Struct WatcherJournalSource 

pub(super) struct WatcherJournalSource {
    roots: Vec<String>,
}

Non-elevated change source for scope mode. Holds the configured roots for Phase 2’s watchers; Phase 1 keeps them only to log/identify the session.

Fields

roots: Vec<String>

Implementations

impl WatcherJournalSource

pub(super) const fn new(roots: Vec<String>) -> Self

Trait Implementations

impl JournalSource for WatcherJournalSource

fn open(&mut self) -> Result<(), UsnError>

(Re)open the journal, creating it when missing. Positions the cursor at the journal’s current end. Called once per establish cycle: at start and after every journal-gone rescan.

fn query(&mut self) -> Result<JournalView, UsnError>

Live journal identity/retention for checkpoint validation.

fn read_blocking(&mut self, _buf: &mut Vec<u8>) -> Result<ReadOutcome, UsnError>

Blocking read of the next batch. Semantics the worker relies on: blocks until records exist, then returns them and advances the cursor past the batch; Gone when the journal died under us; Err on an unrecoverable failure. An empty Records batch is a benign wakeup — the worker re-checks its stop flag and reads again (fakes use this to unblock on stop; the live read returns on the next volume write, which is what keeps Engine::shutdown’s join prompt).

fn journal_id(&self) -> u64

Journal id of the open session (checkpoint identity).

fn next_usn(&self) -> i64

Cursor the next read starts at (the value persisted in checkpoints).

fn set_next_usn(&mut self, _usn: i64)

Reposition the cursor (a snapshot restore replays from its persisted checkpoint instead of the journal’s current end).

fn open_stat_fetcher(&self) -> Result<Box<dyn StatFetcher>, UsnError>

Size/mtime fetcher bound to the same volume, opened once per tail session. Per-record stat calls were already dynamic (usn::apply).