test failure with libxml2 2.10.4

I've tried to change the tests like this:

    def test_html_prefix_nsmap(self):
        etree = self.etree
        html = etree.HTML('<hha:page-description>aa</hha:page-description>')
        el = html.find('.//page-description')
        if etree.LIBXML_VERSION < (2, 10, 4):
            if etree.LIBXML_VERSION < (2, 9, 11):
                self.assertEqual({'hha': None}, el.nsmap)
            else:
                self.assertEqual({}, el.nsmap)
        else:
            self.assertIsNone(el)
            el = html.find('.//hha:page-description') # <-- breaks
            self.assertEqual({}, el.nsmap)

But the marked line does not work as I expected:

======================================================================
ERROR: test_html_prefix_nsmap (lxml.tests.test_etree.ETreeOnlyTestCase.test_html_prefix_nsmap)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "src/lxml/_elementpath.py", line 85, in lxml._elementpath.xpath_tokenizer
    raise KeyError
KeyError

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/usr/lib64/python3.11/unittest/case.py", line 57, in testPartExecutor
    yield
  File "/usr/lib64/python3.11/unittest/case.py", line 623, in run
    self._callTestMethod(testMethod)
  File "/usr/lib64/python3.11/unittest/case.py", line 579, in _callTestMethod
    if method() is not None:
       ^^^^^^^^
  File "/builddir/build/BUILD/lxml-4.9.2/src/lxml/tests/test_etree.py", line 3080, in test_html_prefix_nsmap
    el = html.find('.//hha:page-description')
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "src/lxml/etree.pyx", line 1550, in lxml.etree._Element.find
    return _elementpath.find(self, path, namespaces)
  File "src/lxml/_elementpath.py", line 323, in lxml._elementpath.find
    it = iterfind(elem, path, namespaces)
  File "src/lxml/_elementpath.py", line 312, in lxml._elementpath.iterfind
    selector = _build_path_iterator(path, namespaces)
  File "src/lxml/_elementpath.py", line 295, in lxml._elementpath._build_path_iterator
    selector.append(ops[token[0]](_next, token))
  File "src/lxml/_elementpath.py", line 120, in lxml._elementpath.prepare_descendant
    token = next()
  File "src/lxml/_elementpath.py", line 88, in xpath_tokenizer
    raise SyntaxError("prefix %r not found in prefix map" % prefix)
SyntaxError: prefix 'hha' not found in prefix map

----------------------------------------------------------------------
Ran 1979 tests in 5.332s

I'm targeting this to lxml 5.0 since it's a behavioural change in libxml2 2.10.4 and later. I'll just remove the test and declare the parsing of HTML tag "prefixes" as version dependent and not future proof.

I changed the behaviour in lxml 5.0 to consider "prefixes" in HTML documents as part of the tag name when searching for tags.
Let's see if this works for our users.

See https://github.com/lxml/lxml/commit/72f5a287a4016ecb405f2e8a4a03ae22a5b0b496