\Dom\Document::getElementById() is inconsistent after nodes are removed

[cscott]

Description

The following code:

<?php
 $d = \Dom\HTMLDocument::createFromString('<p id="a">b</p><p id="a">c</p>');
$p = $d->getElementById('a');
$p->remove();
echo($d->getElementById('a')->innerHTML);

Resulted in this output:

Attempt to read property "innerHTML" on null

But I expected this output instead:

c

Compare to the following JS code in a browser:

> document.body.innerHTML = '<p id=a>b</p><p id=a>c</p>'
'<p id=a>b</p><p id=a>c</p>'
> p = document.getElementById('a')
<p id=​"a">​b​</p>​
> p.remove()
undefined
> document.getElementById('a').innerHTML
'c'

PHP Version

PHP 8.4.13 (cli) (built: Oct  1 2025 20:34:15) (NTS)
Copyright (c) The PHP Group
Built by Debian
Zend Engine v4.4.13, Copyright (c) Zend Technologies
    with Zend OPcache v8.4.13, Copyright (c), by Zend Technologies

Operating System

No response

[cscott]

The not-terribly-well-documented secret is that the DOM implementation needs to maintain the ID-to-nodes map as a one-to-many map (not one-to-one) so that cases like this work properly.