362 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1363 doesn't handle exponential notation. Returns true on success,
1364 false on failure. In case of success, stores its result to
1368 simple_atof (const char *beg, const char *end, double *dest)
1372 bool negative = false;
1373 bool seen_dot = false;
1374 bool seen_digit = false;
1377 const char *p = beg;
1379 while (p < end && ISSPACE (*p))
1381 if (p < end && (*p == '-' || *p == '+'))
1383 negative = (*p == '-');
1387 for (; p < end; p++)
1393 result = (10 * result) + (ch - '0');
1395 result += (ch - '0') / (divider *= 10);
1417 /* Verify that the user-specified header in S is valid. It must
1418 contain a colon preceded by non-white-space characters and must not
1419 contain newlines. */
1422 check_user_specified_header (const char *s)
1426 for (p = s; *p && *p != ':' && !ISSPACE (*p); p++)
1428 /* The header MUST contain `:' preceded by at least one
1429 non-whitespace character. */
1430 if (*p != ':' || p == s)
1432 /* The header MUST NOT contain newlines. */
1433 if (strchr (s, '\n'))
1438 /* Decode VAL into a number, according to ITEMS. */
1441 decode_string (const char *val, const struct decode_item *items, int itemcount,
1445 for (i = 0; i < itemcount; i++)
1446 if (0 == strcasecmp (val, items[i].name))
1448 *place = items[i].code;
1455 void cleanup_html_url (void);
1458 /* Free the memory allocated by global variables. */
1462 /* Free external resources, close files, etc. */
1465 fclose (output_stream);
1466 /* No need to check for error because Wget flushes its output (and
1467 checks for errors) after any data arrives. */
1469 /* We're exiting anyway so there's no real need to call free()
1470 hundreds of times. Skipping the frees will make Wget exit
1473 However, when detecting leaks, it's crucial to free() everything
1474 because then you can find the real leaks, i.e. the allocated
1475 memory which grows with the size of the program. */
1481 cleanup_html_url ();
1486 extern acc_t *netrc_list;
1487 free_netrc (netrc_list);
1489 xfree_null (opt.lfilename);
1490 xfree_null (opt.dir_prefix);
1491 xfree_null (opt.input_filename);
1492 xfree_null (opt.output_document);
1493 free_vec (opt.accepts);
1494 free_vec (opt.rejects);
1495 free_vec (opt.excludes);
1496 free_vec (opt.includes);
1497 free_vec (opt.domains);
1498 free_vec (opt.follow_tags);
1499 free_vec (opt.ignore_tags);
1500 xfree_null (opt.progress_type);
1501 xfree_null (opt.ftp_user);
1502 xfree_null (opt.ftp_passwd);
1503 xfree_null (opt.ftp_proxy);
1504 xfree_null (opt.https_proxy);
1505 xfree_null (opt.http_proxy);
1506 free_vec (opt.no_proxy);
1507 xfree_null (opt.useragent);
1508 xfree_null (opt.referer);
1509 xfree_null (opt.http_user);
1510 xfree_null (opt.http_passwd);
1511 free_vec (opt.user_headers);
1513 xfree_null (opt.cert_file);
1514 xfree_null (opt.private_key);
1515 xfree_null (opt.ca_directory);
1516 xfree_null (opt.ca_cert);
1517 xfree_null (opt.random_file);
1518 xfree_null (opt.egd_file);
1520 xfree_null (opt.bind_address);
1521 xfree_null (opt.cookies_input);
1522 xfree_null (opt.cookies_output);
1523 xfree_null (opt.user);
1524 xfree_null (opt.passwd);
1525 #endif /* DEBUG_MALLOC */
1528 /* Unit testing routines. */
1533 test_cmd_spec_restrict_file_names()
1538 int expected_restrict_files_os;
1539 int expected_restrict_files_ctrl;
1540 int expected_restrict_files_case;
1543 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1544 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1545 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1546 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1549 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1554 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1557 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1558 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1559 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1560 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1562 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1563 res == test_array[i].result
1564 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1565 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1566 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1572 #endif /* TESTING */