+// Sort query-strings if possible (Destructive and rough)
+// [OK] &&&&f=d&b&d&c&a=0dd => a=0dd&b&c&d&f=d
+// [OK] nothing==&eg=dummy&eg=padding&eg=foobar => eg=foobar
+function query_normalize($string = '', $equal = TRUE, $equal_cutempty = TRUE, $stortolower = TRUE)
+{
+ if (! is_string($string)) return '';
+ if ($stortolower) $string = strtolower($string);
+
+ $array = explode('&', $string);
+
+ // Remove '&' paddings
+ foreach(array_keys($array) as $key) {
+ if ($array[$key] == '') {
+ unset($array[$key]);
+ }
+ }
+
+ // Consider '='-sepalated input and paddings
+ if ($equal) {
+ $equals = $not_equals = array();
+ foreach ($array as $part) {
+ if (strpos($part, '=') === FALSE) {
+ $not_equals[] = $part;
+ } else {
+ list($key, $value) = explode('=', $part, 2);
+ $value = ltrim($value, '=');
+ if (! $equal_cutempty || $value != '') {
+ $equals[$key] = $value;
+ }
+ }
+ }
+
+ $array = & $not_equals;
+ foreach ($equals as $key => $value) {
+ $array[] = $key . '=' . $value;
+ }
+ unset($equals);
+ }
+
+ natsort($array);
+ return implode('&', $array);
+}
+
+// ---------------------
+// Area pickup
+
+// Pickup all of markup areas
+function area_pickup($string = '', $method = array())
+{
+ $area = array();
+ if (empty($method)) return $area;
+
+ // Anchor tag pair by preg_match and preg_match_all()
+ // [OK] <a href></a>
+ // [OK] <a href= >Good site!</a>
+ // [OK] <a href= "#" >test</a>
+ // [OK] <a href="http://nasty.example.com">visit http://nasty.example.com/</a>
+ // [OK] <a href=\'http://nasty.example.com/\' >discount foobar</a>
+ // [NG] <a href="http://ng.example.com">visit http://ng.example.com _not_ended_
+ $regex = '#<a\b[^>]*\bhref\b[^>]*>.*?</a\b[^>]*(>)#is';
+ if (isset($method['area_anchor'])) {
+ $areas = array();
+ $count = isset($method['asap']) ?
+ preg_match($regex, $string) :
+ preg_match_all($regex, $string, $areas);
+ if (! empty($count)) $area['area_anchor'] = $count;
+ }
+ if (isset($method['uri_anchor'])) {
+ $areas = array();
+ preg_match_all($regex, $string, $areas, PREG_SET_ORDER | PREG_OFFSET_CAPTURE);
+ foreach(array_keys($areas) as $_area) {
+ $areas[$_area] = array(
+ $areas[$_area][0][1], // Area start (<a href>)
+ $areas[$_area][1][1], // Area end (</a>)
+ );
+ }
+ if (! empty($areas)) $area['uri_anchor'] = $areas;
+ }
+
+ // phpBB's "BBCode" pair by preg_match and preg_match_all()
+ // [OK] [url][/url]
+ // [OK] [url]http://nasty.example.com/[/url]
+ // [OK] [link]http://nasty.example.com/[/link]
+ // [OK] [url=http://nasty.example.com]visit http://nasty.example.com/[/url]
+ // [OK] [link http://nasty.example.com/]buy something[/link]
+ $regex = '#\[(url|link|img|email)\b[^\]]*\].*?\[/\1\b[^\]]*(\])#is';
+ if (isset($method['area_bbcode'])) {
+ $areas = array();
+ $count = isset($method['asap']) ?
+ preg_match($regex, $string) :
+ preg_match_all($regex, $string, $areas, PREG_SET_ORDER);
+ if (! empty($count)) $area['area_bbcode'] = $count;
+ }
+ if (isset($method['uri_bbcode'])) {
+ $areas = array();
+ preg_match_all($regex, $string, $areas, PREG_SET_ORDER | PREG_OFFSET_CAPTURE);
+ foreach(array_keys($areas) as $_area) {
+ $areas[$_area] = array(
+ $areas[$_area][0][1], // Area start ([url])
+ $areas[$_area][2][1], // Area end ([/url])
+ );
+ }
+ if (! empty($areas)) $area['uri_bbcode'] = $areas;
+ }
+
+ // Various Wiki syntax
+ // [text_or_uri>text_or_uri]
+ // [text_or_uri:text_or_uri]
+ // [text_or_uri|text_or_uri]
+ // [text_or_uri->text_or_uri]
+ // [text_or_uri text_or_uri] // MediaWiki
+ // MediaWiki: [http://nasty.example.com/ visit http://nasty.example.com/]
+
+ return $area;
+}
+
+// If in doubt, it's a little doubtful
+// if (Area => inside <= Area) $brief += -1
+function area_measure($areas, & $array, $belief = -1, $a_key = 'area', $o_key = 'offset')
+{
+ if (! is_array($areas) || ! is_array($array)) return;
+
+ $areas_keys = array_keys($areas);
+ foreach(array_keys($array) as $u_index) {
+ $offset = isset($array[$u_index][$o_key]) ?
+ intval($array[$u_index][$o_key]) : 0;
+ foreach($areas_keys as $a_index) {
+ if (isset($array[$u_index][$a_key])) {
+ $offset_s = intval($areas[$a_index][0]);
+ $offset_e = intval($areas[$a_index][1]);
+ // [Area => inside <= Area]
+ if ($offset_s < $offset && $offset < $offset_e) {
+ $array[$u_index][$a_key] += $belief;
+ }
+ }
+ }
+ }
+}
+
+
+// ---------------------
+// Spam-uri pickup
+
+// Preprocess: Removing/Modifying uninterest part for URI detection
+function spam_uri_removing_hocus_pocus($binary = '', $method = array())
+{
+ $from = $to = array();
+
+ // Remove sequential spaces and too short lines
+ $length = 4 ; // 'http'(1) and '://'(2) and 'fqdn'(1)
+ if (is_array($method)) {
+ // '<a'(2) or 'href='(5) or '>'(1) or '</a>'(4)
+ // '[uri'(4) or ']'(1) or '[/uri]'(6)
+ if (isset($method['area_anchor']) || isset($method['uri_anchor']) ||
+ isset($method['area_bbcode']) || isset($method['uri_bbcode']))
+ $length = 1; // Seems not effective
+ }
+ $binary = strings($binary, $length, TRUE, FALSE); // Multibyte NOT needed
+
+ // Remove/Replace quoted-spaces within tags
+ $from[] = '#(<\w+ [^<>]*?\w ?= ?")([^"<>]*? [^"<>]*)("[^<>]*?>)#ie';
+ $to[] = "'$1' . str_replace(' ' , '%20' , trim('$2')) . '$3'";
+
+ // Remove words (has no '<>[]:') between spaces
+ $from[] = '/[ \t][\w.,()\ \t]+[ \t]/';
+ $to[] = ' ';
+
+ return preg_replace($from, $to, $binary);
+}
+
+// Preprocess: Domain exposure callback (See spam_uri_pickup_preprocess())
+// http://victim.example.org/?foo+site:nasty.example.com+bar
+// => http://nasty.example.com/?refer=victim.example.org
+// NOTE: 'refer=' is not so good for (at this time).
+// Consider about using IP address of the victim, try to avoid that.
+function _preg_replace_callback_domain_exposure($matches = array())
+{
+ $result = '';
+
+ // Preserve the victim URI as a complicity or ...
+ if (isset($matches[5])) {
+ $result =
+ $matches[1] . '://' . // scheme
+ $matches[2] . '/' . // victim.example.org
+ $matches[3]; // The rest of all (before victim)
+ }
+
+ // Flipped URI
+ if (isset($matches[4])) {
+ $result =
+ $matches[1] . '://' . // scheme
+ $matches[4] . // nasty.example.com
+ '/?refer=' . strtolower($matches[2]) . // victim.example.org
+ ' ' . $result;
+ }
+
+ return $result;
+}
+
+// Preprocess: minor-rawurldecode() and adding space(s) and something
+// to detect/count some URIs _if possible_
+// NOTE: It's maybe danger to var_dump(result). [e.g. 'javascript:']
+// [OK] http://victim.example.org/?site:nasty.example.org
+// [OK] http://victim.example.org/nasty.example.org
+// [OK] http://victim.example.org/go?http%3A%2F%2Fnasty.example.org
+// [OK] http://victim.example.org/http://nasty.example.org
+function spam_uri_pickup_preprocess($string = '', $method = array())
+{
+ if (! is_string($string)) return '';
+
+ // rawurldecode(), just to catch encoded 'http://path/to/file', not to change '%20' to ' '
+ $string = strtr(
+ $string,
+ array(
+ '%3A' => ':',
+ '%3a' => ':',
+ '%2F' => '/',
+ '%2f' => '/',
+ '%5C' => '\\',
+ '%5c' => '\\',
+ )
+ );
+
+ $string = spam_uri_removing_hocus_pocus($string, $method);
+
+ // Domain exposure (simple)
+ // http://victim.example.org/nasty.example.org/path#frag
+ // => http://nasty.example.org/?refer=victim.example.org and original
+ $string = preg_replace(
+ '#h?ttp://' .
+ '(' .
+ 'a9\.com/' . '|' .
+ 'aboutus\.org/' . '|' .
+ 'alexa\.com/data/details\?url=' . '|' .
+ 'ime\.(?:nu|st)/' . '|' . // 2ch.net
+ 'link\.toolbot\.com/' . '|' .
+ 'urlx\.org/' . '|' .
+ 'big5.51job.com/gate/big5/' . '|' .
+ 'big5.china.com/gate/big5/' . '|' .
+ 'big5.shippingchina.com:8080/' . '|' .
+ 'big5.xinhuanet.com/gate/big5/' . '|' .
+ 'bhomiyo.com/en.xliterate/' . '|' .
+ 'google.com/translate_c\?u=(?:http://)?' . '|' .
+ 'web.archive.org/web/2[^/]*/(?:http://)?' . '|' .
+ 'technorati.com/blogs/' .
+ ')' .
+ '([a-z0-9.%_-]+\.[a-z0-9.%_-]+)' . // nasty.example.org
+ '#i',
+ 'http://$2/?refer=$1 $0', // Preserve $0 or remove?
+ $string
+ );
+
+ // Domain exposure (site:) See _preg_replace_callback_domain_exposure()
+ $string = preg_replace_callback(
+ array(
+ '#(h?ttp)://' . // 1:Scheme
+ // 2:Host
+ '(' .
+ '(?:[a-z0-9_.-]+\.)?[a-z0-9_-]+\.[a-z0-9_-]+' .
+ // Something Google: http://www.google.com/supported_domains
+ // AltaVista: http://es.altavista.com/web/results?q=site%3Anasty.example.org+foobar
+ // Live Search: search.live.com
+ // MySpace: http://sads.myspace.com/Modules/Search/Pages/Search.aspx?_snip_&searchString=site:nasty.example.org
+ // (also searchresults.myspace.com)
+ // alltheweb.com
+ // search.bbc.co.uk
+ // search.orange.co.uk
+ // ...
+ ')' .
+ '/' .
+ //TODO: Specify URL-enable characters
+ '([a-z0-9?=&.,%_/\'\\\+-]+)' . // 3:path/?query=foo+bar+
+ '(?:\b|%20)site:([a-z0-9.%_-]+\.[a-z0-9.%_-]+)' . // 4:site:nasty.example.com
+ '()' . // 5:Preserve or remove?
+ '#i',
+ ),
+ '_preg_replace_callback_domain_exposure',
+ $string
+ );
+
+ // URI exposure (uriuri => uri uri)
+ $string = preg_replace(
+ array(
+ '#(?<! )(?:https?|ftp):/#i',
+ // '#[a-z][a-z0-9.+-]{1,8}://#i',
+ // '#[a-z][a-z0-9.+-]{1,8}://#i'
+ ),
+ ' $0',
+ $string
+ );
+
+ return $string;