* op-supervisor: Introduce thread-unsafe log database

* op-supervisor: Add simple r/w locking

* op-supervisor: Add comment

* op-supervisor: Start switching to multi-entry database format

* op-supervisor: Improve test to cover the case where a new block starts at a search checkpoint boundary (other than at the start of the file)

* op-supervisor: Use a flag to indicate when log index should increment rather than a 1 byte increment amount.

* op-supervisor: Comment out unused stuff to make lint happy.

* op-supervisor: Load correct block number and log idx on init

* op-supervisor: Refactor state to only hold context that can always be kept up to date.

* op-supervisor: Support rewinding

* op-supervisor: Remove TODO that probably won't be done there

* op-supervisor: Require first log in block to have logIdx 0

* op-supervisor: Remove completed TODO.

* op-supervisor: Improve testing for logs not existing

* op-supervisor: Fix typo

* op-supervisor: Tidy up TODOs and pending tests.

* op-supervisor: Add invariant assertions for db data

* op-supervisor: Lock db in ClosestBlockInfo

* op-supervisor: Label alerts

* op-supervisor: Use a TruncatedHash for logs everywhere and make it a fixed size array.

* op-supervisor: Separate serialization of initating events

* op-supervisor: Separate serialization of other event types and enforce type code.

* op-supervisor: Introduce entry type

* op-supervisor: Split out an entry database

* op-supervisor: Introduce structs for entry types

* op-supervisor: Use a struct for CanonicalHash too