Code review comment for lp://staging/~rodsmith/hwcert-tools/get-ytd

The only last comment is on handling the .encode bit. That's still totally broken. There are well-known methods for handling bytes and unicode in python2. The general rule is to *always* encode at boundaries (on the outside world is bytes, on the inside text is unicode, bytes are str).

You can try this:
 - assume all the api query functions return unicode strings
 - add __from__ future import unicode_literals
 - you will get everything as unicode internally
 - each file IO needs to have the mode specified as either text with encoding= keyword argument or as 'b' and never treat that as text.

This will give you a clear upgrade path to python3 in any situation. It also means that you never have to encode anything.