Is it worth having `flushTimeouts` in memory only to optimise cancellation (and one debug check in `register()`)?
Currently, this is used in two places: - in `register()` to check whether we have already added the reader - we could instead go through the list in `periodicFlushes` if we really wanted to - in `cancel()` to retrieve the right flush timeout so that we can easily remove the reader from `periodicFlushes` - we could iterate over it instead -> both alternative are a bit slower, but only happen during registration (if even) and cancellation and are therefore not performance critical [ Full content available at: https://github.com/apache/flink/pull/6698 ] This message was relayed via gitbox.apache.org for devnull@infra.apache.org
