See https://reviews.apache.org/r/25016/, this is the consolidated patch for subtasks HBASE-11763 and HBASE-11764.
Is the approach reasonable? Implementation ok? Or objectionable parts? Or nits? I'd like to commit for 0.98.7, which is coming up next week. Otherwise I'll push it to .8 and I don't know how much bandwidth I'll have to carry the patch set going forward. I've been fixing up rejects when rebasing every couple of days (whenever we replace more uses of KeyValue with Cell on master). -- Best regards, - Andy Problems worthy of attack prove their worth by hitting back. - Piet Hein (via Tom White)