QTextEdit HTML parser broken in presence of non-BMP unicode characters

The HTML parser QTextEdit uses is broken in the presence of some non-BMP characters in the HTML even when those characters are encoded using HTML entities. A minimal one-line HTML snippet to reproduce the issue is below alongwith simple wrapper PySide code to reproduce.

from PySide6.QtWidgets import QTextEdit, QApplication
app = QApplication([])
w = QTextEdit()
w.setHtml('<p id="x">&#x1f41a;&#x1f41a;</p>')
w.show()
app.exec() 

This displays as two question marks followed by the spiral shell character U+1f41a (🐚) instead of two spiral shell characters. Note that id="x" is needed, it doesn't reproduce without that, which tells me its dependent on some parser internal state. I have used Python as a convenient way to reproduce the issue but it exists in C++ as well, since the string being passed to setHTML() is pure ASCII. I have also attatched a screenshot of the buggy rendering on my system.