Struct WriteEncoder 

pub struct WriteEncoder<'a> { /* private fields */ }

A borrowed write producer. Before emitting, it encodes one record at a time to measure each part’s byte length and construct the header; emitting a part retains at most one record frame and one 64-KiB chunk.

Implementations

impl<'a> WriteEncoder<'a>

pub fn new(record: &'a WriteRecord) -> Result<Self, Error>

Validate the complete in-memory write and measure each part’s stream.

pub fn with_records( record: &'a WriteRecord, encoded: &'a [Vec<Vec<u8>>], ) -> Result<Self, Error>

new for a write whose records the caller encoded already, a part’s WritePart::encode_records each, in the parts’ order: they are emitted as given, validated and encoded once, and the bytes are those new produces.

pub fn header(&self) -> &WriteHeaderFrame

Metadata that lets a receiver delimit each part without opening it.

pub fn header_frame(&self) -> &[u8]

The header’s complete kind-1 plaintext frame.

pub fn plaintext_length(&self) -> u64

The queued plaintext length, for the database’s SQLite value limit.

pub fn encode_plaintext(&self, output: &mut [u8]) -> Result<(), Error>

Encode the upload-queue value into an exactly sized buffer. The database checks Self::plaintext_length against SQLite’s limit before allocating this buffer. A different buffer length is refused before writing bytes.

pub fn part_chunks( &self, index: usize, ) -> Result<PlaintextChunks<impl Iterator<Item = Result<Vec<u8>, Error>> + '_>, Error>

Cut the selected audience’s change and dismissal frames into canonical plaintext chunks.