Make .show() work on unicode strings. Fixes #38

[dourvaris]

Not sure if testing outputted bytes like this is the most elegant way.

Should also have a side effect of testing unicode identifiers in general across the lib.

[Natim]

Usually in Python we are trying to do the reverse thing.

Decoding when getting inside Python and encoding when getting outside of Python.

Here you are using BytesIO but I think you should be using StringIO instead.

[dourvaris]

Since .show() was 'printing' to stdout which would want bytes, thats the end result format to check.

[Natim]

Good point.