published

published::<T>(name) opens the record of one destination on a broker: every message published there, with its name and headers.

T is the payload type that with and decoded decode to; the count and raw assertions ignore it. The assertions are those of PublishedAssertions, such as assert_called_once. In process the record is the broker's own publish log, so it holds replies, Out slot publishes, retry copies and the test's own publishes. out::<M>() shows what left through one slot, across brokers, and subscriber(name) is the record of a handler. Against a running stand it holds the test's publishes and everything a publisher wired by the runtime sent. The record is a snapshot taken when the method is called: call it after the publish that drives the reaction.

Example

Macros

examples/testing_records.rs

#[tokio::test]
async fn a_receipt_is_published() -> Result<(), Box<dyn Error>> {
    let tb = TestApp::start(app()).await?;
    let broker = tb.broker::<MemoryBroker>();
    broker
        .message(&Order { id: 7 })
        .to("orders")
        .publish()
        .await?;

    broker
        .published::<Receipt>("receipts")
        .assert_called_once()
        .with(&Receipt { order_id: 7 });
    Ok(())
}

Manual

examples/manual/testing_records.rs

#[tokio::test]
async fn a_receipt_is_published() -> Result<(), Box<dyn Error>> {
    let tb = TestApp::start(app()).await?;
    let broker = tb.broker::<MemoryBroker>();
    broker
        .message(&Order { id: 7 })
        .to("orders")
        .publish()
        .await?;

    broker
        .published::<Receipt>("receipts")
        .assert_called_once()
        .with(&Receipt { order_id: 7 });
    Ok(())
}