array(), 'i' => array(), 'a' => array('href' => 1, /*'title' => array('valueless' => 'n'),*/ 'class' => 1, 'target' => 1), 'div' => array('class' => 1), 'p' => array('align' => 1, 'dummy' => array('valueless' => 'y')), 'img' => array('src' => 1), # FIXME 'font' => array('size' => array('minval' => 4, 'maxval' => 20)), 'br' => array()); return $allowed; } static function kses($string, $allowed_html, $allowed_protocols = array('http', 'https', 'ftp', 'news', 'nntp', 'telnet', 'gopher', 'mailto')) ############################################################################### # This function makes sure that only the allowed HTML element names, attribute # names and attribute values plus only sane HTML entities will occur in # $string. You have to remove any slashes from PHP's magic quotes before you # call this function. ############################################################################### { $string = CKses::kses_no_null($string); $string = CKses::kses_js_entities($string); $string = CKses::kses_normalize_entities($string); $string = CKses::kses_hook($string); $allowed_html_fixed = CKses::kses_array_lc($allowed_html); /* @since 2.6, replace the checking with the regex below */ $string = preg_replace('#<(/*\s*)(alert|applet|basefont|base|behavior|bgsound|blink|body|embed|expression|form|frameset|frame|head|html|ilayer|iframe|input|layer|link|meta|object|plaintext|style|script|textarea|title|xml|xss)([^>]*)>#is', "<\\1\\2\\3>", $string); return $string; //return CKses::kses_split($string, $allowed_html_fixed, $allowed_protocols); } # function kses public static function kses_hook($string) ############################################################################### # You add any kses hooks here. ############################################################################### { return $string; } # function kses_hook function kses_version() ############################################################################### # This function returns kses' version number. ############################################################################### { return '0.2.2'; } # function kses_version function kses_split($string, $allowed_html, $allowed_protocols) ############################################################################### # This function searches for HTML tags, no matter how malformed. It also # matches stray ">" characters. ############################################################################### { return preg_replace('%(<'. # EITHER: < '[^>]*'. # things that aren't > '(>|$)'. # > or end of string '|>)%e', # OR: just a > "CKses::kses_split2('\\1', \$allowed_html, ". '$allowed_protocols)', $string); } # function kses_split function kses_split2($string, $allowed_html, $allowed_protocols) ############################################################################### # This function does a lot of work. It rejects some very malformed things # like <:::>. It returns an empty string, if the element isn't allowed (look # ma, no strip_tags()!). Otherwise it splits the tag into an element and an # attribute list. ############################################################################### { $string = CKses::kses_stripslashes($string); if (substr($string, 0, 1) != '<') return '>'; # It matched a ">" character if (!preg_match('%^<\s*(/\s*)?([a-zA-Z0-9]+)([^>]*)>?$%', $string, $matches)) return ''; # It's seriously malformed $slash = trim($matches[1]); $elem = $matches[2]; $attrlist = $matches[3]; if (!@isset($allowed_html[strtolower($elem)])) return ''; # They are using a not allowed HTML element if ($slash != '') return "<$slash$elem>"; # No attributes are allowed for closing elements return CKses::kses_attr("$slash$elem", $attrlist, $allowed_html, $allowed_protocols); } # function kses_split2 function kses_attr($element, $attr, $allowed_html, $allowed_protocols) ############################################################################### # This function removes all attributes, if none are allowed for this element. # If some are allowed it calls CKses::kses_hair() to split them further, and then it # builds up new HTML code from the data that CKses::kses_hair() returns. It also # removes "<" and ">" characters, if there are any left. One more thing it # does is to check if the tag has a closing XHTML slash, and if it does, # it puts one in the returned code as well. ############################################################################### { # Is there a closing XHTML slash at the end of the attributes? $xhtml_slash = ''; if (preg_match('%\s/\s*$%', $attr)) $xhtml_slash = ' /'; # Are any attributes allowed at all for this element? if (@count($allowed_html[strtolower($element)]) == 0) return "<$element$xhtml_slash>"; # Split it $attrarr = CKses::kses_hair($attr, $allowed_protocols); # Go through $attrarr, and save the allowed attributes for this element # in $attr2 $attr2 = ''; foreach ($attrarr as $arreach) { if (!@isset($allowed_html[strtolower($element)] [strtolower($arreach['name'])])) continue; # the attribute is not allowed $current = $allowed_html[strtolower($element)] [strtolower($arreach['name'])]; if (!is_array($current)) $attr2 .= ' '.$arreach['whole']; # there are no checks else { # there are some checks $ok = true; foreach ($current as $currkey => $currval) if (!CKses::kses_check_attr_val($arreach['value'], $arreach['vless'], $currkey, $currval)) { $ok = false; break; } if ($ok) $attr2 .= ' '.$arreach['whole']; # it passed them } # if !is_array($current) } # foreach # Remove any "<" or ">" characters $attr2 = preg_replace('/[<>]/', '', $attr2); return "<$element$attr2$xhtml_slash>"; } # function kses_attr function kses_hair($attr, $allowed_protocols) ############################################################################### # This function does a lot of work. It parses an attribute list into an array # with attribute data, and tries to do the right thing even if it gets weird # input. It will add quotes around attribute values that don't have any quotes # or apostrophes around them, to make it easier to produce HTML code that will # conform to W3C's HTML specification. It will also remove bad URL protocols # from attribute values. ############################################################################### { $attrarr = array(); $mode = 0; $attrname = ''; # Loop through the whole attribute list while (strlen($attr) != 0) { $working = 0; # Was the last operation successful? switch ($mode) { case 0: # attribute name, href for instance if (preg_match('/^([-a-zA-Z]+)/', $attr, $match)) { $attrname = $match[1]; $working = $mode = 1; $attr = preg_replace('/^[-a-zA-Z]+/', '', $attr); } break; case 1: # equals sign or valueless ("selected") if (preg_match('/^\s*=\s*/', $attr)) # equals sign { $working = 1; $mode = 2; $attr = preg_replace('/^\s*=\s*/', '', $attr); break; } if (preg_match('/^\s+/', $attr)) # valueless { $working = 1; $mode = 0; $attrarr[] = array ('name' => $attrname, 'value' => '', 'whole' => $attrname, 'vless' => 'y'); $attr = preg_replace('/^\s+/', '', $attr); } break; case 2: # attribute value, a URL after href= for instance if (preg_match('/^"([^"]*)"(\s+|$)/', $attr, $match)) # "value" { $thisval = CKses::kses_bad_protocol($match[1], $allowed_protocols); $attrarr[] = array ('name' => $attrname, 'value' => $thisval, 'whole' => "$attrname=\"$thisval\"", 'vless' => 'n'); $working = 1; $mode = 0; $attr = preg_replace('/^"[^"]*"(\s+|$)/', '', $attr); break; } if (preg_match("/^'([^']*)'(\s+|$)/", $attr, $match)) # 'value' { $thisval = CKses::kses_bad_protocol($match[1], $allowed_protocols); $attrarr[] = array ('name' => $attrname, 'value' => $thisval, 'whole' => "$attrname='$thisval'", 'vless' => 'n'); $working = 1; $mode = 0; $attr = preg_replace("/^'[^']*'(\s+|$)/", '', $attr); break; } if (preg_match("%^([^\s\"']+)(\s+|$)%", $attr, $match)) # value { $thisval = CKses::kses_bad_protocol($match[1], $allowed_protocols); $attrarr[] = array ('name' => $attrname, 'value' => $thisval, 'whole' => "$attrname=\"$thisval\"", 'vless' => 'n'); # We add quotes to conform to W3C's HTML spec. $working = 1; $mode = 0; $attr = preg_replace("%^[^\s\"']+(\s+|$)%", '', $attr); } break; } # switch if ($working == 0) # not well formed, remove and try again { $attr = CKses::kses_html_error($attr); $mode = 0; } } # while if ($mode == 1) # special case, for when the attribute list ends with a valueless # attribute like "selected" $attrarr[] = array ('name' => $attrname, 'value' => '', 'whole' => $attrname, 'vless' => 'y'); return $attrarr; } # function kses_hair function kses_check_attr_val($value, $vless, $checkname, $checkvalue) ############################################################################### # This function performs different checks for attribute values. The currently # implemented checks are "maxlen", "minlen", "maxval", "minval" and "valueless" # with even more checks to come soon. ############################################################################### { $ok = true; switch (strtolower($checkname)) { case 'maxlen': # The maxlen check makes sure that the attribute value has a length not # greater than the given value. This can be used to avoid Buffer Overflows # in WWW clients and various Internet servers. if (strlen($value) > $checkvalue) $ok = false; break; case 'minlen': # The minlen check makes sure that the attribute value has a length not # smaller than the given value. if (strlen($value) < $checkvalue) $ok = false; break; case 'maxval': # The maxval check does two things: it checks that the attribute value is # an integer from 0 and up, without an excessive amount of zeroes or # whitespace (to avoid Buffer Overflows). It also checks that the attribute # value is not greater than the given value. # This check can be used to avoid Denial of Service attacks. if (!preg_match('/^\s{0,6}[0-9]{1,6}\s{0,6}$/', $value)) $ok = false; if ($value > $checkvalue) $ok = false; break; case 'minval': # The minval check checks that the attribute value is a positive integer, # and that it is not smaller than the given value. if (!preg_match('/^\s{0,6}[0-9]{1,6}\s{0,6}$/', $value)) $ok = false; if ($value < $checkvalue) $ok = false; break; case 'valueless': # The valueless check checks if the attribute has a value # (like ) or not (). If the given value # is a "y" or a "Y", the attribute must not have a value. # If the given value is an "n" or an "N", the attribute must have one. if (strtolower($checkvalue) != $vless) $ok = false; break; } # switch return $ok; } # function kses_check_attr_val function kses_bad_protocol($string, $allowed_protocols) ############################################################################### # This function removes all non-allowed protocols from the beginning of # $string. It ignores whitespace and the case of the letters, and it does # understand HTML entities. It does its work in a while loop, so it won't be # fooled by a string like "javascript:javascript:alert(57)". ############################################################################### { $string = CKses::kses_no_null($string); $string = preg_replace('/\xad+/', '', $string); # deals with Opera "feature" $string2 = $string.'a'; while ($string != $string2) { $string2 = $string; $string = CKses::kses_bad_protocol_once($string, $allowed_protocols); } # while return $string; } # function kses_bad_protocol public static function kses_no_null($string) ############################################################################### # This function removes any NULL characters in $string. ############################################################################### { $string = preg_replace('/\0+/', '', $string); $string = preg_replace('/(\\\\0)+/', '', $string); return $string; } # function kses_no_null function kses_stripslashes($string) ############################################################################### # This function changes the character sequence \" to just " # It leaves all other slashes alone. It's really weird, but the quoting from # preg_replace(//e) seems to require this. ############################################################################### { return preg_replace('%\\\\"%', '"', $string); } # function kses_stripslashes public static function kses_array_lc($inarray) ############################################################################### # This function goes through an array, and changes the keys to all lower case. ############################################################################### { $outarray = array(); foreach ($inarray as $inkey => $inval) { $outkey = strtolower($inkey); $outarray[$outkey] = array(); foreach ($inval as $inkey2 => $inval2) { $outkey2 = strtolower($inkey2); $outarray[$outkey][$outkey2] = $inval2; } # foreach $inval } # foreach $inarray return $outarray; } # function kses_array_lc public static function kses_js_entities($string) ############################################################################### # This function removes the HTML JavaScript entities found in early versions of # Netscape 4. ############################################################################### { return preg_replace('%&\s*\{[^}]*(\}\s*;?|$)%', '', $string); } # function kses_js_entities function kses_html_error($string) ############################################################################### # This function deals with parsing errors in CKses::kses_hair(). The general plan is # to remove everything to and including some whitespace, but it deals with # quotes and apostrophes as well. ############################################################################### { return preg_replace('/^("[^"]*("|$)|\'[^\']*(\'|$)|\S)*\s*/', '', $string); } # function kses_html_error function kses_bad_protocol_once($string, $allowed_protocols) ############################################################################### # This function searches for URL protocols at the beginning of $string, while # handling whitespace and HTML entities. ############################################################################### { return preg_replace('/^((&[^;]*;|[\sA-Za-z0-9])*)'. '(:|:|&#[Xx]3[Aa];)\s*/e', 'CKses::kses_bad_protocol_once2("\\1", $allowed_protocols)', $string); } # function kses_bad_protocol_once function kses_bad_protocol_once2($string, $allowed_protocols) ############################################################################### # This function processes URL protocols, checks to see if they're in the white- # list or not, and returns different data depending on the answer. ############################################################################### { $string2 = CKses::kses_decode_entities($string); $string2 = preg_replace('/\s/', '', $string2); $string2 = CKses::kses_no_null($string2); $string2 = preg_replace('/\xad+/', '', $string2); # deals with Opera "feature" $string2 = strtolower($string2); $allowed = false; foreach ($allowed_protocols as $one_protocol) if (strtolower($one_protocol) == $string2) { $allowed = true; break; } if ($allowed) return "$string2:"; else return ''; } # function kses_bad_protocol_once2 public static function kses_normalize_entities($string) ############################################################################### # This function normalizes HTML entities. It will convert "AT&T" to the correct # "AT&T", ":" to ":", "&#XYZZY;" to "&#XYZZY;" and so on. ############################################################################### { # Disarm all entities by converting & to & $string = str_replace('&', '&', $string); # Change back the allowed entities in our entity whitelist $string = preg_replace('/&([A-Za-z][A-Za-z0-9]{0,19});/', '&\\1;', $string); $string = preg_replace_callback('/&#0*([0-9]{1,5});/', 'CKses::kses_normalize_entities2', $string); $string = preg_replace('/&#([Xx])0*(([0-9A-Fa-f]{2}){1,2});/', '&#\\1\\2;', $string); return $string; } # function kses_normalize_entities public static function kses_normalize_entities2($i) ############################################################################### # This function helps CKses::kses_normalize_entities() to only accept 16 bit values # and nothing more for &#number; entities. ############################################################################### { $i = $i[1]; return (($i > 65535) ? "&#$i;" : "&#$i;"); } # function kses_normalize_entities2 public static function kses_decode_entities($string) ############################################################################### # This function decodes numeric HTML entities (A and A). It doesn't # do anything with other entities like ä, but we don't need them in the # URL protocol whitelisting system anyway. ############################################################################### { $string = preg_replace('/&#([0-9]+);/e', 'chr("\\1")', $string); $string = preg_replace('/&#[Xx]([0-9A-Fa-f]+);/e', 'chr(hexdec("\\1"))', $string); return $string; } # function kses_decode_entities }