diff options
Diffstat (limited to 'doc/src/sgml/html/release-13-2.html')
| -rw-r--r-- | doc/src/sgml/html/release-13-2.html | 591 |
1 files changed, 591 insertions, 0 deletions
diff --git a/doc/src/sgml/html/release-13-2.html b/doc/src/sgml/html/release-13-2.html new file mode 100644 index 0000000..5b754e3 --- /dev/null +++ b/doc/src/sgml/html/release-13-2.html @@ -0,0 +1,591 @@ +<?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.3. Release 13.2</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-13-3.html" title="E.2. Release 13.3" /><link rel="next" href="release-13-1.html" title="E.4. Release 13.1" /></head><body id="docContent" class="container-fluid col-10"><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">E.3. Release 13.2</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="release-13-3.html" title="E.2. Release 13.3">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 13.4 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="release-13-1.html" title="E.4. Release 13.1">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="RELEASE-13-2"><div class="titlepage"><div><div><h2 class="title" style="clear: both">E.3. Release 13.2</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="release-13-2.html#id-1.11.6.7.4">E.3.1. Migration to Version 13.2</a></span></dt><dt><span class="sect2"><a href="release-13-2.html#id-1.11.6.7.5">E.3.2. Changes</a></span></dt></dl></div><p><strong>Release date: </strong>2021-02-11</p><p> + This release contains a variety of fixes from 13.1. + For information about new features in major release 13, see + <a class="xref" href="release-13.html" title="E.5. Release 13">Section E.5</a>. + </p><div class="sect2" id="id-1.11.6.7.4"><div class="titlepage"><div><div><h3 class="title">E.3.1. Migration to Version 13.2</h3></div></div></div><p> + A dump/restore is not required for those running 13.X. + </p><p> + However, see the first changelog item below concerning a possible + need to update stored views. Also see the third and fourth changelog + items, which describe cases in which reindexing indexes after the + upgrade may be advisable. + </p></div><div class="sect2" id="id-1.11.6.7.5"><div class="titlepage"><div><div><h3 class="title">E.3.2. Changes</h3></div></div></div><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p> + Fix failure to check per-column <code class="literal">SELECT</code> privileges + in some join queries (Tom Lane) + </p><p> + In some cases involving joins, the parser failed to record all the + columns read by a query in the column-usage bitmaps that are used + for permissions checking. Although the executor would still insist + on some sort of <code class="literal">SELECT</code> privilege to run the + query, this meant that a user having <code class="literal">SELECT</code> + privilege on only one column of a table could nonetheless read all + its columns through a suitably crafted query. + </p><p> + A stored view that is subject to this problem will have incomplete + column-usage bitmaps, and thus permissions will still not be + enforced properly on the view after updating. In installations that + depend on column-level permissions for security, it is recommended + to <code class="command">CREATE OR REPLACE</code> all user-defined views to + cause them to be re-parsed. + </p><p> + The <span class="productname">PostgreSQL</span> Project thanks + Sven Klemm for reporting this problem. + (CVE-2021-20229) + </p></li><li class="listitem"><p> + Fix information leakage in constraint-violation error messages + (Heikki Linnakangas) + </p><p> + If an <code class="command">UPDATE</code> command attempts to move a row to a + different partition but finds that it violates some constraint on + the new partition, and the columns in that partition are in + different physical positions than in the parent table, the error + message could reveal the contents of columns that the user does not + have <code class="literal">SELECT</code> privilege on. + (CVE-2021-3393) + </p></li><li class="listitem"><p> + Fix incorrect detection of concurrent page splits while inserting + into a GiST index (Heikki Linnakangas) + </p><p> + Concurrent insertions could lead to a corrupt index with entries + placed in the wrong pages. It's recommended to reindex any GiST + index that's been subject to concurrent insertions. + </p></li><li class="listitem"><p> + Fix <code class="command">CREATE INDEX CONCURRENTLY</code> to wait for + concurrent prepared transactions (Andrey Borodin) + </p><p> + At the point where <code class="command">CREATE INDEX CONCURRENTLY</code> + waits for all concurrent transactions to complete so that it can see + rows they inserted, it must also wait for all prepared transactions + to complete, for the same reason. Its failure to do so meant that + rows inserted by prepared transactions might be omitted from the new + index, causing queries relying on the index to miss such rows. + In installations that have enabled prepared transactions + (<code class="varname">max_prepared_transactions</code> > 0), + it's recommended to reindex any concurrently-built indexes in + case this problem occurred when they were built. + </p></li><li class="listitem"><p> + Avoid crash when trying to rescan an aggregation plan node + that has both hashed and sorted grouping sets + (Jeff Davis) + </p></li><li class="listitem"><p> + Fix possible incorrect query results when a hash aggregation node + spills some tuples to disk (Tom Lane) + </p><p> + It was possible for aggregation grouping values to be replaced + by nulls when the tuples are read back in, leading to wrong answers. + </p></li><li class="listitem"><p> + Fix edge case in incremental sort (Neil Chen) + </p><p> + If the last tuple of a sort batch chanced to be the first tuple of + the next group of already-sorted tuples, the code did the wrong + thing. This could lead to <span class="quote">“<span class="quote">retrieved too many tuples in a + bounded sort</span>”</span> error messages, or to silently-wrong sorting + results. + </p></li><li class="listitem"><p> + Avoid crash when a <code class="literal">CALL</code> or <code class="literal">DO</code> + statement that performs a transaction rollback is executed via + extended query protocol (Thomas Munro, Tom Lane) + </p><p> + In <span class="productname">PostgreSQL</span> 13, this case reliably + caused a null-pointer dereference. In earlier versions the bug + seems to have no visible symptoms, but it's not quite clear that + it could never cause a problem. + </p></li><li class="listitem"><p> + Avoid unnecessary errors with <code class="literal">BEFORE UPDATE</code> + triggers on partitioned tables (Álvaro Herrera) + </p><p> + A <code class="literal">BEFORE UPDATE FOR EACH ROW</code> trigger that + modified the row in any way prevented <code class="command">UPDATE</code> from + moving the row to another partition when needed; but there is no + longer any reason for this restriction. + </p></li><li class="listitem"><p> + Fix partition pruning logic to handle asymmetric hash partition sets + (Tom Lane) + </p><p> + If a hash-partitioned table has unequally-sized partitions (that is, + varying modulus values), or it lacks partitions for some remainder + values, then the planner's pruning logic could mistakenly conclude + that some partitions don't need to be scanned, leading to failure to + find rows that the query should find. + </p></li><li class="listitem"><p> + Avoid incorrect results when <code class="literal">WHERE CURRENT OF</code> is + applied to a cursor whose plan contains a MergeAppend node (Tom + Lane) + </p><p> + This case is unsupported (in general, a cursor using <code class="literal">ORDER + BY</code> is not guaranteed to be simply updatable); but the code + previously did not reject it, and could silently give false matches. + </p></li><li class="listitem"><p> + Fix crash when <code class="literal">WHERE CURRENT OF</code> is applied to a + cursor whose plan contains a custom scan node (David Geier) + </p></li><li class="listitem"><p> + Fix planner's mishandling of placeholders whose evaluation should be + delayed by an outer join (Tom Lane) + </p><p> + This occurs in particular with trivial subqueries containing lateral + references to outer-join outputs. The mistake could result in a + malformed plan. The known cases trigger a <span class="quote">“<span class="quote">failed to assign + all NestLoopParams to plan nodes</span>”</span> error, but other symptoms + may be possible. + </p></li><li class="listitem"><p> + Fix planner's handling of placeholders during removal of useless + RESULT RTEs (Tom Lane) + </p><p> + This oversight could lead to <span class="quote">“<span class="quote">no relation entry for + relid <em class="replaceable"><code>N</code></em></span>”</span> planner errors. + </p></li><li class="listitem"><p> + Fix planner's handling of a placeholder that is computed at some + join level and used only at that same level (Tom Lane) + </p><p> + This oversight could lead to <span class="quote">“<span class="quote">failed to build + any <em class="replaceable"><code>N</code></em>-way joins</span>”</span> planner errors. + </p></li><li class="listitem"><p> + Consider unsorted subpaths when planning a Gather Merge operation + (James Coleman) + </p><p> + It's possible to use such a path by adding an explicit Sort node, + and in some cases that gives rise to a superior plan. + </p></li><li class="listitem"><p> + Do not consider <code class="literal">ORDER BY</code> expressions involving + parallel-restricted functions or set-returning functions when trying + to parallelize sorts (James Coleman) + </p><p> + Such cases cannot safely be pushed into worker processes, but the + incremental sort feature accidentally made us consider them. + </p></li><li class="listitem"><p> + Be more careful about whether index AMs support mark/restore + (Andrew Gierth) + </p><p> + This prevents errors about missing support functions in rare edge + cases. + </p></li><li class="listitem"><p> + Fix overestimate of the amount of shared memory needed for parallel + queries (Takayuki Tsunakawa) + </p></li><li class="listitem"><p> + Fix <code class="command">ALTER DEFAULT PRIVILEGES</code> to handle duplicated + arguments safely (Michael Paquier) + </p><p> + Duplicate role or schema names within the same command could lead + to <span class="quote">“<span class="quote">tuple already updated by self</span>”</span> errors or + unique-constraint violations. + </p></li><li class="listitem"><p> + Flush ACL-related caches when <code class="structname">pg_authid</code> + changes (Noah Misch) + </p><p> + This change ensures that permissions-related decisions will promptly + reflect the results of <code class="command">ALTER ROLE ... [NO] INHERIT</code>. + </p></li><li class="listitem"><p> + Fix failure to detect <span class="quote">“<span class="quote">snapshot too old</span>”</span> conditions + in tables rewritten in the current transaction (Kyotaro Horiguchi, + Noah Misch) + </p><p> + This is only a hazard when <code class="varname">wal_level</code> is set + to <code class="literal">minimal</code> and the rewrite is performed + by <code class="command">ALTER TABLE SET TABLESPACE</code>. + </p></li><li class="listitem"><p> + Fix spurious failure of <code class="command">CREATE PUBLICATION</code> + when applied to a table created or rewritten in the current + transaction (Kyotaro Horiguchi) + </p><p> + This is only a hazard when <code class="varname">wal_level</code> is set + to <code class="literal">minimal</code>. + </p></li><li class="listitem"><p> + Prevent misprocessing of ambiguous <code class="command">CREATE TABLE + LIKE</code> clauses (Tom Lane) + </p><p> + A <code class="literal">LIKE</code> clause is re-examined after initial + creation of the new table, to handle importation of indexes and + such. It was possible for this re-examination to find a different + table of the same name, causing unexpected behavior; one example is + where the new table is a temporary table of the same name as + the <code class="literal">LIKE</code> target. + </p></li><li class="listitem"><p> + Rearrange order of operations in <code class="command">CREATE TABLE + LIKE</code> so that indexes are cloned before building foreign + key constraints (Tom Lane) + </p><p> + This fixes the case where a self-referential foreign key constraint + declared in the outer <code class="command">CREATE TABLE</code> depends on an + index that's coming from the <code class="literal">LIKE </code> clause. + </p></li><li class="listitem"><p> + Disallow <code class="command">CREATE STATISTICS</code> on system catalogs + (Tomas Vondra) + </p></li><li class="listitem"><p> + Disallow converting an inheritance child table to a view + (Tom Lane) + </p></li><li class="listitem"><p> + Ensure that disk space allocated for a dropped relation is released + promptly at commit (Thomas Munro) + </p><p> + Previously, if the dropped relation spanned multiple 1GB segments, + only the first segment was truncated immediately. Other segments + were simply unlinked, which doesn't authorize the kernel to release + the storage so long as any other backends still have the files open. + </p></li><li class="listitem"><p> + Prevent dropping a tablespace that is referenced by a partitioned + relation, but is not used for any actual storage (Álvaro + Herrera) + </p><p> + Previously this was allowed, but subsequent operations on the + partitioned relation would fail. + </p></li><li class="listitem"><p> + Fix progress reporting for <code class="command">CLUSTER</code> (Matthias van + de Meent) + </p></li><li class="listitem"><p> + Fix handling of backslash-escaped multibyte characters + in <code class="command">COPY FROM</code> (Heikki Linnakangas) + </p><p> + A backslash followed by a multibyte character was not handled + correctly. In some client character encodings, this could lead to + misinterpreting part of a multibyte character as a field separator + or end-of-copy-data marker. + </p></li><li class="listitem"><p> + Avoid preallocating executor hash tables + in <code class="command">EXPLAIN</code> without <code class="literal">ANALYZE</code> + (Alexey Bashtanov) + </p></li><li class="listitem"><p> + Fix recently-introduced race condition + in <code class="command">LISTEN</code>/<code class="command">NOTIFY</code> queue + handling (Tom Lane) + </p><p> + A newly-listening backend could attempt to read SLRU pages that + were in process of being truncated, possibly causing an error. + </p></li><li class="listitem"><p> + Allow the <code class="type">jsonb</code> concatenation operator to handle all + combinations of JSON data types (Tom Lane) + </p><p> + We can concatenate two JSON objects or two JSON arrays. Handle + other cases by wrapping non-array inputs in one-element arrays, + then performing an array concatenation. Previously, some + combinations of inputs followed this rule but others arbitrarily + threw an error. + </p></li><li class="listitem"><p> + Fix use of uninitialized value while parsing a <code class="literal">*</code> + quantifier in a BRE-mode regular expression (Tom Lane) + </p><p> + This error could cause the quantifier to act non-greedy, that is + behave like a <code class="literal">*?</code> quantifier would do in full + regular expressions. + </p></li><li class="listitem"><p> + Fix numeric <code class="function">power()</code> for the case where the + exponent is exactly <code class="literal">INT_MIN</code> (-2147483648) + (Dean Rasheed) + </p><p> + Previously, a result with no significant digits was produced. + </p></li><li class="listitem"><p> + Fix integer-overflow cases in <code class="function">substring()</code> + functions (Tom Lane, Pavel Stehule) + </p><p> + If the specified starting index and length overflow an integer when + added together, <code class="function">substring()</code> misbehaved, either + throwing a bogus <span class="quote">“<span class="quote">negative substring length</span>”</span> error for + a case that should succeed, or failing to complain that a negative + length is negative (and instead returning the whole string, in most + cases). + </p></li><li class="listitem"><p> + Prevent possible data loss from incorrect detection of the + wraparound point of an SLRU log + (Noah Misch) + </p><p> + The wraparound point typically falls in the middle of a page, which + must be rounded off to a page boundary, and that was not done + correctly. No issue could arise unless an installation had gotten + to within one page of SLRU overflow, which is unlikely in a + properly-functioning system. If this did happen, it would manifest + in later <span class="quote">“<span class="quote">apparent wraparound</span>”</span> or <span class="quote">“<span class="quote">could not + access status of transaction</span>”</span> errors. + </p></li><li class="listitem"><p> + Fix WAL-reading logic to handle timeline switches correctly (Kyotaro + Horiguchi, Fujii Masao) + </p><p> + Previously, if WAL archiving is enabled, a standby could fail to + follow a primary running on a newer timeline, with errors + like <span class="quote">“<span class="quote">requested WAL segment has already been removed</span>”</span>. + </p></li><li class="listitem"><p> + Fix memory leak in walsender processes while sending new snapshots + for logical decoding (Amit Kapila) + </p></li><li class="listitem"><p> + Fix relation cache leak in walsender processes while sending row + changes via the root of a partitioned relation during logical + replication (Amit Langote, Mark Zhao) + </p></li><li class="listitem"><p> + Fix walsender to accept additional commands after + terminating replication (Jeff Davis) + </p></li><li class="listitem"><p> + Ensure detection of deadlocks between hot standby backends and the + startup (WAL-application) process (Fujii Masao) + </p><p> + The startup process did not run the deadlock detection code, so that + in situations where the startup process is last to join a circular + wait situation, the deadlock might never be recognized. + </p></li><li class="listitem"><p> + Fix possible failure to detect recovery conflicts while deleting an + index entry that references a HOT chain (Peter Geoghegan) + </p><p> + The code failed to traverse the HOT chain and might thus compute a + too-old XID horizon, which could lead to incorrect conflict + processing in hot standby. The practical impact of this bug is + limited; in most cases the correct XID horizon would be found anyway + from nearby operations. + </p></li><li class="listitem"><p> + Ensure that a nonempty value + of <code class="varname">krb_server_keyfile</code> always overrides any + setting of <code class="varname">KRB5_KTNAME</code> in the server's + environment (Tom Lane) + </p><p> + Previously, which setting took precedence depended on whether the + client requests GSS encryption. + </p></li><li class="listitem"><p> + In server log messages about failing to match connections + to <code class="filename">pg_hba.conf</code> entries, include details about + whether GSS encryption has been activated (Kyotaro Horiguchi, Tom + Lane) + </p><p> + This is relevant data if <code class="literal">hostgssenc</code> + or <code class="literal">hostnogssenc</code> entries exist. + </p></li><li class="listitem"><p> + Fix assorted issues in server's support for GSS encryption (Tom Lane) + </p><p> + Remove pointless restriction that only GSS authentication + can be used on a GSS-encrypted connection. + Add GSS encryption information to connection-authorized log messages. + Include GSS-related space when computing the required size of + shared memory (this omission could have caused problems with very + high <code class="varname">max_connections</code> settings). + Avoid possible infinite recursion when reporting an unrecoverable + GSS encryption error. + </p></li><li class="listitem"><p> + Ensure that unserviced requests for background workers are cleaned + up when the postmaster begins a <span class="quote">“<span class="quote">smart</span>”</span> + or <span class="quote">“<span class="quote">fast</span>”</span> shutdown sequence (Tom Lane) + </p><p> + Previously, there was a race condition whereby a child process that + had requested a background worker just before shutdown could wait + indefinitely, preventing shutdown from completing. + </p></li><li class="listitem"><p> + Fix portability problem in parsing + of <code class="varname">recovery_target_xid</code> values (Michael Paquier) + </p><p> + The target XID is potentially 64 bits wide, but it was parsed + with <code class="function">strtoul()</code>, causing misbehavior on + platforms where <code class="type">long</code> is 32 bits (such as Windows). + </p></li><li class="listitem"><p> + Avoid trying to use parallel index build in a standalone backend + (Yulin Pei) + </p></li><li class="listitem"><p> + Allow index AMs to support included columns without necessarily + supporting multiple key columns (Tom Lane) + </p></li><li class="listitem"><p> + While taking a base backup, avoid executing any SHA256 code if a + backup manifest is not needed (Michael Paquier) + </p><p> + When using OpenSSL operating in FIPS mode, SHA256 hashing is + rejected, leading to an error. This change makes it possible to + take a base backup on such a platform, so long + as <code class="option">--no-manifest</code> is specified. + </p></li><li class="listitem"><p> + Avoid assertion failure during parallel aggregation of an + aggregate with a non-strict deserialization function (Andrew Gierth) + </p><p> + No such aggregate functions exist in + core <span class="productname">PostgreSQL</span>, but some extensions + such as PostGIS provide some. The mistake is harmless anyway in a + non-assert build. + </p></li><li class="listitem"><p> + Avoid assertion failure in <code class="function">pg_get_functiondef()</code> + when examining a function with a <code class="literal">TRANSFORM</code> option + (Tom Lane) + </p></li><li class="listitem"><p> + Fix data structure misallocation in + PL/pgSQL's <code class="command">CALL</code> statement (Tom Lane) + </p><p> + A <code class="command">CALL</code> in a PL/pgSQL procedure, to another + procedure that has OUT parameters, would fail if the called + procedure did a <code class="command">COMMIT</code> + or <code class="command">ROLLBACK</code>. + </p></li><li class="listitem"><p> + In <span class="application">libpq</span>, do not skip trying SSL after + GSS encryption (Tom Lane) + </p><p> + If we successfully made a GSS-encrypted connection, but then failed + during authentication, we would fall back to an unencrypted + connection rather than next trying an SSL-encrypted connection. + This could lead to unexpected connection failure, or to silently + getting an unencrypted connection where an encrypted one is + expected. Fortunately, GSS encryption could only succeed if both + client and server hold valid tickets in the same Kerberos + infrastructure. It seems unlikely for that to be true in an + environment that requires SSL encryption instead. + </p></li><li class="listitem"><p> + Make <span class="application">libpq</span>'s + <code class="function">PQconndefaults()</code> function report the correct + default value for <code class="varname">channel_binding</code> + (Daniele Varrazzo) + </p></li><li class="listitem"><p> + In <span class="application">psql</span>, re-allow including a password + in a <em class="replaceable"><code>connection_string</code></em> argument of a + <code class="command">\connect</code> command (Tom Lane) + </p><p> + This used to work, but a recent bug fix caused the password to be + ignored (resulting in prompting for a password). + </p></li><li class="listitem"><p> + In <span class="application">psql</span>'s <code class="command">\d</code> commands, + don't truncate the display of column default values (Tom Lane) + </p><p> + Formerly, they were arbitrarily truncated at 128 characters. + </p></li><li class="listitem"><p> + Fix assorted bugs + in <span class="application">psql</span>'s <code class="command">\help</code> + command (Kyotaro Horiguchi, Tom Lane) + </p><p> + <code class="command">\help</code> with two argument words failed to find a + command description using only the first word, for + example <code class="literal">\help reset all</code> should show the help + for <code class="command">RESET</code> but did not. + Also, <code class="command">\help</code> often failed to invoke the pager when + it should. It also leaked memory. + </p></li><li class="listitem"><p> + Fix <span class="application">pg_dump</span>'s dumping of inherited + generated columns (Peter Eisentraut) + </p><p> + The previous behavior resulted in (harmless) errors during restore. + </p></li><li class="listitem"><p> + In <span class="application">pg_dump</span>, ensure that the restore + script runs <code class="command">ALTER PUBLICATION ADD TABLE</code> commands + as the owner of the publication, and similarly runs <code class="command">ALTER + INDEX ATTACH PARTITION</code> commands as the owner of the + partitioned index (Tom Lane) + </p><p> + Previously, these commands would be run by the role that started the + restore script; which will usually work, but in corner cases that + role might not have adequate permissions. + </p></li><li class="listitem"><p> + Fix <span class="application">pg_dump</span> to handle <code class="literal">WITH + GRANT OPTION</code> in an extension's initial privileges + (Noah Misch) + </p><p> + If an extension's script creates an object and grants privileges + on it with grant option, then later the user revokes such + privileges, <span class="application">pg_dump</span> would generate + incorrect SQL for reproducing the situation. (Few if any extensions + do this today.) + </p></li><li class="listitem"><p> + In <span class="application">pg_rewind</span>, ensure that all WAL is + accounted for when rewinding a standby server + (Ian Barwick, Heikki Linnakangas) + </p></li><li class="listitem"><p> + In <span class="application">pgbench</span>, disallow a digit as the first + character of a variable name (Fabien Coelho) + </p><p> + This prevents trying to substitute variables into timestamp literal + values, which may contain strings like <code class="literal">12:34</code>. + </p></li><li class="listitem"><p> + Report the correct database name in connection failure error + messages from some client programs (Álvaro Herrera) + </p><p> + If the database name was defaulted rather than given on the command + line, <span class="application">pg_dumpall</span>, + <span class="application">pgbench</span>, <span class="application">oid2name</span>, + and <span class="application">vacuumlo</span> would produce misleading + error messages after a connection failure. + </p></li><li class="listitem"><p> + Fix memory leak in <code class="filename">contrib/auto_explain</code> + (Japin Li) + </p><p> + Memory consumed while producing the <code class="literal">EXPLAIN</code> + output was not freed until the end of the current transaction (for a + top-level statement) or the end of the surrounding statement (for a + nested statement). This was particularly a problem + with <code class="varname">log_nested_statements</code> enabled. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/postgres_fdw</code>, avoid leaking open + connections to remote servers when a user mapping or foreign server + object is dropped (Bharath Rupireddy) + </p><p> + Open connections that depend on a dropped user mapping or foreign + server can no longer be referenced, but formerly they were kept + around anyway for the duration of the local session. + </p></li><li class="listitem"><p> + Fix faulty assertion in <code class="filename">contrib/postgres_fdw</code> + (Etsuro Fujita) + </p></li><li class="listitem"><p> + In <code class="filename">contrib/pgcrypto</code>, check for error returns + from OpenSSL's EVP functions (Michael Paquier) + </p><p> + We do not really expect errors here, but this change silences + warnings from static analysis tools. + </p></li><li class="listitem"><p> + Make <code class="filename">contrib/pg_prewarm</code> more robust when the + cluster is shut down before prewarming is complete (Tom Lane) + </p><p> + Previously, autoprewarm would rewrite its status file with only the + block numbers that it had managed to load so far, thus perhaps + largely disabling the prewarm functionality in the next startup. + Instead, suppress status file updates until the initial loading pass + is complete. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/pg_trgm</code>'s GiST index support, avoid + crash in the rare case that picksplit is called on exactly two index + items (Andrew Gierth, Alexander Korotkov) + </p></li><li class="listitem"><p> + Fix miscalculation of timeouts + in <code class="filename">contrib/pg_prewarm</code> + and <code class="filename">contrib/postgres_fdw</code> + (Alexey Kondratov, Tom Lane) + </p><p> + The main loop in <code class="filename">contrib/pg_prewarm</code>'s + autoprewarm parent process underestimated its desired sleep time by + a factor of 1000, causing it to consume much more CPU than intended. + When waiting for a result from a remote + server, <code class="filename">contrib/postgres_fdw</code> overestimated the + desired timeout by a factor of 1000 (though this error had been + mitigated by imposing a clamp to 60 seconds). + </p><p> + Both of these errors stemmed from incorrectly converting + seconds-and-microseconds to milliseconds. Introduce a new + API <code class="function">TimestampDifferenceMilliseconds()</code> + to make it easier to get this right in the future. + </p></li><li class="listitem"><p> + Improve <span class="application">configure</span>'s heuristics for + selecting <code class="varname">PG_SYSROOT</code> on macOS (Tom Lane) + </p><p> + The new method is more likely to produce desirable results when + Xcode is newer than the underlying operating system. Choosing + a sysroot that does not match the OS version may result in + nonfunctional executables. + </p></li><li class="listitem"><p> + While building on macOS, specify <code class="option">-isysroot</code> in + link steps as well as compile steps (James Hilliard) + </p><p> + This likewise improves the results when Xcode is out of sync with + the operating system. + </p></li><li class="listitem"><p> + Fix JIT compilation to be compatible with LLVM 11 and LLVM 12 + (Andres Freund) + </p></li><li class="listitem"><p> + Fix potential mishandling of references to boolean variables in + JIT expression compilation (Andres Freund) + </p><p> + No field reports attributable to this have been seen, but it seems + likely that it could cause problems on some architectures. + </p></li><li class="listitem"><p> + Fix compile failure with ICU 68 and later (Tom Lane) + </p></li><li class="listitem"><p> + Avoid <code class="function">memcpy()</code> with a NULL source pointer and + zero count during partitioned index creation (Álvaro Herrera) + </p><p> + While such a call is not known to cause problems in itself, some + compilers assume that the arguments of <code class="function">memcpy()</code> + are never NULL, which could result in incorrect optimization of + nearby code. + </p></li><li class="listitem"><p> + Update time zone data files to <span class="application">tzdata</span> + release 2021a for DST law changes in Russia (Volgograd zone) and + South Sudan, plus historical corrections for Australia, Bahamas, + Belize, Bermuda, Ghana, Israel, Kenya, Nigeria, Palestine, + Seychelles, and Vanuatu. + </p><p> + Notably, the Australia/Currie zone has been corrected to the point + where it is identical to Australia/Hobart. + </p></li></ul></div></div></div><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navfooter"><hr></hr><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="release-13-3.html" title="E.2. Release 13.3">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="release.html" title="Appendix E. Release Notes">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="release-13-1.html" title="E.4. Release 13.1">Next</a></td></tr><tr><td width="40%" align="left" valign="top">E.2. Release 13.3 </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 13.4 Documentation">Home</a></td><td width="40%" align="right" valign="top"> E.4. Release 13.1</td></tr></table></div></body></html>
\ No newline at end of file |