Cardano Updates

Signed-off-by: Juhi Bhandari <[email protected]>

Nixpkgs expects to drop support for intel macs in 26.11.

Consolidates golden tests from the separate `cardano-ledger-alonzo-test`
package into the main `cardano-ledger-alonzo` impl package, following the
same pattern already completed in other eras (allegra, mary, conway, babbage).

Picks up the pin to upstream feat/tx-generator-n2c-reconnect
(commit 3e2623d) so publish-images resolves the tag and
the smoke-test runs against the supervised daemon.

Pin 'cardano-node-clients' to
3e2623d438ca5062e381171538a0f4da927d0679 — the head of
'feat/tx-generator-n2c-reconnect' upstream
(lambdasistemi/cardano-node-clients#105) which wires the
tx-generator daemon's N2C connection through
runReconnectLoop, fixing the 210x bearer-close container
exits Antithesis surfaced on ed6666d.

Per the 'Pins main only' rule this temporary
feature-branch pin is OK on a downstream feature branch;
will be repointed to the upstream main merge SHA
before merging this PR to main.

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

The asteria-stub container becomes the long-running chain-following
indexer. Its main process is upstream `utxo-indexer` from
lambdasistemi/cardano-node-clients PR #98
(035-indexer-n2c-reconnect at 5707836b), pinned via a flake input
and baked into the image by `nix build .#docker-image`. The indexer
follows relay1's chain over N2C, persists state to RocksDB at
/idx-db, exposes ready/utxos_at/await on /tmp/idx.sock, and now
auto-reconnects in-process when the N2C peer closes the socket
(supervisor + exponential-backoff retry from upstream issue #97).

Replaces the per-component Dockerfile + apt + cardano-cli setup
with a pure-nix build (flake.nix + nix/docker-image.nix). cardano-cli
is no longer needed in this image because composer scripts now
query the indexer's socket directly.

Composer scripts (kept in their stub shape):
  - parallel_driver_heartbeat.sh — pings indexer's `ready`,
    Sometimes-true on processedSlot==tipSlot
  - eventually_alive.sh — post-fault: same probe with retries
  - finally_alive.sh — post-workload: same probe with retries

Compose: mount relay1-state:/state:ro for the indexer's N2C peer
and add asteria-stub-db named volume for RocksDB persistence. No
restart: always — the supervisor handles transient peer churn.

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

chore(deps-dev): update Docusaurus packages to 3.10.0 and add new ses…

Nixpkgs expects to drop support for intel macs in 26.11.

* mithril-client-cli from `0.13.9` to `0.13.10`
* mithril-common from `0.6.70` to `0.6.71`
* mithril-end-to-end from `0.4.131` to `0.4.132`
* mithril-test-lab/mithril-devnet/VERSION from `0.4.31` to `0.4.32`