Chris@0: <?php
Chris@0: 
Chris@0: /*
Chris@0:  * This file is part of the Symfony package.
Chris@0:  *
Chris@0:  * (c) Fabien Potencier <fabien@symfony.com>
Chris@0:  *
Chris@0:  * For the full copyright and license information, please view the LICENSE
Chris@0:  * file that was distributed with this source code.
Chris@0:  */
Chris@0: 
Chris@0: namespace Symfony\Component\CssSelector\Parser\Tokenizer;
Chris@0: 
Chris@0: /**
Chris@0:  * CSS selector tokenizer escaping applier.
Chris@0:  *
Chris@0:  * This component is a port of the Python cssselect library,
Chris@0:  * which is copyright Ian Bicking, @see https://github.com/SimonSapin/cssselect.
Chris@0:  *
Chris@0:  * @author Jean-François Simon <jeanfrancois.simon@sensiolabs.com>
Chris@0:  *
Chris@0:  * @internal
Chris@0:  */
Chris@0: class TokenizerEscaping
Chris@0: {
Chris@0:     private $patterns;
Chris@0: 
Chris@0:     public function __construct(TokenizerPatterns $patterns)
Chris@0:     {
Chris@0:         $this->patterns = $patterns;
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * @param string $value
Chris@0:      *
Chris@0:      * @return string
Chris@0:      */
Chris@0:     public function escapeUnicode($value)
Chris@0:     {
Chris@0:         $value = $this->replaceUnicodeSequences($value);
Chris@0: 
Chris@0:         return preg_replace($this->patterns->getSimpleEscapePattern(), '$1', $value);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * @param string $value
Chris@0:      *
Chris@0:      * @return string
Chris@0:      */
Chris@0:     public function escapeUnicodeAndNewLine($value)
Chris@0:     {
Chris@0:         $value = preg_replace($this->patterns->getNewLineEscapePattern(), '', $value);
Chris@0: 
Chris@0:         return $this->escapeUnicode($value);
Chris@0:     }
Chris@0: 
Chris@0:     /**
Chris@0:      * @param string $value
Chris@0:      *
Chris@0:      * @return string
Chris@0:      */
Chris@0:     private function replaceUnicodeSequences($value)
Chris@0:     {
Chris@0:         return preg_replace_callback($this->patterns->getUnicodeEscapePattern(), function ($match) {
Chris@0:             $c = hexdec($match[1]);
Chris@0: 
Chris@0:             if (0x80 > $c %= 0x200000) {
Chris@17:                 return \chr($c);
Chris@0:             }
Chris@0:             if (0x800 > $c) {
Chris@17:                 return \chr(0xC0 | $c >> 6).\chr(0x80 | $c & 0x3F);
Chris@0:             }
Chris@0:             if (0x10000 > $c) {
Chris@17:                 return \chr(0xE0 | $c >> 12).\chr(0x80 | $c >> 6 & 0x3F).\chr(0x80 | $c & 0x3F);
Chris@0:             }
Chris@0:         }, $value);
Chris@0:     }
Chris@0: }