@hachikuji @lindong28 Here is the update. As you can see, the first commit reverts the down-conversion logic and the second commit re-implements it. Line 170 of `LazyDownConversionRecords` handles the compressionType change in the middle of the batches (case 2) and Line 63 handles the case where the first batch is compressed with Zstd (case 1). Updated `KafkaApis` handles `IllegalArgumentException` exception thrown by `LazyDownConversionRecords`.
However, as you can see in `FetchRequestTest`, it seems like the case 1 handling is not working properly. Could you have a look if you have some spare time? Thanks. [ Full content available at: https://github.com/apache/kafka/pull/2267 ] This message was relayed via gitbox.apache.org for [email protected]
