PG19: migrate get_relation_info_hook to build_simple_rel_hook

[ihalatci]

What

Restores partitioned-index stripping on PostgreSQL 19, which regressed when PG19 removed the get_relation_info_hook that Citus relied on. This is a PG19-only planner-hook migration and a no-op on PG ≤ 18.

The columnar scan_analyze_next_tuple signature item from the original scope was already fixed in the build-foundation PR #8601 and is intentionally not touched here.

Why

PG19 dropped get_relation_info_hook in favour of build_simple_rel_hook. Citus used that hook (multi_get_relation_info_hook) to strip partitioned indexes from partitioned-table plans. PR #8601 disabled it on PG19 with a TODO(PG19 Phase 2) marker, so index-stripping was temporarily absent on PG19, causing a client-backend crash on partitioned-table planning paths inside the regression suite (make check).

How

• Add a PG19-only multi_build_simple_rel_hook in distributed_planner.c (prototype in distributed_planner.h) matching the build_simple_rel_hook signature (PlannerInfo *root, RelOptInfo *rel, RangeTblEntry *rte). It delegates straight to the existing multi_get_relation_info_hook body, which re-derives the RangeTblEntry from rel->relid — so the removed relationObjectId / inhparent arguments are passed as InvalidOid / false and intentionally ignored, keeping behaviour identical across versions.
• Register it via build_simple_rel_hook under #if PG_VERSION_NUM >= PG_VERSION_19 in shared_library_init.c; the old get_relation_info_hook assignment is kept under the #else arm for PG ≤ 18.

Files

• src/backend/distributed/planner/distributed_planner.c
• src/include/distributed/distributed_planner.h
• src/backend/distributed/shared_library_init.c

Item 1 (src/backend/columnar/columnar_tableam.c) is untouched — already fixed in #8601.

Scope note

An earlier revision of this PR also carried a distribution_column.c AccessShareLock change ("item 3"). That fix addresses a PG16+ relation_open() assertion / relcache-lifetime issue that is not PG19-specific, so it has been split out into a separate PR targeting main. This PR is now scoped solely to the PG19 build_simple_rel_hook migration.

Stacking

Stacked on #8619 — base branch is pg19-columnar-relinfo. GitHub will auto-retarget this PR to pg19-support once #8619 merges.

Cross-version gate

• PG19 (19beta1): full -Werror build green; partitioned_indexes_create (+ setup) regress passed, zero diff — exercises the partitioned-index planning path and proves the hook is active.
• PG18 (18.4): full -Werror build green; same focused regress passed, zero diff — the pre-PG19 #else arm is unchanged and neutral.
• PG17 (17.10): the migration is preprocessor-excluded (#if PG_VERSION_NUM >= PG_VERSION_19), so PG17 is structurally unaffected; the earlier full-scope gate confirmed a green PG17 -Werror build + regress.
• check-style: citus_indent --check clean on all touched files.

All arms build clean under full -Werror (no flag demotion). The change is behaviourally confined to PG19; PG ≤ 18 is preprocessor-excluded and proven regression-neutral.

Part of #8597. Relates to #8608.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.80%. Comparing base (f43b3b6) to head (6117eac).

Additional details and impacted files

@@                    Coverage Diff                    @@
##           pg19-columnar-relinfo    #8622      +/-   ##
=========================================================
- Coverage                  88.94%   88.80%   -0.15%     
=========================================================
  Files                        288      288              
  Lines                      64403    64406       +3     
  Branches                    8098     8099       +1     
=========================================================
- Hits                       57282    57193      -89     
- Misses                      4785     4873      +88     
- Partials                    2336     2340       +4     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[ihalatci]

Cross-version gate results ✅

All three majors build clean under full -Werror (no flag demotion):

Major	Build -Werror	Targeted regress
PG17 (17.10)	✅ green	✅ partitioned_indexes_create (+ setup): all passed, zero diff
PG18 (18.4)	✅ green	✅ partitioned_indexes_create (+ setup): all passed, zero diff
PG19 (19beta1)	✅ green	✅ create_distributed_table_concurrently passed, no crashes

Regression-neutrality (PG17/PG18): Item 2 is preprocessor-excluded on PG ≤ 18 (#if PG_VERSION_NUM >= PG_VERSION_19). Item 3 (the only unconditional change) is proven neutral by the green partitioned_indexes_create run on both PG17 and PG18 — that test drives create_distributed_table → BuildDistributionKeyFromColumnName, the exact lock-upgrade path.

PG19 targeted notes (against the known port-WIP baseline — diffs expected):

• create_distributed_table_concurrently — passed clean.
• multi_partitioning — 2 cosmetic baseline diffs only (psql \d+ partition-list formatting; upstream error-wording unique constraint → PRIMARY KEY constraint); no crash, ran to completion (positive signal item 2's hook is working).
• columnar_analyze — diff is a setup artifact of the standalone targeted schedule (columnar AM/tables not pre-created by this minimal schedule); not a code regression. Item 1 is untouched (already fixed in PG19: build foundation (source-level compat, scaffold ruleutils_19.c, CI image bump) #8601).

Caveat: the pgenv PG19 (19beta1) build is not --enable-cassert, so item 3's specific assertion (cassert-only relation_open() check) cannot be tripped here. The targeted run confirms no regression on that path; a cassert PG19 build would be needed to demonstrate the original crash → fix directly.

[ihalatci]

Item 3 — direct crash→fix evidence on a --enable-cassert PG19 build

The standard gate ran against a non-cassert PG19 (matching CI: the Citus exttester image installs the PGDG postgresql-19 package as its test server, which is not --enable-cassert). Item 3 guards a cassert-only Assert() in PG19's relation_open(), so neither CI nor the non-cassert gate can exercise it. To close that gap I built a local --enable-cassert --enable-debug PG19 (19beta1) and demonstrated the crash → fix directly.

Environment: PostgreSQL 19beta1, --enable-cassert --enable-debug (separate prefix; the gate's non-cassert 19beta1 left untouched). Identical citus source, only distribution_column.c toggled between base (NoLock) and the fix.

❌ BEFORE (item-3 hunk reverted — BuildDistributionKeyFromColumnName opens with NoLock)

make check focused schedule → select create_distributed_table_concurrently(...) aborts the backend:

TRAP: failed Assert("lockmode != NoLock || CheckRelationLockedByMe(r, AccessShareLock, true)"),
      File: "relation.c", Line: 119, PID: 36559
postgres ... SELECT(ExceptionalCondition+0x71)
postgres ... SELECT(try_relation_open+0x107)
.../citus.so(+0x131566)                                   <- BuildDistributionKeyFromColumnName
.../citus.so(create_distributed_table_concurrently+0x19f)
...
LOG:  client backend (PID 36559) was terminated by signal 6: Aborted
DETAIL:  Failed process was running: select create_distributed_table_concurrently('test','key', colocate_with := 'nocolo');

Result: not ok 4 - create_distributed_table_concurrently (backend killed by SIGABRT) → 1 of 4 tests failed.

Call path: create_distributed_table_concurrently → AddDistributionColumnForRelation(..., NoLock) (distribution_column_map.c:74) → BuildDistributionKeyFromColumnName(..., NoLock) → try_relation_open(rel, NoLock) → PG19 assert (relation.c:119).

✅ AFTER (item-3 fix applied — NoLock upgraded to AccessShareLock for the existence check)

Same cassert PG19, same schedule:

ok 4  - create_distributed_table_concurrently
# All 4 tests passed.

No TRAP, no Assert, no signal 6 anywhere in the server logs. The assertion is satisfied because the existence-check now holds AccessShareLock across relation_open/relation_close, while remaining a no-op for callers that already hold a lock (and on PG17/PG18, proven regression-neutral in the gate above).

This directly demonstrates the crash the fix prevents — the one piece that, by design, cannot show up in CI.