Brief Analysis of Coral IP Library_PHP Tutorial-PHP Tutorial-php.cn

这不是什么新鲜事情了，很早之前就已经有人做出来了。
就是使用PHP操作纯真IP库或珊瑚虫IP库，根据来访者的IP得到所在的物理位置。

我先帖出代码。然后再慢慢一步步浅析出来。希望对想了解这一块的朋友们有帮助。

Only For PHP5的代码。会继续优化代码的。

class IpLocation{
    private $fp;
    private $wrydat;
    private $wrydat_version;
    private $ipnumber;
    private $firstip;
    private $lastip;
    private $ip_range_begin;
    private $ip_range_end;
    private $country;
    private $area;
    const REDIRECT_MODE_0 = 0;
    const REDIRECT_MODE_1 = 1;
    const REDIRECT_MODE_2 = 2;
    function __construct(){
        $args = func_get_args();
        $this->wrydat = func_num_args()>0?$args[0]:'CoralWry.dat';
        $this->initialize();
    }
    function __destruct(){
        fclose($this->fp);
    }
    private function initialize(){
        if(file_exists($this->wrydat))
            $this->fp = fopen($this->wrydat,'rb');
        $this->getipnumber();
        $this->getwryversion();
    }
    public function get($str){
        return $this->$str;
    }
    public function set($str,$val){
        $this->$str = $val;
    }
    private function getbyte($length,$offset=null){
        if(!is_null($offset)){
            fseek($this->fp,$offset,SEEK_SET);
        }
        $b = fread($this->fp,$length);
        return $b;
    }
/**
* Pack the IP address into binary data in big endian (high endian) format
* The data storage format is little endian (low endian), such as:
* 00 28 C6 DA 218.198.40.0 little endian
* 3F 28 C6 DA 218.198.40.0 little endian
* Such data cannot be compared by binary search, so the obtained IP data must first be converted to big endian
* @param using strrev $ip
* @return binary data in big endian format
*/
    private function packip($ip){
        return pack( "N", intval( ip2long( $ip)));
    }

    private function getlong($length=4, $offset=null){
        $chr=null;
        for($c=0;$length%4!=0&&$c<(4-$length%4);$c++){
            $chr .= chr(0);
        }
        $var = unpack( "Vlong", $this->getbyte($length, $offset).$chr);
        return $var['long'];
    }

    private function getwryversion(){
        $length = preg_match("/coral/i",$this->wrydat)?26:30;
        $this->wrydat_version = $this->getbyte($length, $this->firstip-$length);
    }

    private function getipnumber(){
        $this->firstip = $this->getlong();
        $this->lastip = $this->getlong();
        $this->ipnumber = ($this->lastip-$this->firstip)/7+1;
    }

    private function getstring($data="",$offset=null){
        $char = $this->getbyte(1,$offset);
        while(ord($char) > 0){
            $data .= $char;
            $char = $this->getbyte(1);
        }
        return $data;
    }

    private function iplocaltion($ip){
        $ip = $this->packip($ip);
        $low = 0;
        $high = $this->ipnumber-1;
        $ipposition = $this->lastip;
        while($low <= $high){
            $t = floor(($low+$high)/2);
            if($ip < strrev($this->getbyte(4,$this->firstip+$t*7))){
                $high = $t - 1;
            } else {
                if($ip > strrev($this->getbyte(4,$this->getlong(3)))){
                    $low = $t + 1;
                }else{
                    $ipposition = $this->firstip+$t*7;
                    break;
                }
            }
        }
        return $ipposition;
    }
    private function getarea(){
        $b = $this->getbyte(1);
        switch(ord($b)){
            case self::REDIRECT_MODE_0 :
                return "未知";
                break;
            case self::REDIRECT_MODE_1:
            case self::REDIRECT_MODE_2:
                return $this->getstring("",$this->getlong(3));
                break;
            default:
                return $this->getstring($b);
                break;
        }
    }
    public function getiplocation($ip){
        $ippos = $this->iplocaltion($ip);
        $this->ip_range_begin = long2ip($this->getlong(4,$ippos));
        $this->ip_range_end = long2ip($this->getlong(4,$this->getlong(3)));
        $b = $this->getbyte(1);
        switch (ord($b)){
            case self::REDIRECT_MODE_1:
                $b = $this->getbyte(1,$this->getlong(3));
                if(ord($b) == REDIRECT_MODE_2){
                    $countryoffset = $this->getlong(3);
                    $this->area = $this->getarea();
                    $this->country = $this->getstring("",$countryoffset);
                }else{
                    $this->country = $this->getstring($b);
                    $this->area    = $this->getarea();
                }
                break;

            case self::REDIRECT_MODE_2:
                    $countryoffset = $this->getlong(3);
                    $this->area = $this->getarea();
                    $this->country = $this->getstring("",$countryoffset);
                break;

            default:
                $this->country = $this->getstring($b);
                $this->area    = $this->getarea();
                break;
        }
    }
}
/* */
echo microtime();
echo "n";
$iploca = new IpLocation;
//$iploca = new IpLocation('QQWry.dat');
echo $iploca->get('wrydat_version');
echo "n";
echo $iploca->get('ipnumber');
echo "n";
$iploca->getiplocation('211.44.32.34');
/**/
echo $iploca->get('ip_range_begin');
echo "n";
echo $iploca->get('ip_range_end');
echo "n";
echo $iploca->get('country');
echo "n";
echo $iploca->get('area');

echo "n";
echo $iploca->get('lastip');
echo "n";
echo microtime();
echo "n";
unset($iploca);

参考资料：LumaQQ的纯真IP数据库格式详解

CoralWry.dat文件结构上分为3个区域：

文件头［固定8个字节］
数据区［不固定长度，记录IP的地址信息］
索引区［大小由文件头决定］

该文件数据的存储方式是：little endian。
在这里引用了谈谈Unicode编码里的关于little endian 与 big endian的区别

引用：

big endian和little endian是CPU处理多字节数的不同方式。例如“汉”字的Unicode编码是6C49。那么写到文件里时，究竟是将6C写在前面，还是将49写在前面？如果将6C写在前面，就是big endian。还是将49写在前面，就是little endian。

　　“endian”这个词出自《格列佛游记》。小人国的内战就源于吃鸡蛋时是究竟从大头(Big-Endian)敲开还是从小头(Little-Endian)敲开，由此曾发生过六次叛乱，其中一个皇帝送了命，另一个丢了王位。

　　我们一般将endian翻译成“字节序”，将big endian和little endian称作“大尾”和“小尾”。

文件头：
红色框框里的就是文件头，前4个字节是索引区的开始地址，后4个字节是索引区的结束地址。

如下图所示：

Brief Analysis of Coral IP Library_PHP Tutorial
点击放大

由于数据库是使用了little endian的字节库，所以我们需要把它倒过来。
把文件头的0-3的字节读取出来，再使用 unpack 函数把二进制数据转换为big endian格式的无符号整型。
处理后，索引区的开始地址位置是：00077450 ；索引区的结束地址位置是：000CE17C。
如果你手头上有UltraEdit的软件，可以打开CoralWry.dat文件，查找地址为：00077450 的位置，那就是IP地址索引区的开始。
如下图所示：

Brief Analysis of Coral IP Library_PHP Tutorial
点击放大

红色框框住那就是索引区的开始位置。