| 621 |
lars |
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\Yaml;
|
|
|
13 |
|
|
|
14 |
use Symfony\Component\Yaml\Exception\ParseException;
|
|
|
15 |
|
|
|
16 |
/**
|
|
|
17 |
* Unescaper encapsulates unescaping rules for single and double-quoted
|
|
|
18 |
* YAML strings.
|
|
|
19 |
*
|
|
|
20 |
* @author Matthew Lewinski <matthew@lewinski.org>
|
|
|
21 |
*
|
|
|
22 |
* @internal
|
|
|
23 |
*/
|
|
|
24 |
class Unescaper
|
|
|
25 |
{
|
|
|
26 |
/**
|
|
|
27 |
* Regex fragment that matches an escaped character in a double quoted string.
|
|
|
28 |
*/
|
|
|
29 |
public const REGEX_ESCAPED_CHARACTER = '\\\\(x[0-9a-fA-F]{2}|u[0-9a-fA-F]{4}|U[0-9a-fA-F]{8}|.)';
|
|
|
30 |
|
|
|
31 |
/**
|
|
|
32 |
* Unescapes a single quoted string.
|
|
|
33 |
*
|
|
|
34 |
* @param string $value A single quoted string
|
|
|
35 |
*/
|
|
|
36 |
public function unescapeSingleQuotedString(string $value): string
|
|
|
37 |
{
|
|
|
38 |
return str_replace('\'\'', '\'', $value);
|
|
|
39 |
}
|
|
|
40 |
|
|
|
41 |
/**
|
|
|
42 |
* Unescapes a double quoted string.
|
|
|
43 |
*
|
|
|
44 |
* @param string $value A double quoted string
|
|
|
45 |
*/
|
|
|
46 |
public function unescapeDoubleQuotedString(string $value): string
|
|
|
47 |
{
|
|
|
48 |
$callback = function ($match) {
|
|
|
49 |
return $this->unescapeCharacter($match[0]);
|
|
|
50 |
};
|
|
|
51 |
|
|
|
52 |
// evaluate the string
|
|
|
53 |
return preg_replace_callback('/'.self::REGEX_ESCAPED_CHARACTER.'/u', $callback, $value);
|
|
|
54 |
}
|
|
|
55 |
|
|
|
56 |
/**
|
|
|
57 |
* Unescapes a character that was found in a double-quoted string.
|
|
|
58 |
*
|
|
|
59 |
* @param string $value An escaped character
|
|
|
60 |
*/
|
|
|
61 |
private function unescapeCharacter(string $value): string
|
|
|
62 |
{
|
|
|
63 |
return match ($value[1]) {
|
|
|
64 |
'0' => "\x0",
|
|
|
65 |
'a' => "\x7",
|
|
|
66 |
'b' => "\x8",
|
|
|
67 |
't' => "\t",
|
|
|
68 |
"\t" => "\t",
|
|
|
69 |
'n' => "\n",
|
|
|
70 |
'v' => "\xB",
|
|
|
71 |
'f' => "\xC",
|
|
|
72 |
'r' => "\r",
|
|
|
73 |
'e' => "\x1B",
|
|
|
74 |
' ' => ' ',
|
|
|
75 |
'"' => '"',
|
|
|
76 |
'/' => '/',
|
|
|
77 |
'\\' => '\\',
|
|
|
78 |
// U+0085 NEXT LINE
|
|
|
79 |
'N' => "\xC2\x85",
|
|
|
80 |
// U+00A0 NO-BREAK SPACE
|
|
|
81 |
'_' => "\xC2\xA0",
|
|
|
82 |
// U+2028 LINE SEPARATOR
|
|
|
83 |
'L' => "\xE2\x80\xA8",
|
|
|
84 |
// U+2029 PARAGRAPH SEPARATOR
|
|
|
85 |
'P' => "\xE2\x80\xA9",
|
|
|
86 |
'x' => self::utf8chr(hexdec(substr($value, 2, 2))),
|
|
|
87 |
'u' => self::utf8chr(hexdec(substr($value, 2, 4))),
|
|
|
88 |
'U' => self::utf8chr(hexdec(substr($value, 2, 8))),
|
|
|
89 |
default => throw new ParseException(sprintf('Found unknown escape character "%s".', $value)),
|
|
|
90 |
};
|
|
|
91 |
}
|
|
|
92 |
|
|
|
93 |
/**
|
|
|
94 |
* Get the UTF-8 character for the given code point.
|
|
|
95 |
*/
|
|
|
96 |
private static function utf8chr(int $c): string
|
|
|
97 |
{
|
|
|
98 |
if (0x80 > $c %= 0x200000) {
|
|
|
99 |
return \chr($c);
|
|
|
100 |
}
|
|
|
101 |
if (0x800 > $c) {
|
|
|
102 |
return \chr(0xC0 | $c >> 6).\chr(0x80 | $c & 0x3F);
|
|
|
103 |
}
|
|
|
104 |
if (0x10000 > $c) {
|
|
|
105 |
return \chr(0xE0 | $c >> 12).\chr(0x80 | $c >> 6 & 0x3F).\chr(0x80 | $c & 0x3F);
|
|
|
106 |
}
|
|
|
107 |
|
|
|
108 |
return \chr(0xF0 | $c >> 18).\chr(0x80 | $c >> 12 & 0x3F).\chr(0x80 | $c >> 6 & 0x3F).\chr(0x80 | $c & 0x3F);
|
|
|
109 |
}
|
|
|
110 |
}
|