Cardano Updates

Direct tests for ObjectDiffusion mini protocol

Move tx counter tracing to drainRejectionThread where it will be run at
most once per second.

If state change wait at most 25ms for additional changes in order to
coalece TX events together.

and introduce `perasSupportToBool` and `perasSupportFromBool` functions.

This will make it easier to transition to `contra-tracer-0.2` which
requires to replace `Tracer` with `Tracer . emit`.  Also removed usage
of `showTracing` which is not part of `contra-tracer-0.2`

Mini-protocol job handler should not throw synchronous exceptions.  If
it does, the job result won't be written to completion var and mux
monitoring loop won't be notified.

The `BlockedOnCompletionVar` exception was added only as a safety
measure that the `putMVar` should never block.  However, it's only safe
to do it in the `jobAction`.  We instead use `tryPutMVar` in the
`jobHandler` as a better alternative.

* fixed `prop_timeouts_enforced` in
  `ouroboros-network:framework-sim-tests`
* verify that `prop_timeouts_enforced` in
  `cardano-diffusion:cardano-diffusion-sim-tests`
  evaluates events

This patch makes it easier to change the result type of
`diffusionSimulation`.  For that one just needs to change the
`DiffSimResult` type alias, all the properties should compile right
away.

Let's acknowledge that we don't expose all possible smart constructors.
The one we expose are used by `cardano-db-sync`, but not by
`ouroboros-consensus-diffusion`, `trace-forward`, `tx-generator` or
`ekg-forward`.

The tests-lib should only exposed generic traces which can be combined
as needed.   `debugTracerG` is removed, we only expose:

* debugTracer - a tracer to `stderr`
* dynamicTracer - a tracer using `traceM` (IOSim only)
* sayTracer - a tracer via `say` (IO & IOSim)

`debugTracerG` is removed.  It was only used in testnet simulation,
where now instead we use `dynamicTracer` and `sayTracer`.

`label`s are easier to read, since `QuickCheck` will list each group in
a separate list.  This is important for the sim-net tests where we have
multiple nodes.  With a label we will get the output grouped by node,
since every node returns a labelled `Property`.

`PrettyShow` allows to to pretty print addresses, thread labels etc.
This makes the simulation output more readable.

Send lists of TXs to the mempool when possible. This mimics the
behaviour of the V1 tx submission.