query_has_tx fails with Err(Plexer(AgentDequeue))

[jonathangenlambda]

I have added the txmonitor::client to network facades, (see https://github.com/txpipe/pallas/pull/442/files) and am trying to use the mempool querying functionality now.
query_size_and_capacity works repeatedly (when doing it in a loop, for example each block) perfectly fine, but query_has_tx fails with Err(Plexer(AgentDequeue)).

Here is my testcode:

async fn check_pending_tx_mempool(node_client: &mut NodeClient) {
    let monitor_client = node_client.monitor();

    let aquire_result = monitor_client.acquire().await;
    info!("aquire_result {:?}", aquire_result);
    
    let sac_result = monitor_client.query_size_and_capacity().await;
    info!("sac_result {:?}", sac_result);
    
    // this returns Err(Plexer(AgentDequeue)) 
    let has_tx_result = monitor_client.query_has_tx("fc43da88c808bbd84c394cf43d81679a4c319c21f64f9f6d9bbab57dc8bc5ff3".to_string()).await;
    info!("has_tx_result {:?}", has_tx_result);
}

[jonathangenlambda]

Also it seems that query_next_tx blocks forever if there is no TX in the mempool - not sure if this is the intended behaviour.

[jonathangenlambda]

Using a tokio timeout also doesnt work with query_next_tx as it renders the client in an invalid state ending up all subsequent requests in Err(AgencyIsTheirs):

let query_next_tx_result = monitor_client.query_next_tx();
if let Err(_) = tokio::time::timeout(Duration::from_millis(10), query_next_tx_result).await {
    info!("query_next_tx_result didnt return");
}

[scarmuega]

@jonathangenlambda thanks for reporting. I'll try to reproduce and then get back to you.