Still more testing to do, but I want to get this panic fixed ASAP.

All new tests pass with -race

Append when there's nothing blocking between appending and sending an
update to the server.

Make it match TaskRunner update hook behavior

* UpdateState: set state, append event, persist, update servers
* EmitEvent: append event, persist, update servers
* AppendEvent: append event, persist

AppendEvent may not even have to persist, but for the sake of
correctness I'm going with that for now.

Also remove initial alloc from broadcaster as it just caused useless
extra processing.

minimize passed dependencies to ease testing

And test thoroughly.

Leaving fingerprinters in case that interface changes with plugins.

* Stopping an alloc is implemented via Updates but update hooks are
  *not* run.
* Destroying an alloc is a best effort cleanup.
* AllocRunner destroy hooks implemented.
* Disk migration and blocking on a previous allocation exiting moved to
  its own package to avoid cycles. Now only depends on alloc broadcaster
  instead of also using a waitch.
* AllocBroadcaster now only drops stale allocations and always keeps the
  latest version.
* Made AllocDir safe for concurrent use

Lots of internal contexts that are currently unused. Unsure if they
should be used or removed.

Allow alloc and task runners to persist their own state when something
changes instead of periodically syncing all state.

Saves a tiny bit of cpu and some IO. Sadly doesn't prevent all IO on
duplicate writes as the transactions are still created and committed.

$ go test -bench=. -benchmem
goos: linux
goarch: amd64
pkg: github.com/hashicorp/nomad/helper/boltdd
BenchmarkWriteDeduplication_On-4             500           4059591 ns/op           23736 B/op         56 allocs/op
BenchmarkWriteDeduplication_Off-4            300           4115319 ns/op           25942 B/op         55 allocs/op

finished but doesn't handle deleting deeply nested buckets

Still WIP as it does not handle deletions.

Updates are applied asynchronously but sequentially

Should probably be moved to an Inspect method in the Driver Plugin world

Deregistration works but difficult to test due to terminal updates not
being fully implemented in the new client/ar/tr.

also updated comments

Restoring calls NewAR -> Restore -> Run

NewAR now calls NewTR
AR.Restore calls TR.Restore
AR.Run calls TR.Run