Maximilian Hoffman

related: - https://github.com/dolthub/go-mysql-server/pull/1875 - https://github.com/dolthub/go-mysql-server/pull/1874

@jycor How are we feeling about this? Can we make more specific statements about certain queries that fail related to this?

It would be helpful to supplement this description with a specific query that fails and with expected/current behavior.

#benchmark

So I guess the main question is, why did we ever do this? I can eliminate most of the addresses we're writing in a transaction (haven't done all, but current...

I'm seeing the GC and dangling ref implications of not checking child references. Can we bisect the addresses into new chunks, and child addresses that need reference checks?

> Definitely run this by @reltuk. Seems like a promising avenue for optimization even if this might be a bit too much. The TLDR from @reltuk is that this ref...

Closing for now, adding GC gen watermarks to sessions isn't my first choice for TPC-C throughput at the moment

I may be overcomplicating this, but I think the transformation needs to happen between building the memo and creating the join reorder graph. This is because we need the following...

#benchmark