adopt dig-node-control-interface 0.22 — serve control.spends.list

[MichaelTaylor3d]

Task

dig-node-control-interface 0.22.0 publishes control.spends.list, and dig-node does not serve it.
Adopt 0.22 and serve the method, so the node and the contract agree again.

Parent epic: https://github.com/DIG-Network/dig_ecosystem/issues/3166
Discovered by: #382 (PR #383)

Context

§2.4b requires a PR that touches a crate to bring its dig-* deps to latest. PR #383 did that, and
crates/dig-node-service/tests/control_contract_conformance.rs:54 failed:

the contract publishes methods this node does not serve: ["control.spends.list"]

That guard is doing exactly its job — this is contract drift, not a test problem. The capability itself
already exists behind the CLI (dign spends list, crates/dig-node-service/src/spend_audit_cli.rs), so
the work is exposing it over the control plane and matching the shape the contract declares, not building
a new feature.

PR #383 therefore held the dep at 0.21 rather than shipping a node that fails its own conformance
guard, and split the adoption here. Say so plainly if the shape 0.22 declares turns out to differ from
what the CLI already produces.

Scope

Bump dig-node-control-interface to 0.22 in both crates that declare it (dig-wallet,
dig-node-service) — they are one workspace, and bumping one alone resolves two semver-incompatible
copies of the control types — then serve control.spends.list.

Evidence

control_contract_conformance green with the dep at 0.22, plus a test of the method's own response shape.

Report

Whether the contract's declared shape matched the CLI's existing output, or had to be reconciled.

[MichaelTaylor3d]

Duplicate of #385, filed a few hours earlier for the same work.

Consolidating there because #385 already carries the measurement that makes this bigger than a dispatch arm: the PR#31 gate found SpendQuery carries no after_id and SpendLog::query cannot report complete, so the paginated contract cannot be satisfied by the existing query type without extending it and its callers.

Your finding that control_contract_conformance catches the drift is a good one and is worth carrying over — that test is doing real work, and holding the dep at 0.21 until the method is served is the right call rather than a §2.4b miss.