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 { "authnochallenge", &opt.auth_without_challenge,
118 { "background", &opt.background, cmd_boolean },
119 { "backupconverted", &opt.backup_converted, cmd_boolean },
120 { "backups", &opt.backups, cmd_number },
121 { "base", &opt.base_href, cmd_string },
122 { "bindaddress", &opt.bind_address, cmd_string },
124 { "cacertificate", &opt.ca_cert, cmd_file },
126 { "cache", &opt.allow_cache, cmd_boolean },
128 { "cadirectory", &opt.ca_directory, cmd_directory },
129 { "certificate", &opt.cert_file, cmd_file },
130 { "certificatetype", &opt.cert_type, cmd_cert_type },
131 { "checkcertificate", &opt.check_cert, cmd_boolean },
133 { "connecttimeout", &opt.connect_timeout, cmd_time },
134 { "contentdisposition", &opt.content_disposition, cmd_boolean },
135 { "continue", &opt.always_rest, cmd_boolean },
136 { "convertlinks", &opt.convert_links, cmd_boolean },
137 { "cookies", &opt.cookies, cmd_boolean },
138 { "cutdirs", &opt.cut_dirs, cmd_number },
140 { "debug", &opt.debug, cmd_boolean },
142 { "deleteafter", &opt.delete_after, cmd_boolean },
143 { "dirprefix", &opt.dir_prefix, cmd_directory },
144 { "dirstruct", NULL, cmd_spec_dirstruct },
145 { "dnscache", &opt.dns_cache, cmd_boolean },
146 { "dnstimeout", &opt.dns_timeout, cmd_time },
147 { "domains", &opt.domains, cmd_vector },
148 { "dotbytes", &opt.dot_bytes, cmd_bytes },
149 { "dotsinline", &opt.dots_in_line, cmd_number },
150 { "dotspacing", &opt.dot_spacing, cmd_number },
151 { "dotstyle", &opt.dot_style, cmd_string },
153 { "egdfile", &opt.egd_file, cmd_file },
155 { "excludedirectories", &opt.excludes, cmd_directory_vector },
156 { "excludedomains", &opt.exclude_domains, cmd_vector },
157 { "followftp", &opt.follow_ftp, cmd_boolean },
158 { "followtags", &opt.follow_tags, cmd_vector },
159 { "forcehtml", &opt.force_html, cmd_boolean },
160 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
161 { "ftppassword", &opt.ftp_passwd, cmd_string },
162 { "ftpproxy", &opt.ftp_proxy, cmd_string },
163 { "ftpuser", &opt.ftp_user, cmd_string },
164 { "glob", &opt.ftp_glob, cmd_boolean },
165 { "header", NULL, cmd_spec_header },
166 { "htmlextension", &opt.html_extension, cmd_boolean },
167 { "htmlify", NULL, cmd_spec_htmlify },
168 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
169 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
170 { "httppassword", &opt.http_passwd, cmd_string },
171 { "httpproxy", &opt.http_proxy, cmd_string },
172 { "httpsproxy", &opt.https_proxy, cmd_string },
173 { "httpuser", &opt.http_user, cmd_string },
174 { "ignorecase", &opt.ignore_case, cmd_boolean },
175 { "ignorelength", &opt.ignore_length, cmd_boolean },
176 { "ignoretags", &opt.ignore_tags, cmd_vector },
177 { "includedirectories", &opt.includes, cmd_directory_vector },
179 { "inet4only", &opt.ipv4_only, cmd_boolean },
180 { "inet6only", &opt.ipv6_only, cmd_boolean },
182 { "input", &opt.input_filename, cmd_file },
183 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
184 { "limitrate", &opt.limit_rate, cmd_bytes },
185 { "loadcookies", &opt.cookies_input, cmd_file },
186 { "logfile", &opt.lfilename, cmd_file },
187 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
188 { "maxredirect", &opt.max_redirect, cmd_number },
189 { "mirror", NULL, cmd_spec_mirror },
190 { "netrc", &opt.netrc, cmd_boolean },
191 { "noclobber", &opt.noclobber, cmd_boolean },
192 { "noparent", &opt.no_parent, cmd_boolean },
193 { "noproxy", &opt.no_proxy, cmd_vector },
194 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
195 { "outputdocument", &opt.output_document, cmd_file },
196 { "pagerequisites", &opt.page_requisites, cmd_boolean },
197 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
198 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
199 { "password", &opt.passwd, cmd_string },
200 { "postdata", &opt.post_data, cmd_string },
201 { "postfile", &opt.post_file_name, cmd_file },
202 { "preferfamily", NULL, cmd_spec_prefer_family },
203 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
205 { "privatekey", &opt.private_key, cmd_file },
206 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
208 { "progress", &opt.progress_type, cmd_spec_progress },
209 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
210 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
211 { "proxypassword", &opt.proxy_passwd, cmd_string },
212 { "proxyuser", &opt.proxy_user, cmd_string },
213 { "quiet", &opt.quiet, cmd_boolean },
214 { "quota", &opt.quota, cmd_bytes_sum },
216 { "randomfile", &opt.random_file, cmd_file },
218 { "randomwait", &opt.random_wait, cmd_boolean },
219 { "readtimeout", &opt.read_timeout, cmd_time },
220 { "reclevel", &opt.reclevel, cmd_number_inf },
221 { "recursive", NULL, cmd_spec_recursive },
222 { "referer", &opt.referer, cmd_string },
223 { "reject", &opt.rejects, cmd_vector },
224 { "relativeonly", &opt.relative_only, cmd_boolean },
225 { "removelisting", &opt.remove_listing, cmd_boolean },
226 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
227 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
228 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
229 { "robots", &opt.use_robots, cmd_boolean },
230 { "savecookies", &opt.cookies_output, cmd_file },
231 { "saveheaders", &opt.save_headers, cmd_boolean },
233 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
235 { "serverresponse", &opt.server_response, cmd_boolean },
236 { "spanhosts", &opt.spanhost, cmd_boolean },
237 { "spider", &opt.spider, cmd_boolean },
238 { "strictcomments", &opt.strict_comments, cmd_boolean },
239 { "timeout", NULL, cmd_spec_timeout },
240 { "timestamping", &opt.timestamping, cmd_boolean },
241 { "tries", &opt.ntry, cmd_number_inf },
242 { "useproxy", &opt.use_proxy, cmd_boolean },
243 { "user", &opt.user, cmd_string },
244 { "useragent", NULL, cmd_spec_useragent },
245 { "verbose", NULL, cmd_spec_verbose },
246 { "wait", &opt.wait, cmd_time },
247 { "waitretry", &opt.waitretry, cmd_time },
249 { "wdebug", &opt.wdebug, cmd_boolean },
253 /* Look up CMDNAME in the commands[] and return its position in the
254 array. If CMDNAME is not found, return -1. */
257 command_by_name (const char *cmdname)
259 /* Use binary search for speed. Wget has ~100 commands, which
260 guarantees a worst case performance of 7 string comparisons. */
261 int lo = 0, hi = countof (commands) - 1;
265 int mid = (lo + hi) >> 1;
266 int cmp = strcasecmp (cmdname, commands[mid].name);
277 /* Reset the variables to default values. */
283 /* Most of the default values are 0 (and 0.0, NULL, and false).
284 Just reset everything, and fill in the non-zero values. Note
285 that initializing pointers to NULL this way is technically
286 illegal, but porting Wget to a machine where NULL is not all-zero
287 bit pattern will be the least of the implementors' worries. */
294 opt.add_hostdir = true;
298 opt.http_keep_alive = true;
299 opt.use_proxy = true;
300 tmp = getenv ("no_proxy");
302 opt.no_proxy = sepstring (tmp);
303 opt.allow_cache = true;
305 opt.read_timeout = 900;
306 opt.use_robots = true;
308 opt.remove_listing = true;
310 opt.dot_bytes = 1024;
311 opt.dot_spacing = 10;
312 opt.dots_in_line = 50;
314 opt.dns_cache = true;
318 opt.check_cert = true;
321 /* The default for file name restriction defaults to the OS type. */
322 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
323 opt.restrict_files_os = restrict_windows;
325 opt.restrict_files_os = restrict_unix;
327 opt.restrict_files_ctrl = true;
328 opt.restrict_files_case = restrict_no_case_restriction;
330 opt.max_redirect = 20;
333 /* Return the user's home directory (strdup-ed), or NULL if none is
338 char *home = getenv ("HOME");
343 /* Under MSDOS, if $HOME isn't defined, use the directory where
344 `wget.exe' resides. */
345 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
346 char *p, buf[PATH_MAX];
348 strcpy (buf, _w32_get_argv0 ());
349 p = strrchr (buf, '/'); /* djgpp */
351 p = strrchr (buf, '\\'); /* others */
355 #elif !defined(WINDOWS)
356 /* If HOME is not defined, try getting it from the password
358 struct passwd *pwd = getpwuid (getuid ());
359 if (!pwd || !pwd->pw_dir)
363 /* Under Windows, if $HOME isn't defined, use the directory where
364 `wget.exe' resides. */
369 return home ? xstrdup (home) : NULL;
372 /* Return the path to the user's .wgetrc. This is either the value of
373 `WGETRC' environment variable, or `$HOME/.wgetrc'.
375 If the `WGETRC' variable exists but the file does not exist, the
376 function will exit(). */
378 wgetrc_file_name (void)
383 /* Try the environment. */
384 env = getenv ("WGETRC");
387 if (!file_exists_p (env))
389 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
393 return xstrdup (env);
396 /* If that failed, try $HOME/.wgetrc. */
399 file = aprintf ("%s/.wgetrc", home);
403 /* Under Windows, if we still haven't found .wgetrc, look for the file
404 `wget.ini' in the directory where `wget.exe' resides; we do this for
405 backward compatibility with previous versions of Wget.
406 SYSTEM_WGETRC should not be defined under WINDOWS. */
407 if (!file || !file_exists_p (file))
413 file = aprintf ("%s/wget.ini", home);
419 if (!file_exists_p (file))
427 /* Return values of parse_line. */
435 static enum parse_line parse_line (const char *, char **, char **, int *);
436 static bool setval_internal (int, const char *, const char *);
438 /* Initialize variables from a wgetrc file. Returns zero (failure) if
439 there were errors in the file. */
442 run_wgetrc (const char *file)
449 fp = fopen (file, "rb");
452 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
453 file, strerror (errno));
454 return true; /* not a fatal error */
456 enable_tilde_expansion = true;
458 while ((line = read_whole_line (fp)) != NULL)
460 char *com = NULL, *val = NULL;
463 /* Parse the line. */
464 switch (parse_line (line, &com, &val, &comind))
467 /* If everything is OK, set the value. */
468 if (!setval_internal (comind, com, val))
470 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
471 exec_name, file, ln);
475 case line_syntax_error:
476 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
477 exec_name, file, ln);
480 case line_unknown_command:
481 fprintf (stderr, _("%s: Unknown command `%s' in %s at line %d.\n"),
482 exec_name, com, file, ln);
495 enable_tilde_expansion = false;
501 /* Initialize the defaults and run the system wgetrc and user's own
509 /* Load the hard-coded defaults. */
512 /* If SYSTEM_WGETRC is defined, use it. */
514 if (file_exists_p (SYSTEM_WGETRC))
515 ok &= run_wgetrc (SYSTEM_WGETRC);
517 /* Override it with your own, if one exists. */
518 file = wgetrc_file_name ();
521 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
522 something like realpath() before comparing them with `strcmp' */
524 if (!strcmp (file, SYSTEM_WGETRC))
526 fprintf (stderr, _("\
527 %s: Warning: Both system and user wgetrc point to `%s'.\n"),
532 ok &= run_wgetrc (file);
534 /* If there were errors processing either `.wgetrc', abort. */
542 /* Remove dashes and underscores from S, modifying S in the
548 char *t = s; /* t - tortoise */
549 char *h = s; /* h - hare */
551 if (*h == '_' || *h == '-')
558 /* Parse the line pointed by line, with the syntax:
559 <sp>* command <sp>* = <sp>* value <sp>*
560 Uses malloc to allocate space for command and value.
562 Returns one of line_ok, line_empty, line_syntax_error, or
563 line_unknown_command.
565 In case of line_ok, *COM and *VAL point to freshly allocated
566 strings, and *COMIND points to com's index. In case of error or
567 empty line, their values are unmodified. */
569 static enum parse_line
570 parse_line (const char *line, char **com, char **val, int *comind)
573 const char *end = line + strlen (line);
574 const char *cmdstart, *cmdend;
575 const char *valstart, *valend;
580 /* Skip leading and trailing whitespace. */
581 while (*line && c_isspace (*line))
583 while (end > line && c_isspace (end[-1]))
586 /* Skip empty lines and comments. */
587 if (!*line || *line == '#')
593 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
597 /* Skip '=', as well as any space before or after it. */
598 while (p < end && c_isspace (*p))
600 if (p == end || *p != '=')
601 return line_syntax_error;
603 while (p < end && c_isspace (*p))
609 /* The syntax is valid (even though the command might not be). Fill
610 in the command name and value. */
611 *com = strdupdelim (cmdstart, cmdend);
612 *val = strdupdelim (valstart, valend);
614 /* The line now known to be syntactically correct. Check whether
615 the command is valid. */
616 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
618 ind = command_by_name (cmdcopy);
620 return line_unknown_command;
622 /* Report success to the caller. */
627 /* Run commands[comind].action. */
630 setval_internal (int comind, const char *com, const char *val)
632 assert (0 <= comind && comind < countof (commands));
633 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
634 return commands[comind].action (com, val, commands[comind].place);
637 /* Run command COM with value VAL. If running the command produces an
638 error, report the error and exit.
640 This is intended to be called from main() to modify Wget's behavior
641 through command-line switches. Since COM is hard-coded in main(),
642 it is not canonicalized, and this aborts when COM is not found.
644 If COMIND's are exported to init.h, this function will be changed
645 to accept COMIND directly. */
648 setoptval (const char *com, const char *val, const char *optname)
650 /* Prepend "--" to OPTNAME. */
651 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
654 strcpy (dd_optname + 2, optname);
656 assert (val != NULL);
657 if (!setval_internal (command_by_name (com), dd_optname, val))
661 /* Parse OPT into command and value and run it. For example,
662 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
663 This is used by the `--execute' flag in main.c. */
666 run_command (const char *opt)
670 switch (parse_line (opt, &com, &val, &comind))
673 if (!setval_internal (comind, com, val))
679 fprintf (stderr, _("%s: Invalid --execute command `%s'\n"),
685 /* Generic helper functions, for use with `commands'. */
687 /* Forward declarations: */
692 static bool decode_string (const char *, const struct decode_item *, int, int *);
693 static bool simple_atoi (const char *, const char *, int *);
694 static bool simple_atof (const char *, const char *, double *);
696 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
698 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
699 && c_tolower((p)[1]) == (c1) \
702 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
703 && c_tolower((p)[1]) == (c1) \
704 && c_tolower((p)[2]) == (c2) \
708 /* Store the boolean value from VAL to PLACE. COM is ignored,
709 except for error messages. */
711 cmd_boolean (const char *com, const char *val, void *place)
715 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
716 /* "on", "yes" and "1" mean true. */
718 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
719 /* "off", "no" and "0" mean false. */
724 _("%s: %s: Invalid boolean `%s'; use `on' or `off'.\n"),
725 exec_name, com, val);
729 *(bool *) place = value;
733 /* Set the non-negative integer value from VAL to PLACE. With
734 incorrect specification, the number remains unchanged. */
736 cmd_number (const char *com, const char *val, void *place)
738 if (!simple_atoi (val, val + strlen (val), place)
739 || *(int *) place < 0)
741 fprintf (stderr, _("%s: %s: Invalid number `%s'.\n"),
742 exec_name, com, val);
748 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
750 cmd_number_inf (const char *com, const char *val, void *place)
752 if (!strcasecmp (val, "inf"))
757 return cmd_number (com, val, place);
760 /* Copy (strdup) the string at COM to a new location and place a
761 pointer to *PLACE. */
763 cmd_string (const char *com, const char *val, void *place)
765 char **pstring = (char **)place;
767 xfree_null (*pstring);
768 *pstring = xstrdup (val);
772 #if defined(WINDOWS) || defined(MSDOS)
773 # define ISSEP(c) ((c) == '/' || (c) == '\\')
775 # define ISSEP(c) ((c) == '/')
778 /* Like the above, but handles tilde-expansion when reading a user's
779 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
780 gets expanded to the user's home directory. */
782 cmd_file (const char *com, const char *val, void *place)
784 char **pstring = (char **)place;
786 xfree_null (*pstring);
788 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
790 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
793 *pstring = xstrdup (val);
798 char *home = home_dir ();
802 homelen = strlen (home);
803 while (homelen && ISSEP (home[homelen - 1]))
804 home[--homelen] = '\0';
806 /* Skip the leading "~/". */
807 for (++val; ISSEP (*val); val++)
810 *pstring = concat_strings (home, "/", val, (char *) 0);
813 #if defined(WINDOWS) || defined(MSDOS)
814 /* Convert "\" to "/". */
817 for (s = *pstring; *s; s++)
825 /* Like cmd_file, but strips trailing '/' characters. */
827 cmd_directory (const char *com, const char *val, void *place)
831 /* Call cmd_file() for tilde expansion and separator
832 canonicalization (backslash -> slash under Windows). These
833 things should perhaps be in a separate function. */
834 if (!cmd_file (com, val, place))
839 while (t > s && *--t == '/')
845 /* Split VAL by space to a vector of values, and append those values
846 to vector pointed to by the PLACE argument. If VAL is empty, the
847 PLACE vector is cleared instead. */
850 cmd_vector (const char *com, const char *val, void *place)
852 char ***pvec = (char ***)place;
855 *pvec = merge_vecs (*pvec, sepstring (val));
865 cmd_directory_vector (const char *com, const char *val, void *place)
867 char ***pvec = (char ***)place;
871 /* Strip the trailing slashes from directories. */
874 seps = sepstring (val);
875 for (t = seps; t && *t; t++)
877 int len = strlen (*t);
878 /* Skip degenerate case of root directory. */
881 if ((*t)[len - 1] == '/')
882 (*t)[len - 1] = '\0';
885 *pvec = merge_vecs (*pvec, seps);
895 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
896 "100k" or "2.5G" to a floating point number. */
899 parse_bytes_helper (const char *val, double *result)
902 const char *end = val + strlen (val);
904 /* Check for "inf". */
905 if (0 == strcmp (val, "inf"))
911 /* Strip trailing whitespace. */
912 while (val < end && c_isspace (end[-1]))
917 switch (c_tolower (end[-1]))
920 --end, mult = 1024.0;
923 --end, mult = 1048576.0;
926 --end, mult = 1073741824.0;
929 --end, mult = 1099511627776.0;
932 /* Not a recognized suffix: assume it's a digit. (If not,
933 simple_atof will raise an error.) */
937 /* Skip leading and trailing whitespace. */
938 while (val < end && c_isspace (*val))
940 while (val < end && c_isspace (end[-1]))
945 if (!simple_atof (val, end, &number) || number < 0)
948 *result = number * mult;
952 /* Parse VAL as a number and set its value to PLACE (which should
955 By default, the value is assumed to be in bytes. If "K", "M", or
956 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
957 1<<30, respectively. Floating point values are allowed and are
958 cast to integer before use. The idea is to be able to use things
959 like 1.5k instead of "1536".
961 The string "inf" is returned as 0.
963 In case of error, false is returned and memory pointed to by PLACE
964 remains unmodified. */
967 cmd_bytes (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 *(wgint *)place = (wgint)byte_value;
980 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
981 SIZE_SUM. It works by converting the string to double, therefore
982 working with values up to 2^53-1 without loss of precision. This
983 value (8192 TB) is large enough to serve for a while. */
986 cmd_bytes_sum (const char *com, const char *val, void *place)
989 if (!parse_bytes_helper (val, &byte_value))
991 fprintf (stderr, _("%s: %s: Invalid byte value `%s'\n"),
992 exec_name, com, val);
995 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
999 /* Store the value of VAL to *OUT. The value is a time period, by
1000 default expressed in seconds, but also accepting suffixes "m", "h",
1001 "d", and "w" for minutes, hours, days, and weeks respectively. */
1004 cmd_time (const char *com, const char *val, void *place)
1006 double number, mult;
1007 const char *end = val + strlen (val);
1009 /* Strip trailing whitespace. */
1010 while (val < end && c_isspace (end[-1]))
1016 fprintf (stderr, _("%s: %s: Invalid time period `%s'\n"),
1017 exec_name, com, val);
1021 switch (c_tolower (end[-1]))
1024 --end, mult = 1; /* seconds */
1027 --end, mult = 60; /* minutes */
1030 --end, mult = 3600; /* hours */
1033 --end, mult = 86400.0; /* days */
1036 --end, mult = 604800.0; /* weeks */
1039 /* Not a recognized suffix: assume it belongs to the number.
1040 (If not, simple_atof will raise an error.) */
1044 /* Skip leading and trailing whitespace. */
1045 while (val < end && c_isspace (*val))
1047 while (val < end && c_isspace (end[-1]))
1052 if (!simple_atof (val, end, &number))
1055 *(double *)place = number * mult;
1061 cmd_cert_type (const char *com, const char *val, void *place)
1063 static const struct decode_item choices[] = {
1064 { "pem", keyfile_pem },
1065 { "der", keyfile_asn1 },
1066 { "asn1", keyfile_asn1 },
1068 int ok = decode_string (val, choices, countof (choices), place);
1070 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1075 /* Specialized helper functions, used by `commands' to handle some
1076 options specially. */
1078 static bool check_user_specified_header (const char *);
1081 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1083 if (!cmd_boolean (com, val, &opt.dirstruct))
1085 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1086 must be affected inversely. */
1088 opt.no_dirstruct = false;
1090 opt.no_dirstruct = true;
1095 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1097 /* Empty value means reset the list of headers. */
1100 free_vec (opt.user_headers);
1101 opt.user_headers = NULL;
1105 if (!check_user_specified_header (val))
1107 fprintf (stderr, _("%s: %s: Invalid header `%s'.\n"),
1108 exec_name, com, val);
1111 opt.user_headers = vec_append (opt.user_headers, val);
1116 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1118 int flag = cmd_boolean (com, val, &opt.htmlify);
1119 if (flag && !opt.htmlify)
1120 opt.remove_listing = false;
1124 /* Set the "mirror" mode. It means: recursive download, timestamping,
1125 no limit on max. recursion depth, and don't remove listings. */
1128 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1132 if (!cmd_boolean (com, val, &mirror))
1136 opt.recursive = true;
1137 if (!opt.no_dirstruct)
1138 opt.dirstruct = true;
1139 opt.timestamping = true;
1140 opt.reclevel = INFINITE_RECURSION;
1141 opt.remove_listing = false;
1146 /* Validate --prefer-family and set the choice. Allowed values are
1147 "IPv4", "IPv6", and "none". */
1150 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1152 static const struct decode_item choices[] = {
1153 { "IPv4", prefer_ipv4 },
1154 { "IPv6", prefer_ipv6 },
1155 { "none", prefer_none },
1157 int prefer_family = prefer_ipv4;
1158 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1160 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1161 opt.prefer_family = prefer_family;
1165 /* Set progress.type to VAL, but verify that it's a valid progress
1166 implementation before that. */
1169 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1171 if (!valid_progress_implementation_p (val))
1173 fprintf (stderr, _("%s: %s: Invalid progress type `%s'.\n"),
1174 exec_name, com, val);
1177 xfree_null (opt.progress_type);
1179 /* Don't call set_progress_implementation here. It will be called
1180 in main() when it becomes clear what the log output is. */
1181 opt.progress_type = xstrdup (val);
1185 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1186 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1190 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1192 if (!cmd_boolean (com, val, &opt.recursive))
1196 if (opt.recursive && !opt.no_dirstruct)
1197 opt.dirstruct = true;
1203 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1205 int restrict_os = opt.restrict_files_os;
1206 int restrict_ctrl = opt.restrict_files_ctrl;
1207 int restrict_case = opt.restrict_files_case;
1211 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1215 end = strchr (val, ',');
1217 end = val + strlen (val);
1219 if (VAL_IS ("unix"))
1220 restrict_os = restrict_unix;
1221 else if (VAL_IS ("windows"))
1222 restrict_os = restrict_windows;
1223 else if (VAL_IS ("lowercase"))
1224 restrict_case = restrict_lowercase;
1225 else if (VAL_IS ("uppercase"))
1226 restrict_case = restrict_uppercase;
1227 else if (VAL_IS ("nocontrol"))
1228 restrict_ctrl = false;
1232 _("%s: %s: Invalid restriction `%s', use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1233 exec_name, com, val);
1240 while (*val && *end);
1244 opt.restrict_files_os = restrict_os;
1245 opt.restrict_files_ctrl = restrict_ctrl;
1246 opt.restrict_files_case = restrict_case;
1253 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1255 static const struct decode_item choices[] = {
1256 { "auto", secure_protocol_auto },
1257 { "sslv2", secure_protocol_sslv2 },
1258 { "sslv3", secure_protocol_sslv3 },
1259 { "tlsv1", secure_protocol_tlsv1 },
1261 int ok = decode_string (val, choices, countof (choices), place);
1263 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"), exec_name, com, val);
1268 /* Set all three timeout values. */
1271 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1274 if (!cmd_time (com, val, &value))
1276 opt.read_timeout = value;
1277 opt.connect_timeout = value;
1278 opt.dns_timeout = value;
1283 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1285 /* Disallow embedded newlines. */
1286 if (strchr (val, '\n'))
1288 fprintf (stderr, _("%s: %s: Invalid value `%s'.\n"),
1289 exec_name, com, val);
1292 xfree_null (opt.useragent);
1293 opt.useragent = xstrdup (val);
1297 /* The "verbose" option cannot be cmd_boolean because the variable is
1298 not bool -- it's of type int (-1 means uninitialized because of
1299 some random hackery for disallowing -q -v). */
1302 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1305 if (cmd_boolean (com, val, &flag))
1313 /* Miscellaneous useful routines. */
1315 /* A very simple atoi clone, more useful than atoi because it works on
1316 delimited strings, and has error reportage. Returns true on success,
1317 false on failure. If successful, stores result to *DEST. */
1320 simple_atoi (const char *beg, const char *end, int *dest)
1323 bool negative = false;
1324 const char *p = beg;
1326 while (p < end && c_isspace (*p))
1328 if (p < end && (*p == '-' || *p == '+'))
1330 negative = (*p == '-');
1336 /* Read negative numbers in a separate loop because the most
1337 negative integer cannot be represented as a positive number. */
1340 for (; p < end && c_isdigit (*p); p++)
1342 int next = (10 * result) + (*p - '0');
1344 return false; /* overflow */
1348 for (; p < end && c_isdigit (*p); p++)
1350 int next = (10 * result) - (*p - '0');
1352 return false; /* underflow */
1363 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1364 doesn't handle exponential notation. Returns true on success,
1365 false on failure. In case of success, stores its result to
1369 simple_atof (const char *beg, const char *end, double *dest)
1373 bool negative = false;
1374 bool seen_dot = false;
1375 bool seen_digit = false;
1378 const char *p = beg;
1380 while (p < end && c_isspace (*p))
1382 if (p < end && (*p == '-' || *p == '+'))
1384 negative = (*p == '-');
1388 for (; p < end; p++)
1394 result = (10 * result) + (ch - '0');
1396 result += (ch - '0') / (divider *= 10);
1418 /* Verify that the user-specified header in S is valid. It must
1419 contain a colon preceded by non-white-space characters and must not
1420 contain newlines. */
1423 check_user_specified_header (const char *s)
1427 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1429 /* The header MUST contain `:' preceded by at least one
1430 non-whitespace character. */
1431 if (*p != ':' || p == s)
1433 /* The header MUST NOT contain newlines. */
1434 if (strchr (s, '\n'))
1439 /* Decode VAL into a number, according to ITEMS. */
1442 decode_string (const char *val, const struct decode_item *items, int itemcount,
1446 for (i = 0; i < itemcount; i++)
1447 if (0 == strcasecmp (val, items[i].name))
1449 *place = items[i].code;
1456 void cleanup_html_url (void);
1459 /* Free the memory allocated by global variables. */
1463 /* Free external resources, close files, etc. */
1466 fclose (output_stream);
1467 /* No need to check for error because Wget flushes its output (and
1468 checks for errors) after any data arrives. */
1470 /* We're exiting anyway so there's no real need to call free()
1471 hundreds of times. Skipping the frees will make Wget exit
1474 However, when detecting leaks, it's crucial to free() everything
1475 because then you can find the real leaks, i.e. the allocated
1476 memory which grows with the size of the program. */
1482 cleanup_html_url ();
1487 extern acc_t *netrc_list;
1488 free_netrc (netrc_list);
1490 xfree_null (opt.lfilename);
1491 xfree_null (opt.dir_prefix);
1492 xfree_null (opt.input_filename);
1493 xfree_null (opt.output_document);
1494 free_vec (opt.accepts);
1495 free_vec (opt.rejects);
1496 free_vec (opt.excludes);
1497 free_vec (opt.includes);
1498 free_vec (opt.domains);
1499 free_vec (opt.follow_tags);
1500 free_vec (opt.ignore_tags);
1501 xfree_null (opt.progress_type);
1502 xfree_null (opt.ftp_user);
1503 xfree_null (opt.ftp_passwd);
1504 xfree_null (opt.ftp_proxy);
1505 xfree_null (opt.https_proxy);
1506 xfree_null (opt.http_proxy);
1507 free_vec (opt.no_proxy);
1508 xfree_null (opt.useragent);
1509 xfree_null (opt.referer);
1510 xfree_null (opt.http_user);
1511 xfree_null (opt.http_passwd);
1512 free_vec (opt.user_headers);
1514 xfree_null (opt.cert_file);
1515 xfree_null (opt.private_key);
1516 xfree_null (opt.ca_directory);
1517 xfree_null (opt.ca_cert);
1518 xfree_null (opt.random_file);
1519 xfree_null (opt.egd_file);
1521 xfree_null (opt.bind_address);
1522 xfree_null (opt.cookies_input);
1523 xfree_null (opt.cookies_output);
1524 xfree_null (opt.user);
1525 xfree_null (opt.passwd);
1526 #endif /* DEBUG_MALLOC */
1529 /* Unit testing routines. */
1534 test_cmd_spec_restrict_file_names()
1539 int expected_restrict_files_os;
1540 int expected_restrict_files_ctrl;
1541 int expected_restrict_files_case;
1544 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1545 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1546 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1547 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1550 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1555 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1558 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1559 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1560 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1561 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1563 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1564 res == test_array[i].result
1565 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1566 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1567 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1573 #endif /* TESTING */