#622 helped with this, but it's still overly laconic about what's going
on. I think this is better.

Informal pinging has been broken since #623, luckily this was
unreleased, but existed edge.

The PR replaces vere's current (buggy, insufficient) lockfile mechanism
with an advisory `fcntl()` write lock (`F_SETLK`, `F_WRLCK`) on the
first byte of vere's PID file (`$pier/.vere.lock`). It does not (yet?)
include the functionality to kill the lock-holder and take over the
pier, as I'm not sure that's desirable.

There have been many problems with vere's current lockfile
implementation. The biggest advantage of this approach is that the lock
is automatically released if the process is killed or the host system
crashes. The biggest question is: does it work reliably across docker
shared volumes, virtual machines, &c.