root/tags/2.1.1/wp-includes/kses.php

Revision 4654, 22.9 kB (checked in by markjaquith, 2 years ago)

Reverse backport (?) from 2.0.x of KSES filters

  • Property svn:eol-style set to native
  • Property svn:keywords set to Author Date Id Revision
Line 
1 <?php
2
3 // Added wp_ prefix to avoid conflicts with existing kses users
4 # kses 0.2.2 - HTML/XHTML filter that only allows some elements and attributes
5 # Copyright (C) 2002, 2003, 2005  Ulf Harnhammar
6 # *** CONTACT INFORMATION ***
7 #
8 # E-mail:      metaur at users dot sourceforge dot net
9 # Web page:    http://sourceforge.net/projects/kses
10 # Paper mail:  Ulf Harnhammar
11 #              Ymergatan 17 C
12 #              753 25  Uppsala
13 #              SWEDEN
14 #
15 # [kses strips evil scripts!]
16 if (!defined('CUSTOM_TAGS'))
17     define('CUSTOM_TAGS', false);
18
19 // You can override this in your my-hacks.php file
20 if (!CUSTOM_TAGS) {
21     $allowedposttags = array ('address' => array (), 'a' => array ('href' => array (), 'title' => array (), 'rel' => array (), 'rev' => array (), 'name' => array ()), 'abbr' => array ('title' => array ()), 'acronym' => array ('title' => array ()), 'b' => array (), 'big' => array (), 'blockquote' => array ('cite' => array ()), 'br' => array (), 'button' => array ('disabled' => array (), 'name' => array (), 'type' => array (), 'value' => array ()), 'caption' => array ('align' => array ()), 'code' => array (), 'col' => array ('align' => array (), 'char' => array (), 'charoff' => array (), 'span' => array (), 'valign' => array (), 'width' => array ()), 'del' => array ('datetime' => array ()), 'dd' => array (), 'div' => array ('align' => array ()), 'dl' => array (), 'dt' => array (), 'em' => array (), 'fieldset' => array (), 'font' => array ('color' => array (), 'face' => array (), 'size' => array ()), 'form' => array ('action' => array (), 'accept' => array (), 'accept-charset' => array (), 'enctype' => array (), 'method' => array (), 'name' => array (), 'target' => array ()), 'h1' => array ('align' => array ()), 'h2' => array ('align' => array ()), 'h3' => array ('align' => array ()), 'h4' => array ('align' => array ()), 'h5' => array ('align' => array ()), 'h6' => array ('align' => array ()), 'hr' => array ('align' => array (), 'noshade' => array (), 'size' => array (), 'width' => array ()), 'i' => array (), 'img' => array ('alt' => array (), 'align' => array (), 'border' => array (), 'height' => array (), 'hspace' => array (), 'longdesc' => array (), 'vspace' => array (), 'src' => array (), 'width' => array ()), 'ins' => array ('datetime' => array (), 'cite' => array ()), 'kbd' => array (), 'label' => array ('for' => array ()), 'legend' => array ('align' => array ()), 'li' => array (), 'p' => array ('align' => array ()), 'pre' => array ('width' => array ()), 'q' => array ('cite' => array ()), 's' => array (), 'strike' => array (), 'strong' => array (), 'sub' => array (), 'sup' => array (), 'table' => array ('align' => array (), 'bgcolor' => array (), 'border' => array (), 'cellpadding' => array (), 'cellspacing' => array (), 'rules' => array (), 'summary' => array (), 'width' => array ()), 'tbody' => array ('align' => array (), 'char' => array (), 'charoff' => array (), 'valign' => array ()), 'td' => array ('abbr' => array (), 'align' => array (), 'axis' => array (), 'bgcolor' => array (), 'char' => array (), 'charoff' => array (), 'colspan' => array (), 'headers' => array (), 'height' => array (), 'nowrap' => array (), 'rowspan' => array (), 'scope' => array (), 'valign' => array (), 'width' => array ()), 'textarea' => array ('cols' => array (), 'rows' => array (), 'disabled' => array (), 'name' => array (), 'readonly' => array ()), 'tfoot' => array ('align' => array (), 'char' => array (), 'charoff' => array (), 'valign' => array ()), 'th' => array ('abbr' => array (), 'align' => array (), 'axis' => array (), 'bgcolor' => array (), 'char' => array (), 'charoff' => array (), 'colspan' => array (), 'headers' => array (), 'height' => array (), 'nowrap' => array (), 'rowspan' => array (), 'scope' => array (), 'valign' => array (), 'width' => array ()), 'thead' => array ('align' => array (), 'char' => array (), 'charoff' => array (), 'valign' => array ()), 'title' => array (), 'tr' => array ('align' => array (), 'bgcolor' => array (), 'char' => array (), 'charoff' => array (), 'valign' => array ()), 'tt' => array (), 'u' => array (), 'ul' => array (), 'ol' => array (), 'var' => array () );
22     $allowedtags = array ('a' => array ('href' => array (), 'title' => array ()), 'abbr' => array ('title' => array ()), 'acronym' => array ('title' => array ()), 'b' => array (), 'blockquote' => array ('cite' => array ()),
23         //    'br' => array(),
24     'code' => array (),
25         //    'del' => array('datetime' => array()),
26         //    'dd' => array(),
27         //    'dl' => array(),
28         //    'dt' => array(),
29     'em' => array (), 'i' => array (),
30         //    'ins' => array('datetime' => array(), 'cite' => array()),
31         //    'li' => array(),
32         //    'ol' => array(),
33         //    'p' => array(),
34         //    'q' => array(),
35     'strike' => array (), 'strong' => array (),
36         //    'sub' => array(),
37         //    'sup' => array(),
38         //    'u' => array(),
39         //    'ul' => array(),
40     );
41 }
42 function wp_kses($string, $allowed_html, $allowed_protocols = array ('http', 'https', 'ftp', 'ftps', 'mailto', 'news', 'irc', 'gopher', 'nntp', 'feed', 'telnet'))
43     ###############################################################################
44         # This function makes sure that only the allowed HTML element names, attribute
45         # names and attribute values plus only sane HTML entities will occur in
46         # $string. You have to remove any slashes from PHP's magic quotes before you
47         # call this function.
48         ###############################################################################
49     {
50     $string = wp_kses_no_null($string);
51     $string = wp_kses_js_entities($string);
52     $string = wp_kses_normalize_entities($string);
53     $string = wp_kses_hook($string);
54     $allowed_html_fixed = wp_kses_array_lc($allowed_html);
55     return wp_kses_split($string, $allowed_html_fixed, $allowed_protocols);
56 } # function wp_kses
57
58 function wp_kses_hook($string)
59 ###############################################################################
60 # You add any kses hooks here.
61 ###############################################################################
62 {
63     return $string;
64 } # function wp_kses_hook
65
66 function wp_kses_version()
67 ###############################################################################
68 # This function returns kses' version number.
69 ###############################################################################
70 {
71     return '0.2.2';
72 } # function wp_kses_version
73
74 function wp_kses_split($string, $allowed_html, $allowed_protocols)
75 ###############################################################################
76 # This function searches for HTML tags, no matter how malformed. It also
77 # matches stray ">" characters.
78 ###############################################################################
79 {
80     return preg_replace('%((<!--.*?(-->|$))|(<[^>]*(>|$)|>))%e',
81     "wp_kses_split2('\\1', \$allowed_html, ".'$allowed_protocols)', $string);
82 } # function wp_kses_split
83
84 function wp_kses_split2($string, $allowed_html, $allowed_protocols)
85 ###############################################################################
86 # This function does a lot of work. It rejects some very malformed things
87 # like <:::>. It returns an empty string, if the element isn't allowed (look
88 # ma, no strip_tags()!). Otherwise it splits the tag into an element and an
89 # attribute list.
90 ###############################################################################
91 {
92     $string = wp_kses_stripslashes($string);
93
94     if (substr($string, 0, 1) != '<')
95         return '&gt;';
96     # It matched a ">" character
97
98     if (preg_match('%^<!--(.*?)(-->)?$%', $string, $matches)) {
99         $string = str_replace(array('<!--', '-->'), '', $matches[1]);
100         while ( $string != $newstring = wp_kses($string, $allowed_html, $allowed_protocols) )
101             $string = $newstring;
102         if ( $string == '' )
103             return '';
104         return "<!--{$string}-->";
105     }
106     # Allow HTML comments
107
108     if (!preg_match('%^<\s*(/\s*)?([a-zA-Z0-9]+)([^>]*)>?$%', $string, $matches))
109         return '';
110     # It's seriously malformed
111
112     $slash = trim($matches[1]);
113     $elem = $matches[2];
114     $attrlist = $matches[3];
115
116     if (!@isset($allowed_html[strtolower($elem)]))
117         return '';
118     # They are using a not allowed HTML element
119
120     if ($slash != '')
121         return "<$slash$elem>";
122     # No attributes are allowed for closing elements
123
124     return wp_kses_attr("$slash$elem", $attrlist, $allowed_html, $allowed_protocols);
125 } # function wp_kses_split2
126
127 function wp_kses_attr($element, $attr, $allowed_html, $allowed_protocols)
128 ###############################################################################
129 # This function removes all attributes, if none are allowed for this element.
130 # If some are allowed it calls wp_kses_hair() to split them further, and then it
131 # builds up new HTML code from the data that kses_hair() returns. It also
132 # removes "<" and ">" characters, if there are any left. One more thing it
133 # does is to check if the tag has a closing XHTML slash, and if it does,
134 # it puts one in the returned code as well.
135 ###############################################################################
136 {
137     # Is there a closing XHTML slash at the end of the attributes?
138
139     $xhtml_slash = '';
140     if (preg_match('%\s/\s*$%', $attr))
141         $xhtml_slash = ' /';
142
143     # Are any attributes allowed at all for this element?
144
145     if (@ count($allowed_html[strtolower($element)]) == 0)
146         return "<$element$xhtml_slash>";
147
148     # Split it
149
150     $attrarr = wp_kses_hair($attr, $allowed_protocols);
151
152     # Go through $attrarr, and save the allowed attributes for this element
153     # in $attr2
154
155     $attr2 = '';
156
157     foreach ($attrarr as $arreach) {
158         if (!@ isset ($allowed_html[strtolower($element)][strtolower($arreach['name'])]))
159             continue; # the attribute is not allowed
160
161         $current = $allowed_html[strtolower($element)][strtolower($arreach['name'])];
162         if ($current == '')
163             continue; # the attribute is not allowed
164
165         if (!is_array($current))
166             $attr2 .= ' '.$arreach['whole'];
167         # there are no checks
168
169         else {
170             # there are some checks
171             $ok = true;
172             foreach ($current as $currkey => $currval)
173                 if (!wp_kses_check_attr_val($arreach['value'], $arreach['vless'], $currkey, $currval)) {
174                     $ok = false;
175                     break;
176                 }
177
178             if ($ok)
179                 $attr2 .= ' '.$arreach['whole']; # it passed them
180         } # if !is_array($current)
181     } # foreach
182
183     # Remove any "<" or ">" characters
184
185     $attr2 = preg_replace('/[<>]/', '', $attr2);
186
187     return "<$element$attr2$xhtml_slash>";
188 } # function wp_kses_attr
189
190 function wp_kses_hair($attr, $allowed_protocols)
191 ###############################################################################
192 # This function does a lot of work. It parses an attribute list into an array
193 # with attribute data, and tries to do the right thing even if it gets weird
194 # input. It will add quotes around attribute values that don't have any quotes
195 # or apostrophes around them, to make it easier to produce HTML code that will
196 # conform to W3C's HTML specification. It will also remove bad URL protocols
197 # from attribute values.
198 ###############################################################################
199 {
200     $attrarr = array ();
201     $mode = 0;
202     $attrname = '';
203
204     # Loop through the whole attribute list
205
206     while (strlen($attr) != 0) {
207         $working = 0; # Was the last operation successful?
208
209         switch ($mode) {
210             case 0 : # attribute name, href for instance
211
212                 if (preg_match('/^([-a-zA-Z]+)/', $attr, $match)) {
213                     $attrname = $match[1];
214                     $working = $mode = 1;
215                     $attr = preg_replace('/^[-a-zA-Z]+/', '', $attr);
216                 }
217
218                 break;
219
220             case 1 : # equals sign or valueless ("selected")
221
222                 if (preg_match('/^\s*=\s*/', $attr)) # equals sign
223                     {
224                     $working = 1;
225                     $mode = 2;
226                     $attr = preg_replace('/^\s*=\s*/', '', $attr);
227                     break;
228                 }
229
230                 if (preg_match('/^\s+/', $attr)) # valueless
231                     {
232                     $working = 1;
233                     $mode = 0;
234                     $attrarr[] = array ('name' => $attrname, 'value' => '', 'whole' => $attrname, 'vless' => 'y');
235                     $attr = preg_replace('/^\s+/', '', $attr);
236                 }
237
238                 break;
239
240             case 2 : # attribute value, a URL after href= for instance
241
242                 if (preg_match('/^"([^"]*)"(\s+|$)/', $attr, $match))
243                     # "value"
244                     {
245                     $thisval = wp_kses_bad_protocol($match[1], $allowed_protocols);
246
247                     $attrarr[] = array ('name' => $attrname, 'value' => $thisval, 'whole' => "$attrname=\"$thisval\"", 'vless' => 'n');
248                     $working = 1;
249                     $mode = 0;
250                     $attr = preg_replace('/^"[^"]*"(\s+|$)/', '', $attr);
251                     break;
252                 }
253
254                 if (preg_match("/^'([^']*)'(\s+|$)/", $attr, $match))
255                     # 'value'
256                     {
257                     $thisval = wp_kses_bad_protocol($match[1], $allowed_protocols);
258
259                     $attrarr[] = array ('name' => $attrname, 'value' => $thisval, 'whole' => "$attrname='$thisval'", 'vless' => 'n');
260                     $working = 1;
261                     $mode = 0;
262                     $attr = preg_replace("/^'[^']*'(\s+|$)/", '', $attr);
263                     break;
264                 }
265
266                 if (preg_match("%^([^\s\"']+)(\s+|$)%", $attr, $match))
267                     # value
268                     {
269                     $thisval = wp_kses_bad_protocol($match[1], $allowed_protocols);
270
271                     $attrarr[] = array ('name' => $attrname, 'value' => $thisval, 'whole' => "$attrname=\"$thisval\"", 'vless' => 'n');
272                     # We add quotes to conform to W3C's HTML spec.
273                     $working = 1;
274                     $mode = 0;
275                     $attr = preg_replace("%^[^\s\"']+(\s+|$)%", '', $attr);
276                 }
277
278                 break;
279         } # switch
280
281         if ($working == 0) # not well formed, remove and try again
282             {
283             $attr = wp_kses_html_error($attr);
284             $mode = 0;
285         }
286     } # while
287
288     if ($mode == 1)
289         # special case, for when the attribute list ends with a valueless
290         # attribute like "selected"
291         $attrarr[] = array ('name' => $attrname, 'value' => '', 'whole' => $attrname, 'vless' => 'y');
292
293     return $attrarr;
294 } # function wp_kses_hair
295
296 function wp_kses_check_attr_val($value, $vless, $checkname, $checkvalue)
297 ###############################################################################
298 # This function performs different checks for attribute values. The currently
299 # implemented checks are "maxlen", "minlen", "maxval", "minval" and "valueless"
300 # with even more checks to come soon.
301 ###############################################################################
302 {
303     $ok = true;
304
305     switch (strtolower($checkname)) {
306         case 'maxlen' :
307             # The maxlen check makes sure that the attribute value has a length not
308             # greater than the given value. This can be used to avoid Buffer Overflows
309             # in WWW clients and various Internet servers.
310
311             if (strlen($value) > $checkvalue)
312                 $ok = false;
313             break;
314
315         case 'minlen' :
316             # The minlen check makes sure that the attribute value has a length not
317             # smaller than the given value.
318
319             if (strlen($value) < $checkvalue)
320                 $ok = false;
321             break;
322
323         case 'maxval' :
324             # The maxval check does two things: it checks that the attribute value is
325             # an integer from 0 and up, without an excessive amount of zeroes or
326             # whitespace (to avoid Buffer Overflows). It also checks that the attribute
327             # value is not greater than the given value.
328             # This check can be used to avoid Denial of Service attacks.
329
330             if (!preg_match('/^\s{0,6}[0-9]{1,6}\s{0,6}$/', $value))
331                 $ok = false;
332             if ($value > $checkvalue)
333                 $ok = false;
334             break;
335
336         case 'minval' :
337             # The minval check checks that the attribute value is a positive integer,
338             # and that it is not smaller than the given value.
339
340             if (!preg_match('/^\s{0,6}[0-9]{1,6}\s{0,6}$/', $value))
341                 $ok = false;
342             if ($value < $checkvalue)
343                 $ok = false;
344             break;
345
346         case 'valueless' :
347             # The valueless check checks if the attribute has a value
348             # (like <a href="blah">) or not (<option selected>). If the given value
349             # is a "y" or a "Y", the attribute must not have a value.
350             # If the given value is an "n" or an "N", the attribute must have one.
351
352             if (strtolower($checkvalue) != $vless)
353                 $ok = false;
354             break;
355     } # switch
356
357     return $ok;
358 } # function wp_kses_check_attr_val
359
360 function wp_kses_bad_protocol($string, $allowed_protocols)
361 ###############################################################################
362 # This function removes all non-allowed protocols from the beginning of
363 # $string. It ignores whitespace and the case of the letters, and it does
364 # understand HTML entities. It does its work in a while loop, so it won't be
365 # fooled by a string like "javascript:javascript:alert(57)".
366 ###############################################################################
367 {
368     $string = wp_kses_no_null($string);
369     $string = preg_replace('/\xad+/', '', $string); # deals with Opera "feature"
370     $string2 = $string.'a';
371
372     while ($string != $string2) {
373         $string2 = $string;
374         $string = wp_kses_bad_protocol_once($string, $allowed_protocols);
375     } # while
376
377     return $string;
378 } # function wp_kses_bad_protocol
379
380 function wp_kses_no_null($string)
381 ###############################################################################
382 # This function removes any NULL characters in $string.
383 ###############################################################################
384 {
385     $string = preg_replace('/\0+/', '', $string);
386     $string = preg_replace('/(\\\\0)+/', '', $string);
387
388     return $string;
389 } # function wp_kses_no_null
390
391 function wp_kses_stripslashes($string)
392 ###############################################################################
393 # This function changes the character sequence  \"  to just  "
394 # It leaves all other slashes alone. It's really weird, but the quoting from
395 # preg_replace(//e) seems to require this.
396 ###############################################################################
397 {
398     return preg_replace('%\\\\"%', '"', $string);
399 } # function wp_kses_stripslashes
400
401 function wp_kses_array_lc($inarray)
402 ###############################################################################
403 # This function goes through an array, and changes the keys to all lower case.
404 ###############################################################################
405 {
406     $outarray = array ();
407
408     foreach ($inarray as $inkey => $inval) {
409         $outkey = strtolower($inkey);
410         $outarray[$outkey] = array ();
411
412         foreach ($inval as $inkey2 => $inval2) {
413             $outkey2 = strtolower($inkey2);
414             $outarray[$outkey][$outkey2] = $inval2;
415         } # foreach $inval
416     } # foreach $inarray
417
418     return $outarray;
419 } # function wp_kses_array_lc
420
421 function wp_kses_js_entities($string)
422 ###############################################################################
423 # This function removes the HTML JavaScript entities found in early versions of
424 # Netscape 4.
425 ###############################################################################
426 {
427     return preg_replace('%&\s*\{[^}]*(\}\s*;?|$)%', '', $string);
428 } # function wp_kses_js_entities
429
430 function wp_kses_html_error($string)
431 ###############################################################################
432 # This function deals with parsing errors in wp_kses_hair(). The general plan is
433 # to remove everything to and including some whitespace, but it deals with
434 # quotes and apostrophes as well.
435 ###############################################################################
436 {
437     return preg_replace('/^("[^"]*("|$)|\'[^\']*(\'|$)|\S)*\s*/', '', $string);
438 } # function wp_kses_html_error
439
440 function wp_kses_bad_protocol_once($string, $allowed_protocols)
441 ###############################################################################
442 # This function searches for URL protocols at the beginning of $string, while
443 # handling whitespace and HTML entities.
444 ###############################################################################
445 {
446     return preg_replace('/^((&[^;]*;|[\sA-Za-z0-9])*)'.'(:|&#58;|&#[Xx]3[Aa];)\s*/e', 'wp_kses_bad_protocol_once2("\\1", $allowed_protocols)', $string);
447 } # function wp_kses_bad_protocol_once
448
449 function wp_kses_bad_protocol_once2($string, $allowed_protocols)
450 ###############################################################################
451 # This function processes URL protocols, checks to see if they're in the white-
452 # list or not, and returns different data depending on the answer.
453 ###############################################################################
454 {
455     $string2 = wp_kses_decode_entities($string);
456     $string2 = preg_replace('/\s/', '', $string2);
457     $string2 = wp_kses_no_null($string2);
458     $string2 = preg_replace('/\xad+/', '', $string2);
459     # deals with Opera "feature"
460     $string2 = strtolower($string2);
461
462     $allowed = false;
463     foreach ($allowed_protocols as $one_protocol)
464         if (strtolower($one_protocol) == $string2) {
465             $allowed = true;
466             break;
467         }
468
469     if ($allowed)
470         return "$string2:";
471     else
472         return '';
473 } # function wp_kses_bad_protocol_once2
474
475 function wp_kses_normalize_entities($string)
476 ###############################################################################
477 # This function normalizes HTML entities. It will convert "AT&T" to the correct
478 # "AT&amp;T", "&#00058;" to "&#58;", "&#XYZZY;" to "&amp;#XYZZY;" and so on.
479 ###############################################################################
480 {
481     # Disarm all entities by converting & to &amp;
482