In order to properly get pids etc. we need to pass bytes to MD5 and SHA1
instead of unicode strings. Also ord() is no longer needed since
data is bytes value gets int and we need chr() to get characters from
the mapping bytearrays.
THIS IS ON THE MASTER BRANCH. The Master branch will be Python 3.0 from now on. While Python 2.7 support will not be deliberately broken, all efforts should now focus on Python 3.0 compatibility.
I can see a lot of work has been done. There's more to do. I've bumped the version number of everything I came across to the next major number for Python 3.0 compatibility indication.
Thanks everyone. I hope to update here at least once a week until we have a stable 7.0 release for calibre 5.0