release-9.2.sgml

<!-- doc/src/sgml/release-9.2.sgml -->
<!-- See header comment in release.sgml about typical markup -->

 <sect1 id="release-9-2-11">
  <title>Release 9.2.11</title>

  <note>
  <title>Release Date</title>
  <simpara>2015-05-21</simpara>
  </note>

  <para>
   This release contains a variety of fixes from 9.2.10.
   For information about new features in the 9.2 major release, see
   <xref linkend="release-9-2">.
  </para>

  <sect2>
   <title>Migration to Version 9.2.11</title>

   <para>
    A dump/restore is not required for those running 9.2.X.
   </para>

   <para>
    However, if you use <filename>contrib/citext</>'s
    <function>regexp_matches()</> functions, see the changelog entry below
    about that.
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 9.2.10,
    see <xref linkend="release-9-2-10">.
   </para>

  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
     <para>
      Fix incorrect declaration of <filename>contrib/citext</>'s
      <function>regexp_matches()</> functions (Tom Lane)
     </para>

     <para>
      These functions should return <type>setof text[]</>, like the core
      functions they are wrappers for; but they were incorrectly declared as
      returning just <type>text[]</>.  This mistake had two results: first,
      if there was no match you got a scalar null result, whereas what you
      should get is an empty set (zero rows).  Second, the <literal>g</> flag
      was effectively ignored, since you would get only one result array even
      if there were multiple matches.
     </para>

     <para>
      While the latter behavior is clearly a bug, there might be applications
      depending on the former behavior; therefore the function declarations
      will not be changed by default until <productname>PostgreSQL</> 9.5.
      In pre-9.5 branches, the old behavior exists in version 1.0 of
      the <literal>citext</> extension, while we have provided corrected
      declarations in version 1.1 (which is <emphasis>not</> installed by
      default).  To adopt the fix in pre-9.5 branches, execute
      <literal>ALTER EXTENSION citext UPDATE TO '1.1'</> in each database in
      which <literal>citext</> is installed.  (You can also <quote>update</>
      back to 1.0 if you need to undo that.)  Be aware that either update
      direction will require dropping and recreating any views or rules that
      use <filename>citext</>'s <function>regexp_matches()</> functions.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix incorrect checking of deferred exclusion constraints after a HOT
      update (Tom Lane)
     </para>

     <para>
      If a new row that potentially violates a deferred exclusion constraint
      is HOT-updated (that is, no indexed columns change and the row can be
      stored back onto the same table page) later in the same transaction,
      the exclusion constraint would be reported as violated when the check
      finally occurred, even if the row(s) the new row originally conflicted
      with had been deleted.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix planning of star-schema-style queries (Tom Lane)
     </para>

     <para>
      Sometimes, efficient scanning of a large table requires that index
      parameters be provided from more than one other table (commonly,
      dimension tables whose keys are needed to index a large fact table).
      The planner should be able to find such plans, but an overly
      restrictive search heuristic prevented it.
     </para>
    </listitem>

    <listitem>
     <para>
      Prevent improper reordering of antijoins (NOT EXISTS joins) versus
      other outer joins (Tom Lane)
     </para>

     <para>
      This oversight in the planner has been observed to cause <quote>could
      not find RelOptInfo for given relids</> errors, but it seems possible
      that sometimes an incorrect query plan might get past that consistency
      check and result in silently-wrong query output.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix incorrect matching of subexpressions in outer-join plan nodes
      (Tom Lane)
     </para>

     <para>
      Previously, if textually identical non-strict subexpressions were used
      both above and below an outer join, the planner might try to re-use
      the value computed below the join, which would be incorrect because the
      executor would force the value to NULL in case of an unmatched outer row.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix GEQO planner to cope with failure of its join order heuristic
      (Tom Lane)
     </para>

     <para>
      This oversight has been seen to lead to <quote>failed to join all
      relations together</> errors in queries involving <literal>LATERAL</>,
      and that might happen in other cases as well.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible deadlock at startup
      when <literal>max_prepared_transactions</> is too small
      (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Don't archive useless preallocated WAL files after a timeline switch
      (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Recursively <function>fsync()</> the data directory after a crash
      (Abhijit Menon-Sen, Robert Haas)
     </para>

     <para>
      This ensures consistency if another crash occurs shortly later.  (The
      second crash would have to be a system-level crash, not just a database
      crash, for there to be a problem.)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix autovacuum launcher's possible failure to shut down, if an error
      occurs after it receives SIGTERM (&Aacute;lvaro Herrera)
     </para>
    </listitem>

    <listitem>
     <para>
      Cope with unexpected signals in <function>LockBufferForCleanup()</>
      (Andres Freund)
     </para>

     <para>
      This oversight could result in spurious errors about <quote>multiple
      backends attempting to wait for pincount 1</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix crash when doing <literal>COPY IN</> to a table with check
      constraints that contain whole-row references (Tom Lane)
     </para>

     <para>
      The known failure case only crashes in 9.4 and up, but there is very
      similar code in 9.3 and 9.2, so back-patch those branches as well.
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid waiting for WAL flush or synchronous replication during commit of
      a transaction that was read-only so far as the user is concerned
      (Andres Freund)
     </para>

     <para>
      Previously, a delay could occur at commit in transactions that had
      written WAL due to HOT page pruning, leading to undesirable effects
      such as sessions getting stuck at startup if all synchronous replicas
      are down.  Sessions have also been observed to get stuck in catchup
      interrupt processing when using synchronous replication; this will fix
      that problem as well.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix crash when manipulating hash indexes on temporary tables
      (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible failure during hash index bucket split, if other processes
      are modifying the index concurrently (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Check for interrupts while analyzing index expressions (Jeff Janes)
     </para>

     <para>
      <command>ANALYZE</> executes index expressions many times; if there are
      slow functions in such an expression, it's desirable to be able to
      cancel the <command>ANALYZE</> before that loop finishes.
     </para>
    </listitem>

    <listitem>
     <para>
      Ensure <structfield>tableoid</> of a foreign table is reported
      correctly when a <literal>READ COMMITTED</> recheck occurs after
      locking rows in <command>SELECT FOR UPDATE</>, <command>UPDATE</>,
      or <command>DELETE</> (Etsuro Fujita)
     </para>
    </listitem>

    <listitem>
     <para>
      Add the name of the target server to object description strings for
      foreign-server user mappings (&Aacute;lvaro Herrera)
     </para>
    </listitem>

    <listitem>
     <para>
      Recommend setting <literal>include_realm</> to 1 when using
      Kerberos/GSSAPI/SSPI authentication (Stephen Frost)
     </para>

     <para>
      Without this, identically-named users from different realms cannot be
      distinguished.  For the moment this is only a documentation change, but
      it will become the default setting in <productname>PostgreSQL</> 9.5.
     </para>
    </listitem>

    <listitem>
     <para>
      Remove code for matching IPv4 <filename>pg_hba.conf</> entries to
      IPv4-in-IPv6 addresses (Tom Lane)
     </para>

     <para>
      This hack was added in 2003 in response to a report that some Linux
      kernels of the time would report IPv4 connections as having
      IPv4-in-IPv6 addresses.  However, the logic was accidentally broken in
      9.0.  The lack of any field complaints since then shows that it's not
      needed anymore.  Now we have reports that the broken code causes
      crashes on some systems, so let's just remove it rather than fix it.
      (Had we chosen to fix it, that would make for a subtle and potentially
      security-sensitive change in the effective meaning of
      IPv4 <filename>pg_hba.conf</> entries, which does not seem like a good
      thing to do in minor releases.)
     </para>
    </listitem>

    <listitem>
     <para>
      Report WAL flush, not insert, position in <literal>IDENTIFY_SYSTEM</>
      replication command (Heikki Linnakangas)
     </para>

     <para>
      This avoids a possible startup failure
      in <application>pg_receivexlog</>.
     </para>
    </listitem>

    <listitem>
     <para>
      While shutting down service on Windows, periodically send status
      updates to the Service Control Manager to prevent it from killing the
      service too soon; and ensure that <application>pg_ctl</> will wait for
      shutdown (Krystian Bigaj)
     </para>
    </listitem>

    <listitem>
     <para>
      Reduce risk of network deadlock when using <application>libpq</>'s
      non-blocking mode (Heikki Linnakangas)
     </para>

     <para>
      When sending large volumes of data, it's important to drain the input
      buffer every so often, in case the server has sent enough response data
      to cause it to block on output.  (A typical scenario is that the server
      is sending a stream of NOTICE messages during <literal>COPY FROM
      STDIN</>.)  This worked properly in the normal blocking mode, but not
      so much in non-blocking mode.  We've modified <application>libpq</>
      to opportunistically drain input when it can, but a full defense
      against this problem requires application cooperation: the application
      should watch for socket read-ready as well as write-ready conditions,
      and be sure to call <function>PQconsumeInput()</> upon read-ready.
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>libpq</>, fix misparsing of empty values in URI
      connection strings (Thomas Fanghaenel)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix array handling in <application>ecpg</> (Michael Meskes)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <application>psql</> to sanely handle URIs and conninfo strings as
      the first parameter to <command>\connect</>
      (David Fetter, Andrew Dunstan, &Aacute;lvaro Herrera)
     </para>

     <para>
      This syntax has been accepted (but undocumented) for a long time, but
      previously some parameters might be taken from the old connection
      instead of the given string, which was agreed to be undesirable.
     </para>
    </listitem>

    <listitem>
     <para>
      Suppress incorrect complaints from <application>psql</> on some
      platforms that it failed to write <filename>~/.psql_history</> at exit
      (Tom Lane)
     </para>

     <para>
      This misbehavior was caused by a workaround for a bug in very old
      (pre-2006) versions of <application>libedit</>.  We fixed it by
      removing the workaround, which will cause a similar failure to appear
      for anyone still using such versions of <application>libedit</>.
      Recommendation: upgrade that library, or use <application>libreadline</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <application>pg_dump</>'s rule for deciding which casts are
      system-provided casts that should not be dumped (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>pg_dump</>, fix failure to honor <literal>-Z</>
      compression level option together with <literal>-Fd</>
      (Michael Paquier)
     </para>
    </listitem>

    <listitem>
     <para>
      Make <application>pg_dump</> consider foreign key relationships
      between extension configuration tables while choosing dump order
      (Gilles Darold, Michael Paquier, Stephen Frost)
     </para>

     <para>
      This oversight could result in producing dumps that fail to reload
      because foreign key constraints are transiently violated.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix dumping of views that are just <literal>VALUES(...)</> but have
      column aliases (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>pg_upgrade</>, force timeline 1 in the new cluster
      (Bruce Momjian)
     </para>

     <para>
      This change prevents upgrade failures caused by bogus complaints about
      missing WAL history files.
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>pg_upgrade</>, check for improperly non-connectable
      databases before proceeding
      (Bruce Momjian)
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>pg_upgrade</>, quote directory paths
      properly in the generated <literal>delete_old_cluster</> script
      (Bruce Momjian)
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>pg_upgrade</>, preserve database-level freezing info
      properly
      (Bruce Momjian)
     </para>

     <para>
      This oversight could cause missing-clog-file errors for tables within
      the <literal>postgres</> and <literal>template1</> databases.
     </para>
    </listitem>

    <listitem>
     <para>
      Run <application>pg_upgrade</> and <application>pg_resetxlog</> with
      restricted privileges on Windows, so that they don't fail when run by
      an administrator (Muhammad Asif Naeem)
     </para>
    </listitem>

    <listitem>
     <para>
      Improve handling of <function>readdir()</> failures when scanning
      directories in <application>initdb</> and <application>pg_basebackup</>
      (Marco Nenciarini)
     </para>
    </listitem>

<!--
Author: Andres Freund <andres@anarazel.de>
Branch: REL9_2_STABLE [6b700301c] 2015-02-17 16:03:00 +0100
-->

    <listitem>
     <para>
      Fix failure in <application>pg_receivexlog</> (Andres Freund)
     </para>

     <para>
      A patch merge mistake in 9.2.10 led to <quote>could not create archive
      status file</> errors.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix slow sorting algorithm in <filename>contrib/intarray</> (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix compile failure on Sparc V8 machines (Rob Rowan)
     </para>
    </listitem>

    <listitem>
     <para>
      Update time zone data files to <application>tzdata</> release 2015d
      for DST law changes in Egypt, Mongolia, and Palestine, plus historical
      changes in Canada and Chile.  Also adopt revised zone abbreviations for
      the America/Adak zone (HST/HDT not HAST/HADT).
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-9-2-10">
  <title>Release 9.2.10</title>

  <note>
  <title>Release Date</title>
  <simpara>2015-02-05</simpara>
  </note>

  <para>
   This release contains a variety of fixes from 9.2.9.
   For information about new features in the 9.2 major release, see
   <xref linkend="release-9-2">.
  </para>

  <sect2>
   <title>Migration to Version 9.2.10</title>

   <para>
    A dump/restore is not required for those running 9.2.X.
   </para>

   <para>
    However, if you are a Windows user and are using the <quote>Norwegian
    (Bokm&aring;l)</> locale, manual action is needed after the upgrade to
    replace any <quote>Norwegian (Bokm&aring;l)_Norway</> locale names stored
    in <productname>PostgreSQL</> system catalogs with the plain-ASCII
    alias <quote>Norwegian_Norway</>.  For details see
    <ulink url="http://wiki.postgresql.org/wiki/Changes_To_Norwegian_Locale"></>
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 9.2.9,
    see <xref linkend="release-9-2-9">.
   </para>

  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
     <para>
      Fix buffer overruns in <function>to_char()</>
      (Bruce Momjian)
     </para>

     <para>
      When <function>to_char()</> processes a numeric formatting template
      calling for a large number of digits, <productname>PostgreSQL</>
      would read past the end of a buffer.  When processing a crafted
      timestamp formatting template, <productname>PostgreSQL</> would write
      past the end of a buffer.  Either case could crash the server.
      We have not ruled out the possibility of attacks that lead to
      privilege escalation, though they seem unlikely.
      (CVE-2015-0241)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix buffer overrun in replacement <function>*printf()</> functions
      (Tom Lane)
     </para>

     <para>
      <productname>PostgreSQL</> includes a replacement implementation
      of <function>printf</> and related functions.  This code will overrun
      a stack buffer when formatting a floating point number (conversion
      specifiers <literal>e</>, <literal>E</>, <literal>f</>, <literal>F</>,
      <literal>g</> or <literal>G</>) with requested precision greater than
      about 500.  This will crash the server, and we have not ruled out the
      possibility of attacks that lead to privilege escalation.
      A database user can trigger such a buffer overrun through
      the <function>to_char()</> SQL function.  While that is the only
      affected core <productname>PostgreSQL</> functionality, extension
      modules that use printf-family functions may be at risk as well.
     </para>

     <para>
      This issue primarily affects <productname>PostgreSQL</> on Windows.
      <productname>PostgreSQL</> uses the system implementation of these
      functions where adequate, which it is on other modern platforms.
      (CVE-2015-0242)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix buffer overruns in <filename>contrib/pgcrypto</>
      (Marko Tiikkaja, Noah Misch)
     </para>

     <para>
      Errors in memory size tracking within the <filename>pgcrypto</>
      module permitted stack buffer overruns and improper dependence on the
      contents of uninitialized memory.  The buffer overrun cases can
      crash the server, and we have not ruled out the possibility of
      attacks that lead to privilege escalation.
      (CVE-2015-0243)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible loss of frontend/backend protocol synchronization after
      an error
      (Heikki Linnakangas)
     </para>

     <para>
      If any error occurred while the server was in the middle of reading a
      protocol message from the client, it could lose synchronization and
      incorrectly try to interpret part of the message's data as a new
      protocol message.  An attacker able to submit crafted binary data
      within a command parameter might succeed in injecting his own SQL
      commands this way.  Statement timeout and query cancellation are the
      most likely sources of errors triggering this scenario.  Particularly
      vulnerable are applications that use a timeout and also submit
      arbitrary user-crafted data as binary query parameters.  Disabling
      statement timeout will reduce, but not eliminate, the risk of
      exploit.  Our thanks to Emil Lenngren for reporting this issue.
      (CVE-2015-0244)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix information leak via constraint-violation error messages
      (Stephen Frost)
     </para>

     <para>
      Some server error messages show the values of columns that violate
      a constraint, such as a unique constraint.  If the user does not have
      <literal>SELECT</> privilege on all columns of the table, this could
      mean exposing values that the user should not be able to see.  Adjust
      the code so that values are displayed only when they came from the SQL
      command or could be selected by the user.
      (CVE-2014-8161)
     </para>
    </listitem>

    <listitem>
     <para>
      Lock down regression testing's temporary installations on Windows
      (Noah Misch)
     </para>

     <para>
      Use SSPI authentication to allow connections only from the OS user
      who launched the test suite.  This closes on Windows the same
      vulnerability previously closed on other platforms, namely that other
      users might be able to connect to the test postmaster.
      (CVE-2014-0067)
     </para>
    </listitem>

    <listitem>
     <para>
      Cope with the Windows locale named <quote>Norwegian (Bokm&aring;l)</>
      (Heikki Linnakangas)
     </para>

     <para>
      Non-ASCII locale names are problematic since it's not clear what
      encoding they should be represented in.  Map the troublesome locale
      name to a plain-ASCII alias, <quote>Norwegian_Norway</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid possible data corruption if <command>ALTER DATABASE SET
      TABLESPACE</> is used to move a database to a new tablespace and then
      shortly later move it back to its original tablespace (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid corrupting tables when <command>ANALYZE</> inside a transaction
      is rolled back (Andres Freund, Tom Lane, Michael Paquier)
     </para>

     <para>
      If the failing transaction had earlier removed the last index, rule, or
      trigger from the table, the table would be left in a corrupted state
      with the relevant <structname>pg_class</> flags not set though they
      should be.
     </para>
    </listitem>

    <listitem>
     <para>
      Ensure that unlogged tables are copied correctly
      during <command>CREATE DATABASE</> or <command>ALTER DATABASE SET
      TABLESPACE</> (Pavan Deolasee, Andres Freund)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <command>DROP</>'s dependency searching to correctly handle the
      case where a table column is recursively visited before its table
      (Petr Jelinek, Tom Lane)
     </para>

     <para>
      This case is only known to arise when an extension creates both a
      datatype and a table using that datatype.  The faulty code might
      refuse a <command>DROP EXTENSION</> unless <literal>CASCADE</> is
      specified, which should not be required.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix use-of-already-freed-memory problem in EvalPlanQual processing
      (Tom Lane)
     </para>

     <para>
      In <literal>READ COMMITTED</> mode, queries that lock or update
      recently-updated rows could crash as a result of this bug.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix planning of <command>SELECT FOR UPDATE</> when using a partial
      index on a child table (Kyotaro Horiguchi)
     </para>

     <para>
      In <literal>READ COMMITTED</> mode, <command>SELECT FOR UPDATE</> must
      also recheck the partial index's <literal>WHERE</> condition when
      rechecking a recently-updated row to see if it still satisfies the
      query's <literal>WHERE</> condition.  This requirement was missed if the
      index belonged to an inheritance child table, so that it was possible
      to incorrectly return rows that no longer satisfy the query condition.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix corner case wherein <command>SELECT FOR UPDATE</> could return a row
      twice, and possibly miss returning other rows (Tom Lane)
     </para>

     <para>
      In <literal>READ COMMITTED</> mode, a <command>SELECT FOR UPDATE</>
      that is scanning an inheritance tree could incorrectly return a row
      from a prior child table instead of the one it should return from a
      later child table.
     </para>
    </listitem>

    <listitem>
     <para>
      Reject duplicate column names in the referenced-columns list of
      a <literal>FOREIGN KEY</> declaration (David Rowley)
     </para>

     <para>
      This restriction is per SQL standard.  Previously we did not reject
      the case explicitly, but later on the code would fail with
      bizarre-looking errors.
     </para>
    </listitem>

    <listitem>
     <para>
      Restore previous behavior of conversion of domains to JSON
      (Tom Lane)
     </para>

     <para>
      This change causes domains over numeric and boolean to be treated
      like their base types for purposes of conversion to JSON.  It worked
      like that before 9.3.5 and 9.2.9, but was unintentionally changed
      while fixing a related problem.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix bugs in raising a <type>numeric</> value to a large integral power
      (Tom Lane)
     </para>

     <para>
      The previous code could get a wrong answer, or consume excessive
      amounts of time and memory before realizing that the answer must
      overflow.
     </para>
    </listitem>

    <listitem>
     <para>
      In <function>numeric_recv()</>, truncate away any fractional digits
      that would be hidden according to the value's <literal>dscale</> field
      (Tom Lane)
     </para>

     <para>
      A <type>numeric</> value's display scale (<literal>dscale</>) should
      never be less than the number of nonzero fractional digits; but
      apparently there's at least one broken client application that
      transmits binary <type>numeric</> values in which that's true.
      This leads to strange behavior since the extra digits are taken into
      account by arithmetic operations even though they aren't printed.
      The least risky fix seems to be to truncate away such <quote>hidden</>
      digits on receipt, so that the value is indeed what it prints as.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix incorrect search for shortest-first regular expression matches
      (Tom Lane)
     </para>

     <para>
      Matching would often fail when the number of allowed iterations is
      limited by a <literal>?</> quantifier or a bound expression.
     </para>
    </listitem>

    <listitem>
     <para>
      Reject out-of-range numeric timezone specifications (Tom Lane)
     </para>

     <para>
      Simple numeric timezone specifications exceeding +/- 168 hours (one
      week) would be accepted, but could then cause null-pointer dereference
      crashes in certain operations.  There's no use-case for such large UTC
      offsets, so reject them.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix bugs in <type>tsquery</> <literal>@&gt;</> <type>tsquery</>
      operator (Heikki Linnakangas)
     </para>

     <para>
      Two different terms would be considered to match if they had the same
      CRC.  Also, if the second operand had more terms than the first, it
      would be assumed not to be contained in the first; which is wrong
      since it might contain duplicate terms.
     </para>
    </listitem>

    <listitem>
     <para>
      Improve ispell dictionary's defenses against bad affix files (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Allow more than 64K phrases in a thesaurus dictionary (David Boutin)
     </para>

     <para>
      The previous coding could crash on an oversize dictionary, so this was
      deemed a back-patchable bug fix rather than a feature addition.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix namespace handling in <function>xpath()</> (Ali Akbar)
     </para>

     <para>
      Previously, the <type>xml</> value resulting from
      an <function>xpath()</> call would not have namespace declarations if
      the namespace declarations were attached to an ancestor element in the
      input <type>xml</> value, rather than to the specific element being
      returned.  Propagate the ancestral declaration so that the result is
      correct when considered in isolation.
     </para>
    </listitem>

    <listitem>
     <para>
      Ensure that whole-row variables expose nonempty column names
      to functions that pay attention to column names within composite
      arguments (Tom Lane)
     </para>

     <para>
      In some contexts, constructs like <literal>row_to_json(tab.*)</> may
      not produce the expected column names.  This is fixed properly as of
      9.4; in older branches, just ensure that we produce some nonempty
      name.  (In some cases this will be the underlying table's column name
      rather than the query-assigned alias that should theoretically be
      visible.)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix mishandling of system columns,
      particularly <structfield>tableoid</>, in FDW queries (Etsuro Fujita)
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid doing <literal><replaceable>indexed_column</> = ANY
      (<replaceable>array</>)</literal> as an index qualifier if that leads
      to an inferior plan (Andrew Gierth)
     </para>

     <para>
      In some cases, <literal>= ANY</> conditions applied to non-first index
      columns would be done as index conditions even though it would be
      better to use them as simple filter conditions.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix planner problems with nested append relations, such as inherited
      tables within <literal>UNION ALL</> subqueries (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fail cleanly when a GiST index tuple doesn't fit on a page, rather
      than going into infinite recursion (Andrew Gierth)
     </para>
    </listitem>

    <listitem>
     <para>
      Exempt tables that have per-table <varname>cost_limit</>
      and/or <varname>cost_delay</> settings from autovacuum's global cost
      balancing rules (&Aacute;lvaro Herrera)
     </para>

     <para>
      The previous behavior resulted in basically ignoring these per-table
      settings, which was unintended.  Now, a table having such settings
      will be vacuumed using those settings, independently of what is going
      on in other autovacuum workers.  This may result in heavier total I/O
      load than before, so such settings should be re-examined for sanity.
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid wholesale autovacuuming when autovacuum is nominally off
      (Tom Lane)
     </para>

     <para>
      Even when autovacuum is nominally off, we will still launch autovacuum
      worker processes to vacuum tables that are at risk of XID wraparound.
      However, such a worker process then proceeded to vacuum all tables in
      the target database, if they met the usual thresholds for
      autovacuuming.  This is at best pretty unexpected; at worst it delays
      response to the wraparound threat.  Fix it so that if autovacuum is
      turned off, workers <emphasis>only</> do anti-wraparound vacuums and
      not any other work.
     </para>
    </listitem>

    <listitem>
     <para>
      During crash recovery, ensure that unlogged relations are rewritten as
      empty and are synced to disk before recovery is considered complete
      (Abhijit Menon-Sen, Andres Freund)
     </para>

     <para>
      This prevents scenarios in which unlogged relations might contain
      garbage data following database crash recovery.
     </para>