Code review comment for lp://staging/~brian.curtin/ubuntu-sso-client/py3-unicode

This series of py3k branches is looking awesome, thanks for working on it!

One thing we should consider: in every file where we add the "from __future__ import unicode_literals" we should be extra careful about the string literals that we are not turning from u"..." to "...".

I can see two different cases happening here:
 * In some places the old code assumes that conversions when comparing unicode with str are automatic, so to make this work on 3 we should leave old string literal as they are.
 * In some other places some old string literals should be changed from "..." to "...".encode(...), because they are truly using "byte literals". For instance in many of the tests in ubuntu_sso/keyring/tests/test_common.py that fake some functions that will still return bytes on py3.