Fix build fail

[dvush]

📝 Summary

1. add --all-targets to makefile checks (otherwise we miss linting tests and benches)
2. fix compilation errors and warnings in tests

💡 Motivation and Context

---

✅ I have completed the following steps:

• Run make lint
• Run make test
• Added tests (if applicable)

[claude]

Claude finished @dvush's task in 1m 15s —— View job

---

Review of PR #916

• Gather context and read changed files
• Review Makefile changes
• Review benchmark file changes
• Review deleted order_dumper.rs
• Review block_orders/mod.rs changes
• Post review summary

---

LGTM — this is a clean, straightforward fix. All four changes are correct and well-motivated:

Makefile (--all-targets): Adding --all-targets to both lint and fmt targets ensures clippy catches issues in tests, benches, and examples — not just lib/bin targets. The fmt target also correctly gains --workspace which it was previously missing (the lint target already had it).

Benchmark fix (txpool_fetcher.rs): The subscribe_to_txpool_with_blobs signature was updated upstream to require Arc<MempoolTxsDetector> as the third parameter. The benchmark now passes it correctly. No issues.

order_dumper.rs deletion: OrderDumper and OrderStoreAction are not referenced anywhere else in the codebase — confirmed via grep. The comment "Soon to be replaced my mockall" suggests it was already slated for removal. Clean delete.

block_orders/mod.rs: Simply removes the #[cfg(test)] mod order_dumper; declaration to match the file deletion. No other changes to this file.

No bugs, no security concerns, no performance issues.

[claude]

Claude finished @dvush's task in 1m 32s —— View job

---

PR Review

• Read changed files
• Review Makefile changes
• Review benchmark changes
• Review block_orders changes and order_dumper deletion
• Review bundle_tests/setup.rs changes
• Post review summary

---

No blocking issues found. A few minor observations:

Makefile — Adding --all-targets to lint is clearly correct. For the fmt target, note that cargo clippy --all-targets --fix will now auto-apply fixes to tests/benches/examples in addition to lib/bin. This is almost certainly desired, but worth being aware of the wider blast radius when running make fmt.

Also the test target (Makefile:160) still doesn't compile/run bench code — only cargo test on the default targets. --all-targets on clippy catches bench compile errors at lint time, which is probably sufficient, but there's no test-time coverage ensuring benches build. Not a regression, just noting.

bundle_tests/setup.rs:215 — #[allow(clippy::result_large_err)] is a workaround, not a fix. ExecutionError::LowerInsertedValue { before: SimValue, inplace: SimValue } (crates/rbuilder/src/building/mod.rs:576-577) is the large variant pushing the enum over the clippy threshold. Since this is test code called on the happy path, the allow is fine — but if similar warnings surface elsewhere on production paths, boxing SimValue inside the variant would be the proper fix.

txpool_fetcher.rs bench — Correctly matches the updated subscribe_to_txpool_with_blobs signature (crates/rbuilder/src/live_builder/order_input/txpool_fetcher.rs:23-28).

order_dumper.rs deletion — Verified no remaining references to OrderDumper or OrderStoreAction anywhere in the tree. Clean removal.
| Branch: fix-test-build

[claude]

Claude finished @dvush's task in 46s —— View job

---

PR Review

• Read newly changed files (since last review)
• Review parallel_builder/conflict_task_generator.rs changes
• Review parallel_builder/groups.rs changes
• Review live_builder/mod.rs changes
• Review utils/receipts.rs changes
• Post review summary

---

No blocking issues. Latest commit 62b0adb is four mechanical Rust 1.95 clippy fixes, all semantically equivalent to the originals:

conflict_task_generator.rs:55 — sort_by(|a,b| b.len().cmp(&a.len())) → sort_by_key(|b| Reverse(b.orders.len())). Equivalent descending sort; Reverse newtype flips Ord. ✓

groups.rs:196 — .chain(x.into_iter()) → .chain(x). Iterator::chain takes IntoIterator, so the explicit conversion was redundant. ✓

live_builder/mod.rs:374-375 — .iter().filter_map(|(_, r)| r.adjustment_fee_payer) → .values().filter_map(|r| r.adjustment_fee_payer). Same result; .values() is the idiomatic form. ✓

receipts.rs:94 — .zip(x.into_iter()) → .zip(x). Iterator::zip takes IntoIterator. ✓

Earlier-reviewed items from prior commits still stand: bundle_tests/setup.rs:215's #[allow(clippy::result_large_err)] is a workaround (root cause is the LowerInsertedValue variant embedding two SimValues — box the variant if this appears on production paths), and make test still doesn't exercise bench compilation — --all-targets on clippy is the only bench-build gate.

No bugs, no security concerns, no performance regressions.
| Branch: fix-test-build