<!-- doc/src/sgml/release-9.3.sgml -->

<!-- See header comment in release.sgml about typical markup -->

<sect1 id="release-9-3-23">

<title>Release 9.3.23</title>

<formalpara>

<title>Release date:</title>

<para>2018-05-10</para>

</formalpara>

<para>

This release contains a variety of fixes from 9.3.22.

For information about new features in the 9.3 major release, see

<xref linkend="release-9-3">.

</para>

<sect2>

<title>Migration to Version 9.3.23</title>

<para>

A dump/restore is not required for those running 9.3.X.

</para>

<para>

However, if the function marking mistakes mentioned in the first

changelog entry below affect you, you will want to take steps to

correct your database catalogs.

</para>

<para>

Also, if you are upgrading from a version earlier than 9.3.22,

see <xref linkend="release-9-3-22">.

</para>

</sect2>

<sect2>

<title>Changes</title>

<itemizedlist>

<listitem>

<para>

Fix incorrect volatility markings on a few built-in functions

(Thomas Munro, Tom Lane)

</para>

<para>

The functions

<function>query_to_xml</function>,

<function>cursor_to_xml</function>,

<function>cursor_to_xmlschema</function>,

<function>query_to_xmlschema</function>, and

<function>query_to_xml_and_xmlschema</function>

should be marked volatile because they execute user-supplied queries

that might contain volatile operations. They were not, leading to a

risk of incorrect query optimization. This has been repaired for new

installations by correcting the initial catalog data, but existing

installations will continue to contain the incorrect markings.

Practical use of these functions seems to pose little hazard, but in

case of trouble, it can be fixed by manually updating these

functions' <structname>pg_proc</structname> entries. (Note that that

will need to be done in each database of the installation.) Another

option is to <application>pg_upgrade</application> the database to a

version containing the corrected initial data.

</para>

</listitem>

<listitem>

<para>

Avoid re-using TOAST value OIDs that match dead-but-not-yet-vacuumed

TOAST entries (Pavan Deolasee)

</para>

<para>

Once the OID counter has wrapped around, it's possible to assign a

TOAST value whose OID matches a previously deleted entry in the same

TOAST table. If that entry were not yet vacuumed away, this resulted

in <quote>unexpected chunk number 0 (expected 1) for toast

value <replaceable>nnnnn</replaceable></quote> errors, which would

persist until the dead entry was removed

by <command>VACUUM</command>. Fix by not selecting such OIDs when

creating a new TOAST entry.

</para>

</listitem>

<listitem>

<para>

Change <command>ANALYZE</command>'s algorithm for updating

<structname>pg_class</structname>.<structfield>reltuples</structfield>

(David Gould)

</para>

<para>

Previously, pages not actually scanned by <command>ANALYZE</command>

were assumed to retain their old tuple density. In a large table

where <command>ANALYZE</command> samples only a small fraction of the

pages, this meant that the overall tuple density estimate could not

change very much, so that <structfield>reltuples</structfield> would

change nearly proportionally to changes in the table's physical size

(<structfield>relpages</structfield>) regardless of what was actually

happening in the table. This has been observed to result

in <structfield>reltuples</structfield> becoming so much larger than

reality as to effectively shut off autovacuuming. To fix, assume

that <command>ANALYZE</command>'s sample is a statistically unbiased

sample of the table (as it should be), and just extrapolate the

density observed within those pages to the whole table.

</para>

</listitem>

<listitem>

<para>

Fix <literal>UPDATE/DELETE ... WHERE CURRENT OF</literal> to not fail

when the referenced cursor uses an index-only-scan plan (Yugo Nagata,

Tom Lane)

</para>

</listitem>

<listitem>

<para>

Fix incorrect planning of join clauses pushed into parameterized

paths (Andrew Gierth, Tom Lane)

</para>

<para>

This error could result in misclassifying a condition as

a <quote>join filter</quote> for an outer join when it should be a

plain <quote>filter</quote> condition, leading to incorrect join

output.

</para>

</listitem>

<listitem>

<para>

Fix misoptimization of <literal>CHECK</literal> constraints having

provably-NULL subclauses of

top-level <literal>AND</literal>/<literal>OR</literal> conditions

(Tom Lane, Dean Rasheed)

</para>

<para>

This could, for example, allow constraint exclusion to exclude a

child table that should not be excluded from a query.

</para>

</listitem>

<listitem>

<para>

Avoid failure if a query-cancel or session-termination interrupt

occurs while committing a prepared transaction (Stas Kelvich)

</para>

</listitem>

<listitem>

<para>

Fix query-lifespan memory leakage in repeatedly executed hash joins

(Tom Lane)

</para>

</listitem>

<listitem>

<para>

Fix overly strict sanity check

in <function>heap_prepare_freeze_tuple</function>

(&Aacute;lvaro Herrera)

</para>

<para>

This could result in incorrect <quote>cannot freeze committed

xmax</quote> failures in databases that have

been <application>pg_upgrade</application>'d from 9.2 or earlier.

</para>

</listitem>

<listitem>

<para>

Prevent dangling-pointer dereference when a C-coded before-update row

trigger returns the <quote>old</quote> tuple (Rushabh Lathia)

</para>

</listitem>

<listitem>

<para>

Reduce locking during autovacuum worker scheduling (Jeff Janes)

</para>

<para>

The previous behavior caused drastic loss of potential worker

concurrency in databases with many tables.

</para>

</listitem>

<listitem>

<para>

Ensure client hostname is copied while copying

<structname>pg_stat_activity</structname> data to local memory

(Edmund Horner)

</para>

<para>

Previously the supposedly-local snapshot contained a pointer into

shared memory, allowing the client hostname column to change

unexpectedly if any existing session disconnected.

</para>

</listitem>

<listitem>

<para>

Fix incorrect processing of multiple compound affixes

in <literal>ispell</literal> dictionaries (Arthur Zakirov)

</para>

</listitem>

<listitem>

<para>

Fix collation-aware searches (that is, indexscans using inequality

operators) in SP-GiST indexes on text columns (Tom Lane)

</para>

<para>

Such searches would return the wrong set of rows in most non-C

locales.

</para>

</listitem>

<listitem>

<para>

Count the number of index tuples correctly during initial build of an

SP-GiST index (Tomas Vondra)

</para>

<para>

Previously, the tuple count was reported to be the same as that of

the underlying table, which is wrong if the index is partial.

</para>

</listitem>

<listitem>

<para>

Count the number of index tuples correctly during vacuuming of a

GiST index (Andrey Borodin)

</para>

<para>

Previously it reported the estimated number of heap tuples,

which might be inaccurate, and is certainly wrong if the

index is partial.

</para>

</listitem>

<listitem>

<para>

Allow <function>scalarltsel</function>

and <function>scalargtsel</function> to be used on non-core datatypes

(Tomas Vondra)

</para>

</listitem>

<listitem>

<para>

Reduce <application>libpq</application>'s memory consumption when a

server error is reported after a large amount of query output has

been collected (Tom Lane)

</para>

<para>

Discard the previous output before, not after, processing the error

message. On some platforms, notably Linux, this can make a

difference in the application's subsequent memory footprint.

</para>

</listitem>

<listitem>

<para>

Fix double-free crashes in <application>ecpg</application>

(Patrick Krecker, Jeevan Ladhe)

</para>

</listitem>

<listitem>

<para>

Fix <application>ecpg</application> to handle <type>long long

int</type> variables correctly in MSVC builds (Michael Meskes,

Andrew Gierth)

</para>

</listitem>

<listitem>

<para>

Fix mis-quoting of values for list-valued GUC variables in dumps

(Michael Paquier, Tom Lane)

</para>

<para>

The <varname>local_preload_libraries</varname>,

<varname>session_preload_libraries</varname>,

<varname>shared_preload_libraries</varname>,

and <varname>temp_tablespaces</varname> variables were not correctly

quoted in <application>pg_dump</application> output. This would

cause problems if settings for these variables appeared in

<command>CREATE FUNCTION ... SET</command> or <command>ALTER

DATABASE/ROLE ... SET</command> clauses.

</para>

</listitem>

<listitem>

<para>

Fix overflow handling in <application>PL/pgSQL</application>

integer <command>FOR</command> loops (Tom Lane)

</para>

<para>

The previous coding failed to detect overflow of the loop variable

on some non-gcc compilers, leading to an infinite loop.

</para>

</listitem>

<listitem>

<para>

Adjust <application>PL/Python</application> regression tests to pass

under Python 3.7 (Peter Eisentraut)

</para>

</listitem>

<listitem>

<para>

Support testing <application>PL/Python</application> and related

modules when building with Python 3 and MSVC (Andrew Dunstan)

</para>

</listitem>

<listitem>

<para>

Rename internal <function>b64_encode</function>

and <function>b64_decode</function> functions to avoid conflict with

Solaris 11.4 built-in functions (Rainer Orth)

</para>

</listitem>

<listitem>

<para>

Sync our copy of the timezone library with IANA tzcode release 2018e

(Tom Lane)

</para>

<para>

This fixes the <application>zic</application> timezone data compiler

to cope with negative daylight-savings offsets. While

the <productname>PostgreSQL</productname> project will not

immediately ship such timezone data, <application>zic</application>

might be used with timezone data obtained directly from IANA, so it

seems prudent to update <application>zic</application> now.

</para>

</listitem>

<listitem>

<para>

Update time zone data files to <application>tzdata</application>

release 2018d for DST law changes in Palestine and Antarctica (Casey

Station), plus historical corrections for Portugal and its colonies,

as well as Enderbury, Jamaica, Turks &amp; Caicos Islands, and

Uruguay.

</para>

</listitem>

</itemizedlist>

</sect2>

</sect1>

<sect1 id="release-9-3-22">

<title>Release 9.3.22</title>