use std::time::{Duration, Instant};
use std::ops::Deref;
-/// BackgroundProcessor takes care of tasks that (1) need to happen periodically to keep
+/// `BackgroundProcessor` takes care of tasks that (1) need to happen periodically to keep
/// Rust-Lightning running properly, and (2) either can or should be run in the background. Its
/// responsibilities are:
-/// * Monitoring whether the ChannelManager needs to be re-persisted to disk, and if so,
+/// * Processing [`Event`]s with a user-provided [`EventHandler`].
+/// * Monitoring whether the [`ChannelManager`] needs to be re-persisted to disk, and if so,
/// writing it to disk/backups by invoking the callback given to it at startup.
-/// ChannelManager persistence should be done in the background.
-/// * Calling `ChannelManager::timer_tick_occurred()` and
-/// `PeerManager::timer_tick_occurred()` every minute (can be done in the
-/// background).
+/// [`ChannelManager`] persistence should be done in the background.
+/// * Calling [`ChannelManager::timer_tick_occurred`] and [`PeerManager::timer_tick_occurred`]
+/// at the appropriate intervals.
///
-/// Note that if ChannelManager persistence fails and the persisted manager becomes out-of-date,
-/// then there is a risk of channels force-closing on startup when the manager realizes it's
-/// outdated. However, as long as `ChannelMonitor` backups are sound, no funds besides those used
-/// for unilateral chain closure fees are at risk.
+/// It will also call [`PeerManager::process_events`] periodically though this shouldn't be relied
+/// upon as doing so may result in high latency.
+///
+/// # Note
+///
+/// If [`ChannelManager`] persistence fails and the persisted manager becomes out-of-date, then
+/// there is a risk of channels force-closing on startup when the manager realizes it's outdated.
+/// However, as long as [`ChannelMonitor`] backups are sound, no funds besides those used for
+/// unilateral chain closure fees are at risk.
+///
+/// [`ChannelMonitor`]: lightning::chain::channelmonitor::ChannelMonitor
+/// [`Event`]: lightning::util::events::Event
#[must_use = "BackgroundProcessor will immediately stop on drop. It should be stored until shutdown."]
pub struct BackgroundProcessor {
stop_thread: Arc<AtomicBool>,
/// `persist_manager` returns an error. In case of an error, the error is retrieved by calling
/// either [`join`] or [`stop`].
///
- /// Typically, users should either implement [`ChannelManagerPersister`] to never return an
- /// error or call [`join`] and handle any error that may arise. For the latter case, the
- /// `BackgroundProcessor` must be restarted by calling `start` again after handling the error.
+ /// # Data Persistence
///
/// `persist_manager` is responsible for writing out the [`ChannelManager`] to disk, and/or
/// uploading to one or more backup services. See [`ChannelManager::write`] for writing out a
/// [`ChannelManager`]. See [`FilesystemPersister::persist_manager`] for Rust-Lightning's
/// provided implementation.
///
+ /// Typically, users should either implement [`ChannelManagerPersister`] to never return an
+ /// error or call [`join`] and handle any error that may arise. For the latter case,
+ /// `BackgroundProcessor` must be restarted by calling `start` again after handling the error.
+ ///
+ /// # Event Handling
+ ///
+ /// `event_handler` is responsible for handling events that users should be notified of (e.g.,
+ /// payment failed). A user's [`EventHandler`] may be decorated with other handlers to implement
+ /// common functionality. See individual [`Event`]s for further details.
+ ///
/// [top-level documentation]: Self
/// [`join`]: Self::join
/// [`stop`]: Self::stop
/// [`ChannelManager`]: lightning::ln::channelmanager::ChannelManager
/// [`ChannelManager::write`]: lightning::ln::channelmanager::ChannelManager#impl-Writeable
/// [`FilesystemPersister::persist_manager`]: lightning_persister::FilesystemPersister::persist_manager
+ /// [`Event`]: lightning::util::events::Event
pub fn start<
Signer: 'static + Sign,
CF: 'static + Deref + Send + Sync,