async fn republish_commitments<S>(
session_id: SessionId,
pseudonym: HoprPseudonym,
routing: &DestinationRouting,
ssa_state: &SessionSsaState,
pix_toolbox: &PixToolbox,
msg_sender: &mut S,
members: Vec<SsaIndex>,
) -> Result<()>where
S: Sink<(DestinationRouting, ApplicationDataOut)> + Unpin,
S::Error: Error + Send + Sync + Clone + 'static,Expand description
Sends the whole burst of each of members again, rebuilt from the material the generator
retains, and announces nothing.
This answers a repeat that arrives too late to have crossed the batch’s bursts (see
SSA_REPEAT_RESEND_GRACE): some member’s burst went out and none of it reached the Exit. The
repeat names the whole batch, not the member that is missing, so every announced member is sent
again, and a member the Exit already holds absorbs the copy. There is no ReadyToDeposit,
because each of members was announced when its publication completed (see
OwedPublications::announced).
Each re-send spends one of the cycle’s
MAX_COMMITMENT_RETRANSMISSIONS, the same
budget a scoped re-request draws on. Once the generator refuses to rebuild a cycle, because that
budget is spent or its material is gone, the cycle is no longer counted as announced, so later
repeats do not ask again. A transport failure stops the re-send where it is and is the non-fatal
SessionManagerError::NotAdmitted. The Exit’s next repeat after the grace tries again.