Internally we do hex-escaping of non-ascii characters. Why?
I would think it was more efficient to not do that, and only do it when we print strings to logs and other debug output.
I concur.
Ciao, Michael.