refactor(ergo-node): split oversized files (wallet_bridge, boot, snapshot, scala_compat, messaging)

[arkadianet]

Summary

Splits the largest/most tangled files in ergo-node into cohesive submodule directories, following the same methodology used for the ergo-compiler splits (#200, #202): incremental extraction with test verification after each step, red-team review of every commit, one bundled PR.

• node/wallet_bridge.rs (5355 → 1487 lines): extracted the business-logic layer called by the already-split commands/*.rs handlers into node/wallet_bridge/support/ (8 files: hints_codec, key_derivation, dto, multisig_helpers, tx_build, sign_submit, generate_sign, sweep). run_wallet_writer and its dispatch match stay in wallet_bridge.rs.
  • Includes a real bug fix: derive_next_key_impl duplicated persist_tracked_pubkey's redb-write logic inline instead of calling it, and separately advanced WALLET_DERIVATION_HEAD outside that transaction. Unified into one atomic write_txn — a crash between the two writes previously could wedge derivation permanently (tracked pubkey persisted, head stuck, every future call hitting the dedup check and returning DerivationPathExists).
• node/boot.rs (1834 lines): split run_inner_with_backend's ~1280-line backend-agnostic boot tail into node/boot/{mod,peers,sync_setup,mining,api_wiring}.rs. Each phase function takes the prior phase's outputs and returns a small struct of the values it produces, with run_inner_with_backend itself as the orchestrating body. api_wiring and mining each split into two functions since the real dependency order interleaves them (scaffold → mining subsystem → bind; subsystem → NodeState → engine spawn). Manually boot-tested against real testnet peers (synced header chain to height 400, clean shutdown).
• node/snapshot_emit.rs + snapshot.rs: split into node/snapshot_emit/{mod,recent_blocks,bootstrap_panel,mempool_projection,events_projection}.rs (mod.rs keeps publish_snapshot as orchestrator) and snapshot/{mod,build,publisher}.rs (the NodeSnapshot/SnapshotParts DTOs stay whole in mod.rs — splitting a single wide DTO struct across files hurts more than it helps).
• api_bridge/scala_compat.rs: light touch only — the trait impl (NodeChainQuery for ScalaCompatBridge) is one large block of small independent read-only methods, not worth forcing apart. Only the ~300-line fee-per-byte pool-ranking cluster moved to scala_compat/pool_fee_stats.rs.
• node/messaging.rs: extracted the three sizable per-batch match arms (CODE_INV, CODE_MODIFIER ~190 lines, CODE_PEERS) into named helpers (handle_inv, handle_modifier_batch, handle_peers_response) following the same &mut NodeState -> Vec<Action> convention as handle_message/run_wallet_writer. Split into node/messaging/{dispatch,manifest,utxo_chunk,popow}.rs.

Every commit was independently verified (cargo test -p ergo-node including all integration tests, cargo clippy -p ergo-node --all-targets --all-features -- -D warnings, cargo fmt --all -- --check) and red-team reviewed by an independent Opus pass diffing against the pre-refactor original — no behavior-changing findings survived across any of the 5 splits. The whole-workspace gate (cargo clippy --workspace --all-targets --all-features -- -D warnings and cargo test --workspace) is also green.

Purely structural — no functional changes intended anywhere except the one explicitly-called-out derive_next_key_impl atomicity fix.

Test plan

• cargo test -p ergo-node (500 lib tests + full integration suite, all green) after each of the 5 splits
• cargo clippy -p ergo-node --all-targets --all-features -- -D warnings clean after each split
• cargo fmt --all -- --check clean after each split
• cargo clippy --workspace --all-targets --all-features -- -D warnings clean (whole-workspace gate)
• cargo test --workspace clean (whole-workspace gate, every crate)
• Manual boot test against real testnet peers after the boot.rs split (header sync to height 400, clean SIGTERM shutdown)
• Independent Opus red-team review per commit, diffing every moved function against the pre-refactor original

🤖 Generated with Claude Code

https://claude.ai/code/session_01Kw7uCwqmiGna8fpg6TvxGi

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 58c2c1e2-9718-464f-8ece-6be5602f072b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch refactor/ergo-node-split

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[arkadianet]

Splitting into two smaller PRs to stay under coderabbitai's 25-files-per-PR review limit: #205 (wallet_bridge + boot) and #206 (snapshot + scala_compat + messaging).