release-9.4.sgml

<!-- doc/src/sgml/release-9.4.sgml -->
<!-- See header comment in release.sgml about typical markup -->

 <sect1 id="release-9-4">
  <title>Release 9.4</title>

  <note>
   <title>Release Date</title>
   <simpara>2014-??-??</simpara>
   <simpara>Current as of 2014-08-17</simpara>
  </note>

  <sect2>
   <title>Overview</title>

   <para>
    Major enhancements in <productname>PostgreSQL</> 9.4 include:
   </para>

   <!-- This list duplicates items below, but without authors or details-->

   <itemizedlist>

     <listitem>
      <para>
       Allow <link linkend="rules-materializedviews">materialized views</>
       to be refreshed without blocking reads
      </para>
     </listitem>

     <listitem>
      <para>
       Add support for <link linkend="logicaldecoding">logical decoding</>
       of WAL data, to allow database changes to be streamed out in a
       customizable format
      </para>
     </listitem>

     <listitem>
      <para>
        Allow <link linkend="bgworker">background worker processes</>
        to be dynamically registered, started and terminated
      </para>
     </listitem>

     <listitem>
      <para>
       Add <link linkend="datatype-json"><type>jsonb</></link>, a more
       capable and efficient data type for for storing <acronym>JSON</> data
      </para>
     </listitem>

     <listitem>
      <para>
       Add new <acronym>SQL</> command <xref linkend="SQL-ALTERSYSTEM">
       for updating <filename>postgresql.conf</> configuration file entries
      </para>
     </listitem>

     <listitem>
      <para>
       Reduce lock strength for some <xref linkend="SQL-ALTERTABLE">
       commands
      </para>
     </listitem>

   </itemizedlist>

   <para>
    The above items are explained in more detail in the sections below.
   </para>

  </sect2>

  <sect2>

  <title>Migration to Version 9.4</title>

   <para>
    A dump/restore using <xref linkend="app-pg-dumpall">, or use
    of <xref linkend="pgupgrade">, is required for those wishing to migrate
    data from any previous release.
   </para>

   <para>
    Version 9.4 contains a number of changes that may affect compatibility
    with previous releases.  Observe the following incompatibilities:
   </para>

   <itemizedlist>

    <listitem>
     <para>
      Tighten checks for multidimensional <link
      linkend="arrays">array</link> input (Bruce Momjian)
     </para>

     <para>
      Previously, an input array string that started with a single-element
      array dimension could later contain multidimensional segments,
      e.g. <literal>'{{1}, {2,3}}'::int[]</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Unicode escapes in <link linkend="datatype-json"><type>JSON</type></link>
      text values are no longer rendered with the backslash escaped
      (Andrew Dunstan)
     </para>

     <para>
      Previously, all backslashes in text values being formed into JSON
      were escaped. Now a backslash followed by <literal>u</> and four
      hexadecimal digits is not escaped, as this is a legal sequence in a
      JSON string value, and escaping the backslash led to some perverse
      results.
     </para>
    </listitem>

    <listitem>
     <para>
      When converting values of type <type>date</>, <type>timestamp</>
      or <type>timestamptz</>
      to <link linkend="datatype-json"><type>JSON</type></link>, render the
      values in a format compliant with ISO 8601 (Andrew Dunstan)
     </para>

     <para>
      Previously such values were rendered according to the current
      <xref linkend="guc-datestyle"> setting; but many JSON processors
      require timestamps to be in ISO 8601 format.  If necessary, the
      previous behavior can be obtained by explicitly casting the datetime
      value to <type>text</> before passing it to the JSON conversion
      function.
     </para>
    </listitem>

    <listitem>
     <para>
      Cause consecutive whitespace in <link
      linkend="functions-formatting-table"><function>to_timestamp()</></link>
      and <function>to_date()</> format strings to consume a corresponding
      number of characters in the input string (whitespace or not), then
      conditionally consume adjacent whitespace, if not in <literal>FX</>
      mode (Jeevan Chalke)
     </para>

     <para>
      Previously, consecutive whitespace characters in a non-<literal>FX</>
      format string behaved like a single whitespace character and consumed
      all adjacent whitespace in the input string.  For example, previously
      a format string of three spaces would consume only the first space in
      <literal>' 12'</>, but it will now consume all three characters.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <link
      linkend="textsearch-functions-table"><function>ts_rank_cd()</></link>
      to ignore stripped lexemes (Alex Hill)
     </para>

     <para>
      Previously, stripped lexemes were treated as if they had a default
      location, producing a rank of dubious usefulness.
     </para>
    </listitem>

    <listitem>
     <para>
      For functions declared to
      take <link linkend="xfunc-sql-variadic-functions"><literal>VARIADIC
      "any"</></link>, an actual parameter marked as <literal>VARIADIC</>
      must be of a determinable array type (Pavel Stehule)
     </para>

     <para>
      Such parameters can no longer be written as an undecorated string
      literal or <literal>NULL</>; a cast to an appropriate array data type
      will now be required.  Note that this does not affect parameters not
      marked <literal>VARIADIC</>.
     </para>
    </listitem>

    <listitem>
     <para>
      <xref linkend="sql-discard"> now also discards sequence-related state
      (Fabr&iacute;zio de Royes Mello, Robert Haas)
     </para>
    </listitem>

    <listitem>
     <para>
      Writable foreign data wrappers must return all columns when the foreign
      table has an <literal>AFTER ROW</> trigger (Noah Misch)
     </para>

     <para>
      Previously, foreign tables never had triggers, and
      the <literal>RETURNING</> clause alone dictated the columns required.
     </para>
    </listitem>

    <listitem>
     <para>
      Rename <link linkend="SQL-EXPLAIN"><command>EXPLAIN
      ANALYZE</></link>'s <quote>total runtime</quote> output
      to <quote>execution time</quote> (Tom Lane)
     </para>

     <para>
      Now that planning time is also reported, the previous name was
      confusing.
     </para>
    </listitem>

    <listitem>
     <para>
      <link linkend="SQL-SHOW"><command>SHOW TIME ZONE</></link> now
      outputs simple numeric UTC offsets in <acronym>POSIX</> timezone
      format (Tom Lane)
     </para>

     <para>
      Previously, such timezone settings were displayed as <link
      linkend="datatype-interval-output"><type>interval</></link> values.
      The new output is properly interpreted by <command>SET TIME ZONE</>
      when passed as a simple string, whereas the old output required
      special treatment to be re-parsed correctly.
     </para>
    </listitem>

    <listitem>
     <para>
      Prevent <link
      linkend="ddl-constraints-check-constraints"><literal>CHECK</></link>
      constraints from referencing system columns, except
      <structfield>tableoid</> (Amit Kapila)
     </para>

     <para>
      Previously such check constraints were allowed, but they would often
      cause errors during restores.
     </para>
    </listitem>

    <listitem>
     <para>
      Use the last specified <xref linkend="recovery-target"> if multiple
      values are specified (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      On Windows, automatically preserve quotes in command strings supplied
      by the user (Heikki Linnakangas)
     </para>

     <para>
      User commands that did their own quote preservation might need
      adjustment.  This is likely to be an issue for commands used in
      <xref linkend="guc-archive-command">, <xref linkend="restore-command">,
      and <link linkend="sql-copy"><command>COPY TO/FROM PROGRAM</></link>.
     </para>
    </listitem>

    <listitem>
     <para>
      Remove catalog column <link
      linkend="catalog-pg-class"><structfield>pg_class.reltoastidxid</></link>
      (Michael Paquier)
     </para>
    </listitem>

    <listitem>
     <para>
      Remove catalog column <link
      linkend="catalog-pg-rewrite"><structfield>pg_rewrite.ev_attr</></link>
      (Kevin Grittner)
     </para>

     <para>
      Per-column rules have not been supported since
      <application>PostgreSQL</> 7.3.
     </para>
    </listitem>

    <listitem>
     <para>
      Remove native support for <application>Kerberos</> authentication
      (<option>--with-krb5</>, etc)
      (Magnus Hagander)
     </para>

     <para>
      The supported way to use <application>Kerberos</> authentication is
      with <acronym>GSSAPI</>.  The native code has been deprecated since
      <productname>PostgreSQL</> 8.3.
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>PL/Python</>, handle domains over arrays like the
      underlying array type (Rodolfo Campero)
     </para>

     <para>
      Previously such values were treated as strings.
     </para>
    </listitem>

    <listitem>
     <para>
      Make libpq's <link
      linkend="libpq-pqconnectdbparams"><function>PQconnectdbParams()</></link>
      and <link
      linkend="libpq-pqpingparams"><function>PQpingParams()</></link>
      functions process zero-length strings as defaults (Adrian
      Vondendriesch)
     </para>

     <para>
      Previously, these functions treated zero-length string values as
      selecting the default in only some cases.
     </para>
    </listitem>

    <listitem>
     <para>
      Change empty arrays returned by the <xref linkend="intarray"> module
      to be zero-dimensional arrays (Bruce Momjian)
     </para>

     <para>
      Previously, empty arrays were returned as zero-length one-dimensional
      arrays, whose text representation looked the same as zero-dimensional
      arrays (<literal>{}</>), but they acted differently in array
      operations.  <application>intarray</>'s behavior in this area now
      matches the built-in array operators.
     </para>
    </listitem>

    <listitem>
     <para>
      <xref linkend="pgupgrade"> now uses <option>-U</>
      or <option>--username</> to specify the user name (Bruce Momjian)
     </para>

     <para>
      Previously this option was spelled <option>-u</> or <option>--user</>,
      but that was inconsistent with other tools.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>

  <sect2>
   <title>Changes</title>

   <para>
    Below you will find a detailed account of the changes between
    <productname>PostgreSQL</productname> 9.4 and the previous major
    release.
   </para>

   <sect3>
    <title>Server</title>

     <itemizedlist>

      <listitem>
       <para>
        Allow <link linkend="bgworker">background worker processes</link>
        to be dynamically registered, started and terminated (Robert Haas)
       </para>

       <para>
        The <filename>contrib/worker_spi</> module shows an example of use
        of this feature.
       </para>
      </listitem>

      <listitem>
       <para>
        Allow dynamic allocation of shared memory segments (Robert Haas,
        Amit Kapila)
       </para>

       <para>
        This feature is illustrated in <xref linkend="test-shm-mq">.
       </para>
      </listitem>

      <listitem>
       <para>
        During crash recovery or immediate shutdown, send uncatchable
        termination signals (<systemitem>SIGKILL</>) to child processes
        that do not shut down promptly (MauMau, &Aacute;lvaro Herrera)
       </para>

       <para>
        This reduces the likelihood of leaving orphaned child processes
        behind after <xref linkend="app-postmaster"> shutdown, as well
        as ensuring that crash recovery can proceed if some child processes
        have become <quote>stuck</>.
       </para>
      </listitem>

      <listitem>
       <para>
        Improve randomness of the database system identifier (Tom Lane)
       </para>
      </listitem>

      <listitem>
       <para>
        Make <xref linkend="SQL-VACUUM"> properly report dead but
        not-yet-removable rows to the statistics collector (Hari Babu)
       </para>

       <para>
        Previously these were reported as live rows.
       </para>
      </listitem>

     </itemizedlist>

    <sect4>
     <title>Indexes</title>

     <itemizedlist>

      <listitem>
       <para>
        Reduce <link linkend="GIN"><acronym>GIN</></link> index size
        (Alexander Korotkov, Heikki Linnakangas)
       </para>

       <para>
        Indexes upgraded via <xref linkend="pgupgrade"> will work fine
        but will still be in the old, larger <acronym>GIN</> format.
        Use <xref linkend="SQL-REINDEX"> to recreate such an index in the
        new format.
       </para>
      </listitem>

      <listitem>
       <para>
        Improve speed of multi-key <link
        linkend="GIN"><acronym>GIN</></link> lookups (Alexander Korotkov,
        Heikki Linnakangas)
       </para>
      </listitem>

      <listitem>
       <para>
        Add <link linkend="GiST"><acronym>GiST</></link> index support
        for <link linkend="datatype-inet"><type>inet</></link> and
        <link linkend="datatype-cidr"><type>cidr</></link> data types
        (Emre Hasegeli)
       </para>

       <para>
        Such indexes improve <link
        linkend="cidr-inet-operators-table">subnet and supernet</link>
        lookups and ordering comparisons.
       </para>
      </listitem>

      <listitem>
       <para>
        Fix rare race condition in B-tree page deletion (Heikki Linnakangas)
       </para>
      </listitem>

      <listitem>
       <para>
        Make the handling of interrupted B-tree page splits more robust
        (Heikki Linnakangas)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>General Performance</title>

     <itemizedlist>

      <listitem>
       <para>
        Allow multiple backends to insert
        into <link linkend="wal"><acronym>WAL</></link> buffers
        concurrently (Heikki Linnakangas)
       </para>

       <para>
        This improves parallel write performance.
       </para>
      </listitem>

      <listitem>
       <para>
        Conditionally write only the modified portion of updated rows to
        <link linkend="wal"><acronym>WAL</></link> (Amit Kapila)
       </para>
      </listitem>

      <listitem>
       <para>
        Improve performance of aggregate functions used as <link
        linkend="syntax-window-functions">window functions</link>
        (David Rowley, Florian Pflug, Tom Lane)
       </para>
      </listitem>

      <listitem>
       <para>
        Improve speed of aggregates that
        use <link linkend="datatype-numeric"><type>numeric</></link> state
        values (Hadi Moshayedi)
       </para>
      </listitem>

      <listitem>
       <para>
        Attempt to <link linkend="vacuum-for-wraparound">freeze</link>
        tuples when tables are rewritten with <xref
        linkend="SQL-CLUSTER"> or <link
        linkend="SQL-VACUUM"><command>VACUUM FULL</></link> (Robert Haas,
        Andres Freund)
       </para>

       <para>
        This can avoid the need to freeze the tuples in the future.
       </para>
      </listitem>

      <listitem>
       <para>
        Improve speed of <xref linkend="SQL-COPY">
        with <literal>DEFAULT</> <link
        linkend="functions-sequence-table"><function>nextval()</></link>
        columns (Simon Riggs)
       </para>
      </listitem>

      <listitem>
       <para>
        Improve speed of accessing many different <link
        linkend="SQL-CREATESEQUENCE">sequences</link> in the same session
        (David Rowley)
       </para>
      </listitem>

      <listitem>
       <para>
        Raise hard limit on the number of tuples held in memory during sorting
        and B-tree index builds (Noah Misch)
       </para>
      </listitem>

      <listitem>
       <para>
        Reduce memory allocated by <application>PL/pgSQL</>
        <xref linkend="SQL-DO"> blocks (Tom Lane)
       </para>
      </listitem>

      <listitem>
       <para>
        Make the planner more aggressive in extracting restriction clauses
        from mixed <literal>AND</>/<literal>OR</> clauses (Tom Lane)
       </para>
      </listitem>

      <listitem>
       <para>
        Disallow pushing volatile <literal>WHERE</> clauses down
        into <literal>DISTINCT</> subqueries (Tom Lane)
       </para>

       <para>
        Pushing down a <literal>WHERE</> clause can produce a more
        efficient plan overall, but at the cost of evaluating the clause
        more often than is implied by the text of the query; so don't do it
        if the clause contains any volatile functions.
       </para>
      </listitem>

      <listitem>
       <para>
        Auto-resize the catalog caches (Heikki Linnakangas)
       </para>

       <para>
        This reduces memory consumption for sessions accessing only a few
        tables, and improves performance for sessions accessing many tables.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Monitoring</title>

     <itemizedlist>

      <listitem>
       <para>
        Add <xref linkend="pg-stat-archiver-view"> system view to
        report <link linkend="wal"><acronym>WAL</></link> archiver activity
        (Gabriele Bartolini)
       </para>
      </listitem>

      <listitem>
       <para>
        Add <structfield>n_mod_since_analyze</> columns to
        <xref linkend="pg-stat-all-tables-view"> and related system views
        (Mark Kirkwood)
       </para>

       <para>
        These columns expose the system's estimate of the number of changed
        tuples since the table's last <xref linkend="sql-analyze">.  This
        estimate drives decisions about when to auto-analyze.
       </para>
      </listitem>

      <listitem>
       <para>
        Add <structfield>backend_xid</> and <structfield>backend_xmin</>
        columns to the system view <xref linkend="pg-stat-activity-view">,
        and a <structfield>backend_xmin</> column to
        <xref linkend="pg-stat-replication-view"> (Christian Kruse)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title><acronym>SSL</></title>

     <itemizedlist>

      <listitem>
       <para>
        Add support for <acronym>SSL</> <acronym>ECDH</> key exchange
        (Marko Kreen)
       </para>

       <para>
        This allows use of Elliptic Curve keys for server authentication.
        Such keys are faster and have better security than <acronym>RSA</>
        keys. The new configuration parameter
        <xref linkend="guc-ssl-ecdh-curve">
        controls which curve is used for <acronym>ECDH</>.
       </para>
      </listitem>

      <listitem>
       <para>
        Improve the default <xref linkend="guc-ssl-ciphers"> setting
        (Marko Kreen)
       </para>
      </listitem>

      <listitem>
       <para>
        By default, the server not the client now controls the preference
        order of <acronym>SSL</> ciphers
        (Marko Kreen)
       </para>

       <para>
        Previously, the order specified by <xref linkend="guc-ssl-ciphers">
        was usually ignored in favor of client-side defaults, which are not
        configurable in most <productname>PostgreSQL</> clients.  If
        desired, the old behavior can be restored via the new configuration
        parameter <xref linkend="guc-ssl-prefer-server-ciphers">.
       </para>
      </listitem>

      <listitem>
       <para>
        Make <xref linkend="guc-log-connections"> show <acronym>SSL</>
        encryption information (Andreas Kunert)
       </para>
      </listitem>

      <listitem>
       <para>
        Improve <acronym>SSL</> renegotiation handling (&Aacute;lvaro
        Herrera)
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

    <sect4>
     <title>Server Settings</title>

     <itemizedlist>

      <listitem>
       <para>
        Add new <acronym>SQL</> command <xref linkend="SQL-ALTERSYSTEM">
        for updating <filename>postgresql.conf</> configuration file entries
        (Amit Kapila)
       </para>

       <para>
        Previously such settings could only be changed by manually
        editing <filename>postgresql.conf</>.
       </para>
      </listitem>

      <listitem>
       <para>
        Add <xref linkend="guc-autovacuum-work-mem"> configuration parameter
        to control the amount of memory used by autovacuum workers
        (Peter Geoghegan)
       </para>
      </listitem>

      <listitem>
       <para>
        Add <xref linkend="guc-huge-pages"> parameter to allow using huge
        memory pages on Linux (Christian Kruse, Richard Poole, Abhijit
        Menon-Sen)
       </para>

       <para>
        This can improve performance on large-memory systems.
       </para>
      </listitem>

      <listitem>
       <para>
        Add <xref linkend="guc-max-worker-processes"> parameter
        to limit the number of background workers (Robert Haas)
       </para>

       <para>
        This is helpful in configuring a standby server to have the
        required number of worker processes (the same as the primary).
       </para>
      </listitem>

      <listitem>
       <para>
        Add superuser-only <xref linkend="guc-session-preload-libraries">
        parameter to load libraries at session start (Peter Eisentraut)
       </para>

       <para>
        In contrast to <xref linkend="guc-local-preload-libraries">, this
        parameter can load any shared library, not just those in
        the <filename>$libdir/plugins</> directory.
       </para>
      </listitem>

      <listitem>
       <para>
        Add <xref linkend="guc-wal-log-hints"> parameter to enable WAL
        logging of hint-bit changes (Sawada Masahiko)
       </para>

       <para>
        Hint bit changes are not normally logged, except when checksums are
        enabled.  This is useful for external tools
        like <application>pg_rewind</>.
       </para>
      </listitem>

      <listitem>
       <para>
        Increase the default settings of <xref linkend="guc-work-mem">
        and <xref linkend="guc-maintenance-work-mem"> by four times (Bruce
        Momjian)
       </para>

       <para>
        The new defaults are 4MB and 64MB respectively.
       </para>
      </listitem>

      <listitem>
       <para>
        Increase the default setting of <xref
        linkend="guc-effective-cache-size">
        to 4GB (Bruce Momjian, Tom Lane)
       </para>
      </listitem>

      <listitem>
       <para>
        Allow <function>printf</function>-style space padding to be
        specified in <xref linkend="guc-log-line-prefix"> (David Rowley)
       </para>
      </listitem>

      <listitem>
       <para>
        Allow terabyte units (<literal>TB</>) to be used when specifying
        configuration variable values (Simon Riggs)
       </para>
      </listitem>

      <listitem>
       <para>
        Show <acronym>PID</>s of lock holders and waiters and improve
        information about relations in <xref linkend="guc-log-lock-waits">
        log messages (Christian Kruse)
       </para>
      </listitem>

      <listitem>
       <para>
        Reduce server logging level when loading shared libraries (Peter
        Geoghegan)
       </para>

       <para>
        The previous level was <literal>LOG</>, which was too verbose
        for per-session libraries.
       </para>
      </listitem>

      <listitem>
       <para>
        On Windows, make <literal>SQL_ASCII</>-encoded databases and server
        processes (e.g., <xref linkend="app-postmaster">) emit messages in
        the character encoding of the server's Windows user locale
        (Alexander Law, Noah Misch)
       </para>

       <para>
        Previously these messages were output in the Windows
        <acronym>ANSI</> code page.
       </para>
      </listitem>

     </itemizedlist>

    </sect4>

   </sect3>

   <sect3>
    <title>Replication and Recovery</title>

     <itemizedlist>

      <listitem>
       <para>
        Add <link linkend="streaming-replication-slots">replication
        slots</link> to coordinate activity on streaming standbys with the
        node they are streaming from (Andres Freund, Robert Haas)
       </para>

       <para>
        Replication slots allow preservation of resources like
        <acronym>WAL</> files on the primary until they are no longer
        needed by standby servers.
       </para>
      </listitem>

      <listitem>
       <para>
        Add <link linkend="recovery-config"><filename>recovery.conf</></link>
        parameter <xref linkend="recovery-min-apply-delay">
        to delay replication (Robert Haas, Fabr&iacute;zio de Royes Mello,
        Simon Riggs)
       </para>

       <para>
        Delaying replay on standby servers can be useful for recovering
        from user errors.
       </para>
      </listitem>

      <listitem>
       <para>
        Add <xref linkend="recovery-target">
        option <option>immediate</> to stop <link
        linkend="wal"><acronym>WAL</></link> recovery as soon as a
        consistent state is reached (MauMau, Heikki Linnakangas)
       </para>
      </listitem>

      <listitem>
       <para>
        Improve recovery target processing (Heikki Linnakangas)
       </para>

       <para>
        The timestamp reported
        by <link linkend="functions-recovery-info-table"><function>pg_last_xact_replay_timestamp()</></link>
        now reflects already-committed records, not transactions about to
        be committed. Recovering to a restore point now replays the restore
        point, rather than stopping just before the restore point.
       </para>
      </listitem>

      <listitem>
       <para>
        <link
        linkend="functions-admin-backup-table"><function>pg_switch_xlog()</></link>
        now clears any unused trailing space in the old <acronym>WAL</> file
        (Heikki Linnakangas)
       </para>

       <para>
        This improves the compression ratio for <acronym>WAL</> files.
       </para>
      </listitem>

      <listitem>
       <para>
        Report failure return codes from <link
        linkend="archive-recovery-settings">external recovery commands</>
        (Peter Eisentraut)
       </para>
      </listitem>

      <listitem>
       <para>
        Reduce spinlock contention during <acronym>WAL</> replay (Heikki
        Linnakangas)
       </para>
      </listitem>

      <listitem>
       <para>
        Write <acronym>WAL</> records of running transactions more
        frequently (Andres Freund)
       </para>

       <para>
        This allows standby servers to start faster and clean up resources
        more aggressively.
       </para>
      </listitem>

     </itemizedlist>

     <sect4>
      <title><link linkend="logicaldecoding">Logical Decoding</></title>

      <para>
       Logical decoding allows database changes to be streamed in a
       configurable format. The data is read from
       the <link linkend="wal"><acronym>WAL</></link> and transformed into the
       desired target format. To implement this feature, the following changes
       were made:
      </para>

      <itemizedlist>

       <listitem>
        <para>
         Add support for <link linkend="logicaldecoding">logical decoding</>
         of WAL data, to allow database changes to be streamed out in a
         customizable format
         (Andres Freund)
        </para>
       </listitem>

       <listitem>
        <para>
         Add new <xref linkend="guc-wal-level"> setting <option>logical</>
         to enable logical change-set encoding in <acronym>WAL</> (Andres
         Freund)
        </para>
       </listitem>

       <listitem>
        <para>
         Add table-level parameter <link
         linkend="catalog-pg-class"><literal>REPLICA IDENTITY</></link>
         to control logical replication (Andres Freund)
        </para>
       </listitem>