Postgres 14.12, 15.7, and 16.3 updates

[tristan957]

No description provided.

[tristan957]

@hlinnaka I didn't do the merge strategy because I think we should unsquash the v15 and v16 branches, though it seemed to work when I tried it out.

[github-actions]

3102 tests run: 2975 passed, 0 failed, 127 skipped (full report)

---

Flaky tests (5)

Postgres 16

• test_secondary_background_downloads: debug

Postgres 15

• test_hot_standby: debug
• test_timeline_delete_resumed_on_attach: release
• test_vm_bit_clear_on_heap_lock: debug

Postgres 14

• test_secondary_background_downloads: debug

Code coverage* (full report)

• functions: 31.3% (6383 of 20404 functions)
• lines: 47.7% (48639 of 101920 lines)

* collected from Rust tests only

---

_{The comment gets automatically updated with the latest test results
98aba0a at 2024-05-20T16:41:33.682Z :recycle:}

[tristan957]

FYI: neondatabase/postgres@3be8940

[hlinnaka]

Per the discussion at #7688 (comment), let's:

• keep the test for an unlogged index, and
• add an explicit test for an unlogged sequence.

With those changes, LGTM.

PS. Thanks for the very thorough comment on the history of unlogged sequences, that's very helpful!

[hlinnaka]

Per the discussion at #7688 (comment), let's:

• keep the test for an unlogged index, and

Tristan pointed on slack that that has been addressed.

• add an explicit test for an unlogged sequence.

This would still be nice for clarity, but not essential. The test does cover unlogged sequences as it is