Repository navigation
UTF-8 corruption in \Dom\HTMLDocumentΒ #17481
Description
Description
The following code:
<?php
$Repeated = str_repeat( 'β', 4096 );
//$Repeated = str_repeat( 'π', 4096 );
$Data = '<!DOCTYPE HTML><html>' . $Repeated . '</html>';
$Document = \Dom\HTMLDocument::createFromString( $Data, 0, 'UTF-8' );
echo $Document->saveHTML();
// var_dump($Document->body->textContent);The resulting string contains random invalid UTF-8 sequences like with the οΏ½ character. With the repeated emoji, emojis become corrupted. If you repeat the string for longer, there are more corrupted bytes in random places.
I initially spotted this bug when parsing a real HTML document and used textContent (innerHTML produces the same issue) on an element I found with xpath.
PHP Version
PHP 8.4.2
Operating System
Windows 11 and Debian 12
Activity
xPaw commentedon Jan 15, 2025
FYI this is reproducible with 2-byte utf-8 sequences too, but not ascii.
ndossche commentedon Jan 15, 2025
I'll try to have a look at the encoding code tomorrow.
Fix phpGH-17481: UTF-8 corruption in \Dom\HTMLDocument
Merge branch 'PHP-8.4'
Fix GH-17481: UTF-8 corruption in \Dom\HTMLDocument
Fix potential read of uninitialized padding data in DOM
Fix potential read of uninitialized padding data in DOM (#17628)
Fix potential read of uninitialized padding data in DOM (#17628)
Fix potential read of uninitialized padding data in DOM (#17628)