Struct Acu 

pub struct Acu<T: Transport, C: Clock> {
    pub reply_delay_ms: u32,
    pub retry: RetryConfig,
    /* private fields */
}

ACU driver.

Fields

reply_delay_ms: u32

Reply-delay budget per attempt, in milliseconds.

retry: RetryConfig

Retry policy applied by Acu::exchange.

Implementations

impl<T: Transport, C: Clock> Acu<T, C>

pub fn new(transport: T, clock: C) -> Self

New driver with default reply delay and retry policy.

pub fn transport(&mut self) -> &mut T

Borrow the underlying transport.

pub fn clock(&self) -> &C

Borrow the clock.

pub fn send_to( &mut self, pd_addr: u8, pd: &mut PdState, command: &Command, ) -> Result<Vec<u8>, Error>

Encode and send command to pd_addr once. Returns the bytes written.

pub fn receive(&mut self, pd: &mut PdState) -> Result<Reply, Error>

Drain whatever bytes the transport has, then attempt to parse one reply. Returns Err(Error::Timeout) once the per-attempt reply-delay budget elapses without a complete packet.

The loop reads up to a small fixed number of times per call to drain a transport that may return short reads. When the underlying transport returns Ok(0) (no more bytes immediately ready), we check the deadline and either return Timeout or immediately return — the caller is expected to call us again later.

pub fn exchange( &mut self, pd_addr: u8, pd: &mut PdState, command: &Command, ) -> Result<ExchangeOutcome, Error>

Run a command/reply round-trip with full retry & off-line policy.

• On a clean reply, returns ExchangeOutcome::Reply.
• On osdp_BUSY, returns ExchangeOutcome::Busy without bumping SQN (per Annex A.2: BUSY’s SQN is always 0). The caller decides whether to retry now or service other PDs first.
• On timeout, retries up to RetryConfig::max_retries additional times re-using the same SQN — that asks the PD to repeat its prior reply, per §5.7 / Table 2.
• When the PD has been silent for ≥ crate::OFFLINE_THRESHOLD_MS, returns ExchangeOutcome::Offline.