Thanks Konstantin, I have a minor comment on the wiki page's "Proposed Solution" section:
"The solution is to ignore that problem and it is a choice of the user to..": ---------- I'd suggest we rephrase it and admit "-1 is a special value in Kafka, which means we do not have a valid way to express Wednesday, December 31, 1969 11:59:59 PM UTC." rather than letting users to choose how to interpret it, since as mentioned, both at Kafka broker side, as well as at Kafka client side (Streams, Producer) we have already using this protocol to set -1 as a special value of `unknown`, so users choosing how to interpret it freely would lead to confusing behaviors. Instead, we should well document this caveat, and educate users in the document that, given this situation, if you do need to express Wednesday, December 31, 1969 11:59:59 PM UTC, consider shifting it by one millisecond. Otherwise, I'm +1 on this KIP. Guozhang On Wed, May 23, 2018 at 2:40 PM, Konstantin Chukhlomin <chuhlo...@gmail.com> wrote: > All, > > Thanks for the feedback on KIP-228. I've updated the KIP, and would like > to start to start a vote. > > KIP: https://cwiki.apache.org/confluence/display/KAFKA/KIP- > 228+Negative+record+timestamp+support <https://cwiki.apache.org/ > confluence/display/KAFKA/KIP-228+Negative+record+timestamp+support> > > Pull Request: https://github.com/apache/kafka/pull/5072/files < > https://github.com/apache/kafka/pull/5072/files> > > Discussion Thread: http://mail-archives.apache. > org/mod_mbox/kafka-dev/201712.mbox/%3c178138FA-CA30-44A8- > b92a-b966b8ca1...@gmail.com%3e <http://mail-archives.apache. > org/mod_mbox/kafka-dev/201712.mbox/%3C178138FA-CA30-44A8- > b92a-b966b8ca1...@gmail.com%3E> > > Thanks! -- -- Guozhang
