Struct PtyReadChunk 

struct PtyReadChunk {
    data: Vec<u8>,
    osc_events: Vec<OscEvent>,
    ack: Option<Sender<()>>,
}

One PTY read’s worth of already-OSC-cleaned output, on its way from the blocking read loop to the coalescing flusher (start(), below). See docs/reports/REPORT_RENDERER_CPU_UNBATCHED_PTY_OUTPUT_2026_09_11.md.

Fields

data: Vec<u8>

osc_events: Vec<OscEvent>

ack: Option<Sender<()>>

Flush barrier. When set, the flusher signals it AFTER committing the batch this chunk lands in — so a sender that enqueues a barrier and awaits it knows everything queued before it has reached the FileStore and the broker.

Exists because “the flusher task ended” is NOT a usable completion signal on Windows: ConPTY’s console host outlives the direct child, so the read loop never sees EOF, never drops its sender, and the channel never closes. Waiting on the task therefore always burned the full FLUSHER_DRAIN_TIMEOUT (see its doc comment) even though the flusher itself had been idle for most of that second. A barrier asks the question actually being asked — “is prior output committed yet?” — and is answered as soon as the batch it lands in commits.