Ask for a resync

The app asks a source for a pass at once by sending a Resync to the inbox that spawn or spawn_checked returned:

use pito_sync::Resync;

asks.send(Resync("feed".into(), None).into())?;

Resync(name, id) becomes a Message::Resync with .into(). The thread runs a pass and hands id to pass as its second argument. A source decides what an id means: "full" for a full resync instead of an incremental one, for example, or the id of the one item to fetch again.

fn pass(&mut self, conn: &mut Connection, id: Option<&str>) -> Result<usize, Failure> {
    let since = match id {
        Some("full") => None,
        _ => pito_sync::sync(conn, "feed").ok().flatten().and_then(|s| s.cursor),
    };
    let items = self.fetch(since.as_deref())?;
    self.write(conn, items)
}

fetch and write stand for the source’s own code.

It always runs

An explicit resync is never dropped. It runs right after another pass, and it runs while a retry_after holds back the timed passes, so a user’s refresh key always does something. A resync took 33 µs from the send to its closing Changed in the measured run.

What the UI sees

Like every pass, a resync sends Changed(name) when it starts and when it ends. The feed example sends three resyncs, the second one with the id "offline", which its source treats as a server that doesn’t answer:

for id in [None, Some("offline"), None] {
    pause();
    let _ = asks.send(Resync("feed".into(), id.map(String::from)).into());
}

feed      syncing…
feed      6 posts, newest "One file per app"
feed      syncing…
feed      unreachable: the server did not answer (6 posts kept)
feed      syncing…
feed      9 posts, newest "The sync table"

When the source’s thread has ended, because the file couldn’t be opened or the Changed receiver was dropped, send returns Err.