Implement sending commands via Unix sockets with optional systemd socket activation.

[davnotdev]

This PR addresses #33 and #34. Though, there are a few places that need to be smoothed out:

1. Logs are done using eprintln. Perhaps I should use the log crate?
2. The files in systemd/ assume that burrow is installed in /usr/local/bin/burrow, but this is not ideal. As @Muirrum mentioned, those files should be up to distro packagers.
3. There are a few nasty unwraps I'd like to replace with proper error handling, but I feel like that's outside the scope of this PR.

[davnotdev]

I merged main into my daemon branch, and now everything is red and failing to compile. Woopsies.
Never mind, I think I've undone it.

[conradev]

Sorry for taking so long to review this – it looks great!

1. We should absolutely use logging macros instead of eprintln
2. Assumptions in those files are okay for now, the service files are really suggestions for the time being
3. Agreed

There are a few random items:

• Requests could consider having an ID so you can send multiple over the same connection at once
• The daemon expects and depends on systemd right now, but ideally the daemon does some runtime discovery to figure out what to do: if the socket is free, grab it, if it isn't, check if systemd has it, else, error

[conradev]

also the commits do need to be squashed into one