Cardano Updates

Upgrade pumba flake so it builds on modern mac hardware.

`goldenExampleEraTxCborSpec` allows us to generate a `Spec` for
generating a `golden/tx.cbor` golden file out of some `Tx` provided as a
parameter.

Then, for each era's test-suite, we call this function to generate the
golden file using the example transaction we have defined for each era.

We also extended `EraTest` so that we don't need to manually call
`getDataFileName` in order to get the fullpath for each era's cabal
project.

We also create new `Binary.Golden` modules for Mary and Babbage (previously
they reused Allegra's and Alonzo's respectively, which used the wrong era's
example transactions).

Additionally, fix `exampleDijkstraBasedTopTx` to not add PlutusV4 scripts
to `scriptTxWitsL`: PlutusV4 is not included in Dijkstra's
`transaction_witness_set` CDDL, so those scripts were
silently dropped during serialization, causing a roundtrip failure.

In `conwayGovTransition`, we remove the condition on
`hardforkConwayBootstrapPhase` for checking `ZeroTreasuryWithdrawals`.

We then replayed the preview, preprod and mainnet public chains with
this change which guarentees there were no empty treasury withdrawals
pre-Conway era.

From https://github.com/IntersectMBO/cardano-node at 97036a66bcf8c89f687ae57a048eecc0389977ef

`goldenExampleEraTxCborSpec` allows us to generate a `Spec` for
generating a `golden/tx.cbor` golden file out of some `Tx` provided as a
parameter.

Then, for each era's test-suite, we call this function to generate the
golden file using the example transaction we have defined for each era.

We also extended `EraTest` so that we don't need to manually call
`getDataFileName` in order to get the fullpath for each era's cabal
project.

We also create new `Binary.Golden` modules for Mary and Babbage (previously
they reused Allegra's and Alonzo's respectively, which used the wrong era's
example transactions).

Additionally, fix `exampleDijkstraBasedTopTx` to not add PlutusV4 scripts
to `scriptTxWitsL`: PlutusV4 is not included in Dijkstra's
`transaction_witness_set` CDDL, so those scripts were
silently dropped during serialization, causing a roundtrip failure.

Pulls in the depth checklist (cluster_fork_depth_N_observed) and
adversary_observed_as_peer Sometimes assertions onto the
cardano_node_adversary testnet.

Two new probes on top of the existing fork-tree state machine.

Layer 3 — depth checklist
  - Tracks maxForkDepth in tracer-sidecar State.
  - Emits one Sometimes assertion per depth threshold from 2 up to
    k - 1: cluster_fork_depth_2_observed, _3_observed, ... _k-1.
  - Reading the report, the highest passed threshold tells you how
    deep the cluster forked under fault injection without drilling
    into per-event details.

Layer 2 — peer-observability probe
  - Adds adversary_observed_as_peer Sometimes assertion.
  - Fires on the first Net.InboundGovernor.* log line whose JSON
    contains the substring "adversary.example", proving the
    adversary's chain-sync sessions reached a producer's peer-
    selection layer at least once during the run (not just the
    kernel TCP backlog).
  - Uses a substring check against the encoded JSON, so we don't
    pin the message schema — the producer's tracer can rename
    fields without breaking the probe.

State now carries maxForkDepth and adversaryObservedAsPeer; both
land alongside the existing forkObserved / forkExceededK flags
inside updateForkTree's refresh function and a new generic
mkSpec rule.

feat: add EDR on switching to our own mini-protocol implementations

Signed-off-by: Jenita <[email protected]>
Co-authored-by: Jenita <[email protected]>

Signed-off-by: Chris Gianelloni <[email protected]>

Signed-off-by: Chris Gianelloni <[email protected]>