Cardano Updates

Each transaction's `txDirectDeposits` are now applied to the threaded
`CertState` immediately after that transaction's `CERTS` step, replacing
the batch-wide application introduced earlier on this branch.

  * `SUBLEDGER-V` applies `DirectDepositsOf stx` to `certState₁` after
    `CERTS` and emits the result.
  * `LEDGER-V` still applies the top-level transaction's deposits to
    `certState₂` after the top-level `CERTS`, but the premise now bounds
    `dom (DirectDepositsOf tx)` rather than the batch aggregate
    `dom (allDirectDeposits tx)`.
  * `certStateWithDDeps` is generalised to
    `∀ {ℓ} → Tx ℓ → CertState → CertState` so both rules can invoke it
    with the local `DirectDepositsOf`.

The original draft of this branch argued that batch-wide application
made CIP-159's phantom-asset prohibition manifest in the rule
structure.  Per-transaction application is equally safe: the prohibition
is enforced separately by `NoPhantomWithdrawals` in the `Utxo` module,
which bounds batch-wide withdrawal totals against the `accountBalances`
field of `UTxOEnv`/`SubUTxOEnv` — fixed at the pre-batch snapshot
`RewardsOf certState₀` regardless of when (or how) deposits are applied
to the threaded `CertState`.  Per-transaction application also mirrors
the Haskell ledger executor's sequential per-transaction processing,
which simplifies the conformance proof and avoids a structural gap
between the spec and the implementation.

A direct-deposit registration premise is added to each rule to prevent
`applyDirectDeposits` from silently re-introducing a credential into
`rewards` after deregistration earlier in the same transaction:

  * `SUBLEDGER-V` requires
    `dom (DirectDepositsOf stx) ⊆ dom (RewardsOf certState₁)`.
  * `LEDGER-V` requires
    `dom (DirectDepositsOf tx)  ⊆ dom (RewardsOf certState₂)`.

The check is local to each transaction's post-`CERTS` state, so a
sub-transaction whose deposit targets a credential deregistered by an
earlier sub-transaction in the same batch will fail the premise.

`Computational-SUBLEDGER` and `Computational-LEDGER` are updated to
dispatch on these decidable premises in `computeProof` and
`completeness`.  `depositsChange` is unaffected: it reads only the
protocol-deposit fields of `DState`/`PState`/`GState`, which
`applyDirectDeposits` does not touch.

Refs: #1122

After all sub-rule transitions (`SUBLEDGERS`, `CERTS`, `GOVS`, `UTXOW`),
apply batch-wide direct deposits to the final CertState via
`applyDirectDeposits` and `allDirectDeposits`.

`Ledger.lagda.md`:
+  Update `LEDGER-V` output: compute `certStateFinal` by applying
   `allDirectDeposits` to `certState₂`, use `certStateFinal` in the
   output `LedgerState` and in `rmOrphanDRepVotes`;
+  `LEDGER-I` unchanged (invalid batches don't apply deposits);
+  Document direct deposit application ordering and phantom asset
   prevention rationale.

`Ledger/Properties/Computational.lagda.md`:
+  Update `computeProof` valid branch to compute `certStateFinal` and use
   it in the output `LedgerState`.
+  Clean up redundant code.

`applyDirectDeposits` uses `∪⁺` (union-with-addition) on the `rewards`
map, so a credential not present in `RewardsOf certState₂` is added as a
new entry rather than rejected.  The pre-existing SUBUTXO/UTXO premise
`dom (DirectDepositsOf tx) ⊆ dom (AccountBalancesOf Γ)` only checks
against the pre-batch balances `RewardsOf certState₀`, so it does not
catch the case where a credential is registered pre-batch, deregistered
during `CERTS`, and direct-deposited to in the same batch.  Without the
new premise, the resulting `certStateFinal` is internally inconsistent:
the credential has a positive `rewards` entry but no corresponding
`voteDelegs`, `stakeDelegs`, or `deposits` entry — effectively re-
registering the credential without a key deposit.

Add the phase-1 premise

    dom (allDirectDeposits tx) ⊆ dom (RewardsOf certState₂)

to `LEDGER-V` to rule this out.  Filtering `allDirectDeposits tx` by
`dom (RewardsOf certState₂)` would be an alternative, but it would
silently destroy user funds when a deposit and a deregistration race in
the same batch.  The premise is symmetric to (and complements) the
existing pre-batch SUBUTXO/UTXO check.

Update `Ledger/Properties/Computational.lagda.md` for the new premise,
and pass `UTxOStateOf s₁` explicitly to `computeUtxow` in the LEDGER
valid branch (the previous `_` resolved by chained unification through
the LEDGER-V constructor, which was fragile and obscured intent).

TODO: fix `Ledger/Properties/Computational.lagda.md`; it's choking on
the new premise.

Nixpkgs expects to drop support for intel macs in 26.11.

* Add Dijkstra Foreign modules

* NativeScripts; add DecEq-Network as an explicit parameter to avoid agda bug

for bug report see https://github.com/agda/agda/issues/8532

for bug report see https://github.com/agda/agda/issues/8532

Add batch-wide withdrawal bound check to prevent phantom asset attacks
when nested transactions combine deposits and withdrawals.

`Transaction.lagda.md`:
+  Define allWithdrawals batch aggregation helper (mirrors
   allDirectDeposits)

`Utxo.lagda.md`:
+  Define NoPhantomWithdrawals predicate using allWithdrawals
+  Add NoPhantomWithdrawals premise to UTXO rule
+  Document phantom asset attack and spend-side safety analogy

`Utxo/Properties/Computational.lagda.md`:
+  Update Computational-UTXO for new premise tuple arity (21+h → 22+h)

Nixpkgs expects to drop support for intel macs in 26.11.