release-9.2.sgml

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

 <sect1 id="release-9-2-4">
  <title>Release 9.2.4</title>

  <note>
  <title>Release Date</title>
  <simpara>2013-04-04</simpara>
  </note>

  <para>
   This release contains a variety of fixes from 9.2.3.
   For information about new features in the 9.2 major release, see
   <xref linkend="release-9-2">.
  </para>

  <sect2>
   <title>Migration to Version 9.2.4</title>

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

   <para>
    However, this release corrects several errors in management of GiST
    indexes.  After installing this update, it is advisable to
    <command>REINDEX</> any GiST indexes that meet one or more of the
    conditions described below.
   </para>

   <para>
    Also, if you are upgrading from a version earlier than 9.2.2,
    see the release notes for 9.2.2.
   </para>

  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
     <para>
      Fix insecure parsing of server command-line switches (Mitsumasa
      Kondo, Kyotaro Horiguchi)
     </para>

     <para>
      A connection request containing a database name that begins with
      <quote><literal>-</></quote> could be crafted to damage or destroy
      files within the server's data directory, even if the request is
      eventually rejected.  (CVE-2013-1899)
     </para>
    </listitem>

    <listitem>
     <para>
      Reset OpenSSL randomness state in each postmaster child process
      (Marko Kreen)
     </para>

     <para>
      This avoids a scenario wherein random numbers generated by
      <filename>contrib/pgcrypto</> functions might be relatively easy for
      another database user to guess.  The risk is only significant when
      the postmaster is configured with <varname>ssl</> = <literal>on</>
      but most connections don't use SSL encryption.  (CVE-2013-1900)
     </para>
    </listitem>

    <listitem>
     <para>
      Make REPLICATION privilege checks test current user not authenticated
      user (Noah Misch)
     </para>

     <para>
      An unprivileged database user could exploit this mistake to call
      <function>pg_start_backup()</> or <function>pg_stop_backup()</>,
      thus possibly interfering with creation of routine backups.
      (CVE-2013-1901)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix GiST indexes to not use <quote>fuzzy</> geometric comparisons when
      it's not appropriate to do so (Alexander Korotkov)
     </para>

     <para>
      The core geometric types perform comparisons using <quote>fuzzy</>
      equality, but <function>gist_box_same</> must do exact comparisons,
      else GiST indexes using it might become inconsistent.  After installing
      this update, users should <command>REINDEX</> any GiST indexes on
      <type>box</>, <type>polygon</>, <type>circle</>, or <type>point</>
      columns, since all of these use <function>gist_box_same</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix erroneous range-union and penalty logic in GiST indexes that use
      <filename>contrib/btree_gist</> for variable-width data types, that is
      <type>text</>, <type>bytea</>, <type>bit</>, and <type>numeric</>
      columns (Tom Lane)
     </para>

     <para>
      These errors could result in inconsistent indexes in which some keys
      that are present would not be found by searches, and also in useless
      index bloat.  Users are advised to <command>REINDEX</> such indexes
      after installing this update.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix bugs in GiST page splitting code for multi-column indexes
      (Tom Lane)
     </para>

     <para>
      These errors could result in inconsistent indexes in which some keys
      that are present would not be found by searches, and also in indexes
      that are unnecessarily inefficient to search.  Users are advised to
      <command>REINDEX</> multi-column GiST indexes after installing this
      update.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <function>gist_point_consistent</>
      to handle fuzziness consistently (Alexander Korotkov)
     </para>

     <para>
      Index scans on GiST indexes on <type>point</> columns would sometimes
      yield results different from a sequential scan, because
      <function>gist_point_consistent</> disagreed with the underlying
      operator code about whether to do comparisons exactly or fuzzily.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix buffer leak in WAL replay (Heikki Linnakangas)
     </para>

     <para>
      This bug could result in <quote>incorrect local pin count</> errors
      during replay, making recovery impossible.
     </para>
    </listitem>

    <listitem>
     <para>
      Ensure we do crash recovery before entering archive recovery, if the
      database was not stopped cleanly and a <filename>recovery.conf</> file
      is present (Heikki Linnakangas, Kyotaro Horiguchi, Mitsumasa Kondo)
     </para>

     <para>
      This is needed to ensure that the database is consistent in certain
      scenarios, such as initializing a standby server with a filesystem
      snapshot from a running server.
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid deleting not-yet-archived WAL files during crash recovery
      (Heikki Linnakangas, Fujii Masao)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix race condition in <command>DELETE RETURNING</> (Tom Lane)
     </para>

     <para>
      Under the right circumstances, <command>DELETE RETURNING</> could
      attempt to fetch data from a shared buffer that the current process
      no longer has any pin on.  If some other process changed the buffer
      meanwhile, this would lead to garbage <literal>RETURNING</> output, or
      even a crash.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix infinite-loop risk in regular expression compilation (Tom Lane,
      Don Porter)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix potential null-pointer dereference in regular expression compilation
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <function>to_char()</> to use ASCII-only case-folding rules where
      appropriate (Tom Lane)
     </para>

     <para>
      This fixes misbehavior of some template patterns that should be
      locale-independent, but mishandled <quote><literal>I</></quote> and
      <quote><literal>i</></quote> in Turkish locales.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix unwanted rejection of timestamp <literal>1999-12-31 24:00:00</>
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix SQL-language functions to be safely usable as support
      functions for range types (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix logic error when a single transaction does <command>UNLISTEN</>
      then <command>LISTEN</> (Tom Lane)
     </para>

     <para>
      The session wound up not listening for notify events at all, though it
      surely should listen in this case.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible planner crash after columns have been added to a view
      that's depended on by another view (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix performance issue in <literal>EXPLAIN (ANALYZE, TIMING OFF)</>
      (Pavel Stehule)
     </para>
    </listitem>

    <listitem>
     <para>
      Remove useless <quote>picksplit doesn't support secondary split</> log
      messages (Josh Hansen, Tom Lane)
     </para>

     <para>
      This message seems to have been added in expectation of code that was
      never written, and probably never will be, since GiST's default
      handling of secondary splits is actually pretty good.  So stop nagging
      end users about it.
     </para>
    </listitem>

    <listitem>
     <para>
      Remove vestigial secondary-split support in
      <function>gist_box_picksplit()</> (Tom Lane)
     </para>

     <para>
      Not only was this implementation of secondary-split not better than the
      default implementation, it's actually worse.  So remove it and let the
      default code path handle the case.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible failure to send a session's last few transaction
      commit/abort counts to the statistics collector (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Eliminate memory leaks in PL/Perl's <function>spi_prepare()</> function
      (Alex Hunsaker, Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <application>pg_dumpall</> to handle database names containing
      <quote><literal>=</></quote> correctly (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Avoid crash in <application>pg_dump</> when an incorrect connection
      string is given (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Ignore invalid indexes in <application>pg_dump</> and
      <application>pg_upgrade</> (Michael Paquier, Bruce Momjian)
     </para>

     <para>
      Dumping invalid indexes can cause problems at restore time, for example
      if the reason the index creation failed was because it tried to enforce
      a uniqueness condition not satisfied by the table's data.  Also, if the
      index creation is in fact still in progress, it seems reasonable to
      consider it to be an uncommitted DDL change, which
      <application>pg_dump</> wouldn't be expected to dump anyway.
      <application>pg_upgrade</> now also skips invalid indexes rather than
      failing.
     </para>
    </listitem>

    <listitem>
     <para>
      In <application>pg_basebackup</>, include only the current server
      version's subdirectory when backing up a tablespace (Heikki
      Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Add a server version check in <application>pg_basebackup</> and
      <application>pg_receivexlog</>, so they fail cleanly with version
      combinations that won't work (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <filename>contrib/dblink</> to handle inconsistent settings of
      <varname>DateStyle</> or <varname>IntervalStyle</> safely (Daniel
      Farina, Tom Lane)
     </para>

     <para>
      Previously, if the remote server had different settings of these
      parameters, ambiguous dates might be read incorrectly.  This fix
      ensures that datetime and interval columns fetched by a
      <filename>dblink</> query will be interpreted correctly.  Note however
      that inconsistent settings are still risky, since literal values
      appearing in SQL commands sent to the remote server might be
      interpreted differently than they would be locally.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <filename>contrib/pg_trgm</>'s <function>similarity()</> function
      to return zero for trigram-less strings (Tom Lane)
     </para>

     <para>
      Previously it returned <literal>NaN</> due to internal division by zero.
     </para>
    </listitem>

    <listitem>
     <para>
      Enable building <productname>PostgreSQL</> with Microsoft Visual
      Studio 2012 (Brar Piening, Noah Misch)
     </para>
    </listitem>

    <listitem>
     <para>
      Update time zone data files to <application>tzdata</> release 2013b
      for DST law changes in Chile, Haiti, Morocco, Paraguay, and some
      Russian areas.  Also, historical zone data corrections for numerous
      places.
     </para>

     <para>
      Also, update the time zone abbreviation files for recent changes in
      Russia and elsewhere: <literal>CHOT</>, <literal>GET</>,
      <literal>IRKT</>, <literal>KGT</>, <literal>KRAT</>, <literal>MAGT</>,
      <literal>MAWT</>, <literal>MSK</>, <literal>NOVT</>, <literal>OMST</>,
      <literal>TKT</>, <literal>VLAT</>, <literal>WST</>, <literal>YAKT</>,
      <literal>YEKT</> now follow their current meanings, and
      <literal>VOLT</> (Europe/Volgograd) and <literal>MIST</>
      (Antarctica/Macquarie) are added to the default abbreviations list.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-9-2-3">
  <title>Release 9.2.3</title>

  <note>
  <title>Release Date</title>
  <simpara>2013-02-07</simpara>
  </note>

  <para>
   This release contains a variety of fixes from 9.2.2.
   For information about new features in the 9.2 major release, see
   <xref linkend="release-9-2">.
  </para>

  <sect2>
   <title>Migration to Version 9.2.3</title>

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

   <para>
    However, if you are upgrading from a version earlier than 9.2.2,
    see the release notes for 9.2.2.
   </para>

  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
     <para>
      Prevent execution of <function>enum_recv</> from SQL (Tom Lane)
     </para>

     <para>
      The function was misdeclared, allowing a simple SQL command to crash the
      server.  In principle an attacker might be able to use it to examine the
      contents of server memory.  Our thanks to Sumit Soni (via Secunia SVCRP)
      for reporting this issue.  (CVE-2013-0255)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix multiple problems in detection of when a consistent database
      state has been reached during WAL replay (Fujii Masao, Heikki
      Linnakangas, Simon Riggs, Andres Freund)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix detection of end-of-backup point when no actual redo work is
      required (Heikki Linnakangas)
     </para>

     <para>
      This mistake could result in incorrect <quote>WAL ends before end of
      online backup</> errors.
     </para>
    </listitem>

    <listitem>
     <para>
      Update minimum recovery point when truncating a relation file (Heikki
      Linnakangas)
     </para>

     <para>
      Once data has been discarded, it's no longer safe to stop recovery at
      an earlier point in the timeline.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix recycling of WAL segments after changing recovery target timeline
      (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Properly restore timeline history files from archive on cascading
      standby servers (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix lock conflict detection on hot-standby servers (Andres Freund,
      Robert Haas)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix missing cancellations in hot standby mode (Noah Misch, Simon Riggs)
     </para>

     <para>
      The need to cancel conflicting hot-standby queries would sometimes be
      missed, allowing those queries to see inconsistent data.
     </para>
    </listitem>

    <listitem>
     <para>
      Prevent recovery pause feature from pausing before users can connect
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix SQL grammar to allow subscripting or field selection from a
      sub-SELECT result (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix performance problems with autovacuum truncation in busy workloads
      (Jan Wieck)
     </para>

     <para>
      Truncation of empty pages at the end of a table requires exclusive
      lock, but autovacuum was coded to fail (and release the table lock)
      when there are conflicting lock requests.  Under load, it is easily
      possible that truncation would never occur, resulting in table bloat.
      Fix by performing a partial truncation, releasing the lock, then
      attempting to re-acquire the lock and continue.  This fix also greatly
      reduces the average time before autovacuum releases the lock after a
      conflicting request arrives.
     </para>
    </listitem>

    <listitem>
     <para>
      Improve performance of <function>SPI_execute</> and related
      functions, thereby improving PL/pgSQL's <literal>EXECUTE</>
      (Heikki Linnakangas, Tom Lane)
     </para>

     <para>
      Remove some data-copying overhead that was added in 9.2 as a
      consequence of revisions in the plan caching mechanism.  This
      eliminates a performance regression compared to 9.1, and also saves
      memory, especially when the query string to be executed contains many
      SQL statements.
     </para>

     <para>
      A side benefit is that multi-statement query strings are now
      processed fully serially, that is we complete execution of earlier
      statements before running parse analysis and planning on the
      following ones.  This eliminates a long-standing issue, in that DDL
      that should affect the behavior of a later statement will now behave as
      expected.
     </para>
    </listitem>

    <listitem>
     <para>
      Restore pre-9.2 cost estimates for index usage (Tom Lane)
     </para>

     <para>
      An ill-considered change of a fudge factor led to undesirably high
      cost estimates for use of very large indexes.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix intermittent crash in <literal>DROP INDEX CONCURRENTLY</> (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix potential corruption of shared-memory lock table during
      <command>CREATE/DROP INDEX CONCURRENTLY</> (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <command>COPY</>'s multiple-tuple-insertion code for the case of
      a tuple larger than page size minus fillfactor (Heikki Linnakangas)
     </para>

     <para>
      The previous coding could get into an infinite loop.
     </para>
    </listitem>

    <listitem>
     <para>
      Protect against race conditions when scanning
      <structname>pg_tablespace</> (Stephen Frost, Tom Lane)
     </para>

     <para>
      <command>CREATE DATABASE</> and <command>DROP DATABASE</> could
      misbehave if there were concurrent updates of
      <structname>pg_tablespace</> entries.
     </para>
    </listitem>

    <listitem>
     <para>
      Prevent <command>DROP OWNED</> from trying to drop whole databases or
      tablespaces (&Aacute;lvaro Herrera)
     </para>

     <para>
      For safety, ownership of these objects must be reassigned, not dropped.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix error in <link
      linkend="guc-vacuum-freeze-table-age"><varname>vacuum_freeze_table_age</></link>
      implementation (Andres Freund)
     </para>

     <para>
      In installations that have existed for more than <link
      linkend="guc-vacuum-freeze-min-age"><varname>vacuum_freeze_min_age</></link>
      transactions, this mistake prevented autovacuum from using partial-table
      scans, so that a full-table scan would always happen instead.
     </para>
    </listitem>

    <listitem>
     <para>
      Prevent misbehavior when a <symbol>RowExpr</> or <symbol>XmlExpr</>
      is parse-analyzed twice (Andres Freund, Tom Lane)
     </para>

     <para>
      This mistake could be user-visible in contexts such as
      <literal>CREATE TABLE LIKE INCLUDING INDEXES</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Improve defenses against integer overflow in hashtable sizing
      calculations (Jeff Davis)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix some bugs associated with privileges on datatypes (Tom Lane)
     </para>

     <para>
      There were some issues with default privileges for types, and
      <application>pg_dump</> failed to dump such privileges at all.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix failure to ignore leftover temporary tables after a server crash
      (Tom Lane)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix failure to rotate postmaster log files for size reasons on
      Windows (Jeff Janes, Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Reject out-of-range dates in <function>to_date()</> (Hitoshi Harada)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <function>pg_extension_config_dump()</> to handle
      extension-update cases properly (Tom Lane)
     </para>

     <para>
      This function will now replace any existing entry for the target
      table, making it usable in extension update scripts.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix PL/pgSQL's reporting of plan-time errors in possibly-simple
      expressions (Tom Lane)
     </para>

     <para>
      The previous coding resulted in sometimes omitting the first line in
      the <literal>CONTEXT</> traceback for the error.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix PL/Python's handling of functions used as triggers on multiple
      tables (Andres Freund)
     </para>
    </listitem>

    <listitem>
     <para>
      Ensure that non-ASCII prompt strings are translated to the correct
      code page on Windows (Alexander Law, Noah Misch)
     </para>

     <para>
      This bug affected <application>psql</> and some other client programs.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible crash in <application>psql</>'s <command>\?</> command
      when not connected to a database (Meng Qingzhong)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix possible error if a relation file is removed while
      <application>pg_basebackup</> is running (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Tolerate timeline switches while <literal>pg_basebackup -X fetch</>
      is backing up a standby server (Heikki Linnakangas)
     </para>
    </listitem>

    <listitem>
     <para>
      Make <application>pg_dump</> exclude data of unlogged tables when
      running on a hot-standby server (Magnus Hagander)
     </para>

     <para>
      This would fail anyway because the data is not available on the standby
      server, so it seems most convenient to assume
      <option>--no-unlogged-table-data</> automatically.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <application>pg_upgrade</> to deal with invalid indexes safely
      (Bruce Momjian)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix <application>pg_upgrade</>'s -O/-o options (Marti Raudsepp)
     </para>
    </listitem>

    <listitem>
     <para>
      Fix one-byte buffer overrun in <application>libpq</>'s
      <function>PQprintTuples</> (Xi Wang)
     </para>

     <para>
      This ancient function is not used anywhere by
      <productname>PostgreSQL</> itself, but it might still be used by some
      client code.
     </para>
    </listitem>

    <listitem>
     <para>
      Make <application>ecpglib</> use translated messages properly
      (Chen Huajun)
     </para>
    </listitem>

    <listitem>
     <para>
      Properly install <application>ecpg_compat</> and
      <application>pgtypes</> libraries on MSVC (Jiang Guiqing)
     </para>
    </listitem>

    <listitem>
     <para>
      Include our version of <function>isinf()</> in
      <application>libecpg</> if it's not provided by the system
      (Jiang Guiqing)
     </para>
    </listitem>

    <listitem>
     <para>
      Rearrange configure's tests for supplied functions so it is not
      fooled by bogus exports from libedit/libreadline (Christoph Berg)
     </para>
    </listitem>

    <listitem>
     <para>
      Ensure Windows build number increases over time (Magnus Hagander)
     </para>
    </listitem>

    <listitem>
     <para>
      Make <application>pgxs</> build executables with the right
      <literal>.exe</> suffix when cross-compiling for Windows
      (Zoltan Boszormenyi)
     </para>
    </listitem>

    <listitem>
     <para>
      Add new timezone abbreviation <literal>FET</> (Tom Lane)
     </para>

     <para>
      This is now used in some eastern-European time zones.
     </para>
    </listitem>

   </itemizedlist>

  </sect2>
 </sect1>

 <sect1 id="release-9-2-2">
  <title>Release 9.2.2</title>

  <note>
  <title>Release Date</title>
  <simpara>2012-12-06</simpara>
  </note>

  <para>
   This release contains a variety of fixes from 9.2.1.
   For information about new features in the 9.2 major release, see
   <xref linkend="release-9-2">.
  </para>

  <sect2>
   <title>Migration to Version 9.2.2</title>

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

   <para>
    However, you may need to perform <command>REINDEX</> operations to
    correct problems in concurrently-built indexes, as described in the first
    changelog item below.
   </para>

   <para>
    Also, if you are upgrading from version 9.2.0,
    see the release notes for 9.2.1.
   </para>

  </sect2>

  <sect2>
   <title>Changes</title>

   <itemizedlist>

    <listitem>
     <para>
      Fix multiple bugs associated with <command>CREATE/DROP INDEX
      CONCURRENTLY</> (Andres Freund, Tom Lane, Simon Riggs, Pavan Deolasee)
     </para>

     <para>
      An error introduced while adding <command>DROP INDEX CONCURRENTLY</>
      allowed incorrect indexing decisions to be made during the initial
      phase of <command>CREATE INDEX CONCURRENTLY</>; so that indexes built
      by that command could be corrupt.  It is recommended that indexes
      built in 9.2.X with <command>CREATE INDEX CONCURRENTLY</> be rebuilt
      after applying this update.
     </para>

     <para>
      In addition, fix <command>CREATE/DROP INDEX CONCURRENTLY</> to use
      in-place updates when changing the state of an index's
      <structname>pg_index</> row.  This prevents race conditions that could
      cause concurrent sessions to miss updating the target index, thus
      again resulting in corrupt concurrently-created indexes.
     </para>

     <para>
      Also, fix various other operations to ensure that they ignore
      invalid indexes resulting from a failed <command>CREATE INDEX
      CONCURRENTLY</> command.  The most important of these is
      <command>VACUUM</>, because an auto-vacuum could easily be launched
      on the table before corrective action can be taken to fix or remove
      the invalid index.
     </para>

     <para>
      Also fix <command>DROP INDEX CONCURRENTLY</> to not disable
      insertions into the target index until all queries using it are done.
     </para>

     <para>
      Also fix misbehavior if <command>DROP INDEX CONCURRENTLY</> is
      canceled: the previous coding could leave an un-droppable index behind.
     </para>
    </listitem>

    <listitem>
     <para>
      Correct predicate locking for <command>DROP INDEX CONCURRENTLY</>
      (Kevin Grittner)
     </para>

     <para>
      Previously, SSI predicate locks were processed at the wrong time,
      possibly leading to incorrect behavior of serializable transactions
      executing in parallel with the <command>DROP</>.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix buffer locking during WAL replay (Tom Lane)
     </para>

     <para>
      The WAL replay code was insufficiently careful about locking buffers
      when replaying WAL records that affect more than one page.  This could
      result in hot standby queries transiently seeing inconsistent states,
      resulting in wrong answers or unexpected failures.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix an error in WAL generation logic for GIN indexes (Tom Lane)
     </para>

     <para>
      This could result in index corruption, if a torn-page failure occurred.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix an error in WAL replay logic for SP-GiST indexes (Tom Lane)
     </para>

     <para>
      This could result in index corruption after a crash, or on a standby
      server.
     </para>
    </listitem>

    <listitem>
     <para>
      Fix incorrect detection of end-of-base-backup location during WAL
      recovery (Heikki Linnakangas)
     </para>

     <para>
      This mistake allowed hot standby mode to start up before the database
      reaches a consistent state.
     </para>