So I have a program in Python for use on a Linux machine ... If my data has a zero byte, this is taken as the "end-of-string" terminator ('\0') by os.write() and nothing is written past that.

I think I can assume I have a utf-8 encoded byte array, and try and work with that. I've been using the ctypes module in Python to interact with a .DLL file that contains the encrypt/decrypt ...

They are accepted with Python 2. An application might want to use bytes filenames to avoid problems with undecodable characters. $ python2.7 -c 'import PIL.Image; PIL.Image.open(b"file.png")' $ ...