I apologize I'm a little late on chiming in on flight but I had some questions/comments that a quick search of the mailing list didn't seem to turn up anything and I didn't see comment on the initial pull request [1]
1. What is meant by "sidecar patterns" [2] on the data buffer bytes? 2. Was using "repeated bytes data_body" considered instead of a single body value? If we put back the "page id" in the buffer metadata (removed in [3]), it seems like we would get more flexibility for managing the data coming off the wire (e.g. if there was a buffer per repeated element, we could free memory immediately on column projections). 3. We might be able to simplify the implementation on the C++ side once Google open-sources more of there zero-copy API [3] "soon". (I suppose we might need to keep the low level implementation to deal with alignment though?) Thanks, Micah [1] https://github.com/apache/arrow/pull/2547 [2] https://github.com/apache/arrow/blob/master/format/Flight.proto#L289 [3] https://github.com/protocolbuffers/protobuf/issues/1896
