Sometimes some gai_strerror will throw unwanted bullshit characters into the mix, incase this happens again, I have tested a little snippet of code and these nonsense characters appear to be bypass-able.

e = b'\xdd'
print(e.decode('utf-8',"surrogateescape"))

• 'Fixes error parse bytes #32'

@IvanKirpichnikov Here's that pull request you've been looking for.