Closed
Description
we should have a test that actually works on a unicode char > 0xFFFF and then checks for its length / do some regexp matching on it.
This test will correctly fail on narrow python builds, which is yet another way of notifying developers of something being wrong with their system.