Added redirections

[pukiwiki/pukiwiki_sandbox.git] / spam / spam_pickup.php

diff --git a/spam/spam_pickup.php b/spam/spam_pickup.php
index b0ecc99..443e69a 100644 (file)
--- a/spam/spam_pickup.php
+++ b/spam/spam_pickup.php
@@ -1,10 +1,14 @@
 <?php
-// $Id: spam_pickup.php,v 1.66 2009/01/02 09:30:50 henoheno Exp $
-// Copyright (C) 2006-2007 PukiWiki Developers Team
+// $Id: spam_pickup.php,v 1.71 2009/01/04 08:56:07 henoheno Exp $
+// Copyright (C) 2006-2009 PukiWiki Developers Team
 // License: GPL v2 or (at your option) any later version
 //
 // Functions for Concept-work of spam-uri metrics
 //
+// (PHP 4 >= 4.3.0): preg_match_all(PREG_OFFSET_CAPTURE): $method['uri_XXX'] related feature
+//
+
+if (! defined('DOMAIN_INI_FILE')) define('DOMAIN_INI_FILE', 'domain.ini.php');
 
 // ---------------------
 // URI pickup
@@ -14,18 +18,18 @@
 // [OK] http://nasty.example.org:80/foo/xxx#nasty_string/bar
 // [OK] ftp://nasty.example.org:80/dfsdfs
 // [OK] ftp://cnn.example.com&story=breaking_news@10.0.0.1/top_story.htm (from RFC3986)
+// Not available for: IDN(ignored)
 function uri_pickup($string = '')
 {
        if (! is_string($string)) return array();
 
-       // Not available for: IDN(ignored)
        $array = array();
        preg_match_all(
                // scheme://userinfo@host:port/path/or/pathinfo/maybefile.and?query=string#fragment
                // Refer RFC3986 (Regex below is not strict)
                '#(\b[a-z][a-z0-9.+-]{1,8}):[/\\\]+' .          // 1: Scheme
                '(?:' .
-                       '([^\s<>"\'\[\]/\#?@]*)' .              // 2: Userinfo (Username)
+                       '([^\s<>"\'\[\]/\#?@]*)' .              // 2: Userinfo (Username and/or password)
                '@)?' .
                '(' .
                        // 3: Host
@@ -34,7 +38,7 @@ function uri_pickup($string = '')
                        '[a-z0-9_-][a-z0-9_.-]+[a-z0-9_-]' .    // hostname(FQDN) : foo.example.org
                ')' .
                '(?::([0-9]*))?' .                                      // 4: Port
-               '((?:/+[^\s<>"\'\[\]/\#]+)*/+)?' .      // 5: Directory path
+               '((?:/+[^\s<>"\'\[\]/\#?]+)*/+)?' .     // 5: Directory path
                '([^\s<>"\'\[\]\#?]+)?' .                       // 6: File?
                '(?:\?([^\s<>"\'\[\]\#]+))?' .          // 7: Query string
                '(?:\#([a-z0-9._~%!$&\'()*+,;=:@-]*))?' .       // 8: Fragment
@@ -86,27 +90,43 @@ function uri_pickup_implode($uri = array())
                $tmp[] = & $uri['scheme'];
                $tmp[] = '://';
        }
+
        if (isset($uri['userinfo']) && $uri['userinfo'] !== '') {
                $tmp[] = & $uri['userinfo'];
                $tmp[] = '@';
+       } else if (isset($uri['user']) || isset($uri['pass'])) {
+               if (isset($uri['user']) && $uri['user'] !== '') {
+                       $tmp[] = & $uri['user'];
+               }
+               $tmp[] = ':';
+               if (isset($uri['pass']) && $uri['pass'] !== '') {
+                       $tmp[] = & $uri['pass'];
+               }
+               $tmp[] = '@';
        }
+
        if (isset($uri['host']) && $uri['host'] !== '') {
                $tmp[] = & $uri['host'];
        }
+
        if (isset($uri['port']) && $uri['port'] !== '') {
                $tmp[] = ':';
                $tmp[] = & $uri['port'];
        }
+
        if (isset($uri['path']) && $uri['path'] !== '') {
                $tmp[] = & $uri['path'];
        }
+
        if (isset($uri['file']) && $uri['file'] !== '') {
                $tmp[] = & $uri['file'];
        }
+
        if (isset($uri['query']) && $uri['query'] !== '') {
                $tmp[] = '?';
                $tmp[] = & $uri['query'];
        }
+
        if (isset($uri['fragment']) && $uri['fragment'] !== '') {
                $tmp[] = '#';
                $tmp[] = & $uri['fragment'];
@@ -115,6 +135,7 @@ function uri_pickup_implode($uri = array())
        return implode('', $tmp);
 }
 
+
 // ---------------------
 // URI normalization
 
@@ -821,4 +842,76 @@ function spam_uri_pickup($string = '', $method = array())
        return $array;
 }
 
+// Rough hostname checker
+// TODO: Strict digit, 0x, CIDR, '999.999.999.999', ':', '::G'
+function is_ip($string = '')
+{
+       if (! is_string($string)) return FALSE;
+
+       if (strpos($string, ':') !== FALSE) {
+               return 6;       // Seems IPv6
+       }
+
+       if (preg_match('/^' .
+               '(?:[0-9]{1,3}\.){3}[0-9]{1,3}' . '|' .
+               '(?:[0-9]{1,3}\.){1,3}'         . '$/',
+               $string)) {
+               return 4;       // Seems IPv4(dot-decimal)
+       }
+
+       return FALSE;   // Seems not IP
+}
+
+// Check responsibility-root of the FQDN
+// 'foo.bar.example.com'        => 'example.com'        (.com        has the last whois for it)
+// 'foo.bar.example.au'         => 'example.au'         (.au         has the last whois for it)
+// 'foo.bar.example.edu.au'     => 'example.edu.au'     (.edu.au     has the last whois for it)
+// 'foo.bar.example.act.edu.au' => 'example.act.edu.au' (.act.edu.au has the last whois for it)
+function whois_responsibility($fqdn = 'foo.bar.example.com', $parent = FALSE, $implicit = TRUE)
+{
+       static $domain;
+
+       if ($fqdn === NULL) {
+               $domain = NULL; // Unset
+               return '';
+       }
+       if (! is_string($fqdn)) return '';
+
+       if (is_ip($fqdn)) return $fqdn;
+
+       if (! isset($domain)) {
+               $domain = array();
+               if (file_exists(DOMAIN_INI_FILE)) {
+                       include(DOMAIN_INI_FILE);       // Set
+               }
+       }
+
+       $result  = array();
+       $dcursor = & $domain;
+       $array   = array_reverse(explode('.', $fqdn));
+       $i = 0;
+       while(TRUE) {
+               if (! isset($array[$i])) break;
+               $acursor = $array[$i];
+               if (is_array($dcursor) && isset($dcursor[$acursor])) {
+                       $result[] = & $array[$i];
+                       $dcursor  = & $dcursor[$acursor];
+               } else {
+                       if (! $parent && isset($acursor)) {
+                               $result[] = & $array[$i];       // Whois servers must know this subdomain
+                       }
+                       break;
+               }
+               ++$i;
+       }
+
+       // Implicit responsibility: Top-Level-Domains must not be yours
+       // 'bar.foo.something' => 'foo.something'
+       if ($implicit && count($result) == 1 && count($array) > 1) {
+               $result[] = & $array[1];
+       }
+
+       return $result ? implode('.', array_reverse($result)) : '';
+}
+
 ?>