[tech-debt] Hardcoded sync batch limit of 500 events lacks configuration and blocks protocol migration

[intendednull]

Problem

The sync response in crates/client/src/listeners.rs hard-codes a batch size of 500 events with an explicit TODO noting this is a temporary workaround:

// crates/client/src/listeners.rs:293-304
// TODO: migrate to heads-based sync — can't filter by state_hash in DAG model
events.into_iter().take(500).collect()

There are three inconsistent hardcoded limits:

• 500 events in the sync response batch (listeners.rs:296)
• 10_000 as a max batch in listeners.rs:256
• 10_000 in crates/storage/src/store.rs:287

None of these are exposed through ClientConfig or a named constant.

Impact

• Cannot tune for different network conditions without changing code
• Inconsistent limits across the sync path
• The TODO signals this is blocking full DAG-based sync protocol migration

Fix

1. Extract to named constants:

pub const SYNC_BATCH_SIZE: usize = 500;
pub const SYNC_MAX_BATCH: usize = 10_000;

2. Expose through ClientConfig for runtime tuning
3. Track the heads-based sync migration as a follow-up (related to issue Worker sync protocol uses bulk SyncBatch instead of per-author Advertise/Request/Response #65 on worker sync protocol)

Location

• crates/client/src/listeners.rs:256, 296-304
• crates/storage/src/store.rs:287

Found during

Codebase audit at commit 18119e343ec53a9393312006d9d3b205129f696d