Hi all， In order to make hot compaction strategy more flexible, and split its code from flush part. I remove the vm-style compaction and add a hot compaction structure with the format of tsfile list manager. Now we have two strategies, normal strategy and level strategy, the normal strategy disables hot compaction. The level strategy enables size-tiering compaction. What do you think of the design above? And the following pr[1] add max chunk point limitation in level strategy and merge read limit. Can someone review and help merge it to master?
[1] https://github.com/apache/iotdb/pull/1758 LingZhe ZHANG School of Software, Tsinghua University
