Consistently use byte/str everywhere

Created by: Lekensteyn

Python2 and python3 differ in the handling of bytes and unicode. In python2, unicode is not equal to str. In Python3, bytes are not equal to str:

$ for v in 2 3; do python$v -c "print(type(u''), type(b''), type(''))";done
(<type 'unicode'>, <type 'str'>, <type 'str'>)
<class 'str'> <class 'bytes'> <class 'str'>
$ for v in 2 3; do echo|python$v -c $'import os\nprint(type(os.read(0, 1)))';done
<type 'str'>
<class 'bytes'>

Todo:

• read in hidapi/udev.py returns type bytes on Python3 instead of str.
• Change every check for type(foo) == unicode or type(foo) == str (or isinstance). unicode has gone in Python3, this became str. isinstance('', bytes) is True for python2, but False in python3.
• Currently bytes are printed verbatim (in hidconsole at least). On Python 2 one could see '\xffX' where python 3 shows b'\xffX'. Changing it to '\xff\x58' would give a more consistent output, but I think you did this to spot text easier in the protocol. Does it matter that the output is different?

Possible helpful resources:

• http://stackoverflow.com/questions/7585435/best-way-to-convert-string-to-bytes-in-python-3
• http://stackoverflow.com/questions/606191/convert-byte-array-to-python-string