function reportResult(ISblamPost $post, $score, $cert)
{
if (!$this->add) {
return;
}
if ($score > 0.66 && $cert > 0.75) {
$this->plonker->addIPs($post->getAuthorIPs(), $score);
} else {
if ($score < -0.6 && $cert > 0.7) {
$this->plonker->removeIPs($post->getAuthorIPs());
}
}
}
function preTestPost(ISblamPost $p)
{
$this->addedhosts = array();
$this->ips = array();
// Check sender's IPs
foreach ($p->getAuthorIPs() as $ip) {
$this->checkIP($ip, "sender");
}
// Check all IPs of all linked websites
$links = $p->getLinks();
if ($links) {
foreach ($links as $link) {
$host = $link->getHostname();
$domain = $link->getDomain();
if ($host && $domain && $host !== $domain) {
$this->checkHost($host, "link", 0.75);
$this->checkHost($domain, "link", 0.75);
} else {
if ($host) {
$this->checkHost($host, "link");
}
if ($domain) {
$this->checkHost($domain, "link");
}
}
}
}
$this->checkHostList();
}
function testPost(ISblamPost $p)
{
$isWhiteIP = false;
$out = array();
foreach ($p->getAuthorIPs() as $ip) {
if (array_search($ip, $this->whitelist) !== false) {
$isWhiteIP = true;
} else {
$isWhiteIP = false;
}
}
if ($isWhiteIP) {
$out[] = array(-1.0, self::CERTAINITY_HIGH, "Sent from whitelisted IP");
}
return $out;
}
function testPost(ISblamPost $p)
{
// whitelist only direct connection (because other can be forged) and only when there aren't any objectionable hosts there
$out = array();
$firstIP = true;
$whitelisted = false;
foreach ($p->getAuthorIPs() as $ip) {
$rev = SblamURI::gethostbyaddr($ip);
if (!$rev) {
continue;
}
if (is_array($rev)) {
warn($rev, 'gethostbyaddr returned array');
$rev = reset($rev);
}
// WTF?
if (preg_match('!(?:\\.|^)(?:' . $this->isps . ')$!', $rev)) {
$out[] = array(0.5, self::CERTAINITY_LOW, "Sent from blacklisted ISP ({$rev})");
} else {
if ($firstIP && preg_match('!\\.(?:' . $this->whitelist . ')$!', $rev)) {
$whitelisted = true;
} else {
if (preg_match('!\\.(?:' . $this->blacklist . ')$!', $rev)) {
$out[] = array(0.35, self::CERTAINITY_LOW, "Sent from blacklisted TLD ({$rev})");
}
}
}
$firstIP = false;
}
if (!count($out) && $whitelisted) {
return array(-0.25, self::CERTAINITY_LOW, "Sent from whitelisted TLD ({$rev})");
}
if (count($out)) {
return $out;
}
}
function testPost(ISblamPost $p)
{
$h = $p->getHeaders();
if (!$h || count($h) < 2) {
return NULL;
}
// HTTP_HOST is hardcoded! :///
$out = array();
if (!empty($h['HTTP_MOD_SECURITY_MESSAGE'])) {
$out[] = array(1, self::CERTAINITY_HIGH, "mod_security warning");
}
// Buggy .Net always adds header which is only needed for large forms (and browsers tend not to use it)
if (!empty($h["HTTP_EXPECT"]) && false !== strpos($h['HTTP_EXPECT'], '100-') && strlen($p->getRawContent()) < 5000) {
$out[] = array(0.3, self::CERTAINITY_NORMAL, "100-expect .Net header");
}
// Bots tend to send these
if (!empty($h["HTTP_PRAGMA"])) {
$out[] = array(empty($h["HTTP_VIA"]) ? 0.3 : 0.1, self::CERTAINITY_LOW, "Pragma header");
}
if (!empty($h["HTTP_RANGE"])) {
$out[] = array(0.5, self::CERTAINITY_HIGH, "Range header");
}
if (!empty($h["HTTP_PROXY_CONNECTION"])) {
$out[] = array(0.2, self::CERTAINITY_LOW, "Proxy-Connection header");
}
if (!empty($h["HTTP_REFERER"]) && ($cnt = substr_count($h["HTTP_REFERER"], "http://")) > 1) {
$out[] = array(min(1.5, 0.5 + $cnt / 6), self::CERTAINITY_HIGH, "Multiple links in referrer");
}
if (($cnt = count($p->getAuthorIPs())) > 4) {
$out[] = array(($cnt - 2) / 10, $cnt > 7 ? self::CERTAINITY_HIGH : self::CERTAINITY_NORMAL, "Insane number of relays ({$cnt})");
}
// Unpatched IE!?
if (!empty($h["HTTP_USER_AGENT"]) && preg_match('/MSIE [456]\\.[0-9]; Windows (?:9|NT 5)/', $h['HTTP_USER_AGENT'])) {
$out[] = array(0.3, self::CERTAINITY_NORMAL, "Unpatched IE");
}
// Browsers almost always send these
if (empty($h["HTTP_ACCEPT"])) {
$out[] = array(0.7, self::CERTAINITY_NORMAL, "Missing Accept header");
}
if (empty($h["HTTP_USER_AGENT"])) {
$out[] = array(1, self::CERTAINITY_NORMAL, "Missing UA header");
}
if (empty($h["HTTP_ACCEPT_LANGUAGE"])) {
$out[] = array(0.5, self::CERTAINITY_NORMAL, "Missing Accept-Language header");
}
if (empty($h["HTTP_ACCEPT_ENCODING"]) && empty($h["HTTP_VIA"]) && (empty($h["HTTP_USER_AGENT"]) || false === strpos($h["HTTP_USER_AGENT"], 'Mozilla/4.0 (compatible; MSIE '))) {
$out[] = array(0.4, self::CERTAINITY_LOW, "Missing Accept-Encoding header");
}
if (!empty($h["HTTP_ACCEPT_CHARSET"])) {
$out[] = array(-0.2, self::CERTAINITY_LOW, "Has Accept-Charset header");
}
// Non-transparent proxy must add Via header
if (empty($h["HTTP_VIA"]) && (!empty($h['HTTP_X_FORWARDED_FOR']) || !empty($h['HTTP_MAX_FORWARDS']))) {
$out[] = array(0.2, self::CERTAINITY_LOW, "Lame proxy");
}
// TE: requires Connection:TE
if (!empty($h["HTTP_TE"]) && (empty($h['HTTP_CONNECTION']) || !preg_match('!\\bTE\\b!', $h['HTTP_CONNECTION']))) {
$out[] = array(0.2, self::CERTAINITY_NORMAL, "Invalid TE header");
}
// Googlebot doesn't post comments!
if (!empty($h['HTTP_USER_AGENT']) && preg_match('!Googlebot[/ -]|Slurp|Wget/|W3C_Validator|Advertise\\.com|nicebot|MMCrawler/|MSIECrawler|ia_archiver|WebaltBot/|nutbot\\.com|\\+http://search\\.!', $h['HTTP_USER_AGENT'])) {
$out[] = array(1, self::CERTAINITY_NORMAL, "Bots don't post comments");
}
// Headless browsers no thanks
if (!empty($h['HTTP_USER_AGENT']) && preg_match('!PhantomJS|CasperJS!', $h['HTTP_USER_AGENT'])) {
$out[] = array(1, self::CERTAINITY_HIGH, "Nice try, PhantomJS");
}
if (!empty($h['HTTP_USERAGENT']) || !empty($h['HTTP_USER_AGENT']) && preg_match('!^User-Agent!i', $h['HTTP_USER_AGENT'])) {
$out[] = array(1, self::CERTAINITY_NORMAL, "Really badly written bot");
}
// I assume multipart forms are too tricky for most bots
if (!empty($h['HTTP_CONTENT_LENGTH']) && !empty($h['HTTP_CONTENT_TYPE']) && preg_match('!^\\s*multipart/form-data\\s*;\\s*boundary\\s*=!i', $h['HTTP_CONTENT_TYPE'])) {
$out[] = array(-0.2, self::CERTAINITY_LOW, "Multipart form");
}
// browsers nicely decode and normalize paths, remove fragment part
if (($path = $p->getPath()) && preg_match('!&|^https?://|^//|/%7e|#|\\.\\./!i', $path)) {
$out[] = array(0.3, self::CERTAINITY_NORMAL, "Improperly encoded path");
}
if (!empty($h["HTTP_REFERER"]) && preg_match('!&|/%7e|\\.\\./!i', $h["HTTP_REFERER"])) {
$out[] = array(0.25, self::CERTAINITY_LOW, "Improperly encoded referer");
}
if (count($out)) {
return $out;
}
}
function reportResult(ISblamPost $p, $score, $cert)
{
if (!function_exists('apc_store')) {
throw new Exception("NO APC");
}
if ($score > 1.2 && $cert > 0.95) {
foreach ($p->getAuthorIPs() as $ip) {
apc_store('ip-ban:' . $ip, time() + 5, 5);
// block for 5 sec
}
}
}