Ashwin Agrawal

@adam8157 this PR has been sitting for many months... please can you work with the required stakeholders to drive closure for the PR and report on what's the plan for...

Do you wish to create a github issue then in-place to track the problem and close this PR if this is not intended solution?

Closing this PR for now given no movement forward on it. Once concrete plan and path to move forward is know re-open the PR.

Side notes while reviewing this PR (which should be tackled as separate PR) - GUC `gp_enable_slow_writer_testmode` should be replaced by fault injector mechanism to test - GUC `gp_instrument_shmem_size` still relevant?...

Thanks @QingMa-gp for starting this effort on GUCs, really important GPDB7 readiness task. Once this PR is wrapped up, next we should look into flags assigned for each and every...

@x4m please address the comments posted by Jimmy and resolve the conflicts reported for the PR, then we can go ahead to commit the same.

Curious to know what exact purpose this package or dependency is used?

> > Curious to know what exact purpose this package or dependency is used? > > @ashwinstar We do not use this package directly in our code, but seems like...

I think by design AO and specially current CO tables might not work with zero columns. So this case we should detect and ERROR out I believe.

Question (outside the scope of this PR): Do we need distributed snapshot for parse analyze phase?