ijuma commented on pull request #10123:
URL: https://github.com/apache/kafka/pull/10123#issuecomment-780326554
I'm not sure it's worth it since the cost of compressing on the broker
during fetches is significantly higher than compressing during produce (the
data is already on the heap for
ijuma commented on pull request #10123:
URL: https://github.com/apache/kafka/pull/10123#issuecomment-778799711
Merged to trunk and 2.8.
This is an automated message from the Apache Git Service.
To respond to the message,
ijuma commented on pull request #10123:
URL: https://github.com/apache/kafka/pull/10123#issuecomment-778799315
JDK 15 build passed, JDK 8 had unrelated flaky failures and JDK 11 died
abruptly.
This is an automated message
ijuma commented on pull request #10123:
URL: https://github.com/apache/kafka/pull/10123#issuecomment-778797029
@chia7712 with these changes, older Android clients should be able to start
(before they would fail when the method handles code was reached even if no
compression was actually
ijuma commented on pull request #10123:
URL: https://github.com/apache/kafka/pull/10123#issuecomment-77877
@chia7712 Both solutions should behave the same at runtime. At compile time,
the previous solution did not require snappy libraries in the classpath (but it
did require lz4 and
ijuma commented on pull request #10123:
URL: https://github.com/apache/kafka/pull/10123#issuecomment-778637526
@chia7712 Thoughts on this change?
This is an automated message from the Apache Git Service.
To respond to the
