This prepares PostgreSQL for compatibility with Neon's storage.
Significant changes compared to the PostgreSQL 15 patchset include:

- Backported changes for users and roles are no longer required
- Use RM_NEON_ID for changes in WAL, instead of modifying core WAL records

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

* Neon logical replication support for PG16

* Log heap rewrite file after creation.

---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
Co-authored-by: Arseny Sher <sher-ars@yandex.ru>

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

* Update WAL buffers when restoring WAL at compute needed for LR

* Fix copying data in WAL buffers

---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

PG16 adds new function to SMGR: zeroextend
It's implementation in Neon actually wal-log zero pages of extended relation.
This zero page is wal-logged using XLOG_FPI.
As far as page is zero, the hole optimization (excluding from the image everything between pg_upper and pd_lower) doesn't work.

This PR allows to set hole size to BLCKSZ in case of zero page (PageIsNull() returns true).
---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

* Prevent output callbacks from hearing about neon-file messages

* Load SLRU segments on demand

refer #8763

* Fix errors in downloading SLRU segments

* Fix build problems

* Undo occcasional changes

* Remove unintenmded changes

* Fix smgr_read_slru_segment

* Determine SLRU kind in extension

* Use ctl->PagePrecedes for SLRU page comparison in SimpleLruDownloadSegment to address wraparround

---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

Fixes bug #18341. Backpatch to all supported versions.

Discussion: https://www.postgresql.org/message-id/18341-ce16599e7fd6228c@postgresql.org

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

…mary is not alive (#363)

* Set wasShutdown=true during hot-standby replica startup only when primary is not alive
* Report fatal error if hot standaby replica is started with oldestAcriveXid=0

Postgres part of neondatabase/neon#6705
---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

…d for oldestActiveXid while replica startup (#387)

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

This keeps the walproposer processes alive at shutdown, until after
the shutdown checkpoint has been written. That gives the walproposers
a chance to stream it to the safekeepers.

…on' error message (#390)

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

This is done in other Postgres headers.

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>

* Revert "Add comment explaining why it is safe to use FirstNormalTransactionXid for oldestActiveXid while replica startup (#387)"

This reverts commit 67240b5.

* Revert "Set wasShutdown=true during hot-standby replica startup only when primary is not alive (#363)"

This reverts commit 61c280c.

* Remember last written LSN when it is first requested

* Return updated LSN in SetLastWrittenLSN

---------

Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>