scratch – Blame information for rev 87
?pathlinks?
Rev | Author | Line No. | Line |
---|---|---|---|
87 | office | 1 | <?php |
2 | |||
3 | /* |
||
4 | * This file is part of the Symfony package. |
||
5 | * |
||
6 | * (c) Fabien Potencier <fabien@symfony.com> |
||
7 | * |
||
8 | * For the full copyright and license information, please view the LICENSE |
||
9 | * file that was distributed with this source code. |
||
10 | */ |
||
11 | |||
12 | namespace Symfony\Component\CssSelector\Parser\Tokenizer; |
||
13 | |||
14 | /** |
||
15 | * CSS selector tokenizer escaping applier. |
||
16 | * |
||
17 | * This component is a port of the Python cssselect library, |
||
18 | * which is copyright Ian Bicking, @see https://github.com/SimonSapin/cssselect. |
||
19 | * |
||
20 | * @author Jean-François Simon <jeanfrancois.simon@sensiolabs.com> |
||
21 | * |
||
22 | * @internal |
||
23 | */ |
||
24 | class TokenizerEscaping |
||
25 | { |
||
26 | /** |
||
27 | * @var TokenizerPatterns |
||
28 | */ |
||
29 | private $patterns; |
||
30 | |||
31 | /** |
||
32 | * @param TokenizerPatterns $patterns |
||
33 | */ |
||
34 | public function __construct(TokenizerPatterns $patterns) |
||
35 | { |
||
36 | $this->patterns = $patterns; |
||
37 | } |
||
38 | |||
39 | /** |
||
40 | * @param string $value |
||
41 | * |
||
42 | * @return string |
||
43 | */ |
||
44 | public function escapeUnicode($value) |
||
45 | { |
||
46 | $value = $this->replaceUnicodeSequences($value); |
||
47 | |||
48 | return preg_replace($this->patterns->getSimpleEscapePattern(), '$1', $value); |
||
49 | } |
||
50 | |||
51 | /** |
||
52 | * @param string $value |
||
53 | * |
||
54 | * @return string |
||
55 | */ |
||
56 | public function escapeUnicodeAndNewLine($value) |
||
57 | { |
||
58 | $value = preg_replace($this->patterns->getNewLineEscapePattern(), '', $value); |
||
59 | |||
60 | return $this->escapeUnicode($value); |
||
61 | } |
||
62 | |||
63 | /** |
||
64 | * @param string $value |
||
65 | * |
||
66 | * @return string |
||
67 | */ |
||
68 | private function replaceUnicodeSequences($value) |
||
69 | { |
||
70 | return preg_replace_callback($this->patterns->getUnicodeEscapePattern(), function ($match) { |
||
71 | $c = hexdec($match[1]); |
||
72 | |||
73 | if (0x80 > $c %= 0x200000) { |
||
74 | return chr($c); |
||
75 | } |
||
76 | if (0x800 > $c) { |
||
77 | return chr(0xC0 | $c >> 6).chr(0x80 | $c & 0x3F); |
||
78 | } |
||
79 | if (0x10000 > $c) { |
||
80 | return chr(0xE0 | $c >> 12).chr(0x80 | $c >> 6 & 0x3F).chr(0x80 | $c & 0x3F); |
||
81 | } |
||
82 | }, $value); |
||
83 | } |
||
84 | } |