Add join method to BackgroundProcessor
authorJeffrey Czyz <jkczyz@gmail.com>
Mon, 19 Jul 2021 17:50:56 +0000 (12:50 -0500)
committerJeffrey Czyz <jkczyz@gmail.com>
Wed, 28 Jul 2021 21:30:51 +0000 (16:30 -0500)
The previous commit wraps the background thread's JoinHandle in an
Option. Providing a dedicated method to join hides this implementation
detail from users.

lightning-background-processor/src/lib.rs

index cc4c9e635025117c6b87be05f6615c4788f7092b..55ac14c048c412e8a9d0a7f17318d62b7569b3f6 100644 (file)
@@ -41,9 +41,7 @@ use std::ops::Deref;
 /// for unilateral chain closure fees are at risk.
 pub struct BackgroundProcessor {
        stop_thread: Arc<AtomicBool>,
-       /// May be used to retrieve and handle the error if `BackgroundProcessor`'s thread
-       /// exits due to an error while persisting.
-       pub thread_handle: Option<JoinHandle<Result<(), std::io::Error>>>,
+       thread_handle: Option<JoinHandle<Result<(), std::io::Error>>>,
 }
 
 #[cfg(not(test))]
@@ -84,21 +82,25 @@ ChannelManagerPersister<Signer, M, T, K, F, L> for Fun where
 }
 
 impl BackgroundProcessor {
-       /// Start a background thread that takes care of responsibilities enumerated in the top-level
-       /// documentation.
+       /// Start a background thread that takes care of responsibilities enumerated in the [top-level
+       /// documentation].
        ///
-       /// If `persist_manager` returns an error, then this thread will return said error (and
-       /// `start()` will need to be called again to restart the `BackgroundProcessor`). Users should
-       /// wait on [`thread_handle`]'s `join()` method to be able to tell if and when an error is
-       /// returned, or implement `persist_manager` such that an error is never returned to the
-       /// `BackgroundProcessor`
+       /// The thread runs indefinitely unless the object is dropped, [`stop`] is called, or
+       /// `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.
        ///
        /// `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.
        ///
-       /// [`thread_handle`]: BackgroundProcessor::thread_handle
+       /// [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
@@ -161,7 +163,29 @@ impl BackgroundProcessor {
                Self { stop_thread: stop_thread_clone, thread_handle: Some(handle) }
        }
 
-       /// Stop `BackgroundProcessor`'s thread.
+       /// Join `BackgroundProcessor`'s thread, returning any error that occurred while persisting
+       /// [`ChannelManager`].
+       ///
+       /// # Panics
+       ///
+       /// This function panics if the background thread has panicked such as while persisting or
+       /// handling events.
+       ///
+       /// [`ChannelManager`]: lightning::ln::channelmanager::ChannelManager
+       pub fn join(mut self) -> Result<(), std::io::Error> {
+               assert!(self.thread_handle.is_some());
+               self.join_thread()
+       }
+
+       /// Stop `BackgroundProcessor`'s thread, returning any error that occurred while persisting
+       /// [`ChannelManager`].
+       ///
+       /// # Panics
+       ///
+       /// This function panics if the background thread has panicked such as while persisting or
+       /// handling events.
+       ///
+       /// [`ChannelManager`]: lightning::ln::channelmanager::ChannelManager
        pub fn stop(mut self) -> Result<(), std::io::Error> {
                assert!(self.thread_handle.is_some());
                self.stop_and_join_thread()
@@ -169,6 +193,10 @@ impl BackgroundProcessor {
 
        fn stop_and_join_thread(&mut self) -> Result<(), std::io::Error> {
                self.stop_thread.store(true, Ordering::Release);
+               self.join_thread()
+       }
+
+       fn join_thread(&mut self) -> Result<(), std::io::Error> {
                match self.thread_handle.take() {
                        Some(handle) => handle.join().unwrap(),
                        None => Ok(()),
@@ -430,7 +458,7 @@ mod tests {
                let persister = |_: &_| Err(std::io::Error::new(std::io::ErrorKind::Other, "test"));
                let event_handler = |_| {};
                let bg_processor = BackgroundProcessor::start(persister, event_handler, nodes[0].chain_monitor.clone(), nodes[0].node.clone(), nodes[0].peer_manager.clone(), nodes[0].logger.clone());
-               match bg_processor.stop() {
+               match bg_processor.join() {
                        Ok(_) => panic!("Expected error persisting manager"),
                        Err(e) => {
                                assert_eq!(e.kind(), std::io::ErrorKind::Other);