| 1 |
<?php |
| 2 |
if ( ! defined( 'ABSPATH' ) ) { |
| 3 |
die( 'You are not allowed to call this page directly.' ); |
| 4 |
} |
| 5 |
|
| 6 |
/** |
| 7 |
* Sanitize HTML attribute values to prevent stored XSS. |
| 8 |
* |
| 9 |
* @since 6.34 |
| 10 |
*/ |
| 11 |
class FrmHtmlSanitizer { |
| 12 |
|
| 13 |
/** |
| 14 |
* Sanitize href and src attribute values to valid URLs only. |
| 15 |
* |
| 16 |
* Decodes HTML entities in the attribute value before validating, |
| 17 |
* so entity-encoded payloads are rejected. |
| 18 |
* |
| 19 |
* @since 6.34 |
| 20 |
* |
| 21 |
* @param string $value HTML string to process. |
| 22 |
* |
| 23 |
* @return string |
| 24 |
*/ |
| 25 |
public static function sanitize_url_attributes( $value ) { |
| 26 |
if ( '' === $value || ( ! str_contains( $value, 'href' ) && ! str_contains( $value, 'src' ) ) ) { |
| 27 |
return $value; |
| 28 |
} |
| 29 |
|
| 30 |
$sanitized = preg_replace_callback( |
| 31 |
'/\b(href|src)\s*=\s*"([^"]*)"/', |
| 32 |
array( self::class, 'sanitize_url_attribute_value' ), |
| 33 |
$value |
| 34 |
); |
| 35 |
|
| 36 |
return $sanitized ?? ''; |
| 37 |
} |
| 38 |
|
| 39 |
/** |
| 40 |
* Callback to sanitize a single URL attribute match. |
| 41 |
* |
| 42 |
* @since 6.34 |
| 43 |
* |
| 44 |
* @param array $matches Regex matches with attribute name and value. |
| 45 |
* |
| 46 |
* @return string Rebuilt attribute with a safe URL value. |
| 47 |
*/ |
| 48 |
private static function sanitize_url_attribute_value( $matches ) { |
| 49 |
$url = trim( html_entity_decode( $matches[2], ENT_QUOTES, 'UTF-8' ) ); |
| 50 |
|
| 51 |
if ( str_starts_with( $url, '#' ) ) { |
| 52 |
return $matches[1] . '="' . esc_attr( $url ) . '"'; |
| 53 |
} |
| 54 |
|
| 55 |
if ( 'src' === $matches[1] && self::is_png_data_uri( $url ) ) { |
| 56 |
return $matches[1] . '="' . esc_attr( $url ) . '"'; |
| 57 |
} |
| 58 |
|
| 59 |
if ( ! preg_match( '/^(https?:\/\/|mailto:|tel:)/i', $url ) ) { |
| 60 |
return $matches[1] . '=""'; |
| 61 |
} |
| 62 |
|
| 63 |
$safe = esc_url( $url, array( 'http', 'https', 'mailto', 'tel' ) ); |
| 64 |
|
| 65 |
if ( '' === $safe ) { |
| 66 |
return $matches[1] . '=""'; |
| 67 |
} |
| 68 |
|
| 69 |
$host = wp_parse_url( $safe, PHP_URL_HOST ); |
| 70 |
|
| 71 |
if ( $host && preg_match( '/%[0-9a-f]{2}/i', $host ) ) { |
| 72 |
return $matches[1] . '=""'; |
| 73 |
} |
| 74 |
|
| 75 |
return $matches[1] . '="' . esc_attr( $safe ) . '"'; |
| 76 |
} |
| 77 |
|
| 78 |
/** |
| 79 |
* Check for a PNG data URI that contains only base64 characters, like a drawn signature image src. |
| 80 |
* |
| 81 |
* The pattern allows only base64 characters after the prefix, so the value cannot carry a media |
| 82 |
* type of its own or any markup into the attribute. |
| 83 |
* |
| 84 |
* @since 6.34 |
| 85 |
* |
| 86 |
* @param string $url Decoded URL value to check. |
| 87 |
* |
| 88 |
* @return bool |
| 89 |
*/ |
| 90 |
private static function is_png_data_uri( $url ) { |
| 91 |
return 1 === preg_match( '#^data:image/png;base64,[A-Za-z0-9+/]+={0,2}$#D', $url ); |
| 92 |
} |
| 93 |
} |
| 94 |
|
