Github user maochf commented on a diff in the pull request:
https://github.com/apache/twill/pull/16#discussion_r94559673
--- Diff:
twill-core/src/test/java/org/apache/twill/kafka/client/KafkaTest.java ---
@@ -170,11 +174,128 @@ public void testKafkaClient() throws Exception {
Cancellable cancel = kafkaClient.getConsumer().prepare().add(topic, 0,
0).consume(new KafkaConsumer
.MessageCallback() {
@Override
- public void onReceived(Iterator<FetchedMessage> messages) {
+ public long onReceived(Iterator<FetchedMessage> messages) {
+ long nextOffset = Long.MIN_VALUE;
+ while (messages.hasNext()) {
+ FetchedMessage message = messages.next();
+ LOG.info(Charsets.UTF_8.decode(message.getPayload()).toString());
+ latch.countDown();
+ }
+ return nextOffset;
+ }
+
+ @Override
+ public void finished() {
+ stopLatch.countDown();
+ }
+ });
+
+ Assert.assertTrue(latch.await(5, TimeUnit.SECONDS));
+ cancel.cancel();
+ Assert.assertTrue(stopLatch.await(1, TimeUnit.SECONDS));
+ }
+
+ @Test
+ public void testKafkaClientReadFromIdx() throws Exception {
+ String topic = "testClient";
+
+ // Publish 30 messages with indecies the same as offsets within the
range 0 - 29
+ Thread t1 = createPublishThread(kafkaClient, topic, Compression.GZIP,
"GZIP Testing message", 10);
+ t1.start();
+ t1.join();
+ Thread t2 = createPublishThread(kafkaClient, topic, Compression.NONE,
"Testing message", 10, 10);
+ t2.start();
+ t2.join();
+ Thread t3 = createPublishThread(kafkaClient, topic,
Compression.SNAPPY, "Snappy Testing message", 10, 20);
+ t3.start();
+ t3.join();
+
+ final int startIdx = 15;
+ final CountDownLatch latch = new CountDownLatch(30 - startIdx);
+ final CountDownLatch stopLatch = new CountDownLatch(1);
+ final BlockingQueue<Long> offsetQueue = new LinkedBlockingQueue<>();
+ // Creater a consumer
+ final SimpleKafkaConsumer consumer = (SimpleKafkaConsumer)
kafkaClient.getConsumer();
+ Cancellable initCancel =
kafkaClient.getConsumer().prepare().add(topic, 0, 0).consume(new KafkaConsumer
+ .MessageCallback() {
+ long minOffset = -2; // earliest msg
+ long maxOffset = -1; // latest msg
+ @Override
+ // Use binary search to find the offset of the message with the
index matching startIdx. Returns the next offset
--- End diff --
Right, since the offsets and indices of the messages are known, it can
directly jump to the desired idx without binary search.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
