Chris@0: <?php
Chris@0: /**
Chris@0:  * Zend Framework (http://framework.zend.com/)
Chris@0:  *
Chris@0:  * @link      http://github.com/zendframework/zf2 for the canonical source repository
Chris@0:  * @copyright Copyright (c) 2005-2015 Zend Technologies USA Inc. (http://www.zend.com)
Chris@0:  * @license   http://framework.zend.com/license/new-bsd New BSD License
Chris@0:  */
Chris@0: 
Chris@0: namespace Zend\Stdlib;
Chris@0: 
Chris@0: use Zend\Stdlib\StringWrapper\StringWrapperInterface;
Chris@0: 
Chris@0: /**
Chris@0:  * Utility class for handling strings of different character encodings
Chris@0:  * using available PHP extensions.
Chris@0:  *
Chris@0:  * Declared abstract, as we have no need for instantiation.
Chris@0:  */
Chris@0: abstract class StringUtils
Chris@0: {
Chris@0:     /**
Chris@0:      * Ordered list of registered string wrapper instances
Chris@0:      *
Chris@0:      * @var StringWrapperInterface[]
Chris@0:      */
Chris@0:     protected static $wrapperRegistry = null;
Chris@0: 
Chris@0:     /**
Chris@0:      * A list of known single-byte character encodings (upper-case)
Chris@0:      *
Chris@0:      * @var string[]
Chris@0:      */
Chris@0:     protected static $singleByteEncodings = [
Chris@0:         'ASCII', '7BIT', '8BIT',
Chris@0:         'ISO-8859-1', 'ISO-8859-2', 'ISO-8859-3', 'ISO-8859-4', 'ISO-8859-5',
Chris@0:         'ISO-8859-6', 'ISO-8859-7', 'ISO-8859-8', 'ISO-8859-9', 'ISO-8859-10',
Chris@0:         'ISO-8859-11', 'ISO-8859-13', 'ISO-8859-14', 'ISO-8859-15', 'ISO-8859-16',
Chris@0:         'CP-1251', 'CP-1252',
Chris@0:         // TODO
Chris@0:     ];
Chris@0: 
Chris@0:     /**
Chris@0:      * Is PCRE compiled with Unicode support?
Chris@0:      *
Chris@0:      * @var bool
Chris@0:      **/
Chris@0:     protected static $hasPcreUnicodeSupport = null;
Chris@0: 
Chris@0:     /**
Chris@0:      * Get registered wrapper classes
Chris@0:      *
Chris@0:      * @return string[]
Chris@0:      */
Chris@0:     public static function getRegisteredWrappers()
Chris@0:     {
Chris@0:         if (static::$wrapperRegistry === null) {
Chris@0:             static::$wrapperRegistry = [];
Chris@0: 
Chris@0:             if (extension_loaded('intl')) {
Chris@0:                 static::$wrapperRegistry[] = 'Zend\Stdlib\StringWrapper\Intl';
Chris@0:             }
Chris@0: 
Chris@0:             if (extension_loaded('mbstring')) {
Chris@0:                 static::$wrapperRegistry[] = 'Zend\Stdlib\StringWrapper\MbString';
Chris@0:             }
Chris@0: 
Chris@0:             if (extension_loaded('iconv')) {
Chris@0:                 static::$wrapperRegistry[] = 'Zend\Stdlib\StringWrapper\Iconv';
Chris@0:             }
Chris@0: 
Chris@0:             static::$wrapperRegistry[] = 'Zend\Stdlib\StringWrapper\Native';
Chris@0:         }
Chris@0: 
Chris@0:         return static::$wrapperRegistry;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Register a string wrapper class
Chris@0:      *
Chris@0:      * @param string $wrapper
Chris@0:      * @return void
Chris@0:      */
Chris@0:     public static function registerWrapper($wrapper)
Chris@0:     {
Chris@0:         $wrapper = (string) $wrapper;
Chris@12:         if (! in_array($wrapper, static::$wrapperRegistry, true)) {
Chris@0:             static::$wrapperRegistry[] = $wrapper;
Chris@0:         }
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Unregister a string wrapper class
Chris@0:      *
Chris@0:      * @param string $wrapper
Chris@0:      * @return void
Chris@0:      */
Chris@0:     public static function unregisterWrapper($wrapper)
Chris@0:     {
Chris@0:         $index = array_search((string) $wrapper, static::$wrapperRegistry, true);
Chris@0:         if ($index !== false) {
Chris@0:             unset(static::$wrapperRegistry[$index]);
Chris@0:         }
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Reset all registered wrappers so the default wrappers will be used
Chris@0:      *
Chris@0:      * @return void
Chris@0:      */
Chris@0:     public static function resetRegisteredWrappers()
Chris@0:     {
Chris@0:         static::$wrapperRegistry = null;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get the first string wrapper supporting the given character encoding
Chris@0:      * and supports to convert into the given convert encoding.
Chris@0:      *
Chris@0:      * @param string      $encoding        Character encoding to support
Chris@0:      * @param string|null $convertEncoding OPTIONAL character encoding to convert in
Chris@0:      * @return StringWrapperInterface
Chris@0:      * @throws Exception\RuntimeException If no wrapper supports given character encodings
Chris@0:      */
Chris@0:     public static function getWrapper($encoding = 'UTF-8', $convertEncoding = null)
Chris@0:     {
Chris@0:         foreach (static::getRegisteredWrappers() as $wrapperClass) {
Chris@0:             if ($wrapperClass::isSupported($encoding, $convertEncoding)) {
Chris@0:                 $wrapper = new $wrapperClass($encoding, $convertEncoding);
Chris@0:                 $wrapper->setEncoding($encoding, $convertEncoding);
Chris@0:                 return $wrapper;
Chris@0:             }
Chris@0:         }
Chris@0: 
Chris@0:         throw new Exception\RuntimeException(
Chris@0:             'No wrapper found supporting "' . $encoding . '"'
Chris@0:             . (($convertEncoding !== null) ? ' and "' . $convertEncoding . '"' : '')
Chris@0:         );
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Get a list of all known single-byte character encodings
Chris@0:      *
Chris@0:      * @return string[]
Chris@0:      */
Chris@0:     public static function getSingleByteEncodings()
Chris@0:     {
Chris@0:         return static::$singleByteEncodings;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Check if a given encoding is a known single-byte character encoding
Chris@0:      *
Chris@0:      * @param string $encoding
Chris@0:      * @return bool
Chris@0:      */
Chris@0:     public static function isSingleByteEncoding($encoding)
Chris@0:     {
Chris@0:         return in_array(strtoupper($encoding), static::$singleByteEncodings);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Check if a given string is valid UTF-8 encoded
Chris@0:      *
Chris@0:      * @param string $str
Chris@0:      * @return bool
Chris@0:      */
Chris@0:     public static function isValidUtf8($str)
Chris@0:     {
Chris@0:         return is_string($str) && ($str === '' || preg_match('/^./su', $str) == 1);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * Is PCRE compiled with Unicode support?
Chris@0:      *
Chris@0:      * @return bool
Chris@0:      */
Chris@0:     public static function hasPcreUnicodeSupport()
Chris@0:     {
Chris@0:         if (static::$hasPcreUnicodeSupport === null) {
Chris@0:             ErrorHandler::start();
Chris@0:             static::$hasPcreUnicodeSupport = defined('PREG_BAD_UTF8_OFFSET_ERROR') && preg_match('/\pL/u', 'a') == 1;
Chris@0:             ErrorHandler::stop();
Chris@0:         }
Chris@0:         return static::$hasPcreUnicodeSupport;
Chris@0:     }
Chris@0: }