chevaris commented on code in PR #1264: URL: https://github.com/apache/curator/pull/1264#discussion_r2046278696
########## curator-recipes/src/test/java/org/apache/curator/framework/recipes/nodes/TestPersistentTtlNode.java: ########## @@ -237,4 +241,88 @@ void touch() { } } } + + @Test + public void testInternalExecutorClose() throws Exception { + final String mainPath = "/parent/main"; + final String touchPath = ZKPaths.makePath(mainPath, PersistentTtlNode.DEFAULT_CHILD_NODE_NAME); + final CountDownLatch touchCreatedLatch = new CountDownLatch(1); + try (CuratorFramework client = + CuratorFrameworkFactory.newClient(server.getConnectString(), new RetryOneTime(1))) { + client.start(); + assertTrue(client.blockUntilConnected(1, TimeUnit.SECONDS)); + try (PersistentWatcher watcher = new PersistentWatcher(client, mainPath, true)) { + final Watcher listener = event -> { + final String path = event.getPath(); + if (touchPath.equals(path)) { + touchCreatedLatch.countDown(); + } + }; + watcher.getListenable().addListener(listener); + watcher.start(); + final AtomicLong executorThreadId = new AtomicLong(); + try (PersistentTtlNode node = new PersistentTtlNode(client, mainPath, ttlMs, new byte[0])) { + node.start(); + assertTrue(touchCreatedLatch.await(5 * ttlMs, TimeUnit.MILLISECONDS)); + node.getCloseableScheduledExecutorService() + .submit(() -> + executorThreadId.set(Thread.currentThread().getId())); + assertFalse(getThreadsWithIdAndName(executorThreadId, PersistentTtlNode.TOUCH_THREAD_NAME) + .isEmpty()); + } + Thread.sleep(10L); Review Comment: I have my doubts here what is best, because awaitTermination assures you that tasks were executed, BUT the Javadoc does NOT confirm that the thread has been destroyed, so we can only rely that will happen eventually After closing the PersistentTTLNode: - When executor is internal, probably best is a loop with small sleeps to check that thread disappears. Notice also that CloseExecutorService is in another package and extracting methods for testing requires them to be public that I really do NOT like. - When executor is provided, probably the best is to use awaitTermination and after that sleep for instance 100 msecs and check that the thread is still there What do you think? -- 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: commits-unsubscr...@curator.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org