Hi, I'm trying to use MockConsumer to test my application code but I've faced a couple of limitations and I want to know if there are workarounds or something that I'm overlooking. Note: I'm using kafka-clients v 0.11.0.2
1. Why the addRecord <https://github.com/apache/kafka/blob/73be1e1168f91ee2a9d68e1d1c75c14018cf7d3a/clients/src/main/java/org/apache/kafka/clients/consumer/MockConsumer.java#L179> requires that the consumer has assigned partitions? Given that this is just simulating records being produced or existing records. 2. Why the poll <https://github.com/apache/kafka/blob/73be1e1168f91ee2a9d68e1d1c75c14018cf7d3a/clients/src/main/java/org/apache/kafka/clients/consumer/MockConsumer.java#L132> clear the map of records? It should not be cleared after commit? 3. Why the commitAsync <https://github.com/apache/kafka/blob/73be1e1168f91ee2a9d68e1d1c75c14018cf7d3a/clients/src/main/java/org/apache/kafka/clients/consumer/MockConsumer.java#L198> doesn't check for an exception and always succeed? Due to items (2) and (3) I'm not be able to test scenarios where the commits fails and the consumer should poll again the same elements. If someone knows about other scenarios that can't be tested with MockConsumer, please let me know. Thanks.
