Ajax software     Free javascripts 
↑


Main Page

{

// use cURL to read the data from $url

// NOTE: additional settings are required when accessing the web through a proxy

$ch = curl_init();

curl_setopt ($ch, CURLOPT_URL, $url);

curl_setopt ($ch, CURLOPT_HEADER, 1);

curl_setopt ($ch, CURLOPT_RETURNTRANSFER, 1);

curl_setopt ($ch, CURLOPT_FOLLOWLOCATION, 1);

curl_setopt ($ch, CURLOPT_TIMEOUT, 60);

$result = curl_exec($ch);

curl_close($ch);

// use _parseListURL to parse the list of IPs and user agents

$lists = SimpleCloak::_parseListURL($result, $ua_regex, $ip_regex);

// if the user agents and IPs weren’t retrieved, we cancel the update

if (!$lists[‘ua_list’] || !$lists[‘ip_list’]) return;

// lock the cloack_data table to avoid concurrency problems

mysql_query(‘LOCK TABLES cloak_data WRITE’);

// delete all the existing data for $spider_name

SimpleCloak::_deleteSpiderData($spider_name);

// insert the list of user agents for the spider

foreach ($lists[‘ua_list’] as $ua) {

SimpleCloak::_insertSpiderData($spider_name, ‘UA’, $ua);

}

// insert the list of IPs for the spider

foreach ($lists[‘ip_list’] as $ip) {

SimpleCloak::_insertSpiderData($spider_name, ‘IP’, $ip);

}

// release the table lock

mysql_query(‘UNLOCK TABLES’);

}

// helper function used to parse lists of user agents and IPs

function _parseListURL($data, $ua_regex, $ip_regex)

{

$ua_list_ret = preg_match_all($ua_regex, $data, $ua_list);

$ip_list_ret = preg_match_all($ip_regex, $data, $ip_list);

return array(‘ua_list’ => $ua_list[1], ‘ip_list’ => $ip_list[1]);

}

// inserts a new row of data to the cloaking table

function _insertSpiderData($spider_name, $record_type, $value)

{

// escape input data

$spider_name = mysql_escape_string($spider_name);

$record_type = mysql_escape_string($record_type);

$value = mysql_escape_string($value);

229

Chapter 11: Cloaking, Geo-Targeting, and IP Delivery

c11.qxd:c11 11:01 229

Ajax software     Free javascripts 
→