Trait SupervisedStream 

pub trait SupervisedStream: Send {
    type Item: Send;
    type ConnectError: Display + Send;
    type StreamError: Display + Send;
    type Stream: Stream<Item = Result<Self::Item, Self::StreamError>> + Send;

    // Required methods
    fn name(&self) -> &str;
    fn backoff_policy(&self) -> BackoffPolicy;
    fn connect<'life0, 'async_trait>(
        &'life0 mut self,
    ) -> Pin<Box<dyn Future<Output = Result<Self::Stream, Self::ConnectError>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait;
    fn on_message<'life0, 'async_trait>(
        &'life0 mut self,
        item: Self::Item,
    ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait;

    // Provided methods
    fn on_shutdown<'life0, 'async_trait>(
        &'life0 mut self,
    ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait { ... }
    fn supervise<'life0, 'async_trait, S>(
        &'life0 mut self,
        shutdown: S,
    ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>
       where S: Future<Output = ()> + Send + 'async_trait,
             Self: 'async_trait,
             'life0: 'async_trait { ... }
}

A supervised, self-reconnecting streaming subscription.

Implementors own the connection state (clients, channels, publishers) as fields, so the reconnect/backoff/shutdown loop can hand each item to on_message without the per-item state-cloning a closure-based supervisor forces. The provided supervise method owns that loop; an implementor supplies only how to connect, how to handle an item, and (optionally) how to tear down.

Required Associated Types

type Item: Send

Item the stream yields and on_message consumes.

type ConnectError: Display + Send

Error a failed connect attempt yields. Logged, then retried.

type StreamError: Display + Send

Error the stream may yield per item. Logged, then reconnected.

type Stream: Stream<Item = Result<Self::Item, Self::StreamError>> + Send

The stream a successful connect opens.

Required Methods

fn name(&self) -> &str

Names this subscription in the supervisor’s reconnect/close logs.

fn backoff_policy(&self) -> BackoffPolicy

Backoff policy for reconnect attempts.

fn connect<'life0, 'async_trait>( &'life0 mut self, ) -> Pin<Box<dyn Future<Output = Result<Self::Stream, Self::ConnectError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Open a fresh stream of items.

fn on_message<'life0, 'async_trait>( &'life0 mut self, item: Self::Item, ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Handle one delivered item. Awaited to completion, so a slow handler holds the read loop; offload work that must not block it.

Provided Methods

fn on_shutdown<'life0, 'async_trait>( &'life0 mut self, ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Teardown run once before supervise returns, on every exit path. Cancel tokens or release resources here.

fn supervise<'life0, 'async_trait, S>( &'life0 mut self, shutdown: S, ) -> Pin<Box<dyn Future<Output = ()> + Send + 'async_trait>>

where S: Future<Output = ()> + Send + 'async_trait, Self: 'async_trait, 'life0: 'async_trait,

Keep the subscription alive across reconnects until shutdown resolves.

Every item connect yields is handed to on_message. Reconnect timing follows BackoffPolicy: an opened stream that later closes or errors reconnects at the floor, since the connection itself was healthy.

shutdown stops the supervisor promptly whenever it is waiting: to connect, for the next item, or during a backoff. It does not interrupt an in-flight on_message. on_shutdown runs on every exit path.

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors