Eike Hein has kindly donated a working Mpris2 adaptor with instructions
on how to integrate into JuK, so I've done so.

This is just at the "compiles and links" stage and the track ID will
give a lot of weirdness due to the strictness of the D-Bus Object Path
name spec.

The reason we don't just use an identifier for a track is that the
"primary key" JuK uses is the canonical file name. JuK has no internal
RDBMS or anything similar where primary keys would be assigned. Because
of this we encode track IDs using z-Base-32 so they will not violate the
D-Bus spec.

Thanks to Eike for the large code contribution and integration support!