Обсуждение: [Commitfest 2022-07] Patch Triage: Needs Review, Part 2

Поиск
Список
Период
Сортировка

[Commitfest 2022-07] Patch Triage: Needs Review, Part 2

От
Jacob Champion
Дата:
Hello,

Part 2 should include entries four commitfests and older. (For the rest,
it's probably too early to call something "stalled", so I don't plan to
do any more triage there.) Patch authors CC'd.

= Stalled Patches, Recommend Return =

I plan to return these with a note saying "needs more interest".

- Extended statistics in EXPLAIN
  https://commitfest.postgresql.org/38/3050/

  There's interest, but it seemed controversial. And a reviewer
  attempted to revive this in January, but there hasn't been new
  engagement or response from the original people involved in a year.

- Map WAL segment files on PMEM as WAL buffers
  https://commitfest.postgresql.org/38/3181/

  Stalled out; last review was in January and it needs a rebase.

- Support pg_ident mapping for LDAP
  https://commitfest.postgresql.org/38/3314/

  This one's mine; I think it's clear that there's not enough interest
  in this idea yet and I think I'd rather put effort into SASL, which
  would ideally do the same thing natively.

- Improve logging when using Huge Pages
  https://commitfest.postgresql.org/38/3310/

  There was interest last year but no mails this year, so this probably
  needs some buy-in first.

- functions to compute size of schemas/AMs (and maybe \dn++ and \dA++)
  https://commitfest.postgresql.org/38/3256/

  Been rebasing without review since September 2021. Seems like there's
  an idea in here that people want, though. Any sponsors for a future
  CF?

- Upgrade pgcrypto to crypt_blowfish 1.3
  https://commitfest.postgresql.org/38/3338/

  No conversation on this since October 2021. Like above, seems like a
  reasonable feature, so maybe someone can sponsor it and quickly get it
  resurrected in a future CF?

= Stalled Patches, Need Help =

I plan to move these forward unless someone says otherwise, but they
look stuck to me and need assistance.

- pgbench: add multiconnect support
  https://commitfest.postgresql.org/38/3227/

  Seems to be interest. Not much review, though.

- pg_stats and range statistics
  https://commitfest.postgresql.org/38/3184/

  There was immediate agreement that this feature was desirable, and
  then the reviews dried up. Anyone want to bump this?

- Asymmetric partition-wise JOIN
  https://commitfest.postgresql.org/38/3099/

  There was a rebase in January by a reviewer, so there's definite
  interest, but work hasn't progressed in a while. I've marked Waiting
  on Author in the meantime.

- Logging plan of the currently running query
  https://commitfest.postgresql.org/38/3142/

  Last review in Febrary and currently in a rebase loop.

- schema change not getting invalidated, both renamed table and new
  table data were getting replicated
  https://commitfest.postgresql.org/38/3262/

  This looks like a bug fix that should not be closed out, but it's been
  in a rebase loop without review for... a year? Any takers? Should we
  make an open issue?

- pgbench: using prepared BEGIN statement in a pipeline could cause an
  error
  https://commitfest.postgresql.org/38/3260/

  A bug fix, but maybe the approach taken for the fix is controversial?

- Atomic rename feature for Windows
  https://commitfest.postgresql.org/38/3347/

  I think this got derailed by a committer conversation about platform
  deprecation? I have no idea where the patch stands after that
  exchange; can someone recap?

= Active Patches =

These will be moved ahead:

- Lazy JIT IR code generation to increase JIT speed with partitions
- Logical replication failure "ERROR: could not map filenode
  "base/13237/442428" to relation OID" with catalog modifying txns
- Add proper planner support for ORDER BY / DISTINCT aggregates
- Fix ExecRTCheckPerms() inefficiency with many prunable partitions
- Using each rel as both outer and inner for anti-joins
- Postgres picks suboptimal index after building extended statistics
- Cache tuple routing info during bulk loads into partitioned tables
- postgres_fdw: commit remote (sub)transactions in parallel during
  pre-commit
- add checkpoint stats of snapshot and mapping files of pg_logical dir
- Allows database-specific role memberships

Thanks,
--Jacob