build(deps): bump dig-rewards-coin to 0.11 (dig_ecosystem#3450)

[MichaelTaylor3d]

Refs DIG-Network/dig_ecosystem#3450

Bumps dig-rewards-coin 0.10 -> 0.11 and refreshes the caret-compatible dig-* floors (§2.4b): dig-mirror-collateral 0.3.1, dig-mirror-coin 0.9.1, dig-chainsource-interface 0.3.3 (normal + dev), dig-cert 0.1.3, dig-logging 0.2.2, dig-urn-resolver 0.5.4, dig-constants 0.13.1. Cargo.lock refreshed. No version bump (lane PR into develop, §2.4).

No source change. dig-rewards-coin 0.11's new CommitmentEpochStarted variant is returned only by clawback::withdraw_committed_incentives, which dig-node never calls (guarded by the existing absence test in rewards/chain_port.rs); the only match on the error enum (reader_error_to_port_error) has a wildcard arm. The not-recoverable state stays covered by tests/rewards_chain_port_a3.rs (an_epoch_started_commitment_is_reported_as_none_not_zero), green on this head. Mapping criterion ruled MOOT on the ticket.

🤖 Generated with Claude Code

[MichaelTaylor3d]

Verdict: PASS at a1a94a1 (independent review; no local build, CI green on this head).

1. Lock: dig-rewards-coin 0.11.0 only (single entry, checksum changed); dig-mirror-collateral 0.3.1, dig-mirror-coin 0.9.1, dig-chainsource-interface 0.3.3, dig-cert 0.1.3, dig-logging 0.2.2, dig-urn-resolver 0.5.4, dig-constants 0.13.1 all resolve. chia-protocol entries unchanged (0.26.0, 0.36.1): no new major line. Only other lock churn is windows-sys/socket2 dependency-edge reshuffles, no new packages.
2. Comments: chainsource "caret-matched" still true (single 0.3.3). Non-blocking drift: Cargo.toml dig-constants comment says "lockstep with dig-node-core's pin ... 0.11.2", but core pins "0.13.0" (crates/dig-node-core/Cargo.toml:390) vs service "0.13.1". The comment was already stale before this PR (0.11.2); both resolve to the one 0.13.1 lock entry, so no behavioural drift. Optional follow-up: refresh the comment or raise core to 0.13.1.
3. MOOT ruling holds: withdraw_committed_incentives appears only in the absence guard test (chain_port.rs:570-580) and comments; reader_error_to_port_error has a wildcard other => ChainPortError::Other(...) arm.
4. PR body accurate against the diff.

No blocking findings; no inline threads opened (0 open).

[MichaelTaylor3d]

loop-security verdict: PASS at head a1a94a1 (dig-node#634, base develop f1e9f80).

1. Registry-only: the lock diff changes 0 name/source lines and adds 0 packages. The only package-record change is dig-rewards-coin 0.10.0 -> 0.11.0, from the crates.io registry. Git-source count in Cargo.lock is 11 before and after (unchanged, pre-existing). The other dig-* floor refreshes are Cargo.toml-only because the lock already resolved those versions.
2. Checksum present on the changed entry (e08c8bcf...01db24). The 0.11.0 source is in the local registry cache and its dependency list is identical to 0.10.0.
3. 0.11 change is read-path state.rs (recoverable predicate now shares distributor_epoch_started) plus a new CommitmentEpochStarted error variant used only by clawback::withdraw_committed_incentives. dig-node never references that function (guard test chain_port.rs:570-580). recoverable_base_units: Option<u64> is passed through unchanged (chain_port.rs:232, None = not recoverable, never 0). The wildcard arm at chain_port.rs:310-317 absorbs the new variant. No spend/custody path touched.
4. No new crate, so no second rustls CryptoProvider. rustls 0.23.43 is unchanged, and no aws-lc line is in the diff (aws-lc-rs/sys are pre-existing).

Defence-in-depth, not gating: the lock also re-points ~20 dependency edges (windows-sys 0.61.2 -> 0.59/0.60/0.48; socket2 0.6.5 -> 0.5.10 in quinn/quinn-udp). All target versions already existed in the lock and are within each parent's declared range. They are Windows-only or socket-level, with no new package. Looks like resolver churn from the lock refresh. Worth knowing, no action.

Not covered: no local build, no cargo-audit run, no source diff of the transitive dig-* floor bumps beyond the lock.