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. */
31 #define USE_GNULIB_ALLOC
53 #include "recur.h" /* for INFINITE_RECURSION */
54 #include "convert.h" /* for convert_cleanup */
55 #include "res.h" /* for res_cleanup */
56 #include "http.h" /* for http_cleanup */
57 #include "retr.h" /* for output_stream */
63 /* We want tilde expansion enabled only when reading `.wgetrc' lines;
64 otherwise, it will be performed by the shell. This variable will
65 be set by the wgetrc-reading function. */
67 static bool enable_tilde_expansion;
70 #define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
72 CMD_DECLARE (cmd_boolean);
73 CMD_DECLARE (cmd_bytes);
74 CMD_DECLARE (cmd_bytes_sum);
76 CMD_DECLARE (cmd_cert_type);
78 CMD_DECLARE (cmd_directory_vector);
79 CMD_DECLARE (cmd_number);
80 CMD_DECLARE (cmd_number_inf);
81 CMD_DECLARE (cmd_string);
82 CMD_DECLARE (cmd_file);
83 CMD_DECLARE (cmd_directory);
84 CMD_DECLARE (cmd_time);
85 CMD_DECLARE (cmd_vector);
87 CMD_DECLARE (cmd_spec_dirstruct);
88 CMD_DECLARE (cmd_spec_header);
89 CMD_DECLARE (cmd_spec_htmlify);
90 CMD_DECLARE (cmd_spec_mirror);
91 CMD_DECLARE (cmd_spec_prefer_family);
92 CMD_DECLARE (cmd_spec_progress);
93 CMD_DECLARE (cmd_spec_recursive);
94 CMD_DECLARE (cmd_spec_restrict_file_names);
96 CMD_DECLARE (cmd_spec_secure_protocol);
98 CMD_DECLARE (cmd_spec_timeout);
99 CMD_DECLARE (cmd_spec_useragent);
100 CMD_DECLARE (cmd_spec_verbose);
102 /* List of recognized commands, each consisting of name, place and
103 function. When adding a new command, simply add it to the list,
104 but be sure to keep the list sorted alphabetically, as
105 command_by_name's binary search depends on it. Also, be sure to
106 add any entries that allocate memory (e.g. cmd_string and
107 cmd_vector) to the cleanup() function below. */
109 static const struct {
112 bool (*action) (const char *, const char *, void *);
114 /* KEEP THIS LIST ALPHABETICALLY SORTED */
115 { "accept", &opt.accepts, cmd_vector },
116 { "addhostdir", &opt.add_hostdir, cmd_boolean },
117 { "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
118 { "authnochallenge", &opt.auth_without_challenge,
120 { "background", &opt.background, cmd_boolean },
121 { "backupconverted", &opt.backup_converted, cmd_boolean },
122 { "backups", &opt.backups, cmd_number },
123 { "base", &opt.base_href, cmd_string },
124 { "bindaddress", &opt.bind_address, cmd_string },
126 { "cacertificate", &opt.ca_cert, cmd_file },
128 { "cache", &opt.allow_cache, cmd_boolean },
130 { "cadirectory", &opt.ca_directory, cmd_directory },
131 { "certificate", &opt.cert_file, cmd_file },
132 { "certificatetype", &opt.cert_type, cmd_cert_type },
133 { "checkcertificate", &opt.check_cert, cmd_boolean },
135 { "connecttimeout", &opt.connect_timeout, cmd_time },
136 { "contentdisposition", &opt.content_disposition, cmd_boolean },
137 { "continue", &opt.always_rest, cmd_boolean },
138 { "convertlinks", &opt.convert_links, cmd_boolean },
139 { "cookies", &opt.cookies, cmd_boolean },
140 { "cutdirs", &opt.cut_dirs, cmd_number },
142 { "debug", &opt.debug, cmd_boolean },
144 { "deleteafter", &opt.delete_after, cmd_boolean },
145 { "dirprefix", &opt.dir_prefix, cmd_directory },
146 { "dirstruct", NULL, cmd_spec_dirstruct },
147 { "dnscache", &opt.dns_cache, cmd_boolean },
148 { "dnstimeout", &opt.dns_timeout, cmd_time },
149 { "domains", &opt.domains, cmd_vector },
150 { "dotbytes", &opt.dot_bytes, cmd_bytes },
151 { "dotsinline", &opt.dots_in_line, cmd_number },
152 { "dotspacing", &opt.dot_spacing, cmd_number },
153 { "dotstyle", &opt.dot_style, cmd_string },
155 { "egdfile", &opt.egd_file, cmd_file },
157 { "excludedirectories", &opt.excludes, cmd_directory_vector },
158 { "excludedomains", &opt.exclude_domains, cmd_vector },
159 { "followftp", &opt.follow_ftp, cmd_boolean },
160 { "followtags", &opt.follow_tags, cmd_vector },
161 { "forcehtml", &opt.force_html, cmd_boolean },
162 { "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
163 { "ftppassword", &opt.ftp_passwd, cmd_string },
164 { "ftpproxy", &opt.ftp_proxy, cmd_string },
165 { "ftpuser", &opt.ftp_user, cmd_string },
166 { "glob", &opt.ftp_glob, cmd_boolean },
167 { "header", NULL, cmd_spec_header },
168 { "htmlextension", &opt.html_extension, cmd_boolean },
169 { "htmlify", NULL, cmd_spec_htmlify },
170 { "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
171 { "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
172 { "httppassword", &opt.http_passwd, cmd_string },
173 { "httpproxy", &opt.http_proxy, cmd_string },
174 { "httpsproxy", &opt.https_proxy, cmd_string },
175 { "httpuser", &opt.http_user, cmd_string },
176 { "ignorecase", &opt.ignore_case, cmd_boolean },
177 { "ignorelength", &opt.ignore_length, cmd_boolean },
178 { "ignoretags", &opt.ignore_tags, cmd_vector },
179 { "includedirectories", &opt.includes, cmd_directory_vector },
181 { "inet4only", &opt.ipv4_only, cmd_boolean },
182 { "inet6only", &opt.ipv6_only, cmd_boolean },
184 { "input", &opt.input_filename, cmd_file },
185 { "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186 { "limitrate", &opt.limit_rate, cmd_bytes },
187 { "loadcookies", &opt.cookies_input, cmd_file },
188 { "logfile", &opt.lfilename, cmd_file },
189 { "login", &opt.ftp_user, cmd_string },/* deprecated*/
190 { "maxredirect", &opt.max_redirect, cmd_number },
191 { "mirror", NULL, cmd_spec_mirror },
192 { "netrc", &opt.netrc, cmd_boolean },
193 { "noclobber", &opt.noclobber, cmd_boolean },
194 { "noparent", &opt.no_parent, cmd_boolean },
195 { "noproxy", &opt.no_proxy, cmd_vector },
196 { "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
197 { "outputdocument", &opt.output_document, cmd_file },
198 { "pagerequisites", &opt.page_requisites, cmd_boolean },
199 { "passiveftp", &opt.ftp_pasv, cmd_boolean },
200 { "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
201 { "password", &opt.passwd, cmd_string },
202 { "postdata", &opt.post_data, cmd_string },
203 { "postfile", &opt.post_file_name, cmd_file },
204 { "preferfamily", NULL, cmd_spec_prefer_family },
205 { "preservepermissions", &opt.preserve_perm, cmd_boolean },
207 { "privatekey", &opt.private_key, cmd_file },
208 { "privatekeytype", &opt.private_key_type, cmd_cert_type },
210 { "progress", &opt.progress_type, cmd_spec_progress },
211 { "protocoldirectories", &opt.protocol_directories, cmd_boolean },
212 { "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
213 { "proxypassword", &opt.proxy_passwd, cmd_string },
214 { "proxyuser", &opt.proxy_user, cmd_string },
215 { "quiet", &opt.quiet, cmd_boolean },
216 { "quota", &opt.quota, cmd_bytes_sum },
218 { "randomfile", &opt.random_file, cmd_file },
220 { "randomwait", &opt.random_wait, cmd_boolean },
221 { "readtimeout", &opt.read_timeout, cmd_time },
222 { "reclevel", &opt.reclevel, cmd_number_inf },
223 { "recursive", NULL, cmd_spec_recursive },
224 { "referer", &opt.referer, cmd_string },
225 { "reject", &opt.rejects, cmd_vector },
226 { "relativeonly", &opt.relative_only, cmd_boolean },
227 { "removelisting", &opt.remove_listing, cmd_boolean },
228 { "restrictfilenames", NULL, cmd_spec_restrict_file_names },
229 { "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
230 { "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
231 { "robots", &opt.use_robots, cmd_boolean },
232 { "savecookies", &opt.cookies_output, cmd_file },
233 { "saveheaders", &opt.save_headers, cmd_boolean },
235 { "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
237 { "serverresponse", &opt.server_response, cmd_boolean },
238 { "spanhosts", &opt.spanhost, cmd_boolean },
239 { "spider", &opt.spider, cmd_boolean },
240 { "strictcomments", &opt.strict_comments, cmd_boolean },
241 { "timeout", NULL, cmd_spec_timeout },
242 { "timestamping", &opt.timestamping, cmd_boolean },
243 { "tries", &opt.ntry, cmd_number_inf },
244 { "useproxy", &opt.use_proxy, cmd_boolean },
245 { "user", &opt.user, cmd_string },
246 { "useragent", NULL, cmd_spec_useragent },
247 { "verbose", NULL, cmd_spec_verbose },
248 { "wait", &opt.wait, cmd_time },
249 { "waitretry", &opt.waitretry, cmd_time },
251 { "wdebug", &opt.wdebug, cmd_boolean },
255 /* Look up CMDNAME in the commands[] and return its position in the
256 array. If CMDNAME is not found, return -1. */
259 command_by_name (const char *cmdname)
261 /* Use binary search for speed. Wget has ~100 commands, which
262 guarantees a worst case performance of 7 string comparisons. */
263 int lo = 0, hi = countof (commands) - 1;
267 int mid = (lo + hi) >> 1;
268 int cmp = strcasecmp (cmdname, commands[mid].name);
279 /* Reset the variables to default values. */
285 /* Most of the default values are 0 (and 0.0, NULL, and false).
286 Just reset everything, and fill in the non-zero values. Note
287 that initializing pointers to NULL this way is technically
288 illegal, but porting Wget to a machine where NULL is not all-zero
289 bit pattern will be the least of the implementors' worries. */
296 opt.add_hostdir = true;
300 opt.http_keep_alive = true;
301 opt.use_proxy = true;
302 tmp = getenv ("no_proxy");
304 opt.no_proxy = sepstring (tmp);
305 opt.allow_cache = true;
307 opt.read_timeout = 900;
308 opt.use_robots = true;
310 opt.remove_listing = true;
312 opt.dot_bytes = 1024;
313 opt.dot_spacing = 10;
314 opt.dots_in_line = 50;
316 opt.dns_cache = true;
320 opt.check_cert = true;
323 /* The default for file name restriction defaults to the OS type. */
324 #if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
325 opt.restrict_files_os = restrict_windows;
327 opt.restrict_files_os = restrict_unix;
329 opt.restrict_files_ctrl = true;
330 opt.restrict_files_case = restrict_no_case_restriction;
332 opt.max_redirect = 20;
335 /* Return the user's home directory (strdup-ed), or NULL if none is
340 char *home = getenv ("HOME");
345 /* Under MSDOS, if $HOME isn't defined, use the directory where
346 `wget.exe' resides. */
347 const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
348 char *p, buf[PATH_MAX];
350 strcpy (buf, _w32_get_argv0 ());
351 p = strrchr (buf, '/'); /* djgpp */
353 p = strrchr (buf, '\\'); /* others */
357 #elif !defined(WINDOWS)
358 /* If HOME is not defined, try getting it from the password
360 struct passwd *pwd = getpwuid (getuid ());
361 if (!pwd || !pwd->pw_dir)
365 /* Under Windows, if $HOME isn't defined, use the directory where
366 `wget.exe' resides. */
371 return home ? xstrdup (home) : NULL;
374 /* Return the path to the user's .wgetrc. This is either the value of
375 `WGETRC' environment variable, or `$HOME/.wgetrc'.
377 If the `WGETRC' variable exists but the file does not exist, the
378 function will exit(). */
380 wgetrc_file_name (void)
385 /* Try the environment. */
386 env = getenv ("WGETRC");
389 if (!file_exists_p (env))
391 fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
395 return xstrdup (env);
398 /* If that failed, try $HOME/.wgetrc. */
401 file = aprintf ("%s/.wgetrc", home);
405 /* Under Windows, if we still haven't found .wgetrc, look for the file
406 `wget.ini' in the directory where `wget.exe' resides; we do this for
407 backward compatibility with previous versions of Wget.
408 SYSTEM_WGETRC should not be defined under WINDOWS. */
409 if (!file || !file_exists_p (file))
415 file = aprintf ("%s/wget.ini", home);
421 if (!file_exists_p (file))
429 /* Return values of parse_line. */
437 static enum parse_line parse_line (const char *, char **, char **, int *);
438 static bool setval_internal (int, const char *, const char *);
440 /* Initialize variables from a wgetrc file. Returns zero (failure) if
441 there were errors in the file. */
444 run_wgetrc (const char *file)
451 fp = fopen (file, "rb");
454 fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
455 file, strerror (errno));
456 return true; /* not a fatal error */
458 enable_tilde_expansion = true;
460 while ((line = read_whole_line (fp)) != NULL)
462 char *com = NULL, *val = NULL;
465 /* Parse the line. */
466 switch (parse_line (line, &com, &val, &comind))
469 /* If everything is OK, set the value. */
470 if (!setval_internal (comind, com, val))
472 fprintf (stderr, _("%s: Error in %s at line %d.\n"),
473 exec_name, file, ln);
477 case line_syntax_error:
478 fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
479 exec_name, file, ln);
482 case line_unknown_command:
483 fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
484 exec_name, quote (com), file, ln);
497 enable_tilde_expansion = false;
503 /* Initialize the defaults and run the system wgetrc and user's own
511 /* Load the hard-coded defaults. */
514 /* If SYSTEM_WGETRC is defined, use it. */
516 if (file_exists_p (SYSTEM_WGETRC))
517 ok &= run_wgetrc (SYSTEM_WGETRC);
519 /* Override it with your own, if one exists. */
520 file = wgetrc_file_name ();
523 /* #### We should canonicalize `file' and SYSTEM_WGETRC with
524 something like realpath() before comparing them with `strcmp' */
526 if (!strcmp (file, SYSTEM_WGETRC))
528 fprintf (stderr, _("\
529 %s: Warning: Both system and user wgetrc point to %s.\n"),
530 exec_name, quote (file));
534 ok &= run_wgetrc (file);
536 /* If there were errors processing either `.wgetrc', abort. */
544 /* Remove dashes and underscores from S, modifying S in the
550 char *t = s; /* t - tortoise */
551 char *h = s; /* h - hare */
553 if (*h == '_' || *h == '-')
560 /* Parse the line pointed by line, with the syntax:
561 <sp>* command <sp>* = <sp>* value <sp>*
562 Uses malloc to allocate space for command and value.
564 Returns one of line_ok, line_empty, line_syntax_error, or
565 line_unknown_command.
567 In case of line_ok, *COM and *VAL point to freshly allocated
568 strings, and *COMIND points to com's index. In case of error or
569 empty line, their values are unmodified. */
571 static enum parse_line
572 parse_line (const char *line, char **com, char **val, int *comind)
575 const char *end = line + strlen (line);
576 const char *cmdstart, *cmdend;
577 const char *valstart, *valend;
582 /* Skip leading and trailing whitespace. */
583 while (*line && c_isspace (*line))
585 while (end > line && c_isspace (end[-1]))
588 /* Skip empty lines and comments. */
589 if (!*line || *line == '#')
595 while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
599 /* Skip '=', as well as any space before or after it. */
600 while (p < end && c_isspace (*p))
602 if (p == end || *p != '=')
603 return line_syntax_error;
605 while (p < end && c_isspace (*p))
611 /* The syntax is valid (even though the command might not be). Fill
612 in the command name and value. */
613 *com = strdupdelim (cmdstart, cmdend);
614 *val = strdupdelim (valstart, valend);
616 /* The line now known to be syntactically correct. Check whether
617 the command is valid. */
618 BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
620 ind = command_by_name (cmdcopy);
622 return line_unknown_command;
624 /* Report success to the caller. */
629 /* Run commands[comind].action. */
632 setval_internal (int comind, const char *com, const char *val)
634 assert (0 <= comind && comind < countof (commands));
635 DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
636 return commands[comind].action (com, val, commands[comind].place);
639 /* Run command COM with value VAL. If running the command produces an
640 error, report the error and exit.
642 This is intended to be called from main() to modify Wget's behavior
643 through command-line switches. Since COM is hard-coded in main(),
644 it is not canonicalized, and this aborts when COM is not found.
646 If COMIND's are exported to init.h, this function will be changed
647 to accept COMIND directly. */
650 setoptval (const char *com, const char *val, const char *optname)
652 /* Prepend "--" to OPTNAME. */
653 char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
656 strcpy (dd_optname + 2, optname);
658 assert (val != NULL);
659 if (!setval_internal (command_by_name (com), dd_optname, val))
663 /* Parse OPT into command and value and run it. For example,
664 run_command("foo=bar") is equivalent to setoptval("foo", "bar").
665 This is used by the `--execute' flag in main.c. */
668 run_command (const char *opt)
672 switch (parse_line (opt, &com, &val, &comind))
675 if (!setval_internal (comind, com, val))
681 fprintf (stderr, _("%s: Invalid --execute command %s\n"),
682 exec_name, quote (opt));
687 /* Generic helper functions, for use with `commands'. */
689 /* Forward declarations: */
694 static bool decode_string (const char *, const struct decode_item *, int, int *);
695 static bool simple_atoi (const char *, const char *, int *);
696 static bool simple_atof (const char *, const char *, double *);
698 #define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
700 #define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
701 && c_tolower((p)[1]) == (c1) \
704 #define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
705 && c_tolower((p)[1]) == (c1) \
706 && c_tolower((p)[2]) == (c2) \
710 /* Store the boolean value from VAL to PLACE. COM is ignored,
711 except for error messages. */
713 cmd_boolean (const char *com, const char *val, void *place)
717 if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
718 /* "on", "yes" and "1" mean true. */
720 else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
721 /* "off", "no" and "0" mean false. */
726 _("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
727 exec_name, com, quote (val));
731 *(bool *) place = value;
735 /* Set the non-negative integer value from VAL to PLACE. With
736 incorrect specification, the number remains unchanged. */
738 cmd_number (const char *com, const char *val, void *place)
740 if (!simple_atoi (val, val + strlen (val), place)
741 || *(int *) place < 0)
743 fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
744 exec_name, com, quote (val));
750 /* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
752 cmd_number_inf (const char *com, const char *val, void *place)
754 if (!strcasecmp (val, "inf"))
759 return cmd_number (com, val, place);
762 /* Copy (strdup) the string at COM to a new location and place a
763 pointer to *PLACE. */
765 cmd_string (const char *com, const char *val, void *place)
767 char **pstring = (char **)place;
769 xfree_null (*pstring);
770 *pstring = xstrdup (val);
774 #if defined(WINDOWS) || defined(MSDOS)
775 # define ISSEP(c) ((c) == '/' || (c) == '\\')
777 # define ISSEP(c) ((c) == '/')
780 /* Like the above, but handles tilde-expansion when reading a user's
781 `.wgetrc'. In that case, and if VAL begins with `~', the tilde
782 gets expanded to the user's home directory. */
784 cmd_file (const char *com, const char *val, void *place)
786 char **pstring = (char **)place;
788 xfree_null (*pstring);
790 /* #### If VAL is empty, perhaps should set *PLACE to NULL. */
792 if (!enable_tilde_expansion || !(*val == '~' && ISSEP (val[1])))
795 *pstring = xstrdup (val);
800 char *home = home_dir ();
804 homelen = strlen (home);
805 while (homelen && ISSEP (home[homelen - 1]))
806 home[--homelen] = '\0';
808 /* Skip the leading "~/". */
809 for (++val; ISSEP (*val); val++)
812 *pstring = concat_strings (home, "/", val, (char *) 0);
815 #if defined(WINDOWS) || defined(MSDOS)
816 /* Convert "\" to "/". */
819 for (s = *pstring; *s; s++)
827 /* Like cmd_file, but strips trailing '/' characters. */
829 cmd_directory (const char *com, const char *val, void *place)
833 /* Call cmd_file() for tilde expansion and separator
834 canonicalization (backslash -> slash under Windows). These
835 things should perhaps be in a separate function. */
836 if (!cmd_file (com, val, place))
841 while (t > s && *--t == '/')
847 /* Split VAL by space to a vector of values, and append those values
848 to vector pointed to by the PLACE argument. If VAL is empty, the
849 PLACE vector is cleared instead. */
852 cmd_vector (const char *com, const char *val, void *place)
854 char ***pvec = (char ***)place;
857 *pvec = merge_vecs (*pvec, sepstring (val));
867 cmd_directory_vector (const char *com, const char *val, void *place)
869 char ***pvec = (char ***)place;
873 /* Strip the trailing slashes from directories. */
876 seps = sepstring (val);
877 for (t = seps; t && *t; t++)
879 int len = strlen (*t);
880 /* Skip degenerate case of root directory. */
883 if ((*t)[len - 1] == '/')
884 (*t)[len - 1] = '\0';
887 *pvec = merge_vecs (*pvec, seps);
897 /* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
898 "100k" or "2.5G" to a floating point number. */
901 parse_bytes_helper (const char *val, double *result)
904 const char *end = val + strlen (val);
906 /* Check for "inf". */
907 if (0 == strcmp (val, "inf"))
913 /* Strip trailing whitespace. */
914 while (val < end && c_isspace (end[-1]))
919 switch (c_tolower (end[-1]))
922 --end, mult = 1024.0;
925 --end, mult = 1048576.0;
928 --end, mult = 1073741824.0;
931 --end, mult = 1099511627776.0;
934 /* Not a recognized suffix: assume it's a digit. (If not,
935 simple_atof will raise an error.) */
939 /* Skip leading and trailing whitespace. */
940 while (val < end && c_isspace (*val))
942 while (val < end && c_isspace (end[-1]))
947 if (!simple_atof (val, end, &number) || number < 0)
950 *result = number * mult;
954 /* Parse VAL as a number and set its value to PLACE (which should
957 By default, the value is assumed to be in bytes. If "K", "M", or
958 "G" are appended, the value is multiplied with 1<<10, 1<<20, or
959 1<<30, respectively. Floating point values are allowed and are
960 cast to integer before use. The idea is to be able to use things
961 like 1.5k instead of "1536".
963 The string "inf" is returned as 0.
965 In case of error, false is returned and memory pointed to by PLACE
966 remains unmodified. */
969 cmd_bytes (const char *com, const char *val, void *place)
972 if (!parse_bytes_helper (val, &byte_value))
974 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
975 exec_name, com, quote (val));
978 *(wgint *)place = (wgint)byte_value;
982 /* Like cmd_bytes, but PLACE is interpreted as a pointer to
983 SIZE_SUM. It works by converting the string to double, therefore
984 working with values up to 2^53-1 without loss of precision. This
985 value (8192 TB) is large enough to serve for a while. */
988 cmd_bytes_sum (const char *com, const char *val, void *place)
991 if (!parse_bytes_helper (val, &byte_value))
993 fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
994 exec_name, com, quote (val));
997 *(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1001 /* Store the value of VAL to *OUT. The value is a time period, by
1002 default expressed in seconds, but also accepting suffixes "m", "h",
1003 "d", and "w" for minutes, hours, days, and weeks respectively. */
1006 cmd_time (const char *com, const char *val, void *place)
1008 double number, mult;
1009 const char *end = val + strlen (val);
1011 /* Strip trailing whitespace. */
1012 while (val < end && c_isspace (end[-1]))
1018 fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1019 exec_name, com, quote (val));
1023 switch (c_tolower (end[-1]))
1026 --end, mult = 1; /* seconds */
1029 --end, mult = 60; /* minutes */
1032 --end, mult = 3600; /* hours */
1035 --end, mult = 86400.0; /* days */
1038 --end, mult = 604800.0; /* weeks */
1041 /* Not a recognized suffix: assume it belongs to the number.
1042 (If not, simple_atof will raise an error.) */
1046 /* Skip leading and trailing whitespace. */
1047 while (val < end && c_isspace (*val))
1049 while (val < end && c_isspace (end[-1]))
1054 if (!simple_atof (val, end, &number))
1057 *(double *)place = number * mult;
1063 cmd_cert_type (const char *com, const char *val, void *place)
1065 static const struct decode_item choices[] = {
1066 { "pem", keyfile_pem },
1067 { "der", keyfile_asn1 },
1068 { "asn1", keyfile_asn1 },
1070 int ok = decode_string (val, choices, countof (choices), place);
1072 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1077 /* Specialized helper functions, used by `commands' to handle some
1078 options specially. */
1080 static bool check_user_specified_header (const char *);
1083 cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1085 if (!cmd_boolean (com, val, &opt.dirstruct))
1087 /* Since dirstruct behaviour is explicitly changed, no_dirstruct
1088 must be affected inversely. */
1090 opt.no_dirstruct = false;
1092 opt.no_dirstruct = true;
1097 cmd_spec_header (const char *com, const char *val, void *place_ignored)
1099 /* Empty value means reset the list of headers. */
1102 free_vec (opt.user_headers);
1103 opt.user_headers = NULL;
1107 if (!check_user_specified_header (val))
1109 fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1110 exec_name, com, quote (val));
1113 opt.user_headers = vec_append (opt.user_headers, val);
1118 cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1120 int flag = cmd_boolean (com, val, &opt.htmlify);
1121 if (flag && !opt.htmlify)
1122 opt.remove_listing = false;
1126 /* Set the "mirror" mode. It means: recursive download, timestamping,
1127 no limit on max. recursion depth, and don't remove listings. */
1130 cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1134 if (!cmd_boolean (com, val, &mirror))
1138 opt.recursive = true;
1139 if (!opt.no_dirstruct)
1140 opt.dirstruct = true;
1141 opt.timestamping = true;
1142 opt.reclevel = INFINITE_RECURSION;
1143 opt.remove_listing = false;
1148 /* Validate --prefer-family and set the choice. Allowed values are
1149 "IPv4", "IPv6", and "none". */
1152 cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1154 static const struct decode_item choices[] = {
1155 { "IPv4", prefer_ipv4 },
1156 { "IPv6", prefer_ipv6 },
1157 { "none", prefer_none },
1159 int prefer_family = prefer_ipv4;
1160 int ok = decode_string (val, choices, countof (choices), &prefer_family);
1162 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1163 opt.prefer_family = prefer_family;
1167 /* Set progress.type to VAL, but verify that it's a valid progress
1168 implementation before that. */
1171 cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1173 if (!valid_progress_implementation_p (val))
1175 fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1176 exec_name, com, quote (val));
1179 xfree_null (opt.progress_type);
1181 /* Don't call set_progress_implementation here. It will be called
1182 in main() when it becomes clear what the log output is. */
1183 opt.progress_type = xstrdup (val);
1187 /* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1188 set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1192 cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1194 if (!cmd_boolean (com, val, &opt.recursive))
1198 if (opt.recursive && !opt.no_dirstruct)
1199 opt.dirstruct = true;
1205 cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1207 int restrict_os = opt.restrict_files_os;
1208 int restrict_ctrl = opt.restrict_files_ctrl;
1209 int restrict_case = opt.restrict_files_case;
1213 #define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1217 end = strchr (val, ',');
1219 end = val + strlen (val);
1221 if (VAL_IS ("unix"))
1222 restrict_os = restrict_unix;
1223 else if (VAL_IS ("windows"))
1224 restrict_os = restrict_windows;
1225 else if (VAL_IS ("lowercase"))
1226 restrict_case = restrict_lowercase;
1227 else if (VAL_IS ("uppercase"))
1228 restrict_case = restrict_uppercase;
1229 else if (VAL_IS ("nocontrol"))
1230 restrict_ctrl = false;
1234 _("%s: %s: Invalid restriction %s, use [unix|windows],[lowercase|uppercase],[nocontrol].\n"),
1235 exec_name, com, quote (val));
1242 while (*val && *end);
1246 opt.restrict_files_os = restrict_os;
1247 opt.restrict_files_ctrl = restrict_ctrl;
1248 opt.restrict_files_case = restrict_case;
1255 cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1257 static const struct decode_item choices[] = {
1258 { "auto", secure_protocol_auto },
1259 { "sslv2", secure_protocol_sslv2 },
1260 { "sslv3", secure_protocol_sslv3 },
1261 { "tlsv1", secure_protocol_tlsv1 },
1263 int ok = decode_string (val, choices, countof (choices), place);
1265 fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1270 /* Set all three timeout values. */
1273 cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1276 if (!cmd_time (com, val, &value))
1278 opt.read_timeout = value;
1279 opt.connect_timeout = value;
1280 opt.dns_timeout = value;
1285 cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1287 /* Disallow embedded newlines. */
1288 if (strchr (val, '\n'))
1290 fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1291 exec_name, com, quote (val));
1294 xfree_null (opt.useragent);
1295 opt.useragent = xstrdup (val);
1299 /* The "verbose" option cannot be cmd_boolean because the variable is
1300 not bool -- it's of type int (-1 means uninitialized because of
1301 some random hackery for disallowing -q -v). */
1304 cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1307 if (cmd_boolean (com, val, &flag))
1315 /* Miscellaneous useful routines. */
1317 /* A very simple atoi clone, more useful than atoi because it works on
1318 delimited strings, and has error reportage. Returns true on success,
1319 false on failure. If successful, stores result to *DEST. */
1322 simple_atoi (const char *beg, const char *end, int *dest)
1325 bool negative = false;
1326 const char *p = beg;
1328 while (p < end && c_isspace (*p))
1330 if (p < end && (*p == '-' || *p == '+'))
1332 negative = (*p == '-');
1338 /* Read negative numbers in a separate loop because the most
1339 negative integer cannot be represented as a positive number. */
1342 for (; p < end && c_isdigit (*p); p++)
1344 int next = (10 * result) + (*p - '0');
1346 return false; /* overflow */
1350 for (; p < end && c_isdigit (*p); p++)
1352 int next = (10 * result) - (*p - '0');
1354 return false; /* underflow */
1365 /* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1366 doesn't handle exponential notation. Returns true on success,
1367 false on failure. In case of success, stores its result to
1371 simple_atof (const char *beg, const char *end, double *dest)
1375 bool negative = false;
1376 bool seen_dot = false;
1377 bool seen_digit = false;
1380 const char *p = beg;
1382 while (p < end && c_isspace (*p))
1384 if (p < end && (*p == '-' || *p == '+'))
1386 negative = (*p == '-');
1390 for (; p < end; p++)
1396 result = (10 * result) + (ch - '0');
1398 result += (ch - '0') / (divider *= 10);
1420 /* Verify that the user-specified header in S is valid. It must
1421 contain a colon preceded by non-white-space characters and must not
1422 contain newlines. */
1425 check_user_specified_header (const char *s)
1429 for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1431 /* The header MUST contain `:' preceded by at least one
1432 non-whitespace character. */
1433 if (*p != ':' || p == s)
1435 /* The header MUST NOT contain newlines. */
1436 if (strchr (s, '\n'))
1441 /* Decode VAL into a number, according to ITEMS. */
1444 decode_string (const char *val, const struct decode_item *items, int itemcount,
1448 for (i = 0; i < itemcount; i++)
1449 if (0 == strcasecmp (val, items[i].name))
1451 *place = items[i].code;
1458 void cleanup_html_url (void);
1461 /* Free the memory allocated by global variables. */
1465 /* Free external resources, close files, etc. */
1468 fclose (output_stream);
1469 /* No need to check for error because Wget flushes its output (and
1470 checks for errors) after any data arrives. */
1472 /* We're exiting anyway so there's no real need to call free()
1473 hundreds of times. Skipping the frees will make Wget exit
1476 However, when detecting leaks, it's crucial to free() everything
1477 because then you can find the real leaks, i.e. the allocated
1478 memory which grows with the size of the program. */
1484 cleanup_html_url ();
1489 extern acc_t *netrc_list;
1490 free_netrc (netrc_list);
1492 xfree_null (opt.lfilename);
1493 xfree_null (opt.dir_prefix);
1494 xfree_null (opt.input_filename);
1495 xfree_null (opt.output_document);
1496 free_vec (opt.accepts);
1497 free_vec (opt.rejects);
1498 free_vec (opt.excludes);
1499 free_vec (opt.includes);
1500 free_vec (opt.domains);
1501 free_vec (opt.follow_tags);
1502 free_vec (opt.ignore_tags);
1503 xfree_null (opt.progress_type);
1504 xfree_null (opt.ftp_user);
1505 xfree_null (opt.ftp_passwd);
1506 xfree_null (opt.ftp_proxy);
1507 xfree_null (opt.https_proxy);
1508 xfree_null (opt.http_proxy);
1509 free_vec (opt.no_proxy);
1510 xfree_null (opt.useragent);
1511 xfree_null (opt.referer);
1512 xfree_null (opt.http_user);
1513 xfree_null (opt.http_passwd);
1514 free_vec (opt.user_headers);
1516 xfree_null (opt.cert_file);
1517 xfree_null (opt.private_key);
1518 xfree_null (opt.ca_directory);
1519 xfree_null (opt.ca_cert);
1520 xfree_null (opt.random_file);
1521 xfree_null (opt.egd_file);
1523 xfree_null (opt.bind_address);
1524 xfree_null (opt.cookies_input);
1525 xfree_null (opt.cookies_output);
1526 xfree_null (opt.user);
1527 xfree_null (opt.passwd);
1528 #endif /* DEBUG_MALLOC */
1531 /* Unit testing routines. */
1536 test_cmd_spec_restrict_file_names()
1541 int expected_restrict_files_os;
1542 int expected_restrict_files_ctrl;
1543 int expected_restrict_files_case;
1546 { "windows", restrict_windows, true, restrict_no_case_restriction, true },
1547 { "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1548 { "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1549 { "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1552 for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1557 res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1560 fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1561 fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1562 fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1563 fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1565 mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1566 res == test_array[i].result
1567 && opt.restrict_files_os == test_array[i].expected_restrict_files_os
1568 && opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1569 && opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1575 #endif /* TESTING */