Struct TestPeer 

pub struct TestPeer { /* private fields */ }

Implementations

impl TestPeer

pub fn new() -> (Self, Builder<Client, impl HandleDispatchFrom<Agent>, NullRun>)

Build a TestPeer plus a pre-wired Client.builder() whose notification handlers route session/mcp/elicitation traffic into the peer. The caller decides whether to run the builder via connect_to (drop the agent-side cx) or connect_with (capture the agent-side cx).

pub async fn next_session_notification(&mut self) -> SessionNotification

pub async fn next_mcp_notification(&mut self) -> McpNotification

pub async fn next_elicitation_request(&mut self) -> ElicitationParams

pub fn queue_elicitation_response(&self, response: ElicitationResponse)

Queue a response the peer will hand back for the next incoming elicitation request. If the queue is empty when a request arrives, the peer responds with a protocol error, which exercises the cancel_result() fallback path in the caller.

pub async fn fake_elicitation( &mut self, cx: &ConnectionTo<Client>, ) -> (Responder<ElicitationResponse>, Receiver<ElicitationResponse>)

Kick off a placeholder elicitation request from the agent side of cx, hand back the Responder<ElicitationResponse> captured on the client side, and return a receiver that resolves when the responder is consumed.

Use in tests that pass a Responder<ElicitationResponse> into code under test and want to observe the response without driving a full ACP round-trip themselves.