-Re: Review Request 15805: KAFKA-1140.v2: addressed Jun's comments
Jun Rao 2013-11-26, 18:19
Yes, we will have to pass in the decoders to MessageAndMetadata.
On Tue, Nov 26, 2013 at 10:09 AM, Guozhang Wang <[EMAIL PROTECTED]> wrote:
> Does this enforce the key/value decoder to be passed also into
> *From:* Jun Rao [[EMAIL PROTECTED]] on behalf of Jun Rao [
> [EMAIL PROTECTED]]
> *Sent:* Tuesday, November 26, 2013 9:30 AM
> *To:* kafka; Jun Rao; Guozhang Wang
> *Subject:* Re: Review Request 15805: KAFKA-1140.v2: addressed Jun's
> This is an automatically generated e-mail. To reply, visit:
> Thinking about this a bit more. It seems that a better approach is to move the decoding into MessageAndMetadata. We define two methods key() and message() that do the decoding and throw a RuntimeException back to the caller if decoding fails. This way, the client can still get the metadata (offset, partitionId, etc) associated with a message even when decoding fails.
> - Jun Rao
> On November 25th, 2013, 8:55 p.m. UTC, Guozhang Wang wrote:
> Review request for kafka.
> By Guozhang Wang.
> *Updated Nov. 25, 2013, 8:55 p.m.*
> *Bugs: *KAFKA-1140 <https://issues.apache.org/jira/browse/KAFKA-1140>
> *Repository: *kafka
> - core/src/main/scala/kafka/consumer/ConsumerIterator.scala
> - core/src/test/scala/unit/kafka/consumer/ConsumerIteratorTest.scala
> View Diff <https://reviews.apache.org/r/15805/diff/>