path: root/doc/src/sgml/release-15.sgml

<!-- doc/src/sgml/release-15.sgml -->
<!-- See header comment in release.sgml about typical markup -->

 <sect1 id="release-15-7">
  <title>Release 15.7</title>

  <formalpara>
  <title>Release date:</title>
  <para>2024-05-09</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.6.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.7</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, a security vulnerability was found in the system
    views <structname>pg_stats_ext</structname>
    and <structname>pg_stats_ext_exprs</structname>, potentially allowing
    authenticated database users to see data they shouldn't.  If this is
    of concern in your installation, follow the steps in the first
    changelog entry below to rectify it.
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 15.6,
    see <xref linkend="release-15-6"/>.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Nathan Bossart <nathan@postgresql.org>
Branch: master [521a7156a] 2024-05-06 09:00:00 -0500
Branch: REL_16_STABLE [2485a85e9] 2024-05-06 09:00:07 -0500
Branch: REL_15_STABLE [9cc2b6289] 2024-05-06 09:00:13 -0500
Branch: REL_14_STABLE [c3425383b] 2024-05-06 09:00:19 -0500
-->
     <para>
      Restrict visibility of <structname>pg_stats_ext</structname> and
      <structname>pg_stats_ext_exprs</structname> entries to the table
      owner (Nathan Bossart)
     </para>

     <para>
      These views failed to hide statistics for expressions that involve
      columns the accessing user does not have permission to read.  View
      columns such as <structfield>most_common_vals</structfield> might
      expose security-relevant data.  The potential interactions here are
      not fully clear, so in the interest of erring on the side of safety,
      make rows in these views visible only to the owner of the associated
      table.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks
      Lukas Fittl for reporting this problem.
      (CVE-2024-4317)
     </para>

     <para>
      By itself, this fix will only fix the behavior in newly initdb'd
      database clusters.  If you wish to apply this change in an existing
      cluster, you will need to do the following:
     </para>

     <procedure>
      <step>
       <para>
        Find the SQL script <filename>fix-CVE-2024-4317.sql</filename> in
        the <replaceable>share</replaceable> directory of
        the <productname>PostgreSQL</productname> installation (typically
        located someplace like <filename>/usr/share/postgresql/</filename>).
        Be sure to use the script appropriate to
        your <productname>PostgreSQL</productname> major version.
        If you do not see this file, either your version is not vulnerable
        (only v14&ndash;v16 are affected) or your minor version is too
        old to have the fix.
       </para>
      </step>

      <step>
       <para>
        In <emphasis>each</emphasis> database of the cluster, run
        the <filename>fix-CVE-2024-4317.sql</filename> script as superuser.
        In <application>psql</application> this would look like
<programlisting>
\i /usr/share/postgresql/fix-CVE-2024-4317.sql
</programlisting>
        (adjust the file path as appropriate).  Any error probably indicates
        that you've used the wrong script version.  It will not hurt to run
        the script more than once.
       </para>
      </step>

      <step>
       <para>
        Do not forget to include the <literal>template0</literal>
        and <literal>template1</literal> databases, or the vulnerability
        will still exist in databases you create later.  To
        fix <literal>template0</literal>, you'll need to temporarily make
        it accept connections.  Do that with
<programlisting>
ALTER DATABASE template0 WITH ALLOW_CONNECTIONS true;
</programlisting>
        and then after fixing <literal>template0</literal>, undo it with
<programlisting>
ALTER DATABASE template0 WITH ALLOW_CONNECTIONS false;
</programlisting>
       </para>
      </step>
     </procedure>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [b4a71cf65] 2024-03-14 14:57:16 -0400
Branch: REL_16_STABLE [52898c63e] 2024-03-14 14:57:16 -0400
Branch: REL_15_STABLE [7c61d2342] 2024-03-14 14:57:16 -0400
Branch: REL_14_STABLE [3621ffd9f] 2024-03-14 14:57:16 -0400
Branch: REL_13_STABLE [0200398dd] 2024-03-14 14:57:16 -0400
Branch: REL_12_STABLE [82c87af7a] 2024-03-14 14:57:16 -0400
-->
     <para>
      Fix <command>INSERT</command> from
      multiple <command>VALUES</command> rows into a target column that is
      a domain over an array or composite type (Tom Lane)
     </para>

     <para>
      Such cases would either fail with surprising complaints about
      mismatched datatypes, or insert unexpected coercions that could lead
      to odd results.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [4989ce726] 2024-02-21 17:18:52 +0100
Branch: REL_16_STABLE [a3f5d2056] 2024-02-21 17:18:52 +0100
Branch: REL_15_STABLE [90ad85db6] 2024-02-21 17:18:52 +0100
-->
     <para>
      Require <literal>SELECT</literal> privilege on the target table
      for <command>MERGE</command> with a <literal>DO NOTHING</literal>
      clause (&Aacute;lvaro Herrera)
     </para>

     <para>
      <literal>SELECT</literal> privilege would be required in all
      practical cases anyway, but require it even if the query reads no
      columns of the target table.  This avoids an edge case in
      which <command>MERGE</command> would require no privileges whatever,
      which seems undesirable even when it's a do-nothing command.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [29ef1dd19] 2024-03-07 09:57:02 +0000
Branch: REL_16_STABLE [dd73d10ad] 2024-03-07 09:55:39 +0000
Branch: REL_15_STABLE [b5c645d2a] 2024-03-07 09:53:31 +0000
-->
     <para>
      Fix handling of self-modified tuples in <command>MERGE</command>
      (Dean Rasheed)
     </para>

     <para>
      Throw an error if a target row joins to more than one source row, as
      required by the SQL standard.  (The previous coding could silently
      ignore this condition if a concurrent update was involved.)  Also,
      throw a non-misleading error if a target row is already updated by a
      later command in the current transaction, thanks to
      a <literal>BEFORE</literal> trigger or a volatile function used in
      the query.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [4c2369ac5] 2024-02-20 12:49:37 +1300
Branch: REL_16_STABLE [fb95cc72b] 2024-02-20 12:50:09 +1300
Branch: REL_15_STABLE [1b3495e29] 2024-02-20 12:50:34 +1300
Branch: REL_14_STABLE [f9c8f7ccd] 2024-02-20 12:50:57 +1300
Branch: REL_13_STABLE [3850fcca6] 2024-02-20 12:51:17 +1300
Branch: REL_12_STABLE [3ffcd24c2] 2024-02-20 12:51:38 +1300
-->
     <para>
      Fix incorrect pruning of NULL partition when a table is partitioned
      on a boolean column and the query has a boolean <literal>IS
      NOT</literal> clause (David Rowley)
     </para>

     <para>
      A NULL value satisfies a clause such
      as <literal><replaceable>boolcol</replaceable> IS NOT
      FALSE</literal>, so pruning away a partition containing NULLs
      yielded incorrect answers.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [fad3b5b5a] 2024-03-26 15:28:31 -0400
Branch: REL_16_STABLE [7445f0928] 2024-03-26 15:28:16 -0400
Branch: REL_15_STABLE [b48eda4e5] 2024-03-26 15:28:16 -0400
Branch: REL_14_STABLE [66bbad581] 2024-03-26 15:28:16 -0400
Branch: REL_13_STABLE [97de2a159] 2024-03-26 15:28:16 -0400
Branch: REL_12_STABLE [a8b740868] 2024-03-26 15:28:16 -0400
-->
     <para>
      Make <command>ALTER FOREIGN TABLE SET SCHEMA</command> move any
      owned sequences into the new schema (Tom Lane)
     </para>

     <para>
      Moving a regular table to a new schema causes any sequences owned by
      the table to be moved to that schema too (along with indexes and
      constraints).  This was overlooked for foreign tables, however.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [6743c5ae6] 2024-02-09 08:09:22 +0100
Branch: REL_16_STABLE [86d2b434c] 2024-02-09 08:09:59 +0100
Branch: REL_15_STABLE [d17a3a4c6] 2024-02-09 08:15:27 +0100
-->
     <para>
      Make <command>ALTER TABLE ... ADD COLUMN</command> create
      identity/serial sequences with the same persistence as their owning
      tables (Peter Eisentraut)
     </para>

     <para>
      <command>CREATE UNLOGGED TABLE</command> will make any owned
      sequences be unlogged too.  <command>ALTER TABLE</command> missed
      that consideration, so that an added identity column would have a
      logged sequence, which seems pointless.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [91e7115b1] 2024-05-02 17:36:31 -0400
Branch: REL_16_STABLE [11d40a44b] 2024-05-02 17:36:31 -0400
Branch: REL_15_STABLE [5f4a1a0a7] 2024-05-02 17:36:31 -0400
Branch: master [42b041243] 2024-04-28 14:34:21 -0400
Branch: REL_16_STABLE [b19255ca6] 2024-04-28 14:34:21 -0400
Branch: REL_15_STABLE [9b41d1d63] 2024-04-28 14:34:21 -0400
Branch: REL_14_STABLE [617a23927] 2024-04-28 14:34:21 -0400
-->
     <para>
      Improve <command>ALTER TABLE ... ALTER COLUMN TYPE</command>'s error
      message when there is a dependent function or publication (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [8c239ee15] 2024-04-21 21:21:26 +0200
Branch: REL_16_STABLE [9e6faeb32] 2024-04-21 21:21:55 +0200
Branch: REL_15_STABLE [276b7888f] 2024-04-21 21:22:11 +0200
-->
     <para>
      In <command>CREATE DATABASE</command>, recognize strategy keywords
      case-insensitively for consistency with other options (Tomas Vondra)
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [0960ae196] 2024-03-18 14:03:58 +0200
Branch: REL_16_STABLE [1f4eb7342] 2024-03-18 14:04:13 +0200
Branch: REL_15_STABLE [d3d95f583] 2024-03-18 14:04:17 +0200
Branch: REL_14_STABLE [262757b73] 2024-03-18 14:04:19 +0200
Branch: REL_13_STABLE [992189a3e] 2024-03-18 14:04:24 +0200
Branch: REL_12_STABLE [f3e4581ac] 2024-03-18 14:04:28 +0200
-->
     <para>
      Fix <command>EXPLAIN</command>'s counting of heap pages accessed by
      a bitmap heap scan (Melanie Plageman)
     </para>

     <para>
      Previously, heap pages that contain no visible tuples were not
      counted; but it seems more consistent to count all pages returned by
      the bitmap index scan.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [33e729c51] 2024-03-17 10:17:11 +0000
Branch: REL_16_STABLE [34c854b93] 2024-03-17 10:19:31 +0000
Branch: REL_15_STABLE [89ee14a2f] 2024-03-17 10:20:20 +0000
-->
     <para>
      Fix <command>EXPLAIN</command>'s output for subplans
      in <command>MERGE</command> (Dean Rasheed)
     </para>

     <para>
      <command>EXPLAIN</command> would sometimes fail to properly display
      subplan Params referencing variables in other parts of the plan tree.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [6faca9ae2] 2024-04-02 14:59:32 -0400
Branch: REL_16_STABLE [cbfbb14bd] 2024-04-02 14:59:04 -0400
Branch: REL_15_STABLE [4fb56a734] 2024-04-02 14:59:04 -0400
Branch: REL_14_STABLE [ca392df8d] 2024-04-02 14:59:04 -0400
Branch: REL_13_STABLE [4afc2c219] 2024-04-02 14:59:04 -0400
Branch: REL_12_STABLE [f5d9212e5] 2024-04-02 14:59:04 -0400
-->
     <para>
      Avoid deadlock during removal of orphaned temporary tables
      (Mikhail Zhilin)
     </para>

     <para>
      If the session that creates a temporary table crashes without
      removing the table, autovacuum will eventually try to remove the
      orphaned table.  However, an incoming session that's been assigned
      the same temporary namespace will do that too.  If a temporary table
      has a dependency (such as an owned sequence) then a deadlock could
      result between these two cleanup attempts.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [f65ab862e] 2024-04-29 10:24:56 -0700
Branch: REL_16_STABLE [92685c389] 2024-04-29 10:24:59 -0700
Branch: REL_15_STABLE [7c5915c4b] 2024-04-29 10:24:59 -0700
Branch: REL_14_STABLE [2ca19aa81] 2024-04-29 10:25:00 -0700
Branch: REL_13_STABLE [70cadfba0] 2024-04-29 10:25:00 -0700
Branch: REL_12_STABLE [f222349c4] 2024-04-29 10:25:00 -0700
-->
     <para>
      Avoid race condition while examining per-relation frozen-XID values
      (Noah Misch)
     </para>

     <para>
      <command>VACUUM</command>'s computation of per-database frozen-XID
      values from per-relation values could get confused by a concurrent
      update of those values by another <command>VACUUM</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Masahiko Sawada <msawada@postgresql.org>
Branch: master [5cd72cc0c] 2024-05-01 12:34:06 +0900
Branch: REL_16_STABLE [f199436c1] 2024-05-01 12:34:04 +0900
Branch: REL_15_STABLE [faba2f8f3] 2024-05-01 12:34:01 +0900
-->
     <para>
      Fix buffer usage reporting for parallel vacuuming (Anthonin Bonnefoy)
     </para>

     <para>
      Buffer accesses performed by parallel workers were not getting
      counted in the statistics reported in <literal>VERBOSE</literal>
      mode.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: REL_15_STABLE [bf379b555] 2024-04-30 15:22:55 -0400
Branch: REL_14_STABLE [51189f98a] 2024-04-30 15:22:55 -0400
Branch: REL_13_STABLE [1ee22d1e8] 2024-04-30 15:22:55 -0400
Branch: REL_12_STABLE [56d30fb10] 2024-04-30 15:22:56 -0400
-->
     <para>
      Disallow converting a table to a view within an outer SQL command
      that is using that table (Tom Lane)
     </para>

     <para>
      This avoids possible crashes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [03107b4ed] 2024-04-16 11:22:51 -0400
Branch: REL_16_STABLE [91800af13] 2024-04-16 11:03:43 -0400
Branch: REL_15_STABLE [5aacfa64e] 2024-04-16 11:22:39 -0400
Branch: REL_14_STABLE [ab2402268] 2024-04-16 11:22:39 -0400
Branch: REL_13_STABLE [d9e4ee74f] 2024-04-16 11:22:39 -0400
Branch: REL_12_STABLE [f502849d4] 2024-04-16 11:22:39 -0400
-->
     <para>
      Ensure that join conditions generated from equivalence classes are
      applied at the correct plan level (Tom Lane)
     </para>

     <para>
      In versions before <productname>PostgreSQL</productname> 16, it was
      possible for generated conditions to be evaluated below outer joins
      when they should be evaluated above (after) the outer join, leading
      to incorrect query results.  All versions have a similar hazard when
      considering joins to <command>UNION ALL</command> trees that have
      constant outputs for the join column in
      some <command>SELECT </command> arms.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: REL_16_STABLE [9d36b883b] 2024-05-01 16:35:05 +1200
Branch: REL_15_STABLE [7e5d20bbd] 2024-05-01 16:35:37 +1200
-->
     <para>
      Prevent potentially-incorrect optimization of some window functions
      (David Rowley)
     </para>

     <para>
      Disable <quote>run condition</quote> optimization
      of <function>ntile()</function> and <function>count()</function>
      with non-constant arguments.  This avoids possible misbehavior with
      sub-selects, typically leading to errors like <quote>WindowFunc not
      found in subplan target lists</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [a767cdc84] 2024-03-27 13:39:03 -0400
Branch: REL_16_STABLE [a94f51a7b] 2024-03-27 13:39:03 -0400
Branch: REL_15_STABLE [03561a6c7] 2024-03-27 13:39:03 -0400
Branch: REL_14_STABLE [0d30e48c2] 2024-03-27 13:39:03 -0400
Branch: REL_13_STABLE [de3c5b187] 2024-03-27 13:39:03 -0400
Branch: REL_12_STABLE [25675c474] 2024-03-27 13:39:03 -0400
-->
     <para>
      Avoid unnecessary use of moving-aggregate mode with a non-moving
      window frame (Vallimaharajan G)
     </para>

     <para>
      When a plain aggregate is used as a window function, and the window
      frame start is specified as <literal>UNBOUNDED PRECEDING</literal>,
      the frame's head cannot move so we do not need to use the special
      (and more expensive) moving-aggregate mode.  This optimization was
      intended all along, but due to a coding error it never triggered.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [a6b2a51e1] 2024-02-23 15:21:53 -0500
Branch: REL_16_STABLE [ef0333e67] 2024-02-23 15:21:53 -0500
Branch: REL_15_STABLE [37bbe3d3a] 2024-02-23 15:21:53 -0500
Branch: REL_14_STABLE [cbeb45527] 2024-02-23 15:21:53 -0500
Branch: REL_13_STABLE [9061fd23c] 2024-02-23 15:21:53 -0500
Branch: REL_12_STABLE [cf807eba5] 2024-02-23 15:21:53 -0500
-->
     <para>
      Avoid use of already-freed data while planning partition-wise joins
      under GEQO (Tom Lane)
     </para>

     <para>
      This would typically end in a crash or unexpected error message.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [e62984647] 2024-03-11 18:19:56 +1300
Branch: REL_16_STABLE [348233cb1] 2024-03-11 18:20:39 +1300
Branch: REL_15_STABLE [74530804f] 2024-03-11 18:21:23 +1300
Branch: REL_14_STABLE [72b8507db] 2024-03-11 18:21:48 +1300
-->
     <para>
      Avoid freeing still-in-use data in Memoize (Tender Wang, Andrei
      Lepikhov)
     </para>

     <para>
      In production builds this error frequently didn't cause any
      problems, as the freed data would most likely not get overwritten
      before it was used.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [a37a3e2b3] 2024-03-05 16:17:02 +1300
Branch: REL_16_STABLE [ac7e6a01c] 2024-03-05 16:17:53 +1300
Branch: REL_15_STABLE [164fe7a6e] 2024-03-05 16:18:19 +1300
Branch: REL_14_STABLE [0c2dda109] 2024-03-05 16:18:42 +1300
Branch: REL_13_STABLE [421dfb41a] 2024-03-05 16:19:05 +1300
Branch: REL_12_STABLE [94246405d] 2024-03-05 16:19:26 +1300
-->
     <para>
      Fix incorrectly-reported statistics kind codes in <quote>requested
      statistics kind <replaceable>X</replaceable> is not yet
      built</quote> error messages (David Rowley)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [e0df80828] 2024-04-15 12:56:56 -0400
Branch: REL_16_STABLE [cc1eb6a3c] 2024-04-15 12:56:56 -0400
Branch: REL_15_STABLE [09989ba84] 2024-04-15 12:56:56 -0400
Branch: REL_14_STABLE [78e81e14d] 2024-04-15 12:56:56 -0400
Branch: REL_13_STABLE [b6e21cef7] 2024-04-15 12:56:56 -0400
Branch: REL_12_STABLE [e0970862e] 2024-04-15 12:56:56 -0400
Branch: master [2ed8f9a01] 2024-03-06 14:41:13 -0500
Branch: REL_16_STABLE [1b3029be5] 2024-03-06 14:41:13 -0500
Branch: REL_15_STABLE [3b671dcf5] 2024-03-06 14:41:13 -0500
Branch: REL_14_STABLE [a595c3075] 2024-03-06 14:41:13 -0500
Branch: REL_13_STABLE [d769f9d97] 2024-03-06 14:41:13 -0500
Branch: REL_12_STABLE [466376c9f] 2024-03-06 14:41:13 -0500
-->
     <para>
      Be more careful with <type>RECORD</type>-returning functions
      in <literal>FROM</literal> (Tom Lane)
     </para>

     <para>
      The output columns of such a function call must be defined by
      an <literal>AS</literal> clause that specifies the column names and
      data types.  If the actual function output value doesn't match that,
      an error is supposed to be thrown at runtime.  However, some code
      paths would examine the actual value prematurely, and potentially
      issue strange errors or suffer assertion failures if it doesn't
      match expectations.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [6ee3261e9] 2024-03-12 18:16:25 -0400
Branch: REL_16_STABLE [40d1bdeb7] 2024-03-12 18:16:10 -0400
Branch: REL_15_STABLE [6f66fadad] 2024-03-12 18:16:10 -0400
Branch: REL_14_STABLE [649bbba11] 2024-03-12 18:16:10 -0400
Branch: REL_13_STABLE [28184f039] 2024-03-12 18:16:10 -0400
Branch: REL_12_STABLE [dc1503d5b] 2024-03-12 18:16:10 -0400
-->
     <para>
      Fix confusion about the return rowtype of SQL-language procedures
      (Tom Lane)
     </para>

     <para>
      A procedure implemented in SQL language that returns a single
      composite-type column would cause an assertion failure or core dump.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [d57b7cc33] 2024-02-16 16:02:00 +0200
Branch: master [75bcba6cb] 2024-02-21 02:51:41 +0200
Branch: REL_16_STABLE [760767182] 2024-03-11 03:06:55 +0200
Branch: REL_15_STABLE [84788ee5b] 2024-03-11 03:06:45 +0200
Branch: REL_14_STABLE [84cc1a552] 2024-03-11 03:06:15 +0200
Branch: REL_13_STABLE [445c7e38f] 2024-03-11 03:06:10 +0200
Branch: REL_12_STABLE [98bfb7558] 2024-03-11 02:53:07 +0200
-->
     <para>
      Add protective stack depth checks to some recursive functions
      (Egor Chindyaskin)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d163fdbfe] 2024-02-28 14:00:30 -0500
Branch: REL_16_STABLE [17db5436e] 2024-02-28 14:00:30 -0500
Branch: REL_15_STABLE [db8855b66] 2024-02-28 14:00:30 -0500
Branch: REL_14_STABLE [fe3b1b575] 2024-02-28 14:00:30 -0500
-->
     <para>
      Fix mis-rounding and overflow hazards
      in <function>date_bin()</function> (Moaaz Assali)
     </para>

     <para>
      In the case where the source timestamp is before the origin
      timestamp and their difference is already an exact multiple of the
      stride, the code incorrectly subtracted the stride anyway.  Also,
      detect some integer-overflow cases that would have produced
      incorrect results.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [4019285c0] 2024-04-28 13:42:13 -0400
Branch: REL_16_STABLE [3752e3d21] 2024-04-28 13:42:13 -0400
Branch: REL_15_STABLE [e6e3ee5b7] 2024-04-28 13:42:13 -0400
Branch: REL_14_STABLE [1748379b6] 2024-04-28 13:42:13 -0400
Branch: REL_13_STABLE [440b6251b] 2024-04-28 13:42:13 -0400
Branch: REL_12_STABLE [cb0ccefa0] 2024-04-28 13:42:13 -0400
-->
     <para>
      Detect integer overflow when adding or subtracting
      an <type>interval</type> to/from a <type>timestamp</type>
      (Joseph Koshakow)
     </para>

     <para>
      Some cases that should cause an out-of-range error produced an
      incorrect result instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [ce571434a] 2024-02-09 12:29:41 -0500
Branch: REL_16_STABLE [4eb261165] 2024-02-09 12:29:41 -0500
Branch: REL_15_STABLE [26c89d105] 2024-02-09 12:29:41 -0500
Branch: REL_14_STABLE [d21690edb] 2024-02-09 12:29:41 -0500
Branch: REL_13_STABLE [ceb224b62] 2024-02-09 12:29:41 -0500
Branch: REL_12_STABLE [f38903d1e] 2024-02-09 12:29:41 -0500
-->
     <para>
      Avoid race condition in <function>pg_get_expr()</function>
      (Tom Lane)
     </para>

     <para>
      If the relation referenced by the argument is dropped concurrently,
      the function's intention is to return NULL, but sometimes it failed
      instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [165d921c9] 2024-02-08 12:45:26 +0200
Branch: REL_16_STABLE [e3e05adde] 2024-02-09 12:37:21 +0200
Branch: REL_15_STABLE [503299b7f] 2024-02-09 12:38:32 +0200
Branch: REL_14_STABLE [18388291a] 2024-02-09 12:39:42 +0200
Branch: REL_13_STABLE [4efaf4b09] 2024-02-09 12:39:54 +0200
Branch: REL_12_STABLE [d44060cfc] 2024-02-09 12:55:43 +0200
-->
     <para>
      Fix detection of old transaction IDs in XID status functions
      (Karina Litskevich)
     </para>

     <para>
      Transaction IDs more than 2<superscript>31</superscript>
      transactions in the past could be misidentified as recent,
      leading to misbehavior of <function>pg_xact_status()</function>
      or <function>txid_status()</function>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [935829743] 2024-04-13 08:34:20 -0700
Branch: REL_16_STABLE [4e62ba21a] 2024-04-13 08:35:06 -0700
Branch: REL_15_STABLE [7c490a18b] 2024-04-13 08:35:20 -0700
Branch: REL_14_STABLE [08059fc04] 2024-04-13 08:35:32 -0700
-->
     <para>
      Ensure that a table's freespace map won't return a page that's past
      the end of the table (Ronan Dunklau)
     </para>

     <para>
      Because the freespace map isn't WAL-logged, this was possible in
      edge cases involving an OS crash, a replica promote, or a PITR
      restore.  The result would be a <quote>could not read block</quote>
      error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Etsuro Fujita <efujita@postgresql.org>
Branch: REL_16_STABLE [e79ceafe9] 2024-04-11 19:05:00 +0900
Branch: REL_15_STABLE [b82dca2a5] 2024-04-11 19:05:02 +0900
Branch: REL_14_STABLE [b714bc40c] 2024-04-11 19:05:04 +0900
Branch: REL_13_STABLE [01b01a77f] 2024-04-11 19:05:05 +0900
Branch: REL_12_STABLE [0341d4b10] 2024-04-11 19:05:07 +0900
-->
     <para>
      Fix file descriptor leakage when an error is thrown while waiting
      in <function>WaitEventSetWait</function> (Etsuro Fujita)
     </para>
    </listitem>

    <listitem>
<!--
Author: Etsuro Fujita <efujita@postgresql.org>
Branch: REL_16_STABLE [f6f61a4bd] 2024-04-04 17:25:00 +0900
Branch: REL_15_STABLE [3f96d113f] 2024-04-04 17:25:02 +0900
Branch: REL_14_STABLE [e10ca95ff] 2024-04-04 17:25:04 +0900
-->
     <para>
      Avoid corrupting exception stack if an FDW implements async append
      but doesn't configure any wait conditions for the Append plan node
      to wait for (Alexander Pyhalov)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f5a465f1a] 2024-02-25 16:15:07 -0500
Branch: REL_16_STABLE [8c785d354] 2024-02-25 16:15:07 -0500
Branch: REL_15_STABLE [940489b46] 2024-02-25 16:15:07 -0500
Branch: REL_14_STABLE [09f09884c] 2024-02-25 16:15:07 -0500
Branch: REL_13_STABLE [43cca9de9] 2024-02-25 16:15:07 -0500
Branch: REL_12_STABLE [c0b4dad38] 2024-02-25 16:15:07 -0500
-->
     <para>
      Throw an error if an index is accessed while it is being reindexed
      (Tom Lane)
     </para>

     <para>
      Previously this was just an assertion check, but promote it into a
      regular runtime error.  This will provide a more on-point error
      message when reindexing a user-defined index expression that
      attempts to access its own table.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [a63224be4] 2024-05-01 13:21:21 +1200
Branch: REL_16_STABLE [68d358545] 2024-05-01 13:21:50 +1200
Branch: REL_15_STABLE [52f21f928] 2024-05-01 13:22:16 +1200
Branch: REL_14_STABLE [e6b0efc65] 2024-05-01 13:22:41 +1200
Branch: REL_13_STABLE [0a34bcd0c] 2024-05-01 13:23:05 +1200
Branch: REL_12_STABLE [e3f9dcabd] 2024-05-01 13:23:25 +1200
-->
     <para>
      Ensure that index-only scans on <type>name</type> columns return a
      fully-padded value (David Rowley)
     </para>

     <para>
      The value physically stored in the index is truncated, and
      previously a pointer to that value was returned to callers.  This
      provoked complaints when testing under valgrind.  In theory it could
      result in crashes, though none have been reported.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [aa79bde72] 2024-04-25 10:40:52 +0530
Branch: REL_16_STABLE [a9155efc7] 2024-04-25 10:52:34 +0530
Branch: REL_15_STABLE [28a8cc457] 2024-04-25 10:33:04 +0530
-->
     <para>
      Fix race condition in deciding whether a table sync operation is
      needed in logical replication (Vignesh C)
     </para>

     <para>
      An invalidation event arriving while a subscriber identifies which
      tables need to be synced would be forgotten about, so that any
      tables newly in need of syncing might not get processed in a timely
      fashion.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [fbf9a7ac4] 2024-02-13 21:23:41 +0200
Branch: REL_16_STABLE [f2f09b825] 2024-02-13 21:24:45 +0200
Branch: REL_15_STABLE [d46c26961] 2024-02-13 21:25:27 +0200
Branch: REL_14_STABLE [9b8550fbd] 2024-02-13 21:25:39 +0200
Branch: REL_13_STABLE [e92375692] 2024-02-13 21:25:48 +0200
Branch: REL_12_STABLE [95cc48ca0] 2024-02-13 21:25:59 +0200
-->
     <para>
      Fix crash with DSM allocations larger than 4GB (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [f8c5317d0] 2024-03-12 10:18:32 +0200
Branch: REL_16_STABLE [539e328b1] 2024-03-12 10:18:44 +0200
Branch: REL_15_STABLE [4fce5f970] 2024-03-12 10:18:50 +0200
Branch: REL_14_STABLE [bf1f593e8] 2024-03-12 10:18:51 +0200
Branch: REL_13_STABLE [ad5cd55e6] 2024-03-12 10:18:53 +0200
Branch: REL_12_STABLE [df27d76d3] 2024-03-12 10:18:54 +0200
-->
     <para>
      Disconnect if a new server session's client socket cannot be put
      into non-blocking mode (Heikki Linnakangas)
     </para>

     <para>
      It was once theoretically possible for us to operate with a socket
      that's in blocking mode; but that hasn't worked fully in a long
      time, so fail at connection start rather than misbehave later.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [453c46873] 2024-03-07 19:38:17 -0500
Branch: REL_16_STABLE [6a2c80e95] 2024-03-07 19:37:51 -0500
Branch: REL_15_STABLE [0fe82e45c] 2024-03-07 19:37:51 -0500
Branch: REL_14_STABLE [473babd42] 2024-03-07 19:37:51 -0500
Branch: REL_13_STABLE [9fbe07275] 2024-03-07 19:37:51 -0500
Branch: REL_12_STABLE [c42e5fdcf] 2024-03-07 19:37:51 -0500
-->
     <para>
      Fix inadequate error reporting
      with <application>OpenSSL</application> 3.0.0 and later (Heikki
      Linnakangas, Tom Lane)
     </para>

     <para>
      System-reported errors passed through by OpenSSL were reported with
      a numeric error code rather than anything readable.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [a584d03ce] 2024-02-09 11:21:08 -0500
Branch: REL_16_STABLE [52afe5632] 2024-02-09 11:21:08 -0500
Branch: REL_15_STABLE [806f98951] 2024-02-09 11:21:08 -0500
Branch: REL_14_STABLE [8ead39e38] 2024-02-09 11:21:08 -0500
Branch: REL_13_STABLE [8c5da20db] 2024-02-09 11:21:08 -0500
Branch: REL_12_STABLE [9fb1396a9] 2024-02-09 11:21:08 -0500
Branch: master [0028b55bc] 2024-02-09 11:11:39 -0500
Branch: REL_16_STABLE [9440d23a0] 2024-02-09 11:11:39 -0500
Branch: REL_15_STABLE [9f041b041] 2024-02-09 11:11:39 -0500
Branch: REL_14_STABLE [7d7cc7fd6] 2024-02-09 11:11:39 -0500
Branch: REL_13_STABLE [67f8cf0f0] 2024-02-09 11:11:39 -0500
Branch: REL_12_STABLE [95e960e81] 2024-02-09 11:11:39 -0500
-->
     <para>
      Avoid concurrent calls to <function>bindtextdomain()</function>
      in <application>libpq</application>
      and <application>ecpglib</application> (Tom Lane)
     </para>

     <para>
      Although GNU <application>gettext</application>'s implementation
      seems to be fine with concurrent calls, the version available on
      Windows is not.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [6f0cef935] 2024-04-16 12:31:42 -0400
Branch: REL_16_STABLE [392e6e9e6] 2024-04-16 12:31:32 -0400
Branch: REL_15_STABLE [25f937217] 2024-04-16 12:31:32 -0400
Branch: REL_14_STABLE [463164639] 2024-04-16 12:31:32 -0400
Branch: REL_13_STABLE [02531e8ca] 2024-04-16 12:31:32 -0400
Branch: REL_12_STABLE [2b6a74afe] 2024-04-16 12:31:32 -0400
Branch: REL_16_STABLE [0018f0af5] 2024-04-18 20:47:37 -0400
Branch: REL_15_STABLE [1e7b1b026] 2024-04-18 20:47:37 -0400
Branch: REL_14_STABLE [df66319f7] 2024-04-18 20:47:37 -0400
Branch: REL_13_STABLE [481597fc6] 2024-04-18 20:47:37 -0400
Branch: REL_12_STABLE [61dd815e0] 2024-04-18 20:47:37 -0400
Branch: REL_16_STABLE [dd3fddc85] 2024-04-19 01:07:16 -0400
Branch: REL_15_STABLE [f7e891748] 2024-04-19 01:07:32 -0400
Branch: REL_14_STABLE [de84608e2] 2024-04-19 01:07:41 -0400
Branch: REL_13_STABLE [c6bfeab42] 2024-04-19 01:07:47 -0400
Branch: REL_12_STABLE [cd26f08e4] 2024-04-19 01:07:52 -0400
-->
     <para>
      Fix crash in <application>ecpg</application>'s preprocessor if
      the program tries to redefine a macro that was defined on the
      preprocessor command line (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [096a761d6] 2024-04-04 15:31:53 -0400
Branch: REL_16_STABLE [118558e6d] 2024-04-04 15:31:53 -0400
Branch: REL_15_STABLE [f159f1814] 2024-04-04 15:31:53 -0400
Branch: REL_14_STABLE [d3167ed3f] 2024-04-04 15:31:53 -0400
Branch: REL_13_STABLE [5ba29e945] 2024-04-04 15:31:53 -0400
Branch: REL_12_STABLE [360d007e3] 2024-04-04 15:31:53 -0400
-->
     <para>
      In <application>ecpg</application>, avoid issuing
      false <quote>unsupported feature will be passed to server</quote>
      warnings (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [e77a1c58e] 2024-02-19 11:38:18 +0900
Branch: REL_16_STABLE [88e03d055] 2024-02-19 11:38:44 +0900
Branch: REL_15_STABLE [b5cb6022b] 2024-02-19 11:38:47 +0900
Branch: REL_14_STABLE [a05bb9add] 2024-02-19 11:38:49 +0900
Branch: REL_13_STABLE [c031ce97b] 2024-02-19 11:38:52 +0900
Branch: REL_12_STABLE [771240f97] 2024-02-19 11:38:54 +0900
-->
     <para>
      Ensure that the string result
      of <application>ecpg</application>'s <function>intoasc()</function>
      function is correctly zero-terminated (Oleg Tselebrovskiy)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f463de59d] 2024-04-08 17:00:07 -0400
Branch: REL_16_STABLE [a85e3ba1c] 2024-04-08 17:00:07 -0400
Branch: REL_15_STABLE [4f1d33d70] 2024-04-08 17:00:07 -0400
-->
     <para>
      In <application>psql</application>, avoid leaking a query result
      after the query is cancelled (Tom Lane)
     </para>

     <para>
      This happened only when cancelling a non-last query in a query
      string made with <literal>\;</literal> separators.
     </para>
    </listitem>

    <listitem>
<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
Branch: master [7e65ad197] 2024-03-21 23:31:57 +0100
Branch: REL_16_STABLE [5863bacb8] 2024-03-21 23:31:57 +0100
Branch: REL_15_STABLE [12128be62] 2024-03-21 23:31:57 +0100
Branch: REL_14_STABLE [be01c8c34] 2024-03-21 23:31:57 +0100
Branch: REL_13_STABLE [affc46b76] 2024-03-21 23:31:57 +0100
Branch: REL_12_STABLE [d82cb467b] 2024-03-21 23:31:57 +0100
Branch: REL_14_STABLE [6ebd43725] 2024-03-22 01:01:30 +0100
Branch: REL_13_STABLE [d5c6affb8] 2024-03-22 01:01:30 +0100
Branch: REL_12_STABLE [82c2192d9] 2024-03-22 01:01:30 +0100
-->
     <para>
      Fix <application>pg_dumpall</application> so that role comments, if
      present, will be dumped regardless of the setting
      of <option>--no-role-passwords</option> (Daniel Gustafsson,
      &Aacute;lvaro Herrera)
     </para>
    </listitem>

    <listitem>
<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
Branch: master [c1fc502f5] 2024-02-13 13:47:12 +0100
Branch: REL_16_STABLE [103235888] 2024-02-13 13:47:12 +0100
Branch: REL_15_STABLE [29f005238] 2024-02-13 13:47:12 +0100
-->
     <para>
      Skip files named <filename>.DS_Store</filename>
      in <application>pg_basebackup</application>,
      <application>pg_checksums</application>,
      and <application>pg_rewind</application> (Daniel Gustafsson)
     </para>

     <para>
      This avoids problems on macOS, where the Finder may create such
      files.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5392dd3d2] 2024-04-10 15:45:58 -0400
Branch: REL_16_STABLE [48f216dc6] 2024-04-10 15:45:58 -0400
Branch: REL_15_STABLE [d85db0a8e] 2024-04-10 15:45:59 -0400
Branch: REL_14_STABLE [dc5824a06] 2024-04-10 15:45:59 -0400
Branch: REL_13_STABLE [f5cee411a] 2024-04-10 15:45:59 -0400
Branch: REL_12_STABLE [5e9d8bed0] 2024-04-10 15:45:59 -0400
-->
     <para>
      Fix <application>PL/pgSQL</application>'s parsing of single-line
      comments (<literal>--</literal>-style comments) following
      expressions (Erik Wienhold, Tom Lane)
     </para>

     <para>
      This mistake caused parse errors if such a comment followed
      a <literal>WHEN</literal> expression in
      a <application>PL/pgSQL</application> <command>CASE</command>
      statement.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [b1fe8efdf] 2024-03-24 00:09:24 +0200
Branch: REL_16_STABLE [3676b846b] 2024-03-23 23:02:30 +0200
Branch: REL_15_STABLE [0d466bce9] 2024-03-23 23:02:43 +0200
Branch: REL_14_STABLE [4e8529da4] 2024-03-23 23:03:10 +0200
Branch: REL_13_STABLE [5cc1f2626] 2024-03-23 23:03:12 +0200
Branch: REL_12_STABLE [d603e6744] 2024-03-24 00:08:13 +0200
Branch: master [ab65dfb0f] 2024-03-24 00:09:24 +0200
Branch: REL_16_STABLE [a6ddb8ad0] 2024-03-23 23:02:30 +0200
Branch: REL_15_STABLE [54e6184db] 2024-03-23 23:02:43 +0200
Branch: REL_14_STABLE [5df5d9cd7] 2024-03-23 23:03:10 +0200
Branch: REL_13_STABLE [e2c241416] 2024-03-23 23:03:12 +0200
Branch: REL_12_STABLE [50f8611d0] 2024-03-23 23:03:14 +0200
-->
     <para>
      In <filename>contrib/amcheck</filename>, don't report false match
      failures due to short- versus long-header values (Andrey Borodin,
      Michael Zhilin)
     </para>

     <para>
      A variable-length datum in a heap tuple or index tuple could have
      either a short or a long header, depending on compression parameters
      that applied when it was made.  Treat these cases as equivalent
      rather than complaining if there's a difference.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [bb616ed3e] 2024-04-14 18:19:58 +0200
Branch: REL_16_STABLE [8cea358b1] 2024-04-14 18:24:38 +0200
Branch: master [2f20ced1e] 2024-04-14 18:07:15 +0200
Branch: REL_16_STABLE [ccd8f0fa1] 2024-04-14 18:17:09 +0200
Branch: REL_15_STABLE [3cd413511] 2024-04-14 18:17:29 +0200
Branch: REL_14_STABLE [ad23af83d] 2024-04-14 18:18:07 +0200
-->
     <para>
      Fix bugs in BRIN output functions (Tomas Vondra)
     </para>

     <para>
      These output functions are only used for displaying index entries
      in <filename>contrib/pageinspect</filename>, so the errors are of
      limited practical concern.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [c399248b3] 2024-03-11 12:27:11 +1300
Branch: REL_16_STABLE [6a9e2cb2b] 2024-03-11 12:27:46 +1300
Branch: REL_15_STABLE [ab64b275a] 2024-03-11 12:28:11 +1300
Branch: REL_14_STABLE [628c3f2e1] 2024-03-11 12:28:40 +1300
Branch: REL_13_STABLE [20b85b3da] 2024-03-11 12:29:03 +1300
Branch: REL_12_STABLE [9301e0f41] 2024-03-11 12:29:24 +1300
-->
     <para>
      In <filename>contrib/postgres_fdw</filename>, avoid emitting
      requests to sort by a constant (David Rowley)
     </para>

     <para>
      This could occur in cases involving <literal>UNION ALL</literal>
      with constant-emitting subqueries.  Sorting by a constant is useless
      of course, but it also risks being misinterpreted by the remote
      server, leading to <quote>ORDER BY
      position <replaceable>N</replaceable> is not in select list</quote>
      errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [a3021aafc] 2024-04-21 13:46:20 -0400
Branch: REL_16_STABLE [75929b6cf] 2024-04-21 13:46:20 -0400
Branch: REL_15_STABLE [6c85e3359] 2024-04-21 13:46:20 -0400
Branch: REL_14_STABLE [3ed6e1698] 2024-04-21 13:46:20 -0400
Branch: REL_13_STABLE [0e56b2b94] 2024-04-21 13:46:20 -0400
Branch: REL_12_STABLE [ce1c30ece] 2024-04-21 13:46:20 -0400
-->
     <para>
      Make <filename>contrib/postgres_fdw</filename> set the remote
      session's time zone to <literal>GMT</literal>
      not <literal>UTC</literal> (Tom Lane)
     </para>

     <para>
      This should have the same results for practical purposes.
      However, <literal>GMT</literal> is recognized by hard-wired code in
      the server, while <literal>UTC</literal> is looked up in the
      timezone database.  So the old code could fail in the unlikely event
      that the remote server's timezone database is missing entries.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: REL_16_STABLE [7c93f31de] 2024-04-16 12:25:48 +0900
Branch: REL_15_STABLE [689ba4f1c] 2024-04-16 12:26:10 +0900
Branch: REL_14_STABLE [6fa5e67e8] 2024-04-16 12:26:15 +0900
Branch: REL_13_STABLE [bb418aeee] 2024-04-16 12:26:17 +0900
Branch: REL_12_STABLE [4b0e5d601] 2024-04-16 12:26:21 +0900
-->
     <para>
      In <filename>contrib/xml2</filename>, avoid use of library functions
      that have been deprecated in recent versions
      of <application>libxml2</application> (Dmitry Koval)
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [53c8d6c9f] 2024-04-10 12:13:46 +1200
Branch: REL_16_STABLE [bf1cfe77e] 2024-04-10 10:46:15 +1200
Branch: REL_15_STABLE [74992929a] 2024-04-10 12:14:04 +1200
Branch: REL_14_STABLE [7fe32eaa4] 2024-04-10 12:15:07 +1200
Branch: REL_13_STABLE [4f90750b5] 2024-04-10 12:15:41 +1200
Branch: REL_12_STABLE [01b55203a] 2024-04-10 12:15:59 +1200
-->
     <para>
      Fix incompatibility with LLVM 18 (Thomas Munro, Dmitry Dolgov)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [8a92b70c1] 2024-03-26 11:44:49 -0400
Branch: REL_16_STABLE [7651fd387] 2024-03-26 11:44:49 -0400
Branch: REL_15_STABLE [3c3f4fd74] 2024-03-26 11:44:49 -0400
Branch: REL_14_STABLE [d82605bcd] 2024-03-26 11:44:49 -0400
Branch: REL_13_STABLE [243e99532] 2024-03-26 11:44:49 -0400
Branch: REL_12_STABLE [7124e7d52] 2024-03-26 11:44:49 -0400
-->
     <para>
      Allow <literal>make check</literal> to work with
      the <application>musl</application> C library (Thomas Munro, Bruce
      Momjian, Tom Lane)
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15-6">
  <title>Release 15.6</title>

  <formalpara>
  <title>Release date:</title>
  <para>2024-02-08</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.5.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.6</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, one bug was fixed that could have resulted in corruption of
    GIN indexes during concurrent updates.  If you suspect such
    corruption, reindex affected indexes after installing this update.
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 15.5,
    see <xref linkend="release-15-5"/>.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [5a9167c39] 2024-02-05 11:01:23 +0200
Branch: REL_16_STABLE [d6a61cb3b] 2024-02-05 11:02:56 +0200
Branch: REL_15_STABLE [f2fdea198] 2024-02-05 11:03:26 +0200
Branch: REL_14_STABLE [f4f288352] 2024-02-05 11:03:43 +0200
Branch: REL_13_STABLE [d541ce3b6] 2024-02-05 11:04:08 +0200
Branch: REL_12_STABLE [2699fc035] 2024-02-05 11:04:22 +0200
Branch: master [b96115acb] 2024-02-05 11:01:30 +0200
Branch: REL_16_STABLE [fb3836855] 2024-02-05 11:03:03 +0200
Branch: REL_15_STABLE [06f36bc01] 2024-02-05 11:03:28 +0200
Branch: REL_14_STABLE [a45c950ae] 2024-02-05 11:03:43 +0200
Branch: REL_13_STABLE [b73d21648] 2024-02-05 11:04:10 +0200
Branch: REL_12_STABLE [add8bc9b8] 2024-02-05 11:04:23 +0200
-->
     <para>
      Tighten security restrictions within <command>REFRESH MATERIALIZED
      VIEW CONCURRENTLY</command> (Heikki Linnakangas)
     </para>

     <para>
      One step of a concurrent refresh command was run under weak security
      restrictions.  If a materialized view's owner could persuade a
      superuser or other high-privileged user to perform a concurrent
      refresh on that view, the view's owner could control code executed
      with the privileges of the user running <command>REFRESH</command>.
      Fix things so that all user-determined code is run as the view's
      owner, as expected.
     </para>

     <para>
      The only known exploit for this error does not work
      in <productname>PostgreSQL</productname> 16.0 and later, so it may
      be that v16 is not vulnerable in practice.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks Pedro
      Gallegos for reporting this problem.
      (CVE-2024-0985) <!-- not CVE-2023-5869 as claimed in commit msg -->
     </para>
    </listitem>

    <listitem>
<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
Branch: master [9dce22033] 2023-09-27 13:02:21 +0200
Branch: REL_16_STABLE [2cf50585e] 2023-11-17 10:18:38 +0100
Branch: REL_15_STABLE [aef521849] 2023-11-17 10:21:34 +0100
Branch: REL_14_STABLE [75a20a4b4] 2023-11-17 10:27:45 +0100
Branch: REL_13_STABLE [10912f7d4] 2023-11-17 11:46:31 +0100
Branch: REL_12_STABLE [3b991f81c] 2023-11-17 11:47:17 +0100
Branch: REL_12_STABLE [c922b2410] 2023-12-11 12:02:01 +0100
-->
     <para>
      Fix memory leak when performing JIT inlining (Andres Freund,
      Daniel Gustafsson)
     </para>

     <para>
      There have been multiple reports of backend processes suffering
      out-of-memory conditions after sufficiently many JIT compilations.
      This fix should resolve that.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: REL_15_STABLE [f374fb4aa] 2024-01-18 11:12:31 +0900
Branch: REL_14_STABLE [5f6ec27a6] 2024-01-18 11:12:38 +0900
Branch: REL_13_STABLE [dc9d424cf] 2024-01-18 11:12:43 +0900
Branch: REL_12_STABLE [810382285] 2024-01-18 11:12:48 +0900
-->
     <para>
      When dequeueing from an LWLock, avoid needing to search the list of
      waiting processes (Andres Freund)
     </para>

     <para>
      This fixes O(N^2) behavior when the list of waiters is long.  In
      some use-cases this results in substantial throughput improvements.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: REL_16_STABLE [62f120203] 2024-02-01 12:34:21 -0500
Branch: REL_15_STABLE [12ec16d11] 2024-02-01 12:34:21 -0500
Branch: REL_14_STABLE [e031995d5] 2024-02-01 12:34:21 -0500
Branch: REL_13_STABLE [7af96a66f] 2024-02-01 12:34:21 -0500
Branch: REL_12_STABLE [2e822a1d6] 2024-02-01 12:34:21 -0500
-->
     <para>
      Avoid generating incorrect partitioned-join plans (Richard Guo)
     </para>

     <para>
      Some uncommon situations involving lateral references could create
      incorrect plans.  Affected queries could produce wrong answers, or
      odd failures such as <quote>variable not found in subplan target
      list</quote>, or executor crashes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [29f114b6f] 2024-01-11 15:28:22 -0500
Branch: REL_16_STABLE [80bece312] 2024-01-11 15:28:13 -0500
Branch: REL_15_STABLE [a0b4fda44] 2024-01-11 15:28:13 -0500
Branch: REL_14_STABLE [ea1d54288] 2024-01-11 15:28:13 -0500
Branch: REL_13_STABLE [bfd28bb07] 2024-01-11 15:28:13 -0500
Branch: REL_12_STABLE [69c12c417] 2024-01-11 15:28:13 -0500
-->
     <para>
      Fix incorrect wrapping of subquery output expressions in
      PlaceHolderVars (Tom Lane)
     </para>

     <para>
      This fixes incorrect results when a subquery is underneath an outer
      join and has an output column that laterally references something
      outside the outer join's scope.  The output column might not appear
      as NULL when it should do so due to the action of the outer join.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d641b827a] 2024-01-10 13:36:33 -0500
Branch: REL_16_STABLE [ee9553218] 2024-01-10 13:36:34 -0500
Branch: REL_15_STABLE [c3f52fd5d] 2024-01-10 13:36:34 -0500
-->
     <para>
      Fix misprocessing of window function run conditions (Richard Guo)
     </para>

     <para>
      This oversight could lead to <quote>WindowFunc not found in subplan
      target lists</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [a4f7d33a9] 2023-11-09 11:23:42 +0000
Branch: REL_16_STABLE [06a546382] 2023-11-09 11:27:04 +0000
Branch: REL_15_STABLE [c0bfdaf2b] 2023-11-09 11:28:25 +0000
-->
     <para>
      Skip inappropriate actions when <command>MERGE</command> causes a
      cross-partition update (Dean Rasheed)
     </para>

     <para>
      When executing a <literal>MERGE UPDATE</literal> action on a
      partitioned table, if the <literal>UPDATE</literal> is turned into
      a <literal>DELETE</literal> and <literal>INSERT</literal> due to
      changing a partition key column, skip firing <literal>AFTER
      UPDATE ROW</literal> triggers, as well as other post-update actions
      such as RLS checks.  These actions would typically fail, which is
      why a regular <literal>UPDATE</literal> doesn't do them in such
      cases; <literal>MERGE</literal> shouldn't either.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [a0ff37173] 2023-12-21 12:55:22 +0000
Branch: REL_16_STABLE [7f07384dc] 2023-12-21 12:53:48 +0000
Branch: REL_15_STABLE [7e8c6d7af] 2023-12-21 12:51:55 +0000
-->
     <para>
      Cope with <literal>BEFORE ROW DELETE</literal> triggers in
      cross-partition <command>MERGE</command> updates (Dean Rasheed)
     </para>

     <para>
      If such a trigger attempted to prevent the update by returning
      NULL, <command>MERGE</command> would suffer an error or assertion
      failure.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [6545ba96c] 2024-01-14 12:38:41 -0500
Branch: REL_16_STABLE [23e0ba59c] 2024-01-14 12:38:41 -0500
Branch: REL_15_STABLE [1a4e54617] 2024-01-14 12:38:41 -0500
Branch: REL_14_STABLE [c7edaeec5] 2024-01-14 12:38:41 -0500
-->
     <para>
      Prevent access to a no-longer-pinned buffer in <literal>BEFORE ROW
      UPDATE</literal> triggers (Alexander Lakhin, Tom Lane)
     </para>

     <para>
      If the tuple being updated had just been updated and moved to
      another page by another session, there was a narrow window where
      we would attempt to fetch data from the new tuple version without
      any pin on its buffer.  In principle this could result in garbage
      data appearing in non-updated columns of the proposed new tuple.
      The odds of problems in practice seem rather low, however.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [2a67b5a60] 2024-01-07 09:10:19 +0200
Branch: REL_16_STABLE [714a987bc] 2024-01-07 09:32:58 +0200
Branch: REL_15_STABLE [1a7c03e6f] 2024-01-07 09:10:40 +0200
Branch: REL_14_STABLE [3bdaa8fc6] 2024-01-07 09:10:49 +0200
Branch: REL_13_STABLE [60de25c6e] 2024-01-07 09:11:00 +0200
Branch: REL_12_STABLE [72d5b2776] 2024-01-07 09:18:43 +0200
Branch: master [6d94cc6ca] 2024-01-08 20:00:20 +0200
Branch: REL_16_STABLE [20a566cd4] 2024-01-08 19:57:09 +0200
Branch: REL_15_STABLE [6eecc3a62] 2024-01-08 19:57:03 +0200
Branch: REL_14_STABLE [97e64911d] 2024-01-08 19:57:58 +0200
Branch: REL_13_STABLE [17b6f39f1] 2024-01-08 19:58:38 +0200
Branch: REL_12_STABLE [a5e2853c3] 2024-01-08 19:58:51 +0200
-->
     <para>
      Avoid requesting an oversize shared-memory area in parallel hash
      join (Thomas Munro, Andrei Lepikhov, Alexander Korotkov)
     </para>

     <para>
      The limiting value was too large, allowing <quote>invalid DSA memory
      alloc request size</quote> errors to occur with sufficiently large
      expected hash table sizes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [60f227316] 2023-11-28 12:00:14 +0200
Branch: REL_16_STABLE [9fee3232a] 2023-11-28 11:59:45 +0200
Branch: REL_15_STABLE [2873fbfe0] 2023-11-28 11:59:50 +0200
Branch: REL_14_STABLE [59c62a21f] 2023-11-28 11:59:51 +0200
Branch: REL_13_STABLE [e434d3620] 2023-11-28 11:59:52 +0200
Branch: REL_12_STABLE [b8a606e21] 2023-11-28 11:59:53 +0200
-->
     <para>
      Avoid assertion failures in <function>heap_update()</function>
      and <function>heap_delete()</function> when a tuple to be updated by
      a foreign-key enforcement trigger fails the extra visibility
      crosscheck (Alexander Lakhin)
     </para>

     <para>
      This error had no impact in non-assert builds.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [1b2c6b756] 2023-11-08 14:06:26 +0900
Branch: REL_16_STABLE [4dccf9575] 2023-11-08 14:06:36 +0900
Branch: REL_15_STABLE [7e18c0bd6] 2023-11-08 14:06:39 +0900
Branch: REL_14_STABLE [59fc39c0d] 2023-11-08 14:06:42 +0900
-->
     <para>
      Fix overly tight assertion
      about <varname>false_positive_rate</varname> parameter of
      BRIN bloom operator classes (Alexander Lakhin)
     </para>

     <para>
      This error had no impact in non-assert builds, either.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [bb812ab09] 2024-01-24 14:20:01 +0900
Branch: REL_16_STABLE [51193e7a7] 2024-01-24 14:20:08 +0900
Branch: REL_15_STABLE [ad6fbbeeb] 2024-01-24 14:20:10 +0900
Branch: REL_14_STABLE [5a7833f49] 2024-01-24 14:20:11 +0900
Branch: REL_13_STABLE [bfec14d06] 2024-01-24 14:20:13 +0900
Branch: REL_12_STABLE [2f7242837] 2024-01-24 14:20:14 +0900
-->
     <para>
      Fix possible failure during <command>ALTER TABLE ADD
      COLUMN</command> on a complex inheritance tree (Tender Wang)
     </para>

     <para>
      If a grandchild table would inherit the new column via multiple
      intermediate parents, the command failed with <quote>tuple already
      updated by self</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [3e91dba8b] 2024-01-31 13:15:21 +0900
Branch: REL_16_STABLE [f33e83285] 2024-01-31 13:16:43 +0900
Branch: REL_15_STABLE [41fa4b31c] 2024-01-31 13:16:46 +0900
Branch: REL_14_STABLE [dde5b01c3] 2024-01-31 13:16:47 +0900
Branch: REL_13_STABLE [4d0e8a008] 2024-01-31 13:16:49 +0900
Branch: REL_12_STABLE [056109782] 2024-01-31 13:16:50 +0900
-->
     <para>
      Fix problems with duplicate token names in <command>ALTER TEXT
      SEARCH CONFIGURATION ... MAPPING</command> commands (Tender Wang,
      Michael Paquier)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [28f84f72f] 2023-11-19 21:03:38 +0100
Branch: REL_16_STABLE [ee32b824d] 2023-11-19 21:03:50 +0100
Branch: REL_15_STABLE [0177fc773] 2023-11-19 21:04:16 +0100
Branch: REL_14_STABLE [9dd50e92e] 2023-11-19 21:04:28 +0100
Branch: REL_13_STABLE [1e468ea4e] 2023-11-19 21:04:39 +0100
Branch: REL_12_STABLE [4f8d3c5b5] 2023-11-19 21:04:47 +0100
-->
     <para>
      Properly lock the associated table during <command>DROP
      STATISTICS</command> (Tomas Vondra)
     </para>

     <para>
      Failure to acquire the lock could result in <quote>tuple
      concurrently deleted</quote> errors if the <command>DROP</command>
      executes concurrently with <command>ANALYZE</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [743ddafc7] 2023-11-16 10:05:14 -0500
Branch: REL_16_STABLE [f07a3039c] 2023-11-16 10:05:14 -0500
Branch: REL_15_STABLE [9057ddbef] 2023-11-16 10:05:14 -0500
Branch: REL_14_STABLE [9bd0f74ea] 2023-11-16 10:05:14 -0500
Branch: REL_13_STABLE [5c11104c7] 2023-11-16 10:05:14 -0500
Branch: REL_12_STABLE [abd1b1325] 2023-11-16 10:05:14 -0500
-->
     <para>
      Fix function volatility checking for <literal>GENERATED</literal>
      and <literal>DEFAULT</literal> expressions (Tom Lane)
     </para>

     <para>
      These places could fail to detect insertion of a volatile function
      default-argument expression, or decide that a polymorphic function
      is volatile although it is actually immutable on the datatype of
      interest.  This could lead to improperly rejecting or accepting
      a <literal>GENERATED</literal> clause, or to mistakenly applying the
      constant-default-value optimization in <command>ALTER TABLE ADD
      COLUMN</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [ad98fb142] 2024-01-13 13:46:27 -0500
Branch: REL_16_STABLE [7e2561e1a] 2024-01-13 13:46:27 -0500
Branch: REL_15_STABLE [2a46a0df4] 2024-01-13 13:46:27 -0500
Branch: REL_14_STABLE [db122d426] 2024-01-13 13:46:27 -0500
Branch: REL_13_STABLE [98e03f957] 2024-01-13 13:46:27 -0500
Branch: REL_12_STABLE [3b4d85cf1] 2024-01-13 13:46:27 -0500
Branch: master [96c019ffa] 2024-01-13 13:54:11 -0500
Branch: REL_16_STABLE [56dcd71de] 2024-01-13 13:54:11 -0500
Branch: REL_15_STABLE [d41358f4b] 2024-01-13 13:54:11 -0500
Branch: REL_14_STABLE [7ceeb57ba] 2024-01-13 13:54:11 -0500
Branch: REL_13_STABLE [475b3ea3c] 2024-01-13 13:54:11 -0500
Branch: REL_12_STABLE [d29a4fbac] 2024-01-13 13:54:11 -0500
-->
     <para>
      Detect that a new catalog cache entry became stale while detoasting
      its fields (Tom Lane)
     </para>

     <para>
      We expand any out-of-line fields in a catalog tuple before inserting
      it into the catalog caches.  That involves database access which
      might cause invalidation of catalog cache entries &mdash; but the
      new entry isn't in the cache yet, so we would miss noticing that it
      should get invalidated.  The result is a race condition in which an
      already-stale cache entry could get made, and then persist
      indefinitely.  This would lead to hard-to-predict misbehavior.
      Fix by rechecking the tuple's visibility after detoasting.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [0e3e8fbd3] 2023-11-09 09:50:23 +0000
Branch: REL_16_STABLE [c396aca2b] 2023-11-09 09:53:05 +0000
Branch: REL_15_STABLE [308a69a98] 2023-11-09 09:54:22 +0000
Branch: REL_14_STABLE [dea12b40d] 2023-11-09 09:55:39 +0000
Branch: REL_13_STABLE [4f4a0010a] 2023-11-09 09:56:31 +0000
Branch: REL_12_STABLE [b17a02be2] 2023-11-09 09:57:52 +0000
-->
     <para>
      Fix edge-case integer overflow detection bug on some platforms (Dean
      Rasheed)
     </para>

     <para>
      Computing <literal>0 - INT64_MIN</literal> should result in an
      overflow error, and did on most platforms.  However, platforms with
      neither integer overflow builtins nor 128-bit integers would fail to
      spot the overflow, instead returning <literal>INT64_MIN</literal>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [25cd2d640] 2024-01-26 13:39:45 -0500
Branch: REL_16_STABLE [7204aea83] 2024-01-26 13:39:37 -0500
Branch: REL_15_STABLE [86b6243a8] 2024-01-26 13:39:37 -0500
Branch: REL_14_STABLE [3eb8a87f6] 2024-01-26 13:39:37 -0500
Branch: REL_13_STABLE [425127bed] 2024-01-26 13:39:37 -0500
Branch: REL_12_STABLE [c3bdb25fb] 2024-01-26 13:39:37 -0500
-->
     <para>
      Detect Julian-date overflow when adding or subtracting
      an <type>interval</type> to/from a <type>timestamp</type> (Tom Lane)
     </para>

     <para>
      Some cases that should cause an out-of-range error produced an
      incorrect result instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [b218fbb7a] 2023-11-18 14:41:20 +0000
Branch: REL_16_STABLE [72d0c135b] 2023-11-18 14:46:02 +0000
Branch: REL_15_STABLE [2851aa7d1] 2023-11-18 14:47:04 +0000
Branch: REL_14_STABLE [2ffcebdba] 2023-11-18 14:48:10 +0000
Branch: REL_13_STABLE [428770aad] 2023-11-18 14:49:18 +0000
Branch: REL_12_STABLE [f499d2b20] 2023-11-18 14:50:00 +0000
-->
     <para>
      Add more checks for overflow in <function>interval_mul()</function>
      and <function>interval_div()</function> (Dean Rasheed)
     </para>

     <para>
      Some cases that should cause an out-of-range error produced an
      incorrect result instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [235c09efb] 2024-02-01 17:12:50 +0900
Branch: REL_16_STABLE [781bc121d] 2024-02-01 17:13:09 +0900
Branch: REL_15_STABLE [171d21f50] 2024-02-01 17:13:11 +0900
-->
     <para>
      Ensure cached statistics are discarded after a change
      to <varname>stats_fetch_consistency</varname> (Shinya Kato)
     </para>

     <para>
      In some code paths, it was possible for stale statistics to be
      returned.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [059de3ca4] 2023-12-26 17:57:48 -0500
Branch: REL_16_STABLE [e87252ceb] 2023-12-26 17:57:48 -0500
Branch: REL_15_STABLE [76dd3d94a] 2023-12-26 17:57:48 -0500
Branch: REL_14_STABLE [bf4f30a0f] 2023-12-26 17:57:48 -0500
Branch: REL_13_STABLE [cb88f44ec] 2023-12-26 17:57:48 -0500
Branch: REL_12_STABLE [ea61b1cf6] 2023-12-26 17:57:48 -0500
-->
     <para>
      Make the <structname>pg_file_settings</structname> view check
      validity of unapplied values for settings
      with <literal>backend</literal>
      or <literal>superuser-backend</literal> context (Tom Lane)
     </para>

     <para>
      Invalid values were not noted in the view as intended.  This escaped
      detection because there are very few settings in these groups.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [a11c9c42e] 2023-12-01 16:05:41 +0100
Branch: REL_16_STABLE [267f33f68] 2023-12-01 16:09:54 +0100
Branch: REL_15_STABLE [15d485921] 2023-12-01 16:14:22 +0100
Branch: REL_14_STABLE [e846fc491] 2023-12-01 16:17:38 +0100
Branch: REL_13_STABLE [3c49fa2af] 2023-12-01 16:23:18 +0100
Branch: REL_12_STABLE [5d40b3c4f] 2023-12-01 16:27:18 +0100
-->
     <para>
      Match collation too when matching an existing index to a new
      partitioned index (Peter Eisentraut)
     </para>

     <para>
      Previously we could accept an index that has a different collation
      from the corresponding element of the partition key, possibly
      leading to misbehavior.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [0ae3b4662] 2024-01-18 16:30:11 +0900
Branch: REL_16_STABLE [c030e263e] 2024-01-18 16:31:38 +0900
Branch: REL_15_STABLE [a0c19de11] 2024-01-18 16:31:46 +0900
Branch: REL_14_STABLE [b6f2e019d] 2024-01-18 16:31:50 +0900
Branch: master [8013850c8] 2024-01-18 15:04:24 +0900
Branch: REL_16_STABLE [7ce65c6f7] 2024-01-18 15:04:31 +0900
Branch: REL_15_STABLE [1cf2dba84] 2024-01-18 15:04:35 +0900
Branch: REL_14_STABLE [b685b41cf] 2024-01-18 15:04:39 +0900
-->
     <para>
      Avoid failure if a child index is dropped concurrently
      with <command>REINDEX INDEX</command> on a partitioned index
      (Fei Changhong)
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [6a1ea02c4] 2024-01-29 13:46:22 +0200
Branch: REL_16_STABLE [b899e00e7] 2024-01-29 13:46:30 +0200
Branch: REL_15_STABLE [e43425f48] 2024-01-29 13:46:42 +0200
Branch: REL_14_STABLE [f120c0872] 2024-01-29 13:46:43 +0200
Branch: REL_13_STABLE [e74c91665] 2024-01-29 13:46:45 +0200
Branch: REL_12_STABLE [e6511fe64] 2024-01-29 13:46:48 +0200
-->
     <para>
      Fix insufficient locking when cleaning up an incomplete split of
      a GIN index's internal page (Fei Changhong, Heikki Linnakangas)
     </para>

     <para>
      The code tried to do this with shared rather than exclusive lock on
      the buffer.  This could lead to index corruption if two processes
      attempted the cleanup concurrently.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5c62ecf6e] 2023-11-13 11:44:35 -0500
Branch: REL_16_STABLE [f76b975d5] 2023-11-13 11:44:48 -0500
Branch: REL_15_STABLE [4c73ec604] 2023-11-13 11:44:57 -0500
Branch: REL_14_STABLE [d900e74e0] 2023-11-13 11:45:04 -0500
Branch: REL_13_STABLE [c859d688b] 2023-11-13 11:45:09 -0500
Branch: REL_12_STABLE [5a6937ec9] 2023-11-13 11:45:13 -0500
-->
     <para>
      Avoid premature release of buffer pin in GIN index insertion
      (Tom Lane)
     </para>

     <para>
      If an index root page split occurs concurrently with our own
      insertion, the code could fail with <quote>buffer NNNN is not owned
      by resource owner</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [903737c5b] 2023-12-21 12:43:36 -0500
Branch: REL_16_STABLE [0977bd64e] 2023-12-21 12:43:36 -0500
Branch: REL_15_STABLE [ab04c1901] 2023-12-21 12:43:36 -0500
Branch: REL_14_STABLE [375f441bd] 2023-12-21 12:43:36 -0500
Branch: REL_13_STABLE [cc3865150] 2023-12-21 12:43:36 -0500
Branch: REL_12_STABLE [1771ec9a8] 2023-12-21 12:43:36 -0500
-->
     <para>
      Avoid failure with partitioned SP-GiST indexes (Tom Lane)
     </para>

     <para>
      Trying to use an index of this kind could lead to <quote>No such
      file or directory</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [59bd34c2f] 2023-12-15 13:55:05 -0500
Branch: REL_16_STABLE [152bfc0af] 2023-12-15 13:55:05 -0500
Branch: REL_15_STABLE [7a99fb6e1] 2023-12-15 13:55:05 -0500
Branch: REL_14_STABLE [f552f2be2] 2023-12-15 13:55:05 -0500
Branch: REL_13_STABLE [55b5c67da] 2023-12-15 13:55:05 -0500
Branch: REL_12_STABLE [ba66f2533] 2023-12-15 13:55:05 -0500
-->
     <para>
      Fix ownership change reporting for large objects (Tom Lane)
     </para>

     <para>
      A no-op <command>ALTER LARGE OBJECT OWNER</command> command (that
      is, one selecting the existing owner) passed the wrong class ID to
      the <varname>PostAlterHook</varname>, probably confusing any
      extension using that hook.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: REL_16_STABLE [db69101a1] 2023-12-14 09:59:47 +0100
Branch: REL_15_STABLE [8dd70828b] 2023-12-14 09:59:52 +0100
-->
     <para>
      Fix reporting of I/O timing data in <literal>EXPLAIN
      (BUFFERS)</literal> (Michael Paquier)
     </para>

     <para>
      The numbers labeled as <quote>shared/local</quote> actually refer
      only to shared buffers, so change that label
      to <quote>shared</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [df220714e] 2024-02-01 13:44:19 -0800
Branch: REL_16_STABLE [6d423e9ff] 2024-02-01 13:44:22 -0800
Branch: REL_15_STABLE [d493bed28] 2024-02-01 13:44:22 -0800
Branch: master [0b6517a3b] 2024-02-01 13:44:19 -0800
Branch: REL_16_STABLE [48a6bf5c4] 2024-02-01 13:44:22 -0800
Branch: REL_15_STABLE [8fa4a1ac6] 2024-02-01 13:44:23 -0800
-->
     <para>
      Ensure durability of <command>CREATE DATABASE</command> (Noah Misch)
     </para>

     <para>
      If an operating system crash occurred during or shortly
      after <command>CREATE DATABASE</command>, recovery could fail, or
      subsequent connections to the new database could fail.  If a base
      backup was taken in that window, similar problems could be observed
      when trying to use the backup.  The symptom would be that the
      database directory, <filename>PG_VERSION</filename> file, or
      <filename>pg_filenode.map</filename> file was missing or empty.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [1d35f705e] 2024-01-25 17:07:56 +0900
Branch: REL_16_STABLE [edbd1b41a] 2024-01-29 09:04:51 +0900
Branch: REL_15_STABLE [8b34cff33] 2024-01-29 09:04:55 +0900
-->
     <para>
      Add more <literal>LOG</literal> messages when starting and ending
      recovery from a backup (Andres Freund)
     </para>

     <para>
      This change provides additional information in the postmaster log
      that may be useful for diagnosing recovery problems.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [8a7cbfce1] 2023-12-12 17:05:18 +0100
Branch: REL_16_STABLE [0e2c05af9] 2023-12-12 17:05:27 +0100
Branch: REL_15_STABLE [f5d8f59ca] 2023-12-12 17:05:29 +0100
Branch: REL_14_STABLE [0cfd3ddfe] 2023-12-12 17:05:33 +0100
Branch: REL_13_STABLE [d99294e77] 2023-12-12 17:05:34 +0100
Branch: REL_12_STABLE [4d45ecc92] 2023-12-12 17:05:36 +0100
-->
     <para>
      Prevent standby servers from incorrectly processing dead index
      tuples during subtransactions (Fei Changhong)
     </para>

     <para>
      The <structfield>startedInRecovery</structfield> flag was not
      correctly set for a subtransaction.  This affects only processing of
      dead index tuples.  It could allow a query in a subtransaction to
      ignore index entries that it should return (if they are already dead
      on the primary server, but not dead to the standby transaction), or
      to prematurely mark index entries as dead that are not yet dead on
      the primary.  It is not clear that the latter case has any serious
      consequences, but it's not the intended behavior.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [cd7f19da3] 2023-12-08 16:09:03 +1300
Branch: REL_16_STABLE [8ca56620c] 2023-12-08 16:10:23 +1300
Branch: REL_15_STABLE [b9f687f5a] 2023-12-08 16:11:12 +1300
-->
     <para>
      Fix integer overflow hazard in checking whether a record will fit
      into the WAL decoding buffer (Thomas Munro)
     </para>

     <para>
      This bug appears to be only latent except when running a
      32-bit <productname>PostgreSQL</productname> build on a 64-bit
      platform.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [8d7d2197f] 2023-12-11 08:50:43 +0530
Branch: REL_16_STABLE [01cc92fa6] 2023-12-11 08:36:17 +0530
Branch: REL_15_STABLE [332b43063] 2023-12-11 08:23:33 +0530
Branch: REL_14_STABLE [3f86867d5] 2023-12-11 08:10:46 +0530
Branch: REL_13_STABLE [b47ea4f3d] 2023-12-11 07:57:50 +0530
Branch: REL_12_STABLE [e81e617f3] 2023-12-11 07:45:45 +0530
-->
     <para>
      Fix deadlock between a logical replication apply worker, its
      tablesync worker, and a session process trying to alter the
      subscription (Shlok Kyal)
     </para>

     <para>
      One edge of the deadlock loop did not involve a lock wait, so the
      deadlock went undetected and would persist until manual
      intervention.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [360392fa2] 2023-11-27 08:49:55 +0530
Branch: REL_16_STABLE [d7ca9209c] 2023-11-27 08:59:12 +0530
Branch: REL_15_STABLE [a77fb8c68] 2023-11-27 09:14:17 +0530
-->
     <para>
      Ensure that column default values are correctly transmitted by
      the <application>pgoutput</application> logical replication plugin
      (Nikhil Benesch)
     </para>

     <para>
      <command>ALTER TABLE ADD COLUMN</command> with a constant default
      value for the new column avoids rewriting existing tuples, instead
      expecting that reading code will insert the correct default into a
      tuple that lacks that column.  If replication was subsequently
      initiated on the table, <application>pgoutput</application> would
      transmit NULL instead of the correct default for such a column,
      causing incorrect replication on the subscriber.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [eeb0ebad7] 2023-11-22 11:44:14 +0530
Branch: REL_16_STABLE [1b6da28e0] 2023-11-22 11:27:38 +0530
Branch: REL_15_STABLE [57aae65ae] 2023-11-22 11:14:35 +0530
-->
     <para>
      Fix failure of logical replication's initial sync for a table with
      no columns (Vignesh C)
     </para>

     <para>
      This case generated an improperly-formatted <command>COPY</command>
      command.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: REL_15_STABLE [b793a416b] 2024-01-29 10:42:41 +0530
Branch: REL_14_STABLE [6b77048e5] 2024-01-29 10:34:03 +0530
-->
     <para>
      Prevent examining system catalogs with the wrong snapshot during
      logical decoding (Fei Changhong)
     </para>

     <para>
      If decoding begins partway through a transaction that modifies
      system catalogs, the decoder may not recognize that, causing it to
      fail to treat that transaction as in-progress for catalog lookups.
      This fix deals with the case that a top-level transaction is already
      marked as containing catalog changes, but its subtransaction(s) are
      not.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [aaf09c592] 2024-01-03 17:40:41 -0500
Branch: REL_16_STABLE [fb464a1ae] 2024-01-03 17:40:38 -0500
Branch: REL_15_STABLE [a0d016393] 2024-01-03 17:40:38 -0500
Branch: REL_14_STABLE [9b042e27e] 2024-01-03 17:40:38 -0500
Branch: REL_13_STABLE [5810888c5] 2024-01-03 17:40:38 -0500
Branch: REL_12_STABLE [c20f2aab6] 2024-01-03 17:40:38 -0500
-->
     <para>
      Return the correct status code when a new client disconnects without
      responding to the server's password challenge (Liu Lang, Tom Lane)
     </para>

     <para>
      In some cases we'd treat this as a loggable error, which was not the
      intention and tends to create log spam, since common clients
      like <application>psql</application> frequently do this.  It may
      also confuse extensions that
      use <varname>ClientAuthentication_hook</varname>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [c82207a54] 2023-11-28 12:34:03 -0500
Branch: REL_16_STABLE [efa8f6064] 2023-11-28 12:34:03 -0500
Branch: REL_15_STABLE [5dd30bb54] 2023-11-28 12:34:03 -0500
Branch: REL_14_STABLE [b2b1f1288] 2023-11-28 12:34:03 -0500
Branch: REL_13_STABLE [efb804682] 2023-11-28 12:34:03 -0500
Branch: REL_12_STABLE [0bd682246] 2023-11-28 12:34:03 -0500
-->
     <para>
      Fix incompatibility with <application>OpenSSL</application> 3.2
      (Tristan Partin, Bo Andreson)
     </para>

     <para>
      Use the BIO <quote>app_data</quote> field for our private storage,
      instead of assuming it's okay to use the <quote>data</quote> field.
      This mistake didn't cause problems before, but with 3.2 it leads
      to crashes and complaints about double frees.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [0a5c46a7a] 2023-12-11 11:51:56 -0500
Branch: REL_16_STABLE [ebbd499d4] 2023-12-11 11:51:56 -0500
Branch: REL_15_STABLE [551d4b28e] 2023-12-11 11:51:56 -0500
Branch: REL_14_STABLE [07ce24326] 2023-12-11 11:51:56 -0500
Branch: REL_13_STABLE [87b46ad90] 2023-12-11 11:51:56 -0500
Branch: REL_12_STABLE [271d24f31] 2023-12-11 11:51:56 -0500
-->
     <para>
      Be more wary about <application>OpenSSL</application> not
      setting <varname>errno</varname> on error (Tom Lane)
     </para>

     <para>
      If <varname>errno</varname> isn't set, assume the cause of the
      reported failure is read EOF.  This fixes rare cases of strange
      error reports like <quote>could not accept SSL connection:
      Success</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: REL_16_STABLE [501cfd07d] 2023-11-23 13:30:13 +0200
Branch: REL_15_STABLE [481d7d1c0] 2023-11-23 13:31:51 +0200
Branch: REL_14_STABLE [555276f85] 2023-11-23 13:31:57 +0200
-->
     <para>
      Fix file descriptor leakage when a foreign data
      wrapper's <function>ForeignAsyncRequest</function> function fails
      (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [e4e63cd98] 2024-02-02 15:34:29 -0500
Branch: REL_16_STABLE [a15378100] 2024-02-02 15:34:29 -0500
Branch: REL_15_STABLE [3766b8b64] 2024-02-02 15:34:29 -0500
Branch: REL_14_STABLE [329b96882] 2024-02-02 15:34:29 -0500
Branch: REL_13_STABLE [29df29dad] 2024-02-02 15:34:29 -0500
Branch: REL_12_STABLE [4493bfb70] 2024-02-02 15:34:29 -0500
-->
     <para>
      Report <systemitem>ENOMEM</systemitem> errors from file-related system
      calls as <literal>ERRCODE_OUT_OF_MEMORY</literal>,
      not <literal>ERRCODE_INTERNAL_ERROR</literal> (Alexander Kuzmenkov)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [57b440ec1] 2024-01-18 16:11:16 -0500
Branch: REL_16_STABLE [00f941356] 2024-01-18 16:10:57 -0500
Branch: REL_15_STABLE [de2d393a8] 2024-01-18 16:10:57 -0500
Branch: REL_14_STABLE [b7e8f27d1] 2024-01-18 16:10:57 -0500
-->
     <para>
      In <application>PL/pgSQL</application>, support SQL commands that
      are <command>CREATE FUNCTION</command>/<command>CREATE
      PROCEDURE</command> with SQL-standard bodies (Tom Lane)
     </para>

     <para>
      Previously, such cases failed with parsing errors due to the
      semicolon(s) appearing in the function body.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [b8ba7344e] 2023-12-05 12:43:24 +0100
Branch: REL_16_STABLE [878aa41f8] 2023-12-05 12:43:24 +0100
Branch: REL_15_STABLE [1171c6e74] 2023-12-05 12:43:24 +0100
Branch: REL_14_STABLE [f89681816] 2023-12-05 12:43:24 +0100
Branch: master [c6605c1bd] 2024-01-16 12:27:52 +0100
Branch: REL_16_STABLE [39aab1108] 2024-01-16 12:27:52 +0100
Branch: REL_15_STABLE [2b656cbd2] 2024-01-16 12:27:52 +0100
Branch: REL_14_STABLE [7a7c8c98a] 2024-01-16 12:27:52 +0100
-->
     <para>
      Fix <application>libpq</application>'s
      handling of errors in pipelines (&Aacute;lvaro Herrera)
     </para>

     <para>
      The pipeline state could get out of sync if an error is returned
      for reasons other than a query problem (for example, if the
      connection is lost).  Potentially this would lead to a busy-loop in
      the calling application.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [1a5594b95] 2023-11-08 16:44:08 +0100
Branch: REL_16_STABLE [42f832685] 2023-11-08 16:44:08 +0100
Branch: REL_15_STABLE [0e28091d5] 2023-11-08 16:44:08 +0100
Branch: REL_14_STABLE [99fa98766] 2023-11-08 16:44:08 +0100
-->
     <para>
      Make <application>libpq</application>'s
      <function>PQsendFlushRequest()</function> function flush the client
      output buffer under the same rules as
      other <literal>PQsend</literal> functions (Jelte Fennema-Nio)
     </para>

     <para>
      In pipeline mode, it may still be necessary to
      call <function>PQflush()</function> as well; but this change removes
      some inconsistency.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [01eca6a91] 2023-11-27 09:40:44 +0900
Branch: REL_16_STABLE [8984480b5] 2023-11-27 09:40:49 +0900
Branch: REL_15_STABLE [b97226815] 2023-11-27 09:40:50 +0900
Branch: REL_14_STABLE [c3b79223f] 2023-11-27 09:40:52 +0900
Branch: REL_13_STABLE [09f680d11] 2023-11-27 09:40:55 +0900
Branch: REL_12_STABLE [0217a7444] 2023-11-27 09:40:57 +0900
-->
     <para>
      Avoid race condition when <application>libpq</application>
      initializes OpenSSL support concurrently in two different threads
      (Willi Mann, Michael Paquier)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d053a879b] 2023-11-23 13:30:18 -0500
Branch: REL_16_STABLE [85eb77185] 2023-11-23 13:30:18 -0500
Branch: REL_15_STABLE [a50053777] 2023-11-23 13:30:18 -0500
Branch: REL_14_STABLE [8f23e6a45] 2023-11-23 13:30:18 -0500
Branch: REL_13_STABLE [5abdfd88f] 2023-11-23 13:30:18 -0500
Branch: REL_12_STABLE [18fad508b] 2023-11-23 13:30:19 -0500
-->
     <para>
      Fix timing-dependent failure in GSSAPI data transmission (Tom Lane)
     </para>

     <para>
      When using GSSAPI encryption in non-blocking
      mode, <application>libpq</application> sometimes failed
      with <quote>GSSAPI caller failed to retransmit all data needing to
      be retried</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [a70f2a57f] 2023-11-13 17:04:26 -0500
Branch: REL_16_STABLE [64d2467fc] 2023-11-13 17:04:26 -0500
Branch: REL_15_STABLE [f15147df6] 2023-11-13 17:04:10 -0500
Branch: REL_14_STABLE [15439205d] 2023-11-13 17:04:10 -0500
Branch: REL_13_STABLE [9e08789d4] 2023-11-13 17:04:10 -0500
Branch: REL_12_STABLE [891afa84c] 2023-11-13 17:04:10 -0500
Branch: master [1e3f461e8] 2023-11-14 00:31:39 -0500
Branch: REL_16_STABLE [f1674ac6b] 2023-11-14 00:31:39 -0500
Branch: REL_15_STABLE [63c1b4d88] 2023-11-14 00:31:39 -0500
Branch: REL_14_STABLE [c532be99d] 2023-11-14 00:31:39 -0500
Branch: REL_13_STABLE [7836f6838] 2023-11-14 00:31:39 -0500
Branch: REL_12_STABLE [e43790342] 2023-11-14 00:31:39 -0500
-->
     <para>
      In <application>pg_dump</application>, don't dump RLS policies or
      security labels for extension member objects (Tom Lane, Jacob
      Champion)
     </para>

     <para>
      Previously, commands would be included in the dump to set these
      properties, which is really incorrect since they should be
      considered as internal affairs of the extension.  Moreover, the
      restoring user might not have adequate privilege to set them, and
      indeed the dumping user might not have enough privilege to dump them
      (since dumping RLS policies requires acquiring lock on their table).
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [7418767f1] 2023-12-29 10:57:11 -0500
Branch: REL_16_STABLE [b2c9936a7] 2023-12-29 10:57:11 -0500
Branch: REL_15_STABLE [1e0841426] 2023-12-29 10:57:11 -0500
Branch: REL_14_STABLE [aa2e323ee] 2023-12-29 10:57:11 -0500
Branch: REL_13_STABLE [8d4e8b746] 2023-12-29 10:57:11 -0500
Branch: REL_12_STABLE [69d7edb06] 2023-12-29 10:57:11 -0500
-->
     <para>
      In <application>pg_dump</application>, don't dump an extended
      statistics object if its underlying table isn't being dumped
      (Rian McGuire, Tom Lane)
     </para>

     <para>
      This conforms to the behavior for other dependent objects such as
      indexes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [49f7c6c44] 2024-01-22 17:48:30 +0100
Branch: REL_16_STABLE [07b53de70] 2024-01-22 17:48:30 +0100
Branch: REL_15_STABLE [3fd36be52] 2024-01-22 17:48:30 +0100
Branch: REL_14_STABLE [85ecff891] 2024-01-22 17:48:30 +0100
-->
     <para>
      Make it an error for a <application>pgbench</application> script to
      end with an open pipeline (Anthonin Bonnefoy)
     </para>

     <para>
      Previously, <application>pgbench</application> would behave oddly if
      a <command>\startpipeline</command> command lacked a
      matching <command>\endpipeline</command>.  This seems like a
      scripting mistake rather than a case
      that <application>pgbench</application> needs to handle nicely, so
      throw an error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [9034a2d51] 2024-01-07 15:19:50 -0500
Branch: REL_16_STABLE [cf6f802bf] 2024-01-07 15:19:50 -0500
Branch: REL_15_STABLE [940ab02b5] 2024-01-07 15:19:50 -0500
Branch: REL_14_STABLE [90e8b86fc] 2024-01-07 15:19:50 -0500
Branch: REL_13_STABLE [9e7432faf] 2024-01-07 15:19:50 -0500
Branch: REL_12_STABLE [1c7443521] 2024-01-07 15:19:50 -0500
-->
     <para>
      Fix crash in <filename>contrib/intarray</filename> if an array with
      an element equal to <literal>INT_MAX</literal> is inserted into
      a <literal>gist__int_ops</literal> index
      (Alexander Lakhin, Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [208470136] 2023-12-19 18:19:05 +0900
Branch: REL_16_STABLE [7f68b1462] 2023-12-19 18:19:15 +0900
Branch: REL_15_STABLE [2e08440d6] 2023-12-19 18:19:16 +0900
Branch: REL_14_STABLE [bfbe4a146] 2023-12-19 18:19:18 +0900
Branch: REL_13_STABLE [586c6a091] 2023-12-19 18:19:19 +0900
Branch: REL_12_STABLE [f610d4f11] 2023-12-19 18:19:21 +0900
-->
     <para>
      Report a better error
      when <filename>contrib/pageinspect</filename>'s
      <function>hash_bitmap_info()</function> function is applied to a
      partitioned hash index (Alexander Lakhin, Michael Paquier)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [a8dd62ef4] 2023-12-19 15:20:39 +0900
Branch: REL_16_STABLE [b4948ed66] 2023-12-19 15:20:48 +0900
Branch: REL_15_STABLE [b745f1680] 2023-12-19 15:20:50 +0900
Branch: REL_14_STABLE [f729fdab4] 2023-12-19 15:20:52 +0900
Branch: REL_13_STABLE [b4c1d255c] 2023-12-19 15:20:53 +0900
Branch: REL_12_STABLE [bd2d3c928] 2023-12-19 15:20:55 +0900
-->
     <para>
      Report a better error
      when <filename>contrib/pgstattuple</filename>'s
      <function>pgstathashindex()</function> function is applied to a
      partitioned hash index (Alexander Lakhin)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [9886744a3] 2024-01-10 10:41:57 +0900
Branch: REL_16_STABLE [714bfb781] 2024-01-12 13:53:07 +0900
Branch: REL_15_STABLE [33d1be06a] 2024-01-12 13:53:10 +0900
Branch: REL_14_STABLE [4c6944223] 2024-01-12 13:53:12 +0900
Branch: REL_13_STABLE [c8aab699e] 2024-01-12 13:53:13 +0900
Branch: REL_12_STABLE [9e70e6564] 2024-01-12 13:53:14 +0900
Branch: master [b83747a8a] 2024-01-11 10:39:58 +0900
Branch: REL_16_STABLE [506c77f9b] 2024-01-12 13:59:51 +0900
Branch: REL_15_STABLE [7e7d827f5] 2024-01-12 13:59:58 +0900
Branch: REL_14_STABLE [4a7475e9a] 2024-01-12 14:00:00 +0900
Branch: REL_13_STABLE [2a0eb80ae] 2024-01-12 14:00:01 +0900
Branch: REL_12_STABLE [e50a52b2b] 2024-01-12 14:00:02 +0900
-->
     <para>
      On Windows, suppress autorun options when launching subprocesses
      in <application>pg_ctl</application>
      and <application>pg_regress</application> (Kyotaro Horiguchi)
     </para>

     <para>
      When launching a child process via <filename>cmd.exe</filename>,
      pass the <option>/D</option> flag to prevent executing any autorun
      commands specified in the registry.  This avoids possibly-surprising
      side effects.
     </para>
    </listitem>

    <listitem>
<!--
Author: Nathan Bossart <nathan@postgresql.org>
Branch: master [97287bdfa] 2024-01-29 12:08:57 -0600
Branch: REL_16_STABLE [1b924a86e] 2024-01-29 12:09:03 -0600
Branch: REL_15_STABLE [3726c1cb0] 2024-01-29 12:09:08 -0600
-->
     <para>
      Move <function>is_valid_ascii()</function>
      from <filename>mb/pg_wchar.h</filename>
      to <filename>utils/ascii.h</filename> (Jubilee Young)
     </para>

     <para>
      This change avoids the need to
      include <filename>&lt;simd.h&gt;</filename>
      in <filename>pg_wchar.h</filename>, which was causing problems for
      some third-party code.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [400928b83] 2024-01-29 12:06:13 -0500
Branch: REL_16_STABLE [e02fea093] 2024-01-29 12:06:07 -0500
Branch: REL_15_STABLE [3f8ac13b1] 2024-01-29 12:06:07 -0500
Branch: REL_14_STABLE [29e25a6b1] 2024-01-29 12:06:08 -0500
Branch: REL_13_STABLE [7c53b1977] 2024-01-29 12:06:08 -0500
Branch: REL_12_STABLE [b2fd1dab9] 2024-01-29 12:06:08 -0500
-->
     <para>
      Fix compilation failures with <application>libxml2</application>
      version 2.12.0 and later (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [7636725b9] 2023-12-06 14:10:39 +0900
Branch: REL_16_STABLE [6248a2bb9] 2023-12-06 14:11:41 +0900
Branch: REL_15_STABLE [87ed81a87] 2023-12-06 14:11:44 +0900
Branch: REL_14_STABLE [c49b6cab1] 2023-12-06 14:11:46 +0900
Branch: REL_13_STABLE [7d5a74033] 2023-12-06 14:11:47 +0900
-->
     <para>
      Fix compilation failure of <literal>WAL_DEBUG</literal> code on
      Windows (Bharath Rupireddy)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [58c3151bb] 2023-12-26 16:16:37 -0500
Branch: REL_16_STABLE [b0115e7e2] 2023-12-26 16:16:29 -0500
Branch: REL_15_STABLE [5f8d6d709] 2023-12-26 16:16:29 -0500
Branch: REL_14_STABLE [7978eee02] 2023-12-26 16:16:29 -0500
Branch: REL_13_STABLE [48e797164] 2023-12-26 16:16:29 -0500
Branch: REL_12_STABLE [1613e2fae] 2023-12-26 16:16:29 -0500
Branch: master [a46972e30] 2023-12-26 17:03:02 -0500
Branch: REL_16_STABLE [c72049dbc] 2023-12-26 17:03:24 -0500
-->
     <para>
      Suppress compiler warnings from Python's header files
      (Peter Eisentraut, Tom Lane)
     </para>

     <para>
      Our preferred compiler options provoke warnings about constructs
      appearing in recent versions of Python's header files.  When using
      <application>gcc</application>, we can suppress these warnings with
      a pragma.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [d282e88e5] 2024-01-25 13:44:54 +1300
Branch: REL_16_STABLE [60ba7cae7] 2024-01-25 13:45:22 +1300
Branch: REL_15_STABLE [67f7aaa38] 2024-01-25 13:46:07 +1300
Branch: REL_14_STABLE [364283c92] 2024-01-25 13:46:38 +1300
Branch: REL_13_STABLE [70a82f40a] 2024-01-25 13:47:20 +1300
Branch: REL_12_STABLE [d060cb658] 2024-01-25 13:47:35 +1300
-->
     <para>
      Avoid deprecation warning when compiling with LLVM 18 (Thomas Munro)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [272a7c303] 2024-02-01 15:57:53 -0500
Branch: REL_16_STABLE [b4fb76fb5] 2024-02-01 15:57:53 -0500
Branch: REL_15_STABLE [970b1aeeb] 2024-02-01 15:57:53 -0500
Branch: REL_14_STABLE [35b8b2c4a] 2024-02-01 15:57:53 -0500
Branch: REL_13_STABLE [ac3afc3f7] 2024-02-01 15:57:53 -0500
Branch: REL_12_STABLE [b59ae79b7] 2024-02-01 15:57:53 -0500
-->
     <para>
      Update time zone data files to <application>tzdata</application>
      release 2024a for DST law changes in Greenland, Kazakhstan, and
      Palestine, plus corrections for the Antarctic stations Casey and
      Vostok.  Also historical corrections for Vietnam, Toronto, and
      Miquelon.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15-5">
  <title>Release 15.5</title>

  <formalpara>
  <title>Release date:</title>
  <para>2023-11-09</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.4.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.5</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, several mistakes have been discovered that could lead to
    certain types of indexes yielding wrong search results or being
    unnecessarily inefficient.  It is advisable
    to <command>REINDEX</command> potentially-affected indexes after
    installing this update.  See the fourth through seventh changelog
    entries below.
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 15.4,
    see <xref linkend="release-15-4"/>.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [3b0776fde] 2023-11-06 10:38:00 -0500
Branch: REL_16_STABLE [d3d1e2509] 2023-11-06 10:38:00 -0500
Branch: REL_15_STABLE [4f4a422fb] 2023-11-06 10:38:00 -0500
Branch: REL_14_STABLE [9146d0d65] 2023-11-06 10:38:00 -0500
Branch: REL_13_STABLE [d3de70fdb] 2023-11-06 10:38:00 -0500
Branch: REL_12_STABLE [e911afd09] 2023-11-06 10:38:00 -0500
Branch: REL_11_STABLE [8c6633f4d] 2023-11-06 10:38:00 -0500
-->
     <para>
      Fix handling of unknown-type arguments
      in <literal>DISTINCT</literal> <type>"any"</type> aggregate
      functions (Tom Lane)
     </para>

     <para>
      This error led to a <type>text</type>-type value being interpreted
      as an <type>unknown</type>-type value (that is, a zero-terminated
      string) at runtime.  This could result in disclosure of server
      memory following the <type>text</type> value.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks Jingzhou Fu
      for reporting this problem.
      (CVE-2023-5868)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [18b585155] 2023-11-06 10:56:43 -0500
Branch: REL_16_STABLE [e24daa94b] 2023-11-06 10:56:43 -0500
Branch: REL_15_STABLE [3bc6bc3ee] 2023-11-06 10:56:43 -0500
Branch: REL_14_STABLE [edc0a8d82] 2023-11-06 10:56:43 -0500
Branch: REL_13_STABLE [26c599beb] 2023-11-06 10:56:43 -0500
Branch: REL_12_STABLE [d267cea24] 2023-11-06 10:56:43 -0500
Branch: REL_11_STABLE [c48008f59] 2023-11-06 10:56:43 -0500
-->
     <para>
      Detect integer overflow while computing new array dimensions
      (Tom Lane)
     </para>

     <para>
      When assigning new elements to array subscripts that are outside the
      current array bounds, an undetected integer overflow could occur in
      edge cases.  Memory stomps that are potentially exploitable for
      arbitrary code execution are possible, and so is disclosure of
      server memory.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks Pedro
      Gallegos for reporting this problem.
      (CVE-2023-5869)
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [3a9b18b30] 2023-11-06 06:14:13 -0800
Branch: REL_16_STABLE [785412731] 2023-11-06 06:14:16 -0800
Branch: REL_15_STABLE [595c988c9] 2023-11-06 06:14:16 -0800
Branch: REL_14_STABLE [508acb901] 2023-11-06 06:14:17 -0800
Branch: REL_13_STABLE [28b609550] 2023-11-06 06:14:17 -0800
Branch: REL_12_STABLE [2893f2f40] 2023-11-06 06:14:17 -0800
Branch: REL_11_STABLE [e082734c8] 2023-11-06 06:14:18 -0800
Author: Noah Misch <noah@leadboat.com>
Branch: master [b72de09a1] 2023-11-06 06:14:13 -0800
Branch: REL_16_STABLE [2c3c5ec49] 2023-11-06 06:14:16 -0800
Branch: REL_15_STABLE [fbc371909] 2023-11-06 06:14:16 -0800
Branch: REL_14_STABLE [ecd5d240c] 2023-11-06 06:14:17 -0800
Branch: REL_13_STABLE [2c7a2a00a] 2023-11-06 06:14:17 -0800
Branch: REL_12_STABLE [7bbf4d037] 2023-11-06 06:14:17 -0800
Branch: REL_11_STABLE [a27be40c1] 2023-11-06 06:14:18 -0800
-->
     <para>
      Prevent the <literal>pg_signal_backend</literal> role from
      signalling background workers and autovacuum processes
      (Noah Misch, Jelte Fennema-Nio)
     </para>

     <para>
      The documentation says that <literal>pg_signal_backend</literal>
      cannot issue signals to superuser-owned processes.  It was able to
      signal these background processes, though, because they advertise a
      role OID of zero.  Treat that as indicating superuser ownership.
      The security implications of cancelling one of these process types
      are fairly small so far as the core code goes (we'll just start
      another one), but extensions might add background workers that are
      more vulnerable.
     </para>

     <para>
      Also ensure that the <varname>is_superuser</varname> parameter is
      set correctly in such processes.  No specific security consequences
      are known for that oversight, but it might be significant for some
      extensions.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks
      Hemanth Sandrana and Mahendrakar Srinivasarao
      for reporting this problem.
      (CVE-2023-5870)
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [28d3c2ddc] 2023-09-26 14:14:49 +0300
Branch: REL_16_STABLE [d7f521325] 2023-09-26 14:15:01 +0300
Branch: REL_15_STABLE [c3c284b37] 2023-09-26 14:15:19 +0300
Branch: REL_14_STABLE [747cef5a5] 2023-09-26 14:15:28 +0300
Branch: REL_13_STABLE [db7394d4d] 2023-09-26 14:15:35 +0300
Branch: REL_12_STABLE [26917ebea] 2023-09-26 15:41:44 +0300
-->
     <para>
      Fix misbehavior during recursive page split in GiST index build
      (Heikki Linnakangas)
     </para>

     <para>
      Fix a case where the location of a page downlink was incorrectly
      tracked, and introduce some logic to allow recovering from such
      situations rather than silently doing the wrong thing.  This error
      could result in incorrect answers from subsequent index searches.
      It may be advisable to reindex all GiST indexes after installing
      this update.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [5f27b5f84] 2023-10-14 16:33:51 -0700
Branch: REL_16_STABLE [bf1c21c4f] 2023-10-14 16:33:54 -0700
Branch: REL_15_STABLE [782be0f71] 2023-10-14 16:33:54 -0700
Branch: REL_14_STABLE [0834df909] 2023-10-14 16:33:54 -0700
Branch: REL_13_STABLE [6fd1dbdb2] 2023-10-14 16:33:54 -0700
-->
     <para>
      Prevent de-duplication of btree index entries
      for <type>interval</type> columns (Noah Misch)
     </para>

     <para>
      There are <type>interval</type> values that are distinguishable but
      compare equal, for example <literal>24:00:00</literal>
      and <literal>1 day</literal>.  This breaks assumptions made by btree
      de-duplication, so <type>interval</type> columns need to be excluded
      from de-duplication.  This oversight can cause incorrect results
      from index-only scans.  Moreover, after
      updating <application>amcheck</application> will report an error for
      almost all such indexes.  Users should reindex any btree indexes
      on <type>interval</type> columns.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [394d51731] 2023-10-27 18:15:37 +0200
Branch: REL_16_STABLE [e7965226d] 2023-10-27 18:28:19 +0200
Branch: REL_15_STABLE [088233f8d] 2023-10-27 18:37:59 +0200
Branch: REL_14_STABLE [d1740e169] 2023-10-27 18:46:49 +0200
-->
     <para>
      Process <type>date</type> values more sanely in
      BRIN <literal>datetime_minmax_multi_ops</literal> indexes
      (Tomas Vondra)
     </para>

     <para>
      The distance calculation for dates was backward, causing poor
      decisions about which entries to merge.  The index still produces
      correct results, but is much less efficient than it should be.
      Reindexing BRIN <literal>minmax_multi</literal> indexes
      on <type>date</type> columns is advisable.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [8da86d62a] 2023-10-27 18:15:37 +0200
Branch: REL_16_STABLE [497fc9208] 2023-10-27 18:28:28 +0200
Branch: REL_15_STABLE [d04a9283b] 2023-10-27 18:38:02 +0200
Branch: REL_14_STABLE [52c934cc1] 2023-10-27 18:46:53 +0200
Branch: master [b5489b75c] 2023-10-27 18:15:37 +0200
Branch: REL_16_STABLE [0635fe02b] 2023-10-27 18:28:04 +0200
Branch: REL_15_STABLE [daa7b0d7c] 2023-10-27 18:37:56 +0200
Branch: REL_14_STABLE [90c4da6d4] 2023-10-27 18:46:46 +0200
-->
     <para>
      Process large <type>timestamp</type> and <type>timestamptz</type>
      values more sanely in
      BRIN <literal>datetime_minmax_multi_ops</literal> indexes
      (Tomas Vondra)
     </para>

     <para>
      Infinities were mistakenly treated as having distance zero rather
      than a large distance from other values, causing poor decisions
      about which entries to merge.  Also, finite-but-very-large values
      (near the endpoints of the representable timestamp range) could
      result in internal overflows, again causing poor decisions.  The
      index still produces correct results, but is much less efficient
      than it should be.  Reindexing BRIN <literal>minmax_multi</literal>
      indexes on <type>timestamp</type> and <type>timestamptz</type>
      columns is advisable if the column contains, or has contained,
      infinities or large finite values.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [c6cf6d353] 2023-10-27 18:15:37 +0200
Branch: REL_16_STABLE [924e0e2ee] 2023-10-27 18:28:37 +0200
Branch: REL_15_STABLE [2fbb2fcb0] 2023-10-27 18:38:05 +0200
Branch: REL_14_STABLE [0fa73c5cd] 2023-10-27 18:46:56 +0200
-->
     <para>
      Avoid calculation overflows in
      BRIN <literal>interval_minmax_multi_ops</literal> indexes with
      extreme interval values (Tomas Vondra)
     </para>

     <para>
      This bug might have caused unexpected failures while trying to
      insert large interval values into such an index.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [d9e46dfb7] 2023-10-13 01:12:31 +1300
Branch: REL_16_STABLE [595db9e9c] 2023-10-13 01:13:07 +1300
Branch: REL_15_STABLE [1e81d3e6e] 2023-10-13 01:13:36 +1300
Branch: REL_14_STABLE [d26f33c32] 2023-10-13 01:13:59 +1300
Branch: REL_13_STABLE [890a73ba3] 2023-10-13 01:14:22 +1300
Branch: REL_12_STABLE [3cc0c2517] 2023-10-13 01:14:46 +1300
Branch: REL_11_STABLE [2759924f3] 2023-10-13 01:15:12 +1300
Author: David Rowley <drowley@postgresql.org>
Branch: master [f0c409d9c] 2023-10-12 19:50:38 +1300
Branch: REL_16_STABLE [6352f1627] 2023-10-12 19:51:26 +1300
Branch: REL_15_STABLE [916adc7c5] 2023-10-12 19:52:05 +1300
Branch: REL_14_STABLE [cd15bff48] 2023-10-12 19:52:31 +1300
Branch: REL_13_STABLE [45c4b2a3a] 2023-10-12 19:52:58 +1300
Branch: REL_12_STABLE [cd259de50] 2023-10-12 19:53:23 +1300
Branch: REL_11_STABLE [07f261b31] 2023-10-12 19:53:50 +1300
-->
     <para>
      Fix partition step generation and runtime partition pruning for
      hash-partitioned tables with multiple partition keys (David Rowley)
     </para>

     <para>
      Some cases involving an <literal>IS NULL</literal> condition on one
      of the partition keys could result in a crash.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [1d5caec22] 2023-09-30 10:52:21 +0100
Branch: REL_16_STABLE [6d2de076c] 2023-09-30 10:54:29 +0100
Branch: REL_15_STABLE [3c1a1af91] 2023-09-30 10:55:24 +0100
-->
     <para>
      Fix inconsistent rechecking of concurrently-updated rows
      during <command>MERGE</command> (Dean Rasheed)
     </para>

     <para>
      In <literal>READ COMMITTED</literal> mode, an update that finds that
      its target row was just updated by a concurrent transaction will
      recheck the query's <literal>WHERE</literal> conditions on the
      updated row.  <command>MERGE</command> failed to ensure that the
      proper rows of other joined tables were used during this recheck,
      possibly resulting in incorrect decisions about whether the
      newly-updated row should be updated again
      by <command>MERGE</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [387f9ed0a] 2023-10-24 14:48:33 -0400
Branch: REL_16_STABLE [b1444a09d] 2023-10-24 14:48:34 -0400
Branch: REL_15_STABLE [1268e7378] 2023-10-24 14:48:34 -0400
Branch: REL_14_STABLE [8f4a6b9e4] 2023-10-24 14:48:34 -0400
Author: Amit Langote <amitlan@postgresql.org>
Branch: master [01575ad78] 2023-10-26 11:53:56 +0900
Branch: REL_16_STABLE [178ee1d85] 2023-10-26 11:53:41 +0900
Author: Amit Langote <amitlan@postgresql.org>
Branch: master [1f06b7fc6] 2023-10-26 17:32:38 +0900
Branch: REL_16_STABLE [2bf99b48d] 2023-10-26 17:29:32 +0900
-->
     <para>
      Correctly identify the target table in an
      inherited <command>UPDATE</command>/<command>DELETE</command>/<command>MERGE</command>
      even when the parent table is excluded by constraints (Amit Langote,
      Tom Lane)
     </para>

     <para>
      If the initially-named table is excluded by constraints, but not all
      its inheritance descendants are, the first non-excluded descendant
      was identified as the primary target table.  This would lead to
      firing statement-level triggers associated with that table, rather
      than the initially-named table as should happen.  In v16, the same
      oversight could also lead to <quote>invalid perminfoindex 0 in RTE
      with relid NNNN</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Geoghegan <pg@bowt.ie>
Branch: master [714780dcd] 2023-09-28 16:29:37 -0700
Branch: REL_16_STABLE [3fa81b62e] 2023-09-28 16:29:35 -0700
Branch: REL_15_STABLE [cac37c1a1] 2023-09-28 16:29:32 -0700
Branch: REL_14_STABLE [41486c4aa] 2023-09-28 16:29:29 -0700
Branch: REL_13_STABLE [d4724e94a] 2023-09-28 16:29:27 -0700
Branch: REL_12_STABLE [b36d72cbc] 2023-09-28 16:29:24 -0700
Branch: REL_11_STABLE [7c07305e6] 2023-09-28 16:29:22 -0700
-->
     <para>
      Fix edge case in btree mark/restore processing of ScalarArrayOpExpr
      clauses (Peter Geoghegan)
     </para>

     <para>
      When restoring an indexscan to a previously marked position, the
      code could miss required setup steps if the scan had advanced
      exactly to the end of the matches for a ScalarArrayOpExpr (that is,
      an <literal>indexcol = ANY(ARRAY[])</literal>) clause.  This could
      result in missing some rows that should have been fetched.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [0b053e78b] 2023-10-05 20:30:47 +1300
Branch: REL_16_STABLE [31b2b2d72] 2023-10-05 20:31:25 +1300
Branch: REL_15_STABLE [689af6db6] 2023-10-05 20:31:51 +1300
Branch: REL_14_STABLE [e4b95b9b0] 2023-10-05 20:32:14 +1300
-->
     <para>
      Fix intra-query memory leak in Memoize execution
      (Orlov Aleksej, David Rowley)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [237f8765d] 2023-10-28 14:05:01 -0400
Branch: REL_16_STABLE [07494a0df] 2023-10-28 14:04:43 -0400
Branch: REL_15_STABLE [592cb11fb] 2023-10-28 14:04:43 -0400
Branch: REL_14_STABLE [5d7515d7d] 2023-10-28 14:04:43 -0400
Branch: REL_13_STABLE [b7684473d] 2023-10-28 14:04:43 -0400
Branch: REL_12_STABLE [d8d7f282f] 2023-10-28 14:04:43 -0400
Branch: REL_11_STABLE [7ab6971c6] 2023-10-28 14:04:43 -0400
-->
     <para>
      Fix intra-query memory leak when a set-returning function repeatedly
      returns zero rows (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f73fa5a47] 2023-09-18 14:28:17 -0400
Branch: REL_16_STABLE [ec693a3f3] 2023-09-18 14:27:47 -0400
Branch: REL_15_STABLE [95f54f0d0] 2023-09-18 14:27:47 -0400
Branch: REL_14_STABLE [9baabfa38] 2023-09-18 14:27:47 -0400
Branch: REL_13_STABLE [e6e50ada0] 2023-09-18 14:27:47 -0400
Branch: REL_12_STABLE [123337079] 2023-09-18 14:27:47 -0400
Branch: REL_11_STABLE [db00be6d7] 2023-09-18 14:27:47 -0400
-->
     <para>
      Don't crash if <function>cursor_to_xmlschema()</function> is applied
      to a non-data-returning Portal (Boyu Yang)
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [73635b6d6] 2023-10-31 16:42:08 +1300
Branch: REL_16_STABLE [a98f01c93] 2023-10-31 16:42:35 +1300
Branch: REL_15_STABLE [136068353] 2023-10-31 16:43:01 +1300
Branch: REL_14_STABLE [c5988e69f] 2023-10-31 16:43:28 +1300
Branch: REL_13_STABLE [b3c8d1d0e] 2023-10-31 16:44:00 +1300
Branch: REL_12_STABLE [efcb601d2] 2023-10-31 16:44:27 +1300
-->
     <para>
      Throw the intended error if <function>pgrowlocks()</function> is
      applied to a partitioned table (David Rowley)
     </para>

     <para>
      Previously, a not-on-point complaint <quote>only heap AM is
      supported</quote> would be raised.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [13503eb59] 2023-10-30 14:46:05 -0700
Branch: REL_16_STABLE [1a368dd3e] 2023-10-30 14:46:08 -0700
Branch: REL_15_STABLE [e633e9b13] 2023-10-30 14:46:08 -0700
Branch: REL_14_STABLE [0a7b183fd] 2023-10-30 14:46:09 -0700
Branch: REL_13_STABLE [123b0d111] 2023-10-30 14:46:09 -0700
Branch: REL_12_STABLE [975ae0553] 2023-10-30 14:46:09 -0700
Branch: REL_11_STABLE [bae063db4] 2023-10-30 14:46:09 -0700
-->
     <para>
      Handle invalid indexes more cleanly in assorted SQL functions
      (Noah Misch)
     </para>

     <para>
      Report an error if <function>pgstatindex()</function>,
      <function>pgstatginindex()</function>,
      <function>pgstathashindex()</function>,
      or <function>pgstattuple()</function> is applied to an invalid
      index.  If <function>brin_desummarize_range()</function>,
      <function>brin_summarize_new_values()</function>,
      <function>brin_summarize_range()</function>,
      or <function>gin_clean_pending_list()</function> is applied to an
      invalid index, do nothing except to report a debug-level message.
      Formerly these functions attempted to process the index, and might
      fail in strange ways depending on what the failed <command>CREATE
      INDEX</command> had left behind.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [6fde2d9a0] 2023-08-21 13:32:14 +0900
Branch: REL_16_STABLE Release: REL_16_0 [d6af45052] 2023-08-21 13:33:04 +0900
Branch: REL_15_STABLE [ad8753a3a] 2023-08-21 13:33:08 +0900
-->
     <para>
      Fix <function>pg_stat_reset_single_table_counters()</function> to do
      the right thing for a shared catalog (Masahiro Ikeda)
     </para>

     <para>
      Previously the reset would be ineffective.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [cf1c65070] 2023-09-25 11:50:28 -0400
Branch: REL_16_STABLE [8465efc1a] 2023-09-25 11:50:28 -0400
Branch: REL_15_STABLE [71bb73f60] 2023-09-25 11:50:28 -0400
Branch: REL_14_STABLE [a2799f53d] 2023-09-25 11:50:28 -0400
Branch: REL_13_STABLE [52a676ef4] 2023-09-25 11:50:28 -0400
Branch: REL_12_STABLE [4a81ed29d] 2023-09-25 11:50:28 -0400
Branch: REL_11_STABLE [0fb91ed2b] 2023-09-25 11:50:28 -0400
-->
     <para>
      Avoid premature memory allocation failure with long inputs
      to <function>to_tsvector()</function> (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5b7b38246] 2023-10-01 13:16:47 -0400
Branch: REL_16_STABLE [5c34a7374] 2023-10-01 13:16:59 -0400
Branch: REL_15_STABLE [55e188a15] 2023-10-01 13:17:06 -0400
Branch: REL_14_STABLE [cf12c4fd2] 2023-10-01 13:17:11 -0400
Branch: REL_13_STABLE [9da153846] 2023-10-01 13:17:15 -0400
Branch: REL_12_STABLE [2296bc362] 2023-10-01 13:17:20 -0400
Branch: REL_11_STABLE [7a310cae0] 2023-10-01 13:17:25 -0400
-->
     <para>
      Fix over-allocation of the constructed <type>tsvector</type>
      in <function>tsvectorrecv()</function> (Denis Erokhin)
     </para>

     <para>
      If the incoming vector includes position data, the binary receive
      function left wasted space (roughly equal to the size of the
      position data) in the finished <type>tsvector</type>.  In extreme
      cases this could lead to <quote>maximum total lexeme length
      exceeded</quote> failures for vectors that were under the length
      limit when emitted.  In any case it could lead to wasted space
      on-disk.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [ae10dbb0c] 2023-09-04 14:55:37 +0900
Branch: REL_16_STABLE Release: REL_16_0 [fac3fa0d5] 2023-09-04 14:55:49 +0900
Branch: REL_15_STABLE [88aa4a049] 2023-09-04 14:55:51 +0900
Branch: REL_14_STABLE [df11421d8] 2023-09-04 14:55:53 +0900
Branch: REL_13_STABLE [82fa8ae5f] 2023-09-04 14:55:55 +0900
Branch: REL_12_STABLE [40ddb8db9] 2023-09-04 14:55:57 +0900
Branch: REL_11_STABLE [358cd2b25] 2023-09-04 14:55:58 +0900
-->
     <para>
      Fix incorrect coding in <function>gtsvector_picksplit()</function>
      (Alexander Lakhin)
     </para>

     <para>
      This could lead to poor page-split decisions in GiST indexes
      on <type>tsvector</type> columns.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [9b103f861] 2023-10-18 20:43:27 -0400
Branch: REL_16_STABLE [cfa4eba02] 2023-10-18 20:43:17 -0400
Branch: REL_15_STABLE [985ac5ce2] 2023-10-18 20:43:17 -0400
Branch: REL_14_STABLE [2c63dc035] 2023-10-18 20:43:17 -0400
Branch: REL_13_STABLE [817669ea2] 2023-10-18 20:43:17 -0400
-->
     <para>
      Improve checks for corrupt PGLZ compressed data (Flavien Guedez)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d8a09939a] 2023-10-01 12:09:26 -0400
Branch: REL_16_STABLE [ea0e7cd6b] 2023-10-01 12:09:26 -0400
Branch: REL_15_STABLE [95fd5c89f] 2023-10-01 12:09:26 -0400
Branch: REL_14_STABLE [a715c0212] 2023-10-01 12:09:26 -0400
-->
     <para>
      In <command>COPY FROM</command>, fail cleanly when an unsupported
      encoding conversion is needed (Tom Lane)
     </para>

     <para>
      Recent refactoring accidentally removed the intended error check for
      this, such that it ended in <quote>cache lookup failed for function
      0</quote> instead of a useful error message.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [7704a1a72] 2023-11-02 11:47:33 -0400
Branch: REL_16_STABLE [82063edd4] 2023-11-02 11:47:33 -0400
Branch: REL_15_STABLE [ae33659d4] 2023-11-02 11:47:33 -0400
Branch: REL_14_STABLE [47c0f00cf] 2023-11-02 11:47:33 -0400
Branch: REL_13_STABLE [137227c6d] 2023-11-02 11:47:33 -0400
Branch: REL_12_STABLE [65810fc6d] 2023-11-02 11:47:33 -0400
Branch: REL_11_STABLE [6e1cca511] 2023-11-02 11:47:33 -0400 !! no live bug
-->
     <para>
      Avoid crash in <command>EXPLAIN</command> if a parameter marked to
      be displayed by <command>EXPLAIN</command> has a NULL boot-time
      value (Xing Guo, Aleksander Alekseev, Tom Lane)
     </para>

     <para>
      No built-in parameter fits this description, but an extension could
      define such a parameter.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [54b208f90] 2023-10-16 14:06:14 -0400
Branch: REL_16_STABLE [57e6e861d] 2023-10-16 14:06:15 -0400
Branch: REL_15_STABLE [0d1a7cd14] 2023-10-16 14:06:15 -0400
Branch: REL_14_STABLE [f6e1ee3cf] 2023-10-16 14:06:11 -0400
Branch: REL_13_STABLE [1e678175c] 2023-10-16 14:06:11 -0400
Branch: REL_12_STABLE [07eb22a77] 2023-10-16 14:06:11 -0400
Branch: REL_11_STABLE [a295684b8] 2023-10-16 14:06:12 -0400
-->
     <para>
      Ensure we have a snapshot while dropping <literal>ON COMMIT
      DROP</literal> temp tables (Tom Lane)
     </para>

     <para>
      This prevents possible misbehavior if any catalog entries for the
      temp tables have fields wide enough to require toasting (such as a
      very complex <literal>CHECK</literal> condition).
     </para>
    </listitem>

    <listitem>
<!--
Author: Nathan Bossart <nathan@postgresql.org>
Branch: master [97550c071] 2023-10-17 10:41:48 -0500
Branch: REL_16_STABLE [ee06199fc] 2023-10-17 10:41:58 -0500
Branch: REL_15_STABLE [c9265ae80] 2023-10-17 10:42:06 -0500
Branch: REL_14_STABLE [54fc9dca5] 2023-10-17 10:42:12 -0500
Branch: REL_13_STABLE [ac1dfc303] 2023-10-17 10:42:17 -0500
Branch: REL_12_STABLE [e2e169042] 2023-10-17 10:42:22 -0500
Branch: REL_11_STABLE [d0e7f95b4] 2023-10-17 10:42:28 -0500
Author: Nathan Bossart <nathan@postgresql.org>
Branch: REL_14_STABLE [555bc89c9] 2023-10-17 16:11:03 -0500
Branch: REL_13_STABLE [ebc093fa6] 2023-10-17 16:11:08 -0500
Branch: REL_12_STABLE [0e32652a7] 2023-10-17 16:11:13 -0500
Branch: REL_11_STABLE [bc322c73c] 2023-10-17 16:11:18 -0500
-->
     <para>
      Avoid improper response to shutdown signals in child processes
      just forked by <function>system()</function> (Nathan Bossart)
     </para>

     <para>
      This fix avoids a race condition in which a child process that has
      been forked off by <function>system()</function>, but hasn't yet
      exec'd the intended child program, might receive and act on a signal
      intended for the parent server process.  That would lead to
      duplicate cleanup actions being performed, which will not end well.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [63a582222] 2023-10-16 17:33:08 +1300
Branch: REL_16_STABLE [5725e4ebe] 2023-10-16 17:21:41 +1300
Branch: REL_15_STABLE [5e39884d3] 2023-10-16 17:23:02 +1300
Branch: REL_14_STABLE [dc75748a9] 2023-10-16 17:23:25 +1300
Branch: REL_13_STABLE [67060be3d] 2023-10-16 17:24:35 +1300
Branch: REL_12_STABLE [43c979086] 2023-10-16 17:25:43 +1300
-->
     <para>
      Cope with torn reads of <filename>pg_control</filename> in frontend
      programs (Thomas Munro)
     </para>

     <para>
      On some file systems, reading <filename>pg_control</filename> may
      not be an atomic action when the server concurrently writes that
      file.  This is detectable via a bad CRC.  Retry a few times to see
      if the file becomes valid before we report error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [c558e6fd9] 2023-10-16 10:43:47 +1300
Branch: REL_16_STABLE [2371432cd] 2023-10-16 10:45:05 +1300
Branch: REL_15_STABLE [606be8a35] 2023-10-16 10:45:18 +1300
Branch: REL_14_STABLE [a56fe5cf0] 2023-10-16 10:47:01 +1300
Branch: REL_13_STABLE [ae9da357b] 2023-10-16 10:47:10 +1300
Branch: REL_12_STABLE [637e86ecc] 2023-10-16 10:47:33 +1300
Branch: REL_11_STABLE [f1634c968] 2023-10-16 10:52:40 +1300
-->
     <para>
      Avoid torn reads of <filename>pg_control</filename> in relevant SQL
      functions (Thomas Munro)
     </para>

     <para>
      Acquire the appropriate lock before
      reading <filename>pg_control</filename>, to ensure we get a
      consistent view of that file.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [a956bd3fa] 2023-10-03 15:37:00 +0900
Branch: REL_16_STABLE [75f31a3f2] 2023-10-03 15:37:18 +0900
Branch: REL_15_STABLE [95e91da66] 2023-10-03 15:37:19 +0900
Branch: REL_14_STABLE [f91c87b31] 2023-10-03 15:37:21 +0900
Branch: REL_13_STABLE [a0b013625] 2023-10-03 15:37:23 +0900
Branch: REL_12_STABLE [9b3900cdb] 2023-10-03 15:37:24 +0900
-->
     <para>
      Avoid integer overflow when computing size of backend activity
      string array (Jakub Wartak)
     </para>

     <para>
      On 64-bit machines we will allow values
      of <varname>track_activity_query_size</varname> large enough to
      cause 32-bit overflow when multiplied by the allowed number of
      connections.  The code actually allocating the per-backend local
      array was careless about this though, and allocated the array
      incorrectly.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [c181f2e2b] 2023-09-30 17:03:50 +0300
Branch: REL_16_STABLE [992d2ca81] 2023-09-30 17:07:24 +0300
Branch: REL_15_STABLE [5ae245664] 2023-09-30 17:07:32 +0300
Branch: REL_14_STABLE [2d4f99bc6] 2023-09-30 17:07:37 +0300
Branch: REL_13_STABLE [ed9247cd7] 2023-09-30 17:07:41 +0300
-->
     <para>
      Fix briefly showing inconsistent progress statistics
      for <command>ANALYZE</command> on inherited tables
      (Heikki Linnakangas)
     </para>

     <para>
      The block-level counters should be reset to zero at the same time we
      update the current-relation field.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [e64c733bb] 2023-10-02 12:39:35 +0300
Branch: REL_16_STABLE [4a97a43a7] 2023-10-02 12:49:25 +0300
Branch: REL_15_STABLE [0684d1949] 2023-10-02 12:49:29 +0300
Branch: REL_14_STABLE [594001864] 2023-10-02 12:50:32 +0300
-->
     <para>
      Fix the background writer to report any WAL writes it makes to the
      statistics counters (Nazir Bilal Yavuz)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [e221c0bef] 2023-09-26 09:29:47 +0900
Branch: REL_16_STABLE [280f70221] 2023-09-26 09:30:36 +0900
Branch: REL_15_STABLE [802fcb9ed] 2023-09-26 09:30:39 +0900
-->
     <para>
      Fix confusion about forced-flush behavior
      in <function>pgstat_report_wal()</function>
      (Ryoga Yoshida, Michael Paquier)
     </para>

     <para>
      This could result in some statistics about WAL I/O being forgotten
      in a shutdown.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [dc8d72c1c] 2023-09-25 14:42:17 -0400
Branch: REL_16_STABLE [055f786ea] 2023-09-25 14:42:17 -0400
Branch: REL_15_STABLE [0e59266a5] 2023-09-25 14:41:57 -0400
Branch: REL_14_STABLE [5cff431bc] 2023-09-25 14:41:57 -0400
Branch: REL_13_STABLE [8cfc3e80b] 2023-09-25 14:41:57 -0400
Branch: REL_12_STABLE [4435a8d5d] 2023-09-25 14:41:57 -0400
Branch: REL_11_STABLE [fdc7cf73b] 2023-09-25 14:41:57 -0400
-->
     <para>
      Track the dependencies of cached <command>CALL</command> statements,
      and re-plan them when needed (Tom Lane)
     </para>

     <para>
      DDL commands, such as replacement of a function that has been
      inlined into a <command>CALL</command> argument, can create the need
      to re-plan a <command>CALL</command> that has been cached by
      PL/pgSQL.  That was not happening, leading to misbehavior or strange
      errors such as <quote>cache lookup failed</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
Branch: master [5f3aa309a] 2023-09-22 11:18:25 +0200
Branch: REL_16_STABLE [f720875a4] 2023-09-22 11:18:25 +0200
Branch: REL_15_STABLE [9dc85806d] 2023-09-22 11:18:25 +0200
-->
     <para>
      Avoid a possible pfree-a-NULL-pointer crash after an error in
      OpenSSL connection setup (Sergey Shinderuk)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [e0e492e5a] 2023-09-15 17:01:52 -0400
Branch: REL_16_STABLE [53630f12d] 2023-09-15 17:01:26 -0400
Branch: REL_15_STABLE [2679a107a] 2023-09-15 17:01:26 -0400
Branch: REL_14_STABLE [ae13f8166] 2023-09-15 17:01:26 -0400
Branch: REL_13_STABLE [c5b7f791d] 2023-09-15 17:01:26 -0400
Branch: REL_12_STABLE [d29812c0c] 2023-09-15 17:01:26 -0400
Branch: REL_11_STABLE [a374f6c61] 2023-09-15 17:01:26 -0400
-->
     <para>
      Track nesting depth correctly when
      inspecting <type>RECORD</type>-type Vars from outer query levels
      (Richard Guo)
     </para>

     <para>
      This oversight could lead to assertion failures, core dumps,
      or <quote>bogus varno</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [ee3a551e9] 2023-09-14 11:27:29 +1200
Branch: REL_16_STABLE [1a6900e58] 2023-09-14 11:27:16 +1200
Branch: REL_15_STABLE [17a3f1c34] 2023-09-14 11:27:00 +1200
Branch: REL_14_STABLE [6341cb0b0] 2023-09-14 11:27:43 +1200
-->
     <para>
      Track hash function and negator function dependencies of
      ScalarArrayOpExpr plan nodes (David Rowley)
     </para>

     <para>
      In most cases this oversight was harmless, since these functions
      would be unlikely to disappear while the node's original operator
      remains present.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [3acd0599b] 2023-09-13 14:58:22 +1200
Branch: REL_16_STABLE [f899c7f1e] 2023-09-13 15:06:07 +1200
Branch: REL_15_STABLE [a26cc0334] 2023-09-13 14:43:04 +1200
Branch: REL_14_STABLE [e2452c2a6] 2023-09-13 14:46:16 +1200
Branch: REL_13_STABLE [189097db8] 2023-09-13 14:48:00 +1200
Branch: REL_12_STABLE [012b80ad2] 2023-09-13 14:50:14 +1200
Branch: REL_11_STABLE [6ae57f190] 2023-09-13 14:52:34 +1200
-->
     <para>
      Fix error-handling bug in <type>RECORD</type> type cache management
      (Thomas Munro)
     </para>

     <para>
      An out-of-memory error occurring at just the wrong point could leave
      behind inconsistent state that would lead to an infinite loop.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: REL_15_STABLE [c7256e656] 2023-09-12 10:23:17 +0530
Branch: REL_14_STABLE [f7d25117b] 2023-09-12 10:12:51 +0530
Branch: REL_13_STABLE [c570bb4d6] 2023-09-12 09:58:50 +0530
Branch: REL_12_STABLE [7e57208ed] 2023-09-12 09:44:37 +0530
Branch: REL_11_STABLE [feb4e218e] 2023-09-12 09:36:56 +0530
-->
     <para>
      Fix assertion failure when logical decoding is retried in the same
      session after an error (Hou Zhijie)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [6b18b3fe2] 2023-10-03 10:21:44 +0900
Branch: REL_16_STABLE [a06efbc3a] 2023-10-03 10:25:12 +0900
Branch: REL_15_STABLE [afc79591d] 2023-10-03 10:25:14 +0900
Branch: REL_14_STABLE [baeb8542c] 2023-10-03 10:25:15 +0900
Branch: REL_13_STABLE [6103d2cca] 2023-10-03 10:25:17 +0900
Branch: REL_12_STABLE [22b2e6e9d] 2023-10-03 10:25:19 +0900
-->
     <para>
      Treat out-of-memory failures as fatal while reading WAL
      (Michael Paquier)
     </para>

     <para>
      Previously this would be treated as a bogus-data condition, leading
      to the conclusion that we'd reached the end of WAL, which is
      incorrect and could lead to inconsistent WAL replay.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [bae868caf] 2023-09-23 10:26:24 +1200
Branch: REL_16_STABLE [ce497f648] 2023-09-23 10:27:02 +1200
Branch: REL_15_STABLE [f4d152edd] 2023-09-23 10:28:12 +1200
Branch: REL_14_STABLE [3ce3b53d7] 2023-09-23 10:28:40 +1200
Branch: REL_13_STABLE [6606c5716] 2023-09-23 10:29:44 +1200
Branch: REL_12_STABLE [e8f3c0687] 2023-09-23 10:29:56 +1200
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [becfbdd6c] 2023-09-26 10:53:38 +1300
Branch: REL_16_STABLE [10d0591ea] 2023-09-26 10:53:52 +1300
Branch: REL_15_STABLE [99d334a18] 2023-09-26 10:54:02 +1300
Branch: REL_14_STABLE [3d413c5a7] 2023-09-26 10:54:24 +1300
Branch: REL_13_STABLE [45d1fe8b5] 2023-09-26 10:54:47 +1300
Branch: REL_12_STABLE [bde2f1847] 2023-09-26 10:59:49 +1300
-->
     <para>
      Fix possible recovery failure due to trying to allocate memory based
      on a bogus WAL record length field (Thomas Munro, Michael Paquier)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [2b8e5273e] 2023-09-04 08:04:22 +0900
Branch: REL_16_STABLE Release: REL_16_0 [c397c98ae] 2023-09-04 08:04:40 +0900
Branch: REL_15_STABLE [5a9325fdd] 2023-09-04 08:04:43 +0900
-->
     <para>
      Fix race condition in database dropping that could lead to the
      autovacuum launcher getting stuck (Andres Freund, Will Mortensen,
      Jacob Speidel)
     </para>

     <para>
      The race could lead to a statistics entry for the removed database
      remaining present, confusing the launcher's selection of which
      database to process.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [52c6c0f19] 2023-08-30 08:03:42 +0900
Branch: REL_16_STABLE Release: REL_16_0 [1d2939b6d] 2023-08-30 08:03:48 +0900
Branch: REL_15_STABLE [5180160c1] 2023-08-30 08:03:51 +0900
Branch: REL_14_STABLE [d2bd4ba30] 2023-08-30 08:03:52 +0900
Branch: REL_13_STABLE [420c8dc7e] 2023-08-30 08:03:54 +0900
-->
     <para>
      Fix datatype size confusion in logical tape management
      (Ranier Vilela)
     </para>

     <para>
      Integer overflow was possible on platforms where long is wider than
      int, although it would take a multiple-terabyte temporary file to
      cause a problem.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [5fec3c870] 2023-08-29 09:09:40 +0300
Branch: REL_16_STABLE Release: REL_16_0 [5f38ff3d0] 2023-08-29 09:10:28 +0300
Branch: REL_15_STABLE [0c1024060] 2023-08-29 09:12:35 +0300
Branch: REL_14_STABLE [3d895f95a] 2023-08-29 09:12:24 +0300
Branch: REL_13_STABLE [a01e479e8] 2023-08-29 09:12:17 +0300
-->
     <para>
      Avoid unintended close of syslogger process's stdin
      (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d8b2fcc9d] 2023-08-24 12:02:46 -0400
Branch: REL_16_STABLE Release: REL_16_0 [ba0d737ca] 2023-08-24 12:02:40 -0400
Branch: REL_15_STABLE [870085135] 2023-08-24 12:02:40 -0400
Branch: REL_14_STABLE [9b2a41db1] 2023-08-24 12:02:40 -0400
Branch: REL_13_STABLE [27566bcf3] 2023-08-24 12:02:40 -0400
Branch: REL_12_STABLE [b808dbf90] 2023-08-24 12:02:40 -0400
Branch: REL_11_STABLE [9c59f3862] 2023-08-24 12:02:40 -0400
-->
     <para>
      Avoid doing plan cache revalidation of utility statements
      that do not receive interesting processing during parse analysis
      (Tom Lane)
     </para>

     <para>
      Aside from saving a few cycles, this prevents failure after a cache
      invalidation for statements that must not set a snapshot, such
      as <command>SET TRANSACTION ISOLATION LEVEL</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andrew Dunstan <andrew@dunslane.net>
Branch: master [a68458108] 2023-08-22 15:17:05 -0400
Branch: REL_16_STABLE Release: REL_16_0 [f938acd68] 2023-08-22 15:18:19 -0400
Branch: REL_15_STABLE [75f323aa1] 2023-08-22 15:16:01 -0400
Branch: REL_14_STABLE [7f4515a58] 2023-08-22 15:15:45 -0400
Branch: REL_13_STABLE [1bb619d4d] 2023-08-22 15:15:31 -0400
Branch: REL_12_STABLE [01993ac74] 2023-08-22 15:15:18 -0400
Branch: REL_11_STABLE [2d13dab04] 2023-08-23 17:22:16 -0400
Branch: REL_11_STABLE [b915bf495] 2023-08-24 16:14:26 -0400
-->
     <para>
      Keep by-reference <structfield>attmissingval</structfield> values in
      a long-lived context while they are being used (Andrew Dunstan)
     </para>

     <para>
      This avoids possible use of dangling pointers when a tuple slot
      outlives the tuple descriptor with which its value was constructed.
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [fa2e87494] 2023-08-09 13:09:25 -0700
Branch: REL_16_STABLE Release: REL_16_0 [0d56c32c8] 2023-08-09 13:09:01 -0700
Branch: REL_15_STABLE [9f3343e40] 2023-08-07 15:12:49 -0700
Branch: REL_14_STABLE [00953f1e2] 2023-08-07 15:12:33 -0700
Branch: REL_13_STABLE [90cadfd49] 2023-08-07 15:12:09 -0700
Branch: REL_12_STABLE [a86d40227] 2023-08-07 15:11:18 -0700
Branch: REL_11_STABLE [43ba5105a] 2023-08-07 15:09:11 -0700
Branch: master [bee263b08] 2023-08-10 10:20:54 -0700
Branch: REL_16_STABLE Release: REL_16_0 [7bed93a46] 2023-08-10 10:24:58 -0700
Branch: REL_15_STABLE [1bc19dfcf] 2023-08-10 10:25:47 -0700
Branch: REL_14_STABLE [5a32af3f2] 2023-08-10 10:26:12 -0700
Branch: REL_13_STABLE [4caa9e33e] 2023-08-10 10:26:46 -0700
Branch: REL_12_STABLE [01bc42634] 2023-08-10 10:27:00 -0700
Branch: REL_11_STABLE [ae1269295] 2023-08-10 10:27:24 -0700
-->
     <para>
      Recalculate the effective value of <varname>search_path</varname>
      after <command>ALTER ROLE</command> (Jeff Davis)
     </para>

     <para>
      This ensures that after renaming a role, the meaning of the special
      string <literal>$user</literal> is re-determined.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [dab889d60] 2023-10-22 10:04:55 +1300
Branch: REL_16_STABLE [174ccda5e] 2023-10-22 10:05:22 +1300
Branch: REL_15_STABLE [f72790b29] 2023-10-22 10:05:40 +1300
Branch: REL_14_STABLE [fb9a16a1a] 2023-10-22 10:05:59 +1300
-->
     <para>
      Fix <quote>could not duplicate handle</quote> error occurring on
      Windows when <varname>min_dynamic_shared_memory</varname> is set
      above zero (Thomas Munro)
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [ef74c7197] 2023-10-10 11:01:13 -0700
Branch: REL_16_STABLE [b8963e8a2] 2023-10-10 11:01:55 -0700
Branch: REL_15_STABLE [b9bb02620] 2023-10-10 11:02:45 -0700
Branch: REL_14_STABLE [6615bb95a] 2023-10-10 11:02:56 -0700
Branch: REL_13_STABLE [469d367be] 2023-10-10 11:03:03 -0700
Branch: REL_12_STABLE [6dbd01b9f] 2023-10-10 11:03:10 -0700
Branch: REL_11_STABLE [04f0baa85] 2023-10-10 11:03:20 -0700
-->
     <para>
      Fix order of operations in <function>GenericXLogFinish</function>
      (Jeff Davis)
     </para>

     <para>
      This code violated the conditions required for crash safety by
      writing WAL before marking changed buffers dirty.  No core code uses
      this function, but extensions do (<filename>contrib/bloom</filename>
      does, for example).
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [af5b3c3d1] 2023-09-19 08:31:06 +0900
Branch: REL_16_STABLE [f171430f0] 2023-09-19 08:31:22 +0900
Branch: REL_15_STABLE [9e0ce80f3] 2023-09-19 08:31:24 +0900
Branch: REL_14_STABLE [bb65f3c1b] 2023-09-19 08:31:26 +0900
Branch: REL_13_STABLE [35e8ba3e5] 2023-09-19 08:31:27 +0900
Branch: REL_12_STABLE [99d27e55c] 2023-09-19 08:31:29 +0900
Branch: REL_11_STABLE [ddded779a] 2023-09-19 08:31:31 +0900
-->
     <para>
      Remove incorrect assertion in PL/Python exception handling
      (Alexander Lakhin)
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [37188cea0] 2023-08-22 12:50:01 -0700
Branch: REL_16_STABLE Release: REL_16_0 [501704e82] 2023-08-22 12:49:40 -0700
Branch: REL_15_STABLE [1d9976d1b] 2023-08-22 12:49:17 -0700
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [f234b8cd1] 2023-09-05 11:39:27 +0200
Branch: REL_16_STABLE Release: REL_16_0 [9ddecd406] 2023-09-05 11:39:17 +0200
Branch: REL_15_STABLE [3a788447d] 2023-09-05 11:36:55 +0200
-->
     <para>
      Fix assertion failure in <application>pg_dump</application> when
      it's asked to dump the <literal>pg_catalog</literal> schema (Peter
      Eisentraut)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [06c0c3688] 2023-10-02 13:27:58 -0400
Branch: REL_16_STABLE [aaaf8fbb6] 2023-10-02 13:27:51 -0400
Branch: REL_15_STABLE [10e705bd2] 2023-10-02 13:27:51 -0400
Branch: REL_14_STABLE [829d91cc6] 2023-10-02 13:27:51 -0400
Branch: REL_13_STABLE [295953f8b] 2023-10-02 13:27:51 -0400
Branch: REL_12_STABLE [c02f29758] 2023-10-02 13:27:51 -0400
Branch: REL_11_STABLE [4f16152d9] 2023-10-02 13:27:51 -0400
-->
     <para>
      Fix <application>pg_restore</application> so that selective restores
      will include both table-level and column-level ACLs for selected
      tables (Euler Taveira, Tom Lane)
     </para>

     <para>
      Formerly, only the table-level ACL would get restored if both types
      were present.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [2e3dc8c14] 2023-09-25 14:27:33 +0200
Branch: REL_16_STABLE [fb9ddd0fa] 2023-09-25 14:34:06 +0200
Branch: REL_15_STABLE [8845d8597] 2023-09-25 14:34:06 +0200
Branch: REL_14_STABLE [26c291a15] 2023-09-25 14:34:06 +0200
Branch: REL_13_STABLE [d3946c4f6] 2023-09-25 14:34:05 +0200
Branch: REL_12_STABLE [af9f6cd1d] 2023-09-25 14:34:05 +0200
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: REL_15_STABLE [04d2d605f] 2023-11-03 12:07:11 -0400
Branch: REL_14_STABLE [f9280e77e] 2023-11-03 12:07:25 -0400
Branch: REL_13_STABLE [2f1f49975] 2023-11-03 12:07:35 -0400
Branch: REL_12_STABLE [d3246a2ad] 2023-11-03 12:07:40 -0400
-->
     <para>
      Add logic to <application>pg_upgrade</application> to check for use
      of <type>abstime</type>, <type>reltime</type>,
      and <type>tinterval</type> data types (&Aacute;lvaro Herrera)
     </para>

     <para>
      These obsolete data types were removed
      in <productname>PostgreSQL</productname> version 12, so check to
      make sure they aren't present in an older database before claiming
      it can be upgraded.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [e1c6db630] 2023-09-07 14:12:18 +0900
Branch: REL_16_STABLE Release: REL_16_0 [105795843] 2023-09-07 14:12:25 +0900
Branch: REL_15_STABLE [574bff7bd] 2023-09-07 14:12:29 +0900
Branch: REL_14_STABLE [9de74ca70] 2023-09-07 14:12:31 +0900
Branch: REL_13_STABLE [6cdd7b805] 2023-09-07 14:12:33 +0900
Branch: REL_12_STABLE [2fe968ce9] 2023-09-07 14:12:34 +0900
Branch: REL_11_STABLE [dbd1a06c9] 2023-09-07 14:12:36 +0900
-->
     <para>
      Avoid generating invalid temporary slot names
      in <application>pg_basebackup</application> (Jelte Fennema)
     </para>

     <para>
      This has only been seen to occur when the server connection runs
      through <application>pgbouncer</application>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [90ebcc32d] 2023-10-14 15:54:46 -0700
Branch: REL_16_STABLE [06ff06484] 2023-10-14 15:54:49 -0700
Branch: REL_15_STABLE [1102f4ece] 2023-10-14 15:54:49 -0700
Branch: REL_14_STABLE [0df88a686] 2023-10-14 15:54:49 -0700
Branch: REL_13_STABLE [3895e9153] 2023-10-14 15:54:50 -0700
Branch: REL_12_STABLE [73cda80a3] 2023-10-14 15:54:50 -0700
-->
     <para>
      Avoid false <quote>too many client connections</quote> errors
      in <application>pgbench</application> on Windows (Noah Misch)
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [6ec9e9975] 2023-10-30 14:46:05 -0700
Branch: REL_16_STABLE [3c6a05b80] 2023-10-30 14:46:08 -0700
Branch: REL_15_STABLE [6f81386a9] 2023-10-30 14:46:08 -0700
Branch: REL_14_STABLE [42496cba6] 2023-10-30 14:46:09 -0700
Branch: REL_13_STABLE [f31ccb575] 2023-10-30 14:46:09 -0700
Branch: REL_12_STABLE [3a205c916] 2023-10-30 14:46:09 -0700
Branch: REL_11_STABLE [e04509f32] 2023-10-30 14:46:09 -0700
-->
     <para>
      In <filename>contrib/amcheck</filename>, do not report interrupted
      page deletion as corruption (Noah Misch)
     </para>

     <para>
      This fix prevents false-positive reports of <quote>the first child
      of leftmost target page is not leftmost of its
      level</quote>, <quote>block NNNN is not leftmost</quote>
      or <quote>left link/right link pair in index XXXX not in
      agreement</quote>.  They appeared
      if <application>amcheck</application> ran after an unfinished btree
      index page deletion and before <command>VACUUM</command> had cleaned
      things up.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [849172ff4] 2023-10-29 11:14:37 +0000
Branch: REL_16_STABLE [ab73a37e9] 2023-10-29 11:14:36 +0000
Branch: REL_15_STABLE [5f0691839] 2023-10-29 11:14:35 +0000
Branch: REL_14_STABLE [162b38a06] 2023-10-29 11:14:34 +0000
Branch: REL_13_STABLE [35e6a5c20] 2023-10-29 11:14:34 +0000
Branch: REL_12_STABLE [f7cd7b265] 2023-10-29 11:14:33 +0000
Branch: REL_11_STABLE [c804ffb56] 2023-10-29 11:14:32 +0000
-->
     <para>
      Fix failure of <filename>contrib/btree_gin</filename> indexes
      on <type>interval</type> columns,
      when an indexscan using the <literal>&lt;</literal>
      or <literal>&lt;=</literal> operator is performed (Dean Rasheed)
     </para>

     <para>
      Such an indexscan failed to return all the entries it should.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [76200e5ee] 2023-10-19 05:13:23 +1300
Branch: REL_16_STABLE [774185056] 2023-10-19 05:14:07 +1300
Branch: REL_15_STABLE [b60e3ac76] 2023-10-19 05:15:38 +1300
Branch: REL_14_STABLE [0a8b7d5c1] 2023-10-19 05:16:00 +1300
Branch: REL_13_STABLE [53c4dabe1] 2023-10-19 05:16:24 +1300
Branch: REL_12_STABLE [7da915e33] 2023-10-19 05:17:27 +1300
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [37d5babb5] 2023-10-18 22:47:23 +1300
Branch: REL_16_STABLE [74d19ec09] 2023-10-18 22:53:56 +1300
Branch: REL_15_STABLE [eed1feb3f] 2023-10-18 22:59:46 +1300
Branch: REL_14_STABLE [82d9a782a] 2023-10-18 23:04:09 +1300
Branch: REL_13_STABLE [f28956b23] 2023-10-18 23:16:48 +1300
Branch: REL_12_STABLE [15ddc9725] 2023-10-18 23:18:16 +1300
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [f90b4a846] 2023-10-19 03:01:55 +1300
Branch: REL_16_STABLE [60596f148] 2023-10-19 03:07:18 +1300
Branch: REL_15_STABLE [b2e097788] 2023-10-19 03:03:27 +1300
Branch: REL_14_STABLE [ee3e4c41f] 2023-10-19 03:03:46 +1300
Branch: REL_13_STABLE [981292c19] 2023-10-19 03:04:00 +1300
Branch: REL_12_STABLE [d701f0d1f] 2023-10-19 03:04:09 +1300
Author: Thomas Munro <tmunro@postgresql.org>
Branch: REL_13_STABLE [7c070e224] 2023-10-24 11:23:08 +1300
Branch: REL_12_STABLE [9ad986276] 2023-10-24 11:23:21 +1300
-->
     <para>
      Add support for LLVM 16 and 17 (Thomas Munro, Dmitry Dolgov)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [06843df4a] 2023-09-29 14:07:30 -0400
Branch: REL_16_STABLE [75c562653] 2023-09-29 14:07:30 -0400
Branch: REL_15_STABLE [be3398ea1] 2023-09-29 14:07:30 -0400
Branch: REL_14_STABLE [e9d1560d3] 2023-09-29 14:07:30 -0400
Branch: REL_13_STABLE [1b64e374c] 2023-09-29 14:07:30 -0400
Branch: REL_12_STABLE [d8ceb98e4] 2023-09-29 14:07:30 -0400
Branch: REL_11_STABLE [4d10fda52] 2023-09-29 14:07:30 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [3aa021b29] 2023-09-26 21:06:21 -0400
Branch: REL_16_STABLE [e73d6a0df] 2023-09-26 21:06:21 -0400
Branch: REL_15_STABLE [78f17fb97] 2023-09-26 21:06:21 -0400
Branch: REL_14_STABLE [5d60e8ed3] 2023-09-26 21:06:21 -0400
Branch: REL_13_STABLE [53cc19bfa] 2023-09-26 21:06:21 -0400
Branch: REL_12_STABLE [abe423712] 2023-09-26 21:06:21 -0400
Branch: REL_11_STABLE [0e0de20c8] 2023-09-26 21:06:21 -0400
-->
     <para>
      Suppress assorted build-time warnings on
      recent <productname>macOS</productname> (Tom Lane)
     </para>

     <para>
      <productname>Xcode 15</productname> (released
      with <productname>macOS Sonoma</productname>) changed the linker's
      behavior in a way that causes many duplicate-library warnings while
      building <productname>PostgreSQL</productname>.  These were
      harmless, but they're annoying so avoid citing the same libraries
      twice.  Also remove use of the <option>-multiply_defined
      suppress</option> linker switch, which apparently has been a no-op
      for a long time, and is now actively complained of.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [6fc3a138b] 2023-09-27 14:40:23 +0900
Branch: REL_16_STABLE [641db601b] 2023-09-27 14:41:15 +0900
Branch: REL_15_STABLE [8a9e4e84e] 2023-09-27 14:41:21 +0900
Branch: REL_14_STABLE [e4d7ad30f] 2023-09-27 14:41:23 +0900
Branch: REL_13_STABLE [a64b8b035] 2023-09-27 14:41:26 +0900
-->
     <para>
      When building <filename>contrib/unaccent</filename>'s rules file,
      fall back to using <literal>python</literal>
      if <literal>--with-python</literal> was not given and make
      variable <literal>PYTHON</literal> was not set (Japin Li)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5fd3e06f6] 2023-10-28 11:54:40 -0400
Branch: REL_16_STABLE [d1537afe3] 2023-10-28 11:54:51 -0400
Branch: REL_15_STABLE [85b98a70b] 2023-10-28 11:54:59 -0400
Branch: REL_14_STABLE [877e7b4b9] 2023-10-28 11:55:06 -0400
Branch: REL_13_STABLE [d94612873] 2023-10-28 11:55:14 -0400
Branch: REL_12_STABLE [7180bb827] 2023-10-28 11:55:22 -0400
Branch: REL_11_STABLE [64fc5e005] 2023-10-28 11:55:30 -0400
-->
     <para>
      Remove <literal>PHOT</literal> (Phoenix Islands Time) from the
      default timezone abbreviations list (Tom Lane)
     </para>

     <para>
      Presence of this abbreviation in the default list can cause failures
      on recent Debian and Ubuntu releases, as they no longer install the
      underlying tzdb entry by default.  Since this is a made-up
      abbreviation for a zone with a total human population of about two
      dozen, it seems unlikely that anyone will miss it.  If someone does,
      they can put it back via a custom abbreviations file.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15-4">
  <title>Release 15.4</title>

  <formalpara>
  <title>Release date:</title>
  <para>2023-08-10</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.3.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.4</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, if you use BRIN indexes, it may be advisable to reindex them;
    see the third changelog entry below.
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 15.1,
    see <xref linkend="release-15-1"/>.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [cd5f2a357] 2023-08-07 06:05:56 -0700
Branch: REL_16_STABLE [f53511010] 2023-08-07 06:05:59 -0700
Branch: REL_15_STABLE [de494ec14] 2023-08-07 06:06:00 -0700
Branch: REL_14_STABLE [d4648a74b] 2023-08-07 06:06:00 -0700
Branch: REL_13_STABLE [b1b585e0f] 2023-08-07 06:06:00 -0700
Branch: REL_12_STABLE [eb044d8f0] 2023-08-07 06:06:00 -0700
Branch: REL_11_STABLE [919ebb023] 2023-08-07 06:06:01 -0700
-->
     <para>
      Disallow substituting a schema or owner name into an extension script
      if the name contains a quote, backslash, or dollar sign (Noah Misch)
     </para>

     <para>
      This restriction guards against SQL-injection hazards for trusted
      extensions.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks Micah Gates,
      Valerie Woolard, Tim Carey-Smith, and Christoph Berg for reporting
      this problem.
      (CVE-2023-39417)
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [c2e08b04c] 2023-08-07 09:28:47 +0100
Branch: REL_16_STABLE [67a007dc0] 2023-08-07 09:27:09 +0100
Branch: REL_15_STABLE [cb2ae5741] 2023-08-07 09:24:27 +0100
-->
     <para>
      Fix <command>MERGE</command> to enforce row security policies
      properly (Dean Rasheed)
     </para>

     <para>
      When <command>MERGE</command> performs an <literal>UPDATE</literal>
      action, it should enforce any <literal>UPDATE</literal> or
      <literal>SELECT</literal> RLS policies defined on the target table,
      to be consistent with the way that a plain <command>UPDATE</command>
      with a <literal>WHERE</literal> clause works.  Instead it was
      enforcing <literal>INSERT</literal> RLS policies for both
      <literal>INSERT</literal> and <literal>UPDATE</literal> actions.
     </para>

     <para>
      In addition, when <command>MERGE</command> performs a <literal>DO
      NOTHING</literal> action, it applied the target table's
      <literal>DELETE</literal> RLS policies to existing rows, even though
      those rows are not being deleted.  While it's not a security
      problem, this could result in unwanted errors.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks
      Dean Rasheed for reporting this problem.
      (CVE-2023-39418)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master Release: REL_16_BR [3581cbdcd] 2023-05-19 01:29:44 +0200
Branch: REL_15_STABLE [e18769323] 2023-05-19 00:15:13 +0200
Branch: REL_14_STABLE [40d465caf] 2023-05-19 00:15:00 +0200
Branch: REL_13_STABLE [6c512fc6e] 2023-05-19 00:14:05 +0200
Branch: REL_12_STABLE [d78a66d92] 2023-05-19 00:16:13 +0200
Branch: REL_11_STABLE [fc7dc728d] 2023-05-19 00:21:05 +0200
Branch: master Release: REL_16_BR [3ec8a3bfb] 2023-05-18 23:33:23 +0200
Branch: REL_15_STABLE [80f64b900] 2023-05-18 23:33:45 +0200
Branch: REL_14_STABLE [3f1356e5d] 2023-05-18 23:34:10 +0200
Branch: REL_13_STABLE [2b1ab28b9] 2023-05-18 23:34:35 +0200
Branch: REL_12_STABLE [d42ffda68] 2023-05-18 23:34:56 +0200
Branch: REL_11_STABLE [b511d7323] 2023-05-18 23:35:16 +0200
-->
     <para>
      Fix confusion between empty (no rows) ranges and all-NULL ranges in
      BRIN indexes, as well as incorrect merging of all-NULL summaries
      (Tomas Vondra)
     </para>

     <para>
      Each of these oversights could result in forgetting that a BRIN
      index range contains any NULL values, potentially allowing
      subsequent queries that should return NULL values to miss doing so.
     </para>

     <para>
      This fix will not in itself correct faulty BRIN entries.
      It's recommended to <command>REINDEX</command> any BRIN indexes that
      may be used to search for nulls.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [c66a7d75e] 2023-07-13 13:03:28 -0700
Branch: REL_16_STABLE [a4b4cc1d6] 2023-07-13 13:03:30 -0700
Branch: REL_15_STABLE [f66403749] 2023-07-13 13:04:45 -0700
Branch: REL_14_STABLE [d11efe830] 2023-07-13 13:03:33 -0700
Branch: REL_13_STABLE [81ce00006] 2023-07-13 13:03:34 -0700
Branch: REL_12_STABLE [034a9fcd2] 2023-07-13 13:03:36 -0700
Branch: REL_11_STABLE [1c38e7ae1] 2023-07-13 13:03:37 -0700
-->
     <para>
      Avoid leaving a corrupted database behind when <command>DROP
      DATABASE</command> is interrupted (Andres Freund)
     </para>

     <para>
      If <command>DROP DATABASE</command> was interrupted after it had
      already begun taking irreversible steps, the target database
      remained accessible (because the removal of
      its <structname>pg_database</structname> row would roll back),
      but it would have corrupt contents.  Fix by marking the database
      as inaccessible before we begin to perform irreversible operations.
      A failure after that will leave the database still partially
      present, but nothing can be done with it except to issue
      another <command>DROP DATABASE</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [cfc43aeb3] 2023-06-30 13:54:48 +0900
Branch: REL_16_STABLE [f4691e2e7] 2023-06-30 13:54:53 +0900
Branch: REL_15_STABLE [cb4ac3e56] 2023-06-30 13:54:55 +0900
Branch: REL_14_STABLE [663b35f2d] 2023-06-30 13:54:56 +0900
Branch: REL_13_STABLE [537b70b82] 2023-06-30 13:54:59 +0900
Branch: REL_12_STABLE [dbe0e5c56] 2023-06-30 13:55:02 +0900
Branch: REL_11_STABLE [c75c33de5] 2023-06-30 13:55:07 +0900
-->
     <para>
      Ensure that partitioned indexes are correctly marked as valid or not
      at creation (Michael Paquier)
     </para>

     <para>
      If a new partitioned index matches an existing but invalid index on
      one of the partitions, the partitioned index could end up being
      marked valid prematurely.  This could lead to misbehavior or
      assertion failures in subsequent queries on the partitioned table.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_16_BR [fc55c7ff8] 2023-06-28 15:57:31 +0900
Branch: REL_15_STABLE [7aa17b498] 2023-06-28 15:57:43 +0900
Branch: REL_14_STABLE [6160e221d] 2023-06-28 15:57:48 +0900
Branch: REL_13_STABLE [f42844069] 2023-06-28 15:57:51 +0900
Branch: REL_12_STABLE [63b292e73] 2023-06-28 15:57:53 +0900
Branch: REL_11_STABLE [e90e9275f] 2023-06-28 15:57:55 +0900
-->
     <para>
      Ignore invalid child indexes when matching partitioned indexes to
      child indexes during <command>ALTER TABLE ATTACH PARTITION</command>
      (Michael Paquier)
     </para>

     <para>
      Such an index will now be ignored, and a new child index created
      instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [38ea6aa90] 2023-07-14 10:12:48 +0900
Branch: REL_16_STABLE [31f9d41d6] 2023-07-14 10:13:14 +0900
Branch: REL_15_STABLE [c0dc97c7b] 2023-07-14 10:13:15 +0900
Branch: REL_14_STABLE [954cc2139] 2023-07-14 10:13:17 +0900
Branch: REL_13_STABLE [c89d74c18] 2023-07-14 10:13:20 +0900
Branch: REL_12_STABLE [f1d6bcdd8] 2023-07-14 10:13:21 +0900
Branch: REL_11_STABLE [ed2b58c15] 2023-07-14 10:13:22 +0900
-->
     <para>
      Fix possible failure when marking a partitioned index valid after
      all of its partitions have been attached (Michael Paquier)
     </para>

     <para>
      The update of the index's <structname>pg_index</structname> entry
      could use stale data for other columns.  One reported symptom is
      an <quote>attempted to update invisible tuple</quote> error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [bd5ddbe86] 2023-07-10 09:40:07 +0900
Branch: REL_16_STABLE [55c95f24c] 2023-07-10 09:40:12 +0900
Branch: REL_15_STABLE [d1e0f408c] 2023-07-10 09:40:14 +0900
Branch: REL_14_STABLE [235e716bc] 2023-07-10 09:40:15 +0900
Branch: REL_13_STABLE [f5b075adc] 2023-07-10 09:40:17 +0900
Branch: REL_12_STABLE [02021f1b2] 2023-07-10 09:40:22 +0900
Branch: REL_11_STABLE [914e72e6e] 2023-07-10 09:40:24 +0900
-->
     <para>
      Fix <command>ALTER EXTENSION SET SCHEMA</command> to complain if the
      extension contains any objects outside the extension's schema
      (Michael Paquier, Heikki Linnakangas)
     </para>

     <para>
      Erroring out if the extension contains objects in multiple schemas
      was always intended; but the check was mis-coded so that it would
      fail to detect some cases, leading to surprising behavior.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_16_BR [97d891010] 2023-06-30 07:49:01 +0900
Branch: REL_15_STABLE [93401ec02] 2023-06-30 07:49:07 +0900
-->
     <para>
      Fix tracking of tables' access method dependencies (Michael Paquier)
     </para>

     <para>
      <command>ALTER TABLE ... SET ACCESS METHOD</command> failed to
      update relevant <structname>pg_depend</structname> entries when
      changing a table's access method.  When using non-built-in access
      methods, this creates a risk that an access method could be dropped
      even though tables still depend on it.  This fix corrects the logic
      in <command>ALTER TABLE</command>, but it will not adjust any
      already-missing <structname>pg_depend</structname> entries.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master Release: REL_16_BR [7fcd7ef2a] 2023-06-19 13:00:42 +1200
Branch: REL_15_STABLE [8f2ec8cc7] 2023-06-19 13:01:29 +1200
Branch: REL_14_STABLE [73f1c17fc] 2023-06-19 13:01:58 +1200
Branch: REL_13_STABLE [06286f8a2] 2023-06-19 13:02:24 +1200
Branch: REL_12_STABLE [dcef5b052] 2023-06-19 13:02:52 +1200
Branch: REL_11_STABLE [f6345f03f] 2023-06-19 13:03:17 +1200
-->
     <para>
      Don't use partial unique indexes for uniqueness proofs in the
      planner (David Rowley)
     </para>

     <para>
      This could give rise to incorrect plans, since the presumed
      uniqueness of rows read from a table might not hold if the index in
      question isn't used to scan the table.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [990c3650c] 2023-08-07 22:14:21 +1200
Branch: REL_16_STABLE [ae89129aa] 2023-08-07 22:14:54 +1200
Branch: REL_15_STABLE [71662373b] 2023-08-07 22:15:23 +1200
Branch: REL_14_STABLE [bf315354e] 2023-08-07 22:15:50 +1200
-->
     <para>
      Don't Memoize lateral joins with volatile join conditions
      (Richard Guo)
     </para>

     <para>
      Applying Memoize to a sub-plan that contains volatile filter
      conditions is likely to lead to wrong answers.  The check to avoid
      doing this missed some cases that can arise when
      using <literal>LATERAL</literal>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Etsuro Fujita <efujita@postgresql.org>
Branch: master [6f80a8d9c] 2023-07-28 15:45:00 +0900
Branch: REL_16_STABLE [695f5deb7] 2023-07-28 15:45:01 +0900
Branch: REL_15_STABLE [d1ef5631e] 2023-07-28 15:45:02 +0900
Branch: REL_14_STABLE [b0e390e6d] 2023-07-28 15:45:04 +0900
Branch: REL_13_STABLE [730f983ef] 2023-07-28 15:45:06 +0900
Branch: REL_12_STABLE [9edf72aa7] 2023-07-28 15:45:08 +0900
Branch: REL_11_STABLE [db01f2696] 2023-07-28 15:45:09 +0900
-->
     <para>
      Avoid producing incorrect plans for foreign joins with
      pseudoconstant join clauses (Etsuro Fujita)
     </para>

     <para>
      The planner currently lacks support for attaching pseudoconstant
      join clauses to a pushed-down remote join, so disable generation
      of remote joins in such cases.  (A better solution will require
      ABI-breaking changes of planner data structures, so it will have to
      wait for a future major release.)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [792213f2e] 2023-06-13 15:58:43 -0400
Branch: REL_15_STABLE [cc6974df1] 2023-06-13 15:58:37 -0400
Branch: REL_14_STABLE [d1423c52e] 2023-06-13 15:58:37 -0400
Branch: REL_13_STABLE [a36d0014f] 2023-06-13 15:58:37 -0400
Branch: REL_12_STABLE [b4110bdbf] 2023-06-13 15:58:37 -0400
Branch: REL_11_STABLE [13192a324] 2023-06-13 15:58:37 -0400
-->
     <para>
      Correctly handle sub-SELECTs in RLS policy expressions and
      security-barrier views when expanding rule actions (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [126552c85] 2023-07-04 09:07:31 +1200
Branch: REL_16_STABLE [12529028a] 2023-07-04 09:04:35 +1200
Branch: REL_15_STABLE [d34aa0a2f] 2023-07-04 09:10:37 +1200
Branch: REL_14_STABLE [fb663f387] 2023-07-04 09:14:16 +1200
Branch: REL_13_STABLE [fc1547384] 2023-07-04 09:20:55 +1200
Branch: REL_12_STABLE [255a925d3] 2023-07-04 09:21:33 +1200
Branch: REL_11_STABLE [13f127800] 2023-07-04 09:40:30 +1200
Branch: master [bcc93a389] 2023-07-04 09:07:31 +1200
Branch: REL_16_STABLE [d03d9a261] 2023-07-04 09:04:35 +1200
Branch: REL_15_STABLE [ab265e985] 2023-07-04 09:10:37 +1200
Branch: REL_14_STABLE [3f7d3a77e] 2023-07-04 09:14:16 +1200
Branch: REL_13_STABLE [8976ac5c5] 2023-07-04 09:20:55 +1200
Branch: REL_12_STABLE [17b8887c2] 2023-07-04 09:21:33 +1200
Branch: REL_11_STABLE [814f3c8e4] 2023-07-04 09:40:30 +1200
Branch: master [f9b7fc651] 2023-07-04 09:07:31 +1200
Branch: REL_16_STABLE [0cb1fb2c9] 2023-07-04 09:04:35 +1200
Branch: REL_15_STABLE [0f275b0ee] 2023-07-04 09:10:37 +1200
Branch: REL_14_STABLE [ae6d536ed] 2023-07-04 09:14:16 +1200
Branch: REL_13_STABLE [8f705d7b9] 2023-07-04 09:20:55 +1200
Branch: REL_12_STABLE [fe88497b4] 2023-07-04 09:21:33 +1200
Branch: REL_11_STABLE [0048c3b51] 2023-07-04 09:27:00 +1200
-->
     <para>
      Fix race conditions in conflict detection
      for <literal>SERIALIZABLE</literal> isolation mode
      (Thomas Munro)
     </para>

     <para>
      Conflicts could be missed when using bitmap heap scans, when using
      GIN indexes, and when examining an initially-empty btree index.
      All these cases could lead to serializability failures due to
      improperly allowing conflicting transactions to commit.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [70b42f279] 2023-05-19 14:26:40 -0400
Branch: REL_15_STABLE [4729d1e8a] 2023-05-19 14:26:34 -0400
Branch: REL_14_STABLE [f8320cc72] 2023-05-19 14:26:34 -0400
-->
     <para>
      Fix misbehavior of EvalPlanQual checks with inherited or partitioned
      target tables (Tom Lane)
     </para>

     <para>
      This oversight could lead to update or delete actions
      in <literal>READ COMMITTED</literal> isolation mode getting
      performed when they should have been skipped because of a
      conflicting concurrent update.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [45392626c] 2023-06-20 17:47:53 -0400
Branch: REL_15_STABLE [c2f974fff] 2023-06-20 17:47:53 -0400
Branch: REL_14_STABLE [d911dce14] 2023-06-20 17:47:36 -0400
Branch: REL_13_STABLE [2f97105e9] 2023-06-20 17:47:36 -0400
Branch: REL_12_STABLE [9529b1eb1] 2023-06-20 17:47:36 -0400
-->
     <para>
      Fix hash join with an inner-side hash key that contains Params
      coming from an outer nested loop (Tom Lane)
     </para>

     <para>
      When rescanning the join after the values of such Params have
      changed, we must rebuild the hash table, but neglected to do so.
      This could result in missing join output rows.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [43af714de] 2023-06-29 10:19:10 -0400
Branch: REL_15_STABLE [cc8cca3c2] 2023-06-29 10:19:10 -0400
Branch: REL_14_STABLE [0789b82a9] 2023-06-29 10:19:10 -0400
Branch: REL_13_STABLE [d0ab203bc] 2023-06-29 10:19:10 -0400
Branch: REL_12_STABLE [53b93e853] 2023-06-29 10:19:10 -0400
Branch: REL_11_STABLE [7f11b7a9c] 2023-06-29 10:19:10 -0400
-->
     <para>
      Fix intermittent failures when trying to update a field of a
      composite column (Tom Lane)
     </para>

     <para>
      If the overall value of the composite column is wide enough to
      require out-of-line toasting, then an unluckily-timed cache flush
      could cause errors or server crashes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master Release: REL_16_BR [f24523672] 2023-06-07 18:54:34 +0200
Branch: REL_15_STABLE [ee87f8b63] 2023-06-07 18:52:21 +0200
Branch: REL_14_STABLE [7f528e96c] 2023-06-07 18:53:04 +0200
Branch: REL_13_STABLE [c504aa857] 2023-06-07 18:53:16 +0200
Branch: REL_12_STABLE [54e1b8587] 2023-06-07 18:53:30 +0200
Branch: master [ce5aaea8c] 2023-07-02 22:21:02 +0200
Branch: REL_16_STABLE [9cf85093b] 2023-07-02 22:22:31 +0200
Branch: REL_15_STABLE [7ae4e7868] 2023-07-02 22:22:50 +0200
Branch: REL_14_STABLE [260dbf19a] 2023-07-02 22:23:04 +0200
Branch: REL_13_STABLE [984c23f6f] 2023-07-02 22:23:20 +0200
-->
     <para>
      Prevent query-lifespan memory leaks in some <command>UPDATE</command>
      queries with triggers (Tomas Vondra)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [98640f960] 2023-07-02 20:03:30 +0200
Branch: REL_16_STABLE [9ae7b5d1f] 2023-07-02 20:04:16 +0200
Branch: REL_15_STABLE [0c5fe4ff6] 2023-07-02 20:04:40 +0200
Branch: REL_14_STABLE [c1affa38c] 2023-07-02 20:05:14 +0200
Branch: REL_13_STABLE [3ce761d5c] 2023-07-02 20:05:35 +0200
-->
     <para>
      Prevent query-lifespan memory leaks when an Incremental Sort plan
      node is rescanned (James Coleman, Laurenz Albe, Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [7398e2722] 2023-06-12 10:54:44 -0400
Branch: REL_15_STABLE [bd590d1fe] 2023-06-12 10:54:28 -0400
Branch: REL_14_STABLE [5eaa05f63] 2023-06-12 10:54:28 -0400
Branch: REL_13_STABLE [6f23b5f74] 2023-06-12 10:54:28 -0400
-->
     <para>
      Accept fractional seconds in the input to <type>jsonpath</type>'s
      <function>datetime()</function> method (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [691594acd] 2023-06-24 17:18:08 -0400
Branch: REL_15_STABLE [a77d90171] 2023-06-24 17:18:08 -0400
Branch: REL_14_STABLE [4c61afa47] 2023-06-24 17:18:08 -0400
Branch: REL_13_STABLE [b6ab18a99] 2023-06-24 17:18:08 -0400
Branch: REL_12_STABLE [3b4580f5c] 2023-06-24 17:18:08 -0400
Branch: REL_11_STABLE [c7f33a197] 2023-06-24 17:18:08 -0400
-->
     <para>
      Prevent stack-overflow crashes with very complex text search
      patterns (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [38df84c65] 2023-07-27 11:56:35 -0400
Branch: REL_16_STABLE [de3f0e3fe] 2023-07-27 11:56:35 -0400
Branch: REL_15_STABLE [313ceda2f] 2023-07-27 12:07:48 -0400
Branch: REL_14_STABLE [341996248] 2023-07-27 12:07:48 -0400
Branch: REL_13_STABLE [288b4288c] 2023-07-27 12:07:48 -0400
Branch: REL_12_STABLE [0660f74e8] 2023-07-27 12:07:48 -0400
Branch: REL_11_STABLE [1d031ad54] 2023-07-27 12:07:48 -0400
-->
     <para>
      Allow tokens up to 10240 bytes long
      in <filename>pg_hba.conf</filename>
      and <filename>pg_ident.conf</filename> (Tom Lane)
     </para>

     <para>
      The previous limit of 256 bytes has been found insufficient for some
      use-cases.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: REL_15_STABLE [a5f312c58] 2023-07-06 13:05:25 +0300
-->
     <para>
      Ensure that all existing placeholders are checked for matches when
      an extension declares its GUC prefix to be reserved (Karina
      Litskevich, Ekaterina Sokolova)
     </para>

     <para>
      Faulty loop logic could cause some entries to be skipped.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [4f4d73466] 2023-07-05 13:13:13 +0300
Branch: REL_16_STABLE [dc0b58417] 2023-07-05 13:25:59 +0300
Branch: REL_15_STABLE [fa96a74a0] 2023-07-05 13:13:30 +0300
Branch: REL_14_STABLE [bfb493dba] 2023-07-05 13:13:35 +0300
Branch: REL_13_STABLE [59c2a6fe9] 2023-07-05 13:13:39 +0300
Branch: REL_12_STABLE [162aa47c3] 2023-07-05 13:14:24 +0300
Branch: REL_11_STABLE [2316ff1ae] 2023-07-05 13:14:33 +0300
-->
     <para>
      Fix mishandling of C++ out-of-memory conditions (Heikki Linnakangas)
     </para>

     <para>
      If JIT is in use, running out of memory in a
      C++ <function>new</function> call would lead to
      a <productname>PostgreSQL</productname> FATAL error, instead of the
      expected C++ exception.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [9089287aa] 2023-07-20 14:23:46 -0400
Branch: REL_16_STABLE [c0f531396] 2023-07-20 14:23:46 -0400
Branch: REL_15_STABLE [fbaf65cd6] 2023-07-20 14:23:46 -0400
Branch: REL_14_STABLE [10fd061bb] 2023-07-20 14:23:46 -0400
Branch: REL_13_STABLE [291c02540] 2023-07-20 14:23:46 -0400
-->
     <para>
      Fix rare null-pointer crash in <filename>plancache.c</filename>
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Masahiko Sawada <msawada@postgresql.org>
Branch: master [68a59f9e9] 2023-07-05 14:49:46 +0900
Branch: REL_16_STABLE [be8cae7e2] 2023-07-05 14:49:58 +0900
Branch: REL_15_STABLE [66f8a1397] 2023-07-05 14:49:53 +0900
-->
     <para>
      Avoid leaking a stats entry for a subscription when it is dropped
      (Masahiko Sawada)
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [03f80daac] 2023-07-04 15:16:47 +1200
Branch: REL_16_STABLE [af8f9ec66] 2023-07-04 15:21:36 +1200
Branch: REL_15_STABLE [9ffb10f18] 2023-07-04 15:24:42 +1200
Branch: REL_14_STABLE [b7ec66731] 2023-07-04 15:26:42 +1200
Branch: REL_13_STABLE [a0003572f] 2023-07-04 15:28:29 +1200
Branch: REL_12_STABLE [74ad9b0d1] 2023-07-04 15:29:41 +1200
Branch: REL_11_STABLE [1605623ec] 2023-07-04 16:13:12 +1200
-->
     <para>
      Avoid losing track of possibly-useful shared memory segments when a
      page free results in coalescing ranges of free space (Dongming Liu)
     </para>

     <para>
      Ensure that the segment is moved into the
      appropriate <quote>bin</quote> for its new amount of free space, so
      that it will be found by subsequent searches.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Geoghegan <pg@bowt.ie>
Branch: master Release: REL_16_BR [5f0762f14] 2023-06-21 17:41:58 -0700
Branch: REL_15_STABLE [642bec1f8] 2023-06-21 17:41:56 -0700
Branch: REL_14_STABLE [63fa0deb3] 2023-06-21 17:41:54 -0700
Branch: REL_13_STABLE [b6311824f] 2023-06-21 17:41:52 -0700
Branch: REL_12_STABLE [355917c07] 2023-06-21 17:41:50 -0700
Branch: REL_11_STABLE [7ddba19eb] 2023-06-21 17:41:48 -0700
Branch: master Release: REL_16_BR [5abff197c] 2023-05-25 15:33:00 -0700
Branch: REL_15_STABLE [6983a5112] 2023-05-25 15:32:57 -0700
Branch: REL_14_STABLE [322c9b340] 2023-05-25 15:32:53 -0700
Branch: REL_13_STABLE [8f876d15c] 2023-05-25 15:32:50 -0700
Branch: REL_12_STABLE [188dad680] 2023-05-25 15:32:48 -0700
Branch: REL_11_STABLE [a72b503ca] 2023-05-25 15:32:45 -0700
-->
     <para>
      Allow <command>VACUUM</command> to continue after detecting certain
      types of b-tree index corruption (Peter Geoghegan)
     </para>

     <para>
      If an invalid sibling-page link is detected, log the issue and press
      on, rather than throwing an error as before.  Nothing short
      of <command>REINDEX</command> will fix the broken index, but
      preventing <command>VACUUM</command> from completing until that is
      done risks making matters far worse.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [83ecfa9fa] 2023-07-13 13:03:28 -0700
Branch: REL_16_STABLE [03ccc9569] 2023-07-13 13:03:30 -0700
Branch: REL_15_STABLE [82e97b864] 2023-07-13 13:03:31 -0700
Branch: REL_14_STABLE [e246fd423] 2023-07-13 13:03:33 -0700
Branch: REL_13_STABLE [53336e8f6] 2023-07-13 13:03:34 -0700
Branch: REL_12_STABLE [7aec84e4c] 2023-07-13 13:03:36 -0700
Branch: REL_11_STABLE [1386f0987] 2023-07-13 13:03:37 -0700
-->
     <para>
      Ensure that <varname>WrapLimitsVacuumLock</varname> is released
      after <command>VACUUM</command> detects invalid data
      in <structname>pg_database</structname>.<structfield>datfrozenxid</structfield>
      or <structname>pg_database</structname>.<structfield>datminmxid</structfield>
      (Andres Freund)
     </para>

     <para>
      Failure to release this lock could lead to a deadlock later,
      although the lock would be cleaned up if the session exits or
      encounters some other error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [cb0cca188] 2023-07-18 13:43:44 +0900
Branch: REL_16_STABLE [f88bc9f38] 2023-07-18 13:44:27 +0900
Branch: REL_15_STABLE [a878eff6b] 2023-07-18 13:44:29 +0900
Branch: REL_14_STABLE [442749100] 2023-07-18 13:44:31 +0900
Branch: REL_13_STABLE [db59108a2] 2023-07-18 13:44:33 +0900
Branch: REL_12_STABLE [d2ee542a2] 2023-07-18 13:44:34 +0900
Branch: REL_11_STABLE [bc0581f8f] 2023-07-18 13:44:35 +0900
Branch: master [4e465aac3] 2023-07-18 14:04:31 +0900
Branch: REL_16_STABLE [926aa6d11] 2023-07-18 14:04:46 +0900
Branch: REL_15_STABLE [f6ecd2622] 2023-07-18 14:04:48 +0900
Branch: REL_14_STABLE [763d26205] 2023-07-18 14:04:50 +0900
Branch: REL_13_STABLE [b3ca4f0a5] 2023-07-18 14:04:51 +0900
Branch: REL_12_STABLE [410a0d6bd] 2023-07-18 14:04:52 +0900
Branch: REL_11_STABLE [6c7bffc09] 2023-07-18 14:04:54 +0900
-->
     <para>
      Avoid double replay of prepared transactions during crash
      recovery (suyu.cmj, Michael Paquier)
     </para>

     <para>
      After a crash partway through a checkpoint with some two-phase
      transaction state data already flushed to disk by this checkpoint,
      crash recovery could attempt to replay the prepared transaction(s)
      twice, leading to a fatal error such as <quote>lock is already
      held</quote> in the startup process.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [4b4798e13] 2023-07-04 17:57:03 +0300
Branch: REL_16_STABLE [d431644b4] 2023-07-04 18:04:27 +0300
Branch: REL_15_STABLE [e24c02e4d] 2023-07-04 18:07:27 +0300
Branch: REL_14_STABLE [d85bf0719] 2023-07-04 18:07:46 +0300
Branch: REL_13_STABLE [acc8cdff4] 2023-07-04 18:08:08 +0300
Branch: REL_12_STABLE [1b4f1c6f8] 2023-07-04 18:08:26 +0300
Branch: REL_11_STABLE [6377f705c] 2023-07-04 18:08:40 +0300
-->
     <para>
      Ensure that a newly created, but still empty table
      is <function>fsync</function>'ed at the next checkpoint (Heikki
      Linnakangas)
     </para>

     <para>
      Without this, if there is an operating system crash causing the
      empty file to disappear, subsequent operations on the table might
      fail with <quote>could not open file</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Branch: master [3142a8845] 2023-07-06 17:25:29 +0300
Branch: REL_16_STABLE [bf7b70e6b] 2023-07-06 17:28:50 +0300
Branch: REL_15_STABLE [25624c5d3] 2023-07-06 17:29:12 +0300
Branch: REL_14_STABLE [32f327f68] 2023-07-06 17:29:13 +0300
Branch: REL_13_STABLE [c50b869ed] 2023-07-06 17:29:14 +0300
Branch: REL_12_STABLE [80abec387] 2023-07-06 17:29:15 +0300
Branch: REL_11_STABLE [988719b88] 2023-07-06 17:29:16 +0300
-->
     <para>
      Ensure that creation of the init fork of an unlogged index is
      WAL-logged (Heikki Linnakangas)
     </para>

     <para>
      While an unlogged index's main data fork is not WAL-logged, its init
      fork should be, to ensure that we have a consistent state to restore
      the index to after a crash.  This step was missed if the init fork
      contains no data, which is a case not used by any standard index AM;
      but perhaps some extension behaves that way.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [4637a6ac0] 2023-07-03 11:16:27 +1200
Branch: REL_16_STABLE [76d5966b3] 2023-07-03 11:21:56 +1200
Branch: REL_15_STABLE [f50200c01] 2023-07-03 11:22:10 +1200
-->
     <para>
      Silence bogus <quote>missing contrecord</quote> errors (Thomas Munro)
     </para>

     <para>
      Treat this case as plain end-of-WAL to avoid logging inaccurate
      complaints from <application>pg_waldump</application>
      and <application>walsender</application>.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [deae1657e] 2023-08-02 01:39:47 +1200
Branch: REL_16_STABLE [b25acc302] 2023-08-02 01:40:27 +1200
Branch: REL_15_STABLE [67f3a697b] 2023-08-02 01:40:56 +1200
Branch: REL_14_STABLE [f457f2ef1] 2023-08-02 01:41:21 +1200
Branch: REL_13_STABLE [74a5bf1b6] 2023-08-02 01:41:55 +1200
Branch: REL_12_STABLE [668990980] 2023-08-02 01:44:31 +1200
-->
     <para>
      Fix overly strict assertion in <type>jsonpath</type> code
      (David Rowley)
     </para>

     <para>
      This assertion failed if a query applied
      the <literal>.type()</literal> operator to
      a <literal>like_regex</literal> result.
      There was no bug in non-assert builds.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [555b929bb] 2023-06-21 11:07:24 -0400
Branch: REL_15_STABLE [cb74f7bec] 2023-06-21 11:07:11 -0400
Branch: REL_14_STABLE [120ea65b8] 2023-06-21 11:07:11 -0400
Branch: REL_13_STABLE [d1fc0f382] 2023-06-21 11:07:11 -0400
Branch: REL_12_STABLE [a98a04005] 2023-06-21 11:07:11 -0400
Branch: REL_11_STABLE [a8be2356c] 2023-06-21 11:07:11 -0400
-->
     <para>
      Avoid assertion failure when processing an empty statement via the
      extended query protocol in an already-aborted transaction (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_16_BR [605994651] 2023-05-10 11:24:30 +0900
Branch: REL_15_STABLE [ccd21e1cf] 2023-05-10 11:24:40 +0900
-->
     <para>
      Avoid assertion failure when
      the <varname>stats_fetch_consistency</varname> setting is changed
      intra-transaction (Kyotaro Horiguchi)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [1d369c9e9] 2023-05-16 10:53:42 -0400
Branch: REL_15_STABLE [eaf99e4c4] 2023-05-16 10:53:42 -0400
Branch: REL_14_STABLE [ccd362325] 2023-05-16 10:53:42 -0400
Branch: REL_13_STABLE [0409c7fc7] 2023-05-16 10:53:42 -0400
Branch: REL_12_STABLE [0966291a4] 2023-05-16 10:53:42 -0400
Branch: REL_11_STABLE [8084bf9a4] 2023-05-16 10:53:42 -0400
-->
     <para>
      Fix <filename>contrib/fuzzystrmatch</filename>'s
      Soundex <function>difference()</function> function to handle empty
      input sanely (Alexander Lakhin, Tom Lane)
     </para>

     <para>
      An input string containing no alphabetic characters resulted in
      unpredictable output.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_16_BR [d522b05c8] 2023-06-12 09:14:03 +0900
Branch: REL_15_STABLE [3a5222a43] 2023-06-12 09:14:13 +0900
Branch: REL_14_STABLE [e0e682945] 2023-06-12 09:14:14 +0900
Branch: REL_13_STABLE [78bf0a256] 2023-06-12 09:14:17 +0900
Branch: REL_12_STABLE [edf1de65e] 2023-06-12 09:14:19 +0900
Branch: REL_11_STABLE [bbfc26d86] 2023-06-12 09:14:20 +0900
-->
     <para>
      Tighten whitespace checks in <filename>contrib/hstore</filename>
      input (Evan Jones)
     </para>

     <para>
      In some cases, characters would be falsely recognized as whitespace
      and hence discarded.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_16_BR [c00fbe89d] 2023-06-15 13:45:34 +0900
Branch: REL_15_STABLE [4be308ede] 2023-06-15 13:45:38 +0900
Branch: REL_14_STABLE [019a40d61] 2023-06-15 13:45:40 +0900
Branch: REL_13_STABLE [ae9aac64a] 2023-06-15 13:45:41 +0900
Branch: REL_12_STABLE [6ecc1c02a] 2023-06-15 13:45:42 +0900
Branch: REL_11_STABLE [ab40b0395] 2023-06-15 13:45:44 +0900
-->
     <para>
      Disallow oversize input arrays
      with <filename>contrib/intarray</filename>'s
      <literal>gist__int_ops</literal> index opclass (Ankit Kumar Pandey,
      Alexander Lakhin)
     </para>

     <para>
      Previously this code would report a <literal>NOTICE</literal> but
      press on anyway, creating an invalid index entry that presents a
      risk of crashes when the index is read.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [8aac9759b] 2023-07-13 13:07:51 -0400
Branch: REL_16_STABLE [e27f3f52c] 2023-07-13 13:08:08 -0400
Branch: REL_15_STABLE [5cb461989] 2023-07-13 13:08:17 -0400
Branch: REL_14_STABLE [a6991f763] 2023-07-13 13:08:23 -0400
Branch: REL_13_STABLE [7fffcc2ee] 2023-07-13 13:08:28 -0400
Branch: REL_12_STABLE [8d8f37149] 2023-07-13 13:08:33 -0400
Branch: REL_11_STABLE [9f70f6d4c] 2023-07-13 13:08:40 -0400
-->
     <para>
      Avoid useless double decompression of GiST index entries
      in <filename>contrib/intarray</filename> (Konstantin Knizhnik,
      Matthias van de Meent, Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_16_BR [e7bff46e5] 2023-05-19 12:37:58 +0900
Branch: REL_15_STABLE [2dd778221] 2023-05-19 12:38:15 +0900
Branch: REL_14_STABLE [e72580232] 2023-05-19 12:38:18 +0900
-->
     <para>
      Fix <filename>contrib/pageinspect</filename>'s
      <function>gist_page_items()</function> function to work when there
      are included index columns (Alexander Lakhin, Michael Paquier)
     </para>

     <para>
      Previously, if the index has included
      columns, <function>gist_page_items()</function> would fail to
      display those values on index leaf pages, or crash outright on
      non-leaf pages.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [51b2c0879] 2023-05-12 16:11:14 -0400
Branch: REL_15_STABLE [bc478a0a8] 2023-05-12 16:11:14 -0400
-->
     <para>
      In <application>psql</application>, ignore
      the <envar>PSQL_WATCH_PAGER</envar> environment variable when
      stdin/stdout are not a terminal (Tom Lane)
     </para>

     <para>
      This corresponds to the treatment of <envar>PSQL_PAGER</envar> in
      commands besides <command>\watch</command>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [016107478] 2023-06-04 13:05:54 -0400
Branch: REL_15_STABLE [ca9e79274] 2023-06-04 13:05:54 -0400
Branch: REL_14_STABLE [d6f549d7a] 2023-06-04 13:05:54 -0400
-->
     <para>
      Fix <application>pg_dump</application> to correctly handle new-style
      SQL-language functions whose bodies require parse-time dependencies
      on unique indexes (Tom Lane)
     </para>

     <para>
      Such cases can arise from <literal>GROUP BY</literal>
      and <literal>ON CONFLICT</literal> clauses, for example.  The
      function must then be postponed until after the unique index in the
      dump output, but <application>pg_dump</application> did not do that
      and instead printed a warning about <quote>could not resolve
      dependency loop</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master Release: REL_16_BR [b3f32a6c3] 2023-06-04 11:22:05 -0400
Branch: REL_15_STABLE [751ba1a7c] 2023-06-04 11:22:05 -0400
-->
     <para>
      Improve <application>pg_dump</application>'s display of details
      about dependency-loop problems (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master Release: REL_16_BR [8f5e42d33] 2023-05-25 12:36:18 +0200
Branch: REL_15_STABLE [34f511965] 2023-05-25 12:36:18 +0200
-->
     <para>
      Avoid crash in <application>pgbench</application> with an empty
      pipeline and prepared mode (&Aacute;lvaro Herrera)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [a5ea825f9] 2023-07-14 11:15:34 +0900
Branch: REL_16_STABLE [27da47122] 2023-07-14 11:16:03 +0900
Branch: REL_15_STABLE [eb3abec4b] 2023-07-14 11:16:06 +0900
Branch: REL_14_STABLE [7af65523a] 2023-07-14 11:16:08 +0900
Branch: REL_13_STABLE [bdaaf1bf1] 2023-07-14 11:16:10 +0900
Branch: REL_12_STABLE [7d27493b7] 2023-07-14 11:16:11 +0900
Branch: REL_11_STABLE [db9813819] 2023-07-14 11:16:13 +0900
-->
     <para>
      Ensure
      that <structname>pg_index</structname>.<structfield>indisreplident</structfield>
      is kept up-to-date in relation cache entries (Shruthi Gowda)
     </para>

     <para>
      This value could be stale in some cases.  There is no core code that
      relies on the relation cache's copy, so this is only a latent bug as
      far as Postgres itself is concerned; but there may be extensions for
      which it is a live bug.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tatsuo Ishii <ishii@postgresql.org>
Branch: master Release: REL_16_BR [ae66716bf] 2023-06-14 11:02:50 +0900
Branch: REL_15_STABLE [af26f28b9] 2023-06-14 11:11:18 +0900
-->
     <para>
      Fix <application>make_etags</application> script to work with
      non-Exuberant <application>ctags</application> (Masahiko Sawada)
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15-3">
  <title>Release 15.3</title>

  <formalpara>
  <title>Release date:</title>
  <para>2023-05-11</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.2.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.3</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, if you are upgrading from a version earlier than 15.1,
    see <xref linkend="release-15-1"/>.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [681d9e462] 2023-05-08 06:14:07 -0700
Branch: REL_15_STABLE [dbd5795e7] 2023-05-08 06:14:11 -0700
Branch: REL_14_STABLE [01e8182c7] 2023-05-08 06:14:11 -0700
Branch: REL_13_STABLE [2212f7db8] 2023-05-08 06:14:12 -0700
Branch: REL_12_STABLE [78119a0bf] 2023-05-08 06:14:12 -0700
Branch: REL_11_STABLE [23cb8eaeb] 2023-05-08 06:14:12 -0700
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [8d525d7b9] 2023-05-08 11:24:47 -0400
Branch: REL_15_STABLE [1b761d896] 2023-05-08 11:24:47 -0400
Branch: REL_14_STABLE [1913f63dc] 2023-05-08 11:24:47 -0400
Branch: REL_13_STABLE [feb9e7fbb] 2023-05-08 11:24:47 -0400
Branch: REL_12_STABLE [2cd843cc9] 2023-05-08 11:24:47 -0400
Branch: REL_11_STABLE [766e06140] 2023-05-08 11:24:47 -0400
-->
     <para>
      Prevent <command>CREATE SCHEMA</command> from defeating changes
      in <varname>search_path</varname> (Alexander Lakhin)
     </para>

     <para>
      Within a <command>CREATE SCHEMA</command> command, objects in the
      prevailing <varname>search_path</varname>, as well as those in the
      newly-created schema, would be visible even within a called
      function or script that attempted to set a
      secure <varname>search_path</varname>.  This could allow any user
      having permission to create a schema to hijack the privileges of a
      security definer function or extension script.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks
      Alexander Lakhin for reporting this problem.
      (CVE-2023-2454)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [ca73753b0] 2023-05-08 10:12:44 -0400
Branch: REL_15_STABLE [04e560604] 2023-05-08 10:12:44 -0400
Branch: REL_14_STABLE [f8d799eda] 2023-05-08 10:12:44 -0400
Branch: REL_13_STABLE [b8e28f04f] 2023-05-08 10:12:44 -0400
Branch: REL_12_STABLE [ee87b482c] 2023-05-08 10:12:45 -0400
Branch: REL_11_STABLE [473626cf0] 2023-05-08 10:12:45 -0400
-->
     <para>
      Enforce row-level security policies correctly after inlining a
      set-returning function (Stephen Frost, Tom Lane)
     </para>

     <para>
      If a set-returning SQL-language function refers to a table having
      row-level security policies, and it can be inlined into a calling
      query, those RLS policies would not get enforced properly in some
      cases involving re-using a cached plan under a different role.
      This could allow a user to see or modify rows that should have been
      invisible.
     </para>

     <para>
      The <productname>PostgreSQL</productname> Project thanks
      Wolfgang Walther for reporting this problem.
      (CVE-2023-2455)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [8a8661828] 2023-02-22 10:14:52 +0900
Branch: REL_15_STABLE [fa5dd460c] 2023-02-22 10:14:56 +0900
-->
     <para>
      Fix potential corruption of the template (source) database after
      <command>CREATE DATABASE</command> with the <literal>STRATEGY
      WAL_LOG</literal> option (Nathan Bossart, Ryo Matsumura)
     </para>

     <para>
      Improper buffer handling created a risk that any later modification
      of the template's <structname>pg_class</structname> catalog would be
      lost.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [5df319f3d] 2023-03-22 09:20:34 -0700
Branch: REL_15_STABLE [560bb56c6] 2023-03-22 09:26:23 -0700
-->
     <para>
      Fix memory leakage and unnecessary disk reads
      during <command>CREATE DATABASE</command> with the <literal>STRATEGY
      WAL_LOG</literal> option (Andres Freund)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [4dadd660f] 2023-04-28 19:29:12 +0900
Branch: REL_15_STABLE [b9ad73ad2] 2023-04-28 19:29:36 +0900
Branch: REL_14_STABLE [d29eba198] 2023-04-28 19:29:38 +0900
Branch: REL_13_STABLE [7e95a33b4] 2023-04-28 19:29:40 +0900
Branch: REL_12_STABLE [63f7e91ec] 2023-04-28 19:29:42 +0900
Branch: REL_11_STABLE [a9212716b] 2023-04-28 19:29:44 +0900
-->
     <para>
      Avoid crash when the new schema name is omitted
      in <command>CREATE SCHEMA</command> (Michael Paquier)
     </para>

     <para>
      The SQL standard allows writing <literal>CREATE SCHEMA AUTHORIZATION
      <replaceable>owner_name</replaceable></literal>, with the schema
      name defaulting to <replaceable>owner_name</replaceable>.  However
      some code paths expected the schema name to be present and would
      fail.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [326a33a28] 2023-03-28 11:39:24 -0400
Branch: REL_15_STABLE [bf5c4b3d9] 2023-03-28 11:36:50 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [e3ac85014] 2023-03-15 11:59:18 -0400
Branch: REL_15_STABLE [3908d6ae1] 2023-03-15 11:59:18 -0400
-->
     <para>
      Fix various planner failures with <command>MERGE</command>
      commands (Tom Lane)
     </para>

     <para>
      Planning could fail with errors like <quote>variable not found in
      subplan target list</quote> or <quote>PlaceHolderVar found where not
      expected</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [7b14e20b1] 2023-03-13 11:12:20 +0000
Branch: REL_15_STABLE [da6257eee] 2023-03-13 11:11:10 +0000
Branch: master [80a48e0f2] 2023-02-22 09:39:09 +0000
Branch: REL_15_STABLE [018af1cc1] 2023-02-22 09:41:28 +0000
-->
     <para>
      Fix the row count reported by <command>MERGE</command> for some
      corner cases (Dean Rasheed)
     </para>

     <para>
      The row count reported in the command tag counted rows that actually
      hadn't been modified due to a <literal>BEFORE ROW</literal> trigger
      returning NULL.  This is inconsistent with what happens in
      plain <command>UPDATE</command> or <command>DELETE</command>, so
      change it to not count such rows.  Also, avoid counting a row twice
      when <command>MERGE</command> moves it into a different partition of
      a partitioned table.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [9321c79c8] 2023-03-13 10:22:22 +0000
Branch: REL_15_STABLE [7d9a75713] 2023-03-13 10:23:42 +0000
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [fd923b5de] 2023-02-15 20:37:44 +0100
Branch: REL_15_STABLE [5d8ec1b9f] 2023-02-15 20:37:44 +0100
-->
     <para>
      Fix <command>MERGE</command> problems with concurrent updates
      (Dean Rasheed, &Aacute;lvaro Herrera)
     </para>

     <para>
      Some cases misbehaved if a row to be updated or deleted
      by <command>MERGE</command> had just been updated by a concurrent
      transaction.  This could lead to a crash, or the wrong merge action
      being executed, or no action at all.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [41e2c52fd] 2023-05-07 11:01:15 -0400
Branch: REL_15_STABLE [f200b9695] 2023-05-07 11:01:15 -0400
-->
     <para>
      Add support for decompiling <command>MERGE</command>
      commands (&Aacute;lvaro Herrera)
     </para>

     <para>
      This was overlooked when <command>MERGE</command> was added, but
      it's essential support for <command>MERGE</command> in new-style SQL
      functions.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [6949b921d] 2023-03-04 13:32:35 -0500
Branch: REL_15_STABLE [f61e60102] 2023-03-04 13:32:35 -0500
-->
     <para>
      Fix enabling/disabling of foreign-key triggers in partitioned tables
      (Tom Lane)
     </para>

     <para>
      <command>ALTER TABLE ... ENABLE/DISABLE TRIGGER</command> failed if
      applied to a partitioned table's foreign-key enforcement triggers,
      because it tried to locate the clone triggers for the partitions by
      name, and they do not have the same name.  Locate them by
      parent-trigger OID instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [a3c9d35ae] 2023-03-27 15:04:15 -0400
Branch: REL_15_STABLE [d90d59e25] 2023-03-27 15:04:02 -0400
Branch: REL_14_STABLE [334cc4c96] 2023-03-27 15:04:02 -0400
Branch: REL_13_STABLE [29a20ff06] 2023-03-27 15:04:02 -0400
Branch: REL_12_STABLE [cd07163c0] 2023-03-27 15:04:02 -0400
Branch: REL_11_STABLE [78838bc3d] 2023-03-27 15:04:02 -0400
-->
     <para>
      Disallow altering composite types that are stored in indexes
      (Tom Lane)
     </para>

     <para>
      <command>ALTER TYPE</command> disallows non-binary-compatible
      modifications of composite types if they are stored in any table
      columns.  (Perhaps that will be allowed someday, but it hasn't
      happened yet; the locking implications of rewriting many tables are
      daunting.)  We overlooked the possibility that an index might
      contain a composite type that doesn't also appear in its table.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f0d65c0ea] 2023-03-31 11:18:49 -0400
Branch: REL_15_STABLE [6e3698173] 2023-03-31 11:18:49 -0400
Branch: REL_14_STABLE [b0b55d8b8] 2023-03-31 11:18:49 -0400
Branch: REL_13_STABLE [bfb993b1b] 2023-03-31 11:18:49 -0400
Branch: REL_12_STABLE [e8d74aac5] 2023-03-31 11:18:49 -0400
-->
     <para>
      Disallow system columns as elements of foreign keys (Tom Lane)
     </para>

     <para>
      Since the removal of OID as a system column, there is no plausible
      use-case for this, and various bits of code no longer support it.
      Disallow it rather than trying to fix all the cases.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d66bb048c] 2023-03-10 13:52:44 -0500
Branch: REL_15_STABLE [59947bac7] 2023-03-10 13:52:28 -0500
Branch: REL_14_STABLE [53a53ea33] 2023-03-10 13:52:28 -0500
Branch: REL_13_STABLE [866fd004d] 2023-03-10 13:52:28 -0500
Branch: REL_12_STABLE [a30310833] 2023-03-10 13:52:28 -0500
Branch: REL_11_STABLE [6e2674d77] 2023-03-10 13:52:28 -0500
-->
     <para>
      Ensure that <command>COPY TO</command> from an RLS-enabled parent
      table does not copy any rows from child tables (Antonin Houska)
     </para>

     <para>
      The documentation is quite clear that <command>COPY TO</command>
      copies rows from only the named table, not any inheritance children
      it may have.  However, if row-level security was enabled on the table
      then this stopped being true.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [4c40995f6] 2023-05-04 11:48:23 -0400
Branch: REL_15_STABLE [ccb479e76] 2023-05-04 11:48:23 -0400
Branch: REL_14_STABLE [d5de344a5] 2023-05-04 11:48:23 -0400
Branch: REL_13_STABLE [9a72f499a] 2023-05-04 11:48:23 -0400
Branch: REL_12_STABLE [580df5078] 2023-05-04 11:48:23 -0400
Branch: REL_11_STABLE [4624aad61] 2023-05-04 11:48:23 -0400
-->
     <para>
      Avoid possible crash when <function>array_position()</function>
      or <function>array_positions()</function> is passed an empty array
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [b081fe419] 2023-03-14 19:17:31 -0400
Branch: REL_15_STABLE [a67c75f82] 2023-03-14 19:17:31 -0400
Branch: REL_14_STABLE [7cac19105] 2023-03-14 19:17:31 -0400
Branch: REL_13_STABLE [386a26023] 2023-03-14 19:17:31 -0400
Branch: REL_12_STABLE [6d3a9a60f] 2023-03-14 19:17:31 -0400
Branch: REL_11_STABLE [8e33fb9ef] 2023-03-14 19:17:31 -0400
-->
     <para>
      Fix possible out-of-bounds fetch in <function>to_char()</function>
      (Tom Lane)
     </para>

     <para>
      With bad luck this could have resulted in a server crash.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d7056bc1c] 2023-03-01 11:30:31 -0500
Branch: REL_15_STABLE [eae09137d] 2023-03-01 11:30:17 -0500
Branch: REL_14_STABLE [1a9356f65] 2023-03-01 11:30:17 -0500
Branch: REL_13_STABLE [3b37e8442] 2023-03-01 11:30:17 -0500
Branch: REL_12_STABLE [b162660d3] 2023-03-01 11:30:17 -0500
Branch: REL_11_STABLE [b1a9d8ef2] 2023-03-01 11:30:17 -0500
-->
     <para>
      Avoid buffer overread in <function>translate()</function> function
      (Daniil Anisimov)
     </para>

     <para>
      When using the deletion feature, the function might fetch the byte
      just after the input string, creating a small risk of crash.
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [f413941f4] 2023-03-17 12:08:46 -0700
Branch: REL_15_STABLE [8b87e9291] 2023-03-17 12:07:47 -0700
-->
     <para>
      Adjust text-search-related character classification logic to
      correctly detect whether the prevailing locale
      is <literal>C</literal> (Jeff Davis)
     </para>

     <para>
      This code got confused if the database's default collation uses ICU.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5e80d3515] 2023-02-12 12:50:55 -0500
Branch: REL_15_STABLE [0ef65d0f5] 2023-02-12 12:50:55 -0500
-->
     <para>
      Avoid possible crash on empty input for type <type>interval</type>
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f0d0394e8] 2023-02-20 16:55:59 -0500
Branch: REL_15_STABLE [ded5ede27] 2023-02-20 16:55:59 -0500
-->
     <para>
      Re-allow exponential notation in ISO-8601 interval fields
      (Tom Lane)
     </para>

     <para>
      Interval input like <literal>P0.1e10D</literal> isn't officially
      sanctioned by ISO-8601, but we accepted it for a long time before
      version 15, so re-allow it.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [25a7812cd] 2023-03-13 15:19:00 -0400
Branch: REL_15_STABLE [74a1a36d7] 2023-03-13 15:19:00 -0400
Branch: REL_14_STABLE [0ee9d685d] 2023-03-13 15:19:00 -0400
Branch: REL_13_STABLE [52e9a7816] 2023-03-13 15:19:00 -0400
Branch: REL_12_STABLE [c25a929a6] 2023-03-13 15:19:00 -0400
Branch: REL_11_STABLE [234941a3b] 2023-03-13 15:19:00 -0400
-->
     <para>
      Fix error cursor setting for parse errors in JSON string literals
      (Tom Lane)
     </para>

     <para>
      Most cases in which a syntax error is detected in a string literal
      within a JSON value failed to set the error cursor appropriately.
      This led at least to an unhelpful error message (pointing to the
      token before the string, rather than the actual trouble spot), and
      could even result in a crash in v14 and later.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [be504a3e9] 2023-03-07 21:52:32 -0800
Branch: REL_15_STABLE [391f08fd6] 2023-03-07 21:36:48 -0800
Branch: REL_14_STABLE [324281fd5] 2023-03-07 21:36:49 -0800
Branch: REL_13_STABLE [e6d77f22c] 2023-03-07 21:36:51 -0800
Branch: REL_12_STABLE [3c92f7e9d] 2023-03-07 21:36:52 -0800
-->
     <para>
      Fix data corruption due to <varname>vacuum_defer_cleanup_age</varname>
      being larger than the current 64-bit xid (Andres Freund)
     </para>

     <para>
      In v14 and later with non-default settings
      of <varname>vacuum_defer_cleanup_age</varname>, it was possible to
      compute a very large vacuum cleanup horizon xid, leading to vacuum
      removing rows that are still live.  v12 and v13 have a lesser form
      of the same problem affecting only GiST indexes, which could lead to
      index pages getting recycled too early.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [30dbdbe75] 2023-03-13 12:40:28 -0400
Branch: REL_15_STABLE [5fd61bdc1] 2023-03-13 12:40:28 -0400
Branch: REL_14_STABLE [096e70805] 2023-03-13 12:40:28 -0400
Branch: REL_13_STABLE [bc0bcce2e] 2023-03-13 12:40:28 -0400
Branch: REL_12_STABLE [62a91a1b0] 2023-03-13 12:40:28 -0400
Branch: REL_11_STABLE [0736b1131] 2023-03-13 12:40:28 -0400
-->
     <para>
      Fix parser's failure to detect some cases of improperly-nested
      aggregates (Tom Lane)
     </para>

     <para>
      This oversight could lead to executor failures for queries that
      should have been rejected as invalid.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [6c3b697b1] 2023-03-31 12:13:05 +1300
Branch: REL_15_STABLE [df567fbf6] 2023-03-31 12:13:34 +1300
Branch: REL_14_STABLE [211016220] 2023-03-31 12:14:04 +1300
Branch: REL_13_STABLE [8d684c445] 2023-03-31 12:14:31 +1300
Branch: REL_12_STABLE [33510bc64] 2023-03-31 12:15:07 +1300
Branch: REL_11_STABLE [07554c99d] 2023-03-31 12:15:39 +1300
-->
     <para>
      Fix data structure corruption during parsing of
      serial <literal>SEQUENCE NAME</literal> options (David Rowley)
     </para>

     <para>
      This can lead to trouble if an event trigger captures the corrupted
      parse tree.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [88ceac5d7] 2023-04-12 10:46:38 -0400
Branch: REL_15_STABLE [f4badbcf4] 2023-04-12 10:46:30 -0400
Branch: REL_14_STABLE [0dd55ef9b] 2023-04-12 10:46:30 -0400
Branch: REL_13_STABLE [96c698e3f] 2023-04-12 10:46:30 -0400
Branch: REL_12_STABLE [953ff99c2] 2023-04-12 10:46:30 -0400
Branch: REL_11_STABLE [60c8aeaf6] 2023-04-12 10:46:30 -0400
-->
     <para>
      Correctly update plan nodes' parallel-safety markings when moving
      initplans from one node to another (Tom Lane)
     </para>

     <para>
      This planner oversight could lead to <quote>subplan was not
      initialized</quote> errors at runtime.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [684ffac8c] 2023-03-14 11:10:45 -0400
Branch: REL_15_STABLE [3b4594443] 2023-03-14 11:10:45 -0400
Branch: REL_14_STABLE [7c509f7e5] 2023-03-14 11:10:45 -0400
-->
     <para>
      Avoid failure with PlaceHolderVars in extended-statistics code
      (Tom Lane)
     </para>

     <para>
      Use of dependency-type extended statistics could fail with
      <quote>PlaceHolderVar found where not expected</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [eb7d043c9] 2023-03-17 15:49:53 +1300
Branch: REL_15_STABLE [371e3daaa] 2023-03-17 15:51:00 +1300
-->
     <para>
      Fix incorrect tests for whether a qual clause applied to a subquery
      can be transformed into a window aggregate <quote>run
      condition</quote> within the subquery (David Rowley)
     </para>

     <para>
      A SubPlan within such a clause would cause assertion failures or
      incorrect answers, as would some other unusual cases.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [836c31ba5] 2023-02-13 17:11:03 +1300
Branch: REL_15_STABLE [a9fa6d79a] 2023-02-13 17:10:31 +1300
Branch: REL_14_STABLE [4aa43ba21] 2023-02-13 17:09:55 +1300
Branch: REL_13_STABLE [301eb3ee4] 2023-02-13 17:09:26 +1300
Branch: REL_12_STABLE [ac55abd33] 2023-02-13 17:08:46 +1300
Branch: REL_11_STABLE [8d2a8581b] 2023-02-13 17:07:04 +1300
-->
     <para>
      Disable the inverse-transition optimization for window aggregates
      when the call contains sub-SELECTs (David Rowley)
     </para>

     <para>
      This optimization requires that the aggregate's argument expressions
      have repeatable results, which might not hold for a sub-SELECT.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [554841699] 2023-03-26 13:41:06 -0400
Branch: REL_15_STABLE [7c4873438] 2023-03-26 13:41:06 -0400
Branch: REL_14_STABLE [11213d446] 2023-03-26 13:41:06 -0400
Branch: REL_13_STABLE [1bbbe1460] 2023-03-26 13:41:06 -0400
Branch: REL_12_STABLE [ad5fe7420] 2023-03-26 13:41:06 -0400
Branch: REL_11_STABLE [ae320fc21] 2023-03-26 13:41:06 -0400
-->
     <para>
      Fix oversights in execution of nested <literal>ARRAY[]</literal>
      constructs (Alexander Lakhin, Tom Lane)
     </para>

     <para>
      Correctly detect overflow of the total space needed for the result
      array, avoiding a possible crash due to undersized output
      allocation.  Also ensure that any trailing padding space in the
      result array is zeroed; while leaving garbage there is harmless for
      most purposes, it can result in odd behavior later.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [064eb89e8] 2023-04-15 12:01:39 -0400
Branch: REL_15_STABLE [c53ed26ea] 2023-04-15 12:01:39 -0400
Branch: REL_14_STABLE [9b104a27c] 2023-04-15 12:01:39 -0400
Branch: REL_13_STABLE [7428aecdd] 2023-04-15 12:01:39 -0400
Branch: REL_12_STABLE [048caf8d7] 2023-04-15 12:01:39 -0400
-->
     <para>
      Prevent crash when updating a field within an
      array-of-domain-over-composite-type column (Dmitry Dolgov)
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [e0693faf7] 2023-04-14 16:20:27 +1200
Branch: REL_15_STABLE [0c09160e1] 2023-04-14 16:21:07 +1200
Branch: REL_14_STABLE [ae85fb828] 2023-04-14 16:21:42 +1200
Branch: REL_13_STABLE [6848f0c67] 2023-04-14 16:22:11 +1200
Branch: REL_12_STABLE [0b2e77ce2] 2023-04-14 16:22:46 +1200
Branch: REL_11_STABLE [1c19e2863] 2023-04-14 16:23:11 +1200
-->
     <para>
      Fix partition pruning logic for partitioning on boolean columns
      (David Rowley)
     </para>

     <para>
      Pruning with a condition like <literal>boolcol IS NOT TRUE</literal>
      was done incorrectly, leading to possibly not returning rows in
      which <literal>boolcol</literal> is NULL.  Also, the rather unlikely
      case of partitioning on <literal>NOT boolcol</literal> was handled
      incorrectly.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [8d578b9b2] 2023-03-21 14:29:34 +1300
Branch: REL_15_STABLE [c03c6e8cf] 2023-03-21 14:32:14 +1300
Branch: REL_14_STABLE [1b9e42e82] 2023-03-21 14:37:33 +1300
Branch: REL_13_STABLE [6e94d62e3] 2023-03-21 14:38:59 +1300
Branch: REL_12_STABLE [44d44aa97] 2023-03-21 14:41:35 +1300
Branch: REL_11_STABLE [ef16d2724] 2023-03-21 14:43:07 +1300
-->
     <para>
      Fix race condition in per-batch cleanup during parallel hash join
      (Thomas Munro, Melanie Plageman)
     </para>

     <para>
      A crash was possible given unlucky timing and
      <varname>parallel_leader_participation</varname>
      = <literal>off</literal> (which is not the default).
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [7fee7871b] 2023-03-06 18:31:27 -0500
Branch: REL_15_STABLE [70ef50954] 2023-03-06 18:31:16 -0500
Branch: REL_14_STABLE [1e05ea51d] 2023-03-06 18:31:16 -0500
Branch: REL_13_STABLE [4a94cbd02] 2023-03-06 18:31:16 -0500
Branch: REL_12_STABLE [23b75dd03] 2023-03-06 18:31:16 -0500
-->
     <para>
      Recalculate <literal>GENERATED</literal> columns after an
      EvalPlanQual check (Tom Lane)
     </para>

     <para>
      In <literal>READ COMMITTED</literal> isolation mode, the effects of
      a row update might need to get reapplied to a newer version of the
      row than the query found originally.  If so, we need to recompute
      any <literal>GENERATED</literal> columns, in case they depend on
      columns that were changed by the concurrent update.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [579ee5df1] 2023-03-20 13:28:47 +1300
Branch: REL_15_STABLE [8de4660a5] 2023-03-20 13:30:15 +1300
Branch: REL_14_STABLE [f654f343c] 2023-03-20 13:30:55 +1300
-->
     <para>
      Fix memory leak in Memoize plan execution (David Rowley)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [806fad757] 2023-04-25 09:42:19 +0900
Branch: REL_15_STABLE [aa6177c88] 2023-04-25 09:42:33 +0900
Branch: REL_14_STABLE [4cc56f8ed] 2023-04-25 09:42:36 +0900
-->
     <para>
      Fix buffer refcount leak when using batched inserts for a foreign
      table included in a partitioned tree (Alexander Pyhalov)
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [720de00af] 2023-03-15 13:58:18 +1300
Branch: REL_15_STABLE [d9c9c43af] 2023-03-15 14:02:49 +1300
Branch: REL_14_STABLE [2bef57ee8] 2023-03-15 14:05:27 +1300
-->
     <para>
      Restore support for
      sub-millisecond <varname>vacuum_cost_delay</varname> settings
      (Thomas Munro)
     </para>
    </listitem>

    <listitem>
<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
Branch: master [bfac8f8bc] 2023-04-25 13:54:10 +0200
Branch: REL_15_STABLE [0319b306e] 2023-04-25 13:54:10 +0200
Branch: REL_14_STABLE [0e8e5e856] 2023-04-25 13:54:10 +0200
Branch: REL_13_STABLE [b95f36f86] 2023-04-25 13:54:10 +0200
Branch: REL_12_STABLE [cba3c8f6d] 2023-04-25 13:54:10 +0200
Branch: REL_11_STABLE [0151d2c5f] 2023-04-25 13:54:10 +0200
-->
     <para>
      Don't balance vacuum cost delay when a table has a
      per-relation <varname>vacuum_cost_delay</varname> setting of zero
      (Masahiko Sawada)
     </para>

     <para>
      Delay balancing is supposed to be disabled whenever autovacuum is
      processing a table with a
      per-relation <varname>vacuum_cost_delay</varname> setting, but this
      was done only for positive settings, not zero.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [99be6feec] 2023-03-07 18:21:53 -0500
Branch: REL_15_STABLE [76d2177fb] 2023-03-07 18:21:53 -0500
Branch: REL_14_STABLE [9f1e51b59] 2023-03-07 18:21:37 -0500
Branch: REL_13_STABLE [695b34ab3] 2023-03-07 18:21:37 -0500
Branch: REL_12_STABLE [5a19da58e] 2023-03-07 18:21:37 -0500
Branch: REL_11_STABLE [721626cb5] 2023-03-07 18:21:37 -0500
-->
     <para>
      Fix corner-case crashes when columns have been added to the end of a
      view (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [87f3667ec] 2023-02-25 14:44:14 -0500
Branch: REL_15_STABLE [a033f9165] 2023-02-25 14:44:14 -0500
Branch: REL_14_STABLE [9eaba0602] 2023-02-25 14:44:14 -0500
Branch: REL_13_STABLE [1e199c259] 2023-02-25 14:44:14 -0500
Branch: REL_12_STABLE [904b171a4] 2023-02-25 14:44:14 -0500
Branch: REL_11_STABLE [ffec64ba8] 2023-02-25 14:44:14 -0500
-->
     <para>
      Repair rare failure of MULTIEXPR_SUBLINK subplans in partitioned
      updates (Andres Freund, Tom Lane)
     </para>

     <para>
      Use of the syntax <literal>INSERT ... ON CONFLICT DO UPDATE SET (c1,
      ...) = (SELECT ...)</literal> with a partitioned target table could
      result in failure if any child table is dissimilar from the parent
      (for example, different physical column order).
      This typically manifested as failure of consistency checks in the
      executor; but a crash or incorrect data updates are also possible.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [75c737636] 2023-02-23 10:53:01 +0000
Branch: REL_15_STABLE [940b54743] 2023-02-23 10:54:51 +0000
Branch: REL_14_STABLE [f0423bea7] 2023-02-23 10:55:48 +0000
Branch: REL_13_STABLE [226da3d47] 2023-02-23 10:56:41 +0000
Branch: REL_12_STABLE [98b83b734] 2023-02-23 10:57:46 +0000
Branch: REL_11_STABLE [e68b133c3] 2023-02-23 10:58:43 +0000
-->
     <para>
      Fix handling of <literal>DEFAULT</literal> markers within a
      multi-row <literal>INSERT ... VALUES</literal> query on a view that
      has a <literal>DO ALSO INSERT ... SELECT</literal> rule (Dean
      Rasheed)
     </para>

     <para>
      Such cases typically failed with <quote>unrecognized node
      type</quote> errors or assertion failures.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [a7d71c41d] 2023-02-25 14:41:12 +0000
Branch: REL_15_STABLE [8e5b4e001] 2023-02-25 14:43:57 +0000
Branch: REL_14_STABLE [27ff93d18] 2023-02-25 14:44:49 +0000
Branch: REL_13_STABLE [39ad791e8] 2023-02-25 14:45:44 +0000
Branch: REL_12_STABLE [4fd093af7] 2023-02-25 14:47:03 +0000
Branch: REL_11_STABLE [79f194cc0] 2023-02-25 14:48:08 +0000
-->
     <para>
      Support references to <literal>OLD</literal>
      and <literal>NEW</literal> within subqueries in rule actions
      (Dean Rasheed, Tom Lane)
     </para>

     <para>
      Such references are really lateral references, but the server could
      crash if the subquery wasn't explicitly marked
      with <literal>LATERAL</literal>.  Arrange to do that implicitly when
      necessary.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [393430f57] 2023-02-17 16:40:34 -0500
Branch: REL_15_STABLE [c8a5f1685] 2023-02-17 16:40:34 -0500
Branch: REL_14_STABLE [14345f3c6] 2023-02-17 16:40:34 -0500
Branch: REL_13_STABLE [4efb4f0d4] 2023-02-17 16:40:34 -0500
Branch: REL_12_STABLE [3dd287c14] 2023-02-17 16:40:34 -0500
Branch: REL_11_STABLE [df931e9ab] 2023-02-17 16:40:34 -0500
-->
     <para>
      When decompiling a rule or SQL function body
      containing <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
      within <command>WITH</command>, take care to print the correct alias
      for the target table (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [47c0accbe] 2023-03-06 15:07:15 +1300
Branch: REL_15_STABLE [055990904] 2023-03-06 16:05:47 +1300
Branch: REL_14_STABLE [e9051ecd5] 2023-03-06 16:17:22 +1300
Branch: REL_13_STABLE [a0f55fc86] 2023-03-06 16:28:34 +1300
Branch: REL_12_STABLE [afa122e41] 2023-03-06 16:41:34 +1300
Branch: master [65e388d41] 2023-03-09 16:33:24 +1300
Branch: REL_15_STABLE [af397c6c2] 2023-03-09 16:56:51 +1300
Branch: REL_14_STABLE [d811d74be] 2023-03-09 16:57:11 +1300
Branch: REL_13_STABLE [ae632f7a3] 2023-03-09 17:09:57 +1300
Branch: REL_12_STABLE [e30fd0942] 2023-03-09 17:25:20 +1300
Branch: REL_11_STABLE [d1c0f81e7] 2023-03-09 17:26:06 +1300
-->
     <para>
      Fix glitches in <literal>SERIALIZABLE READ ONLY</literal>
      optimization (Thomas Munro)
     </para>

     <para>
      Transactions already marked as <quote>doomed</quote> confused the
      safe-snapshot optimization for <literal>SERIALIZABLE READ
      ONLY</literal> transactions.  The optimization was unnecessarily
      skipped in some cases.  In other cases an assertion failure occurred
      (but there was no problem in non-assert builds).
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [05172f1f3] 2023-02-23 15:40:42 -0500
Branch: REL_15_STABLE [cef1c9c0c] 2023-02-23 15:40:28 -0500
Branch: REL_14_STABLE [0f78df719] 2023-02-23 15:40:28 -0500
Branch: REL_13_STABLE [861e9e486] 2023-02-23 15:40:28 -0500
Branch: REL_12_STABLE [95558bc8f] 2023-02-23 15:40:28 -0500
Branch: REL_11_STABLE [44dbc960f] 2023-02-23 15:40:28 -0500
-->
     <para>
      Avoid leaking cache callback slots in
      the <literal>pgoutput</literal> logical decoding plugin (Shi Yu)
     </para>

     <para>
      Multiple cycles of starting up and shutting down the plugin within a
      single session would eventually lead to an <quote>out of
      relcache_callback_list slots</quote> error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [cd115c353] 2023-04-23 13:58:41 +0300
Branch: REL_15_STABLE [6e7361c85] 2023-04-23 14:00:06 +0300
Branch: REL_14_STABLE [9ef5a3583] 2023-04-23 14:00:13 +0300
Branch: REL_13_STABLE [02191136c] 2023-04-23 14:00:16 +0300
-->
     <para>
      Avoid unnecessary calls to custom validators for index operator
      class options (Alexander Korotkov)
     </para>

     <para>
      This change fixes some cases where an unexpected error was thrown.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [e72910f80] 2023-02-19 01:46:58 +0100
Branch: REL_15_STABLE [305d89ad9] 2023-02-19 01:48:04 +0100
Branch: REL_14_STABLE [f3daa3116] 2023-02-19 01:48:22 +0100
-->
     <para>
      Avoid useless work while scanning a multi-column BRIN index with
      multiple scan keys (Tomas Vondra)
     </para>

     <para>
      The existing code effectively considered only the last scan key
      while deciding whether a range matched, thus usually scanning more
      of the index than it needed to.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [e85831268] 2023-03-20 10:24:14 +0100
Branch: REL_15_STABLE [0c7726c28] 2023-03-20 10:20:35 +0100
Branch: REL_14_STABLE [6a78a42fe] 2023-03-20 10:16:54 +0100
-->
     <para>
      Fix netmask handling in BRIN inet_minmax_multi_ops opclass
      (Tomas Vondra)
     </para>

     <para>
      This error triggered an assertion failure in assert-enabled builds,
      but is mostly harmless in production builds.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [8e5eef50c] 2023-03-29 11:31:30 -0400
Branch: REL_15_STABLE [2dc77adc7] 2023-03-29 11:31:30 -0400
Branch: REL_14_STABLE [a1904c9ce] 2023-03-29 11:31:30 -0400
Branch: REL_13_STABLE [2adb6adad] 2023-03-29 11:31:30 -0400
Branch: REL_12_STABLE [d2a1d4b19] 2023-03-29 11:31:30 -0400
Branch: REL_11_STABLE [b5c6776c1] 2023-03-29 11:31:30 -0400
-->
     <para>
      Fix dereference of dangling pointer during buffering build of a GiST
      index (Alexander Lakhin)
     </para>

     <para>
      This error seems to usually be harmless in production builds, as the
      fetched value is noncritical; but in principle it could cause a
      server crash.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [adedf54e6] 2023-03-23 11:58:36 +0530
Branch: REL_15_STABLE [b6bf90edc] 2023-03-23 11:46:16 +0530
Branch: REL_14_STABLE [9dac02c77] 2023-03-23 11:32:22 +0530
Branch: REL_13_STABLE [be52fff91] 2023-03-23 11:21:13 +0530
Branch: REL_12_STABLE [0f2d4adbf] 2023-03-23 11:08:38 +0530
Branch: master [b797def59] 2023-03-21 09:47:21 +0530
Branch: REL_15_STABLE [3c12407f4] 2023-03-21 09:40:41 +0530
Branch: REL_14_STABLE [65ead7696] 2023-03-21 09:18:51 +0530
Branch: REL_13_STABLE [751d6676d] 2023-03-21 09:07:37 +0530
Branch: REL_12_STABLE [fc63e6ba8] 2023-03-21 08:50:23 +0530
Branch: REL_11_STABLE [4cdaea7a2] 2023-03-21 08:39:00 +0530
-->
     <para>
      Ignore dropped columns and generated columns during logical
      replication of an update or delete action (Onder Kalaci, Shi Yu)
     </para>

     <para>
      Replication with the <literal>REPLICA IDENTITY FULL</literal> option
      failed if the table contained such columns.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [58f5edf84] 2023-05-05 21:25:44 +0900
Branch: REL_15_STABLE [d31dab9a5] 2023-05-05 21:25:50 +0900
Branch: REL_14_STABLE [ae4ffa722] 2023-05-05 21:25:56 +0900
Branch: REL_13_STABLE [cedcc4191] 2023-05-05 21:26:02 +0900
-->
     <para>
      Correct the name of the wait event for SLRU buffer I/O for commit
      timestamps (Alexander Lakhin)
     </para>

     <para>
      This wait event is named <literal>CommitTsBuffer</literal> according
      to the documentation, but the code had it
      as <literal>CommitTSBuffer</literal>.  Change the code to match the
      documentation, as that way is more consistent with the naming of
      related wait events.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [84cc14267] 2023-04-26 07:10:06 +0900
Branch: REL_15_STABLE [1ed1b84bd] 2023-04-26 07:30:42 +0900
Branch: REL_14_STABLE [aeb6f4b3b] 2023-04-26 07:30:47 +0900
-->
     <para>
      Re-activate reporting of wait event <literal>SLRUFlushSync</literal>
      (Thomas Munro)
     </para>

     <para>
      Reporting of this type of wait was accidentally removed in code
      refactoring.
     </para>
    </listitem>

    <listitem>
<!--
Author: Nathan Bossart <nathan@postgresql.org>
Branch: master [b72623671] 2023-04-27 14:31:17 -0700
Branch: REL_15_STABLE [c98b06e2f] 2023-04-27 14:31:33 -0700
Branch: REL_14_STABLE [137003036] 2023-04-27 14:32:40 -0700
Branch: REL_13_STABLE [be40dd63e] 2023-04-27 14:32:55 -0700
-->
     <para>
      Avoid possible underflow when calculating how many WAL segments to
      keep (Kyotaro Horiguchi)
     </para>

     <para>
      This could result in not honoring <varname>wal_keep_size</varname>
      accurately.
     </para>
    </listitem>

    <listitem>
<!--
Author: Robert Haas <rhaas@postgresql.org>
Branch: master [8a2f783cc] 2023-02-06 10:51:08 -0500
Branch: REL_15_STABLE [ecb01e6eb] 2023-02-10 16:27:05 -0500
-->
     <para>
      Disable startup progress reporting overhead in standby mode
      (Bharath Rupireddy)
     </para>

     <para>
      In standby mode, we don't actually report progress of recovery,
      but we were doing work to track it anyway.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [9244c11af] 2023-02-15 10:12:16 +0900
Branch: REL_15_STABLE [5fd61055e] 2023-02-15 10:12:31 +0900
Branch: REL_14_STABLE [864f80fea] 2023-02-15 10:12:33 +0900
Branch: REL_13_STABLE [2eb8e54cc] 2023-02-15 10:12:36 +0900
Branch: REL_12_STABLE [a40e7b75e] 2023-02-15 10:12:38 +0900
Branch: REL_11_STABLE [88d606f7c] 2023-02-15 10:12:40 +0900
-->
     <para>
      Support RSA-PSS certificates with SCRAM-SHA-256 channel binding
      (Jacob Champion, Heikki Linnakangas)
     </para>

     <para>
      This feature requires building with OpenSSL 1.1.1 or newer.  Both
      the server and <application>libpq</application> are affected.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [d41a178b3] 2023-03-15 13:24:47 +1300
Branch: REL_15_STABLE [06066915d] 2023-03-15 13:25:56 +1300
Branch: REL_14_STABLE [9b6e0b9c3] 2023-03-15 13:30:58 +1300
Branch: REL_13_STABLE [9f1c64018] 2023-03-15 13:31:34 +1300
Branch: REL_12_STABLE [836288427] 2023-03-15 13:32:21 +1300
Branch: REL_11_STABLE [5ff8e69d8] 2023-03-15 13:32:41 +1300
Branch: master [6a9229da6] 2023-03-17 10:44:46 +1300
Branch: REL_15_STABLE [75e7378f6] 2023-03-17 10:45:20 +1300
Branch: REL_14_STABLE [00fc4b3a3] 2023-03-17 09:52:45 +1300
Branch: REL_13_STABLE [798dae9f6] 2023-03-17 09:54:03 +1300
Branch: REL_12_STABLE [6f508b8bc] 2023-03-17 09:54:51 +1300
Branch: REL_11_STABLE [9d6c34397] 2023-03-17 09:58:08 +1300
Branch: master [10b6745d3] 2023-03-17 14:44:12 +1300
Branch: REL_15_STABLE [e8a774d00] 2023-03-17 14:46:03 +1300
Branch: REL_14_STABLE [1c0d4affa] 2023-03-17 14:46:50 +1300
Branch: REL_13_STABLE [77a8133c9] 2023-03-17 14:47:16 +1300
Branch: REL_12_STABLE [8fcd1517f] 2023-03-17 14:47:38 +1300
Branch: REL_11_STABLE [b23f2a729] 2023-03-17 14:48:08 +1300
-->
     <para>
      Avoid race condition with process ID tracking on Windows (Thomas Munro)
     </para>

     <para>
      The operating system could recycle a PID before the postmaster
      observed that that child process was gone.  This could lead to
      tracking more than one child with the same PID, resulting in
      confusion.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [e35ded295] 2023-04-20 10:34:46 +1200
Branch: REL_15_STABLE [63a03aea6] 2023-04-21 10:02:25 +1200
-->
     <para>
      Fix <function>list_copy_head()</function> to work correctly on an
      empty List (David Rowley)
     </para>

     <para>
      This case is not known to be reached by any
      core <productname>PostgreSQL</productname> code, but extensions
      might rely on it working.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [d0460a31d] 2023-02-22 13:23:09 +0000
Branch: REL_15_STABLE [576b25bfd] 2023-02-22 13:24:51 +0000
Branch: REL_14_STABLE [482ab3e4f] 2023-02-22 13:26:20 +0000
Branch: REL_13_STABLE [906356cf6] 2023-02-22 13:27:29 +0000
Branch: REL_12_STABLE [52dbd9f84] 2023-02-22 13:28:30 +0000
Branch: REL_11_STABLE [83a54d966] 2023-02-22 13:29:39 +0000
-->
     <para>
      Add missing cases to <function>SPI_result_code_string()</function>
      (Dean Rasheed)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [b3e184a5d] 2023-02-21 18:48:15 -0500
Branch: REL_15_STABLE [f6a55c1d5] 2023-02-21 18:47:46 -0500
Branch: REL_14_STABLE [dc44180f6] 2023-02-21 18:47:47 -0500
Branch: REL_13_STABLE [99e74cd23] 2023-02-21 18:47:47 -0500
Branch: REL_12_STABLE [463bef383] 2023-02-21 18:47:47 -0500
Branch: REL_11_STABLE [21bd818d0] 2023-02-21 18:47:47 -0500
-->
     <para>
      Fix erroneous Valgrind markings
      in <function>AllocSetRealloc()</function> (Karina Litskevich)
     </para>

     <para>
      In the unusual case where the size of a large (&gt;8kB) palloc chunk
      is decreased, a Valgrind-aware build would mismark the defined-ness
      state of the memory released from the chunk, possibly causing
      incorrect results during Valgrind testing.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [0d3b49d4a] 2023-02-22 10:51:34 +0000
Branch: REL_15_STABLE [d8c3b65db] 2023-02-22 10:54:57 +0000
-->
     <para>
      Fix assertion failure for <command>MERGE</command> into a
      partitioned table with row-level security enabled (Dean Rasheed)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [7fe1aa991] 2023-02-22 15:24:18 +0100
Branch: REL_15_STABLE [949ac32e1] 2023-02-22 16:48:30 +0100
Branch: REL_14_STABLE [8b9cbd42b] 2023-02-22 16:25:45 +0100
Branch: REL_13_STABLE [4df581fa0] 2023-02-22 16:09:30 +0100
Branch: REL_12_STABLE [497f863f0] 2023-02-22 15:50:37 +0100
Branch: REL_11_STABLE [8de91ebf2] 2023-02-22 15:35:19 +0100
-->
     <para>
      Avoid assertion failure when decoding a transactional logical
      replication message (Tomas Vondra)
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [c04c6c5d6] 2023-04-21 08:19:41 -0700
Branch: REL_15_STABLE [109363de0] 2023-04-21 08:20:17 -0700
Branch: REL_14_STABLE [dde926b0f] 2023-04-21 08:20:32 -0700
Branch: REL_13_STABLE [79a66c617] 2023-04-21 08:20:47 -0700
Branch: REL_12_STABLE [5bcb15b81] 2023-04-21 08:21:04 -0700
Branch: REL_11_STABLE [106a1bf82] 2023-04-21 08:21:18 -0700
-->
     <para>
      Avoid locale sensitivity when processing regular expression escapes
      (Jeff Davis)
     </para>

     <para>
      A backslash followed by a non-ASCII character could sometimes cause
      an assertion failure, depending on the prevailing locale.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [3e383f9b6] 2023-04-17 14:22:26 -0400
Branch: REL_15_STABLE [2207df7c3] 2023-04-17 14:22:06 -0400
Branch: REL_14_STABLE [72a914e9f] 2023-04-17 14:22:06 -0400
Branch: REL_13_STABLE [86874a6df] 2023-04-17 14:22:06 -0400
Branch: REL_12_STABLE [9b0c1f213] 2023-04-17 14:22:06 -0400
Branch: REL_11_STABLE [c796d7296] 2023-04-17 14:22:06 -0400
-->
     <para>
      Avoid trying to write an empty WAL record
      in <function>log_newpage_range()</function> when the last few pages
      in the specified range are empty (Matthias van de Meent)
     </para>

     <para>
      It is not entirely clear whether this case is reachable in released
      branches, but if it is then an assertion failure could occur.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [441ee1677] 2023-04-24 14:19:46 -0400
Branch: REL_15_STABLE [c1598d85f] 2023-04-24 14:19:46 -0400
Branch: REL_14_STABLE [2ba890ce7] 2023-04-24 14:19:46 -0400
Branch: REL_13_STABLE [bfa691087] 2023-04-24 14:19:46 -0400
Branch: REL_12_STABLE [ee71cad9a] 2023-04-24 14:19:46 -0400
-->
     <para>
      Fix session-lifespan memory leakage in <application>plpgsql</application>
      <literal>DO</literal> blocks that use cast expressions
      (Ajit Awekar, Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f47004add] 2023-04-29 13:06:44 -0400
Branch: REL_15_STABLE [ce9a1a3ea] 2023-04-29 13:06:44 -0400
Branch: REL_14_STABLE [1e868bb6c] 2023-04-29 13:06:44 -0400
Branch: REL_13_STABLE [ee24b5e79] 2023-04-29 13:06:44 -0400
Branch: REL_12_STABLE [900a8d526] 2023-04-29 13:06:44 -0400
Branch: REL_11_STABLE [ea96fbe77] 2023-04-29 13:06:44 -0400
-->
     <para>
      Tighten array dimensionality checks when converting Perl
      list structures to multi-dimensional SQL arrays (Tom Lane)
     </para>

     <para>
      <application>plperl</application> could misbehave when the nesting
      of sub-lists is inconsistent so that the data does not represent a
      rectangular array of values.  Such cases now produce errors, but
      previously they could result in a crash or garbage output.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [0553528e7] 2023-05-04 11:00:33 -0400
Branch: REL_15_STABLE [b7001c6b6] 2023-05-04 11:00:33 -0400
Branch: REL_14_STABLE [9d517339e] 2023-05-04 11:00:33 -0400
Branch: REL_13_STABLE [0e6354ed9] 2023-05-04 11:00:33 -0400
Branch: REL_12_STABLE [b7fcf3824] 2023-05-04 11:00:33 -0400
Branch: REL_11_STABLE [aa7e5e404] 2023-05-04 11:00:33 -0400
Branch: master [81eaaf65e] 2023-04-28 12:24:29 -0400
Branch: REL_15_STABLE [512c55522] 2023-04-28 12:24:29 -0400
Branch: REL_14_STABLE [a1d9aacc4] 2023-04-28 12:24:29 -0400
Branch: REL_13_STABLE [7dcd9998c] 2023-04-28 12:24:29 -0400
Branch: REL_12_STABLE [ff9203f46] 2023-04-28 12:24:29 -0400
Branch: REL_11_STABLE [b7c6af375] 2023-04-28 12:24:29 -0400
-->
     <para>
      Tighten array dimensionality checks when converting Python
      list structures to multi-dimensional SQL arrays (Tom Lane)
     </para>

     <para>
      <application>plpython</application> could misbehave when dealing
      with empty sub-lists, or when the nesting of sub-lists is
      inconsistent so that the data does not represent a rectangular array
      of values.  The former should result in an empty output array, and
      the latter in an error.  But some cases resulted in a crash, and
      others in unexpected output.
     </para>
    </listitem>

    <listitem>
<!--
Author: Nathan Bossart <nathan@postgresql.org>
Branch: master [57d005170] 2023-05-04 16:23:05 -0700
Branch: REL_15_STABLE [825ebc984] 2023-05-04 16:24:48 -0700
Branch: REL_14_STABLE [52c9cf323] 2023-05-04 16:25:05 -0700
Branch: REL_13_STABLE [800531846] 2023-05-04 16:26:00 -0700
Branch: REL_12_STABLE [24964394a] 2023-05-04 16:26:05 -0700
Branch: REL_11_STABLE [0af386b0f] 2023-05-04 16:26:11 -0700
-->
     <para>
      Fix unwinding of exception stack
      in <application>plpython</application> (Xing Guo)
     </para>

     <para>
      Some rare failure cases could return without cleaning up the PG_TRY
      exception stack, risking a crash if another error was raised before
      the next stack level was unwound.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [e0a09d4e3] 2023-03-13 16:36:20 +0900
Branch: REL_15_STABLE [4493256c5] 2023-03-13 16:36:28 +0900
Branch: REL_14_STABLE [7e319231c] 2023-03-13 16:36:31 +0900
Branch: REL_13_STABLE [96bef4374] 2023-03-13 16:36:33 +0900
Branch: REL_12_STABLE [2bc36a56c] 2023-03-13 16:36:34 +0900
-->
     <para>
      Fix inconsistent GSS-encryption error handling
      in <application>libpq</application>'s
      <function>PQconnectPoll()</function>
      (Michael Paquier)
     </para>

     <para>
      With <option>gssencmode</option> set to <literal>require</literal>,
      the connection was not marked dead after a GSS initialization
      failure.  Make it fail immediately, as the equivalent case for TLS
      encryption has long done.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [f18029084] 2023-04-18 11:20:41 +0900
Branch: REL_15_STABLE [8c746be44] 2023-04-18 11:20:47 +0900
Branch: REL_14_STABLE [02f076454] 2023-04-18 11:20:50 +0900
Branch: REL_13_STABLE [e9e457d22] 2023-04-18 11:20:51 +0900
Branch: REL_12_STABLE [a28bd7713] 2023-04-18 11:20:53 +0900
Branch: REL_11_STABLE [9eb44bb04] 2023-04-18 11:20:55 +0900
Branch: REL_11_STABLE [dbd25dd0b] 2023-04-18 12:00:31 +0900
-->
     <para>
      Fix possible data corruption in <application>ecpg</application>
      programs built with the <option>-C ORACLE</option> option
      (Kyotaro Horiguchi)
     </para>

     <para>
      When <function>ecpg_get_data()</function> is called
      with <varname>varcharsize</varname> set to zero, it could write a
      terminating zero character into the last byte of the preceding
      field, truncating the data in that field.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [bc8cd50fe] 2023-03-17 13:31:40 -0400
Branch: REL_15_STABLE [2b216da1e] 2023-03-17 13:31:40 -0400
Branch: REL_14_STABLE [5fc1ac151] 2023-03-17 13:31:40 -0400
Branch: REL_13_STABLE [7e7c5b683] 2023-03-17 13:31:40 -0400
Branch: REL_12_STABLE [8f83ce8c5] 2023-03-17 13:31:40 -0400
Branch: REL_11_STABLE [012ffb365] 2023-03-17 13:31:40 -0400
-->
     <para>
      Fix <application>pg_dump</application> so that partitioned tables
      that are hash-partitioned on an enum-type column can be restored
      successfully (Tom Lane)
     </para>

     <para>
      Since the hash codes for enum values depend on the OIDs assigned to
      the enum, they are typically different after a dump and restore,
      meaning that rows often need to go into a different partition than
      they were in originally.  Users can work around that by specifying
      the <option>--load-via-partition-root</option> option; but since
      there is very little chance of success without that,
      teach <application>pg_dump</application> to apply it automatically
      to such tables.
     </para>

     <para>
      Also, fix <application>pg_restore</application> to not try
      to <command>TRUNCATE</command> target tables before restoring into
      them when <option>--load-via-partition-root</option> mode is used.
      This avoids a hazard of deadlocks and lost data.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [a923e2163] 2023-04-12 09:09:38 +0900
Branch: REL_15_STABLE [5c3254946] 2023-04-12 09:09:53 +0900
Branch: REL_14_STABLE [34105eea6] 2023-04-12 09:09:58 +0900
Branch: master [765f5df72] 2023-03-15 12:55:51 +0900
Branch: REL_15_STABLE [69b6032e0] 2023-03-15 12:56:06 +0900
Branch: REL_14_STABLE [bbf18fe19] 2023-03-15 12:56:10 +0900
-->
     <para>
      Correctly detect non-seekable files on Windows
      (Juan José Santamaría Flecha, Michael Paquier, Daniel Watzinger)
     </para>

     <para>
      This bug led to misbehavior when <application>pg_dump</application>
      writes to a pipe or <application>pg_restore</application> reads from
      one.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [038f586d5] 2023-02-21 10:56:37 +0100
Branch: REL_15_STABLE [108a22bd1] 2023-02-21 10:56:37 +0100
Branch: REL_14_STABLE [663e50e83] 2023-02-21 10:56:37 +0100
-->
     <para>
      In <application>pgbench</application>'s <quote>prepared</quote>
      mode, prepare all the commands in a pipeline before starting the
      pipeline (&Aacute;lvaro Herrera)
     </para>

     <para>
      This avoids a failure when a pgbench script tries to
      start a serializable transaction inside a pipeline.
     </para>
    </listitem>

    <listitem>
<!--
Author: Robert Haas <rhaas@postgresql.org>
Branch: master [e88754a19] 2023-03-24 10:56:10 -0400
Branch: REL_15_STABLE [701ec5557] 2023-03-24 10:59:10 -0400
Branch: REL_14_STABLE [8fd5aa76c] 2023-03-24 11:06:45 -0400
Branch: master [b7cea5882] 2023-03-28 16:16:53 -0400
Branch: REL_15_STABLE [453f53821] 2023-03-28 16:17:03 -0400
Branch: REL_14_STABLE [e3363cda9] 2023-03-28 16:21:29 -0400
-->
     <para>
      In <filename>contrib/amcheck</filename>'s heap checking code, deal
      correctly with tuples having zero xmin or xmax (Robert Haas)
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [4f5d461e0] 2023-03-11 14:12:52 -0800
Branch: REL_15_STABLE [e8a9750d0] 2023-03-11 14:14:50 -0800
Branch: REL_14_STABLE [b3a83055c] 2023-03-11 14:17:51 -0800
-->
     <para>
      In <filename>contrib/amcheck</filename>, deal sanely with xids that
      appear to be before epoch zero (Andres Freund)
     </para>

     <para>
      In cases of corruption we might see a wrapped-around 32-bit xid that
      appears to be before the first xid epoch.  Promoting such a value to
      64-bit form produced a value far in the future, resulting in wrong
      reports.  Return FirstNormalFullTransactionId in such cases so that
      things work reasonably sanely.
     </para>
    </listitem>

    <listitem>
<!--
Author: Robert Haas <rhaas@postgresql.org>
Branch: master [8e82db97b] 2023-04-12 11:37:13 -0400
Branch: REL_15_STABLE [fa83e9e23] 2023-04-12 11:51:09 -0400
-->
     <para>
      In <filename>contrib/basebackup_to_shell</filename>, properly detect
      failure to open a pipe (Robert Haas)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [df38157d9] 2023-04-27 11:55:06 -0400
Branch: REL_15_STABLE [85ec8bcce] 2023-04-27 11:55:06 -0400
Branch: REL_14_STABLE [c74f88c40] 2023-04-27 11:55:06 -0400
Branch: REL_13_STABLE [de2dfa053] 2023-04-27 11:55:06 -0400
Branch: REL_12_STABLE [ce9662598] 2023-04-27 11:55:06 -0400
Branch: REL_11_STABLE [c3c1097dc] 2023-04-27 11:55:06 -0400
-->
     <para>
      In <filename>contrib/hstore_plpython</filename>, avoid crashing if
      the Python value to be transformed isn't a mapping (Dmitry Dolgov,
      Tom Lane)
     </para>

     <para>
      This should give an error, but Python 3 changed some APIs in a way
      that caused the check to misbehave, allowing a crash to ensue.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [8bbd0cce9] 2023-04-23 14:30:09 +0300
Branch: REL_15_STABLE [214495dc5] 2023-04-23 14:30:51 +0300
Branch: REL_14_STABLE [f4a4a18ec] 2023-04-23 14:31:11 +0300
Branch: REL_13_STABLE [48c6825d0] 2023-04-23 14:00:16 +0300
-->
     <para>
      Require the <literal>siglen</literal> option of a GiST index on
      an <type>ltree</type> column, if specified, to be a multiple of 4
      (Alexander Korotkov)
     </para>

     <para>
      Other values result in misaligned accesses to index content, which
      is harmless on Intel-compatible hardware but can cause a crash on
      some other architectures.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [be753639d] 2023-03-02 14:03:02 +0900
Branch: REL_15_STABLE [9d41ecfcd] 2023-03-02 14:03:08 +0900
Branch: REL_14_STABLE [5ad63eee1] 2023-03-02 14:03:21 +0900
-->
     <para>
      In <filename>contrib/pageinspect</filename>, add defenses against
      incorrect input for the <function>gist_page_items()</function> function
      (Dmitry Koval)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [1925ac281] 2023-03-11 12:15:41 -0500
Branch: REL_15_STABLE [6170386c7] 2023-03-11 12:15:41 -0500
Branch: REL_14_STABLE [786528039] 2023-03-11 12:15:41 -0500
Branch: REL_13_STABLE [bc436e4a9] 2023-03-11 12:15:41 -0500
Branch: REL_12_STABLE [1279414bc] 2023-03-11 12:15:41 -0500
Branch: REL_11_STABLE [b18327489] 2023-03-11 12:15:41 -0500
-->
     <para>
      Fix misbehavior in <filename>contrib/pg_trgm</filename> with an
      unsatisfiable regular expression (Tom Lane)
     </para>

     <para>
      A regex such as <literal>$foo</literal> is legal but unsatisfiable;
      the regex compiler recognizes that and produces an empty NFA graph.
      Attempting to optimize such a graph into a pg_trgm GIN or GiST index
      qualification resulted in accessing off the end of a work array,
      possibly leading to crashes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [8427ce4c3] 2023-02-21 20:01:43 +0900
Branch: REL_15_STABLE [5bace41ab] 2023-02-21 20:02:09 +0900
-->
     <para>
      Fix handling of escape sequences
      in <filename>contrib/postgres_fdw</filename>'s
      <varname>application_name</varname> parameter (Kyotaro Horiguchi,
      Michael Paquier)
     </para>

     <para>
      The code to expand these could fail if executed in a background
      process, as for example during auto-analyze of a foreign table.
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [69e8c7cf1] 2023-02-20 11:07:24 -0800
Branch: REL_15_STABLE [da32a99df] 2023-02-20 11:29:31 -0800
-->
     <para>
      In <filename>contrib/pg_walinspect</filename>, limit memory usage
      of <function>pg_get_wal_records_info()</function> (Bharath Rupireddy)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [eab2d3147] 2023-04-20 18:12:32 -0400
Branch: REL_15_STABLE [a14afd3bd] 2023-04-20 18:12:48 -0400
Branch: REL_14_STABLE [6d60b718c] 2023-04-20 18:12:32 -0400
Branch: REL_13_STABLE [6dce37203] 2023-04-20 18:12:32 -0400
Branch: REL_12_STABLE [e2e34dfff] 2023-04-20 18:12:32 -0400
Branch: REL_11_STABLE [c976ccc9e] 2023-04-20 18:12:32 -0400
-->
     <para>
      Use the <option>--strip-unneeded</option> option when stripping
      static libraries with
      GNU-compatible <application>strip</application> (Tom Lane)
     </para>

     <para>
      Previously, <literal>make install-strip</literal> used
      the <option>-x</option> option in this case.  This change avoids
      misbehavior of <application>llvm-strip</application>, and gives
      slightly smaller output as well.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [969509c3f] 2023-02-08 17:15:25 -0500
Branch: REL_15_STABLE [2ee703c9d] 2023-02-08 17:15:23 -0500
Branch: REL_14_STABLE [7f8778fcf] 2023-02-08 17:15:23 -0500
Branch: REL_13_STABLE [c7b608600] 2023-02-08 17:15:23 -0500
Branch: REL_12_STABLE [11f1f9f4f] 2023-02-08 17:15:27 -0500
Branch: REL_11_STABLE [36a646d99] 2023-02-08 17:15:27 -0500
-->
     <para>
      Stop recommending auto-download of DTD files for building the
      documentation, and indeed disable it (Aleksander Alekseev, Peter
      Eisentraut, Tom Lane)
     </para>

     <para>
      It appears no longer possible to build the SGML documentation
      without a local installation of the DocBook DTD files.
      Formerly <application>xsltproc</application> could download those
      files on-the-fly from sourceforge.net; but sourceforge.net now
      permits only HTTPS access, and no common version
      of <application>xsltproc</application> supports that.  Hence, remove
      the bits of our documentation suggesting that that's possible or
      useful, and instead
      add <application>xsltproc</application>'s <option>--nonet</option>
      option to the build recipes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [5d42a9751] 2023-05-05 07:09:52 +0200
Branch: REL_15_STABLE [3d37476f5] 2023-05-05 07:10:15 +0200
Branch: REL_14_STABLE [e07022500] 2023-05-05 07:10:30 +0200
Branch: REL_13_STABLE [670494477] 2023-05-05 07:11:58 +0200
Branch: REL_12_STABLE [14bb2e76c] 2023-05-05 07:12:18 +0200
Branch: REL_11_STABLE [b189f71e5] 2023-05-05 07:12:26 +0200
-->
     <para>
      When running TAP tests in PGXS builds, use a saner location for the
      temporary <filename>portlock</filename> directory (Peter Eisentraut)
     </para>

     <para>
      Place it under <filename>tmp_check</filename> in the build
      directory.  With the previous coding, a PGXS build would try to place
      it in the installation directory, which is not necessarily writable.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [4ddee4d9d] 2023-04-18 14:46:39 -0400
Branch: REL_15_STABLE [62b22caa5] 2023-04-18 14:46:39 -0400
Branch: REL_14_STABLE [b79b36f26] 2023-04-18 14:46:39 -0400
Branch: REL_13_STABLE [87d8ec3e4] 2023-04-18 14:46:39 -0400
Branch: REL_12_STABLE [2ad35cf06] 2023-04-18 14:46:39 -0400
Branch: REL_11_STABLE [7228f2f81] 2023-04-18 14:46:39 -0400
-->
     <para>
      Update time zone data files to <application>tzdata</application>
      release 2023c for DST law changes in Egypt, Greenland, Morocco, and
      Palestine.
     </para>

     <para>
      When observing Moscow time, Europe/Kirov and Europe/Volgograd now
      use the abbreviations MSK/MSD instead of numeric abbreviations,
      for consistency with other timezones observing Moscow time.
      Also, America/Yellowknife is no longer distinct from America/Edmonton;
      this affects some pre-1948 timestamps in that area.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15-2">
  <title>Release 15.2</title>

  <formalpara>
  <title>Release date:</title>
  <para>2023-02-09</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.1.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.2</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, if you are upgrading from a version earlier than 15.1,
    see <xref linkend="release-15-1"/>.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [71c37797d] 2023-02-06 11:20:07 +0900
Branch: REL_15_STABLE [715c345dd] 2023-02-06 11:20:20 +0900
Branch: REL_14_STABLE [626f2c1d6] 2023-02-06 11:20:23 +0900
Branch: REL_13_STABLE [45a945ee9] 2023-02-06 11:20:27 +0900
Branch: REL_12_STABLE [3f7342671] 2023-02-06 11:20:31 +0900
-->
     <para>
      <application>libpq</application> can leak memory contents after
      GSSAPI transport encryption initiation fails (Jacob Champion)
     </para>

     <para>
      A modified server, or an unauthenticated man-in-the-middle, can
      send a not-zero-terminated error message during setup of GSSAPI
      (Kerberos) transport encryption.  <application>libpq</application>
      will then copy that string, as well as following bytes in
      application memory up to the next zero byte, to its error report.
      Depending on what the calling application does with the error
      report, this could result in disclosure of application memory
      contents.  There is also a small probability of a crash due to
      reading beyond the end of memory.  Fix by properly zero-terminating
      the server message.
      (CVE-2022-41862)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [3f7836ff6] 2023-01-05 14:12:17 -0500
Branch: REL_15_STABLE [3706cc97a] 2023-01-05 14:12:17 -0500
Branch: REL_14_STABLE [8cd190e13] 2023-01-05 14:12:17 -0500
Branch: REL_13_STABLE [ad38e2f89] 2023-01-05 14:12:17 -0500
Branch: master [3f244d020] 2023-01-15 13:14:52 -0500
Branch: REL_15_STABLE [a8f7687a0] 2023-01-15 14:06:46 -0500
Branch: REL_14_STABLE [a8b88c26f] 2023-01-15 14:06:46 -0500
Branch: REL_13_STABLE [787db4be9] 2023-01-15 14:06:46 -0500
-->
     <para>
      Fix calculation of which <literal>GENERATED</literal> columns need
      to be updated in child tables during an <command>UPDATE</command> on
      a partitioned table or inheritance tree (Amit Langote, Tom Lane)
     </para>

     <para>
      This fixes failure to update <literal>GENERATED</literal> columns
      that do not exist in the parent table, or that have different
      dependencies than are in the parent column's generation expression.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [fe9e658f4] 2023-01-30 10:04:57 +0000
Branch: REL_15_STABLE [4785af9e6] 2023-01-30 10:07:32 +0000
-->
     <para>
      Fix possible failure of <command>MERGE</command> to compute
      <literal>GENERATED</literal> columns (Dean Rasheed)
     </para>

     <para>
      When the first row-level action of the <command>MERGE</command> was
      an <literal>UPDATE</literal>, any
      subsequent <literal>INSERT</literal> actions would fail to
      compute <literal>GENERATED</literal> columns that were deemed
      unnecessary to compute for the <literal>UPDATE</literal> action
      (due to not depending on any of the <literal>UPDATE</literal> target
      columns).
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [f026c16a2] 2023-01-10 14:17:47 +0000
Branch: REL_15_STABLE [38255f2d0] 2023-01-10 14:16:27 +0000
-->
     <para>
      Fix <command>MERGE</command>'s check for
      unreachable <literal>WHEN</literal> clauses (Dean Rasheed)
     </para>

     <para>
      A <literal>WHEN</literal> clause following an
      unconditional <literal>WHEN</literal> clause should be rejected as
      unreachable, but this case was not always detected.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [7b2ccc5e0] 2022-11-25 13:31:48 +0000
Branch: REL_15_STABLE [04d61bfe6] 2022-11-25 13:29:51 +0000
-->
     <para>
      Fix <command>MERGE</command>'s rule-detection test (Dean Rasheed)
     </para>

     <para>
      <command>MERGE</command> is not supported on tables with rules;
      but it also failed on tables that once had rules but no longer do.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [6ff5aa129] 2022-11-17 18:56:11 +0100
Branch: REL_15_STABLE [3d45edcef] 2022-11-17 18:56:11 +0100
-->
     <para>
      In <command>MERGE</command>, don't count a <literal>DO
      NOTHING</literal> action as a processed tuple (&Aacute;lvaro Herrera)
     </para>

     <para>
      This makes the code's behavior match the documentation.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [935277b24] 2022-12-16 13:07:42 -0500
Branch: REL_15_STABLE [ae98debf7] 2022-12-16 13:07:42 -0500
Branch: REL_14_STABLE [ea5ae4cae] 2022-12-16 13:07:42 -0500
-->
     <para>
      Allow a <literal>WITH RECURSIVE ... CYCLE</literal> CTE
      to access its output column (Tom Lane)
     </para>

     <para>
      A reference to the <literal>SET</literal> column from within the CTE
      would fail with <quote>cache lookup failed for type 0</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Etsuro Fujita <efujita@postgresql.org>
Branch: master [ffbb7e65a] 2022-11-25 17:45:00 +0900
Branch: REL_15_STABLE [fc02019c0] 2022-11-25 17:45:01 +0900
Branch: REL_14_STABLE [e52245228] 2022-11-25 17:45:03 +0900
Branch: master [4b3e37993] 2022-12-08 16:15:00 +0900
Branch: REL_15_STABLE [a0bf7a0ec] 2022-12-08 16:15:01 +0900
Branch: REL_14_STABLE [d43a97ef4] 2022-12-08 16:15:03 +0900
-->
     <para>
      Fix handling of pending inserts when doing a bulk insertion to a
      foreign table (Etsuro Fujita)
     </para>

     <para>
      In some cases pending insertions were not flushed to the FDW soon
      enough, leading to logical inconsistencies, for
      example <literal>BEFORE ROW</literal> triggers not seeing rows they
      should be able to see.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [c9f7f9264] 2023-01-21 13:10:29 -0500
Branch: REL_15_STABLE [9e4288ce6] 2023-01-21 13:10:29 -0500
Branch: REL_14_STABLE [70ec756b0] 2023-01-21 13:10:29 -0500
Branch: REL_13_STABLE [72d611109] 2023-01-21 13:10:29 -0500
Branch: REL_12_STABLE [a5f3f2fce] 2023-01-21 13:10:30 -0500
Branch: REL_11_STABLE [6c122edde] 2023-01-21 13:10:30 -0500
-->
     <para>
      Allow <literal>REPLICA IDENTITY</literal>
      to be set on an index that's not (yet) valid (Tom Lane)
     </para>

     <para>
      When <application>pg_dump</application> dumps a partitioned index
      that's marked <literal>REPLICA IDENTITY</literal>, it generates a
      command sequence that applies <literal>REPLICA IDENTITY</literal>
      before the partitioned index has been marked valid, causing restore
      to fail.  There seems no very good reason to prohibit doing it in
      that order, so allow it.  The marking will have no effect anyway
      until the index becomes valid.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [2605643a3] 2022-12-03 12:11:33 +0000
Branch: REL_15_STABLE [c67204db6] 2022-12-03 12:14:36 +0000
Branch: REL_14_STABLE [2c7ed9f75] 2022-12-03 12:16:07 +0000
Branch: REL_13_STABLE [3bed88123] 2022-12-03 12:17:47 +0000
Branch: REL_12_STABLE [33f600f04] 2022-12-03 12:18:58 +0000
Branch: REL_11_STABLE [30f9b03a0] 2022-12-03 12:20:02 +0000
-->
     <para>
      Fix handling of <literal>DEFAULT</literal> markers in rules that
      perform an <command>INSERT</command> from a
      multi-row <literal>VALUES</literal> list (Dean Rasheed)
     </para>

     <para>
      In some cases a <literal>DEFAULT</literal> marker would not get
      replaced with the proper default-value expression, leading to
      an <quote>unrecognized node type</quote> error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
Branch: master [3161ae86c] 2023-01-12 18:16:34 +0300
Branch: REL_15_STABLE [4dc3f94fa] 2023-01-12 18:17:43 +0300
Branch: REL_14_STABLE [0d9221f1d] 2023-01-12 18:18:04 +0300
Branch: REL_13_STABLE [2ff3ac3b5] 2023-01-12 18:18:31 +0300
Branch: REL_12_STABLE [9e24e4781] 2023-01-12 18:19:19 +0300
-->
     <para>
      Reject uses of undefined variables in <type>jsonpath</type>
      existence checks (Alexander Korotkov, David G. Johnston)
     </para>

     <para>
      While <type>jsonpath</type> match operators threw an error for an
      undefined variable in the path pattern, the existence operators
      silently treated it as a match.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [b0feda79f] 2022-12-12 16:17:54 -0500
Branch: REL_15_STABLE [d79b76b10] 2022-12-12 16:17:49 -0500
Branch: REL_14_STABLE [a18328bb3] 2022-12-12 16:17:49 -0500
-->
     <para>
      Fix <type>jsonb</type> subscripting to cope with toasted subscript
      values (Tom Lane, David G. Johnston)
     </para>

     <para>
      Using a text value fetched directly from a table as
      a <type>jsonb</type> subscript was likely to fail.
      Fetches would usually not find any matching element.
      Assignments could store the value with a garbage key,
      although keys long enough to cause that problem are probably rare in
      the field.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [ffcf6f4cf] 2023-01-26 14:52:19 +1300
Branch: REL_15_STABLE [d9f5345bf] 2023-01-26 14:53:37 +1300
Branch: REL_14_STABLE [2f65b8468] 2023-01-26 14:54:17 +1300
Branch: REL_13_STABLE [1a5afe007] 2023-01-26 14:54:39 +1300
Branch: REL_12_STABLE [b55303792] 2023-01-26 14:55:03 +1300
Branch: REL_11_STABLE [d95dcc9ab] 2023-01-26 14:55:37 +1300
-->
     <para>
      Fix edge-case data corruption in parallel hash joins (Dmitry Astapov)
     </para>

     <para>
      If the final chunk of a large tuple being written out to a temporary
      file was exactly 32760 bytes, it would be corrupted due to a
      fencepost bug.  The query would typically fail later with
      corrupted-data symptoms.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [7e8a80d1f] 2023-01-19 13:13:05 +0900
Branch: REL_15_STABLE [49e3a5e71] 2023-01-19 13:13:27 +0900
Branch: REL_14_STABLE [2e21e2857] 2023-01-19 13:13:28 +0900
Branch: REL_13_STABLE [fed4e92f3] 2023-01-19 13:13:30 +0900
Branch: REL_12_STABLE [162a48287] 2023-01-19 13:13:32 +0900
Branch: REL_11_STABLE [0c2f34af7] 2023-01-19 13:13:34 +0900
-->
     <para>
      Honor non-default settings
      of <varname>checkpoint_completion_target</varname>
      (Bharath Rupireddy)
     </para>

     <para>
      Internal state was not updated after a change
      in <varname>checkpoint_completion_target</varname>, possibly
      resulting in performing checkpoint I/O faster or slower than
      desired, especially if that setting was changed on-the-fly.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [44e9e3426] 2023-01-19 12:23:20 -0500
Branch: REL_15_STABLE [abe203304] 2023-01-19 12:23:20 -0500
Branch: REL_14_STABLE [103450724] 2023-01-19 12:23:20 -0500
Branch: REL_13_STABLE [1b9a0b96a] 2023-01-19 12:23:20 -0500
Branch: REL_12_STABLE [87591db19] 2023-01-19 12:23:20 -0500
Branch: REL_11_STABLE [0a269527f] 2023-01-19 12:23:20 -0500
-->
     <para>
      Log the correct ending timestamp
      in <varname>recovery_target_xid</varname> mode (Tom Lane)
     </para>

     <para>
      When ending recovery based on the <varname>recovery_target_xid</varname>
      setting with <varname>recovery_target_inclusive</varname>
      = <literal>off</literal>, we printed an incorrect timestamp (always
      2000-01-01) in the <quote>recovery stopping before
      ... transaction</quote> log message.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [1561612e3] 2023-01-16 09:44:04 +0100
Branch: REL_15_STABLE [ac01fa647] 2023-01-16 09:45:03 +0100
Branch: REL_14_STABLE [f463335e1] 2023-01-16 09:48:09 +0100
Branch: REL_13_STABLE [cf74b6ead] 2023-01-16 09:54:15 +0100
-->
     <para>
      Improve error reporting for some buffered file read failures
      (Peter Eisentraut)
     </para>

     <para>
      Correctly report a short read, giving the numbers of bytes desired
      and actually read, instead of reporting an irrelevant error code.
      Most places got this right already, but some recently-written
      replication logic did not.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [647fa5005] 2023-01-15 17:32:09 -0500
Branch: REL_15_STABLE [db9127c58] 2023-01-15 17:32:09 -0500
-->
     <para>
      Remove arbitrary limit on number of elements
      in <type>int2vector</type> and <type>oidvector</type> (Tom Lane)
     </para>

     <para>
      The input functions for these types previously rejected more than
      100 elements.  With the introduction of the logical replication
      column list feature, it's necessary to
      accept <type>int2vector</type>s having up to 1600 columns,
      otherwise long column lists cause logical-replication failures.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [20432f873] 2022-12-13 14:23:58 -0500
Branch: REL_15_STABLE [18431ee6f] 2022-12-13 14:23:59 -0500
Branch: REL_14_STABLE [ae47f8a96] 2022-12-13 14:23:59 -0500
Branch: REL_13_STABLE [942cc240f] 2022-12-13 14:23:59 -0500
Branch: REL_12_STABLE [1cca4a75f] 2022-12-13 14:23:59 -0500
Branch: REL_11_STABLE [f48aa5df4] 2022-12-13 14:23:59 -0500
-->
     <para>
      In extended query protocol, avoid an immediate commit
      after <command>ANALYZE</command> if we're running a pipeline
      (Tom Lane)
     </para>

     <para>
      If there's not been an explicit <command>BEGIN
      TRANSACTION</command>, <command>ANALYZE</command> would take it on
      itself to commit, which should not happen within a pipelined series
      of commands.
     </para>
    </listitem>

    <listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [e52daaabf] 2023-01-21 06:08:00 -0800
Branch: REL_15_STABLE [b152bb7b2] 2023-01-21 06:08:03 -0800
Branch: REL_14_STABLE [6900aea67] 2023-01-21 06:08:04 -0800
Branch: REL_13_STABLE [a9bccffe5] 2023-01-21 06:08:04 -0800
Branch: REL_12_STABLE [e75b5c855] 2023-01-21 06:08:04 -0800
Branch: REL_11_STABLE [8f70de7e0] 2023-01-21 06:08:05 -0800
-->
     <para>
      Reject cancel request packets having the wrong length
      (Andrey Borodin)
     </para>

     <para>
      The server would process a cancel request even if its length word
      was too small.  This led to reading beyond the end of the allocated
      buffer.  In theory that could cause a segfault, but it seems quite
      unlikely to happen in practice, since the buffer would have to be
      very close to the end of memory.  The more likely outcome was a bogus
      log message about wrong backend PID or cancel code.  Complain about
      the wrong length, instead.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [94985c210] 2022-12-10 19:27:20 +1300
Branch: REL_15_STABLE [04788ee4c] 2022-12-10 19:27:53 +1300
-->
     <para>
      Fix planner preprocessing oversights for window function run-condition
      expressions (Richard Guo, David Rowley)
     </para>

     <para>
      This could lead to planner errors such as <quote>WindowFunc not
      found in subplan target lists</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: David Rowley <drowley@postgresql.org>
Branch: master [a85832722] 2022-12-07 00:09:36 +1300
Branch: REL_15_STABLE [2a535620c] 2022-12-07 00:10:21 +1300
-->
     <para>
      Fix possible dangling-pointer access during execution of window
      function run-condition expressions (David Rowley)
     </para>

     <para>
      In practice, because the run-condition optimization is only applied
      to certain window functions that happen to all
      return <type>int8</type>, this only manifested as a problem on
      32-bit builds.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5beb7881f] 2022-12-22 10:35:02 -0500
Branch: REL_15_STABLE [1a3daa5bb] 2022-12-22 10:35:02 -0500
Branch: REL_14_STABLE [97431d673] 2022-12-22 10:35:02 -0500
Branch: REL_13_STABLE [4fceb454f] 2022-12-22 10:35:03 -0500
Branch: REL_12_STABLE [d572003f7] 2022-12-22 10:35:03 -0500
Branch: REL_11_STABLE [8cd700cc5] 2022-12-22 10:35:03 -0500
-->
     <para>
      Add recursion and looping defenses in subquery pullup (Tom Lane)
     </para>

     <para>
      A contrived query can result in deep recursion and unreasonable
      amounts of time spent trying to flatten subqueries.  A proper fix
      for that seems unduly invasive for a back-patch, but we can at least
      add stack depth checks and an interrupt check to allow the query to
      be cancelled.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d69d01ba9] 2022-12-05 12:36:40 -0500
Branch: REL_15_STABLE [c959f84c2] 2022-12-05 12:36:41 -0500
Branch: REL_14_STABLE [dc3648f65] 2022-12-05 12:36:41 -0500
Branch: master [e76913802] 2022-12-04 13:48:12 -0500
Branch: REL_15_STABLE [834d97c32] 2022-12-04 13:48:12 -0500
Branch: REL_14_STABLE [ce093aa18] 2022-12-04 13:48:12 -0500
-->
     <para>
      Fix planner issues when combining Memoize nodes with partitionwise
      joins or parameterized nestloops (Richard Guo)
     </para>

     <para>
      These errors could lead to not using Memoize in contexts where it
      would be useful, or possibly to wrong query plans.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [fe12f2f8f] 2022-12-04 13:17:18 -0500
Branch: REL_15_STABLE [bf8fd64ff] 2022-12-04 13:17:18 -0500
Branch: REL_14_STABLE [ec3daeec3] 2022-12-04 13:17:18 -0500
Branch: REL_13_STABLE [4ebca555c] 2022-12-04 13:17:18 -0500
Branch: REL_12_STABLE [1ff549e84] 2022-12-04 13:17:18 -0500
Branch: REL_11_STABLE [2df073313] 2022-12-04 13:17:18 -0500
-->
     <para>
      Fix partitionwise-join code to tolerate failure to produce a plan for
      each partition (Tom Lane)
     </para>

     <para>
      This could result in <quote>could not devise a query plan for the
      given query</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [9c6ad5eaa] 2022-11-22 14:40:20 -0500
Branch: REL_15_STABLE [2debceed2] 2022-11-22 14:40:44 -0500
Branch: REL_14_STABLE [bd06fe4de] 2022-11-22 14:40:45 -0500
Branch: REL_13_STABLE [6e639267a] 2022-11-22 14:40:45 -0500
Branch: REL_12_STABLE [ec10b6139] 2022-11-22 14:40:45 -0500
Branch: REL_11_STABLE [b96a096db] 2022-11-22 14:40:46 -0500
-->
     <para>
      Limit the amount of cleanup work done
      by <function>get_actual_variable_range</function> (Simon Riggs)
     </para>

     <para>
      Planner runs occurring just after deletion of a large number of
      tuples appearing at the end of an index could expend significant
      amounts of work setting the <quote>killed</quote> bits for those
      index entries.  Limit the amount of work done in any one query by
      giving up on this process after examining 100 heap pages.  All the
      cleanup will still happen eventually, but without so large a
      performance hiccup.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [cb2e7ddfe] 2022-12-02 18:10:30 -0800
Branch: REL_15_STABLE [c6a60471a] 2022-12-02 18:07:47 -0800
-->
     <para>
      Prevent the statistics machinery from getting confused when a
      relation's relkind changes (Andres Freund)
     </para>

     <para>
      Converting a table to a view could lead to crashes or assertion
      failures.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [1dd6700f4] 2022-12-01 11:38:14 -0500
Branch: REL_15_STABLE [a711b36e5] 2022-12-01 11:38:15 -0500
Branch: REL_14_STABLE [de0ff6088] 2022-12-01 11:38:15 -0500
-->
     <para>
      Fix under-parenthesized display of <literal>AT TIME ZONE</literal>
      constructs (Tom Lane)
     </para>

     <para>
      This could result in dump/restore failures for rules or views in
      which an argument of <literal>AT TIME ZONE</literal> is itself an
      expression.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [8b47ccb62] 2022-11-29 11:46:33 -0500
Branch: REL_15_STABLE [5dfc2b753] 2022-11-29 11:46:33 -0500
Branch: REL_14_STABLE [7715a3c24] 2022-11-29 11:46:33 -0500
-->
     <para>
      Prevent clobbering of cached parsetrees for utility statements in
      SQL functions (Tom Lane, Daniel Gustafsson)
     </para>

     <para>
      If a SQL-language function executes the same utility command more
      than once within a single calling query, it could crash or report
      strange errors such as <quote>unrecognized node type</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5644d6f90] 2022-11-21 17:07:29 -0500
Branch: REL_15_STABLE [0353db996] 2022-11-21 17:07:07 -0500
Branch: REL_14_STABLE [1b9c04b13] 2022-11-21 17:07:07 -0500
Branch: REL_13_STABLE [74670688f] 2022-11-21 17:07:07 -0500
Branch: REL_12_STABLE [1aed4c4fd] 2022-11-21 17:07:07 -0500
Branch: REL_11_STABLE [c0eed8891] 2022-11-21 17:07:07 -0500
-->
     <para>
      Ensure that execution of full-text-search queries can be cancelled
      while they are performing phrase matches (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [edf12e7bb] 2022-12-01 11:49:15 -0800
Branch: REL_15_STABLE [9377b4f30] 2022-12-01 11:55:59 -0800
Branch: REL_14_STABLE [47e1224d5] 2022-12-01 11:53:53 -0800
Branch: REL_13_STABLE [a844052b5] 2022-12-01 11:52:06 -0800
Branch: REL_12_STABLE [f98c4fb1d] 2022-12-01 11:49:43 -0800
-->
     <para>
      Fix memory leak in hashing strings with nondeterministic collations
      (Jeff Davis)
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [6c6d6ba3e] 2023-01-24 09:25:36 +0530
Branch: REL_15_STABLE [267135d01] 2023-01-24 09:12:04 +0530
Branch: REL_14_STABLE [fd270b728] 2023-01-24 09:02:05 +0530
-->
     <para>
      Fix deadlock between <command>DROP DATABASE</command> and logical
      replication worker process (Hou Zhijie)
     </para>

     <para>
      This was caused by an ill-advised choice to block interrupts while
      creating a logical replication slot in the worker.  In version 15
      that could lead to an undetected deadlock.  In version 14, no
      deadlock has been observed, but it's still a bad idea to block
      interrupts while waiting for network I/O.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [bc54ef4ec] 2023-01-23 18:27:42 -0800
Branch: REL_15_STABLE [704a330a9] 2023-01-23 18:27:45 -0800
Branch: REL_14_STABLE [0a796b8b3] 2023-01-23 18:27:48 -0800
Branch: REL_13_STABLE [c5864805b] 2023-01-23 18:27:51 -0800
Branch: REL_12_STABLE [92fc12787] 2023-01-23 18:27:55 -0800
Branch: REL_11_STABLE [243373159] 2023-01-23 18:27:58 -0800
-->
     <para>
      Clean up the <application>libpq</application> connection object
      after a failed replication connection attempt (Andres Freund)
     </para>

     <para>
      The previous coding leaked the connection object.  In background
      code paths that's pretty harmless because the calling process will
      give up and exit.  But in commands such as <command>CREATE
      SUBSCRIPTION</command>, such a failure resulted in a small
      session-lifespan memory leak.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [8242752f9] 2022-11-29 15:43:17 -0500
Branch: REL_15_STABLE [55fa993d7] 2022-11-29 15:43:17 -0500
Branch: REL_14_STABLE [06998eab1] 2022-11-29 15:43:17 -0500
Branch: REL_13_STABLE [6e8ad1152] 2022-11-29 15:43:17 -0500
Branch: REL_12_STABLE [c4a153d77] 2022-11-29 15:43:17 -0500
Branch: REL_11_STABLE [a6c9e1db2] 2022-11-29 15:43:17 -0500
-->
     <para>
      In hot-standby servers, reduce processing effort for tracking XIDs
      known to be active on the primary (Simon Riggs, Michail Nikolaev)
     </para>

     <para>
      Insufficiently-aggressive cleanup of the KnownAssignedXids array
      could lead to poor performance, particularly
      when <varname>max_connections</varname> is set to a large value on
      the standby.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [0557e1770] 2022-11-22 10:56:07 +0100
Branch: REL_15_STABLE [1ad033df1] 2022-11-22 10:56:07 +0100
Branch: REL_14_STABLE [1b3ed7571] 2022-11-22 10:56:07 +0100
Branch: REL_13_STABLE [36eeb37cd] 2022-11-22 10:56:07 +0100
-->
     <para>
      Ignore invalidated logical-replication slots while determining
      oldest catalog xmin (Sirisha Chamarthi)
     </para>

     <para>
      A replication slot could prevent cleanup of dead tuples in the
      system catalogs even after it becomes invalidated due to
      exceeding <varname>max_slot_wal_keep_size</varname>.  Thus, failure
      of a replication consumer could lead to indefinitely-large catalog
      bloat.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [2b6df0546] 2023-01-07 12:17:14 +0530
Branch: REL_15_STABLE [18b81258a] 2023-01-07 12:04:33 +0530
Branch: REL_14_STABLE [b2cc5b810] 2023-01-07 11:52:41 +0530
-->
     <para>
      In logical decoding, notify the remote node when a transaction is
      detected to have crashed (Hou Zhijie)
     </para>

     <para>
      After a server restart, we'll re-stream the changes for transactions
      occurring shortly before the restart.  Some of these transactions
      probably never completed; when we realize that one didn't we throw
      away the relevant decoding state locally, but we neglected to tell
      the subscriber about it.  That led to the subscriber keeping useless
      streaming files until it's next restarted.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: REL_15_STABLE [898ef41bf] 2022-11-25 09:38:03 +0530
Branch: REL_14_STABLE [9f2cc1a73] 2022-11-25 09:25:50 +0530
Branch: REL_13_STABLE [4ec157c15] 2022-11-25 09:15:31 +0530
Branch: REL_12_STABLE [aa9d916f6] 2022-11-25 09:00:15 +0530
Branch: REL_11_STABLE [9b788aafd] 2022-11-25 08:56:54 +0530
-->
     <para>
      Fix uninitialized-memory usage in logical decoding (Masahiko Sawada)
     </para>

     <para>
      In certain cases, resumption of logical decoding could try to re-use
      XID data that had already been freed, leading to unpredictable
      behavior.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [5ad165d2c] 2023-01-12 13:40:33 +0900
Branch: REL_15_STABLE [6f25e4877] 2023-01-12 13:41:22 +0900
-->
     <para>
      Acquire spinlock while updating shared state during logical decoding
      context creation (Masahiko Sawada)
     </para>

     <para>
      We neglected to acquire the appropriate lock while updating data
      about two-phase transactions, potentially allowing other processes
      to see inconsistent data.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [40b149135] 2022-12-02 10:52:58 +0530
Branch: REL_15_STABLE [ebf87c019] 2022-12-02 11:14:42 +0530
-->
     <para>
      Fix <application>pgoutput</application> replication plug-in to not
      send columns not listed in a table's replication column list
      (Hou Zhijie)
     </para>

     <para>
      <literal>UPDATE</literal> and <literal>DELETE</literal> events did
      not pay attention to the configured column list, thus sending more
      data than expected.  This did not cause a problem when the receiver
      is our built-in logical replication code, but it might confuse other
      receivers, and in any case it wasted network bandwidth.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [e848be60b] 2022-11-14 10:43:33 +0530
Branch: REL_15_STABLE [e49e19181] 2022-11-14 10:32:47 +0530
Branch: REL_14_STABLE [9693f1900] 2022-11-14 10:22:28 +0530
Branch: REL_13_STABLE [20c223336] 2022-11-14 10:11:10 +0530
Branch: REL_12_STABLE [4dccccb37] 2022-11-14 10:01:14 +0530
Branch: REL_11_STABLE [1703033f8] 2022-11-14 09:52:06 +0530
-->
     <para>
      Avoid rare <quote>failed to acquire cleanup lock</quote> panic
      during WAL replay of hash-index page split operations (Robert Haas)
     </para>
    </listitem>

    <listitem>
<!--
Author: Jeff Davis <jdavis@postgresql.org>
Branch: master [d6a3dbe14] 2022-11-11 12:38:29 -0800
Branch: REL_15_STABLE [7bf713dd2] 2022-11-11 12:46:11 -0800
Branch: REL_14_STABLE [f893af496] 2022-11-11 12:46:22 -0800
Branch: REL_13_STABLE [58a45bb1d] 2022-11-11 12:46:34 -0800
Branch: REL_12_STABLE [7dd39e9e8] 2022-11-11 12:46:44 -0800
Branch: REL_11_STABLE [5eaf3e375] 2022-11-11 12:46:52 -0800
-->
     <para>
      Advance a heap page's LSN when setting its all-visible bit during
      WAL replay (Jeff Davis)
     </para>

     <para>
      Failure to do this left the page possibly different on standby
      servers than the primary, and violated some other expectations about
      when the LSN changes.  This seems only a theoretical hazard so
      far as <productname>PostgreSQL</productname> itself is concerned,
      but it could upset third-party tools.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [b2d479289] 2023-02-03 11:13:34 +0000
Branch: REL_15_STABLE [4f74741a5] 2023-02-03 11:11:59 +0000
Branch: REL_14_STABLE [86bfbeab4] 2023-02-03 11:09:15 +0000
-->
     <para>
      Fix <function>int64_div_fast_to_numeric()</function> to work for a
      wider range of inputs (Dean Rasheed)
     </para>

     <para>
      This function misbehaved with some values of its second argument.
      No such usages exist in core <productname>PostgreSQL</productname>,
      but it's clearly a hazard for external modules, so repair.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [f1821b58f] 2023-01-13 11:02:12 +1300
Branch: REL_15_STABLE [8a98523a5] 2023-01-13 11:02:00 +1300
Branch: REL_14_STABLE [547e60b83] 2023-01-13 10:57:29 +1300
Branch: REL_13_STABLE [c159b0383] 2023-01-13 10:54:49 +1300
Branch: REL_12_STABLE [bf388ab82] 2023-01-13 10:54:34 +1300
Branch: REL_11_STABLE [1b40710a8] 2023-01-13 10:54:20 +1300
-->
     <para>
      Fix latent buffer-overrun problem in <literal>WaitEventSet</literal>
      logic (Thomas Munro)
     </para>

     <para>
      The <function>epoll</function>-based
      and <function>kqueue</function>-based implementations could ask the
      kernel for too many events if the size of their internal buffer was
      different from the size of the caller's output buffer.  That case is
      not known to occur in released <productname>PostgreSQL</productname>
      versions, but this error is a hazard for external modules and future
      bug fixes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [8c954168c] 2022-11-19 12:22:04 -0800
Branch: REL_15_STABLE [a0d35ebcc] 2022-11-19 12:33:11 -0800
Branch: REL_14_STABLE [fc4154286] 2022-11-19 12:33:14 -0800
Branch: REL_13_STABLE [c13667b51] 2022-11-19 12:36:33 -0800
Branch: REL_12_STABLE [4cbcb7ed8] 2022-11-19 12:36:48 -0800
Branch: REL_11_STABLE [140c80372] 2022-11-19 12:36:52 -0800
-->
     <para>
      Avoid nominally-undefined behavior when accessing shared memory in
      32-bit builds (Andres Freund)
     </para>

     <para>
      clang's undefined-behavior sanitizer complained about use of a
      pointer that was less aligned than it should be.  It's very unlikely
      that this would cause a problem in non-debug builds, but it's worth
      fixing for testing purposes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [02699bc1f] 2022-12-30 20:49:50 +0100
Branch: REL_15_STABLE [c4f64cfab] 2022-12-30 20:49:11 +0100
Branch: REL_14_STABLE [883dc0214] 2022-12-30 20:47:58 +0100
-->
     <para>
      Fix assertion failure in BRIN minmax-multi opclasses (Tomas Vondra)
     </para>

     <para>
      The assertion was overly strict, so this mistake was harmless in
      non-assert builds.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [51dfaa0b0] 2022-11-29 10:52:44 -0500
Branch: REL_15_STABLE [556c0b913] 2022-11-29 10:52:44 -0500
Branch: REL_14_STABLE [0224646be] 2022-11-29 10:52:44 -0500
Branch: REL_13_STABLE [aca695fb6] 2022-11-29 10:52:44 -0500
Branch: REL_12_STABLE [bb8d48cb9] 2022-11-29 10:52:44 -0500
-->
     <para>
      Remove faulty assertion in useless-RESULT-RTE optimization logic
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [2fcf685f6] 2022-12-23 10:04:18 +0900
Branch: REL_15_STABLE [e3897a3a4] 2022-12-23 10:04:30 +0900
Branch: REL_14_STABLE [7ad458e06] 2022-12-23 10:04:33 +0900
Branch: REL_13_STABLE [4dbe72d61] 2022-12-23 10:04:34 +0900
Branch: REL_12_STABLE [7445869e1] 2022-12-23 10:04:36 +0900
Branch: REL_11_STABLE [df6fea51f] 2022-12-23 10:04:37 +0900
-->
     <para>
      Fix copy-and-paste errors in cache-lookup-failure messages for ACL
      checks (Justin Pryzby)
     </para>

     <para>
      In principle these errors should never be reached.  But if they are,
      some of them reported the wrong type of object.
     </para>
    </listitem>

    <listitem>
<!--
Author: Robert Haas <rhaas@postgresql.org>
Branch: master [45f5c81ad] 2022-12-12 10:26:48 -0500
Branch: REL_15_STABLE [8b5ba2f3f] 2022-12-12 10:33:02 -0500
-->
     <para>
      Fix possible corruption of very large tablespace map files
      in <application>pg_basebackup</application> (Antonin Houska)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [74739d1d3] 2023-01-19 19:32:50 -0500
Branch: REL_15_STABLE [488e89bf7] 2023-01-19 19:32:47 -0500
-->
     <para>
      Avoid harmless warning from <application>pg_dump</application>
      in <option>--if-exists</option> mode (Tom Lane)
     </para>

     <para>
      If the <literal>public</literal> schema has a non-default owner then
      use of <application>pg_dump</application>'s <option>--if-exists</option>
      option resulted in a warning message <quote>warning: could not find
      where to insert IF EXISTS in statement "-- *not* dropping schema,
      since initdb creates it"</quote>.  The dump output was okay, though.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [cabfb8241] 2022-12-02 14:24:44 -0500
Branch: REL_15_STABLE [97299cf99] 2022-12-02 14:24:44 -0500
Branch: REL_14_STABLE [303b26c1b] 2022-12-02 14:24:44 -0500
-->
     <para>
      Fix <application>psql</application>'s <literal>\sf</literal>
      and <literal>\ef</literal> commands to handle SQL-language functions
      that have <acronym>SQL</acronym>-standard function bodies (Tom Lane)
     </para>

     <para>
      These commands misidentified the start of the function body when it
      used new-style syntax.
     </para>
    </listitem>

    <listitem>
<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
Branch: master [e7a59094b] 2023-01-06 11:18:44 +0000
Branch: REL_15_STABLE [2daf4664c] 2023-01-06 11:16:53 +0000
Branch: REL_14_STABLE [48599a18d] 2023-01-06 11:15:22 +0000
Branch: REL_13_STABLE [2ad4abedf] 2023-01-06 11:13:34 +0000
Branch: REL_12_STABLE [274185d11] 2023-01-06 11:11:51 +0000
Branch: REL_11_STABLE [c54b88870] 2023-01-06 11:09:56 +0000
-->
     <para>
      Fix tab completion of <command>ALTER
      FUNCTION/PROCEDURE/ROUTINE</command> ... <command>SET
      SCHEMA</command> (Dean Rasheed)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [aeaaf520f] 2022-11-21 15:37:10 -0500
Branch: REL_15_STABLE [b8988cf1d] 2022-11-21 15:37:10 -0500
-->
     <para>
      Update <filename>contrib/pageinspect</filename> to mark its
      disk-accessing functions as <literal>PARALLEL RESTRICTED</literal>
      (Tom Lane)
     </para>

     <para>
      This avoids possible failure if one of these functions is used to
      examine a temporary table, since a session's temporary tables are not
      accessible from parallel workers.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [701c881f7] 2022-12-21 17:51:50 -0500
Branch: REL_15_STABLE [b87037b37] 2022-12-21 17:51:50 -0500
Branch: REL_14_STABLE [f489b480f] 2022-12-21 17:51:50 -0500
Branch: REL_13_STABLE [d35f1d485] 2022-12-21 17:51:50 -0500
Branch: REL_12_STABLE [c8314d62a] 2022-12-21 17:51:50 -0500
Branch: REL_11_STABLE [0ff4056b8] 2022-12-21 17:51:50 -0500
-->
     <para>
      Fix <filename>contrib/seg</filename> to not crash or print garbage
      if an input number has more than 127 digits (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: REL_15_STABLE [576506303] 2022-11-10 10:23:49 -0500
Branch: REL_14_STABLE [06dca17ad] 2022-11-10 10:23:49 -0500
Branch: REL_13_STABLE [0942acb73] 2022-11-10 10:23:49 -0500
Branch: REL_12_STABLE [cf0f465c0] 2022-11-10 10:23:49 -0500
-->
     <para>
      Fix build on Microsoft Visual Studio 2013 (Tom Lane)
     </para>

     <para>
      A previous patch supposed that all platforms of interest
      have <function>snprintf()</function>, but MSVC 2013 isn't quite
      there yet.  Revert to using <function>sprintf()</function> on that
      platform.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andrew Dunstan <andrew@dunslane.net>
Branch: master [341f4e002] 2022-11-25 15:28:38 -0500
Branch: REL_15_STABLE [fed54fc9a] 2022-11-25 15:37:33 -0500
Branch: REL_14_STABLE [f76191fd9] 2022-11-25 15:37:33 -0500
Branch: REL_13_STABLE [9fe5cff14] 2022-11-25 15:37:33 -0500
Branch: REL_12_STABLE [171c7fffa] 2022-11-25 15:37:34 -0500
Branch: REL_11_STABLE [ae7c51213] 2022-11-25 15:37:34 -0500
-->
     <para>
      Fix compile failure in building PL/Perl with MSVC when using
      Strawberry Perl (Andrew Dunstan)
     </para>
    </listitem>

    <listitem>
<!--
Author: Andrew Dunstan <andrew@dunslane.net>
Branch: master [ccc59a83c] 2022-11-27 09:03:22 -0500
Branch: REL_15_STABLE [b5d8fd418] 2022-11-27 09:18:14 -0500
Branch: REL_14_STABLE [f3f70b8de] 2022-11-27 09:18:20 -0500
Branch: REL_13_STABLE [68d89d80c] 2022-11-27 09:18:32 -0500
Branch: REL_12_STABLE [85565cbca] 2022-11-27 09:18:40 -0500
Branch: REL_11_STABLE [724dd5649] 2022-11-27 09:18:46 -0500
-->
     <para>
      Fix mismatch of PL/Perl built with MSVC versus a Perl library built
      with gcc (Andrew Dunstan)
     </para>

     <para>
      Such combinations could previously fail with <quote>loadable library
      and perl binaries are mismatched</quote> errors.
     </para>
    </listitem>

    <listitem>
<!--
Author: Andres Freund <andres@anarazel.de>
Branch: master [388e80132] 2022-12-29 12:47:29 -0800
Branch: REL_15_STABLE [c6e75e4c2] 2023-01-02 15:49:33 -0800
Branch: REL_14_STABLE [7b5dec760] 2023-01-02 15:50:00 -0800
Branch: REL_13_STABLE [ce073d014] 2023-01-02 15:50:33 -0800
Branch: REL_12_STABLE [f0e13802f] 2023-01-02 15:51:03 -0800
Branch: REL_11_STABLE [99f8bc335] 2023-01-02 15:51:05 -0800
-->
     <para>
      Suppress compiler warnings from Perl's header files (Andres Freund)
     </para>

     <para>
      Our preferred compiler options provoke warnings about constructs
      appearing in recent versions of Perl's header files.  When using
      <application>gcc</application>, we can suppress these warnings with
      a pragma.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: REL_15_STABLE [9a40a0311] 2023-01-20 11:58:12 -0500
Branch: REL_14_STABLE [21c058648] 2023-01-20 11:58:12 -0500
Branch: REL_13_STABLE [c78f109b8] 2023-01-20 11:58:12 -0500
Branch: REL_12_STABLE [6d066d56b] 2023-01-20 11:58:12 -0500
Branch: REL_11_STABLE [b69e9dfab] 2023-01-20 11:58:12 -0500
-->
     <para>
      Fix <application>pg_waldump</application> to build on compilers that
      don't discard unused static-inline functions (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [758f44bc3] 2023-01-31 17:36:55 -0500
Branch: REL_15_STABLE [65f0d9d27] 2023-01-31 17:37:06 -0500
Branch: REL_14_STABLE [7e6154779] 2023-01-31 17:37:14 -0500
Branch: REL_13_STABLE [20d9da107] 2023-01-31 17:37:22 -0500
Branch: REL_12_STABLE [2c95d8776] 2023-01-31 17:37:28 -0500
Branch: REL_11_STABLE [7ddc428ef] 2023-01-31 17:37:34 -0500
-->
     <para>
      Update time zone data files to <application>tzdata</application>
      release 2022g for DST law changes in Greenland and Mexico,
      plus historical corrections for northern Canada, Colombia, and
      Singapore.
     </para>

     <para>
      Notably, a new timezone America/Ciudad_Juarez has been split off
      from America/Ojinaga.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15-1">
  <title>Release 15.1</title>

  <formalpara>
  <title>Release date:</title>
  <para>2022-11-10</para>
  </formalpara>

  <para>
   This release contains a variety of fixes from 15.0.
   For information about new features in major release 15, see
   <xref linkend="release-15"/>.
  </para>

  <sect2>
   <title>Migration to Version 15.1</title>

   <para>
    A dump/restore is not required for those running 15.X.
   </para>

   <para>
    However, if you regularly create and drop tables exceeding 1GB,
    see the first changelog entry below.
   </para>
  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [0e758ae89] 2022-11-07 11:36:45 -0500
Branch: REL_15_STABLE [5fe0ab420] 2022-11-07 11:36:45 -0500
-->
     <para>
      Fix failure to remove non-first segments of large tables
      (Tom Lane)
     </para>

     <para>
      <productname>PostgreSQL</productname> splits large tables into
      multiple files (normally with 1GB per file).  The logic for dropping
      a table was broken and would miss removing all but the first such
      file, in two cases: drops of temporary tables and WAL replay of
      drops of regular tables.  Applications that routinely create
      multi-gigabyte temporary tables could suffer significant disk space
      leakage.
     </para>

     <para>
      Orphaned temporary-table files are removed during postmaster start,
      so the mere act of updating to 15.1 is sufficient to clear any
      leaked temporary-table storage.  However, if you suffered any
      database crashes while using 15.0, and there might have been
      large tables dropped just before such crashes, it's advisable
      to check the database directories for files named according to the
      pattern
      <literal><replaceable>NNNN</replaceable>.<replaceable>NN</replaceable></literal>.
      If there is no matching file named
      just <literal><replaceable>NNNN</replaceable></literal> (without
      the <literal>.<replaceable>NN</replaceable></literal> suffix), these
      files should be removed manually.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [b8f2687fd] 2022-10-11 18:24:14 -0400
Branch: REL_15_STABLE [07ce67698] 2022-10-11 18:24:14 -0400
Branch: REL_14_STABLE [3162bd95c] 2022-10-11 18:24:14 -0400
Branch: REL_13_STABLE [21e042b0b] 2022-10-11 18:24:14 -0400
Branch: REL_12_STABLE [abc510fa2] 2022-10-11 18:24:15 -0400
Branch: REL_11_STABLE [addde9bc6] 2022-10-11 18:24:15 -0400
Branch: REL_10_STABLE [23e2a06ac] 2022-10-11 18:24:15 -0400
-->
     <para>
      Fix handling of <literal>DEFAULT</literal> tokens that appear
      in a multi-row <literal>VALUES</literal> clause of an
      <command>INSERT</command> on an updatable view (Tom Lane)
     </para>

     <para>
      This oversight could lead to <quote>cache lookup failed for
      type</quote> errors, or in older branches even to crashes.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [797e313dc] 2022-10-17 12:14:39 -0400
Branch: REL_15_STABLE [4a41a069e] 2022-10-17 12:14:39 -0400
Branch: REL_14_STABLE [2f26cec48] 2022-10-17 12:14:39 -0400
Branch: REL_13_STABLE [b21615d1e] 2022-10-17 12:14:39 -0400
Branch: REL_12_STABLE [65c1106d8] 2022-10-17 12:14:39 -0400
Branch: REL_11_STABLE [e9377e3e5] 2022-10-17 12:14:39 -0400
Branch: REL_10_STABLE [ecf4ce689] 2022-10-17 12:14:39 -0400
-->
     <para>
      Disallow rules named <literal>_RETURN</literal> that are
      not <literal>ON SELECT</literal> (Tom Lane)
     </para>

     <para>
      This avoids confusion between a view's <literal>ON SELECT</literal>
      rule and any other rules it may have.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d57534740] 2022-10-16 19:18:08 -0400
Branch: REL_15_STABLE [d4abb0bc5] 2022-10-16 19:18:08 -0400
Branch: REL_14_STABLE [8122160ff] 2022-10-16 19:18:08 -0400
-->
     <para>
      Avoid failure in <command>EXPLAIN VERBOSE</command> for a query
      using <literal>SEARCH BREADTH FIRST</literal> with constant
      initial values (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [cba4e78f3] 2022-10-15 19:24:26 +0200
Branch: REL_15_STABLE [16d11d684] 2022-10-15 19:24:26 +0200
-->
     <para>
      Prevent use of <command>MERGE</command> on a partitioned table with
      foreign-table partitions (&Aacute;lvaro Herrera)
     </para>

     <para>
      The case isn't supported, and previously threw an incomprehensible
      error.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [b0284bfb1] 2022-11-03 20:40:21 +0100
Branch: REL_15_STABLE [c301e1c0c] 2022-11-03 20:40:21 +0100
Branch: REL_14_STABLE [18865f4df] 2022-11-03 20:40:21 +0100
Branch: REL_13_STABLE [41b6e7c9a] 2022-11-03 20:40:21 +0100
Branch: REL_12_STABLE [ab70b3a52] 2022-11-03 20:40:21 +0100
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [614a406b4] 2022-10-07 19:37:48 +0200
Branch: REL_15_STABLE Release: REL_15_0 [6083132ab] 2022-10-07 19:37:48 +0200
Branch: REL_14_STABLE [483d26930] 2022-10-07 19:37:48 +0200
Branch: REL_13_STABLE [7d520e68e] 2022-10-07 19:37:48 +0200
Branch: REL_12_STABLE [669803af0] 2022-10-07 19:37:48 +0200
-->
     <para>
      Fix construction of per-partition foreign key constraints while
      doing <command>ALTER TABLE ATTACH PARTITION</command>
      (Jehan-Guillaume de Rorthais, &Aacute;lvaro Herrera)
     </para>

     <para>
      Previously, incorrect or duplicate constraints could be constructed
      for the newly-added partition.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f4857082b] 2022-11-01 14:34:44 -0400
Branch: REL_15_STABLE [1f1865e90] 2022-11-01 14:34:44 -0400
-->
     <para>
      Fix planner failure with extended statistics on partitioned or
      inherited tables (Richard Guo, Justin Pryzby)
     </para>

     <para>
      Some cases failed with <quote>cache lookup failed for statistics
      object</quote>.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [56b662523] 2022-10-13 09:31:57 +0900
Branch: REL_15_STABLE [ca4070f2b] 2022-10-26 09:41:13 +0900
Branch: REL_14_STABLE [5a30d43fa] 2022-10-26 09:41:18 +0900
Branch: REL_13_STABLE [594b97509] 2022-10-26 09:41:22 +0900
Branch: REL_12_STABLE [51c24d9e2] 2022-10-26 09:41:26 +0900
Branch: REL_11_STABLE [341fba2a6] 2022-10-26 09:41:28 +0900
Branch: REL_10_STABLE [b02fc7df1] 2022-10-26 09:41:31 +0900
-->
     <para>
      Fix mis-ordering of WAL operations in fast insert path for GIN
      indexes (Matthias van de Meent, Zhang Mingli)
     </para>

     <para>
      This mistake is not known to have any negative consequences within
      core <productname>PostgreSQL</productname>, but it did cause issues
      for some extensions.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [16b1fe003] 2022-10-20 08:49:48 +0530
Branch: REL_15_STABLE [64ff0fe4e] 2022-10-20 09:43:59 +0530
Branch: REL_14_STABLE [a592ed923] 2022-10-20 09:34:18 +0530
Branch: REL_13_STABLE [25f7be1ca] 2022-10-20 09:25:13 +0530
Branch: REL_12_STABLE [1bf4d9206] 2022-10-20 09:16:28 +0530
Branch: REL_11_STABLE [5f7076cb6] 2022-10-20 09:07:04 +0530
Branch: REL_10_STABLE [10ed7b9e4] 2022-10-20 08:58:11 +0530
Branch: REL_15_STABLE [343afa967] 2022-10-21 10:03:35 +0530
Branch: REL_14_STABLE [4fbe6096b] 2022-10-21 09:52:44 +0530
Branch: REL_13_STABLE [38dbaaf27] 2022-10-21 09:42:24 +0530
Branch: REL_12_STABLE [02600886c] 2022-10-21 09:32:21 +0530
Branch: REL_11_STABLE [216af69ae] 2022-10-21 09:22:20 +0530
-->
     <para>
      Fix bugs in logical decoding when replay starts from a point
      between the beginning of a transaction and the beginning of its
      subtransaction (Masahiko Sawada, Kuroda Hayato)
     </para>

     <para>
      These errors could lead to assertion failures in debug builds, and
      otherwise to memory leaks.
     </para>
    </listitem>

    <listitem>
<!--
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [ce20f8b9f] 2022-10-21 12:57:18 +0530
Branch: REL_15_STABLE [10eaa9750] 2022-10-21 12:43:28 +0530
Branch: REL_14_STABLE [36fc013fa] 2022-10-21 12:33:47 +0530
Branch: REL_13_STABLE [1eed947f9] 2022-10-21 12:22:47 +0530
Branch: REL_12_STABLE [f7f82cf05] 2022-10-21 12:10:11 +0530
Branch: REL_11_STABLE [5c51afe23] 2022-10-21 12:08:14 +0530
Branch: REL_10_STABLE [61838d2dd] 2022-10-21 11:54:34 +0530
Author: Amit Kapila <akapila@postgresql.org>
Branch: master [f972ec5c2] 2022-08-23 10:20:02 +0530
Branch: REL_15_STABLE Release: REL_15_0 [42681dffa] 2022-08-23 10:08:43 +0530
Branch: REL_14_STABLE [6d05d575b] 2022-08-23 09:24:51 +0530
Branch: REL_13_STABLE [4985a4591] 2022-08-23 09:10:28 +0530
Branch: REL_12_STABLE [9415873ae] 2022-08-23 08:51:20 +0530
Branch: REL_11_STABLE [51e9469a4] 2022-08-23 08:42:51 +0530
Branch: REL_10_STABLE [6b50433e8] 2022-08-23 08:39:31 +0530
-->
     <para>
      Accept interrupts in more places during logical decoding (Amit
      Kapila, Masahiko Sawada)
     </para>

     <para>
      This ameliorates problems with slow shutdown of replication workers.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [be541efbf] 2022-11-02 12:29:39 -0400
Branch: REL_15_STABLE [414d29a82] 2022-11-02 12:29:39 -0400
Branch: REL_14_STABLE [a5b7821fc] 2022-11-02 12:29:39 -0400
Branch: REL_13_STABLE [4d3f7e75c] 2022-11-02 12:29:39 -0400
-->
     <para>
      Prevent attempts to replicate into a foreign-table partition in
      replication workers (Shi Yu, Tom Lane)
     </para>

     <para>
      Although partitioned tables can have foreign tables as partitions,
      replicating into such a partition isn't currently supported.
      The logical replication worker process would crash if it was
      attempted.  Now, an error is thrown.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [dea834938] 2022-11-03 12:02:14 -0400
Branch: REL_15_STABLE [f2dc7f9e3] 2022-11-03 12:01:57 -0400
Branch: REL_14_STABLE [2489c38cd] 2022-11-03 12:01:57 -0400
Branch: REL_13_STABLE [b00f342ea] 2022-11-03 12:01:57 -0400
Branch: REL_12_STABLE [d9ffccf8d] 2022-11-03 12:01:57 -0400
Branch: REL_11_STABLE [ed019b5ef] 2022-11-03 12:01:57 -0400
Branch: REL_10_STABLE [5f3cec77b] 2022-11-03 12:01:57 -0400
-->
     <para>
      Avoid crash after function syntax error in replication workers
      (Maxim Orlov, Anton Melnikov, Masahiko Sawada, Tom Lane)
     </para>

     <para>
      If a syntax error occurred in a SQL-language or PL/pgSQL-language
      <command>CREATE FUNCTION</command> or <command>DO</command> command
      executed in a logical replication worker, the worker process would
      crash with a null pointer dereference or assertion failure.
     </para>
    </listitem>

    <listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [9668c4a66] 2022-10-19 14:06:56 +0900
Branch: REL_15_STABLE [5d2a47a29] 2022-10-19 14:07:01 +0900
-->
     <para>
      Avoid double call of the shutdown callback of an archiver module
      (Nathan Bossart, Bharath Rupireddy)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [eec346611] 2022-10-17 11:35:23 -0400
Branch: REL_15_STABLE [2e3326929] 2022-10-17 11:35:23 -0400
Branch: REL_14_STABLE [8c611602b] 2022-10-17 11:35:23 -0400
Branch: REL_13_STABLE [62b263bf7] 2022-10-17 11:35:23 -0400
Branch: REL_12_STABLE [99b6b705d] 2022-10-17 11:35:23 -0400
-->
     <para>
      Add plan-time check for attempted access to a table that has no
      table access method (Tom Lane)
     </para>

     <para>
      This prevents a crash in some catalog-corruption scenarios, for
      example use of a view whose <literal>ON SELECT</literal> rule is
      missing.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [18a4a620e] 2022-10-11 18:54:31 -0400
Branch: REL_15_STABLE [e7b4ff327] 2022-10-11 18:54:31 -0400
Branch: REL_14_STABLE [b10546ecf] 2022-10-11 18:54:31 -0400
Branch: REL_13_STABLE [744270137] 2022-10-11 18:54:31 -0400
Branch: REL_12_STABLE [8f98352b5] 2022-10-11 18:54:31 -0400
Branch: REL_11_STABLE [6c1de98ba] 2022-10-11 18:54:31 -0400
Branch: REL_10_STABLE [ab35b9dd7] 2022-10-11 18:54:31 -0400
-->
     <para>
      Prevent postmaster crash when shared-memory state is corrupted
      (Tom Lane)
     </para>

     <para>
      The postmaster process is supposed to survive and initiate a
      database restart if shared memory becomes corrupted, but one
      bit of code was being insufficiently cautious about that.
     </para>
    </listitem>

    <listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [db1b931a4] 2022-10-14 19:06:26 +0200
Branch: REL_15_STABLE [27ca0bce5] 2022-10-14 19:06:26 +0200
Branch: REL_14_STABLE [b8af4166f] 2022-10-14 19:06:26 +0200
-->
     <para>
      In <application>libpq</application>, handle single-row mode
      correctly when pipelining (Denis Laxalde)
     </para>

     <para>
      The single-row flag was not reset at the correct time if pipeline
      mode was also active.
     </para>
    </listitem>

    <listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [2598b76bf] 2022-10-22 09:42:52 +0200
Branch: REL_15_STABLE [4a6de748d] 2022-10-22 09:41:38 +0200
-->
     <para>
      Fix <application>psql</application>'s exit status when a
      command-line query is canceled (Peter Eisentraut)
     </para>

     <para>
      <literal>psql -c <replaceable>query</replaceable></literal> would
      exit successfully if the query was canceled.  Fix it to exit with
      nonzero status, as in other error cases.
     </para>
    </listitem>

    <listitem>
<!--
Author: Robert Haas <rhaas@postgresql.org>
Branch: master [1f0c4fa25] 2022-10-21 08:21:55 -0400
Branch: REL_15_STABLE [5c013e620] 2022-10-21 08:31:39 -0400
Branch: REL_14_STABLE [aaad8adb0] 2022-10-21 08:39:48 -0400
Branch: REL_13_STABLE [0bf2cd160] 2022-10-21 08:45:08 -0400
Branch: REL_12_STABLE [475e9daf3] 2022-10-21 08:59:26 -0400
Branch: REL_11_STABLE [38214dabd] 2022-10-21 09:05:57 -0400
Branch: REL_10_STABLE [ba58266eb] 2022-10-21 09:11:47 -0400
-->
     <para>
      Allow cross-platform tablespace relocation
      in <application>pg_basebackup</application> (Robert Haas)
     </para>

     <para>
      Allow the remote path in <option>--tablespace-mapping</option> to be
      either a Unix-style or Windows-style absolute path, since the source
      server could be on a different OS than the local system.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [26ee7fb36] 2022-11-02 11:30:04 -0400
Branch: REL_15_STABLE [0eede9625] 2022-11-02 11:30:04 -0400
-->
     <para>
      Fix <application>pg_dump</application>'s failure to dump comments
      attached to some <literal>CHECK</literal> constraints (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [34fa0ddae] 2022-11-04 10:39:52 -0400
Branch: REL_15_STABLE [2c6d43650] 2022-11-04 10:39:52 -0400
-->
     <para>
      Fix <command>CREATE DATABASE</command> to allow
      its <literal>oid</literal> parameter to exceed
      2<superscript>31</superscript> (Tom Lane)
     </para>

     <para>
      This oversight prevented <application>pg_upgrade</application> from
      succeeding when the source installation contained databases with
      OIDs larger than that.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [495e73c20] 2022-11-01 12:48:01 -0400
Branch: REL_15_STABLE [8b0a5cf3f] 2022-11-01 12:48:01 -0400
Branch: REL_14_STABLE [0f2f5645a] 2022-11-01 12:48:01 -0400
Branch: REL_13_STABLE [a9fdb48b7] 2022-11-01 12:48:01 -0400
Branch: REL_12_STABLE [ca4c6764b] 2022-11-01 12:48:01 -0400
Branch: REL_11_STABLE [149e00192] 2022-11-01 12:48:01 -0400
Branch: REL_10_STABLE [56083ff30] 2022-11-01 12:48:01 -0400
-->
     <para>
      In <application>pg_stat_statements</application>, fix access to
      already-freed memory (zhaoqigui)
     </para>

     <para>
      This occurred if <application>pg_stat_statements</application>
      tracked a <command>ROLLBACK</command> command issued via extended
      query protocol.  In debug builds it consistently led to an assertion
      failure.  In production builds there would often be no visible ill
      effect; but if the freed memory had already been reused, the likely
      result would be to store garbage for the query string.
     </para>
    </listitem>

    <listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [c2ae01f69] 2022-10-19 22:18:26 +1300
Branch: REL_15_STABLE [af64846e1] 2022-10-19 22:18:54 +1300
Branch: REL_14_STABLE [d033f8f8b] 2022-10-19 22:32:14 +1300
Branch: REL_13_STABLE [cf94cb586] 2022-10-19 22:38:58 +1300
Branch: REL_12_STABLE [aa34bc4e2] 2022-10-19 22:44:53 +1300
Branch: REL_11_STABLE [da3a6825e] 2022-10-19 22:49:25 +1300
-->
     <para>
      Fix incompatibilities with LLVM 15 (Thomas Munro, Andres Freund)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [1c72d82c2] 2022-11-02 17:37:29 -0400
Branch: REL_15_STABLE [a5737e765] 2022-11-02 17:37:26 -0400
Branch: REL_14_STABLE [058c7b5dd] 2022-11-02 17:37:26 -0400
Branch: REL_13_STABLE [c479492c0] 2022-11-02 17:37:26 -0400
Branch: REL_12_STABLE [5ecf836e9] 2022-11-02 17:37:26 -0400
Branch: REL_11_STABLE [a0f9be1f9] 2022-11-02 17:37:26 -0400
Branch: REL_10_STABLE [19cefebe7] 2022-11-02 17:37:26 -0400
-->
     <para>
      Allow use of <function>__sync_lock_test_and_set()</function> for
      spinlocks on any machine (Tom Lane)
     </para>

     <para>
      This eases porting to new machine architectures, at least if you're
      using a compiler that supports this GCC builtin function.
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [717ec1aae] 2022-10-16 15:27:04 -0400
Branch: REL_15_STABLE [24c4c2617] 2022-10-16 15:27:04 -0400
Branch: REL_14_STABLE [18e60712d] 2022-10-16 15:27:04 -0400
Branch: REL_13_STABLE [bc7a40b42] 2022-10-16 15:27:04 -0400
Branch: REL_12_STABLE [3d7df87c4] 2022-10-16 15:27:04 -0400
Branch: REL_11_STABLE [6618c276b] 2022-10-16 15:27:04 -0400
Branch: REL_10_STABLE [02d074e3e] 2022-10-16 15:27:04 -0400
Branch: REL9_6_STABLE [9bfa043a1] 2022-10-16 15:27:04 -0400
Branch: REL9_5_STABLE [377b37cf7] 2022-10-16 15:27:04 -0400
Branch: REL9_4_STABLE [57dfb6ce6] 2022-10-16 15:27:04 -0400
Branch: REL9_3_STABLE [ee4c44014] 2022-10-16 15:27:04 -0400
Branch: REL9_2_STABLE [fec443414] 2022-10-16 15:27:04 -0400
-->
     <para>
      Rename symbol <literal>REF</literal> to <literal>REF_P</literal> to
      avoid compile failure on recent macOS (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [450ee7012] 2022-10-16 11:47:44 -0400
Branch: REL_15_STABLE [bd4b2926e] 2022-10-16 11:47:44 -0400
Branch: REL_14_STABLE [6fa431d84] 2022-10-16 11:47:44 -0400
Branch: REL_13_STABLE [a2acafc7b] 2022-10-16 11:47:44 -0400
Branch: REL_12_STABLE [d33ac1ec2] 2022-10-16 11:47:44 -0400
-->
     <para>
      Avoid using <function>sprintf</function>, to avoid compile-time
      deprecation warnings (Tom Lane)
     </para>
    </listitem>

    <listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [e7c7605a7] 2022-11-01 17:08:28 -0400
Branch: REL_15_STABLE [c3d16eb3d] 2022-11-01 17:08:49 -0400
Branch: REL_14_STABLE [97bb80b1b] 2022-11-01 17:08:58 -0400
Branch: REL_13_STABLE [ebf48810b] 2022-11-01 17:09:04 -0400
Branch: REL_12_STABLE [ec9a000d8] 2022-11-01 17:09:09 -0400
Branch: REL_11_STABLE [b1cb77bdf] 2022-11-01 17:09:16 -0400
Branch: REL_10_STABLE [b3326a753] 2022-11-01 17:09:21 -0400
-->
     <para>
      Update time zone data files to <application>tzdata</application>
      release 2022f for DST law changes in Chile, Fiji, Iran, Jordan,
      Mexico, Palestine, and Syria, plus historical corrections for Chile,
      Crimea, Iran, and Mexico.
     </para>

     <para>
      Also, the Europe/Kiev zone has been renamed to Europe/Kyiv.
      Also, the following zones have been merged into nearby,
      more-populous zones whose clocks have agreed with them since 1970:
      Antarctica/Vostok, Asia/Brunei,
      Asia/Kuala_Lumpur, Atlantic/Reykjavik, Europe/Amsterdam,
      Europe/Copenhagen, Europe/Luxembourg, Europe/Monaco, Europe/Oslo,
      Europe/Stockholm, Indian/Christmas, Indian/Cocos, Indian/Kerguelen,
      Indian/Mahe, Indian/Reunion, Pacific/Chuuk, Pacific/Funafuti,
      Pacific/Majuro, Pacific/Pohnpei, Pacific/Wake and Pacific/Wallis.
      (This indirectly affects zones that were already links to one of
      these: Arctic/Longyearbyen, Atlantic/Jan_Mayen, Iceland,
      Pacific/Ponape, Pacific/Truk, and Pacific/Yap.)  America/Nipigon,
      America/Rainy_River, America/Thunder_Bay, Europe/Uzhgorod, and
      Europe/Zaporozhye were also merged into nearby zones after
      discovering that their claimed post-1970 differences from those
      zones seem to have been errors.
      In all these cases, the previous zone name remains as an alias;
      but the actual data is that of the zone that was merged into.
     </para>

     <para>
      These zone mergers result in loss of pre-1970 timezone history for
      the merged zones, which may be troublesome for applications
      expecting consistency of <type>timestamptz</type> display.  As an
      example, the stored value <literal>1944-06-01 12:00 UTC</literal>
      would previously display as <literal>1944-06-01
      13:00:00+01</literal> if the Europe/Stockholm zone is selected, but
      now it will read out as <literal>1944-06-01 14:00:00+02</literal>.
     </para>

     <para>
      It is possible to build the time zone data files with options that
      will restore the older zone data, but that choice also inserts a lot
      of other old (and typically poorly-attested) zone data, resulting in
      more total changes from the previous release than accepting these
      upstream changes does.  <productname>PostgreSQL</productname> has
      chosen to ship the <productname>tzdb</productname> data
      as-recommended, and so far as we are aware most major operating
      system distributions are doing likewise.  However, if these changes
      cause significant problems for your application, a possible solution
      is to install a local build of the time zone data files using
      <productname>tzdb</productname>'s backwards-compatibility options
      (see their <literal>PACKRATDATA</literal>
      and <literal>PACKRATLIST</literal> options).
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-15">
  <title>Release 15</title>

  <formalpara>
   <title>Release date:</title>
   <para>2022-10-13</para>
  </formalpara>

  <sect2>
   <title>Overview</title>

   <para>
    <productname>PostgreSQL</productname> 15 contains many new features
    and enhancements, including:
   </para>

   <itemizedlist>
    <listitem>
     <para>
      Support for the <acronym>SQL</acronym>
      <link linkend="sql-merge"><command>MERGE</command></link> command.
     </para>
    </listitem>
    <listitem>
     <para>
      Selective publication of tables' contents within
      <link linkend="logical-replication">logical replication</link>
      publications, through the ability to specify column lists and
      row filter conditions.
     </para>
    </listitem>
    <listitem>
     <para>
      More options for compression, including support for Zstandard (zstd)
      compression.  This includes support for performing compression on
      the server side during
      <link linkend="app-pgbasebackup"><application>pg_basebackup</application></link>.
     </para>
    </listitem>
    <listitem>
     <para>
      Support for structured <link linkend="guc-log-destination">server
      log output</link> using the <acronym>JSON</acronym> format.
     </para>
    </listitem>
    <listitem>
     <para>
      Performance improvements, particularly for in-memory and on-disk
      sorting.
     </para>
    </listitem>
   </itemizedlist>

   <para>
    The above items and other new features of
    <productname>PostgreSQL</productname> 15 are explained in more detail
    in the sections below.
   </para>

  </sect2>

  <sect2>

   <title>Migration to Version 15</title>

    <para>
     A dump/restore using <xref linkend="app-pg-dumpall"/> or use of
     <xref linkend="pgupgrade"/> or logical replication is required for
     those wishing to migrate data from any previous release.  See <xref
     linkend="upgrading"/> for general information on migrating to new
     major releases.
    </para>

    <para>
     Version 15 contains a number of changes that may affect compatibility
     with previous releases.  Observe the following incompatibilities:
    </para>

   <itemizedlist>

<!--
Author: Noah Misch <noah@leadboat.com>
2021-09-09 [b073c3ccd] Revoke PUBLIC CREATE from public schema, now owned by pg
-->

    <listitem>
     <para>
      Remove <literal>PUBLIC</literal> creation permission on the <link
      linkend="ddl-schemas-public"><literal>public</literal> schema</link>
      (Noah Misch)
     </para>

     <para>
      The new default is one of the secure schema usage patterns that <xref
      linkend="ddl-schemas-patterns"/> has recommended since the security
      release for CVE-2018-1058.  The change applies to new database
      clusters and to newly-created databases in existing clusters.
      Upgrading a cluster or restoring a database dump will preserve
      <literal>public</literal>'s existing permissions.
     </para>

     <para>
      For existing databases, especially those having multiple users,
      consider revoking <literal>CREATE</literal> permission on
      the <literal>public</literal> schema to adopt this new default.
      For new databases having no need to defend against insider threats,
      granting <literal>CREATE</literal> permission will yield the behavior
      of prior releases.
     </para>
    </listitem>

<!--
Author: Noah Misch <noah@leadboat.com>
2021-09-09 [b073c3ccd] Revoke PUBLIC CREATE from public schema, now owned by pg
-->

    <listitem>
     <para>
      Change the owner of the <literal>public</literal> schema to be the
      new <literal>pg_database_owner</literal> role (Noah Misch)
     </para>

     <para>
      This allows each database's owner to have ownership privileges on
      the <literal>public</literal> schema within their database.
      Previously it was owned by the bootstrap superuser, so that
      non-superuser database owners could not do anything with it.
     </para>

     <para>
      This change applies to new database clusters and to newly-created
      databases in existing clusters.
      Upgrading a cluster or restoring a database dump will preserve
      <literal>public</literal>'s existing ownership specification.
     </para>
    </listitem>

<!--
Author: Stephen Frost <sfrost@snowman.net>
2022-04-06 [39969e2a1] Remove exclusive backup mode
-->

    <listitem>
     <para>
      Remove long-deprecated <link linkend="backup-base-backup">exclusive
      backup mode</link> (David Steele, Nathan Bossart)
     </para>

     <para>
      If the database server stops abruptly while in this mode, the
      server could fail to start.  The non-exclusive backup mode is
      considered superior for all purposes.  Functions
      <function>pg_start_backup()</function>/<function>pg_stop_backup()</function>
      have been renamed to
      <function>pg_backup_start()</function>/<function>pg_backup_stop()</function>,
      and the functions <function>pg_backup_start_time()</function>
      and <function>pg_is_in_backup()</function> have been removed.
     </para>
    </listitem>

<!--
Author: Peter Geoghegan <pg@bowt.ie>
2022-02-16 [8f388f6f5] Increase hash_mem_multiplier default to 2.0.
-->

    <listitem>
     <para>
      Increase <link
      linkend="guc-hash-mem-multiplier"><varname>hash_mem_multiplier</varname></link>
      default to 2.0 (Peter Geoghegan)
     </para>

     <para>
      This allows query hash operations to use more
      <link linkend="guc-work-mem"><varname>work_mem</varname></link>
      memory than other operations.
     </para>
    </listitem>

<!--
Author: Andres Freund <andres@anarazel.de>
2022-03-07 [76a29adee] plpython: Remove plpythonu, plpython2u and associated tr
-->

    <listitem>
     <para>
      Remove server-side language <link
      linkend="plpython"><literal>plpython2u</literal></link> and generic
      Python language <literal>plpythonu</literal> (Andres Freund)
     </para>

     <para>
      Python 2.x is no longer supported.  While the original intent of
      <literal>plpythonu</literal> was that it could eventually refer
      to <literal>plpython3u</literal>, changing it now seems more likely
      to cause problems than solve them, so it's just been removed.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-11-06 [cbe25dcff] Disallow making an empty lexeme via array_to_tsvector().
-->

    <listitem>
     <para>
      Generate an error if <link
      linkend="textsearch-functions-table"><function>array_to_tsvector()</function></link>
      is passed an empty-string array element (Jean-Christophe Arnu)
     </para>

     <para>
      This is prohibited because lexemes should never be empty.  Users of
      previous Postgres releases should verify that no empty lexemes
      are stored because they can lead to dump/restore failures and
      inconsistent results.
     </para>
    </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2021-12-06 [e9e63b702] Fix inappropriate uses of PG_GETARG_UINT32()
-->

    <listitem>
     <para>
      Generate an error when <link
      linkend="functions-string-other"><function>chr()</function></link>
      is supplied with a negative argument (Peter Eisentraut)
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-15 [2523928b2] Reject change of output-column collation in CREATE OR RE
-->

    <listitem>
     <para>
      Prevent <link linkend="sql-createview"><command>CREATE OR REPLACE
      VIEW</command></link> from changing the collation of an output column
      (Tom Lane)
     </para>
    </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-01-12 [a18b6d2dc] ecpg: Catch zero-length Unicode identifiers correctly
-->

    <listitem>
     <para>
      Disallow zero-length <link
      linkend="sql-syntax-identifiers">Unicode identifiers</link>,
      e.g., <literal>U&amp;""</literal>
      (Peter Eisentraut)
     </para>

     <para>
      Non-Unicode zero-length identifiers were already disallowed.
     </para>
    </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-02-16 [2549f0661] Reject trailing junk after numeric literals
-->

    <listitem>
     <para>
      Prevent <link linkend="sql-syntax-constants-numeric">numeric
      literals</link> from having non-numeric trailing characters (Peter
      Eisentraut)
     </para>

     <para>
      Previously, query text like <literal>123abc</literal> would be
      interpreted as <literal>123</literal> followed
      by a separate token <literal>abc</literal>.
     </para>
    </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-03-28 [e26114c81] Make JSON path numeric literals more correct
-->

    <listitem>
     <para>
      Adjust <link linkend="datatype-json"><acronym>JSON</acronym></link>
      numeric literal processing to match the
      <acronym>SQL</acronym>/<acronym>JSON</acronym>-standard (Peter
      Eisentraut)
     </para>

     <para>
      This accepts numeric formats like <literal>.1</literal> and
      <literal>1.</literal>, and disallows trailing junk after numeric
      literals, like <literal>1.type()</literal>.
     </para>
    </listitem>

<!--
Author: Bruce Momjian <bruce@momjian.us>
2021-08-03 [95ab1e0a9] interval:  round values when spilling to months
-->

    <listitem>
     <para>
      When <link linkend="datatype-datetime"><type>interval</type></link>
      input provides a fractional value for a unit greater than months,
      round to the nearest month (Bruce Momjian)
     </para>

     <para>
      For example, convert <literal>1.99 years</literal> to <literal>2
      years</literal>, not <literal>1 year 11 months</literal> as before.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-03 [591e088dd] Fix portability issues in datetime parsing.
-->

    <listitem>
     <para>
      Improve consistency of <type>interval</type> parsing with trailing
      periods (Tom Lane)
     </para>

     <para>
      Numbers with trailing periods were rejected on some platforms.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-19 [a3d6264bb] interval_out() must be marked STABLE, not IMMUTABLE.
-->

    <listitem>
     <para>
      Mark the <type>interval</type> output
      function as stable, not immutable, since it depends on <link
      linkend="guc-intervalstyle"><varname>IntervalStyle</varname></link>
      (Tom Lane)
     </para>

     <para>
      This will, for example, cause creation of indexes relying on the
      text output of <type>interval</type> values to fail.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-28 [54bd1e43c] Handle integer overflow in interval justification functi
-->

    <listitem>
     <para>
      Detect integer overflow in <link
      linkend="functions-datetime-table">interval justification
      functions</link> (Joe Koshakow)
     </para>

     <para>
      The affected functions are <function>justify_interval()</function>,
      <function>justify_hours()</function>, and
      <function>justify_days()</function>.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-08-02 [c034b629c] Change type "char"'s I/O format for non-ASCII characters
-->

    <listitem>
     <para>
      Change the I/O format of type <type>"char"</type> for non-ASCII
      characters (Tom Lane)
     </para>

     <para>
      Bytes with the high bit set are now output as a backslash and three
      octal digits, to avoid encoding issues.
     </para>
    </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-03-28 [79de9842a] Remove the ability of a role to administer itself.
-->

    <listitem>
     <para>
      Remove the default <link linkend="sql-createrole"><literal>ADMIN
      OPTION</literal></link> privilege a login role has on its own role
      membership (Robert Haas)
     </para>

     <para>
      Previously, a login role could add/remove members of its own role,
      even without <literal>ADMIN OPTION</literal> privilege.
     </para>
    </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2022-01-07 [a2ab9c06e] Respect permissions within logical replication.
-->

    <listitem>
     <para>
      Allow <link linkend="logical-replication">logical replication</link>
      to run as the owner of the subscription (Mark Dilger)
     </para>

     <para>
      Because row-level security policies are not checked, only superusers,
      roles with <literal>bypassrls</literal>, and table owners can
      replicate into tables with row-level security policies.
     </para>
    </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2022-01-08 [96a6f11c0] More cleanup of a2ab9c06ea.
-->

    <listitem>
     <para>
      Prevent <command>UPDATE</command> and <command>DELETE</command>
      <link linkend="logical-replication">logical replication</link>
      operations on tables where the subscription owner does not have
      <command>SELECT</command> permission on the table (Jeff Davis)
     </para>

     <para>
      <command>UPDATE</command> and <command>DELETE</command> commands
      typically involve reading the table as well, so require the
      subscription owner to have table <command>SELECT</command>
      permission.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-07-27 [48c5c9068] Use the "pg_temp" schema alias in EXPLAIN and related ou
-->

    <listitem>
     <para>
      When <link linkend="sql-explain"><command>EXPLAIN</command></link>
      references the session's temporary object schema, refer to it as
      <literal>pg_temp</literal> (Amul Sul)
     </para>

     <para>
      Previously the actual schema name was reported, leading to
      inconsistencies across sessions.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-03-24 [ce95c5437] Fix pg_statio_all_tables view for multiple TOAST indexes
-->

    <listitem>
     <para>
      Fix <link
      linkend="monitoring-pg-statio-all-tables-view"><structname>pg_statio_all_tables</structname></link>
      to sum values for the rare case of <acronym>TOAST</acronym> tables
      with multiple indexes (Andrei Zubkov)
     </para>

     <para>
      Previously such cases would show one row for each index.
     </para>
    </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2021-12-01 [75d22069e] Warning on SET of nonexisting setting with a prefix rese
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-27 [2ed8a8cc5] Rethink handling of settings with a prefix reserved by a
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-21 [88103567c] Disallow setting bogus GUCs within an extension's reserv
-->

    <listitem>
     <para>
      Disallow setting <link linkend="runtime-config-custom">custom
      options</link> that match the name of an installed extension, but
      are not one of the extension's declared variables
      (Florin Irion, Tom Lane)
     </para>

     <para>
      This change causes any such pre-existing variables to be deleted
      during extension load, and then prevents new ones from being created
      later in the session.  The intent is to prevent confusion about
      whether a variable is associated with an extension or not.
     </para>
    </listitem>

<!--
Author: Andres Freund <andres@anarazel.de>
2022-04-06 [6f0cf8787] pgstat: remove stats_temp_directory.
-->

    <listitem>
     <para>
      Remove obsolete server variable
      <varname>stats_temp_directory</varname> (Andres Freund, Kyotaro
      Horiguchi)
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-11-28 [3804539e4] Replace random(), pg_erand48(), etc with a better PRNG A
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-12 [d4f109e4a] Doc: update description of random() function.
-->

    <listitem>
     <para>
      Improve the algorithm used to compute <link
      linkend="functions-math-random-table"><function>random()</function></link>
      (Fabien Coelho)
     </para>

     <para>
      This will cause <function>random()</function>'s results to differ
      from what was emitted by prior versions, even for the same seed
      value.
     </para>
    </listitem>

<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-09-23 [bd8ac900d] Remove PQsendQuery support in pipeline mode
-->

    <listitem>
     <para>
      <application>libpq</application>'s <link
      linkend="libpq-PQsendQuery"><function>PQsendQuery()</function></link>
      function is no longer supported in pipeline mode (Álvaro Herrera)
     </para>

     <para>
      Applications that are using that combination will need to be
      modified to use <function>PQsendQueryParams()</function> instead.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-09 [376ce3e40] Prefer $HOME when looking up the current user's home dir
-->

    <listitem>
     <para>
      On non-Windows platforms, consult the <envar>HOME</envar> environment
      variable to find the user's home directory (Anders Kaseorg)
     </para>

     <para>
      If <envar>HOME</envar> is empty or unset, fall back to the previous
      method of checking the <literal>&lt;pwd.h&gt;</literal> database.
      This change affects <application>libpq</application> (for example,
      while looking up <filename>~/.pgpass</filename>) as well as various
      client application programs.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-15 [2a712066d] Remove pg_dump's - -no-synchronized-snapshots switch.
-->

    <listitem>
     <para>
      Remove <link
      linkend="app-pgdump"><application>pg_dump</application></link>'s
      <option>--no-synchronized-snapshots</option> option (Tom Lane)
     </para>

     <para>
      All still-supported server versions support synchronized snapshots,
      so there's no longer a need for this option.
     </para>
    </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-06-15 [a3ff08e0b] Tweak behavior of psql - -single-transaction depending
-->

      <listitem>
       <para>
        After an error is detected in <link
        linkend="app-psql"><application>psql</application></link>'s
        <option>--single-transaction</option> mode, change the
        final <command>COMMIT</command> command
        to <command>ROLLBACK</command> only
        if <varname>ON_ERROR_STOP</varname> is set (Michael Paquier)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-11-12 [f8abb0f5e] postgres_fdw: suppress casts on constants in limited cas
-->

    <listitem>
     <para>
      Avoid unnecessary casting of constants in queries sent by <link
      linkend="postgres-fdw">postgres_fdw</link> (Dian Fay)
     </para>

     <para>
      When column types are intentionally different between local and
      remote databases, such casts could cause errors.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-14 [fd2abeb7c] Delete contrib/xml2's legacy implementation of xml_is_we
-->

    <listitem>
     <para>
      Remove <link linkend="xml2">xml2</link>'s
      <function>xml_is_well_formed()</function> function (Tom Lane)
     </para>

     <para>
      This function has been implemented in the core backend since
      Postgres 9.1.
     </para>
    </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-07-06 [955b3e0f9] Allow CustomScan providers to say whether they support p
-->

    <listitem>
     <para>
      Allow <link linkend="custom-scan">custom scan providers</link>
      to indicate if they support projections (Sven Klemm)
     </para>

     <para>
      The default is now that custom scan providers are assumed to not
      support projections;  those that do will need to be updated for
      this release.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>

  <sect2>
   <title>Changes</title>

   <para>
    Below you will find a detailed account of the changes between
    <productname>PostgreSQL</productname> 15 and the previous major
    release.
   </para>

   <sect3>
    <title>Server</title>

    <itemizedlist>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-02-14 [37851a8b8] Database-level collation version tracking
-->

     <listitem>
      <para>
       Record and check the collation version of each <link
       linkend="sql-createdatabase">database</link> (Peter Eisentraut)
      </para>

      <para>
       This feature is designed to detect collation version
       changes to avoid index corruption.  Function
       <function>pg_database_collation_actual_version()</function>
       reports the underlying operating system collation version, and
       <command>ALTER DATABASE ...  REFRESH</command> sets the recorded
       database collation version to match the operating system collation
       version.
      </para>
     </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-03-17 [f2553d430] Add option to use ICU as global locale provider
-->

     <listitem>
      <para>
       Allow <link linkend="locale"><acronym>ICU</acronym></link>
       collations to be set as the default for clusters and databases
       (Peter Eisentraut)
      </para>

      <para>
       Previously, only <application>libc</application>-based
       collations could be selected at the cluster and database levels.
       <acronym>ICU</acronym> collations could only be used via explicit
       <literal>COLLATE</literal> clauses.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-03-29 [a2c84990b] Add system view pg_ident_file_mappings
-->

     <listitem>
      <para>
       Add system view <link
       linkend="view-pg-ident-file-mappings"><structname>pg_ident_file_mappings</structname></link>
       to report <filename>pg_ident.conf</filename> information (Julien
       Rouhaud)
      </para>
     </listitem>

       </itemizedlist>

    <sect4>
     <title><link linkend="ddl-partitioning">Partitioning</link></title>

     <itemizedlist>

<!--
Author: David Rowley <drowley@postgresql.org>
2021-08-03 [475dbd0b7] Track a Bitmapset of non-pruned partitions in RelOptInfo
-->

      <listitem>
       <para>
        Improve planning time for queries referencing partitioned tables
        (David Rowley)
       </para>

       <para>
        This change helps when only a few of many partitions are relevant.
       </para>
      </listitem>

<!--
Author: David Rowley <drowley@postgresql.org>
2021-08-03 [db632fbca] Allow ordered partition scans in more cases
-->

      <listitem>
       <para>
        Allow ordered scans of partitions to avoid sorting in more cases
        (David Rowley)
       </para>

       <para>
        Previously, a partitioned table with a <literal>DEFAULT</literal>
        partition or a <literal>LIST</literal> partition containing
        multiple values could not be used for ordered partition scans.
        Now they can be used if such partitions are pruned during planning.
       </para>
      </listitem>

<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-03-20 [ba9a7e392] Enforce foreign key correctly during cross-partition upd
-->

      <listitem>
       <para>
        Improve foreign key behavior of updates on partitioned tables
        that move rows between partitions (Amit Langote)
       </para>

       <para>
        Previously, such updates ran a delete action on the source
        partition and an insert action on the target partition.
        <productname>PostgreSQL</productname> will now run an update action
        on the partition root, providing cleaner semantics.
       </para>
      </listitem>

<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-04-02 [cfdd03f45] Allow CLUSTER on partitioned tables
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-04-14 [3f19e176a] Have CLUSTER ignore partitions not owned by caller
-->

      <listitem>
       <para>
        Allow <link linkend="sql-cluster"><command>CLUSTER</command></link>
        on partitioned tables (Justin Pryzby)
       </para>
      </listitem>

<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2021-07-22 [80ba4bb38] Make ALTER TRIGGER RENAME consistent for partitioned tab
-->

      <listitem>
       <para>
        Fix <link linkend="sql-altertable"><command>ALTER TRIGGER
        RENAME</command></link> on partitioned tables to properly rename
        triggers on all partitions (Arne Roland, Álvaro Herrera)
       </para>

       <para>
        Also prohibit cloned triggers from being renamed.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Indexes</title>

     <itemizedlist>

<!--
Author: Peter Geoghegan <pg@bowt.ie>
2021-10-02 [2903f1404] Enable deduplication in system catalog indexes.
-->

      <listitem>
       <para>
        Allow btree indexes on system and <link
        linkend="storage-toast"><acronym>TOAST</acronym></link>
        tables to efficiently store duplicates (Peter Geoghegan)
       </para>

       <para>
        Previously de-duplication was disabled for these types of indexes.
       </para>
      </listitem>

<!--
Author: Alexander Korotkov <akorotkov@postgresql.org>
2022-02-07 [f1ea98a79] Reduce non-leaf keys overlap in GiST indexes produced by
-->

      <listitem>
       <para>
        Improve lookup performance
        of <link linkend="gist"><acronym>GiST</acronym></link> indexes
        that were built using sorting (Aliaksandr Kalenik, Sergei
        Shoulbakov, Andrey Borodin)
       </para>
      </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-02-03 [94aa7cc5f] Add UNIQUE null treatment option
-->

      <listitem>
       <para>
        Allow unique constraints and indexes to treat
        <literal>NULL</literal> values as not distinct (Peter Eisentraut)
       </para>

       <para>
        Previously <literal>NULL</literal> entries were always treated
        as distinct values, but this can now be changed by creating
        constraints and indexes using <literal>UNIQUE NULLS NOT
        DISTINCT</literal>.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-11-17 [a148f8bc0] Add a planner support function for starts_with().
-->

      <listitem>
       <para>
        Allow the <link
        linkend="functions-string-other"><literal>^@</literal></link>
        starts-with operator and the <function>starts_with()</function>
        function to use btree indexes if using the C collation (Tom Lane)
       </para>

       <para>
        Previously these could only use <link
        linkend="spgist"><acronym>SP-GiST</acronym></link> indexes.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Optimizer</title>

     <itemizedlist>

<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2022-01-16 [269b532ae] Add stxdinherit flag to pg_statistic_ext_data
-->

      <listitem>
       <para>
        Allow <link linkend="sql-createstatistics">extended
        statistics</link> to record statistics for a parent with all its
        children (Tomas Vondra, Justin Pryzby)
       </para>

       <para>
        Regular statistics already tracked parent and
        parent-plus-all-children statistics separately.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-03-24 [0bd7af082] Invent recursive_worktable_factor GUC to replace hard-wi
-->

      <listitem>
       <para>
        Add server variable <link
        linkend="guc-recursive-worktable-factor"><varname>recursive_worktable_factor</varname></link>
        to allow the user to specify the expected size of the working
        table of a <link linkend="queries-with-recursive">recursive
        query</link> (Simon Riggs)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>General Performance</title>

     <itemizedlist>

<!--
Author: David Rowley <drowley@postgresql.org>
2021-07-07 [29f45e299] Use a hash table to speed up NOT IN(values)
-->

      <listitem>
       <para>
        Allow hash lookup for <link
        linkend="functions-subquery-notin"><literal>NOT IN</literal></link>
        clauses with many constants (David Rowley, James Coleman)
       </para>

       <para>
        Previously the code always sequentially scanned the list of values.
       </para>
      </listitem>

<!--
Author: David Rowley <drowley@postgresql.org>
2021-08-22 [22c4e88eb] Allow parallel DISTINCT
-->

     <listitem>
      <para>
       Allow <command>SELECT DISTINCT</command> to be parallelized
       (David Rowley)
      </para>
     </listitem>

<!--
Author: John Naylor <john.naylor@postgresql.org>
2021-12-20 [911588a3f] Add fast path for validating UTF-8 text
-->

      <listitem>
       <para>
        Speed up encoding validation of <acronym>UTF</acronym>-8 text
        by processing 16 bytes at a time
        (John Naylor, Heikki Linnakangas)
       </para>

       <para>
        This will improve text-heavy operations like <link
        linkend="sql-copy"><command>COPY FROM</command></link>.
       </para>
      </listitem>

<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
2021-10-18 [65014000b] Replace polyphase merge algorithm with a simple balanced
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
2021-10-25 [166f94377] Clarify the logic in a few places in the new balanced me
-->

      <listitem>
       <para>
        Improve performance for sorts that exceed <link
        linkend="guc-work-mem"><varname>work_mem</varname></link>
        (Heikki Linnakangas)
       </para>

       <para>
        When the sort data no longer fits in <varname>work_mem</varname>,
        switch to a batch sorting algorithm that uses more output streams
        than before.
       </para>
      </listitem>

<!--
Author: David Rowley <drowley@postgresql.org>
2021-07-22 [91e9e89dc] Make nodeSort.c use Datum sorts for single column sorts
Author: David Rowley <drowley@postgresql.org>
2022-04-04 [40af10b57] Use Generation memory contexts to store tuples in sorts
Author: John Naylor <john.naylor@postgresql.org>
2022-04-02 [697492434] Specialize tuplesort routines for different kinds of abb
-->

      <listitem>
       <para>
        Improve performance and reduce memory consumption of in-memory
        sorts (Ronan Dunklau, David Rowley, Thomas Munro, John Naylor)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-06-29 [4035cd5d4] Add support for LZ4 with compression of full-page writes
Author: Michael Paquier <michael@paquier.xyz>
2022-03-11 [e9537321a] Add support for zstd with compression of full-page write
-->

      <listitem>
       <para>
        Allow <acronym>WAL</acronym> <link
        linkend="guc-full-page-writes">full page writes</link> to use
        LZ4 and Zstandard compression (Andrey Borodin, Justin Pryzby)
       </para>

       <para>
        This is controlled by the <link
        linkend="guc-wal-compression"><varname>wal_compression</varname></link>
        server setting.
       </para>
      </listitem>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2021-07-19 [2dbe89057] Support direct I/O on macOS.
-->

      <listitem>
       <para>
        Add support for writing <acronym>WAL</acronym>
        using <link linkend="guc-wal-sync-method">direct I/O</link> on
        macOS (Thomas Munro)
       </para>

       <para>
        This only works if <literal>max_wal_senders = 0</literal>
        and <literal>wal_level = minimal</literal>.
       </para>
      </listitem>

<!--
Author: Peter Geoghegan <pg@bowt.ie>
2022-04-03 [0b018faba] Set relfrozenxid to oldest extant XID seen by VACUUM.
-->

      <listitem>
       <para>
        Allow <link linkend="routine-vacuuming">vacuum</link> to be more
        aggressive in setting the oldest frozen and multi transaction id
        (Peter Geoghegan)
       </para>
      </listitem>

<!--
Author: Etsuro Fujita <efujita@postgresql.org>
2022-04-06 [c2bb02bc2] Allow asynchronous execution in more cases.
-->

      <listitem>
       <para>
        Allow a query referencing multiple <link
        linkend="ddl-foreign-data">foreign tables</link> to perform
        parallel foreign table scans in more cases (Andrey Lepikhov,
        Etsuro Fujita)
       </para>
      </listitem>

<!--
Author: David Rowley <drowley@postgresql.org>
2022-04-08 [9d9c02ccd] Teach planner and executor about monotonic window funcs
-->

      <listitem>
       <para>
        Improve the performance of <link linkend="functions-window">window
        functions</link> that use <function>row_number()</function>,
        <function>rank()</function>, <function>dense_rank()</function> and
        <function>count()</function>
        (David Rowley)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-06 [a82a5eee3] Use ISB as a spin-delay instruction on ARM64.
-->

      <listitem>
       <para>
        Improve the performance of spinlocks on high-core-count ARM64
        systems (Geoffrey Blake)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Monitoring</title>

     <itemizedlist>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2021-12-13 [64da07c41] Default to log_checkpoints=on, log_autovacuum_min_durati
-->

      <listitem>
       <para>
        Enable default logging of checkpoints and slow autovacuum
        operations (Bharath Rupireddy)
       </para>

       <para>
        This changes the default of <link
        linkend="guc-log-checkpoints"><varname>log_checkpoints</varname></link>
        to <literal>on</literal> and that of <link
        linkend="guc-log-autovacuum-min-duration"><varname>log_autovacuum_min_duration</varname></link>
        to 10 minutes.  This will cause even an idle server to generate
        some log output, which might cause problems on
        resource-constrained servers without log file rotation.  These
        defaults should be changed in such cases.
       </para>
      </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2021-10-25 [9ce346eab] Report progress of startup operations that take a long t
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2021-11-05 [e543906e2] Document default and changeability of log_startup_progre
-->

      <listitem>
       <para>
        Generate progress messages in the server log during slow server
        starts (Nitin Jadhav, Robert Haas)
       </para>

       <para>
        The messages report the cause of the delay.  The time interval for
        notification is controlled by the new server variable <link
        linkend="guc-log-startup-progress-interval"><varname>log_startup_progress_interval</varname></link>.
       </para>
      </listitem>

<!--
Author: Andres Freund <andres@anarazel.de>
2022-04-06 [5891c7a8e] pgstat: store statistics in shared memory.
Author: Andres Freund <andres@anarazel.de>
2022-04-07 [b3abca681] pgstat: Update docs to match the shared memory stats rea
-->

      <listitem>
       <para>
        Store <link linkend="monitoring-stats">cumulative statistics
        system</link> data in shared memory (Kyotaro Horiguchi, Andres
        Freund, Melanie Plageman)
       </para>

       <para>
        Previously this data was sent to a statistics collector process
        via <acronym>UDP</acronym> packets, and could only be read by
        sessions after transferring it via the file system.  There is no
        longer a separate statistics collector process.
       </para>
      </listitem>

<!--
Author: Peter Geoghegan <pg@bowt.ie>
2022-02-11 [872770fd6] Add VACUUM instrumentation for scanned pages, relfrozenx
Author: Peter Geoghegan <pg@bowt.ie>
2022-03-13 [6e20f4600] VACUUM VERBOSE: tweak scanned_pages logic.
Author: Peter Geoghegan <pg@bowt.ie>
2022-04-15 [bdb71dbe8] VACUUM VERBOSE: Show dead items for an empty table.
-->

      <listitem>
       <para>
        Add additional information to <command>VACUUM VERBOSE</command>
        and autovacuum logging messages (Peter Geoghegan)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-04-08 [efb0ef909] Track I/O timing for temporary file blocks in EXPLAIN (B
-->

      <listitem>
       <para>
        Add <link linkend="sql-explain"><command>EXPLAIN
        (BUFFERS)</command></link> output for temporary file block I/O
        (Masahiko Sawada)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-01-17 [dc686681e] Introduce log_destination=jsonlog
-->

      <listitem>
       <para>
        Allow <link linkend="guc-log-destination">log output</link> in
        <acronym>JSON</acronym> format (Sehrope Sarkuni, Michael Paquier)
       </para>

       <para>
        The new setting is <literal>log_destination = jsonlog</literal>.
       </para>
      </listitem>

<!--
Author: Fujii Masao <fujii@postgresql.org>
2021-09-02 [e04267844] Enhance pg_stat_reset_single_table_counters function.
-->

      <listitem>
       <para>
        Allow <link
        linkend="monitoring-stats-funcs-table"><function>pg_stat_reset_single_table_counters()</function></link>
        to reset the counters of relations shared across all databases
        (Sadhuprasad Patro)
       </para>
      </listitem>

<!--
Author: Fujii Masao <fujii@postgresql.org>
2021-11-22 [1b06d7bac] Report wait events for local shell commands like archive
-->

      <listitem>
       <para>
        Add <link linkend="wait-event-table">wait events</link> for local
        shell commands (Fujii Masao)
       </para>

       <para>
        The new wait events are used when calling
        <varname>archive_command</varname>,
        <varname>archive_cleanup_command</varname>,
        <varname>restore_command</varname> and
        <varname>recovery_end_command</varname>.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Privileges</title>

     <itemizedlist>

<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
2022-03-22 [7faa5fc84] Add support for security invoker views.
-->

      <listitem>
       <para>
        Allow table accesses done by
        a <link linkend="sql-createview">view</link> to optionally be
        controlled by privileges of the view's caller (Christoph Heiss)
       </para>

       <para>
        Previously, view accesses were always treated as being done by the
        view's owner.  That's still the default.
       </para>
      </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-28 [7f6772317] Adjust server-side backup to depend on pg_write_server_f
-->

      <listitem>
       <para>
        Allow members of the <link
        linkend="predefined-roles-table"><literal>pg_write_server_files</literal></link>
        predefined role to perform server-side base backups (Dagfinn
        Ilmari Mannsåker)
       </para>

       <para>
        Previously only superusers could perform such backups.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-06 [a0ffa885e] Allow granting SET and ALTER SYSTEM privileges on GUC pa
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-07-19 [a2944d872] Fix missed corner cases for grantable permissions on GUC
-->

      <listitem>
       <para>
        Allow <link linkend="sql-grant"><command>GRANT</command></link>
        to grant permissions to change individual server variables via
        <command>SET</command> and <command>ALTER SYSTEM</command>
        (Mark Dilger)
       </para>

       <para>
        The new function <function>has_parameter_privilege()</function>
        reports on this privilege.
       </para>
      </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2021-11-09 [4168a4745] Add pg_checkpointer predefined role for CHECKPOINT comma
Author: Robert Haas <rhaas@postgresql.org>
2022-07-05 [d3526e59f] Rename pg_checkpointer predefined role to pg_checkpoint.
-->

      <listitem>
       <para>
        Add predefined role <link
        linkend="predefined-roles-table"><literal>pg_checkpoint</literal></link>
        that allows members to run <command>CHECKPOINT</command>
        (Jeff Davis)
       </para>

       <para>
        Previously checkpoints could only be run by superusers.
       </para>
      </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2021-10-27 [77ea4f943] Grant memory views to pg_read_all_stats.
-->

      <listitem>
       <para>
        Allow members of the <link
        linkend="predefined-roles-table"><literal>pg_read_all_stats</literal></link>
        predefined role to access the views <link
        linkend="view-pg-backend-memory-contexts"><structname>pg_backend_memory_contexts</structname></link>
        and <link
        linkend="view-pg-shmem-allocations"><structname>pg_shmem_allocations</structname></link>
        (Bharath Rupireddy)
       </para>

       <para>
        Previously these views could only be accessed by superusers.
       </para>
      </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2021-10-26 [f0b051e32] Allow GRANT on pg_log_backend_memory_contexts().
-->

      <listitem>
       <para>
        Allow <link linkend="sql-grant"><command>GRANT</command></link>
        to grant permissions on <link
        linkend="functions-admin-signal"><function>pg_log_backend_memory_contexts()</function></link>
        (Jeff Davis)
       </para>

       <para>
        Previously this function could only be run by superusers.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Server Configuration</title>

     <itemizedlist>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-09-08 [bd1788051] Introduce GUC shared_memory_size
Author: Michael Paquier <michael@paquier.xyz>
2021-09-09 [3b231596c] Make shared_memory_size a preset option
-->

      <listitem>
       <para>
        Add server variable <link
        linkend="guc-shared-memory-size"><varname>shared_memory_size</varname></link>
        to report the size of allocated shared memory (Nathan Bossart)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-09-21 [43c1c4f65] Introduce GUC shared_memory_size_in_huge_pages
Author: Michael Paquier <michael@paquier.xyz>
2022-03-24 [bbd4951b7] doc: Improve postgres command for shared_memory_size_in_
-->

      <listitem>
       <para>
        Add server variable <link
        linkend="guc-shared-memory-size-in-huge-pages"><varname>shared_memory_size_in_huge_pages</varname></link>
        to report the number of huge memory pages required (Nathan Bossart)
       </para>

       <para>
        This is only supported on Linux.
       </para>
      </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2022-07-20 [2103266a3] Process shared_preload_libraries in single-user mode.
-->

      <listitem>
       <para>
        Honor server variable <link
        linkend="guc-shared-preload-libraries"><varname>shared_preload_libraries</varname></link>
        in single-user mode (Jeff Davis)
       </para>

       <para>
        This change supports use
        of <varname>shared_preload_libraries</varname> to load custom
        access methods and WAL resource managers, which would be essential
        for database access even in single-user mode.
       </para>
      </listitem>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2022-07-02 [a2b0719cc] Default to dynamic_shared_memory_type=sysv on Solaris.
-->

      <listitem>
       <para>
        On Solaris, make the default setting of <link
        linkend="guc-dynamic-shared-memory-type"><varname>dynamic_shared_memory_type</varname></link>
        be <literal>sysv</literal> (Thomas Munro)
       </para>

       <para>
        The previous default choice, <literal>posix</literal>, can result
        in spurious failures on this platform.
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-09-16 [0c39c2920] Support "postgres -C" with runtime-computed GUCs
-->

      <listitem>
       <para>
        Allow <link linkend="app-postgres"><command>postgres
        -C</command></link> to properly report runtime-computed values
        (Nathan Bossart)
       </para>

       <para>
        Previously runtime-computed values <link
        linkend="guc-data-checksums"><varname>data_checksums</varname></link>,
        <link
        linkend="guc-wal-segment-size"><varname>wal_segment_size</varname></link>,
        and <link
        linkend="guc-data-directory-mode"><varname>data_directory_mode</varname></link>
        would report values that would not be accurate on the running
        server.  However, this does not work on a running server.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

   </sect3>

   <sect3>
    <title>Streaming Replication and Recovery</title>

    <itemizedlist>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-02-11 [dab298471] Add suport for server-side LZ4 base backup compression.
Author: Robert Haas <rhaas@postgresql.org>
2022-03-08 [7cf085f07] Add support for zstd base backup compression.
Author: Robert Haas <rhaas@postgresql.org>
2022-03-30 [51c0d186d] Allow parallel zstd compression when taking a base backu
-->

     <listitem>
      <para>
       Add support for LZ4 and Zstandard compression of server-side <link
       linkend="backup-base-backup">base backups</link> (Jeevan Ladhe,
       Robert Haas)
      </para>
     </listitem>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2021-08-02 [7ff23c6d2] Run checkpointer and bgwriter in crash recovery.
-->

     <listitem>
      <para>
       Run the checkpointer and bgwriter processes during crash recovery
       (Thomas Munro)
      </para>

      <para>
       This helps to speed up long crash recoveries.
      </para>
     </listitem>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2022-04-07 [5dc0418fa] Prefetch data referenced by the WAL, take II.
-->

     <listitem>
      <para>
       Allow <acronym>WAL</acronym> processing to pre-fetch needed file
       contents (Thomas Munro)
      </para>

      <para>
       This is controlled by the server variable <link
       linkend="guc-recovery-prefetch"><varname>recovery_prefetch</varname></link>.
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-02-03 [5ef1eefd7] Allow archiving via loadable modules.
-->

     <listitem>
      <para>
       Allow archiving via loadable modules (Nathan Bossart)
      </para>

      <para>
       Previously, archiving was only done by calling shell commands.
       The new server variable <link
       linkend="guc-archive-library"><varname>archive_library</varname></link>
       can be set to specify a library to be called for archiving.
      </para>
     </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2021-07-09 [8e7811e95] Eliminate replication protocol error related to IDENTIFY
-->

     <listitem>
      <para>
       No longer require <link
       linkend="protocol-replication"><literal>IDENTIFY_SYSTEM</literal></link>
       to be run before <literal>START_REPLICATION</literal> (Jeff Davis)
      </para>
     </listitem>

    </itemizedlist>

    <sect4>
     <title><link linkend="logical-replication">Logical Replication</link></title>

     <itemizedlist>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2021-10-27 [5a2832465] Allow publishing the tables of schema.
Author: Amit Kapila <akapila@postgresql.org>
2021-12-08 [1a2aaeb0d] Fix changing the ownership of ALL TABLES IN SCHEMA publi
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-09-22 [f256236fb] Remove ALL keyword from TABLES IN SCHEMA for publication
-->

      <listitem>
       <para>
        Allow <link linkend="sql-createpublication">publication</link> of
        all tables in a schema (Vignesh C, Hou Zhijie, Amit Kapila)
       </para>

       <para>
        For example, this syntax is now supported: <literal>CREATE
        PUBLICATION pub1 FOR TABLES IN SCHEMA s1,s2</literal>.
        <command>ALTER PUBLICATION</command> supports a similar syntax.
        Tables added later to the listed schemas will also be replicated.
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2022-02-22 [52e4f0cd4] Allow specifying row filters for logical replication of
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2022-03-17 [5a0796622] Fix row filters with multiple publications
Author: Amit Kapila <akapila@postgresql.org>
2022-04-18 [676eeb6dd] Add additional documentation for row filters.
-->

      <listitem>
       <para>
        Allow publication content to be filtered using a
        <literal>WHERE</literal> clause (Hou Zhijie, Euler Taveira,
        Peter Smith, Ajin Cherian, Tomas Vondra, Amit Kapila)
       </para>

       <para>
        Rows not satisfying the <literal>WHERE</literal> clause are not
        published.
       </para>
      </listitem>

<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2022-03-26 [923def9a5] Allow specifying column lists for logical replication
-->

      <listitem>
       <para>
        Allow publication content to
        be restricted to specific columns (Tomas Vondra, Álvaro Herrera,
        Rahila Syed)
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2022-03-22 [208c5d65b] Add ALTER SUBSCRIPTION ... SKIP.
-->

      <listitem>
       <para>
        Allow skipping of transactions on a subscriber using <link
        linkend="sql-altersubscription"><command>ALTER SUBSCRIPTION
        ... SKIP</command></link> (Masahiko Sawada)
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2021-06-30 [cda03cfed] Allow enabling two-phase option via replication protocol
Author: Amit Kapila <akapila@postgresql.org>
2021-07-14 [a8fd13cab] Add support for prepared transactions to built-in logica
Author: Amit Kapila <akapila@postgresql.org>
2021-08-04 [63cf61cde] Add prepare API support for streaming transactions in lo
-->

      <listitem>
       <para>
        Add support for prepared (two-phase) transactions to logical
        replication (Peter Smith, Ajin Cherian, Amit Kapila, Nikhil
        Sontakke, Stas Kelvich)
       </para>

       <para>
        The new <link
        linkend="protocol-replication"><literal>CREATE_REPLICATION_SLOT</literal></link>
        option is called <literal>TWO_PHASE</literal>.
        <application>pg_recvlogical</application> now supports a new
        <option>--two-phase</option> option during slot creation.
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2022-03-30 [d5a9d86d8] Skip empty transactions for logical replication.
-->

      <listitem>
       <para>
        Prevent logical replication of empty transactions (Ajin Cherian,
        Hou Zhijie, Euler Taveira)
       </para>

       <para>
        Previously, publishers would send empty transactions to
        subscribers if subscribed tables were not modified.
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-11-23 [1922d7c6e] Add SQL functions to monitor the directory contents of r
-->

      <listitem>
       <para>
        Add <acronym>SQL</acronym> functions to monitor the directory
        contents of logical replication slots (Bharath Rupireddy)
       </para>

       <para>
        The new functions are <link
        linkend="functions-admin-genfile-table"><function>pg_ls_logicalsnapdir()</function></link>,
        <function>pg_ls_logicalmapdir()</function>, and
        <function>pg_ls_replslotdir()</function>.  They can be run by
        members of the predefined <literal>pg_monitor</literal> role.
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2022-03-14 [705e20f85] Optionally disable subscriptions on error.
-->

      <listitem>
       <para>
        Allow subscribers to stop the application of logical replication changes on error
        (Osumi Takamichi, Mark Dilger)
       </para>

       <para>
        This is enabled with the subscriber option <link
        linkend="sql-createsubscription"><literal>disable_on_error</literal></link>
        and avoids possible infinite error loops during stream application.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-11-02 [f3d4019da] Ensure consistent logical replication of datetime and fl
-->

      <listitem>
       <para>
        Adjust subscriber server variables to match the publisher so
        datetime and float8 values are interpreted consistently (Japin Li)
       </para>

       <para>
        Some publishers might be relying on inconsistent behavior.
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2021-11-30 [8d74fc96d] Add a view to show the stats of subscription workers.
Author: Amit Kapila <akapila@postgresql.org>
2022-03-01 [7a8507329] Reconsider pg_stat_subscription_workers view.
-->

      <listitem>
       <para>
        Add system view <link
        linkend="monitoring-pg-stat-subscription-stats"><structname>pg_stat_subscription_stats</structname></link>
        to report on subscriber activity (Masahiko Sawada)
       </para>

       <para>
        The new function <link
        linkend="monitoring-stats-functions"><function>pg_stat_reset_subscription_stats()</function></link>
        allows resetting these statistics counters.
       </para>
      </listitem>

<!--
Author: Amit Kapila <akapila@postgresql.org>
2021-12-08 [a61bff2bf] De-duplicate the result of pg_publication_tables view.
-->

      <listitem>
       <para>
        Suppress duplicate entries in the <link
        linkend="view-pg-publication-tables"><structname>pg_publication_tables</structname></link>
        system view (Hou Zhijie)
       </para>

       <para>
        In some cases a partition could appear more than once.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

   </sect3>

   <sect3>
    <title>Utility Commands</title>

    <itemizedlist>

<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-03-28 [7103ebb7a] Add support for MERGE SQL command
-->

     <listitem>
      <para>
       Add <acronym>SQL</acronym> <link
       linkend="sql-merge"><command>MERGE</command></link>
       command to adjust one table to match another (Simon Riggs, Pavan
       Deolasee, Álvaro Herrera, Amit Langote)
      </para>

      <para>
       This is similar to <command>INSERT ... ON CONFLICT</command>
       but more batch-oriented.
      </para>
     </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-01-28 [43f33dc01] Add HEADER support to COPY text format
Author: Peter Eisentraut <peter@eisentraut.org>
2022-03-30 [072132f04] Add header matching mode to COPY FROM
-->

     <listitem>
      <para>
       Add support for <literal>HEADER</literal> option in <link
       linkend="sql-copy"><command>COPY</command></link> text format
       (Rémi Lapeyre)
      </para>

      <para>
       The new option causes the column names to be output, and optionally
       verified on input.
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-03-29 [9c08aea6a] Add new block-by-block strategy for CREATE DATABASE.
-->

     <listitem>
      <para>
       Add new <acronym>WAL</acronym>-logged method for <link
       linkend="sql-createdatabase">database creation</link> (Dilip Kumar)
      </para>

      <para>
       This is the new default method for copying the template database,
       as it avoids the need for checkpoints during database creation.
       However, it might be slow if the template database is large, so
       the old method is still available.
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-24 [aa0105141] pg_upgrade: Preserve database OIDs.
-->

     <listitem>
      <para>
       Allow <link linkend="sql-createdatabase"><command>CREATE
       DATABASE</command></link> to set the database <acronym>OID</acronym>
       (Shruthi Gowda, Antonin Houska)
      </para>
     </listitem>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2022-02-12 [4eb217631] Fix DROP {DATABASE,TABLESPACE} on Windows.
-->

     <listitem>
      <para>
       Prevent <link linkend="sql-dropdatabase"><command>DROP
       DATABASE</command></link>, <link
       linkend="sql-droptablespace"><command>DROP
       TABLESPACE</command></link>, and <link
       linkend="sql-alterdatabase"><command>ALTER DATABASE SET
       TABLESPACE</command></link> from occasionally failing during
       concurrent use on Windows (Thomas Munro)
      </para>
     </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2021-12-08 [d6f96ed94] Allow specifying column list for foreign key ON DELETE S
-->

     <listitem>
      <para>
       Allow foreign key <link linkend="ddl-constraints-fk"><literal>ON
       DELETE SET</literal></link> actions to affect only specified columns
       (Paul Martinez)
      </para>

      <para>
       Previously, all of the columns in the foreign key were always
       affected.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-07-28 [b0483263d] Add support for SET ACCESS METHOD in ALTER TABLE
-->

     <listitem>
      <para>
       Allow <link linkend="sql-altertable"><command>ALTER
       TABLE</command></link> to modify a table's <literal>ACCESS
       METHOD</literal> (Justin Pryzby, Jeff Davis)
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-08-10 [7b565843a] Add call to object access hook at the end of table rewri
-->

     <listitem>
      <para>
       Properly call object access hooks when <link
       linkend="sql-altertable"><command>ALTER TABLE</command></link>
       causes table rewrites (Michael Paquier)
      </para>
     </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-04-07 [344d62fb9] Unlogged sequences
-->

     <listitem>
      <para>
       Allow creation of unlogged <link
       linkend="sql-createsequence">sequences</link> (Peter Eisentraut)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-07-22 [c2fa113dd] Close old gap in dependency checks for functions returni
-->

     <listitem>
      <para>
       Track dependencies on individual columns in the results of
       functions returning composite types (Tom Lane)
      </para>

      <para>
       Previously, if a view or rule contained a reference to a specific
       column within the result of a composite-returning function, that
       was not noted as a dependency; the view or rule was only considered
       to depend on the composite type as a whole.  This meant that
       dropping the individual column would be allowed, causing problems
       in later use of the view or rule.  The column-level dependency is
       now also noted, so that dropping such a column will be rejected
       unless the view is changed or dropped.
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title>Data Types</title>

    <itemizedlist>

<!--
Author: Dean Rasheed <dean.a.rasheed@gmail.com>
2021-07-26 [085f931f5] Allow numeric scale to be negative or greater than preci
-->

     <listitem>
      <para>
       Allow the scale of
       a <link linkend="datatype-numeric"><type>numeric</type></link>
       value to be negative, or greater than its precision (Dean Rasheed,
       Tom Lane)
      </para>

      <para>
       This allows rounding of values to the left of the decimal point,
       e.g., <literal>'1234'::numeric(4, -2)</literal> returns 1200.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-02 [e39f99046] Fix overflow hazards in interval input and output conver
-->

     <listitem>
      <para>
       Improve overflow detection when casting values to <link
       linkend="datatype-datetime">interval</link> (Joe Koshakow)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-08-02 [c034b629c] Change type "char"'s I/O format for non-ASCII characters
-->

    <listitem>
     <para>
      Change the I/O format of type <type>"char"</type> for non-ASCII
      characters (Tom Lane)
     </para>
    </listitem>

<!--
Author: John Naylor <john.naylor@postgresql.org>
2021-08-26 [bab982161] Update display widths as part of updating Unicode
Author: Peter Eisentraut <peter@eisentraut.org>
2021-09-15 [f7e56f1f5] Update Unicode data to Unicode 14.0.0
-->

     <listitem>
      <para>
       Update the display width information of modern Unicode characters,
       like emojis (Jacob Champion)
      </para>

      <para>
       Also update from Unicode 5.0 to 14.0.0.  There is now an automated
       way to keep Postgres updated with Unicode releases.
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title>Functions</title>

    <itemizedlist>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-03-30 [7ae1619bc] Add range_agg with multirange inputs
-->

     <listitem>
      <para>
       Add multirange input to <link
       linkend="functions-aggregate-table"><function>range_agg()</function></link>
       (Paul Jungwirth)
      </para>
     </listitem>

<!--
Author: Fujii Masao <fujii@postgresql.org>
2022-02-10 [400fc6b64] Add min() and max() aggregates for xid8.
-->

     <listitem>
      <para>
       Add <link linkend="tutorial-agg"><function>MIN()</function></link>
       and <function>MAX()</function> aggregates for the <link
       linkend="datatype-int"><type>xid8</type></link> data type (Ken Kato)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-08-03 [642433707] Add assorted new regexp_xxx SQL functions.
-->

     <listitem>
      <para>
       Add regular expression functions for compatibility with other
       relational systems (Gilles Darold, Tom Lane)
      </para>

      <para>
       The new functions are <link
       linkend="functions-string-other"><function>regexp_count()</function></link>,
       <function>regexp_instr()</function>,
       <function>regexp_like()</function>, and
       <function>regexp_substr()</function>.  Some new optional arguments
       were also added to <function>regexp_replace()</function>.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-13 [c5c192d7b] Implement poly_distance().
-->

     <listitem>
      <para>
       Add the ability to compute the distance between <link
       linkend="datatype-polygon"><type>polygons</type></link> (Tom Lane)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-03-14 [9dde82899] Support "of", "tzh", and "tzm" format codes.
-->

     <listitem>
      <para>
       Add <link
       linkend="functions-formatting-table"><function>to_char()</function></link>
       format codes <literal>of</literal>, <literal>tzh</literal>, and
       <literal>tzm</literal> (Nitin Jadhav)
      </para>

      <para>
       The upper-case equivalents of these were already supported.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-09-06 [388e71af8] Make timetz_zone() stable, and correct a bug for DYNTZ a
-->

     <listitem>
      <para>
       When applying <link
       linkend="functions-datetime-zoneconvert"><literal>AT
       TIME ZONE</literal></link> to a <type>time with time zone</type>
       value, use the transaction start time rather than wall clock time
       to determine whether DST applies (Aleksander Alekseev, Tom Lane)
      </para>

      <para>
       This allows the conversion to be considered stable rather than
       volatile, and it saves a kernel call per invocation.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-11-06 [cbe25dcff] Disallow making an empty lexeme via array_to_tsvector().
-->

     <listitem>
      <para>
       Ignore NULL array elements in <link
       linkend="textsearch-functions-table"><function>ts_delete()</function></link> and
       <function>setweight()</function> functions with array arguments
       (Jean-Christophe Arnu)
      </para>

      <para>
       These functions effectively ignore empty-string array elements
       (since those could never match a valid lexeme).  It seems
       consistent to let them ignore NULL elements too, instead of
       failing.
      </para>
     </listitem>

<!--
Author: David Rowley <drowley@postgresql.org>
2021-07-09 [ca2e4472b] Teach pg_size_pretty and pg_size_bytes about petabytes
-->

     <listitem>
      <para>
       Add support for petabyte units to <link
       linkend="functions-admin-dbsize"><function>pg_size_pretty()</function></link>
       and <function>pg_size_bytes()</function> (David Christensen)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-07-27 [024515cac] In event triggers, use "pg_temp" only for our own temp s
-->

     <listitem>
      <para>
       Change <link
       linkend="pg-event-trigger-ddl-command-end-functions"><function>pg_event_trigger_ddl_commands()</function></link>
       to output references to other sessions' temporary schemas using the
       actual schema name (Tom Lane)
      </para>

      <para>
       Previously this function reported all temporary schemas as
       <literal>pg_temp</literal>, but it's misleading to use that for any
       but the current session's temporary schema.
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title><link linkend="plpgsql">PL/pgSQL</link></title>

    <itemizedlist>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-30 [ccd10a9bf] Tighten enforcement of variable CONSTANT markings in plp
-->

     <listitem>
      <para>
       Fix enforcement of PL/pgSQL variable <literal>CONSTANT</literal>
       markings (Tom Lane)
      </para>

      <para>
       Previously, a variable could be used as a <link
       linkend="plpgsql-statements-calling-procedure"><command>CALL</command></link>
       output parameter or refcursor <command>OPEN</command> variable
       despite being marked <literal>CONSTANT</literal>.
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title><link linkend="libpq">libpq</link></title>

    <itemizedlist>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-04-01 [c1932e542] libpq: Allow IP address SANs in server certificates
-->

     <listitem>
      <para>
       Allow <acronym>IP</acronym> address matching against a server
       certificate's Subject Alternative Name (Jacob Champion)
      </para>
     </listitem>

<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
2022-03-29 [ebc8b7d44] Enable SSL library detection via PQsslAttribute()
-->

     <listitem>
      <para>
       Allow <function>PQsslAttribute()</function> to report the
       <acronym>SSL</acronym> library type without requiring a libpq
       connection (Jacob Champion)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-18 [5987feb70] Make PQcancel use the PGconn's tcp_user_timeout and keep
-->

     <listitem>
      <para>
       Change query cancellations sent by the client to use the same
       <acronym>TCP</acronym> settings as normal client connections
       (Jelte Fennema)
      </para>

      <para>
       This allows configured <acronym>TCP</acronym> timeouts to apply
       to query cancel connections.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-18 [ce1e7a2f7] Don't let libpq "event" procs break the state of PGresul
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-18 [2e372869a] Don't let libpq PGEVT_CONNRESET callbacks break a PGconn
-->

     <listitem>
      <para>
       Prevent libpq event callback failures from forcing an error result
       (Tom Lane)
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title>Client Applications</title>

    <itemizedlist>

<!--
Author: Tatsuo Ishii <ishii@postgresql.org>
2022-03-23 [4a39f87ac] Allow pgbench to retry in some cases.
-->

     <listitem>
      <para>
       Allow <link
       linkend="pgbench"><application>pgbench</application></link> to
       retry after serialization and deadlock failures (Yugo Nagata,
       Marina Polyakova)
      </para>
     </listitem>

    </itemizedlist>

     <sect4>
      <title><xref linkend="app-psql"/></title>

      <itemizedlist>

<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
2021-07-14 [eec57115e] In psql \copy from, send data to server in larger chunks
-->

      <listitem>
       <para>
        Improve performance
        of <application>psql</application>'s <command>\copy</command>
        command, by sending data in larger chunks (Heikki Linnakangas)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-07 [3e707fbb4] psql: add \dconfig command to show server's configuratio
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-11 [5e70d8b5d] Tweak the default behavior of psql's \dconfig.
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-13 [139d46ee2] Further tweak the default behavior of psql's \dconfig.
-->

      <listitem>
       <para>
        Add <command>\dconfig</command> command to report server variables
        (Mark Dilger, Tom Lane)
       </para>

       <para>
        This is similar to the server-side <command>SHOW</command>
        command, but it can process patterns to show multiple variables
        conveniently.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-20 [33d3eeadb] Add a \getenv command to psql.
-->

      <listitem>
       <para>
        Add <command>\getenv</command> command
        to assign the value of an environment variable to a
        <application>psql</application> variable (Tom Lane)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-06 [328dfbdab] Extend psql's \lo_list/\dl to be able to print large obj
-->

      <listitem>
       <para>
        Add <literal>+</literal> option to the
        <literal>\lo_list</literal> and <literal>\dl</literal> commands to
        show large-object privileges (Pavel Luzanov)
       </para>
      </listitem>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2021-07-13 [7c09d2797] Add PSQL_WATCH_PAGER for psql's \watch command.
-->

      <listitem>
       <para>
        Add a pager option for the <command>\watch</command>
        command (Pavel Stehule, Thomas Munro)
       </para>

       <para>
        This is only supported on Unix and is controlled by the
        <envar>PSQL_WATCH_PAGER</envar> environment variable.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-01 [83884682f] psql: include intra-query "- -" comments in what's sent t
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-01 [c2f654930] psql: treat "- -" comments between queries as separate hi
-->

      <listitem>
       <para>
        Make <application>psql</application> include intra-query double-hyphen
        comments in queries sent to the server (Tom Lane, Greg Nancarrow)
       </para>

       <para>
        Previously such comments were removed from the query
        before being sent.  Double-hyphen comments that are before any
        query text are not sent, and are not recorded as separate
        <application>psql</application> history entries.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-01 [3d858af07] psql: initialize comment-begin setting to a useful value
-->

      <listitem>
       <para>
        Adjust <application>psql</application> so
        that <application>Readline</application>'s
        meta-<literal>#</literal> command will insert a double-hyphen
        comment marker (Tom Lane)
       </para>

       <para>
        Previously a pound marker was inserted, unless the user had taken
        the trouble to configure a non-default comment marker.
       </para>
      </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2022-04-04 [7844c9918] psql: Show all query results by default
-->

      <listitem>
       <para>
        Make <application>psql</application> output all results when multiple
        queries are passed to the server at once (Fabien Coelho)
       </para>

       <para>
        Previously, only the last query result was displayed.  The old
        behavior can be restored by setting
        the <literal>SHOW_ALL_RESULTS</literal> <application>psql</application>
        variable to <literal>off</literal>.
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-06-15 [a3ff08e0b] Tweak behavior of psql - -single-transaction depending
-->

      <listitem>
       <para>
        After an error is detected
        in <option>--single-transaction</option> mode, change the
        final <command>COMMIT</command> command
        to <command>ROLLBACK</command> only
        if <varname>ON_ERROR_STOP</varname> is set (Michael Paquier)
       </para>

       <para>
        Previously, detection of an error in a <option>-c</option> command
        or <option>-f</option> script file would lead to
        issuing <command>ROLLBACK</command> at the end, regardless of the
        value of <varname>ON_ERROR_STOP</varname>.
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-08-10 [e2ce88b58] Add tab completion for DECLARE .. ASENSITIVE in psql
Author: Michael Paquier <michael@paquier.xyz>
2021-08-25 [346511313] Add tab completion for EXPLAIN .. EXECUTE in psql
Author: Michael Paquier <michael@paquier.xyz>
2021-08-30 [d3fa87657] Add more tab completion support for ALTER TABLE ADD in p
Author: Michael Paquier <michael@paquier.xyz>
2021-08-31 [f2bbadce6] Add tab completion for data types after ALTER TABLE ADD
Author: Fujii Masao <fujii@postgresql.org>
2021-09-01 [b0c066297] Improve tab-completion for CREATE PUBLICATION.
Author: Fujii Masao <fujii@postgresql.org>
2021-10-05 [0b0d277c3] psql: Improve tab-completion for LOCK TABLE.
Author: Michael Paquier <michael@paquier.xyz>
2021-11-05 [a5b336b8b] Improve psql tab completion for COMMENT
Author: Michael Paquier <michael@paquier.xyz>
2021-11-19 [0cd6d3b3c] Improve psql tab completion for transforms, domains and
Author: Michael Paquier <michael@paquier.xyz>
2021-11-29 [f44ceb46e] Improve psql tab completion for views, FDWs, sequences a
Author: Michael Paquier <michael@paquier.xyz>
2021-12-01 [9270778f4] Improve psql tab completion for various DROP commands
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-03 [dfe67c0e8] Tab completion: don't offer valid constraints in VALIDAT
Author: Fujii Masao <fujii@postgresql.org>
2022-01-15 [74527c3e0] Add tab-completion for CREATE FOREIGN TABLE.
Author: Peter Eisentraut <peter@eisentraut.org>
2022-01-27 [fefce9ef9] psql: Add tab completion for ALTER COLLATION / REFRESH V
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
2022-01-28 [95787e849] Tab-complete ALTER PUBLICATION ADD TABLE with list of ta
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-30 [02b8048ba] psql: improve tab-complete's handling of variant SQL nam
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-01 [020258fbd] Treat case of tab-completion keywords a bit more careful
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-02-09 [f0cd9097c] Further tweaks for psql's new tab-completion logic.
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-03-20 [7fa3db367] psql: handle tab completion of timezone names after "SET
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2022-03-25 [2d2232933] Update tab-completion for CREATE PUBLICATION with sequen
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-04-13 [b5607b074] Fix case sensitivity in psql's tab completion for GUC na
-->

      <listitem>
       <para>
        Improve <application>psql</application>'s tab completion (Shinya
        Kato, Dagfinn Ilmari Mannsåker, Peter Smith, Koyu Tanigawa,
        Ken Kato, David Fetter, Haiying Tang, Peter Eisentraut, Álvaro
        Herrera, Tom Lane, Masahiko Sawada)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-16 [cf0cab868] Remove psql support for server versions preceding 9.2.
-->

      <listitem>
       <para>
        Limit support of <application>psql</application>'s backslash
        commands to servers running <productname>PostgreSQL</productname>
        9.2 or later (Tom Lane)
       </para>

       <para>
        Remove code that was only used when running with an older server.
        Commands that do not require any version-specific adjustments
        compared to 9.2 will still work.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title><link linkend="app-pgdump"><application>pg_dump</application></link></title>

     <itemizedlist>

<!--
Author: Noah Misch <noah@leadboat.com>
2021-06-28 [a7a7be1f2] Dump public schema ownership and security labels.
Author: Noah Misch <noah@leadboat.com>
2021-06-28 [7ac10f692] Dump COMMENT ON SCHEMA public.
-->

      <listitem>
       <para>
        Make <application>pg_dump</application> dump
        <literal>public</literal> schema ownership changes and security
        labels (Noah Misch)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-06 [989596152] Avoid per-object queries in performance-critical paths i
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-06 [be85727a3] Use PREPARE/EXECUTE for repetitive per-object queries in
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-31 [d5e8930f5] pg_dump: minor performance improvements from eliminating
-->

      <listitem>
       <para>
        Improve performance of dumping databases with many objects
        (Tom Lane)
       </para>

       <para>
        This will also improve the performance of <link
        linkend="pgupgrade"><application>pg_upgrade</application></link>.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-06 [65aaed22a] Account for TOAST data while scheduling parallel dumps.
-->

      <listitem>
       <para>
        Improve parallel <application>pg_dump</application>'s performance
        for tables with large <acronym>TOAST</acronym> tables (Tom Lane)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-01-17 [215862886] Add support for - -no-table-access-method in pg_{dump,dum
-->

      <listitem>
       <para>
        Add dump/restore option <option>--no-table-access-method</option>
        to force restore to only use the default table access method
        (Justin Pryzby)
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-14 [30e7c175b] Remove pg_dump/pg_dumpall support for dumping from pre-9
-->

      <listitem>
       <para>
        Limit support of <application>pg_dump</application> and <link
        linkend="app-pg-dumpall"><application>pg_dumpall</application></link>
        to servers running <productname>PostgreSQL</productname> 9.2 or
        later (Tom Lane)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

   </sect3>

   <sect3>
    <title>Server Applications</title>

    <itemizedlist>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-20 [3500ccc39] Support base backup targets.
-->

     <listitem>
      <para>
       Add new <link
       linkend="app-pgbasebackup"><application>pg_basebackup</application></link>
       option <option>--target</option> to control the base backup location
       (Robert Haas)
      </para>

      <para>
       The new options are <literal>server</literal> to write the
       backup locally and <literal>blackhole</literal> to discard the
       backup (for testing).
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-02-11 [751b8d23b] pg_basebackup: Allow client-side LZ4 (de)compression.
Author: Robert Haas <rhaas@postgresql.org>
2022-03-08 [7cf085f07] Add support for zstd base backup compression.
Author: Robert Haas <rhaas@postgresql.org>
2022-01-24 [0ad803291] Server-side gzip compression.
-->

     <listitem>
      <para>
       Allow <application>pg_basebackup</application> to do server-side
       gzip, LZ4, and Zstandard compression and client-side LZ4 and
       Zstandard compression of base backup files (Dipesh Pandit,
       Jeevan Ladhe)
      </para>

      <para>
       Client-side <literal>gzip</literal> compression was already
       supported.
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-28 [d45099425] Allow server-side compression to be used with -Fp.
-->

     <listitem>
      <para>
       Allow <application>pg_basebackup</application> to compress on
       the server side and decompress on the client side before storage
       (Dipesh Pandit)
      </para>

      <para>
       This is accomplished by specifying compression on the server side
       and plain output format.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-01-21 [5c649fe15] Extend the options of pg_basebackup to control compressi
Author: Robert Haas <rhaas@postgresql.org>
2022-01-24 [0ad803291] Server-side gzip compression.
Author: Robert Haas <rhaas@postgresql.org>
2022-01-25 [e1f860f13] Tidy up a few cosmetic issues related to pg_basebackup.
-->

     <listitem>
      <para>
       Allow <application>pg_basebackup</application>'s
       <option>--compress</option> option to control the compression
       location (server or client), compression method, and compression
       options (Michael Paquier, Robert Haas)
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-11-05 [babbbb595] Add support for LZ4 compression in pg_receivewal
Author: Michael Paquier <michael@paquier.xyz>
2022-04-13 [042a923ad] Rework compression options of pg_receivewal
-->

     <listitem>
      <para>
       Add the LZ4 compression method to <link
       linkend="app-pgreceivewal"><application>pg_receivewal</application></link>
       (Georgios Kokolatos)
      </para>

      <para>
       This is enabled via <literal>--compress=lz4</literal> and requires
       binaries to be built using <option>--with-lz4</option>.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-11-04 [d62bcc8b0] Rework compression options of pg_receivewal
-->

     <listitem>
      <para>
       Add additional capabilities to
       <application>pg_receivewal</application>'s
       <option>--compress</option> option (Georgios Kokolatos)
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-10-26 [f61e1dd2c] Allow pg_receivewal to stream from a slot's restart LSN
-->

     <listitem>
      <para>
       Improve <application>pg_receivewal</application>'s ability to
       restart at the proper <acronym>WAL</acronym> location (Ronan
       Dunklau)
      </para>

      <para>
       Previously, <application>pg_receivewal</application> would start
       based on the <acronym>WAL</acronym> file stored in the local archive
       directory, or at the sending server's current <acronym>WAL</acronym>
       flush location.  With this change, if the sending server is running
       Postgres 15 or later, the local archive directory is empty, and
       a replication slot is specified, the replication slot's restart
       point will be used.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-04-07 [0d5c38757] Add option - -config-file to pg_rewind
-->

     <listitem>
      <para>
       Add <link
       linkend="app-pgrewind"><application>pg_rewind</application></link>
       option <option>--config-file</option> to simplify use when server
       configuration files are stored outside the data directory (Gunnar
       Bluth)
      </para>
     </listitem>

    </itemizedlist>

    <sect4>
     <title><link linkend="pgupgrade"><application>pg_upgrade</application></link></title>

     <itemizedlist>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-02-06 [38bfae365] pg_upgrade: Move all the files generated internally to a
Author: Michael Paquier <michael@paquier.xyz>
2022-02-15 [a00849630] Fix thinko with subdirectories generated by pg_upgrade f
Author: Michael Paquier <michael@paquier.xyz>
2022-09-13 [f5047c129] Move any remaining files generated by pg_upgrade into an
-->

      <listitem>
       <para>
        Store <application>pg_upgrade</application>'s log and
        temporary files in a subdirectory of the new cluster called
        <filename>pg_upgrade_output.d</filename> (Justin Pryzby)
       </para>

       <para>
        Previously such files were left in the current directory,
        requiring manual cleanup.  Now they are automatically removed on
        successful completion of <application>pg_upgrade</application>.
       </para>
      </listitem>

<!--
Author: Andres Freund <andres@anarazel.de>
2022-02-21 [27b02e070] pg_upgrade: Don't print progress status when output is n
-->

      <listitem>
       <para>
        Disable default status reporting during
        <application>pg_upgrade</application> operation if the output is
        not a terminal (Andres Freund)
       </para>

       <para>
        The status reporting output can be enabled for non-tty usage by
        using <option>--verbose</option>.
       </para>
      </listitem>

<!--
Author: Daniel Gustafsson <dgustafsson@postgresql.org>
2022-03-24 [26ebb0e28] List offending databases in pg_upgrade datallowconn chec
-->

      <listitem>
       <para>
        Make <application>pg_upgrade</application> report all databases
        with invalid connection settings (Jeevan Ladhe)
       </para>

       <para>
        Previously only the first database with an invalid connection
        setting was reported.
       </para>
      </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-17 [9a974cbcb] pg_upgrade: Preserve relfilenodes and tablespace OIDs.
Author: Robert Haas <rhaas@postgresql.org>
2022-01-24 [aa0105141] pg_upgrade: Preserve database OIDs.
Author: Robert Haas <rhaas@postgresql.org>
2022-07-28 [4ab5dae94] Use TRUNCATE to preserve relfilenode for pg_largeobject
-->

      <listitem>
       <para>
        Make <application>pg_upgrade</application> preserve tablespace
        and database OIDs, as well as relation relfilenode numbers
        (Shruthi Gowda, Antonin Houska)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-12-18 [3d5ffccb6] Add option -N/- -no-sync to pg_upgrade
-->

      <listitem>
       <para>
        Add a <option>--no-sync</option> option to
        <application>pg_upgrade</application> (Michael Paquier)
       </para>

       <para>
        This is recommended only for testing.
       </para>
      </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-14 [e469f0aaf] Remove pg_upgrade support for upgrading from pre-9.2 ser
-->

      <listitem>
       <para>
        Limit support of <application>pg_upgrade</application> to old
        servers running <productname>PostgreSQL</productname> 9.2 or later
        (Tom Lane)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title><link linkend="pgwaldump"><application>pg_waldump</application></link></title>

     <itemizedlist>

<!--
Author: Thomas Munro <tmunro@postgresql.org>
2022-03-24 [127aea2a6] Add additional filtering options to pg_waldump.
Author: Thomas Munro <tmunro@postgresql.org>
2022-03-25 [52b556843] Improve command line options for pg_waldump.
-->

      <listitem>
       <para>
        Allow <application>pg_waldump</application> output to be filtered by
        relation file node, block number, fork number, and full page images
        (David Christensen, Thomas Munro)
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-12-02 [f2c52eeba] pg_waldump: Emit stats summary when interrupted by SIGIN
-->

      <listitem>
       <para>
        Make <application>pg_waldump</application> report statistics
        before an interrupted exit (Bharath Rupireddy)
       </para>

       <para>
        For example, issuing a control-C in a terminal running
        <command>pg_waldump --stats --follow</command> will report the
        current statistics before exiting.  This does not work on Windows.
       </para>
      </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-12-13 [c8b733c4c] Improve description of some WAL records with transaction
-->

      <listitem>
       <para>
        Improve descriptions of some transaction <acronym>WAL</acronym>
        records reported by <application>pg_waldump</application>
        (Masahiko Sawada, Michael Paquier)
       </para>
      </listitem>

<!--
Author: Heikki Linnakangas <heikki.linnakangas@iki.fi>
2021-07-01 [c8bf5098c] Allow specifying pg_waldump - -rmgr option multiple times
-->

      <listitem>
       <para>
        Allow <application>pg_waldump</application> to dump information
        about multiple resource managers (Heikki Linnakangas)
       </para>

       <para>
        This is enabled by specifying the <option>--rmgr</option> option
        multiple times.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

   </sect3>

   <sect3>
    <title>Documentation</title>

    <itemizedlist>

<!--
Author: Fujii Masao <fujii@postgresql.org>
2021-10-05 [f6b5d05ba] doc: Document pg_encoding_to_char() and pg_char_to_encod
-->

     <listitem>
      <para>
       Add documentation for <link
       linkend="functions-info-catalog-table"><function>pg_encoding_to_char()</function></link>
       and <function>pg_char_to_encoding()</function> (Ian Lawrence
       Barwick)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-10-09 [2ae5d72f0] Doc: improve documentation for ^@ starts-with operator.
-->

     <listitem>
      <para>
       Document the <link
       linkend="functions-string-other"><literal>^@</literal></link>
       starts-with operator (Tom Lane)
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title>Source Code</title>

    <itemizedlist>

<!--
Author: Andres Freund <andres@anarazel.de>
2021-12-30 [93d973494] ci: Add continuous integration for github repositories v
-->

     <listitem>
      <para>
       Add support for continuous integration testing using cirrus-ci
       (Andres Freund, Thomas Munro, Melanie Plageman)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-02-18 [6c417bbcc] Add support for building with ZSTD.
-->

     <listitem>
      <para>
       Add configure option <link
       linkend="configure-options-features"><option>--with-zstd</option></link>
       to enable Zstandard builds (Jeevan Ladhe, Robert Haas, Michael
       Paquier)
      </para>
     </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2021-11-22 [d6d1dfcc9] Add ABI extra field to fmgr magic block
-->

     <listitem>
      <para>
       Add an ABI identifier field to the magic block in loadable
       libraries, allowing
       non-community <productname>PostgreSQL</productname> distributions
       to identify libraries that are not compatible with other builds
       (Peter Eisentraut)
      </para>

      <para>
       An ABI field mismatch will generate an error at load time.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-12-11 [07eee5a0d] Create a new type category for "internal use" types.
-->

     <listitem>
      <para>
       Create a new <link
       linkend="catalog-pg-type"><structfield>pg_type.typcategory</structfield></link>
       value for <type>"char"</type> (Tom Lane)
      </para>

      <para>
       Some other internal-use-only types have also been assigned to this
       category.
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-18 [cc333f323] Modify pg_basebackup to use a new COPY subprotocol for b
-->

     <listitem>
      <para>
       Add new protocol message <link
       linkend="protocol-replication-base-backup"><literal>TARGET</literal></link>
       to specify a new <command>COPY</command> method to be used for base
       backups (Robert Haas)
      </para>

      <para>
       <link
       linkend="app-pgbasebackup"><application>pg_basebackup</application></link>
       now uses this method.
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-01-24 [0ad803291] Server-side gzip compression.
Author: Robert Haas <rhaas@postgresql.org>
2022-03-23 [ffd53659c] Replace BASE_BACKUP COMPRESSION_LEVEL option with COMPRE
-->

     <listitem>
      <para>
       Add new protocol message <link
       linkend="protocol-replication-base-backup"><literal>COMPRESSION</literal></link>
       and <literal>COMPRESSION_DETAIL</literal> to specify the compression
       method and options (Robert Haas)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-02-10 [9cd28c2e5] Remove server support for old BASE_BACKUP command syntax
Author: Robert Haas <rhaas@postgresql.org>
2022-02-10 [0d4513b61] Remove server support for the previous base backup proto
-->

     <listitem>
      <para>
       Remove server support for old <literal>BASE_BACKUP</literal>
       command syntax and base backup protocol (Robert Haas)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-03-15 [e4ba69f3f] Allow extensions to add new backup targets.
-->

     <listitem>
      <para>
       Add support for extensions to set custom backup targets (Robert
       Haas)
      </para>
     </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2022-04-06 [5c279a6d3] Custom WAL Resource Managers.
-->

     <listitem>
      <para>
       Allow extensions to define custom <acronym>WAL</acronym>
       resource managers (Jeff Davis)
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-01-31 [d10e41d42] Introduce pg_settings_get_flags() to find flags associat
-->

     <listitem>
      <para>
       Add function <link
       linkend="functions-info-catalog-table"><function>pg_settings_get_flags()</function></link>
       to get the flags of server variables (Justin Pryzby)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-04-08 [8ec569479] Apply PGDLLIMPORT markings broadly.
-->

     <listitem>
      <para>
       On Windows, export all the server's global variables using
       <literal>PGDLLIMPORT</literal> markers (Robert Haas)
      </para>

      <para>
       Previously, only specific variables were accessible to extensions
       on Windows.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-07-23 [3b474a2e6] Increase minimum supported GNU make version to 3.81.
-->

     <listitem>
      <para>
       Require GNU <application>make</application> version 3.81 or later
       to build <productname>PostgreSQL</productname> (Tom Lane)
      </para>
     </listitem>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2021-11-05 [db7d1a7b0] pgcrypto: Remove non-OpenSSL support
-->

     <listitem>
      <para>
       Require OpenSSL to build the <link
       linkend="pgcrypto"><application>pgcrypto</application></link>
       extension (Peter Eisentraut)
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-10-07 [92e6a98c3] Adjust configure to insist on Perl version &gt;= 5.8.3.
-->

     <listitem>
      <para>
       Require <application>Perl</application>
       version 5.8.3 or later (Dagfinn Ilmari Mannsåker)
      </para>
     </listitem>

<!--
Author: Andres Freund <andres@anarazel.de>
2022-02-16 [19252e8ec] plpython: Reject Python 2 during build configuration.
-->

     <listitem>
      <para>
       Require <application>Python</application>
       version 3.2 or later (Andres Freund)
      </para>
     </listitem>

    </itemizedlist>

   </sect3>

   <sect3>
    <title>Additional Modules</title>

    <itemizedlist>

<!--
Author: Peter Eisentraut <peter@eisentraut.org>
2021-09-28 [c3b011d99] Support amcheck of sequences
-->

     <listitem>
      <para>
       Allow <link
       linkend="amcheck"><application>amcheck</application></link> to
       check sequences (Mark Dilger)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2021-11-05 [bd807be69] amcheck: Add additional TOAST pointer checks.
-->

     <listitem>
      <para>
       Improve <application>amcheck</application> sanity checks for
       <acronym>TOAST</acronym> tables (Mark Dilger)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-03-15 [c6306db24] Add 'basebackup_to_shell' contrib module.
Author: Robert Haas <rhaas@postgresql.org>
2022-03-30 [26a0c025e] Document basebackup_to_shell.required_role.
-->

     <listitem>
      <para>
       Add new module <application><link
       linkend="basebackup-to-shell">basebackup_to_shell</link></application>
       as an example of a custom backup target (Robert Haas)
      </para>
     </listitem>

<!--
Author: Robert Haas <rhaas@postgresql.org>
2022-02-03 [5ef1eefd7] Allow archiving via loadable modules.
-->

     <listitem>
      <para>
       Add new module <link
       linkend="basic-archive"><application>basic_archive</application></link>
       as an example of performing archiving via a library (Nathan Bossart)
      </para>
     </listitem>

<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2021-11-06 [57e3c5160] Add bool GiST opclass to btree_gist
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2021-11-08 [e2fbb8837] Fix gist_bool_ops to use gbtreekey2
Author: Tomas Vondra <tomas.vondra@postgresql.org>
2021-12-11 [4c6145b51] Add bool to btree_gist documentation
-->

     <listitem>
      <para>
       Allow <link
       linkend="btree-gist"><application>btree_gist</application></link>
       indexes on boolean columns (Emre Hasegeli)
      </para>

      <para>
       These can be used for exclusion constraints.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2021-07-12 [127404fbe] pageinspect: Improve page_header() for pages of 32kB
-->

     <listitem>
      <para>
       Fix <link
       linkend="pageinspect"><application>pageinspect</application></link>'s
       <function>page_header()</function> to handle 32-kilobyte page sizes
       (Quan Zongliang)
      </para>

      <para>
       Previously, improper negative values could be returned in certain
       cases.
      </para>
     </listitem>

<!--
Author: Michael Paquier <michael@paquier.xyz>
2022-04-08 [76cbf7edb] pg_stat_statements: Track I/O timing for temporary file
-->

     <listitem>
      <para>
       Add counters for temporary file block I/O to <link
       linkend="pgstatstatements"><application>pg_stat_statements</application></link>
       (Masahiko Sawada)
      </para>
     </listitem>

<!--
Author: Magnus Hagander <magnus@hagander.net>
2022-04-08 [57d6aea00] Add JIT counters to pg_stat_statements
-->

     <listitem>
      <para>
       Add <acronym>JIT</acronym> counters to pg_stat_statements (Magnus
       Hagander)
      </para>
     </listitem>

<!--
Author: Jeff Davis <jdavis@postgresql.org>
2022-04-08 [2258e76f9] Add contrib/pg_walinspect.
-->

     <listitem>
      <para>
       Add new module <link
       linkend="pgwalinspect"><application>pg_walinspect</application></link>
       (Bharath Rupireddy)
      </para>

      <para>
       This gives <acronym>SQL</acronym>-level output similar to <link
       linkend="pgwaldump"><application>pg_waldump</application></link>.
      </para>
     </listitem>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2022-01-12 [134d97463] Include permissive/enforcing state in sepgsql log messag
-->

     <listitem>
      <para>
       Indicate the permissive/enforcing state in <link
       linkend="sepgsql"><application>sepgsql</application></link> log
       messages (Dave Page)
      </para>
     </listitem>

    </itemizedlist>

    <sect4>
     <title><link linkend="postgres-fdw"><application>postgres_fdw</application></link></title>

     <itemizedlist>

<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
2021-07-30 [5d44fff01] In postgres_fdw, allow CASE expressions to be pushed to
-->

      <listitem>
       <para>
        Allow postgres_fdw to push down <literal>CASE</literal> expressions
        (Alexander Pyhalov)
       </para>
      </listitem>

<!--
Author: Fujii Masao <fujii@postgresql.org>
2021-09-07 [449ab6350] postgres_fdw: Allow application_name of remote connectio
Author: Fujii Masao <fujii@postgresql.org>
2021-12-24 [6e0cb3dec] postgres_fdw: Allow postgres_fdw.application_name to inc
Author: Fujii Masao <fujii@postgresql.org>
2022-02-18 [94c49d534] postgres_fdw: Make postgres_fdw.application_name support
-->

      <listitem>
       <para>
        Add server variable
        <varname>postgres_fdw.application_name</varname> to control the
        application name of postgres_fdw connections (Hayato Kuroda)
       </para>

       <para>
        Previously the remote session's <link
        linkend="guc-application-name"><varname>application_name</varname></link>
        could only be set on the remote server or via a
        <application>postgres_fdw</application> connection specification.
        <varname>postgres_fdw.application_name</varname> supports some
        escape sequences for customization, making it easier to tell such
        connections apart on the remote server.
       </para>
      </listitem>

<!--
Author: Etsuro Fujita <efujita@postgresql.org>
2022-02-24 [04e706d42] postgres_fdw: Add support for parallel commit.
-->

      <listitem>
       <para>
        Allow parallel commit on <application>postgres_fdw</application>
        servers (Etsuro Fujita)
       </para>

       <para>
        This is enabled with the <literal>CREATE SERVER</literal> option
        <literal>parallel_commit</literal>.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

   </sect3>

  </sect2>

  <sect2 id="release-15-acknowledgements">
   <title>Acknowledgments</title>

   <para>
    The following individuals (in alphabetical order) have contributed
    to this release as patch authors, committers, reviewers, testers,
    or reporters of issues.
   </para>

   <simplelist>
    <member>Abhijit Menon-Sen</member>
    <member>Adam Brusselback</member>
    <member>Adam Mackler</member>
    <member>Adrian Ho</member>
    <member>Ahsan Hadi</member>
    <member>Ajin Cherian</member>
    <member>Alastair McKinley</member>
    <member>Aleksander Alekseev</member>
    <member>Ales Zeleny</member>
    <member>Alex Kingsborough</member>
    <member>Alex Kozhemyakin</member>
    <member>Alexander Korotkov</member>
    <member>Alexander Kukushkin</member>
    <member>Alexander Lakhin</member>
    <member>Alexander Nawratil</member>
    <member>Alexander Pyhalov</member>
    <member>Alexey Borzov</member>
    <member>Alexey Ermakov</member>
    <member>Aliaksandr Kalenik</member>
    <member>Álvaro Herrera</member>
    <member>Amit Kapila</member>
    <member>Amit Khandekar</member>
    <member>Amit Langote</member>
    <member>Amul Sul</member>
    <member>Anastasia Lubennikova</member>
    <member>Anders Kaseorg</member>
    <member>Andreas Dijkman</member>
    <member>Andreas Grob</member>
    <member>Andreas Seltenreich</member>
    <member>Andrei Zubkov</member>
    <member>Andres Freund</member>
    <member>Andrew Alsup</member>
    <member>Andrew Bille</member>
    <member>Andrew Dunstan</member>
    <member>Andrew Gierth</member>
    <member>Andrew Kesper</member>
    <member>Andrey Borodin</member>
    <member>Andrey Lepikhov</member>
    <member>Andrey Sokolov</member>
    <member>Andy Fan</member>
    <member>Anton Melnikov</member>
    <member>Anton Voloshin</member>
    <member>Antonin Houska</member>
    <member>Arjan van de Ven</member>
    <member>Arne Roland</member>
    <member>Arthur Zakirov</member>
    <member>Ashutosh Bapat</member>
    <member>Ashutosh Sharma</member>
    <member>Ashwin Agrawal</member>
    <member>Asif Rehman</member>
    <member>Asim Praveen</member>
    <member>Atsushi Torikoshi</member>
    <member>Aya Iwata</member>
    <member>Bauyrzhan Sakhariyev</member>
    <member>Benoit Lobréau</member>
    <member>Bernd Dorn</member>
    <member>Bertrand Drouvot</member>
    <member>Bharath Rupireddy</member>
    <member>Björn Harrtell</member>
    <member>Boris Kolpackov</member>
    <member>Boris Korzun</member>
    <member>Brad Nicholson</member>
    <member>Brar Piening</member>
    <member>Bruce Momjian</member>
    <member>Bruno da Silva</member>
    <member>Bryn Llewellyn</member>
    <member>Carl Sopchak</member>
    <member>Cary Huang</member>
    <member>Chapman Flack</member>
    <member>Chen Jiaoqian</member>
    <member>Chris Bandy</member>
    <member>Chris Lowder</member>
    <member>Christian Quest</member>
    <member>Christoph Berg</member>
    <member>Christoph Heiss</member>
    <member>Christophe Pettus</member>
    <member>Christopher Painter-Wakefield</member>
    <member>Claudio Freire</member>
    <member>Clemens Zeidler</member>
    <member>Corey Huinker</member>
    <member>Dag Lem</member>
    <member>Dagfinn Ilmari Mannsåker</member>
    <member>Dan Kubb</member>
    <member>Daniel Cherniy</member>
    <member>Daniel Gustafsson</member>
    <member>Daniel Polski</member>
    <member>Daniel Vérité</member>
    <member>Daniel Westermann</member>
    <member>Daniele Varrazzo</member>
    <member>Daniil Anisimov</member>
    <member>Danny Shemesh</member>
    <member>Darafei Praliaskouski</member>
    <member>Daria Lepikhova</member>
    <member>Dave Cramer</member>
    <member>Dave Page</member>
    <member>David Christensen</member>
    <member>David Fetter</member>
    <member>David G. Johnston</member>
    <member>David Rowley</member>
    <member>David Steele</member>
    <member>David Zhang</member>
    <member>Dean Rasheed</member>
    <member>Dian Fay</member>
    <member>Dilip Kumar</member>
    <member>Dipesh Pandit</member>
    <member>Dmitry Dolgov</member>
    <member>Dmitry Koval</member>
    <member>Dmitry Marakasov</member>
    <member>Dominique Devienne</member>
    <member>Dong Wook</member>
    <member>Drew DeVault</member>
    <member>Eduard Català</member>
    <member>Egor Chindyaskin</member>
    <member>Egor Rogov</member>
    <member>Ekaterina Kiryanova</member>
    <member>Elena Indrupskaya</member>
    <member>Elvis Pranskevichus</member>
    <member>Emmanuel Quincerot</member>
    <member>Emre Hasegeli</member>
    <member>Eric Mutta</member>
    <member>Erica Zhang</member>
    <member>Erik Rijkers</member>
    <member>Erki Eessaar</member>
    <member>Etsuro Fujita</member>
    <member>Euler Taveira</member>
    <member>Fabien Coelho</member>
    <member>Fabrice Chapuis</member>
    <member>Fabrice Fontaine</member>
    <member>Fabrízio de Royes Mello</member>
    <member>Feike Steenbergen</member>
    <member>Filip Gospodinov</member>
    <member>Florin Irion</member>
    <member>Floris Van Nee</member>
    <member>Frédéric Yhuel</member>
    <member>Gabriela Serventi</member>
    <member>Gaurab Dey</member>
    <member>Geoff Winkless</member>
    <member>Geoffrey Blake</member>
    <member>Georgios Kokolatos</member>
    <member>Gilles Darold</member>
    <member>Greg Nancarrow</member>
    <member>Greg Rychlewski</member>
    <member>Greg Sabino Mullane</member>
    <member>Greg Stark</member>
    <member>Gregory Smith</member>
    <member>Guillaume Lelarge</member>
    <member>Gunnar Bluth</member>
    <member>Gurjeet Singh</member>
    <member>Haiyang Wang</member>
    <member>Haiying Tang</member>
    <member>Hannu Krosing</member>
    <member>Hans Buschmann</member>
    <member>Hayato Kuroda</member>
    <member>Heath Lord</member>
    <member>Heikki Linnakangas</member>
    <member>Herwig Goemans</member>
    <member>Himanshu Upadhyaya</member>
    <member>Holly Roberts</member>
    <member>Hou Zhijie</member>
    <member>Hubert Lubaczewski</member>
    <member>Ian Barwick</member>
    <member>Ian Campbell</member>
    <member>Ibrar Ahmed</member>
    <member>Ildus Kurbangaliev</member>
    <member>Ilya Anfimov</member>
    <member>Itamar Gafni</member>
    <member>Jacob Champion</member>
    <member>Jaime Casanova</member>
    <member>Jakub Wartak</member>
    <member>James Coleman</member>
    <member>James Hilliard</member>
    <member>James Inform</member>
    <member>Jan Piotrowski</member>
    <member>Japin Li</member>
    <member>Jason Harvey</member>
    <member>Jason Kim</member>
    <member>Jean-Christophe Arnu</member>
    <member>Jeevan Ladhe</member>
    <member>Jeff Davis</member>
    <member>Jeff Janes</member>
    <member>Jehan-Guillaume de Rorthais</member>
    <member>Jelte Fennema</member>
    <member>Jeremy Evans</member>
    <member>Jeremy Schneider</member>
    <member>Jian Guo</member>
    <member>Jian He</member>
    <member>Jimmy Yih</member>
    <member>Jiri Fejfar</member>
    <member>Jitka Plesníková</member>
    <member>Joe Conway</member>
    <member>Joe Wildish</member>
    <member>Joel Jacobson</member>
    <member>Joey Bodoia</member>
    <member>John Naylor</member>
    <member>Jonathan Katz</member>
    <member>Josef Simanek</member>
    <member>Joseph Koshakow</member>
    <member>Josh Soref</member>
    <member>Joshua Brindle</member>
    <member>Juan José Santamaría Flecha</member>
    <member>Julien Rouhaud</member>
    <member>Julien Roze</member>
    <member>Junwang Zhao</member>
    <member>Jürgen Purtz</member>
    <member>Justin Pryzby</member>
    <member>Ken Kato</member>
    <member>Kevin Burke</member>
    <member>Kevin Grittner</member>
    <member>Kevin Humphreys</member>
    <member>Kevin McKibbin</member>
    <member>Kevin Sweet</member>
    <member>Kevin Zheng</member>
    <member>Klaudie Willis</member>
    <member>Konstantin Knizhnik</member>
    <member>Konstantina Skovola</member>
    <member>Kosei Masumura</member>
    <member>Kotaro Kawamoto</member>
    <member>Koyu Tanigawa</member>
    <member>Kuntal Ghosh</member>
    <member>Kyotaro Horiguchi</member>
    <member>Lars Kanis</member>
    <member>Lauren Fliksteen</member>
    <member>Laurent Hasson</member>
    <member>Laurenz Albe</member>
    <member>Leslie Lemaire</member>
    <member>Liam Bowen</member>
    <member>Lingjie Qiang</member>
    <member>Liu Huailing</member>
    <member>Louis Jachiet</member>
    <member>Lukas Fittl</member>
    <member>Ma Liangzhu</member>
    <member>Maciek Sakrejda</member>
    <member>Magnus Hagander</member>
    <member>Mahendra Singh Thalor</member>
    <member>Maksim Milyutin</member>
    <member>Marc Bachmann</member>
    <member>Marcin Krupowicz</member>
    <member>Marcus Gartner</member>
    <member>Marek Szuba</member>
    <member>Marina Polyakova</member>
    <member>Mario Emmenlauer</member>
    <member>Mark Dilger</member>
    <member>Mark Murawski</member>
    <member>Mark Wong</member>
    <member>Markus Wanner</member>
    <member>Markus Winand</member>
    <member>Martijn van Oosterhout</member>
    <member>Martin Jurca</member>
    <member>Martin Kalcher</member>
    <member>Martín Marqués</member>
    <member>Masahiko Sawada</member>
    <member>Masahiro Ikeda</member>
    <member>Masao Fujii</member>
    <member>Masaya Kawamoto</member>
    <member>Masayuki Hirose</member>
    <member>Matthias van de Meent</member>
    <member>Matthijs van der Vleuten</member>
    <member>Maxim Orlov</member>
    <member>Maxim Yablokov</member>
    <member>Melanie Plageman</member>
    <member>Michael Banck</member>
    <member>Michael Harris</member>
    <member>Michael J. Sullivan</member>
    <member>Michael Meskes</member>
    <member>Michael Mühlbeyer</member>
    <member>Michael Paquier</member>
    <member>Michael Powers</member>
    <member>Mike Fiedler</member>
    <member>Mike Oh</member>
    <member>Mikhail Kulagin</member>
    <member>Miles Delahunty</member>
    <member>Naoki Okano</member>
    <member>Nathan Bossart</member>
    <member>Nathan Long</member>
    <member>Nazir Bilal Yavuz</member>
    <member>Neha Sharma</member>
    <member>Neil Chen</member>
    <member>Nicola Contu</member>
    <member>Nicolas Lutic</member>
    <member>Nikhil Benesch</member>
    <member>Nikhil Shetty</member>
    <member>Nikhil Sontakke</member>
    <member>Nikita Glukhov</member>
    <member>Nikolai Berkoff</member>
    <member>Nikolay Samokhvalov</member>
    <member>Nikolay Shaplov</member>
    <member>Nitin Jadhav</member>
    <member>Noah Misch</member>
    <member>Noboru Saito</member>
    <member>Noriyoshi Shinoda</member>
    <member>Olaf Bohlen</member>
    <member>Olly Betts</member>
    <member>Onder Kalaci</member>
    <member>Oskar Stenberg</member>
    <member>Otto Kekalainen</member>
    <member>Paul Guo</member>
    <member>Paul Jungwirth</member>
    <member>Paul Martinez</member>
    <member>Pavan Deolasee</member>
    <member>Pavel Borisov</member>
    <member>Pavel Luzanov</member>
    <member>Pavel Stehule</member>
    <member>Peter Eisentraut</member>
    <member>Peter Geoghegan</member>
    <member>Peter Slavov</member>
    <member>Peter Smith</member>
    <member>Petr Jelínek</member>
    <member>Phil Florent</member>
    <member>Phil Krylov</member>
    <member>Pierre-Aurélien Georges</member>
    <member>Prabhat Sahu</member>
    <member>Quan Zongliang</member>
    <member>Rachel Heaton</member>
    <member>Rahila Syed</member>
    <member>Rajakavitha Kodhandapani</member>
    <member>Rajkumar Raghuwanshi</member>
    <member>Ranier Vilela</member>
    <member>Rei Kamigishi</member>
    <member>Reid Thompson</member>
    <member>Rémi Lapeyre</member>
    <member>Renan Soares Lopes</member>
    <member>Richard Guo</member>
    <member>Richard Wesley</member>
    <member>RKN Sai Krishna</member>
    <member>Robert Haas</member>
    <member>Robert Treat</member>
    <member>Roberto Mello</member>
    <member>Robins Tharakan</member>
    <member>Roger Mason</member>
    <member>Roman Zharkov</member>
    <member>Ronan Dunklau</member>
    <member>Rui Zhao</member>
    <member>Ryan Kelly</member>
    <member>Ryo Matsumura</member>
    <member>Ryohei Takahashi</member>
    <member>Sadhuprasad Patro</member>
    <member>Sait Talha Nisanci</member>
    <member>Sami Imseih</member>
    <member>Sandeep Thakkar</member>
    <member>Sebastian Kemper</member>
    <member>Sehrope Sarkuni</member>
    <member>Sergei Kornilov</member>
    <member>Sergei Shoulbakov</member>
    <member>Sergey Shinderuk</member>
    <member>Shay Rojansky</member>
    <member>Shenhao Wang</member>
    <member>Shi Yu</member>
    <member>Shinya Kato</member>
    <member>Shruthi Gowda</member>
    <member>Simon Perepelitsa</member>
    <member>Simon Riggs</member>
    <member>Sirisha Chamarthi</member>
    <member>Soumyadeep Chakraborty</member>
    <member>Stan Hu</member>
    <member>Stas Kelvich</member>
    <member>Stefen Hillman</member>
    <member>Stephen Frost</member>
    <member>Steve Chavez</member>
    <member>Sumanta Mukherjee</member>
    <member>Suraj Khamkar</member>
    <member>Suraj Kharage</member>
    <member>Sven Klemm</member>
    <member>Takamichi Osumi</member>
    <member>Takayuki Tsunakawa</member>
    <member>Takeshi Ideriha</member>
    <member>Tatsuhiro Nakamori</member>
    <member>Tatsuhito Kasahara</member>
    <member>Tatsuo Ishii</member>
    <member>Tatsuro Yamada</member>
    <member>Teja Mupparti</member>
    <member>Teodor Sigaev</member>
    <member>Thibaud Walkowiak</member>
    <member>Thom Brown</member>
    <member>Thomas McKay</member>
    <member>Thomas Munro</member>
    <member>Tim McNamara</member>
    <member>Timo Stolz</member>
    <member>Timur Khanjanov</member>
    <member>Tom Lane</member>
    <member>Tomas Barton</member>
    <member>Tomas Vondra</member>
    <member>Tony Reix</member>
    <member>Troy Frericks</member>
    <member>Tushar Ahuja</member>
    <member>Victor Wagner</member>
    <member>Victor Yegorov</member>
    <member>Vignesh C</member>
    <member>Vik Fearing</member>
    <member>Vincas Dargis</member>
    <member>Vitaly Burovoy</member>
    <member>Vitaly Voronov</member>
    <member>Vladimir Sitnikov</member>
    <member>Wang Ke</member>
    <member>Wei Sun</member>
    <member>Wei Wang</member>
    <member>Whale Song</member>
    <member>Will Mortensen</member>
    <member>Wolfgang Walther</member>
    <member>Yanliang Lei</member>
    <member>Yaoguang Chen</member>
    <member>Yogendra Suralkar</member>
    <member>YoungHwan Joo</member>
    <member>Yugo Nagata</member>
    <member>Yukun Wang</member>
    <member>Yura Sokolov</member>
    <member>Yusuke Egashira</member>
    <member>Yuzuko Hosoya</member>
    <member>Zhang Mingli</member>
    <member>Zhang Wenjie</member>
    <member>Zhihong Yu</member>
    <member>Zhiyong Wu</member>
   </simplelist>
  </sect2>

 </sect1>