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.1. Release 12.4</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.html" title="Appendix E. Release Notes" /><link rel="next" href="release-12-3.html" title="E.2. Release 12.3" /></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.1. Release 12.4</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="release.html" title="Appendix E. Release Notes">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-12-3.html" title="E.2. Release 12.3">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="RELEASE-12-4"><div class="titlepage"><div><div><h2 class="title" style="clear: both">E.1. Release 12.4</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="release-12-4.html#id-1.11.6.5.4">E.1.1. Migration to Version 12.4</a></span></dt><dt><span class="sect2"><a href="release-12-4.html#id-1.11.6.5.5">E.1.2. Changes</a></span></dt></dl></div><p><strong>Release date: </strong>2020-08-13</p><p>
   This release contains a variety of fixes from 12.3.
   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-1.11.6.5.4"><div class="titlepage"><div><div><h3 class="title">E.1.1. Migration to Version 12.4</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-1.11.6.5.5"><div class="titlepage"><div><div><h3 class="title">E.1.2. Changes</h3></div></div></div><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
      Set a secure <code class="varname">search_path</code> in logical replication
      walsenders and apply workers (Noah Misch)
     </p><p>
      A malicious user of either the publisher or subscriber database
      could potentially cause execution of arbitrary SQL code by the role
      running replication, which is often a superuser.  Some of the risks
      here are equivalent to those described in CVE-2018-1058, and are
      mitigated in this patch by ensuring that the replication sender and
      receiver execute with empty <code class="varname">search_path</code> settings.
      (As with CVE-2018-1058, that change might cause problems for
      under-qualified names used in replicated tables' DDL.)  Other risks
      are inherent in replicating objects that belong to untrusted roles;
      the most we can do is document that there is a hazard to consider.
      (CVE-2020-14349)
     </p></li><li class="listitem"><p>
      Make contrib modules' installation scripts more secure (Tom Lane)
     </p><p>
      Attacks similar to those described in CVE-2018-1058 could be carried
      out against an extension installation script, if the attacker can
      create objects in either the extension's target schema or the schema
      of some prerequisite extension.  Since extensions often require
      superuser privilege to install, this can open a path to obtaining
      superuser privilege.  To mitigate this risk, be more careful about
      the <code class="varname">search_path</code> used to run an installation
      script; disable <code class="varname">check_function_bodies</code> within the
      script; and fix catalog-adjustment queries used in some contrib
      modules to ensure they are secure.  Also provide documentation to
      help third-party extension authors make their installation scripts
      secure.  This is not a complete solution; extensions that depend on
      other extensions can still be at risk if installed carelessly.
      (CVE-2020-14350)
     </p></li><li class="listitem"><p>
      Fix edge cases in partition pruning (Etsuro Fujita, Dmitry Dolgov)
     </p><p>
      When there are multiple partition key columns, generation of pruning
      tests could misbehave if some columns had no
      constraining <code class="literal">WHERE</code> clauses or multiple
      constraining clauses.  This could lead to server crashes, incorrect
      query results, or assertion failures.
     </p></li><li class="listitem"><p>
      Fix construction of parameterized BitmapAnd and BitmapOr index scans
      on the inside of partition-wise nestloop joins (Tom Lane)
     </p><p>
      A plan in which such a scan needed to use a value from the outside
      of the join would usually crash at execution.
     </p></li><li class="listitem"><p>
      Fix incorrect plan execution when a partitioned table is subject to
      both static and run-time partition pruning in the same query, and a
      new partition is added concurrently with the query (Amit Langote,
      Tom Lane)
     </p></li><li class="listitem"><p>
      In logical replication walsender, fix failure to send feedback
      messages after sending a keepalive message (Álvaro Herrera)
     </p><p>
      This is a relatively minor problem when using built-in logical
      replication, because the built-in walreceiver will send a feedback
      reply (which clears the incorrect state) fairly frequently anyway.
      But with some other replication systems, such
      as <span class="application">pglogical</span>, it causes significant
      performance issues.
     </p></li><li class="listitem"><p>
      Fix firing of column-specific <code class="literal">UPDATE</code> triggers in
      logical replication subscribers (Tom Lane)
     </p><p>
      The code neglected to account for the possibility of column numbers
      being different between the publisher and subscriber tables, so that
      if those were indeed different, wrong decisions might be made about
      which triggers to fire.
     </p></li><li class="listitem"><p>
      Update oldest xmin and LSN values
      during <code class="function">pg_replication_slot_advance()</code> (Michael
      Paquier)
     </p><p>
      This function previously failed to do that, possibly preventing
      resource cleanup (such as removal of no-longer-needed WAL segments)
      after manual advancement of a replication slot.
     </p></li><li class="listitem"><p>
      Fix slow execution of <code class="function">ts_headline()</code> (Tom Lane)
     </p><p>
      The phrase-search fix added in our previous set of minor releases
      could cause <code class="function">ts_headline()</code> to take unreasonable
      amounts of time for long documents; to make matters worse, the query
      was not cancellable within the troublesome loop.
     </p></li><li class="listitem"><p>
      Ensure the <code class="function">repeat()</code> function can be interrupted
      by query cancel (Joe Conway)
     </p></li><li class="listitem"><p>
      Fix <code class="function">pg_current_logfile()</code> to not include a
      carriage return (<code class="literal">\r</code>) in its result on Windows
      (Tom Lane)
     </p></li><li class="listitem"><p>
      Ensure that <code class="function">pg_read_file()</code> and related
      functions read until EOF is reached (Joe Conway)
     </p><p>
      Previously, if not given a specific data length to read, these
      functions would stop at whatever file length was reported
      by <code class="function">stat()</code>.  That's unhelpful for pipes and
      other sorts of virtual files.
     </p></li><li class="listitem"><p>
      Forbid numeric <code class="literal">NaN</code> values in <code class="type">jsonpath</code>
      computations (Alexander Korotkov)
     </p><p>
      Neither SQL nor JSON have the concept of <code class="literal">NaN</code>
      (not-a-number), but the <code class="type">jsonpath</code> code attempted to
      allow such values anyway.  This necessarily leads to nonstandard
      behavior, so it seems better to reject such values at the outset.
     </p></li><li class="listitem"><p>
      Handle single <code class="literal">Inf</code> or <code class="literal">NaN</code>
      inputs correctly in floating-point aggregates (Tom Lane)
     </p><p>
      The affected aggregates are
      <code class="function">corr()</code>,
      <code class="function">covar_pop()</code>,
      <code class="function">regr_intercept()</code>,
      <code class="function">regr_r2()</code>,
      <code class="function">regr_slope()</code>,
      <code class="function">regr_sxx()</code>,
      <code class="function">regr_sxy()</code>,
      <code class="function">regr_syy()</code>,
      <code class="function">stddev_pop()</code>, and
      <code class="function">var_pop()</code>.
      The correct answer in such cases is <code class="literal">NaN</code>, but an
      algorithmic change introduced in <span class="productname">PostgreSQL</span>
      v12 had caused these aggregates to produce zero instead.
     </p></li><li class="listitem"><p>
      Fix mis-handling of <code class="literal">NaN</code> inputs during parallel
      aggregation on <code class="type">numeric</code>-type columns (Tom Lane)
     </p><p>
      If some partial aggregation workers found only <code class="literal">NaN</code>s
      while others found only non-<code class="literal">NaN</code>s, the results
      were combined incorrectly, possibly leading to the wrong overall
      result (i.e., not <code class="literal">NaN</code> when it should be).
     </p></li><li class="listitem"><p>
      Reject time-of-day values greater than 24 hours (Tom Lane)
     </p><p>
      The intention of the datetime input code is to
      allow <span class="quote">“<span class="quote">24:00:00</span>”</span> or
      equivalently <span class="quote">“<span class="quote">23:59:60</span>”</span>, but no larger value.
      However, the range check was miscoded so that it would
      accept <span class="quote">“<span class="quote">23:59:60.<em class="replaceable"><code>nnn</code></em></span>”</span> with
      nonzero fractional-second <em class="replaceable"><code>nnn</code></em>.  In
      timestamp values this would result in wrapping into the first second
      of the next day.  In <code class="type">time</code> and <code class="type">timetz</code>
      values, the stored value would actually be more than 24 hours,
      causing dump/reload failures and possibly other misbehavior.
     </p></li><li class="listitem"><p>
      Undo double-quoting of index names in <code class="command">EXPLAIN</code>'s
      non-text output formats (Tom Lane, Euler Taveira)
     </p></li><li class="listitem"><p>
      Fix <code class="command">EXPLAIN</code>'s accounting for resource usage,
      particularly buffer accesses, in parallel workers in a plan
      using <code class="literal">Gather Merge</code> nodes
      (Jehan-Guillaume de Rorthais)
     </p></li><li class="listitem"><p>
      Fix timing of constraint revalidation in <code class="command">ALTER
      TABLE</code> (David Rowley)
     </p><p>
      If <code class="command">ALTER TABLE</code> needs to fully rewrite the table's
      contents (for example, due to change of a column's data type) and
      also needs to scan the table to re-validate foreign keys
      or <code class="literal">CHECK</code> constraints, it sometimes did things in
      the wrong order, leading to odd errors such as <span class="quote">“<span class="quote">could not read
      block 0 in file "base/nnnnn/nnnnn": read only 0 of 8192 bytes</span>”</span>.
     </p></li><li class="listitem"><p>
      Fix <code class="command">REINDEX CONCURRENTLY</code> to preserve the index's
      replication identity flag (Michael Paquier)
     </p><p>
      Previously, reindexing a table's replica identity index caused the
      setting to be lost, preventing old tuple values from being included
      in future logical-decoding output.
     </p></li><li class="listitem"><p>
      Work around incorrect not-null markings for
      <code class="structname">pg_subscription</code>.<code class="structfield">subslotname</code>
      and <code class="structname">pg_subscription_rel</code>.<code class="structfield">srsublsn</code>
      (Tom Lane)
     </p><p>
      The bootstrap catalog data incorrectly marks these two catalog
      columns as always non-null.  There's no easy way to correct that
      mistake in existing installations (though v13 and later will have
      the correct markings).  The main place that depends on that marking
      being correct is JIT-enabled tuple deconstruction, so teach it to
      explicitly ignore the marking for these two columns.  Also adjust
      some C code that accessed <code class="structfield">srsublsn</code> without
      checking to see if it's null; a crash from that is improbable but
      perhaps not impossible.
     </p></li><li class="listitem"><p>
      Cope with <code class="literal">LATERAL</code> references in restriction
      clauses attached to an un-flattened sub-<code class="literal">SELECT</code> in
      the <code class="literal">FROM</code> clause (Tom Lane)
     </p><p>
      This oversight could result in assertion failures or crashes at
      query execution.
     </p></li><li class="listitem"><p>
      Use the query-specified collation for operators invoked during
      selectivity estimation (Tom Lane)
     </p><p>
      Previously, the collation of the underlying database column was
      used.  But using the query's collation is arguably more correct.
      More importantly, now that we have nondeterministic collations,
      there are cases where an operator will fail outright if given a
      nondeterministic collation.  We don't want planning to fail in cases
      where the query itself would work, so this means that we must use
      the query's collation when invoking operators for estimation
      purposes.
     </p></li><li class="listitem"><p>
      Avoid believing that a never-analyzed foreign table has zero tuples
      (Tom Lane)
     </p><p>
      This primarily affected the planner's estimate of the number of
      groups that would be obtained by <code class="literal">GROUP BY</code>.
     </p></li><li class="listitem"><p>
      Remove bogus warning about <span class="quote">“<span class="quote">leftover placeholder tuple</span>”</span>
      in BRIN index de-summarization (Álvaro Herrera)
     </p><p>
      The case can occur legitimately after a cancelled vacuum, so warning
      about it is overly noisy.
     </p></li><li class="listitem"><p>
      Fix selection of tablespaces for <span class="quote">“<span class="quote">shared fileset</span>”</span>
      temporary files (Magnus Hagander, Tom Lane)
     </p><p>
      If <code class="varname">temp_tablespaces</code> is empty or explicitly names
      the database's primary tablespace, such files got placed into
      the <code class="literal">pg_default</code> tablespace rather than the
      database's primary tablespace as expected.
     </p></li><li class="listitem"><p>
      Fix corner-case error in masking of SP-GiST index pages during WAL
      consistency checking (Alexander Korotkov)
     </p><p>
      This could cause false failure reports
      when <code class="varname">wal_consistency_checking</code> is enabled.
     </p></li><li class="listitem"><p>
      Improve error handling in the server's <code class="filename">buffile</code>
      module (Thomas Munro)
     </p><p>
      Fix some cases where I/O errors were indistinguishable from reaching
      EOF, or were not reported at all.  Also add details such as block
      numbers and byte counts where appropriate.
     </p></li><li class="listitem"><p>
      Fix conflict-checking anomalies in <code class="literal">SERIALIZABLE</code>
      isolation mode (Peter Geoghegan)
     </p><p>
      If a concurrently-inserted tuple was updated by a different
      concurrent transaction, and neither tuple version was visible to the
      current transaction's snapshot, serialization conflict checking
      could draw the wrong conclusions about whether the tuple was relevant
      to the results of the current transaction.  This could allow a
      serializable transaction to commit when it should have failed with a
      serialization error.
     </p></li><li class="listitem"><p>
      Avoid repeated marking of dead btree index entries as dead (Masahiko
      Sawada)
     </p><p>
      While functionally harmless, this led to useless WAL traffic when
      checksums are enabled or <code class="varname">wal_log_hints</code> is on.
     </p></li><li class="listitem"><p>
      Fix checkpointer process to discard file sync requests
      when <code class="varname">fsync</code> is off (Heikki Linnakangas)
     </p><p>
      Such requests are treated as no-ops if <code class="varname">fsync</code> is
      off, but we forgot to remove them from the checkpointer's table of
      pending actions.  This would lead to bloat of that table, as well as
      possible assertion failures if <code class="varname">fsync</code> is later
      re-enabled.
     </p></li><li class="listitem"><p>
      Avoid trouble during cleanup of a non-exclusive backup when JIT
      compilation has been activated during the backup (Robert Haas)
     </p></li><li class="listitem"><p>
      Fix failure of some code paths to acquire the correct lock before
      modifying <code class="filename">pg_control</code> (Nathan Bossart, Fujii
      Masao)
     </p><p>
      This oversight could allow <code class="filename">pg_control</code> to be
      written out with an inconsistent checksum, possibly causing trouble
      later, including inability to restart the database if it crashed
      before the next <code class="filename">pg_control</code> update.
     </p></li><li class="listitem"><p>
      Fix errors in <code class="function">currtid()</code>
      and <code class="function">currtid2()</code> (Michael Paquier)
     </p><p>
      These functions (which are undocumented and used only by ancient
      versions of the ODBC driver) contained coding errors that could
      result in crashes, or in confusing error messages such as <span class="quote">“<span class="quote">could
      not open file</span>”</span> when applied to a relation having no storage.
     </p></li><li class="listitem"><p>
      Avoid calling <code class="function">elog()</code>
      or <code class="function">palloc()</code> while holding a spinlock (Michael
      Paquier, Tom Lane)
     </p><p>
      Logic associated with replication slots had several violations of
      this coding rule.  While the odds of trouble are quite low, an error
      in the called function would lead to a stuck spinlock.
     </p></li><li class="listitem"><p>
      Fix assertion in logical replication subscriber to allow use
      of <code class="literal">REPLICA IDENTITY FULL</code> (Euler Taveira)
     </p><p>
      This was just an incorrect assertion, so it has no impact on
      standard production builds.
     </p></li><li class="listitem"><p>
      Ensure that <span class="application">libpq</span> continues to try to
      read from the database connection socket after a write failure
      (Tom Lane)
     </p><p>
      This is important not only to ensure that we collect any final error
      message from a dying server process, but because we do not consider
      the connection lost until we see a read failure.  This oversight
      allowed <span class="application">libpq</span> to continue trying to
      send <code class="command">COPY</code> data indefinitely after a mid-transfer
      loss of connection, rather than reporting failure to the application.
     </p></li><li class="listitem"><p>
      Fix bugs in <span class="application">libpq</span>'s management of GSS
      encryption state (Tom Lane)
     </p><p>
      A connection using GSS encryption could freeze up when attempting to
      reset it after a server restart, or when moving on to the next one of
      a list of candidate servers.
     </p></li><li class="listitem"><p>
      Fix <span class="application">ecpg</span> crash with <code class="type">bytea</code>
      and cursor variables (Jehan-Guillaume de Rorthais)
     </p></li><li class="listitem"><p>
      Report out-of-disk-space errors properly
      in <span class="application">pg_dump</span>
      and <span class="application">pg_basebackup</span> (Justin Pryzby, Tom
      Lane, Álvaro Herrera)
     </p><p>
      Some code paths could produce silly reports like <span class="quote">“<span class="quote">could not
      write file: Success</span>”</span>.
     </p></li><li class="listitem"><p>
      Make <span class="application">pg_restore</span> cope with
      data-offset-less custom-format archive files when it needs to
      restore data items out of order (David Gilman, Tom Lane)
     </p><p>
      <span class="application">pg_dump</span> will produce such files if it
      cannot seek its output (for example, if the output is piped to
      something).  This fix primarily improves the ability to do a
      parallel restore from such a file.
     </p></li><li class="listitem"><p>
      Fix parallel restore of tables having both table-level privileges
      and per-column privileges (Tom Lane)
     </p><p>
      The table-level privilege grants have to be applied first, but a
      parallel restore did not reliably order them that way; this could
      lead to <span class="quote">“<span class="quote">tuple concurrently updated</span>”</span> errors, or to
      disappearance of some per-column privilege grants.  The fix for this
      is to include dependency links between such entries in the archive
      file, meaning that a new dump has to be taken with a
      corrected <span class="application">pg_dump</span> to ensure that the
      problem will not recur.
     </p></li><li class="listitem"><p>
      Ensure that <span class="application">pg_upgrade</span> runs
      with <code class="varname">vacuum_defer_cleanup_age</code> set to zero in the
      target cluster (Bruce Momjian)
     </p><p>
      If the target cluster's configuration has been modified to
      set <code class="varname">vacuum_defer_cleanup_age</code> to a nonzero value,
      that prevented freezing of the system catalogs from working properly,
      which caused the upgrade to fail in confusing ways.  Ensure that any
      such setting is overridden for the duration of the upgrade.
     </p></li><li class="listitem"><p>
      Fix <span class="application">pg_recvlogical</span> to drain pending
      messages before exiting (Noah Misch)
     </p><p>
      Without this, the replication sender might detect a send failure and
      exit without making the expected final update to the replication
      slot's LSN position.  That led to re-transmitting data after the
      next connection.  It was also possible to miss error messages sent
      after the last data that <span class="application">pg_recvlogical</span>
      wants to consume.
     </p></li><li class="listitem"><p>
      Fix <span class="application">pg_rewind</span>'s handling of just-deleted
      files in the source data directory (Justin Pryzby, Michael Paquier)
     </p><p>
      When working with an on-line source database, concurrent file
      deletions are possible, but <span class="application">pg_rewind</span>
      would get confused if deletion happened between seeing a file's
      directory entry and examining it with <code class="function">stat()</code>.
     </p></li><li class="listitem"><p>
      Make <span class="application">pg_test_fsync</span> use binary I/O mode on
      Windows (Michael Paquier)
     </p><p>
      Previously it wrote the test file in text mode, which is not an
      accurate reflection of <span class="productname">PostgreSQL</span>'s
      actual usage.
     </p></li><li class="listitem"><p>
      Fix <code class="filename">contrib/amcheck</code> to not complain about
      deleted index pages that are empty (Alexander Korotkov)
     </p><p>
      This state of affairs is normal during WAL replay.
     </p></li><li class="listitem"><p>
      Fix failure to initialize local state correctly
      in <code class="filename">contrib/dblink</code> (Joe Conway)
     </p><p>
      With the right combination of circumstances, this could lead to
      <code class="function">dblink_close()</code> issuing an unexpected
      remote <code class="command">COMMIT</code>.
     </p></li><li class="listitem"><p>
      Fix <code class="filename">contrib/pgcrypto</code>'s misuse
      of <code class="function">deflate()</code> (Tom Lane)
     </p><p>
      The <code class="function">pgp_sym_encrypt</code> functions could produce
      incorrect compressed data due to mishandling
      of <span class="application">zlib</span>'s API requirements.  We have no
      reports of this error manifesting with
      stock <span class="application">zlib</span>, but it can be seen when using
      IBM's <span class="application">zlibNX</span> implementation.
     </p></li><li class="listitem"><p>
      Fix corner case in decompression logic
      in <code class="filename">contrib/pgcrypto</code>'s
      <code class="function">pgp_sym_decrypt</code> functions (Kyotaro Horiguchi,
      Michael Paquier)
     </p><p>
      A compressed stream can validly end with an empty packet, but the
      decompressor failed to handle this and would complain about corrupt
      data.
     </p></li><li class="listitem"><p>
      Support building our NLS code with Microsoft Visual Studio 2015 or
      later (Juan José Santamaría Flecha, Davinder Singh,
      Amit Kapila)
     </p></li><li class="listitem"><p>
      Avoid possible failure of our MSVC install script when there is a
      file named <code class="filename">configure</code> several levels above the
      source code tree (Arnold Müller)
     </p><p>
      This could confuse some logic that looked
      for <code class="filename">configure</code> to identify the top level of the
      source tree.
     </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.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-12-3.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">Appendix E. Release Notes </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> E.2. Release 12.3</td></tr></table></div></body></html>