reshke

> Hi, please provide cases to verify what you have done. Hi, you mean some regression test-cases? They are as obvious as ``` create table t1(i int); create materialized view...

Seems that matview_optimizer.out depends on matview tests. Pushed commit should fix test

Are there any open items/suggestion/issues?

I can actually do it myself, like cherry-pick series of patches. However, these commit depends on out-of $subj scope work being merged in GPDB main. So, many conflict with applying...

> We welcome contributions from community, if you are interested, please don't hesitate to file PR following [coding contribution guide](https://cloudberrydb.org/contribute/code) HI! How we can observe state of cherry-pick project, progress...

The real question is how you expect this to work. There is no reliable way to determine if query executing would modify data or not. You cannot simply say that...

So you basically suggest query parsing and analyze inside Odyssey? Thats an huge change. We have done it in SPQR and we actually know what to do, but this requires...

Just cherry-picking single commit. It ain't much, but it's honest work

Maybe we should configure CI with more strict compiler flags... LGTM #986

Upstream does this different way: https://github.com/greenplum-db/gpdb-archive/blob/main/src/backend/access/aocs/aocsam_handler.c#L2335 looks like we need to cherry-pick block directory support and restore this logic https://github.com/greenplum-db/gpdb-archive/blob/main/src/backend/access/aocs/aocsam_handler.c#L2196 i will take another look soon