Description of ENT_DISALLOWED in htmlentities is misleading

From manual page: https://php.net/function.htmlentities

Description of ENT_DISALLOWED is misleading.
It states: "Replace invalid code points for the given document type with a Unicode Replacement Character U+FFFD (UTF-8) or &#FFFD; (otherwise) instead of leaving them as is. This may be useful, for instance, to ensure the well-formedness of XML documents with embedded external content."
It should say somenthing like: "Returns an empty string when invalid code points for the given document are found."

Tested using the following code:

$string = "This is a test: \x80\x81";
echo htmlentities($string, ENT_SUBSTITUTE, 'UTF-8');
// Output: This is a test: ��
echo "\n";
$string = "This is a test: \x80\x81";
echo htmlentities($string, ENT_DISALLOWED, 'UTF-8');
// Output: <empty line> 
echo "\n";
$string = "This is a test: xx";
echo htmlentities($string, ENT_DISALLOWED, 'UTF-8');
// Output: This is a test: xx

Start with the ENT_DISALLOWED description on the htmlentities manual page and compare it with the documented behavior shown in the issue’s PHP examples. Update the misleading wording to accurately describe the observed behavior, then review the surrounding ENT_* flag documentation for consistency.