fix some unicode vs ascii codec shit
honestly, the idiotic default behaviour of python with regard to strings
containing non-ascii characters just makes me want to nuke my python install
and never ever touch it anymore.
additionally the functions decode and encode do just the opposite of what one
would expect.
strings should just be bytes until i explicitly request an operation that
*needs* to operate on unicode codepoints, but not for replacing a tab with
spaces.
this fixes display of sources with utf-8 unicode chars in them, like the
(C) copyright symbol \xc2\xa9.