gistfile1.php

<?php
require_once 'Zend/Loader/Autoloader.php';
Zend_Loader_Autoloader::getInstance()->setFallbackAutoloader(true);

//@see http://www.tokumaru.org/d/20090914.html

class Zenmondo_Validate_EncodingUTF8 extends Zend_Validate_Abstract
{
    public function isValid($value)
    {
        if (!mb_check_encoding($value, 'UTF-8')) {
            return false;
        }

        return true;
    }
}

//borrowd from Zend_Validate_Regex
class Zenmondo_Validate_MbEreg extends Zend_Validate_Regex
{
    //@override
    public function isValid($value)
    {
        if (!is_string($value) && !is_int($value) && !is_float($value)) {
            $this->_error(self::INVALID);
            return false;
        }

        $this->_setValue($value);

        $status = mb_ereg($this->_pattern, $value);
        if (false === $status) {
            //require_once 'Zend/Validate/Exception.php';
            throw new Exception("Internal error matching pattern '$this->_pattern' against value '$value'");
        }
        if (!$status) {
            $this->_error(self::NOT_MATCH);
            return false;
        }
        return true;

    }
}


$validator = new Zend_Validate();
$validator->addValidator(new Zenmondo_Validate_EncodingUTF8)
          ->addValidator(new Zenmondo_Validate_MbEreg('\\A[[:^cntrl:]]*\\z'))
          ->addValidator(new Zend_Validate_StringLength(0, 10));

var_dump(mb_regex_encoding()); //EUC-JP
var_dump($validator->isValid("test")); //true
var_dump($validator->isValid("\xc0\xaf")); //false
var_dump($validator->isValid("%C0%AF")); //true
var_dump($validator->isValid(rawurldecode('%C0%AF'))); //false
//var_dump($validator->isValid(rawurldecode('C0%AF'))); //throw Exception

echo '-----', PHP_EOL;

//mb_regex_encoding('ISO-8859-1');
//mb_regex_encoding('UTF-8');
mb_regex_encoding('ASCII');
var_dump($validator->isValid("test"));       //true
var_dump($validator->isValid("日本語"));       //true
var_dump($validator->isValid("\xc0\xaf"));   //false
var_dump($validator->isValid("%C0%AF")); //true
var_dump($validator->isValid(rawurldecode('%C0%AF'))); //false
var_dump($validator->isValid(rawurldecode('C0%AF'))); // false
var_dump($validator->isValid(rawurldecode(rawurlencode("日本語"))));       //true

echo '-----', PHP_EOL;
mb_regex_encoding('ISO-8859-1');
var_dump($validator->isValid(rawurldecode('C0%AF'))); //false