fix String encoding

[tisdall]

fixes #233

[tseaver]

I prefer that the result = text_type(appstruct) to be outside the if, with no else. E.g.:

    result = text_type(appstruct)
    if self.encoding:
        result = result.encode(self.encoding)
    return result

[tisdall]

@tseaver - updated as requested...

[tseaver]

Hmm, this one should start as text and then encode to bytes (Python2 lets us get away with being sloppy, but that doesn't mean we shouldn't be explicit).

[tisdall]

Sorry, I develop everything in Python 3 and then deal with Python 2 only when I need backwards compatibility. I know this is already doing what your describing in Py3, but I don't understand how it's not doing this in Py2. In Py2, should this be u'123'.encode('utf-8') to be explicit?

[tisdall]

Are you saying this should be utf8 = text_type('123').encode('utf-8')?

[tseaver]

That would be the better spelling, yes.

[tseaver]

Or just b'123', in fact (ASCII and UTF-8 are identical encodings for 7-bit values).

[tisdall]

@tseaver - I went with using text_type('123').encode('utf-8') so it's totally explicit that the result is a utf-8 encoded string.

[tisdall]

@tseaver - should I add a note in the CHANGES.rst? Is there anything else I need to change to complete this?

[tseaver]

A note in CHANGES.rst would be great, thanks!

[tisdall]

done

[tseaver]

Thanks!