CTL + C cannot stop

[tingfeng-key]

When a network error occurs, it cannot be stopped through CTL + C, and it will cycle all the time.
And when the network is normal, CTL + C will stop service very slowly. Can you optimize it?

Meta

• teloxide version: 0.5.3
• rustc version: 1.57.0

[p0lunin]

#470 changes a little update processing. Can you check if the bug is reproducible with https://github.com/teloxide/teloxide/tree/dispatching2?

[tingfeng-key]

#470 changes a little update processing. Can you check if the bug is reproducible with https://github.com/teloxide/teloxide/tree/dispatching2?

Still the same

[WaffleLapkin]

So, this is a bug in the polling implementation:

teloxide/src/dispatching/update_listeners/polling.rs

Lines 74 to 82 in c12fae2

	if flag.is_stopped() {
	let mut req = bot.get_updates().offset(*offset).timeout(0).limit(1);
	req.payload_mut().allowed_updates = allowed_updates.take();
	return match req.send().await {
	Ok(_) => None,
	Err(err) => Some((Either::Left(stream::once(ready(Err(err)))), state)),
	};
	}

When stopping we call get_updates one last time to not receive the same updates next time. But if this request fails we return an error and do not stop until the call succeeds.