1
/* Reading/parsing the initialization file.
2
Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
3
2004, 2005, 2006, 2007, 2008, 2009 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 */
63
#define CMD_DECLARE(func) static bool func (const char *, const char *, void *)
65
CMD_DECLARE (cmd_boolean);
66
CMD_DECLARE (cmd_bytes);
67
CMD_DECLARE (cmd_bytes_sum);
69
CMD_DECLARE (cmd_cert_type);
71
CMD_DECLARE (cmd_directory_vector);
72
CMD_DECLARE (cmd_number);
73
CMD_DECLARE (cmd_number_inf);
74
CMD_DECLARE (cmd_string);
75
CMD_DECLARE (cmd_file);
76
CMD_DECLARE (cmd_directory);
77
CMD_DECLARE (cmd_time);
78
CMD_DECLARE (cmd_vector);
80
CMD_DECLARE (cmd_spec_dirstruct);
81
CMD_DECLARE (cmd_spec_header);
82
CMD_DECLARE (cmd_spec_htmlify);
83
CMD_DECLARE (cmd_spec_mirror);
84
CMD_DECLARE (cmd_spec_prefer_family);
85
CMD_DECLARE (cmd_spec_progress);
86
CMD_DECLARE (cmd_spec_recursive);
87
CMD_DECLARE (cmd_spec_restrict_file_names);
89
CMD_DECLARE (cmd_spec_secure_protocol);
91
CMD_DECLARE (cmd_spec_timeout);
92
CMD_DECLARE (cmd_spec_useragent);
93
CMD_DECLARE (cmd_spec_verbose);
95
/* List of recognized commands, each consisting of name, place and
96
function. When adding a new command, simply add it to the list,
97
but be sure to keep the list sorted alphabetically, as
98
command_by_name's binary search depends on it. Also, be sure to
99
add any entries that allocate memory (e.g. cmd_string and
100
cmd_vector) to the cleanup() function below. */
102
static const struct {
105
bool (*action) (const char *, const char *, void *);
107
/* KEEP THIS LIST ALPHABETICALLY SORTED */
108
{ "accept", &opt.accepts, cmd_vector },
109
{ "addhostdir", &opt.add_hostdir, cmd_boolean },
110
{ "adjustextension", &opt.adjust_extension, cmd_boolean },
111
{ "alwaysrest", &opt.always_rest, cmd_boolean }, /* deprecated */
112
{ "askpassword", &opt.ask_passwd, cmd_boolean },
113
{ "authnochallenge", &opt.auth_without_challenge,
115
{ "background", &opt.background, cmd_boolean },
116
{ "backupconverted", &opt.backup_converted, cmd_boolean },
117
{ "backups", &opt.backups, cmd_number },
118
{ "base", &opt.base_href, cmd_string },
119
{ "bindaddress", &opt.bind_address, cmd_string },
121
{ "cacertificate", &opt.ca_cert, cmd_file },
123
{ "cache", &opt.allow_cache, cmd_boolean },
125
{ "cadirectory", &opt.ca_directory, cmd_directory },
126
{ "certificate", &opt.cert_file, cmd_file },
127
{ "certificatetype", &opt.cert_type, cmd_cert_type },
128
{ "checkcertificate", &opt.check_cert, cmd_boolean },
130
{ "connecttimeout", &opt.connect_timeout, cmd_time },
131
{ "contentdisposition", &opt.content_disposition, cmd_boolean },
132
{ "continue", &opt.always_rest, cmd_boolean },
133
{ "convertlinks", &opt.convert_links, cmd_boolean },
134
{ "cookies", &opt.cookies, cmd_boolean },
135
{ "cutdirs", &opt.cut_dirs, cmd_number },
137
{ "debug", &opt.debug, cmd_boolean },
139
{ "defaultpage", &opt.default_page, cmd_string},
140
{ "deleteafter", &opt.delete_after, cmd_boolean },
141
{ "dirprefix", &opt.dir_prefix, cmd_directory },
142
{ "dirstruct", NULL, cmd_spec_dirstruct },
143
{ "dnscache", &opt.dns_cache, cmd_boolean },
144
{ "dnstimeout", &opt.dns_timeout, cmd_time },
145
{ "domains", &opt.domains, cmd_vector },
146
{ "dotbytes", &opt.dot_bytes, cmd_bytes },
147
{ "dotsinline", &opt.dots_in_line, cmd_number },
148
{ "dotspacing", &opt.dot_spacing, cmd_number },
149
{ "dotstyle", &opt.dot_style, cmd_string }, /* deprecated */
151
{ "egdfile", &opt.egd_file, cmd_file },
153
{ "excludedirectories", &opt.excludes, cmd_directory_vector },
154
{ "excludedomains", &opt.exclude_domains, cmd_vector },
155
{ "followftp", &opt.follow_ftp, cmd_boolean },
156
{ "followtags", &opt.follow_tags, cmd_vector },
157
{ "forcehtml", &opt.force_html, cmd_boolean },
158
{ "ftppasswd", &opt.ftp_passwd, cmd_string }, /* deprecated */
159
{ "ftppassword", &opt.ftp_passwd, cmd_string },
160
{ "ftpproxy", &opt.ftp_proxy, cmd_string },
162
{ "ftpstmlf", &opt.ftp_stmlf, cmd_boolean },
163
#endif /* def __VMS */
164
{ "ftpuser", &opt.ftp_user, cmd_string },
165
{ "glob", &opt.ftp_glob, cmd_boolean },
166
{ "header", NULL, cmd_spec_header },
167
{ "htmlextension", &opt.adjust_extension, cmd_boolean }, /* deprecated */
168
{ "htmlify", NULL, cmd_spec_htmlify },
169
{ "httpkeepalive", &opt.http_keep_alive, cmd_boolean },
170
{ "httppasswd", &opt.http_passwd, cmd_string }, /* deprecated */
171
{ "httppassword", &opt.http_passwd, cmd_string },
172
{ "httpproxy", &opt.http_proxy, cmd_string },
173
{ "httpsproxy", &opt.https_proxy, cmd_string },
174
{ "httpuser", &opt.http_user, cmd_string },
175
{ "ignorecase", &opt.ignore_case, cmd_boolean },
176
{ "ignorelength", &opt.ignore_length, cmd_boolean },
177
{ "ignoretags", &opt.ignore_tags, cmd_vector },
178
{ "includedirectories", &opt.includes, cmd_directory_vector },
180
{ "inet4only", &opt.ipv4_only, cmd_boolean },
181
{ "inet6only", &opt.ipv6_only, cmd_boolean },
183
{ "input", &opt.input_filename, cmd_file },
184
{ "iri", &opt.enable_iri, cmd_boolean },
185
{ "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
186
{ "limitrate", &opt.limit_rate, cmd_bytes },
187
{ "loadcookies", &opt.cookies_input, cmd_file },
188
{ "localencoding", &opt.locale, cmd_string },
189
{ "logfile", &opt.lfilename, cmd_file },
190
{ "login", &opt.ftp_user, cmd_string },/* deprecated*/
191
{ "maxredirect", &opt.max_redirect, cmd_number },
192
{ "mirror", NULL, cmd_spec_mirror },
193
{ "netrc", &opt.netrc, cmd_boolean },
194
{ "noclobber", &opt.noclobber, cmd_boolean },
195
{ "noparent", &opt.no_parent, cmd_boolean },
196
{ "noproxy", &opt.no_proxy, cmd_vector },
197
{ "numtries", &opt.ntry, cmd_number_inf },/* deprecated*/
198
{ "outputdocument", &opt.output_document, cmd_file },
199
{ "pagerequisites", &opt.page_requisites, cmd_boolean },
200
{ "passiveftp", &opt.ftp_pasv, cmd_boolean },
201
{ "passwd", &opt.ftp_passwd, cmd_string },/* deprecated*/
202
{ "password", &opt.passwd, cmd_string },
203
{ "postdata", &opt.post_data, cmd_string },
204
{ "postfile", &opt.post_file_name, cmd_file },
205
{ "preferfamily", NULL, cmd_spec_prefer_family },
206
{ "preservepermissions", &opt.preserve_perm, cmd_boolean },/* deprecated */
208
{ "privatekey", &opt.private_key, cmd_file },
209
{ "privatekeytype", &opt.private_key_type, cmd_cert_type },
211
{ "progress", &opt.progress_type, cmd_spec_progress },
212
{ "protocoldirectories", &opt.protocol_directories, cmd_boolean },
213
{ "proxypasswd", &opt.proxy_passwd, cmd_string }, /* deprecated */
214
{ "proxypassword", &opt.proxy_passwd, cmd_string },
215
{ "proxyuser", &opt.proxy_user, cmd_string },
216
{ "quiet", &opt.quiet, cmd_boolean },
217
{ "quota", &opt.quota, cmd_bytes_sum },
219
{ "randomfile", &opt.random_file, cmd_file },
221
{ "randomwait", &opt.random_wait, cmd_boolean },
222
{ "readtimeout", &opt.read_timeout, cmd_time },
223
{ "reclevel", &opt.reclevel, cmd_number_inf },
224
{ "recursive", NULL, cmd_spec_recursive },
225
{ "referer", &opt.referer, cmd_string },
226
{ "reject", &opt.rejects, cmd_vector },
227
{ "relativeonly", &opt.relative_only, cmd_boolean },
228
{ "remoteencoding", &opt.encoding_remote, cmd_string },
229
{ "removelisting", &opt.remove_listing, cmd_boolean },
230
{ "restrictfilenames", NULL, cmd_spec_restrict_file_names },
231
{ "retrsymlinks", &opt.retr_symlinks, cmd_boolean },
232
{ "retryconnrefused", &opt.retry_connrefused, cmd_boolean },
233
{ "robots", &opt.use_robots, cmd_boolean },
234
{ "savecookies", &opt.cookies_output, cmd_file },
235
{ "saveheaders", &opt.save_headers, cmd_boolean },
237
{ "secureprotocol", &opt.secure_protocol, cmd_spec_secure_protocol },
239
{ "serverresponse", &opt.server_response, cmd_boolean },
240
{ "spanhosts", &opt.spanhost, cmd_boolean },
241
{ "spider", &opt.spider, cmd_boolean },
242
{ "strictcomments", &opt.strict_comments, cmd_boolean },
243
{ "timeout", NULL, cmd_spec_timeout },
244
{ "timestamping", &opt.timestamping, cmd_boolean },
245
{ "tries", &opt.ntry, cmd_number_inf },
246
{ "trustservernames", &opt.trustservernames, cmd_boolean },
247
{ "useproxy", &opt.use_proxy, cmd_boolean },
248
{ "user", &opt.user, cmd_string },
249
{ "useragent", NULL, cmd_spec_useragent },
250
{ "verbose", NULL, cmd_spec_verbose },
251
{ "wait", &opt.wait, cmd_time },
252
{ "waitretry", &opt.waitretry, cmd_time },
254
{ "wdebug", &opt.wdebug, cmd_boolean },
258
/* Look up CMDNAME in the commands[] and return its position in the
259
array. If CMDNAME is not found, return -1. */
262
command_by_name (const char *cmdname)
264
/* Use binary search for speed. Wget has ~100 commands, which
265
guarantees a worst case performance of 7 string comparisons. */
266
int lo = 0, hi = countof (commands) - 1;
270
int mid = (lo + hi) >> 1;
271
int cmp = strcasecmp (cmdname, commands[mid].name);
282
/* Reset the variables to default values. */
288
/* Most of the default values are 0 (and 0.0, NULL, and false).
289
Just reset everything, and fill in the non-zero values. Note
290
that initializing pointers to NULL this way is technically
291
illegal, but porting Wget to a machine where NULL is not all-zero
292
bit pattern will be the least of the implementors' worries. */
299
opt.add_hostdir = true;
303
opt.http_keep_alive = true;
304
opt.use_proxy = true;
305
tmp = getenv ("no_proxy");
307
opt.no_proxy = sepstring (tmp);
308
opt.prefer_family = prefer_none;
309
opt.allow_cache = true;
311
opt.read_timeout = 900;
312
opt.use_robots = true;
314
opt.remove_listing = true;
316
opt.dot_bytes = 1024;
317
opt.dot_spacing = 10;
318
opt.dots_in_line = 50;
320
opt.dns_cache = true;
324
opt.check_cert = true;
327
/* The default for file name restriction defaults to the OS type. */
328
#if defined(WINDOWS) || defined(MSDOS) || defined(__CYGWIN__)
329
opt.restrict_files_os = restrict_windows;
331
opt.restrict_files_os = restrict_unix;
333
opt.restrict_files_ctrl = true;
334
opt.restrict_files_nonascii = false;
335
opt.restrict_files_case = restrict_no_case_restriction;
337
opt.max_redirect = 20;
342
opt.enable_iri = true;
344
opt.enable_iri = false;
347
opt.encoding_remote = NULL;
350
/* Return the user's home directory (strdup-ed), or NULL if none is
355
static char buf[PATH_MAX];
360
home = getenv ("HOME");
364
/* Under MSDOS, if $HOME isn't defined, use the directory where
365
`wget.exe' resides. */
366
const char *_w32_get_argv0 (void); /* in libwatt.a/pcconfig.c */
369
strcpy (buf, _w32_get_argv0 ());
370
p = strrchr (buf, '/'); /* djgpp */
372
p = strrchr (buf, '\\'); /* others */
376
#elif !defined(WINDOWS)
377
/* If HOME is not defined, try getting it from the password
379
struct passwd *pwd = getpwuid (getuid ());
380
if (!pwd || !pwd->pw_dir)
382
strcpy (buf, pwd->pw_dir);
385
/* Under Windows, if $HOME isn't defined, use the directory where
386
`wget.exe' resides. */
392
return home ? xstrdup (home) : NULL;
395
/* Check the 'WGETRC' environment variable and return the file name
396
if 'WGETRC' is set and is a valid file.
397
If the `WGETRC' variable exists but the file does not exist, the
398
function will exit(). */
400
wgetrc_env_file_name (void)
402
char *env = getenv ("WGETRC");
405
if (!file_exists_p (env))
407
fprintf (stderr, _("%s: WGETRC points to %s, which doesn't exist.\n"),
411
return xstrdup (env);
416
/* Check for the existance of '$HOME/.wgetrc' and return it's path
417
if it exists and is set. */
419
wgetrc_user_file_name (void)
421
char *home = home_dir ();
423
/* If that failed, try $HOME/.wgetrc (or equivalent). */
426
file = "SYS$LOGIN:.wgetrc";
427
#else /* def __VMS */
430
file = aprintf ("%s/.wgetrc", home);
432
#endif /* def __VMS [else] */
436
if (!file_exists_p (file))
444
/* Return the path to the user's .wgetrc. This is either the value of
445
`WGETRC' environment variable, or `$HOME/.wgetrc'.
447
Additionally, for windows, look in the directory where wget.exe
450
wgetrc_file_name (void)
452
char *file = wgetrc_env_file_name ();
456
file = wgetrc_user_file_name ();
459
/* Under Windows, if we still haven't found .wgetrc, look for the file
460
`wget.ini' in the directory where `wget.exe' resides; we do this for
461
backward compatibility with previous versions of Wget.
462
SYSTEM_WGETRC should not be defined under WINDOWS. */
465
char *home = home_dir ();
471
file = aprintf ("%s/wget.ini", home);
472
if (!file_exists_p (file))
485
/* Return values of parse_line. */
493
static enum parse_line parse_line (const char *, char **, char **, int *);
494
static bool setval_internal (int, const char *, const char *);
495
static bool setval_internal_tilde (int, const char *, const char *);
497
/* Initialize variables from a wgetrc file. Returns zero (failure) if
498
there were errors in the file. */
501
run_wgetrc (const char *file)
508
fp = fopen (file, "r");
511
fprintf (stderr, _("%s: Cannot read %s (%s).\n"), exec_name,
512
file, strerror (errno));
513
return true; /* not a fatal error */
516
while ((line = read_whole_line (fp)) != NULL)
518
char *com = NULL, *val = NULL;
521
/* Parse the line. */
522
switch (parse_line (line, &com, &val, &comind))
525
/* If everything is OK, set the value. */
526
if (!setval_internal_tilde (comind, com, val))
528
fprintf (stderr, _("%s: Error in %s at line %d.\n"),
529
exec_name, file, ln);
533
case line_syntax_error:
534
fprintf (stderr, _("%s: Syntax error in %s at line %d.\n"),
535
exec_name, file, ln);
538
case line_unknown_command:
539
fprintf (stderr, _("%s: Unknown command %s in %s at line %d.\n"),
540
exec_name, quote (com), file, ln);
558
/* Initialize the defaults and run the system wgetrc and user's own
563
char *file, *env_sysrc;
566
/* Load the hard-coded defaults. */
569
/* Run a non-standard system rc file when the according environment
570
variable has been set. For internal testing purposes only! */
571
env_sysrc = getenv ("SYSTEM_WGETRC");
572
if (env_sysrc && file_exists_p (env_sysrc))
573
ok &= run_wgetrc (env_sysrc);
574
/* Otherwise, if SYSTEM_WGETRC is defined, use it. */
576
else if (file_exists_p (SYSTEM_WGETRC))
577
ok &= run_wgetrc (SYSTEM_WGETRC);
579
/* Override it with your own, if one exists. */
580
file = wgetrc_file_name ();
583
/* #### We should canonicalize `file' and SYSTEM_WGETRC with
584
something like realpath() before comparing them with `strcmp' */
586
if (!strcmp (file, SYSTEM_WGETRC))
588
fprintf (stderr, _("\
589
%s: Warning: Both system and user wgetrc point to %s.\n"),
590
exec_name, quote (file));
594
ok &= run_wgetrc (file);
596
/* If there were errors processing either `.wgetrc', abort. */
604
/* Remove dashes and underscores from S, modifying S in the
610
char *t = s; /* t - tortoise */
611
char *h = s; /* h - hare */
613
if (*h == '_' || *h == '-')
620
/* Parse the line pointed by line, with the syntax:
621
<sp>* command <sp>* = <sp>* value <sp>*
622
Uses malloc to allocate space for command and value.
624
Returns one of line_ok, line_empty, line_syntax_error, or
625
line_unknown_command.
627
In case of line_ok, *COM and *VAL point to freshly allocated
628
strings, and *COMIND points to com's index. In case of error or
629
empty line, their values are unmodified. */
631
static enum parse_line
632
parse_line (const char *line, char **com, char **val, int *comind)
635
const char *end = line + strlen (line);
636
const char *cmdstart, *cmdend;
637
const char *valstart, *valend;
642
/* Skip leading and trailing whitespace. */
643
while (*line && c_isspace (*line))
645
while (end > line && c_isspace (end[-1]))
648
/* Skip empty lines and comments. */
649
if (!*line || *line == '#')
655
while (p < end && (c_isalnum (*p) || *p == '_' || *p == '-'))
659
/* Skip '=', as well as any space before or after it. */
660
while (p < end && c_isspace (*p))
662
if (p == end || *p != '=')
663
return line_syntax_error;
665
while (p < end && c_isspace (*p))
671
/* The syntax is valid (even though the command might not be). Fill
672
in the command name and value. */
673
*com = strdupdelim (cmdstart, cmdend);
674
*val = strdupdelim (valstart, valend);
676
/* The line now known to be syntactically correct. Check whether
677
the command is valid. */
678
BOUNDED_TO_ALLOCA (cmdstart, cmdend, cmdcopy);
680
ind = command_by_name (cmdcopy);
682
return line_unknown_command;
684
/* Report success to the caller. */
689
#if defined(WINDOWS) || defined(MSDOS)
690
# define ISSEP(c) ((c) == '/' || (c) == '\\')
692
# define ISSEP(c) ((c) == '/')
695
/* Run commands[comind].action. */
698
setval_internal (int comind, const char *com, const char *val)
700
assert (0 <= comind && ((size_t) comind) < countof (commands));
701
DEBUGP (("Setting %s (%s) to %s\n", com, commands[comind].name, val));
702
return commands[comind].action (com, val, commands[comind].place);
706
setval_internal_tilde (int comind, const char *com, const char *val)
712
ret = setval_internal (comind, com, val);
714
/* We make tilde expansion for cmd_file and cmd_directory */
715
if (((commands[comind].action == cmd_file) ||
716
(commands[comind].action == cmd_directory))
717
&& ret && (*val == '~' && ISSEP (val[1])))
719
pstring = commands[comind].place;
723
homelen = strlen (home);
724
while (homelen && ISSEP (home[homelen - 1]))
725
home[--homelen] = '\0';
727
/* Skip the leading "~/". */
728
for (++val; ISSEP (*val); val++)
730
*pstring = concat_strings (home, "/", val, (char *)0);
736
/* Run command COM with value VAL. If running the command produces an
737
error, report the error and exit.
739
This is intended to be called from main() to modify Wget's behavior
740
through command-line switches. Since COM is hard-coded in main(),
741
it is not canonicalized, and this aborts when COM is not found.
743
If COMIND's are exported to init.h, this function will be changed
744
to accept COMIND directly. */
747
setoptval (const char *com, const char *val, const char *optname)
749
/* Prepend "--" to OPTNAME. */
750
char *dd_optname = (char *) alloca (2 + strlen (optname) + 1);
753
strcpy (dd_optname + 2, optname);
755
assert (val != NULL);
756
if (!setval_internal (command_by_name (com), dd_optname, val))
760
/* Parse OPT into command and value and run it. For example,
761
run_command("foo=bar") is equivalent to setoptval("foo", "bar").
762
This is used by the `--execute' flag in main.c. */
765
run_command (const char *opt)
769
switch (parse_line (opt, &com, &val, &comind))
772
if (!setval_internal (comind, com, val))
778
fprintf (stderr, _("%s: Invalid --execute command %s\n"),
779
exec_name, quote (opt));
784
/* Generic helper functions, for use with `commands'. */
786
/* Forward declarations: */
791
static bool decode_string (const char *, const struct decode_item *, int, int *);
792
static bool simple_atoi (const char *, const char *, int *);
793
static bool simple_atof (const char *, const char *, double *);
795
#define CMP1(p, c0) (c_tolower((p)[0]) == (c0) && (p)[1] == '\0')
797
#define CMP2(p, c0, c1) (c_tolower((p)[0]) == (c0) \
798
&& c_tolower((p)[1]) == (c1) \
801
#define CMP3(p, c0, c1, c2) (c_tolower((p)[0]) == (c0) \
802
&& c_tolower((p)[1]) == (c1) \
803
&& c_tolower((p)[2]) == (c2) \
807
/* Store the boolean value from VAL to PLACE. COM is ignored,
808
except for error messages. */
810
cmd_boolean (const char *com, const char *val, void *place)
814
if (CMP2 (val, 'o', 'n') || CMP3 (val, 'y', 'e', 's') || CMP1 (val, '1'))
815
/* "on", "yes" and "1" mean true. */
817
else if (CMP3 (val, 'o', 'f', 'f') || CMP2 (val, 'n', 'o') || CMP1 (val, '0'))
818
/* "off", "no" and "0" mean false. */
823
_("%s: %s: Invalid boolean %s; use `on' or `off'.\n"),
824
exec_name, com, quote (val));
828
*(bool *) place = value;
832
/* Set the non-negative integer value from VAL to PLACE. With
833
incorrect specification, the number remains unchanged. */
835
cmd_number (const char *com, const char *val, void *place)
837
if (!simple_atoi (val, val + strlen (val), place)
838
|| *(int *) place < 0)
840
fprintf (stderr, _("%s: %s: Invalid number %s.\n"),
841
exec_name, com, quote (val));
847
/* Similar to cmd_number(), only accepts `inf' as a synonym for 0. */
849
cmd_number_inf (const char *com, const char *val, void *place)
851
if (!strcasecmp (val, "inf"))
856
return cmd_number (com, val, place);
859
/* Copy (strdup) the string at COM to a new location and place a
860
pointer to *PLACE. */
862
cmd_string (const char *com, const char *val, void *place)
864
char **pstring = (char **)place;
866
xfree_null (*pstring);
867
*pstring = xstrdup (val);
872
/* Like the above, but handles tilde-expansion when reading a user's
873
`.wgetrc'. In that case, and if VAL begins with `~', the tilde
874
gets expanded to the user's home directory. */
876
cmd_file (const char *com, const char *val, void *place)
878
char **pstring = (char **)place;
880
xfree_null (*pstring);
882
/* #### If VAL is empty, perhaps should set *PLACE to NULL. */
884
*pstring = xstrdup (val);
886
#if defined(WINDOWS) || defined(MSDOS)
887
/* Convert "\" to "/". */
890
for (s = *pstring; *s; s++)
898
/* Like cmd_file, but strips trailing '/' characters. */
900
cmd_directory (const char *com, const char *val, void *place)
904
/* Call cmd_file() for tilde expansion and separator
905
canonicalization (backslash -> slash under Windows). These
906
things should perhaps be in a separate function. */
907
if (!cmd_file (com, val, place))
912
while (t > s && *--t == '/')
918
/* Split VAL by space to a vector of values, and append those values
919
to vector pointed to by the PLACE argument. If VAL is empty, the
920
PLACE vector is cleared instead. */
923
cmd_vector (const char *com, const char *val, void *place)
925
char ***pvec = (char ***)place;
928
*pvec = merge_vecs (*pvec, sepstring (val));
938
cmd_directory_vector (const char *com, const char *val, void *place)
940
char ***pvec = (char ***)place;
944
/* Strip the trailing slashes from directories. */
947
seps = sepstring (val);
948
for (t = seps; t && *t; t++)
950
int len = strlen (*t);
951
/* Skip degenerate case of root directory. */
954
if ((*t)[len - 1] == '/')
955
(*t)[len - 1] = '\0';
958
*pvec = merge_vecs (*pvec, seps);
968
/* Engine for cmd_bytes and cmd_bytes_sum: converts a string such as
969
"100k" or "2.5G" to a floating point number. */
972
parse_bytes_helper (const char *val, double *result)
975
const char *end = val + strlen (val);
977
/* Check for "inf". */
978
if (0 == strcmp (val, "inf"))
984
/* Strip trailing whitespace. */
985
while (val < end && c_isspace (end[-1]))
990
switch (c_tolower (end[-1]))
993
--end, mult = 1024.0;
996
--end, mult = 1048576.0;
999
--end, mult = 1073741824.0;
1002
--end, mult = 1099511627776.0;
1005
/* Not a recognized suffix: assume it's a digit. (If not,
1006
simple_atof will raise an error.) */
1010
/* Skip leading and trailing whitespace. */
1011
while (val < end && c_isspace (*val))
1013
while (val < end && c_isspace (end[-1]))
1018
if (!simple_atof (val, end, &number) || number < 0)
1021
*result = number * mult;
1025
/* Parse VAL as a number and set its value to PLACE (which should
1028
By default, the value is assumed to be in bytes. If "K", "M", or
1029
"G" are appended, the value is multiplied with 1<<10, 1<<20, or
1030
1<<30, respectively. Floating point values are allowed and are
1031
cast to integer before use. The idea is to be able to use things
1032
like 1.5k instead of "1536".
1034
The string "inf" is returned as 0.
1036
In case of error, false is returned and memory pointed to by PLACE
1037
remains unmodified. */
1040
cmd_bytes (const char *com, const char *val, void *place)
1043
if (!parse_bytes_helper (val, &byte_value))
1045
fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1046
exec_name, com, quote (val));
1049
*(wgint *)place = (wgint)byte_value;
1053
/* Like cmd_bytes, but PLACE is interpreted as a pointer to
1054
SIZE_SUM. It works by converting the string to double, therefore
1055
working with values up to 2^53-1 without loss of precision. This
1056
value (8192 TB) is large enough to serve for a while. */
1059
cmd_bytes_sum (const char *com, const char *val, void *place)
1062
if (!parse_bytes_helper (val, &byte_value))
1064
fprintf (stderr, _("%s: %s: Invalid byte value %s\n"),
1065
exec_name, com, quote (val));
1068
*(SUM_SIZE_INT *) place = (SUM_SIZE_INT) byte_value;
1072
/* Store the value of VAL to *OUT. The value is a time period, by
1073
default expressed in seconds, but also accepting suffixes "m", "h",
1074
"d", and "w" for minutes, hours, days, and weeks respectively. */
1077
cmd_time (const char *com, const char *val, void *place)
1079
double number, mult;
1080
const char *end = val + strlen (val);
1082
/* Strip trailing whitespace. */
1083
while (val < end && c_isspace (end[-1]))
1089
fprintf (stderr, _("%s: %s: Invalid time period %s\n"),
1090
exec_name, com, quote (val));
1094
switch (c_tolower (end[-1]))
1097
--end, mult = 1; /* seconds */
1100
--end, mult = 60; /* minutes */
1103
--end, mult = 3600; /* hours */
1106
--end, mult = 86400.0; /* days */
1109
--end, mult = 604800.0; /* weeks */
1112
/* Not a recognized suffix: assume it belongs to the number.
1113
(If not, simple_atof will raise an error.) */
1117
/* Skip leading and trailing whitespace. */
1118
while (val < end && c_isspace (*val))
1120
while (val < end && c_isspace (end[-1]))
1125
if (!simple_atof (val, end, &number))
1128
*(double *)place = number * mult;
1134
cmd_cert_type (const char *com, const char *val, void *place)
1136
static const struct decode_item choices[] = {
1137
{ "pem", keyfile_pem },
1138
{ "der", keyfile_asn1 },
1139
{ "asn1", keyfile_asn1 },
1141
int ok = decode_string (val, choices, countof (choices), place);
1143
fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1148
/* Specialized helper functions, used by `commands' to handle some
1149
options specially. */
1151
static bool check_user_specified_header (const char *);
1154
cmd_spec_dirstruct (const char *com, const char *val, void *place_ignored)
1156
if (!cmd_boolean (com, val, &opt.dirstruct))
1158
/* Since dirstruct behaviour is explicitly changed, no_dirstruct
1159
must be affected inversely. */
1161
opt.no_dirstruct = false;
1163
opt.no_dirstruct = true;
1168
cmd_spec_header (const char *com, const char *val, void *place_ignored)
1170
/* Empty value means reset the list of headers. */
1173
free_vec (opt.user_headers);
1174
opt.user_headers = NULL;
1178
if (!check_user_specified_header (val))
1180
fprintf (stderr, _("%s: %s: Invalid header %s.\n"),
1181
exec_name, com, quote (val));
1184
opt.user_headers = vec_append (opt.user_headers, val);
1189
cmd_spec_htmlify (const char *com, const char *val, void *place_ignored)
1191
int flag = cmd_boolean (com, val, &opt.htmlify);
1192
if (flag && !opt.htmlify)
1193
opt.remove_listing = false;
1197
/* Set the "mirror" mode. It means: recursive download, timestamping,
1198
no limit on max. recursion depth, and don't remove listings. */
1201
cmd_spec_mirror (const char *com, const char *val, void *place_ignored)
1205
if (!cmd_boolean (com, val, &mirror))
1209
opt.recursive = true;
1210
if (!opt.no_dirstruct)
1211
opt.dirstruct = true;
1212
opt.timestamping = true;
1213
opt.reclevel = INFINITE_RECURSION;
1214
opt.remove_listing = false;
1219
/* Validate --prefer-family and set the choice. Allowed values are
1220
"IPv4", "IPv6", and "none". */
1223
cmd_spec_prefer_family (const char *com, const char *val, void *place_ignored)
1225
static const struct decode_item choices[] = {
1226
{ "IPv4", prefer_ipv4 },
1227
{ "IPv6", prefer_ipv6 },
1228
{ "none", prefer_none },
1230
int prefer_family = prefer_none;
1231
int ok = decode_string (val, choices, countof (choices), &prefer_family);
1233
fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1234
opt.prefer_family = prefer_family;
1238
/* Set progress.type to VAL, but verify that it's a valid progress
1239
implementation before that. */
1242
cmd_spec_progress (const char *com, const char *val, void *place_ignored)
1244
if (!valid_progress_implementation_p (val))
1246
fprintf (stderr, _("%s: %s: Invalid progress type %s.\n"),
1247
exec_name, com, quote (val));
1250
xfree_null (opt.progress_type);
1252
/* Don't call set_progress_implementation here. It will be called
1253
in main() when it becomes clear what the log output is. */
1254
opt.progress_type = xstrdup (val);
1258
/* Set opt.recursive to VAL as with cmd_boolean. If opt.recursive is
1259
set to true, also set opt.dirstruct to true, unless opt.no_dirstruct
1263
cmd_spec_recursive (const char *com, const char *val, void *place_ignored)
1265
if (!cmd_boolean (com, val, &opt.recursive))
1269
if (opt.recursive && !opt.no_dirstruct)
1270
opt.dirstruct = true;
1276
cmd_spec_restrict_file_names (const char *com, const char *val, void *place_ignored)
1278
int restrict_os = opt.restrict_files_os;
1279
int restrict_ctrl = opt.restrict_files_ctrl;
1280
int restrict_case = opt.restrict_files_case;
1281
int restrict_nonascii = opt.restrict_files_nonascii;
1285
#define VAL_IS(string_literal) BOUNDED_EQUAL (val, end, string_literal)
1289
end = strchr (val, ',');
1291
end = val + strlen (val);
1293
if (VAL_IS ("unix"))
1294
restrict_os = restrict_unix;
1295
else if (VAL_IS ("windows"))
1296
restrict_os = restrict_windows;
1297
else if (VAL_IS ("lowercase"))
1298
restrict_case = restrict_lowercase;
1299
else if (VAL_IS ("uppercase"))
1300
restrict_case = restrict_uppercase;
1301
else if (VAL_IS ("nocontrol"))
1302
restrict_ctrl = false;
1303
else if (VAL_IS ("ascii"))
1304
restrict_nonascii = true;
1307
fprintf (stderr, _("\
1308
%s: %s: Invalid restriction %s,\n\
1309
use [unix|windows],[lowercase|uppercase],[nocontrol],[ascii].\n"),
1310
exec_name, com, quote (val));
1317
while (*val && *end);
1321
opt.restrict_files_os = restrict_os;
1322
opt.restrict_files_ctrl = restrict_ctrl;
1323
opt.restrict_files_case = restrict_case;
1324
opt.restrict_files_nonascii = restrict_nonascii;
1331
cmd_spec_secure_protocol (const char *com, const char *val, void *place)
1333
static const struct decode_item choices[] = {
1334
{ "auto", secure_protocol_auto },
1335
{ "sslv2", secure_protocol_sslv2 },
1336
{ "sslv3", secure_protocol_sslv3 },
1337
{ "tlsv1", secure_protocol_tlsv1 },
1339
int ok = decode_string (val, choices, countof (choices), place);
1341
fprintf (stderr, _("%s: %s: Invalid value %s.\n"), exec_name, com, quote (val));
1346
/* Set all three timeout values. */
1349
cmd_spec_timeout (const char *com, const char *val, void *place_ignored)
1352
if (!cmd_time (com, val, &value))
1354
opt.read_timeout = value;
1355
opt.connect_timeout = value;
1356
opt.dns_timeout = value;
1361
cmd_spec_useragent (const char *com, const char *val, void *place_ignored)
1363
/* Disallow embedded newlines. */
1364
if (strchr (val, '\n'))
1366
fprintf (stderr, _("%s: %s: Invalid value %s.\n"),
1367
exec_name, com, quote (val));
1370
xfree_null (opt.useragent);
1371
opt.useragent = xstrdup (val);
1375
/* The "verbose" option cannot be cmd_boolean because the variable is
1376
not bool -- it's of type int (-1 means uninitialized because of
1377
some random hackery for disallowing -q -v). */
1380
cmd_spec_verbose (const char *com, const char *val, void *place_ignored)
1383
if (cmd_boolean (com, val, &flag))
1391
/* Miscellaneous useful routines. */
1393
/* A very simple atoi clone, more useful than atoi because it works on
1394
delimited strings, and has error reportage. Returns true on success,
1395
false on failure. If successful, stores result to *DEST. */
1398
simple_atoi (const char *beg, const char *end, int *dest)
1401
bool negative = false;
1402
const char *p = beg;
1404
while (p < end && c_isspace (*p))
1406
if (p < end && (*p == '-' || *p == '+'))
1408
negative = (*p == '-');
1414
/* Read negative numbers in a separate loop because the most
1415
negative integer cannot be represented as a positive number. */
1418
for (; p < end && c_isdigit (*p); p++)
1420
int next = (10 * result) + (*p - '0');
1422
return false; /* overflow */
1426
for (; p < end && c_isdigit (*p); p++)
1428
int next = (10 * result) - (*p - '0');
1430
return false; /* underflow */
1441
/* Trivial atof, with error reporting. Handles "<digits>[.<digits>]",
1442
doesn't handle exponential notation. Returns true on success,
1443
false on failure. In case of success, stores its result to
1447
simple_atof (const char *beg, const char *end, double *dest)
1451
bool negative = false;
1452
bool seen_dot = false;
1453
bool seen_digit = false;
1456
const char *p = beg;
1458
while (p < end && c_isspace (*p))
1460
if (p < end && (*p == '-' || *p == '+'))
1462
negative = (*p == '-');
1466
for (; p < end; p++)
1472
result = (10 * result) + (ch - '0');
1474
result += (ch - '0') / (divider *= 10);
1496
/* Verify that the user-specified header in S is valid. It must
1497
contain a colon preceded by non-white-space characters and must not
1498
contain newlines. */
1501
check_user_specified_header (const char *s)
1505
for (p = s; *p && *p != ':' && !c_isspace (*p); p++)
1507
/* The header MUST contain `:' preceded by at least one
1508
non-whitespace character. */
1509
if (*p != ':' || p == s)
1511
/* The header MUST NOT contain newlines. */
1512
if (strchr (s, '\n'))
1517
/* Decode VAL into a number, according to ITEMS. */
1520
decode_string (const char *val, const struct decode_item *items, int itemcount,
1524
for (i = 0; i < itemcount; i++)
1525
if (0 == strcasecmp (val, items[i].name))
1527
*place = items[i].code;
1534
void cleanup_html_url (void);
1537
/* Free the memory allocated by global variables. */
1541
/* Free external resources, close files, etc. */
1544
fclose (output_stream);
1545
/* No need to check for error because Wget flushes its output (and
1546
checks for errors) after any data arrives. */
1548
/* We're exiting anyway so there's no real need to call free()
1549
hundreds of times. Skipping the frees will make Wget exit
1552
However, when detecting leaks, it's crucial to free() everything
1553
because then you can find the real leaks, i.e. the allocated
1554
memory which grows with the size of the program. */
1560
cleanup_html_url ();
1565
extern acc_t *netrc_list;
1566
free_netrc (netrc_list);
1568
xfree_null (opt.lfilename);
1569
xfree_null (opt.dir_prefix);
1570
xfree_null (opt.input_filename);
1571
xfree_null (opt.output_document);
1572
free_vec (opt.accepts);
1573
free_vec (opt.rejects);
1574
free_vec (opt.excludes);
1575
free_vec (opt.includes);
1576
free_vec (opt.domains);
1577
free_vec (opt.follow_tags);
1578
free_vec (opt.ignore_tags);
1579
xfree_null (opt.progress_type);
1580
xfree_null (opt.ftp_user);
1581
xfree_null (opt.ftp_passwd);
1582
xfree_null (opt.ftp_proxy);
1583
xfree_null (opt.https_proxy);
1584
xfree_null (opt.http_proxy);
1585
free_vec (opt.no_proxy);
1586
xfree_null (opt.useragent);
1587
xfree_null (opt.referer);
1588
xfree_null (opt.http_user);
1589
xfree_null (opt.http_passwd);
1590
free_vec (opt.user_headers);
1592
xfree_null (opt.cert_file);
1593
xfree_null (opt.private_key);
1594
xfree_null (opt.ca_directory);
1595
xfree_null (opt.ca_cert);
1596
xfree_null (opt.random_file);
1597
xfree_null (opt.egd_file);
1599
xfree_null (opt.bind_address);
1600
xfree_null (opt.cookies_input);
1601
xfree_null (opt.cookies_output);
1602
xfree_null (opt.user);
1603
xfree_null (opt.passwd);
1604
xfree_null (opt.base_href);
1606
#endif /* DEBUG_MALLOC */
1609
/* Unit testing routines. */
1614
test_commands_sorted()
1616
int prev_idx = 0, next_idx = 1;
1617
int command_count = countof (commands) - 1;
1619
while (next_idx <= command_count)
1621
cmp = strcasecmp (commands[prev_idx].name, commands[next_idx].name);
1624
mu_assert ("FAILED", false);
1637
test_cmd_spec_restrict_file_names()
1642
int expected_restrict_files_os;
1643
int expected_restrict_files_ctrl;
1644
int expected_restrict_files_case;
1647
{ "windows", restrict_windows, true, restrict_no_case_restriction, true },
1648
{ "windows,", restrict_windows, true, restrict_no_case_restriction, true },
1649
{ "windows,lowercase", restrict_windows, true, restrict_lowercase, true },
1650
{ "unix,nocontrol,lowercase,", restrict_unix, false, restrict_lowercase, true },
1653
for (i = 0; i < sizeof(test_array)/sizeof(test_array[0]); ++i)
1658
res = cmd_spec_restrict_file_names ("dummy", test_array[i].val, NULL);
1661
fprintf (stderr, "test_cmd_spec_restrict_file_names: TEST %d\n", i); fflush (stderr);
1662
fprintf (stderr, "opt.restrict_files_os: %d\n", opt.restrict_files_os); fflush (stderr);
1663
fprintf (stderr, "opt.restrict_files_ctrl: %d\n", opt.restrict_files_ctrl); fflush (stderr);
1664
fprintf (stderr, "opt.restrict_files_case: %d\n", opt.restrict_files_case); fflush (stderr);
1666
mu_assert ("test_cmd_spec_restrict_file_names: wrong result",
1667
res == test_array[i].result
1668
&& opt.restrict_files_os == test_array[i].expected_restrict_files_os
1669
&& opt.restrict_files_ctrl == test_array[i].expected_restrict_files_ctrl
1670
&& opt.restrict_files_case == test_array[i].expected_restrict_files_case);
1676
#endif /* TESTING */