dimas-b commented on code in PR #4648:
URL: https://github.com/apache/polaris/pull/4648#discussion_r3397285127
##########
runtime/service/src/main/java/org/apache/polaris/service/events/PolarisEventListeners.java:
##########
@@ -136,4 +165,55 @@ private void deliverEvent(
public boolean hasListeners(PolarisEventType eventType) {
return eventTypesWithListeners.contains(eventType);
}
+
+ private static class ListenerExecutor implements Executor {
+
+ private final Executor delegate;
+ private final Queue<Runnable> queue;
+ private final AtomicBoolean draining = new AtomicBoolean(false);
+
+ ListenerExecutor(Executor delegate, int capacity) {
+ this.delegate = delegate;
+ this.queue =
+ capacity == -1 ? new ConcurrentLinkedQueue<>() : new
ArrayBlockingQueue<>(capacity);
+ }
+
+ @Override
+ public void execute(@NonNull Runnable task) {
+ if (!queue.offer(task)) {
+ throw new RejectedExecutionException("Event listener queue is full");
+ }
+ if (draining.compareAndSet(false, true)) {
+ try {
+ delegate.execute(this::drain);
+ } catch (Throwable t) {
+ draining.set(false);
+ throw t;
+ }
+ }
+ }
+
+ private void drain() {
+ try {
+ Runnable task;
+ while ((task = queue.poll()) != null) {
+ task.run();
+ }
+ } finally {
+ draining.set(false);
+ // A producer may have enqueued a task between the last poll() and the
set(false) above,
+ // and seen draining=true so it did not schedule a drain. Catch that
here.
+ if (!queue.isEmpty() && draining.compareAndSet(false, true)) {
+ try {
+ // re-schedule for fairness
+ delegate.execute(this::drain);
+ } catch (Exception e) {
Review Comment:
With `CompletableFuture` I guess we have to log task exceptions in our code,
otherwise they will remain in the `CompletableFuture` undiscovered, because
nobody waits on it, right? :thinking:
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
