In case of a literal decoding failure it tries to decode
the result in utf-8. This is handy in python2 environments where
python and the host might be using different charset configurations.
In python3 this issue seams to be solved.
Fixes#829 and bsc#1110871
