rdelaage/postgresql
1
0
Fork 0
mirror of https://git.postgresql.org/git/postgresql.git synced 2024-07-17 19:41:08 +02:00
Code Issues Packages Projects Releases Wiki Activity

First-draft release notes for 13.3.

Browse Source 
As usual, the release notes for older branches will be made by cutting
these down, but put them up for community review first.
This commit is contained in:
Tom Lane 2021-05-07 12:19:30 -04:00
parent ef70b6f8e2
commit 7f4bab7f4a
1 changed files with 979 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

979
doc/src/sgml/release-13.sgml
View File

@ -1,6 +1,985 @@
<!-- doc/src/sgml/release-13.sgml -->
<!-- See header comment in release.sgml about typical markup -->
<sect1 id="release-13-3">
<title>Release 13.3</title>
<formalpara>
<title>Release date:</title>
<para>2021-05-13</para>
</formalpara>
<para>
This release contains a variety of fixes from 13.2.
For information about new features in major release 13, see
<xref linkend="release-13"/>.
</para>
<sect2>
<title>Migration to Version 13.3</title>
<para>
A dump/restore is not required for those running 13.X.
</para>
<para>
However, if you are upgrading from a version earlier than 13.2,
see <xref linkend="release-13-2"/>.
</para>
</sect2>
<sect2>
<title>Changes</title>
<itemizedlist>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: REL_13_STABLE [a71cfc56b] 2021-04-22 11:46:41 -0400
Branch: REL_12_STABLE [3fb93103a] 2021-04-22 11:46:41 -0400
Branch: REL_11_STABLE [27835b547] 2021-04-22 11:46:41 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d479d0028] 2021-04-22 17:30:55 -0400
Branch: REL_13_STABLE [2602ee468] 2021-04-22 17:30:42 -0400
Branch: REL_12_STABLE [05ce4bf8b] 2021-04-22 17:30:42 -0400
-->
<para>
Fix possibly-incorrect computation of <command>UPDATE
... RETURNING</command> outputs for joined cross-partition updates
(Amit Langote, Etsuro Fujita)
</para>
<para>
If an <command>UPDATE</command> for a partitioned table caused a row
to be moved to another partition with a physically different row
type (for example, one with a different set of dropped columns),
computation of <literal>RETURNING</literal> results for that row
could produce errors or wrong answers. No error is observed unless
the <command>UPDATE</command> involves other tables being joined to
the target table.
</para>
</listitem>
<listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [6f70d7ca1] 2021-05-05 12:21:50 -0400
Branch: REL_13_STABLE [923c13520] 2021-05-05 12:14:21 -0400
Branch: REL_12_STABLE [abce26c9e] 2021-05-05 12:14:21 -0400
Branch: REL_11_STABLE [a25b98d2c] 2021-05-05 12:14:21 -0400
-->
<para>
Fix adjustment of constraint deferrability properties in
partitioned tables (&Aacute;lvaro Herrera)
</para>
<para>
When applied to a foreign-key constraint of a partitioned table,
<command>ALTER TABLE ... ALTER CONSTRAINT</command> failed to
adjust the <literal>DEFERRABLE</literal> and/or
<literal>INITIALLY DEFERRED</literal> markings of the constraints
and triggers of leaf partitions. This led to unexpected behavior of
such constraints. After updating to this version, any misbehaving
partitioned tables can be fixed by executing a
new <command>ALTER</command> command to set the desired properties.
</para>
<para>
This change also disallows applying such an <command>ALTER</command>
directly to the constraints of leaf partitions. The only supported
case is for the whole partitioning hierarchy to have identical
constraint properties, so such <command>ALTER</command>s must be
applied at the partition root.
</para>
</listitem>
<listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [a970edbed] 2021-05-04 12:09:08 +0200
Branch: REL_13_STABLE [64190d65f] 2021-05-04 12:10:22 +0200
Branch: REL_12_STABLE [13ff139a2] 2021-05-04 12:10:51 +0200
-->
<para>
When attaching a child table with <command>ALTER TABLE
... INHERIT</command>, insist that any generated columns in the
parent be generated the same way in the child (Peter Eisentraut)
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [f52c5d674] 2021-03-12 11:08:42 -0500
Branch: REL_13_STABLE [8a2297776] 2021-03-12 11:08:42 -0500
Branch: REL_12_STABLE [e7f7950f6] 2021-03-12 11:08:42 -0500
Branch: REL_11_STABLE [6c34f186c] 2021-03-12 11:08:42 -0500
Branch: REL_10_STABLE [e5794cd59] 2021-03-12 11:08:42 -0500
-->
<para>
Forbid marking an identity column as nullable (Vik Fearing)
</para>
<para>
<literal>GENERATED ALWAYS AS IDENTITY</literal> implies <literal>NOT
NULL</literal>, so don't allow it to be combined with an
explicit <literal>NULL</literal> specification.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [69d5ca484] 2021-04-13 15:10:18 -0400
Branch: REL_13_STABLE [c39aa1e87] 2021-04-13 15:10:18 -0400
Branch: REL_12_STABLE [a7fcb6285] 2021-04-13 15:10:18 -0400
Branch: REL_11_STABLE [c7f0275fb] 2021-04-13 15:10:18 -0400
Branch: REL_10_STABLE [46b6635b7] 2021-04-13 15:10:18 -0400
Branch: REL9_6_STABLE [041f4efd2] 2021-04-13 15:10:18 -0400
-->
<para>
Allow <literal>ALTER ROLE/DATABASE ... SET</literal> to set
the <varname>role</varname>, <varname>session_authorization</varname>,
and <varname>temp_buffers</varname> parameters (Tom Lane)
</para>
<para>
Previously, over-eager validity checks might reject these commands,
even if the values would have worked when used later. This created
a command ordering hazard for dump/reload and upgrade scenarios.
</para>
</listitem>
<listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [bd1208098] 2021-02-10 13:06:48 +0900
Branch: REL_13_STABLE [849383138] 2021-02-10 13:09:09 +0900
Branch: REL_12_STABLE [85edb1f26] 2021-02-10 13:09:12 +0900
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [4f4726005] 2021-02-10 16:59:04 +0900
Branch: REL_13_STABLE [c6cd20d91] 2021-02-10 16:59:33 +0900
Branch: REL_12_STABLE [5b2945ec0] 2021-02-10 16:59:43 +0900
-->
<para>
Ensure that <command>REINDEX CONCURRENTLY</command> preserves any
statistics target that's been set for the index (Michael Paquier)
</para>
</listitem>
<listitem>
<!--
Author: Fujii Masao <fujii@postgresql.org>
Branch: master [8a55cb5ba] 2021-02-19 21:57:52 +0900
Branch: REL_13_STABLE [422012c98] 2021-02-19 21:58:43 +0900
Branch: REL_12_STABLE [fadcc4e81] 2021-02-19 21:59:26 +0900
-->
<para>
Fix <command>COMMIT AND CHAIN</command> to work correctly when the
current transaction has live savepoints (Fujii Masao)
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [80ca8464f] 2021-02-25 20:47:32 -0500
Branch: REL_13_STABLE [49076fd3b] 2021-02-25 20:47:32 -0500
-->
<para>
Fix list-manipulation bug in <literal>WITH RECURSIVE</literal>
processing (Michael Paquier, Tom Lane)
</para>
<para>
Sufficiently deep nesting of <literal>WITH</literal> constructs
(at least seven levels) triggered core dumps or incorrect complaints
of faulty <literal>WITH</literal> nesting.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [c402b02b9] 2021-04-12 14:37:49 -0400
Branch: REL_13_STABLE [8a7bd1e6c] 2021-04-12 14:37:22 -0400
Branch: REL_12_STABLE [27011bcff] 2021-04-12 14:37:22 -0400
Branch: REL_11_STABLE [fb2cca828] 2021-04-12 14:37:22 -0400
Branch: REL_10_STABLE [4b0aecee8] 2021-04-12 14:37:22 -0400
Branch: REL9_6_STABLE [a6158a4d9] 2021-04-12 14:37:22 -0400
-->
<para>
Fix bug with coercing the result of a <literal>COLLATE</literal>
expression to a non-collatable type (Tom Lane)
</para>
<para>
This led to a parse tree in which the <literal>COLLATE</literal>
appears to be applied to a non-collatable value. While that
normally has no real impact (since <literal>COLLATE</literal> has no
effect at runtime), it was possible to construct views that would be
rejected during dump/reload.
</para>
</listitem>
<listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [25936fd46] 2021-02-27 18:09:15 -0300
Branch: REL_13_STABLE [2688852a4] 2021-02-27 18:09:15 -0300
Branch: REL_12_STABLE [262eb990c] 2021-02-27 18:09:15 -0300
Branch: REL_11_STABLE [d1c6edd31] 2021-02-27 18:09:15 -0300
-->
<para>
Fix use-after-free bug in saving tuples for <literal>AFTER</literal>
triggers (Amit Langote)
</para>
<para>
This could cause crashes in some situations.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [2efcd502e] 2021-04-30 14:10:26 -0400
Branch: REL_13_STABLE [4d225ba0e] 2021-04-30 14:10:26 -0400
Branch: REL_12_STABLE [89768ee69] 2021-04-30 14:10:26 -0400
Branch: REL_11_STABLE [ebc2ff076] 2021-04-30 14:10:26 -0400
Branch: REL_10_STABLE [0627f3630] 2021-04-30 14:10:26 -0400
Branch: REL9_6_STABLE [73bad52a9] 2021-04-30 14:10:26 -0400
-->
<para>
Disallow calling window functions and procedures via
the <quote>fast path</quote> wire protocol message (Tom Lane)
</para>
<para>
Only plain functions are supported here. While trying to call
an aggregate function failed already, calling a window function
would crash, and calling a procedure would work only if the
procedure did no transaction control.
</para>
</listitem>
<listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [f7aab36d6] 2021-04-28 11:17:58 +0900
Branch: REL_13_STABLE [f3c453787] 2021-04-28 11:18:17 +0900
Branch: REL_12_STABLE [7e76cb1d8] 2021-04-28 11:18:20 +0900
Branch: REL_11_STABLE [5b717e13c] 2021-04-28 11:18:24 +0900
Branch: REL_10_STABLE [90c9bad30] 2021-04-28 11:18:28 +0900
Branch: REL9_6_STABLE [6e41ff056] 2021-04-28 11:18:33 +0900
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [f93f0b5b2] 2021-04-28 11:58:08 +0900
Branch: REL_13_STABLE [a928297bc] 2021-04-28 11:58:43 +0900
Branch: REL_12_STABLE [85a3e0be1] 2021-04-28 11:58:46 +0900
Branch: REL_11_STABLE [669f73666] 2021-04-28 11:58:50 +0900
Branch: REL_10_STABLE [b797918d2] 2021-04-28 11:58:55 +0900
Branch: REL9_6_STABLE [0d05a3a1d] 2021-04-28 11:59:00 +0900
-->
<para>
Extend <function>pg_identify_object_as_address()</function>
to support event triggers (Joel Jacobson)
</para>
</listitem>
<listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [7a3972597] 2021-04-12 11:30:50 +0900
Branch: REL_13_STABLE [be79debd9] 2021-04-12 11:31:26 +0900
Branch: REL_12_STABLE [82dd5706e] 2021-04-12 11:31:30 +0900
Branch: REL_11_STABLE [5656f2c3d] 2021-04-12 11:31:35 +0900
Branch: REL_10_STABLE [1cc110f68] 2021-04-12 11:31:40 +0900
Branch: REL9_6_STABLE [6540322fa] 2021-04-12 11:31:46 +0900
-->
<para>
Fix <function>to_char()</function>'s handling of Roman-numeral month
format codes with negative intervals (Julien Rouhaud)
</para>
<para>
Previously, such cases would usually cause a crash.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [5c06abb9b] 2021-03-08 18:21:51 -0500
Branch: REL_13_STABLE [fe2b5386b] 2021-03-08 18:21:51 -0500
Branch: REL_12_STABLE [8fff8a874] 2021-03-08 18:21:51 -0500
Branch: REL_11_STABLE [ea42ccbce] 2021-03-08 18:21:51 -0500
Branch: REL_10_STABLE [37228ecde] 2021-03-08 18:21:51 -0500
-->
<para>
Check that the argument
of <function>pg_import_system_collations()</function> is a valid
schema OID (Tom Lane)
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [b5a66e735] 2021-02-18 22:38:55 -0500
Branch: REL_13_STABLE [bf9d3a5f8] 2021-02-18 22:38:55 -0500
Branch: REL_12_STABLE [e7cddb5f2] 2021-02-18 22:38:55 -0500
Branch: REL_11_STABLE [6e6fecf49] 2021-02-18 22:38:55 -0500
Branch: REL_10_STABLE [b06450979] 2021-02-18 22:38:55 -0500
Branch: REL9_6_STABLE [ab208421e] 2021-02-18 22:38:55 -0500
-->
<para>
Fix use of uninitialized value while parsing an
<literal>\{<replaceable>m</replaceable>,<replaceable>n</replaceable>\}</literal>
quantifier in a BRE-mode regular expression (Tom Lane)
</para>
<para>
This error could cause the quantifier to act non-greedy, that is
behave like an
<literal>{<replaceable>m</replaceable>,<replaceable>n</replaceable>}?</literal>
quantifier would do in full regular expressions.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [375398244] 2021-04-20 11:32:02 -0400
Branch: REL_13_STABLE [7bfba4f19] 2021-04-20 11:32:02 -0400
-->
<para>
Fix <quote>could not find pathkey item to sort</quote> planner
errors in some situations where the sort key involves an aggregate
or window function (James Coleman, Tom Lane)
</para>
</listitem>
<listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [33e52ad9a] 2021-03-26 22:34:58 +0100
Branch: REL_13_STABLE [67251c82a] 2021-03-26 22:37:45 +0100
Branch: REL_12_STABLE [88acb0135] 2021-03-26 22:40:00 +0100
Branch: REL_11_STABLE [2a307df1b] 2021-03-26 22:44:01 +0100
Branch: REL_10_STABLE [e5eb40eed] 2021-03-26 22:46:15 +0100
-->
<para>
Don't ignore system columns when estimating the number of groups
using extended statistics (Tomas Vondra)
</para>
<para>
This led to strange estimates for queries such as <literal>SELECT
... GROUP BY a, b, ctid</literal>.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [ae4867ec7] 2021-02-12 16:26:47 -0500
Branch: REL_13_STABLE [3a02d68a9] 2021-02-12 16:26:47 -0500
Branch: REL_12_STABLE [0347470b3] 2021-02-12 16:26:47 -0500
Branch: REL_11_STABLE [a28df6fa3] 2021-02-12 16:26:47 -0500
Branch: REL_10_STABLE [374f1cefe] 2021-02-12 16:26:47 -0500
Branch: REL9_6_STABLE [f4d781dae] 2021-02-12 16:26:47 -0500
-->
<para>
Avoid divide-by-zero when estimating selectivity of a regular
expression with a very long fixed prefix (Tom Lane)
</para>
<para>
This typically led to a <literal>NaN</literal> selectivity value,
causing assertion failures or strange planner behavior.
</para>
</listitem>
<listitem>
<!--
Author: Tomas Vondra <tomas.vondra@postgresql.org>
Branch: master [23607a815] 2021-04-07 15:58:36 +0200
Branch: REL_13_STABLE [1aad1d181] 2021-04-07 15:59:30 +0200
Branch: REL_12_STABLE [847612fda] 2021-04-07 16:01:27 +0200
Branch: REL_11_STABLE [7d3a53461] 2021-04-07 16:02:51 +0200
Branch: REL_10_STABLE [e4f251be7] 2021-04-07 16:04:05 +0200
Branch: REL9_6_STABLE [df97ee6ea] 2021-04-07 16:05:21 +0200
-->
<para>
Fix access-off-the-end-of-the-table error in BRIN index bitmap scans
(Tomas Vondra)
</para>
<para>
If the page range size used by a BRIN index isn't a power of two,
there were corner cases in which a bitmap scan could try to fetch
pages past the actual end of the table, leading to <quote>could not
open file</quote> errors.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [38bb3aef3] 2021-02-16 12:07:14 -0500
Branch: REL_13_STABLE [0d779d22a] 2021-02-16 12:07:14 -0500
-->
<para>
Fix potentially wrong answers from GIN <type>tsvector</type> index
searches, when there are many matching tuples (Tom Lane)
</para>
<para>
If the number of index matches became large enough to make the
bitmap holding them become lossy (a threshold that depends
on <varname>work_mem</varname>), the code could get confused about
whether rechecks are required, allowing rows to be returned that
don't actually match the query.
</para>
</listitem>
<listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [909b449e0] 2021-03-22 14:02:26 +0900
Branch: REL_13_STABLE [78c24e97d] 2021-03-22 14:02:36 +0900
-->
<para>
Fix concurrency issues with WAL segment recycling on Windows
(Michael Paquier)
</para>
<para>
This reverts a change that caused intermittent <quote>could not
rename file ...: Permission denied</quote> log messages. While
there were not serious consequences, the log spam was annoying.
</para>
</listitem>
<listitem>
<!--
Author: Michael Paquier <michael@paquier.xyz>
Branch: master [595b9cba2] 2021-03-22 08:30:53 +0900
Branch: REL_13_STABLE [6e5ce888a] 2021-03-22 08:31:01 +0900
Branch: REL_12_STABLE [4b1dd9b1e] 2021-03-22 08:31:05 +0900
Branch: REL_11_STABLE [f1d550f18] 2021-03-22 08:31:09 +0900
Branch: REL_10_STABLE [1ec7162a8] 2021-03-22 08:31:14 +0900
-->
<para>
Avoid incorrect timeline change while recovering uncommitted
two-phase transactions from WAL (Soumyadeep Chakraborty, Jimmy Yih,
Kevin Yeap)
</para>
<para>
This error could lead to subsequent WAL records being written under
the wrong timeline ID, leading to consistency problems, or even
complete failure to be able to restart the server, later on.
</para>
</listitem>
<listitem>
<!--
Author: Fujii Masao <fujii@postgresql.org>
Branch: master [ad8b67492] 2021-04-06 02:25:37 +0900
Branch: REL_13_STABLE [e7bcfd717] 2021-04-06 02:27:11 +0900
Branch: REL_12_STABLE [dd5b7e486] 2021-04-06 02:27:22 +0900
Branch: REL_11_STABLE [b59d3abaf] 2021-04-06 02:27:30 +0900
Branch: REL_10_STABLE [b9cf9d7d3] 2021-04-06 02:27:39 +0900
Branch: REL9_6_STABLE [e3bf96216] 2021-04-06 02:27:48 +0900
-->
<para>
Ensure that locks are released while shutting down a standby
server's startup process (Fujii Masao)
</para>
<para>
When a standby server is shut down while still in recovery, some
locks might be left held. This causes assertion failures in debug
builds; it's unclear whether any serious consequence could occur
in production builds.
</para>
</listitem>
<listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [f900a79ec] 2021-02-15 16:04:59 +1300
Branch: REL_13_STABLE [6c23e5ae9] 2021-02-15 16:05:42 +1300
Branch: REL_12_STABLE [a27f3a7f4] 2021-02-15 16:06:18 +1300
Branch: REL_11_STABLE [1fefe8879] 2021-02-15 16:06:25 +1300
Branch: REL_10_STABLE [800131df7] 2021-02-15 16:08:24 +1300
Branch: REL9_6_STABLE [09a3b19e3] 2021-02-15 16:09:14 +1300
-->
<para>
Ensure we default to <varname>wal_sync_method</varname>
= <literal>fdatasync</literal> on recent FreeBSD (Thomas Munro)
</para>
<para>
FreeBSD 13 supports <literal>open_datasync</literal>, which would
normally become the default choice. However, it's unclear whether
that is actually an improvement for Postgres, so preserve the
existing default for now.
</para>
</listitem>
<listitem>
<!--
Author: Peter Geoghegan <pg@bowt.ie>
Branch: master [9f3665fbf] 2021-03-10 16:27:01 -0800
Branch: REL_13_STABLE [9663d1244] 2021-03-10 16:26:58 -0800
Author: Peter Geoghegan <pg@bowt.ie>
Branch: master [5f8727f5a] 2021-03-10 17:07:57 -0800
Branch: REL_13_STABLE [1fc5a5738] 2021-03-10 17:07:55 -0800
-->
<para>
Disable the <varname>vacuum_cleanup_index_scale_factor</varname>
parameter and storage option (Peter Geoghegan)
</para>
<para>
The notion of tracking <quote>stale</quote> index statistics proved
to interact badly with
the <varname>autovacuum_vacuum_insert_threshold</varname> parameter,
resulting in unnecessary full-index scans and consequent degradation
of autovacuum performance. The latter mechanism seems superior, so
remove the stale-statistics logic. The control parameter for that,
<varname>vacuum_cleanup_index_scale_factor</varname>, will be
removed entirely in v14. In v13, it remains present to avoid
breaking existing configuration files, but it no longer does
anything.
</para>
</listitem>
<listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [e798d095d] 2021-05-04 10:09:12 -0400
Branch: REL_13_STABLE [91a6b3862] 2021-05-04 10:09:11 -0400
Branch: REL_12_STABLE [72fabd4bf] 2021-05-04 10:09:11 -0400
-->
<para>
Pass the correct trigger OID to object post-alter hooks
during <command>ALTER CONSTRAINT</command> (&Aacute;lvaro Herrera)
</para>
<para>
When updating trigger properties during <command>ALTER
CONSTRAINT</command>, the post-alter hook was told that we are
updating a trigger, but the constraint's OID was passed instead of
the trigger's.
</para>
</listitem>
<listitem>
<!--
Author: Thomas Munro <tmunro@postgresql.org>
Branch: master [637668fb1] 2021-02-15 14:27:33 +1300
Branch: REL_13_STABLE [9fe40913c] 2021-02-15 14:20:42 +1300
Branch: REL_12_STABLE [840eda04e] 2021-02-15 14:20:52 +1300
Branch: REL_11_STABLE [acafdd9ed] 2021-02-15 14:21:01 +1300
Branch: REL_10_STABLE [4b426f77c] 2021-02-15 14:22:08 +1300
Branch: REL9_6_STABLE [8a07e5bd8] 2021-02-15 14:22:19 +1300
-->
<para>
Ensure we finish cleaning up when interrupted while detaching a DSM
segment (Thomas Munro)
</para>
<para>
This error could result in temporary files not being cleaned up
promptly after a parallel query.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: REL_12_STABLE [992cba94d] 2021-03-20 12:38:22 -0400
Branch: REL_11_STABLE [c49e287a8] 2021-03-20 12:38:22 -0400
Branch: REL_10_STABLE [7d9629ed2] 2021-03-20 12:38:22 -0400
-->
<para>
Fix memory leak while initializing server's SSL parameters (Michael
Paquier)
</para>
<para>
This is ordinarily insignificant, but if the postmaster is
repeatedly sent <systemitem>SIGHUP</systemitem> signals, the leak
can build up over time.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [e835e89a0] 2021-03-20 12:47:21 -0400
Branch: REL_13_STABLE [4b41f6923] 2021-03-20 12:47:35 -0400
Branch: REL_12_STABLE [2f31414f4] 2021-03-20 12:47:42 -0400
Branch: REL_11_STABLE [8c830148d] 2021-03-20 12:47:48 -0400
Branch: REL_10_STABLE [ad6c19066] 2021-03-20 12:47:53 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [9bacdf9f5] 2021-03-18 22:22:47 -0400
Branch: REL_13_STABLE [12354839e] 2021-03-18 22:21:58 -0400
Branch: REL_12_STABLE [4eca51d44] 2021-03-18 22:21:58 -0400
Branch: REL_11_STABLE [20f11ca0d] 2021-03-18 22:21:58 -0400
Branch: REL_10_STABLE [ba986b7bc] 2021-03-18 22:21:58 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [377b7a830] 2021-03-18 22:22:47 -0400
Branch: REL_13_STABLE [642b0b69b] 2021-03-18 22:09:41 -0400
Branch: REL_12_STABLE [fc552f868] 2021-03-18 22:09:41 -0400
Branch: REL_11_STABLE [26a3ae06d] 2021-03-18 22:09:41 -0400
Branch: REL_10_STABLE [5058e95a6] 2021-03-18 22:09:41 -0400
Branch: REL9_6_STABLE [7e2521770] 2021-03-18 22:09:41 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [d303849b0] 2021-03-18 22:22:47 -0400
Branch: REL_13_STABLE [eba939551] 2021-03-18 21:44:43 -0400
Branch: REL_12_STABLE [92bc14a10] 2021-03-18 21:44:43 -0400
Branch: REL_11_STABLE [099d2914f] 2021-03-18 21:44:43 -0400
Branch: REL_10_STABLE [0b618ddf8] 2021-03-18 21:44:43 -0400
Branch: REL9_6_STABLE [09e961929] 2021-03-18 21:44:43 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [415ffdc22] 2021-03-18 22:22:47 -0400
Branch: REL_13_STABLE [ea3989f34] 2021-03-18 20:50:56 -0400
Branch: REL_12_STABLE [1452a0bb8] 2021-03-18 20:50:56 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [28644fac1] 2021-03-18 22:22:47 -0400
Branch: REL_13_STABLE [536836970] 2021-03-18 20:37:09 -0400
Branch: REL_12_STABLE [fbcc9fe33] 2021-03-18 20:37:09 -0400
Branch: REL_11_STABLE [967b693ea] 2021-03-18 20:37:09 -0400
Branch: REL_10_STABLE [2bed650c4] 2021-03-18 20:37:09 -0400
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [4b12ab18c] 2021-03-16 16:03:06 -0400
Branch: REL_13_STABLE [4d072bf2a] 2021-03-16 16:02:49 -0400
Branch: REL_12_STABLE [6e3475556] 2021-03-16 16:02:49 -0400
Branch: REL_11_STABLE [a2764d87d] 2021-03-16 16:02:50 -0400
Branch: REL_10_STABLE [7ce7f2b79] 2021-03-16 16:02:50 -0400
-->
<para>
Fix assorted minor memory leaks in the server (Tom Lane, Andres Freund)
</para>
</listitem>
<listitem>
<!--
Author: Fujii Masao <fujii@postgresql.org>
Branch: master [614b7f18b] 2021-02-18 23:28:15 +0900
Branch: REL_13_STABLE [d4b667e93] 2021-02-18 23:28:58 +0900
-->
<para>
Fix uninitialized variable in walreceiver's statistics in shared
memory (Fujii Masao)
</para>
<para>
This error was harmless on most platforms, but could cause issues
on platforms lacking atomic variables and/or spinlock support.
</para>
</listitem>
<listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [b94409a02] 2021-05-03 12:18:27 +0200
Branch: REL_13_STABLE [e48ce7ef0] 2021-05-03 21:01:09 +0200
-->
<para>
Reduce the overhead of dtrace probes for LWLock operations, when
dtrace support is compiled in but not active (Peter Eisentraut)
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [c2db458c1] 2021-04-13 13:37:07 -0400
Branch: REL_13_STABLE [97b7ad468] 2021-04-13 13:37:07 -0400
Branch: REL_12_STABLE [6530df6c2] 2021-04-13 13:37:07 -0400
Branch: REL_11_STABLE [22f2a98cf] 2021-04-13 13:37:07 -0400
-->
<para>
Fix failure when a PL/pgSQL <command>DO</command> block makes use of
both composite-type variables and transaction control (Tom Lane)
</para>
<para>
Previously, such cases led to errors about leaked tuple descriptors.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [51c54bb60] 2021-03-11 14:43:45 -0500
Branch: REL_13_STABLE [3580b4a0c] 2021-03-11 14:43:45 -0500
Branch: REL_12_STABLE [56defbdd0] 2021-03-11 14:43:45 -0500
Branch: REL_11_STABLE [d3a557894] 2021-03-11 14:43:45 -0500
Branch: REL_10_STABLE [d2be6cdc5] 2021-03-11 14:43:45 -0500
Branch: REL9_6_STABLE [a98e53e10] 2021-03-11 14:43:45 -0500
-->
<para>
Prevent infinite loop in <application>libpq</application>
if a ParameterDescription message with a corrupt length is received
(Tom Lane)
</para>
</listitem>
<listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [75dbfe4ca] 2021-03-02 15:39:34 -0300
Branch: REL_13_STABLE [b52fd1e7c] 2021-03-02 15:39:34 -0300
Branch: REL_12_STABLE [f92776791] 2021-03-02 15:39:34 -0300
Branch: REL_11_STABLE [ae1c1d84e] 2021-03-02 15:39:34 -0300
Branch: REL_10_STABLE [926139dd0] 2021-03-02 15:39:34 -0300
-->
<para>
When <application>initdb</application> prints instructions about how
to start the server, make the path shown
for <application>pg_ctl</application> use backslash separators on
Windows (Nitin Jadhav)
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [ea8013854] 2021-03-23 14:27:50 -0400
Branch: REL_13_STABLE [c6eac71a8] 2021-03-23 14:27:50 -0400
Branch: REL_12_STABLE [beac4bd65] 2021-03-23 14:27:50 -0400
Branch: REL_11_STABLE [d75edab42] 2021-03-23 14:27:50 -0400
Branch: REL_10_STABLE [d5a905ed5] 2021-03-23 14:27:50 -0400
Branch: REL9_6_STABLE [4f670c64e] 2021-03-23 14:27:50 -0400
-->
<para>
Fix <application>psql</application> to restore the previous behavior
of <literal>\connect
service=<replaceable>something</replaceable></literal> (Tom Lane)
</para>
<para>
A previous bug fix caused environment variables (such
as <varname>PGPORT</varname>) to override entries in the service
file in this context. Previously, and in other contexts, the
priority is the other way around; so restore that behavior.
</para>
</listitem>
<listitem>
<!--
Author: Fujii Masao <fujii@postgresql.org>
Branch: master [fe06819f1] 2021-02-19 22:01:25 +0900
Branch: REL_13_STABLE [be7485a1e] 2021-02-19 22:02:19 +0900
Branch: REL_12_STABLE [67b3ee292] 2021-02-19 22:02:44 +0900
-->
<para>
Fix <application>psql</application>'s <literal>ON_ERROR_ROLLBACK</literal>
feature to handle <literal>COMMIT AND CHAIN</literal> commands
correctly (Arthur Nascimento)
</para>
<para>
Previously, this case failed with <quote>savepoint
"pg_psql_temporary_savepoint" does not exist</quote>.
</para>
</listitem>
<listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [8d645a116] 2021-03-29 18:34:39 -0300
Branch: REL_13_STABLE [f50dc2c72] 2021-03-29 18:34:39 -0300
-->
<para>
In <application>psql</application>, avoid repeated <quote>could not
print result table</quote> failures after the first such error
(&Aacute;lvaro Herrera)
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [48d67fd89] 2021-03-12 12:20:15 -0500
Branch: REL_13_STABLE [6ed059933] 2021-03-12 12:20:15 -0500
Branch: REL_12_STABLE [60e5c2b14] 2021-03-12 12:20:15 -0500
Branch: REL_11_STABLE [33aa7d13d] 2021-03-12 12:20:15 -0500
Branch: REL_10_STABLE [8915e7906] 2021-03-12 12:20:15 -0500
Branch: REL9_6_STABLE [a42c4438b] 2021-03-12 12:20:15 -0500
-->
<para>
Fix race condition in detection of file modification by
<application>psql</application>'s <literal>\e</literal> and related
commands (Laurenz Albe)
</para>
<para>
A very fast typist could fool the code's file-timestamp-based
detection of whether the temporary edit file was changed.
</para>
</listitem>
<listitem>
<!--
Author: Peter Eisentraut <peter@eisentraut.org>
Branch: master [feb270d10] 2021-05-04 14:18:16 +0200
Branch: REL_13_STABLE [a6a3a274a] 2021-05-04 14:18:23 +0200
Branch: REL_12_STABLE [bdaa0d2d7] 2021-05-04 14:18:33 +0200
-->
<para>
Fix <application>pg_dump</application>'s dumping of
generated columns in partitioned tables (Peter Eisentraut)
</para>
<para>
A fix introduced in the previous minor release should not be applied
to partitioned tables, only traditionally-inherited tables.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [ec03f2df1] 2021-04-01 13:34:16 -0400
Branch: REL_13_STABLE [35421a470] 2021-04-01 13:34:16 -0400
Branch: REL_12_STABLE [84c5e0dd2] 2021-04-01 13:34:16 -0400
Branch: REL_11_STABLE [25fe401d5] 2021-04-01 13:34:16 -0400
Branch: REL_10_STABLE [1b6961c8f] 2021-04-01 13:34:16 -0400
Branch: REL9_6_STABLE [2c9b857af] 2021-04-01 13:34:16 -0400
-->
<para>
Fix missed file version check
in <application>pg_restore</application> (Tom Lane)
</para>
<para>
When reading a custom-format archive from a non-seekable source,
<application>pg_restore</application> neglected to check the
archive version. If it was fed a newer archive version than it
can support, it would fail messily later on.
</para>
</listitem>
<listitem>
<!--
Author: Tom Lane <tgl@sss.pgh.pa.us>
Branch: master [57c081de0] 2021-04-29 15:24:37 -0400
Branch: REL_13_STABLE [bbcfee0e5] 2021-04-29 15:24:37 -0400
Branch: REL_12_STABLE [ba86371b9] 2021-04-29 15:24:37 -0400
Branch: REL_11_STABLE [404946d40] 2021-04-29 15:24:37 -0400
Branch: REL_10_STABLE [d5722c927] 2021-04-29 15:24:37 -0400
Branch: REL9_6_STABLE [54a233071] 2021-04-29 15:24:38 -0400
-->
<para>
Add some more checks to <application>pg_upgrade</application> for
user tables containing non-upgradable data types (Tom Lane)
</para>
<para>
Fix detection of some cases where a non-upgradable data type is
embedded within a container type (such as an array or range).
Also disallow upgrading when user tables contain columns of
system-defined composite types, since those types' OIDs are not
stable across versions.
</para>
</listitem>
<listitem>
<!--
Author: Fujii Masao <fujii@postgresql.org>
Branch: master [2eb1fc8b1] 2021-04-03 00:07:00 +0900
Branch: REL_13_STABLE [104164361] 2021-04-03 00:07:49 +0900
Branch: REL_12_STABLE [24bcce3bd] 2021-04-03 00:08:05 +0900
-->
<para>
Fix incorrect progress-reporting calculation
in <application>pg_checksums</application> (Shinya Kato)
</para>
</listitem>
<listitem>
<!--
Author: Fujii Masao <fujii@postgresql.org>
Branch: master [51893c846] 2021-03-23 09:53:08 +0900
Branch: REL_13_STABLE [34279fd4f] 2021-03-23 09:54:38 +0900
Branch: REL_12_STABLE [4220e06f4] 2021-03-23 09:54:59 +0900
Branch: REL_11_STABLE [4a2627c41] 2021-03-23 09:55:21 +0900
Branch: REL_10_STABLE [5386a8506] 2021-03-23 09:55:33 +0900
Branch: REL9_6_STABLE [e73068b07] 2021-03-23 09:55:49 +0900
-->
<para>
Fix <application>pg_waldump</application> to
count <literal>XACT</literal> records correctly when generating
per-record statistics (Kyotaro Horiguchi)
</para>
</listitem>
<listitem>
<!--
Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
Branch: master [8deb6b38d] 2021-02-23 17:30:21 -0300
Branch: REL_13_STABLE [28f4b6108] 2021-02-23 17:30:21 -0300
Branch: REL_12_STABLE [2796ae2ad] 2021-02-23 17:30:21 -0300
Branch: REL_11_STABLE [258391707] 2021-02-23 17:30:21 -0300
Branch: REL_10_STABLE [bf518fefa] 2021-02-23 17:30:21 -0300
Branch: REL9_6_STABLE [0ccebe779] 2021-02-23 17:30:21 -0300
-->
<para>
Fix <filename>contrib/amcheck</filename> to not complain about the
tuple flags <literal>HEAP_XMAX_LOCK_ONLY</literal>
and <literal>HEAP_KEYS_UPDATED</literal> both being set
(Julien Rouhaud)
</para>
<para>
This is a valid state after <literal>SELECT FOR UPDATE</literal>.
</para>
</listitem>
<listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [455dbc010] 2021-04-12 19:24:41 -0700
Branch: REL_13_STABLE [13881199e] 2021-04-12 19:24:58 -0700
Branch: REL_12_STABLE [4749c7ff4] 2021-04-12 19:27:09 -0700
Branch: REL_11_STABLE [c8da16ba1] 2021-04-12 19:27:31 -0700
Branch: REL_10_STABLE [fb9812b6c] 2021-04-12 19:27:36 -0700
Branch: REL9_6_STABLE [f488d19f3] 2021-04-12 19:27:38 -0700
-->
<para>
Adjust VPATH build rules to support recent Oracle Developer Studio
compiler versions (Noah Misch)
</para>
</listitem>
<listitem>
<!--
Author: Noah Misch <noah@leadboat.com>
Branch: master [c3556f6fa] 2021-04-12 19:24:21 -0700
Branch: REL_13_STABLE [766c8fce5] 2021-04-12 19:24:24 -0700
Branch: REL_12_STABLE [e3ac7d176] 2021-04-12 19:24:25 -0700
Branch: REL_11_STABLE [4a3de4092] 2021-04-12 19:24:25 -0700
Branch: REL_10_STABLE [72a9bd047] 2021-04-12 19:24:25 -0700
Branch: REL9_6_STABLE [14652a19f] 2021-04-12 19:24:26 -0700
-->
<para>
Fix testing of PL/Python for Python 3 on Solaris (Noah Misch)
</para>
</listitem>
</itemizedlist>
</sect2>
</sect1>
<sect1 id="release-13-2">
<title>Release 13.2</title>