Struct ConversationActor 

pub struct ConversationActor { /* private fields */ }

Implementations

impl ConversationActor

pub fn handle(&self) -> ConversationHandle

Returns a cloneable command handle.

pub fn pid(&self) -> Result<ParticipantPid, LiminalError>

Returns the current actor PID, restarting after crash when needed.

Errors

Returns LiminalError when the actor is closed or cannot restart.

pub fn state(&self) -> Result<ConversationState, LiminalError>

Queries actor state.

Errors

Returns LiminalError when the actor cannot service the query.

pub fn receive_timeout( &self, timeout: Duration, ) -> Result<Envelope, LiminalError>

Receives the next reply from the conversation, bounded by timeout.

Returns LiminalError::ConversationTimeout if no reply arrives in time, or LiminalError::ParticipantCrashed if a linked participant crashes while waiting (the crash drains the pending receive immediately).

Errors

Returns LiminalError on timeout, participant crash, or actor failure.

pub fn try_take_reply(&self) -> Option<Envelope>

R1(vi)(a): non-blocking drain of one buffered participant reply, if any. The connection polls this on its own slice (woken by the reply-availability notifier) instead of blocking the slice on receive_timeout.

pub fn has_pending_reply(&self) -> bool

Whether a participant reply is buffered without consuming it.

pub fn register_reply_notifier(&self, notifier: Arc<dyn Fn() + Send + Sync>)

R1(vi)(a): installs the reply-availability notifier, fired on the reply queue’s empty→non-empty transition and on terminal actor error. Installed permanently at conversation open; cleared at close/finalize.

pub fn finalize(&self)

Finalizes the conversation without requiring its actor process to run: bounded, non-blocking, and idempotent. Terminates the actor and every participant directly (scheduler tombstone writes, not requests into the actor’s command loop), fails pending receives and queued commands with the typed closed error, and removes both runtime registrations. This is the teardown-path counterpart to ConversationHandle::close: a caller releasing a conversation during ITS OWN teardown must never block on the conversation scheduler being live or responsive.

pub fn notify_on_participant_exit( &self, participant: ParticipantPid, notifier: SyncSender<Instant>, ) -> Result<(), LiminalError>

Registers a one-shot notifier fired the instant participant’s trapped EXIT is processed (carrying the observed Instant — a structural link wakeup, not a poll). If participant is already dead at registration (it crashed before this call), the recorded EXIT instant is replayed immediately, so a crash-before-register is never lost. See [ActorCore::register_exit_notifier].

Errors

Returns LiminalError when a state or registry lock is poisoned.

Trait Implementations

impl Clone for ConversationActor

fn clone(&self) -> ConversationActor

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ConversationActor

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.