<sect1 id="release-9-4-22">

<title>Release 9.4.22</title>

<formalpara>

<title>Release date:</title>

<para>2019-05-09</para>

</formalpara>

<para>

This release contains a variety of fixes from 9.4.21.

For information about new features in the 9.4 major release, see

<xref linkend="release-9-4">.

</para>

<para>

The <productname>PostgreSQL</productname> community will stop

releasing updates for the 9.4.X release series in February 2020.

Users are encouraged to update to a newer release branch soon.

</para>

<sect2>

<title>Migration to Version 9.4.22</title>

<para>

A dump/restore is not required for those running 9.4.X.

</para>

<para>

However, if you are upgrading from a version earlier than 9.4.18,

see <xref linkend="release-9-4-18">.

</para>

</sect2>

<sect2>

<title>Changes</title>

<itemizedlist>

<listitem>

<para>

Fix behavior for an <command>UPDATE</command>

or <command>DELETE</command> on an inheritance tree or partitioned

table in which every table can be excluded (Amit Langote, Tom Lane)

</para>

<para>

In such cases, the query did not report the correct set of output

columns when a <literal>RETURNING</literal> clause was present, and

if there were any statement-level triggers that should be fired, it

didn't fire them.

</para>

</listitem>

<listitem>

<para>

Fix handling of explicit <literal>DEFAULT</literal> items in

an <command>INSERT ... VALUES</command> command with

multiple <literal>VALUES</literal> rows, if the target relation is

an updatable view (Amit Langote, Dean Rasheed)

</para>

<para>

When the updatable view has no default for the column but its

underlying table has one, a single-row <command>INSERT

... VALUES</command> will use the underlying table's default.

In the multi-row case, however, NULL was always used. Correct it to

act like the single-row case.

</para>

</listitem>

<listitem>

<para>

Fix <command>CREATE VIEW</command> to allow zero-column views

(Ashutosh Sharma)

</para>

<para>

We should allow this for consistency with allowing zero-column

tables. Since a table can be converted to a view, zero-column views

could be created even with the restriction in place, leading to

dump/reload failures.

</para>

</listitem>

<listitem>

<para>

Accept XML documents as valid values of type <type>xml</type>

when <varname>xmloption</varname> is set

to <literal>content</literal>, as required by SQL:2006 and later

(Chapman Flack)

</para>

<para>

Previously <productname>PostgreSQL</productname> followed the

SQL:2003 definition, which doesn't allow this. But that creates a

serious problem for dump/restore: there is no setting

of <varname>xmloption</varname> that will accept all valid XML data.

Hence, switch to the 2006 definition.

</para>

<para>

<application>pg_dump</application> is also modified to emit

<literal>SET xmloption = content</literal> while restoring data,

ensuring that dump/restore works even if the prevailing

setting is <literal>document</literal>.

</para>

</listitem>

<listitem>

<para>

Improve server's startup-time checks for whether a pre-existing

shared memory segment is still in use (Noah Misch)

</para>

<para>

The postmaster is now more likely to detect that there are still

active processes from a previous postmaster incarnation, even if

the <filename>postmaster.pid</filename> file has been removed.

</para>

</listitem>

<listitem>

<para>

Fix incompatibility of GIN-index WAL records (Alexander Korotkov)

</para>

<para>

A fix applied in February's minor releases was not sufficiently

careful about backwards compatibility, leading to problems if a

standby server of that vintage reads GIN page-deletion WAL records

generated by a primary server of a previous minor release.

</para>

</listitem>

<listitem>

<para>

Tolerate <literal>EINVAL</literal> and <literal>ENOSYS</literal>

error results, where appropriate, for <function>fsync</function>

and <function>sync_file_range</function> calls

(Thomas Munro, James Sewell)

</para>

<para>

The previous change to panic on file synchronization failures turns

out to have been excessively paranoid for certain cases where a

failure is predictable and essentially means <quote>operation not

supported</quote>.

</para>

</listitem>

<listitem>

<para>

Fix <quote>failed to build any <replaceable>N</replaceable>-way

joins</quote> planner failures with lateral references leading out

of <literal>FULL</literal> outer joins (Tom Lane)

</para>

</listitem>

<listitem>

<para>

Avoid O(N^2) performance issue when rolling back a transaction that

created many tables (Tomas Vondra)

</para>

</listitem>

<listitem>

<para>

Fix race conditions in management of dynamic shared memory

(Thomas Munro)

</para>

<para>

These could lead to <quote>dsa_area could not attach to

segment</quote> or <quote>cannot unpin a segment that is not

pinned</quote> errors.

</para>

</listitem>

<listitem>

<para>

Fix race condition in which a hot-standby postmaster could fail to

shut down after receiving a smart-shutdown request (Tom Lane)

</para>

</listitem>

<listitem>

<para>

Tighten validation of encoded SCRAM-SHA-256 and MD5 passwords

(Jonathan Katz)

</para>

<para>

A password string that had the right initial characters could be

mistaken for one that is correctly hashed into SCRAM-SHA-256 or MD5

format. The password would be accepted but would be unusable later.

</para>

</listitem>

<listitem>

<para>

Fix handling of <varname>lc_time</varname> settings that imply an

encoding different from the database's encoding (Juan Jos&eacute;

Santamar&iacute;a Flecha, Tom Lane)

</para>

<para>

Localized month or day names that include non-ASCII characters

previously caused unexpected errors or wrong output in such locales.

</para>

</listitem>

<listitem>

<para>

Disallow <literal>NaN</literal> as a value for floating-point server

parameters (Tom Lane)

</para>

</listitem>

<listitem>

<para>

Rearrange <command>REINDEX</command> processing to avoid assertion

failures when reindexing individual indexes

of <structname>pg_class</structname> (Andres Freund, Tom Lane)

</para>

</listitem>

<listitem>

<para>

Fix planner assertion failure for parameterized dummy paths (Tom Lane)

</para>

</listitem>

<listitem>

<para>

Insert correct test function in the result

of <function>SnapBuildInitialSnapshot()</function> (Antonin Houska)

</para>

<para>

No core code cares about this, but some extensions do.

</para>

</listitem>

<listitem>

<para>

Fix intermittent <quote>could not reattach to shared memory</quote>

session startup failures on Windows (Noah Misch)

</para>

<para>

A previously unrecognized source of these failures is creation of

thread stacks for a process's default thread pool. Arrange for such

stacks to be allocated in a different memory region.

</para>

</listitem>

<listitem>

<para>

Fix error detection in directory scanning on Windows (Konstantin

Knizhnik)

</para>

<para>

Errors, such as lack of permissions to read the directory, were not

detected or reported correctly; instead the code silently acted as

though the directory were empty.

</para>

</listitem>

<listitem>

<para>

Fix grammar problems in <application>ecpg</application> (Tom Lane)

</para>

<para>

A missing semicolon led to mistranslation

of <literal>SET <replaceable>variable</replaceable> =

DEFAULT</literal> (but

not <literal>SET <replaceable>variable</replaceable> TO

DEFAULT</literal>) in <application>ecpg</application> programs,

producing syntactically invalid output that the server would reject.

Additionally, in a <command>DROP TYPE</command> or <command>DROP

DOMAIN</command> command that listed multiple type names, only the

first type name was actually processed.

</para>

</listitem>

<listitem>

<para>

Fix possible buffer overruns in <application>ecpg</application>'s

processing of include filenames (Liu Huailing, Fei Wu)

</para>

</listitem>

<listitem>

<para>

Avoid crash in <filename>contrib/vacuumlo</filename> if

an <function>lo_unlink()</function> call failed (Tom Lane)

</para>

</listitem>

<listitem>

<para>

Sync our copy of the timezone library with IANA tzcode release 2019a

(Tom Lane)

</para>

<para>

This corrects a small bug in <application>zic</application> that

caused it to output an incorrect year-2440 transition in

the <literal>Africa/Casablanca</literal> zone, and adds support

for <application>zic</application>'s new <option>-r</option> option.

</para>

</listitem>

<listitem>

<para>

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

release 2019a for DST law changes in Palestine and Metlakatla,

plus historical corrections for Israel.

</para>

<para>

<literal>Etc/UCT</literal> is now a backward-compatibility link

to <literal>Etc/UTC</literal>, instead of being a separate zone that

generates the abbreviation <literal>UCT</literal>, which nowadays is

typically a typo. <productname>PostgreSQL</productname> will still

accept <literal>UCT</literal> as an input zone abbreviation, but it

won't output it.

</para>

</listitem>

</itemizedlist>

</sect2>

</sect1>