|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
declare(strict_types=1); |
|
4
|
|
|
|
|
5
|
|
|
/** |
|
6
|
|
|
* This file is part of Wszetko Sitemap. |
|
7
|
|
|
* |
|
8
|
|
|
* (c) Paweł Kłopotek-Główczewski <[email protected]> |
|
9
|
|
|
* |
|
10
|
|
|
* This source file is subject to the MIT license that is bundled |
|
11
|
|
|
* with this source code in the file LICENSE. |
|
12
|
|
|
*/ |
|
13
|
|
|
|
|
14
|
|
|
namespace Wszetko\Sitemap\Helpers; |
|
15
|
|
|
|
|
16
|
|
|
/** |
|
17
|
|
|
* Class Url. |
|
18
|
|
|
* |
|
19
|
|
|
* @package Wszetko\Sitemap\Helpers |
|
20
|
|
|
*/ |
|
21
|
|
|
class Url |
|
22
|
|
|
{ |
|
23
|
|
|
/** |
|
24
|
|
|
* @param string $url |
|
25
|
|
|
* |
|
26
|
|
|
* @return bool|string |
|
27
|
|
|
* |
|
28
|
|
|
* @see https://bugs.php.net/bug.php?id=52923 |
|
29
|
|
|
* @see https://www.php.net/manual/en/function.parse-url.php#114817 |
|
30
|
|
|
*/ |
|
31
|
138 |
|
public static function normalizeUrl(string $url) |
|
32
|
|
|
{ |
|
33
|
138 |
|
$encodedUrl = preg_replace_callback( |
|
34
|
138 |
|
'%[^:/@?&=#]+%usD', |
|
35
|
|
|
function ($matches) { |
|
36
|
138 |
|
return rawurlencode($matches[0]); |
|
37
|
138 |
|
}, |
|
38
|
69 |
|
$url |
|
39
|
|
|
); |
|
40
|
|
|
|
|
41
|
138 |
|
if (null === $encodedUrl) { |
|
42
|
|
|
// @codeCoverageIgnoreStart |
|
43
|
|
|
return false; |
|
44
|
|
|
// @codeCoverageIgnoreEnd |
|
45
|
|
|
} |
|
46
|
|
|
|
|
47
|
138 |
|
$url = parse_url($encodedUrl); |
|
48
|
|
|
|
|
49
|
138 |
|
if (empty($url) || !isset($url['host'])) { |
|
50
|
6 |
|
return false; |
|
51
|
|
|
} |
|
52
|
|
|
|
|
53
|
136 |
|
$url = array_map('urldecode', $url); |
|
54
|
136 |
|
$url['host'] = idn_to_ascii($url['host'], IDNA_DEFAULT, INTL_IDNA_VARIANT_UTS46); |
|
55
|
|
|
|
|
56
|
136 |
|
if (empty($url['scheme']) || !is_string($url['host']) || !self::checkDomain($url['host'])) { |
|
57
|
2 |
|
return false; |
|
58
|
|
|
} |
|
59
|
|
|
|
|
60
|
134 |
|
if (!empty($url['path'])) { |
|
61
|
16 |
|
$url['path'] = explode('/', $url['path']); |
|
62
|
16 |
|
$parts = []; |
|
63
|
|
|
|
|
64
|
16 |
|
foreach ($url['path'] as $element) { |
|
65
|
16 |
|
switch ($element) { |
|
66
|
16 |
|
case '.': |
|
67
|
2 |
|
break; |
|
68
|
16 |
|
case '..': |
|
69
|
2 |
|
array_pop($parts); |
|
70
|
|
|
|
|
71
|
2 |
|
break; |
|
72
|
|
|
default: |
|
73
|
16 |
|
$parts[] = rawurlencode($element); |
|
74
|
|
|
|
|
75
|
16 |
|
break; |
|
76
|
|
|
} |
|
77
|
|
|
} |
|
78
|
|
|
|
|
79
|
16 |
|
$url['path'] = implode('/', $parts); |
|
80
|
|
|
} |
|
81
|
|
|
|
|
82
|
134 |
|
if (isset($url['query']) && is_string($url['query']) && '' !== $url['query']) { |
|
83
|
2 |
|
parse_str($url['query'], $query); |
|
84
|
|
|
array_walk($query, function (&$val, &$var) { |
|
85
|
2 |
|
$var = rawurlencode($var); |
|
86
|
2 |
|
$val = rawurlencode($val); |
|
87
|
2 |
|
}); |
|
88
|
2 |
|
$url['query'] = http_build_query($query); |
|
89
|
|
|
} |
|
90
|
|
|
|
|
91
|
134 |
|
$user = ''; |
|
92
|
|
|
|
|
93
|
134 |
|
if (isset($url['user']) && is_string($url['user']) && '' !== $url['user']) { |
|
94
|
4 |
|
$user = $url['user']; |
|
95
|
|
|
|
|
96
|
4 |
|
if (isset($url['pass']) && is_string($url['pass']) && '' !== $url['pass']) { |
|
97
|
4 |
|
$user .= ':' . $url['pass']; |
|
98
|
|
|
} |
|
99
|
|
|
|
|
100
|
4 |
|
$user .= '@'; |
|
101
|
|
|
} |
|
102
|
|
|
|
|
103
|
|
|
return |
|
104
|
134 |
|
$url['scheme'] . '://' |
|
105
|
134 |
|
. $user |
|
106
|
134 |
|
. mb_strtolower($url['host']) |
|
107
|
134 |
|
. ((isset($url['port']) && is_string($url['port']) && '' !== $url['port']) ? ':' . $url['port'] : '') |
|
108
|
134 |
|
. ((isset($url['path'])) ? $url['path'] : '') |
|
109
|
134 |
|
. (isset($url['query']) && (is_string($url['query']) && '' !== $url['query']) ? '?' . $url['query'] : '') |
|
110
|
134 |
|
. (isset($url['fragment']) && (is_string($url['fragment']) && '' !== $url['fragment']) ? '#' . $url['fragment'] : ''); |
|
111
|
|
|
} |
|
112
|
|
|
|
|
113
|
|
|
/** |
|
114
|
|
|
* @param string $domain |
|
115
|
|
|
* |
|
116
|
|
|
* @return bool |
|
117
|
|
|
*/ |
|
118
|
140 |
|
public static function checkDomain(string $domain): bool |
|
119
|
|
|
{ |
|
120
|
140 |
|
if ('.' == mb_substr($domain, -1) || '.' == mb_substr($domain, 0, 1)) { |
|
121
|
4 |
|
return false; |
|
122
|
|
|
} |
|
123
|
|
|
|
|
124
|
140 |
|
$domain = idn_to_ascii($domain, IDNA_DEFAULT, INTL_IDNA_VARIANT_UTS46); |
|
125
|
|
|
|
|
126
|
140 |
|
return (bool) filter_var($domain, FILTER_VALIDATE_DOMAIN, FILTER_FLAG_HOSTNAME); |
|
127
|
|
|
} |
|
128
|
|
|
} |
|
129
|
|
|
|