[Commitfest 2022-07] Patch Triage: Needs Review, Part 1 - Mailing list pgsql-hackers

From Jacob Champion
Subject [Commitfest 2022-07] Patch Triage: Needs Review, Part 1
Date
Msg-id 0ab66589-2f71-69b3-2002-49e821740b0d@timescale.com
Whole thread Raw
Responses Re: [Commitfest 2022-07] Patch Triage: Needs Review, Part 1
Re: [Commitfest 2022-07] Patch Triage: Needs Review, Part 1
List pgsql-hackers
Hi,

Next up is the large list of Needs Review. This part 1 should include
entries as old or older than seven commitfests running.

My heuristics for classifying these continue to evolve as I go, and
there's a lot to read, so please let me know if I've made any mistakes.

= Stalled Patches, Recommend Return =

These are stalled and I recommend outright that we return them. We don't
have a separate status for "needs more interest" (working on a patch) so
I'd just RwF, with a note explaining that what is actually needed to
continue isn't more code work but more coalition building.

- Implement INSERT SET syntax
  https://commitfest.postgresql.org/38/2218/

A recent author rebase this CF, but unfortunately I think the the real
issue is just a lack of review interest. It's been suggested for return
for a few CFs now.

- Fix up partitionwise join on how equi-join conditions between the
partition keys are identified
  https://commitfest.postgresql.org/38/2266/

It looks like this one was Returned with Feedback but did not actually
have feedback, which may have caused confusion. (Solid motivation for a
new close status.) I don't think there's been any review since 2020.

- New default role allowing to change per-role/database settings
  https://commitfest.postgresql.org/38/2918/

Stalled on review in January, and needs a rebase.

= Stalled Patches, Need Help =

These are stalled but seem to have interest. They need help to either
get them out of the rut, or else be Returned so that the author can try
a different approach instead of perma-rebasing. I plan to move them to
the next CF unless someone speaks up to say otherwise.

- Show shared filesets in pg_ls_tmpdir (pg_ls_* functions for showing
metadata and recurse)
  https://commitfest.postgresql.org/38/2377/

From a quick skim it looks like there was a flurry of initial positive
feedback followed by a stall and then some design whiplash. This thread
needs help to avoid burnout, I think.

- Make message at end-of-recovery less scary
  https://commitfest.postgresql.org/38/2490/

This got marked RfC twice, fell back out, and has been stuck in a rebase
loop.

- Fix behavior of geo_ops when NaN is involved
  https://commitfest.postgresql.org/38/2710/

Stuck in a half-committed state, which is tricky. Could maybe use a
reframing or recap (or a new thread?).

- Add extra statistics to explain for Nested Loop
  https://commitfest.postgresql.org/38/2765/

I think the author is hoping for help with testing and performance
characterization.

- CREATE INDEX CONCURRENTLY on partitioned table
  https://commitfest.postgresql.org/38/2815/

This had an author switch since last CF, so I think it'd be
inappropriate to close it out this time around, but it needs assistance.

- New Table Access Methods for Multi and Single Inserts
  https://commitfest.postgresql.org/38/2871/

Although there was a brief flicker in March, I think this one has
stalled out and is just about ready to be returned.

- Fix pg_rewind race condition just after promotion
  https://commitfest.postgresql.org/38/2864/

Seems like an important fix, but it's silent? Does it need to be
promoted to an Open Issue?

- pg_stat_statements and "IN" conditions
  https://commitfest.postgresql.org/38/2837/

Some good, recent interest. Last review in March.

- Function to log backtrace of postgres processes
  https://commitfest.postgresql.org/38/2863/

This is just starting to stall; I think it needs some assistance.

- Allow batched insert during cross-partition updates
  https://commitfest.postgresql.org/38/2992/

Was RfC (twice), then dropped out, now it's stuck rebasing. Last
substantial review in 2021.

= Active Patches =

The following are actively being worked and I expect to move them to
next CF:

- session variables, LET command
- Remove self join on a unique column
- Incremental Materialized View Maintenance
- More scalable multixacts buffers and locking
- Fast COPY FROM command for the foreign tables
- Extended statistics / estimate Var op Var clauses

Thanks,
--Jacob



pgsql-hackers by date:

Previous
From: "David G. Johnston"
Date:
Subject: Re: pg_auth_members.grantor is bunk
Next
From: Nathan Bossart
Date:
Subject: Re: optimize lookups in snapshot [sub]xip arrays