Philip Dubé

Results 156 comments of Philip Dubé

You would need to implement retries on error. Likely best to do so with reconnection. If you need exactly once semantics things will get more complicated

We've looked at it (it's part of motivation for PG Type System), but it's not on roadmap

This is no longer in scope for PeerDB

this is driven by `binlog_row_metadata = full` in cdc. in qrep we always receive the string info

1. I've recommended a general `interface Tuple { Get(int) any; Len() int }` 2. another option would be to support `Seq[any]` from iter All these will generally involve boxing each...

I've misunderstood these changes changed to storage feature, not sure if you want it as a default feature

Fix would also help postgres avoid slot lag: start reading sync batches right away

seems like we should be running the sync loop as another `waitGroup.Go` so it can return syncErr to cancel the errgroup

this looks a bit wrong since it skips checking table mappings, but in this scenario I doubt table mappings are relevant