Fix multi network persistance

[Sansh2356]

• Braidpool nodes previously ran every network (mainnet, signet, regtest, cpunet) on the same libp2p protocol IDs, so peers from different networks could negotiate substreams against each other. That silently produced bead/hash divergence leading to breakdown of entire network mesh.
• This PR isolates networks end-to-end across the p2p stack and on disk by protocol negotiation during ConnectionEstablishment and adding peers only when negotiation passes as well as persistence in network specific sub-directories having --datadir as parent directory.
• Addresses Bifurcating persistance on basis of Network type #462 and --datadir has no effect inside init_db() #533 .

[Copilot]

Pull request overview

This PR addresses cross-network contamination by scoping Braidpool’s libp2p protocol identifiers and on-disk state to the selected --network, preventing peers from different networks from negotiating substreams and sharing persistence under the same paths.

Changes:

• Scope libp2p protocol IDs (kademlia + bead-sync) and floodsub topic by PoolNetwork, and only admit peers into the mesh/peer manager after Identify confirms expected protocols.
• Make persistence network-specific by creating/using <datadir>/<network>/ as the root for the DB and keystore, and thread the network-scoped path through DB initialization.
• Add a multiaddr routability filter for DHT insertion and expand tests/docs to cover the new isolation model.

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
node/src/utils/mod.rs	Adds is_routable_multiaddr plus unit tests for filtering unroutable listen addrs before DHT insertion.
node/src/stratum.rs	Updates a test to use DBHandler::new_in_memory to avoid persistent DB usage.
node/src/peer_manager/mod.rs	Introduces PendingPeer and pending tracking to defer peer registration until protocol negotiation succeeds.
node/src/main.rs	Implements network-scoped datadir setup, network-scoped topics/protocol checks on Identify, and pending→promoted peer flow.
node/src/db/init_db.rs	Changes DB init to accept a network-scoped directory and warns if a legacy unscoped DB exists.
node/src/db/db_handlers.rs	Threads network-scoped datadir into persistent DB handler initialization.
node/src/db/audit_db_handlers.rs	Threads network-scoped datadir into audit DB handler initialization.
node/src/behaviour/tests.rs	Adds tests asserting protocol/topic names differ across networks and updates swarm construction.
node/src/behaviour/mod.rs	Replaces global protocol constants with network-scoped protocol/topic generators and updates behaviour construction.
node/src/audit.rs	Updates audit DAG initialization to accept the network-scoped datadir.
docs/braidpool_setup.md	Documents the per-network datadir layout and network-scoped on-wire protocol/topic naming.
Cargo.lock	Updates locked dependencies as a result of the build/dependency graph changes in the workspace.

Suppressed comments (1)

node/src/main.rs:133

• If the per-network data path (<datadir>/<network>) already exists but is not a directory, the code logs and continues. That can lead to confusing failures when creating the DB/keystore inside it; returning an error here makes the failure immediate and actionable.

    match fs::metadata(&network_datadir) {
        Ok(m) => {
            if !m.is_dir() {
                error!(path = %network_datadir.display(), "Network data path exists but is not a directory");
            }

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.