mapper: fix phantom updateSentPeers on disconnected nodes

When send() is called on a node with zero active connections
(disconnected but kept for rapid reconnection), it returns nil
(success). handleNodeChange then calls updateSentPeers, recording
peers as delivered when no client received the data.

This corrupts lastSentPeers: future computePeerDiff calculations
produce wrong results because they compare against phantom state.
After reconnection, the node's initial map resets lastSentPeers,
but any changes processed during the disconnect window leave
stale entries that cause asymmetric peer visibility.

Return errNoActiveConnections from send() when there are no
connections. handleNodeChange treats this as a no-op (the change
was generated but not deliverable) and skips updateSentPeers,
keeping lastSentPeers consistent with what clients actually
received.

hscontrol/mapper/batcher.go

@@ -178,10 +178,20 @@ func handleNodeChange(nc nodeConnection, mapper *mapper, r change.Change) error
 	// Send the map response
 	err = nc.send(data)
 	if err != nil {
+		// If the node has no active connections, the data was not
+		// delivered. Do not update lastSentPeers — recording phantom
+		// peer state would corrupt future computePeerDiff calculations,
+		// causing the node to miss peer additions or removals after
+		// reconnection.
+		if errors.Is(err, errNoActiveConnections) {
+			return nil
+		}
+
 		return fmt.Errorf("sending map response to node %d: %w", nodeID, err)
 	}
 
-	// Update peer tracking after successful send
+	// Update peer tracking only after confirmed delivery to at
+	// least one active connection.
 	nc.updateSentPeers(data)
 
 	return nil