XPathEvaluator namespaces regression in 3.6.3+

[pdknsk]

This code used to work.

lxml.etree : (3, 6, 2, 0)
libxml used : (2, 9, 1)
libxml compiled : (2, 9, 1)
libxslt used : (1, 1, 28)
libxslt compiled : (1, 1, 28)

>>> xml = etree.XML('<Response xmlns="http://example.com"><Code>200</Code></Response>')
>>> xpath = etree.XPathEvaluator(xml, namespaces = {'1': xml.nsmap[None]})
>>> xpath('//1:Code/text()')
['200']

It throws an exception now.

lxml.etree : (3, 6, 3, 0)
libxml used : (2, 9, 4)
libxml compiled : (2, 9, 4)
libxslt used : (1, 1, 29)
libxslt compiled : (1, 1, 29)

>>> xpath('//1:Code/text()')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "src/lxml/xpath.pxi", line 307, in lxml.etree.XPathElementEvaluator.__call__ (src/lxml/lxml.etree.c:167085)
  File "src/lxml/xpath.pxi", line 227, in lxml.etree._XPathEvaluatorBase._handle_result (src/lxml/lxml.etree.c:166044)
lxml.etree.XPathEvalError: Invalid expression

[scoder (scoder)]

XML names must not start with a digit. That includes namespace prefixes. It seems that libxml2 got stricter here in recent versions, although I have no idea why that wasn't an error before. Seems like an unusual bug to allow that in the first place.