Struct PublicationQueue 

pub struct PublicationQueue { /* private fields */ }

A bounded FIFO whose render-side operation performs no transport I/O.

A marker is returned only after the complete validated snapshot+commit pair has been encoded and admitted atomically. Full queues consume no revision. A transport failure permanently closes admission.

Implementations

impl PublicationQueue

pub fn new(client: Client, capacity: usize) -> Result<Self, Error>

Move a connected client into a dedicated writer thread.

Construct this only after the handshake; dormant applications never create a queue or thread. Capacity must be positive.

pub fn limits(&self) -> &Limits

Negotiated snapshot limits used by a framework tree builder.

pub fn publish( &mut self, snapshot: &mut Snapshot, ) -> Result<Option<String>, Error>

Validate, encode and non-blockingly admit one complete revision.

pub fn dropped(&self) -> u64

Revisions dropped before admission because the FIFO was full.

pub fn failed(&self) -> bool

Whether transport failed and no later marker may be issued.

pub fn fail(&mut self, code: impl Into<String>, message: impl Into<String>)

Fail closed without writing from the caller’s thread. Already-admitted revisions drain in order, then the worker sends the typed fatal.

pub fn shutdown(self) -> bool

Close admission, drain the ordered worker and wait for its causal completion. Framework lifecycle code calls this outside render hooks so a short-lived one-frame process cannot exit ahead of its semantic data. Returns whether the worker completed without a transport failure.