Baoding-FenShuaJiang
/
Green_GoodERP18_FSJ_Standard


			
				
					
						
						
							
							<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>E.2. Release 12.3</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets V1.79.1" /><link rel="prev" href="release-12-4.html" title="E.1. Release 12.4" /><link rel="next" href="release-12-2.html" title="E.3. Release 12.2" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">E.2. Release 12.3</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="release-12-4.html" title="E.1. Release 12.4">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="release.html" title="Appendix E. Release Notes">Up</a></td><th width="60%" align="center">Appendix E. Release Notes</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 12.4 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="release-12-2.html" title="E.3. Release 12.2">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="RELEASE-12-3"><div class="titlepage"><div><div><h2 class="title" style="clear: both">E.2. Release 12.3</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="release-12-3.html#id-1.11.6.6.4">E.2.1. Migration to Version 12.3</a></span></dt><dt><span class="sect2"><a href="release-12-3.html#id-1.11.6.6.5">E.2.2. Changes</a></span></dt></dl></div><p><strong>Release date: </strong>2020-05-14</p><p>
   This release contains a variety of fixes from 12.2.
   For information about new features in major release 12, see
   <a class="xref" href="release-12.html" title="E.5. Release 12">Section E.5</a>.
  </p><div class="sect2" id="id-1.11.6.6.4"><div class="titlepage"><div><div><h3 class="title">E.2.1. Migration to Version 12.3</h3></div></div></div><p>
    A dump/restore is not required for those running 12.X.
   </p><p>
    However, if you are upgrading from a version earlier than 12.2,
    see <a class="xref" href="release-12-2.html" title="E.3. Release 12.2">Section E.3</a>.
   </p></div><div class="sect2" id="id-1.11.6.6.5"><div class="titlepage"><div><div><h3 class="title">E.2.2. Changes</h3></div></div></div><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
      Fix possible failure with <code class="literal">GENERATED</code> columns
      (David Rowley)
     </p><p>
      If a <code class="literal">GENERATED</code> column's value is an exact copy of
      another column of the table (and it is a pass-by-reference data
      type), it was possible to crash or insert corrupted data into the
      table.  While it would be rather pointless for
      a <code class="literal">GENERATED</code> expression to just duplicate another
      column, an expression using a function that sometimes returns its
      input unchanged could create the situation.
     </p></li><li class="listitem"><p>
      Handle inheritance of generated columns better (Peter Eisentraut)
     </p><p>
      When a table column is inherited during <code class="command">CREATE TABLE
      ... INHERITS</code>, disallow changing any generation properties
      when the parent column is already
      marked <code class="literal">GENERATED</code>; but allow a child column to be
      marked <code class="literal">GENERATED</code> when its parent is not.
     </p></li><li class="listitem"><p>
      Fix cross-column references in <code class="command">CREATE TABLE LIKE INCLUDING
      GENERATED</code> (Peter Eisentraut)
     </p><p>
      <code class="command">CREATE TABLE ... LIKE</code> failed when trying to copy
      a <code class="literal">GENERATED</code> expression that references a
      physically-later column.
     </p></li><li class="listitem"><p>
      Propagate <code class="command">ALTER TABLE ... SET STORAGE</code> to indexes
      (Peter Eisentraut)
     </p><p>
      Non-expression index columns have always copied
      the <code class="structfield">attstorage</code> property of their table
      column at creation.  Update them when <code class="command">ALTER TABLE ... SET
      STORAGE</code> is done, to maintain consistency.
     </p></li><li class="listitem"><p>
      Preserve the <code class="structfield">indisclustered</code> setting of
      indexes rewritten by <code class="command">ALTER TABLE</code> (Amit Langote,
      Justin Pryzby)
     </p><p>
      Previously, <code class="command">ALTER TABLE</code> lost track of which index
      had been used for <code class="command">CLUSTER</code>.
     </p></li><li class="listitem"><p>
      Preserve the replica identity properties of indexes rewritten
      by <code class="command">ALTER TABLE</code> (Quan Zongliang, Peter Eisentraut)
     </p></li><li class="listitem"><p>
      Preserve the <code class="structfield">indisclustered</code> setting of
      indexes rebuilt by <code class="command">REINDEX CONCURRENTLY</code>
      (Justin Pryzby)
     </p></li><li class="listitem"><p>
      Lock objects sooner during <code class="command">DROP OWNED BY</code>
      (Álvaro Herrera)
     </p><p>
      This avoids failures in race-condition cases where another session is
      deleting some of the same objects.
     </p></li><li class="listitem"><p>
      Fix error-case processing for <code class="command">CREATE ROLE ... IN
      ROLE</code> (Andrew Gierth)
     </p><p>
      Some error cases would be reported as <span class="quote">“<span class="quote">unexpected node
      type</span>”</span> or the like, instead of the intended message.
     </p></li><li class="listitem"><p>
      Ensure that when a partition is detached, any triggers cloned from
      its formerly-parent table are removed (Justin Pryzby)
     </p></li><li class="listitem"><p>
      Fix crash when <code class="literal">COLLATE</code> is applied to a
      non-collatable type in a partition bound expression (Dmitry Dolgov)
     </p></li><li class="listitem"><p>
      Ensure that unique indexes over partitioned tables match the
      equality semantics of the partitioning key (Guancheng Luo)
     </p><p>
      This would only be an issue with index opclasses that have unusual
      notions of equality, but it's wrong in theory, so check.
     </p></li><li class="listitem"><p>
      Ensure that members of the <code class="literal">pg_read_all_stats</code> role
      can read all statistics views, as expected (Magnus Hagander)
     </p><p>
      The functions underlying
      the <code class="structname">pg_stat_progress_*</code> views had not gotten
      this memo.
     </p></li><li class="listitem"><p>
      Repair performance regression in
      <code class="structname">information_schema</code>.<code class="structname">triggers</code>
      view (Tom Lane)
     </p><p>
      This patch redefines that view so that an
      outer <code class="literal">WHERE</code> clause constraining the table name
      can be pushed down into the view, allowing its calculations to be
      done only for triggers belonging to the table of interest rather
      than all triggers in the database.  In a database with many triggers
      this would make a significant speed difference for queries of that
      form.  Since things worked that way before v11, this is a potential
      performance regression.  Users who find this to be a problem can fix
      it by replacing the view definition (or, perhaps, just deleting and
      reinstalling the whole <code class="structname">information_schema</code>
      schema).
     </p></li><li class="listitem"><p>
      Repair performance regression in floating point overflow/underflow
      detection (Emre Hasegeli)
     </p><p>
      Previous refactoring had resulted in <code class="function">isinf()</code>
      being called extra times in some hot code paths.
     </p></li><li class="listitem"><p>
      Fix full text search to handle NOT above a phrase search correctly
      (Tom Lane)
     </p><p>
      Queries such as <code class="literal">!(foo&lt;-&gt;bar)</code> failed to find
      matching rows when implemented as a GiST or GIN index search.
     </p></li><li class="listitem"><p>
      Fix full text search for cases where a phrase search includes an
      item with both prefix matching and a weight restriction (Tom Lane)
     </p></li><li class="listitem"><p>
      Fix <code class="function">ts_headline()</code> to make better headline
      selections when working with phrase queries (Tom Lane)
     </p></li><li class="listitem"><p>
      Fix bugs in <code class="varname">gin_fuzzy_search_limit</code> processing
      (Adé Heyward, Tom Lane)
     </p><p>
      A small value of <code class="varname">gin_fuzzy_search_limit</code> could
      result in unexpected slowness due to unintentionally rescanning the
      same index page many times.  Another code path failed to apply the
      intended filtering at all, possibly returning too many values.
     </p></li><li class="listitem"><p>
      Allow input of type <code class="type">circle</code> to accept the format
      <span class="quote">“<span class="quote"><code class="literal">(<em class="replaceable"><code>x</code></em>,<em class="replaceable"><code>y</code></em>),<em class="replaceable"><code>r</code></em></code></span>”</span>
      as the documentation says it does (David Zhang)
     </p></li><li class="listitem"><p>
      Make the <code class="function">get_bit()</code>
      and <code class="function">set_bit()</code> functions cope
      with <code class="type">bytea</code> strings longer than 256MB (Movead Li)
     </p><p>
      Since the bit number argument is only <code class="type">int4</code>, it's
      impossible to use these functions to access bits beyond the first
      256MB of a long <code class="type">bytea</code>.  We'll widen the argument
      to <code class="type">int8</code> in v13, but in the meantime, allow these
      functions to work on the initial substring of a
      long <code class="type">bytea</code>.
     </p></li><li class="listitem"><p>
      Ignore file-not-found errors in <code class="function">pg_ls_waldir()</code>
      and allied functions (Tom Lane)
     </p><p>
      This prevents a race condition failure if a file is removed between
      when we see its directory entry and when we attempt
      to <code class="function">stat()</code> it.
     </p></li><li class="listitem"><p>
      Avoid possibly leaking an open-file descriptor for a directory
      in <code class="function">pg_ls_dir()</code>,
      <code class="function">pg_timezone_names()</code>,
      <code class="function">pg_tablespace_databases()</code>, and allied functions
      (Justin Pryzby)
     </p></li><li class="listitem"><p>
      Fix polymorphic-function type resolution to correctly infer the
      actual type of an <code class="type">anyarray</code> output when given only
      an <code class="type">anyrange</code> input (Tom Lane)
     </p></li><li class="listitem"><p>
      Fix server's connection-startup logic for case where a GSSAPI
      connection is rejected because support is not compiled in, and the
      client then tries SSL instead (Andrew Gierth)
     </p><p>
      This led to a bogus <span class="quote">“<span class="quote">unsupported frontend protocol</span>”</span>
      failure.
     </p></li><li class="listitem"><p>
      Fix memory leakage during GSSAPI encryption (Tom Lane)
     </p><p>
      Both the backend and libpq would leak memory equivalent to the total
      amount of data sent during the session, if GSSAPI encryption is in use.
     </p></li><li class="listitem"><p>
      Fix query-lifespan memory leak for a set-returning function used in
      a query's <code class="literal">FROM</code> clause (Andres Freund)
     </p></li><li class="listitem"><p>
      Avoid leakage of a hashed subplan's hash tables across multiple
      executions (Andreas Karlsson, Tom Lane)
     </p><p>
      This mistake could result in severe memory bloat if a query
      re-executed a hashed subplan enough times.
     </p></li><li class="listitem"><p>
      Improve planner's handling of no-op domain coercions (Tom Lane)
     </p><p>
      Fix some cases where a domain coercion that does nothing was not
      completely removed from expressions.
     </p></li><li class="listitem"><p>
      Avoid unlikely crash when <code class="command">REINDEX</code> is terminated
      by a session-shutdown signal (Tom Lane)
     </p></li><li class="listitem"><p>
      Prevent printout of possibly-incorrect hash join table statistics
      in <code class="command">EXPLAIN</code> (Konstantin Knizhnik, Tom Lane, Thomas
      Munro)
     </p></li><li class="listitem"><p>
      Fix reporting of elapsed time for heap truncation steps
      in <code class="command">VACUUM VERBOSE</code> (Tatsuhito Kasahara)
     </p></li><li class="listitem"><p>
      Fix possible undercounting of deleted B-tree index pages
      in <code class="command">VACUUM VERBOSE</code> output (Peter Geoghegan)
     </p></li><li class="listitem"><p>
      Fix wrong bookkeeping for oldest deleted page in a B-tree index
      (Peter Geoghegan)
     </p><p>
      This could cause subtly wrong decisions about
      when <code class="command">VACUUM</code> can skip an index cleanup scan;
      although it appears there may be no significant user-visible effects
      from that.
     </p></li><li class="listitem"><p>
      Ensure that TimelineHistoryRead and TimelineHistoryWrite wait states
      are reported in all code paths that read or write timeline history
      files (Masahiro Ikeda)
     </p></li><li class="listitem"><p>
      Avoid possibly showing <span class="quote">“<span class="quote">waiting</span>”</span> twice in a process's
      PS status (Masahiko Sawada)
     </p></li><li class="listitem"><p>
      Avoid race condition when <code class="command">ANALYZE</code> replaces the
      catalog tuple for extended statistics data (Dean Rasheed)
     </p></li><li class="listitem"><p>
      Remove ill-considered skip of <span class="quote">“<span class="quote">redundant</span>”</span>
      anti-wraparound vacuums (Michael Paquier)
     </p><p>
      This avoids a corner case where autovacuum could get into a loop of
      repeatedly trying and then skipping the same vacuum job.
     </p></li><li class="listitem"><p>
      Ensure INCLUDE'd columns are always removed from B-tree pivot tuples
      (Peter Geoghegan)
     </p><p>
      This mistake wasted space in some rare cases, but was otherwise
      harmless.
     </p></li><li class="listitem"><p>
      Cope with invalid TOAST indexes that could be left over after a
      failed <code class="command">REINDEX CONCURRENTLY</code> (Julien Rouhaud)
     </p></li><li class="listitem"><p>
      Ensure that valid index dependencies are left behind after a
      failed <code class="command">REINDEX CONCURRENTLY</code> (Michael Paquier)
     </p><p>
      Previously the old index could be left with
      no <code class="structname">pg_depend</code> links at all, so that for
      example it would not get dropped if the parent table is dropped.
     </p></li><li class="listitem"><p>
      Avoid failure if autovacuum tries to access a just-dropped temporary
      schema (Tom Lane)
     </p><p>
      This hazard only arises if a superuser manually drops a temporary
      schema; which isn't normal practice, but should work.
     </p></li><li class="listitem"><p>
      Avoid premature recycling of WAL segments during crash recovery
      (Jehan-Guillaume de Rorthais)
     </p><p>
      WAL segments that become ready to be archived during crash recovery
      were potentially recycled without being archived.
     </p></li><li class="listitem"><p>
      Avoid scanning irrelevant timelines during archive recovery (Kyotaro
      Horiguchi)
     </p><p>
      This can eliminate many attempts to fetch non-existent WAL files from
      archive storage, which is helpful if archive access is slow.
     </p></li><li class="listitem"><p>
      Remove bogus <span class="quote">“<span class="quote">subtransaction logged without previous top-level
      txn record</span>”</span> error check in logical decoding (Arseny Sher,
      Amit Kapila)
     </p><p>
      This condition is legitimately reachable in various scenarios, so
      remove the check.
     </p></li><li class="listitem"><p>
      Avoid possible failure after a replication slot copy, due to
      premature removal of WAL data (Masahiko Sawada, Arseny Sher)
     </p></li><li class="listitem"><p>
      Ensure that a replication
      slot's <code class="literal">io_in_progress_lock</code> is released in failure
      code paths (Pavan Deolasee)
     </p><p>
      This could result in a walsender later becoming stuck waiting for
      the lock.
     </p></li><li class="listitem"><p>
      Ensure that generated columns are correctly handled during updates
      issued by logical replication (Peter Eisentraut)
     </p></li><li class="listitem"><p>
      Fix race conditions in synchronous standby management (Tom Lane)
     </p><p>
      During a change in the <code class="varname">synchronous_standby_names</code>
      setting, there was a window in which wrong decisions could be made
      about whether it is OK to release transactions that are waiting for
      synchronous commit.  Another hazard for similarly wrong decisions
      existed if a walsender process exited and was immediately replaced
      by another.
     </p></li><li class="listitem"><p>
      Add missing SQLSTATE values to a few error reports (Sawada Masahiko)
     </p></li><li class="listitem"><p>
      Fix PL/pgSQL to reliably refuse to execute an event trigger function
      as a plain function (Tom Lane)
     </p></li><li class="listitem"><p>
      Fix memory leak in <span class="application">libpq</span> when
      using <code class="literal">sslmode=verify-full</code> (Roman Peshkurov)
     </p><p>
      Certificate verification during connection startup could leak some
      memory.  This would become an issue if a client process opened many
      database connections during its lifetime.
     </p></li><li class="listitem"><p>
      Fix <span class="application">ecpg</span> to treat an argument of
      just <span class="quote">“<span class="quote"><code class="literal">-</code></span>”</span> as meaning <span class="quote">“<span class="quote">read
      from stdin</span>”</span> on all platforms (Tom Lane)
     </p></li><li class="listitem"><p>
      Fix crash in <span class="application">psql</span> when attempting to
      re-establish a failed connection (Michael Paquier)
     </p></li><li class="listitem"><p>
      Allow tab-completion of the filename argument
      to <span class="application">psql</span>'s <code class="command">\gx</code> command
      (Vik Fearing)
     </p></li><li class="listitem"><p>
      Add <span class="application">pg_dump</span> support for <code class="command">ALTER
      ... DEPENDS ON EXTENSION</code> (Álvaro Herrera)
     </p><p>
      <span class="application">pg_dump</span> previously ignored dependencies added
      this way, causing them to be forgotten during dump/restore or
      <span class="application">pg_upgrade</span>.
     </p></li><li class="listitem"><p>
      Fix <span class="application">pg_dump</span> to dump comments on RLS
      policy objects (Tom Lane)
     </p></li><li class="listitem"><p>
      In <span class="application">pg_dump</span>, postpone restore of event
      triggers till the end (Fabrízio de Royes Mello, Hamid Akhtar,
      Tom Lane)
     </p><p>
      This minimizes the risk that an event trigger could interfere with
      the restoration of other objects.
     </p></li><li class="listitem"><p>
      Ensure that <span class="application">pg_basebackup</span> generates valid
      tar files (Robert Haas)
     </p><p>
      In some cases a partial block of zeroes would be added to the end of
      the file.  While this seems to be harmless with common versions of
      tar, it's not OK per the POSIX file format spec.
     </p></li><li class="listitem"><p>
      Make <span class="application">pg_checksums</span> skip tablespace
      subdirectories that belong to a
      different <span class="productname">PostgreSQL</span> major version
      (Michael Banck, Bernd Helmle)
     </p><p>
      Such subdirectories don't really belong to our database cluster, and
      so must not be processed.
     </p></li><li class="listitem"><p>
      Ignore temporary copies of <code class="filename">pg_internal.init</code>
      in <span class="application">pg_checksums</span> and related
      programs (Michael Paquier)
     </p></li><li class="listitem"><p>
      Fix quoting of <code class="option">--encoding</code>, <code class="option">--lc-ctype</code>
      and <code class="option">--lc-collate</code> values
      in <span class="application">createdb</span> utility (Michael Paquier)
     </p></li><li class="listitem"><p>
      <code class="filename">contrib/lo</code>'s <code class="function">lo_manage()</code>
      function crashed if called directly rather than as a trigger (Tom
      Lane)
     </p></li><li class="listitem"><p>
      In <code class="filename">contrib/ltree</code>,
      protect against overflow of <code class="type">ltree</code>
      and <code class="type">lquery</code> length fields (Nikita Glukhov)
     </p></li><li class="listitem"><p>
      Work around failure in <code class="filename">contrib/pageinspect</code>'s
      <code class="function">bt_metap()</code> function when an oldest_xact value
      exceeds 2^31-1 (Peter Geoghegan)
     </p><p>
      Such XIDs will now be reported as negative integers, which isn't
      great but it beats throwing an error.  v13 will widen the output
      argument to <code class="type">int8</code> to provide saner reporting.
     </p></li><li class="listitem"><p>
      Fix cache reference leak in <code class="filename">contrib/sepgsql</code>
      (Michael Luo)
     </p></li><li class="listitem"><p>
      On Windows, avoid premature creation of postmaster's log file
      during <code class="literal">pg_ctl start</code> (Alexander Lakhin)
     </p><p>
      The previous coding could allow the file to be created with
      permissions that wouldn't allow the postmaster to write on it.
     </p></li><li class="listitem"><p>
      Avoid failures when dealing with Unix-style locale names on
      Windows (Juan José Santamaría Flecha)
     </p></li><li class="listitem"><p>
      On Windows, set console VT100 compatibility mode in
      programs that support <code class="varname">PG_COLOR</code> colorization
      (Juan José Santamaría Flecha)
     </p><p>
      Without this, the colorization option doesn't actually work.
     </p></li><li class="listitem"><p>
      Stop requiring extra parentheses in <code class="function">ereport()</code>
      calls (Andres Freund, Tom Lane)
     </p></li><li class="listitem"><p>
      Use <span class="application">pkg-config</span>, if available, to
      locate <span class="application">libxml2</span>
      during <span class="application">configure</span> (Hugh McMaster, Tom
      Lane, Peter Eisentraut)
     </p><p>
      If <span class="application">pkg-config</span> is not present or lacks
      knowledge of <span class="application">libxml2</span>, we still
      query <span class="application">xml2-config</span> as before.
     </p><p>
      This change could break build processes that try to
      make <span class="productname">PostgreSQL</span> use a non-default version
      of <span class="application">libxml2</span> by putting that
      version's <span class="application">xml2-config</span> into
      the <code class="varname">PATH</code>.  Instead,
      set <code class="varname">XML2_CONFIG</code> to point to the
      non-default <span class="application">xml2-config</span>.  That method
      will work with either older or
      newer <span class="productname">PostgreSQL</span> releases.
     </p></li><li class="listitem"><p>
      Fix Makefile dependencies for <span class="application">libpq</span>
      and <span class="application">ecpg</span> (Dagfinn Ilmari Mannsåker)
     </p></li><li class="listitem"><p>
      In MSVC builds, cope with spaces in the path name for Python
      (Victor Wagner)
     </p></li><li class="listitem"><p>
      In MSVC builds, fix detection of Visual Studio version to work with
      more language settings (Andrew Dunstan)
     </p></li><li class="listitem"><p>
      In MSVC builds, use <code class="literal">-Wno-deprecated</code> with bison
      versions newer than 3.0, as non-Windows builds already do (Andrew
      Dunstan)
     </p></li><li class="listitem"><p>
      Update time zone data files to <span class="application">tzdata</span>
      release 2020a for DST law changes in Morocco and the Canadian Yukon,
      plus historical corrections for Shanghai.
     </p><p>
      The America/Godthab zone has been renamed to America/Nuuk to reflect
      current English usage; however, the old name remains available as a
      compatibility link.
     </p><p>
      Also, update <span class="application">initdb</span>'s list of known
      Windows time zone names to include recent additions, improving the
      odds that it will correctly translate the system time zone setting
      on that platform.
     </p></li></ul></div></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="release-12-4.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="release.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="release-12-2.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">E.1. Release 12.4 </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> E.3. Release 12.2</td></tr></table></div></body></html>