diff options
Diffstat (limited to 'doc/src/sgml/html/release-13-3.html')
| -rw-r--r-- | doc/src/sgml/html/release-13-3.html | 338 |
1 files changed, 338 insertions, 0 deletions
diff --git a/doc/src/sgml/html/release-13-3.html b/doc/src/sgml/html/release-13-3.html new file mode 100644 index 0000000..b443934 --- /dev/null +++ b/doc/src/sgml/html/release-13-3.html @@ -0,0 +1,338 @@ +<?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 13.3</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-13-4.html" title="E.1. Release 13.4" /><link rel="next" href="release-13-2.html" title="E.3. Release 13.2" /></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.2. Release 13.3</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="release-13-4.html" title="E.1. Release 13.4">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 13.4 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="release-13-2.html" title="E.3. Release 13.2">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="RELEASE-13-3"><div class="titlepage"><div><div><h2 class="title" style="clear: both">E.2. Release 13.3</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="release-13-3.html#id-1.11.6.6.4">E.2.1. Migration to Version 13.3</a></span></dt><dt><span class="sect2"><a href="release-13-3.html#id-1.11.6.6.5">E.2.2. Changes</a></span></dt></dl></div><p><strong>Release date: </strong>2021-05-13</p><p> + This release contains a variety of fixes from 13.2. + 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-1.11.6.6.4"><div class="titlepage"><div><div><h3 class="title">E.2.1. Migration to Version 13.3</h3></div></div></div><p> + A dump/restore is not required for those running 13.X. + </p><p> + However, if you are upgrading from a version earlier than 13.2, + see <a class="xref" href="release-13-2.html" title="E.3. Release 13.2">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> + Prevent integer overflows in array subscripting calculations + (Tom Lane) + </p><p> + The array code previously did not complain about cases where an + array's lower bound plus length overflows an integer. This resulted + in later entries in the array becoming inaccessible (since their + subscripts could not be written as integers), but more importantly + it confused subsequent assignment operations. This could lead to + memory overwrites, with ensuing crashes or unwanted data + modifications. + (CVE-2021-32027) + </p></li><li class="listitem"><p> + Fix mishandling of <span class="quote">“<span class="quote">junk</span>”</span> columns in <code class="literal">INSERT + ... ON CONFLICT ... UPDATE</code> target lists (Tom Lane) + </p><p> + If the <code class="literal">UPDATE</code> list contains any multi-column + sub-selects (which give rise to junk columns in addition to the + results proper), the <code class="literal">UPDATE</code> path would end up + storing tuples that include the values of the extra junk columns. + That's fairly harmless in the short run, but if new columns are + added to the table then the values would become accessible, possibly + leading to malfunctions if they don't match the datatypes of the + added columns. + </p><p> + In addition, in versions supporting cross-partition updates, + a cross-partition update triggered by such a case had the reverse + problem: the junk columns were removed from the target list, + typically causing an immediate crash due to malfunction of the + multi-column sub-select mechanism. + (CVE-2021-32028) + </p></li><li class="listitem"><p> + Fix possibly-incorrect computation of <code class="command">UPDATE + ... RETURNING</code> outputs for joined cross-partition updates + (Amit Langote, Etsuro Fujita) + </p><p> + If an <code class="command">UPDATE</code> for a partitioned table caused a row + to be moved to another partition with a physically different row + type (for example, one with a different set of dropped columns), + computation of <code class="literal">RETURNING</code> results for that row + could produce errors or wrong answers. No error is observed unless + the <code class="command">UPDATE</code> involves other tables being joined to + the target table. + (CVE-2021-32029) + </p></li><li class="listitem"><p> + Fix adjustment of constraint deferrability properties in + partitioned tables (Álvaro Herrera) + </p><p> + When applied to a foreign-key constraint of a partitioned table, + <code class="command">ALTER TABLE ... ALTER CONSTRAINT</code> failed to + adjust the <code class="literal">DEFERRABLE</code> and/or + <code class="literal">INITIALLY DEFERRED</code> markings of the constraints + and triggers of leaf partitions. This led to unexpected behavior of + such constraints. After updating to this version, any misbehaving + partitioned tables can be fixed by executing a + new <code class="command">ALTER</code> command to set the desired properties. + </p><p> + This change also disallows applying such an <code class="command">ALTER</code> + directly to the constraints of leaf partitions. The only supported + case is for the whole partitioning hierarchy to have identical + constraint properties, so such <code class="command">ALTER</code>s must be + applied at the partition root. + </p></li><li class="listitem"><p> + When attaching a child table with <code class="command">ALTER TABLE + ... INHERIT</code>, insist that any generated columns in the + parent be generated the same way in the child (Peter Eisentraut) + </p></li><li class="listitem"><p> + Forbid marking an identity column as nullable (Vik Fearing) + </p><p> + <code class="literal">GENERATED ... AS IDENTITY</code> implies <code class="literal">NOT + NULL</code>, so don't allow it to be combined with an + explicit <code class="literal">NULL</code> specification. + </p></li><li class="listitem"><p> + Allow <code class="literal">ALTER ROLE/DATABASE ... SET</code> to set + the <code class="varname">role</code>, <code class="varname">session_authorization</code>, + and <code class="varname">temp_buffers</code> parameters (Tom Lane) + </p><p> + Previously, over-eager validity checks might reject these commands, + even if the values would have worked when used later. This created + a command ordering hazard for dump/reload and upgrade scenarios. + </p></li><li class="listitem"><p> + Ensure that <code class="command">REINDEX CONCURRENTLY</code> preserves any + statistics target that's been set for the index (Michael Paquier) + </p></li><li class="listitem"><p> + Fix <code class="command">COMMIT AND CHAIN</code> to work correctly when the + current transaction has live savepoints (Fujii Masao) + </p></li><li class="listitem"><p> + Fix list-manipulation bug in <code class="literal">WITH RECURSIVE</code> + processing (Michael Paquier, Tom Lane) + </p><p> + Sufficiently deep nesting of <code class="literal">WITH</code> constructs + (at least seven levels) triggered core dumps or incorrect complaints + of faulty <code class="literal">WITH</code> nesting. + </p></li><li class="listitem"><p> + Fix bug with coercing the result of a <code class="literal">COLLATE</code> + expression to a non-collatable type (Tom Lane) + </p><p> + This led to a parse tree in which the <code class="literal">COLLATE</code> + appears to be applied to a non-collatable value. While that + normally has no real impact (since <code class="literal">COLLATE</code> has no + effect at runtime), it was possible to construct views that would be + rejected during dump/reload. + </p></li><li class="listitem"><p> + Fix use-after-free bug in saving tuples for <code class="literal">AFTER</code> + triggers (Amit Langote) + </p><p> + This could cause crashes in some situations. + </p></li><li class="listitem"><p> + Disallow calling window functions and procedures via + the <span class="quote">“<span class="quote">fast path</span>”</span> wire protocol message (Tom Lane) + </p><p> + Only plain functions are supported here. While trying to call + an aggregate function failed already, calling a window function + would crash, and calling a procedure would work only if the + procedure did no transaction control. + </p></li><li class="listitem"><p> + Extend <code class="function">pg_identify_object_as_address()</code> + to support event triggers (Joel Jacobson) + </p></li><li class="listitem"><p> + Fix <code class="function">to_char()</code>'s handling of Roman-numeral month + format codes with negative intervals (Julien Rouhaud) + </p><p> + Previously, such cases would usually cause a crash. + </p></li><li class="listitem"><p> + Check that the argument + of <code class="function">pg_import_system_collations()</code> is a valid + schema OID (Tom Lane) + </p></li><li class="listitem"><p> + Fix use of uninitialized value while parsing an + <code class="literal">\{<em class="replaceable"><code>m</code></em>,<em class="replaceable"><code>n</code></em>\}</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 an + <code class="literal">{<em class="replaceable"><code>m</code></em>,<em class="replaceable"><code>n</code></em>}?</code> + quantifier would do in full regular expressions. + </p></li><li class="listitem"><p> + Fix <span class="quote">“<span class="quote">could not find pathkey item to sort</span>”</span> planner + errors in some situations where the sort key involves an aggregate + or window function (James Coleman, Tom Lane) + </p></li><li class="listitem"><p> + Don't ignore system columns when estimating the number of groups + using extended statistics (Tomas Vondra) + </p><p> + This led to strange estimates for queries such as <code class="literal">SELECT + ... GROUP BY a, b, ctid</code>. + </p></li><li class="listitem"><p> + Avoid divide-by-zero when estimating selectivity of a regular + expression with a very long fixed prefix (Tom Lane) + </p><p> + This typically led to a <code class="literal">NaN</code> selectivity value, + causing assertion failures or strange planner behavior. + </p></li><li class="listitem"><p> + Fix access-off-the-end-of-the-table error in BRIN index bitmap scans + (Tomas Vondra) + </p><p> + If the page range size used by a BRIN index isn't a power of two, + there were corner cases in which a bitmap scan could try to fetch + pages past the actual end of the table, leading to <span class="quote">“<span class="quote">could not + open file</span>”</span> errors. + </p></li><li class="listitem"><p> + Fix potentially wrong answers from GIN <code class="type">tsvector</code> index + searches, when there are many matching tuples (Tom Lane) + </p><p> + If the number of index matches became large enough to make the + bitmap holding them become lossy (a threshold that depends + on <code class="varname">work_mem</code>), the code could get confused about + whether rechecks are required, allowing rows to be returned that + don't actually match the query. + </p></li><li class="listitem"><p> + Fix concurrency issues with WAL segment recycling on Windows + (Michael Paquier) + </p><p> + This reverts a change that caused intermittent <span class="quote">“<span class="quote">could not + rename file ...: Permission denied</span>”</span> log messages. While + there were not serious consequences, the log spam was annoying. + </p></li><li class="listitem"><p> + Avoid incorrect timeline change while recovering uncommitted + two-phase transactions from WAL (Soumyadeep Chakraborty, Jimmy Yih, + Kevin Yeap) + </p><p> + This error could lead to subsequent WAL records being written under + the wrong timeline ID, leading to consistency problems, or even + complete failure to be able to restart the server, later on. + </p></li><li class="listitem"><p> + Ensure that locks are released while shutting down a standby + server's startup process (Fujii Masao) + </p><p> + When a standby server is shut down while still in recovery, some + locks might be left held. This causes assertion failures in debug + builds; it's unclear whether any serious consequence could occur + in production builds. + </p></li><li class="listitem"><p> + Fix crash when a logical replication worker does <code class="command">ALTER + SUBSCRIPTION REFRESH</code> (Peter Smith) + </p><p> + The core code won't do this, but a replica trigger could. + </p></li><li class="listitem"><p> + Ensure we default to <code class="varname">wal_sync_method</code> + = <code class="literal">fdatasync</code> on recent FreeBSD (Thomas Munro) + </p><p> + FreeBSD 13 supports <code class="literal">open_datasync</code>, which would + normally become the default choice. However, it's unclear whether + that is actually an improvement for Postgres, so preserve the + existing default for now. + </p></li><li class="listitem"><p> + Disable the <code class="varname">vacuum_cleanup_index_scale_factor</code> + parameter and storage option (Peter Geoghegan) + </p><p> + The notion of tracking <span class="quote">“<span class="quote">stale</span>”</span> index statistics proved + to interact badly with + the <code class="varname">autovacuum_vacuum_insert_threshold</code> parameter, + resulting in unnecessary full-index scans and consequent degradation + of autovacuum performance. The latter mechanism seems superior, so + remove the stale-statistics logic. The control parameter for that, + <code class="varname">vacuum_cleanup_index_scale_factor</code>, will be + removed entirely in v14. In v13, it remains present to avoid + breaking existing configuration files, but it no longer does + anything. + </p></li><li class="listitem"><p> + Pass the correct trigger OID to object post-alter hooks + during <code class="command">ALTER CONSTRAINT</code> (Álvaro Herrera) + </p><p> + When updating trigger properties during <code class="command">ALTER + CONSTRAINT</code>, the post-alter hook was told that we are + updating a trigger, but the constraint's OID was passed instead of + the trigger's. + </p></li><li class="listitem"><p> + Ensure we finish cleaning up when interrupted while detaching a DSM + segment (Thomas Munro) + </p><p> + This error could result in temporary files not being cleaned up + promptly after a parallel query. + </p></li><li class="listitem"><p> + Fix assorted minor memory leaks in the server (Tom Lane, Andres Freund) + </p></li><li class="listitem"><p> + Fix uninitialized variable in walreceiver's statistics in shared + memory (Fujii Masao) + </p><p> + This error was harmless on most platforms, but could cause issues + on platforms lacking atomic variables and/or spinlock support. + </p></li><li class="listitem"><p> + Reduce the overhead of dtrace probes for LWLock operations, when + dtrace support is compiled in but not active (Peter Eisentraut) + </p></li><li class="listitem"><p> + Fix failure when a PL/pgSQL <code class="command">DO</code> block makes use of + both composite-type variables and transaction control (Tom Lane) + </p><p> + Previously, such cases led to errors about leaked tuple descriptors. + </p></li><li class="listitem"><p> + Prevent infinite loop in <span class="application">libpq</span> + if a ParameterDescription message with a corrupt length is received + (Tom Lane) + </p></li><li class="listitem"><p> + When <span class="application">initdb</span> prints instructions about how + to start the server, make the path shown + for <span class="application">pg_ctl</span> use backslash separators on + Windows (Nitin Jadhav) + </p></li><li class="listitem"><p> + Fix <span class="application">psql</span> to restore the previous behavior + of <code class="literal">\connect + service=<em class="replaceable"><code>something</code></em></code> (Tom Lane) + </p><p> + A previous bug fix caused environment variables (such + as <code class="varname">PGPORT</code>) to override entries in the service + file in this context. Restore the previous behavior, in which the + priority is the other way around. + </p></li><li class="listitem"><p> + Fix <span class="application">psql</span>'s <code class="literal">ON_ERROR_ROLLBACK</code> + feature to handle <code class="literal">COMMIT AND CHAIN</code> commands + correctly (Arthur Nascimento) + </p><p> + Previously, this case failed with <span class="quote">“<span class="quote">savepoint + "pg_psql_temporary_savepoint" does not exist</span>”</span>. + </p></li><li class="listitem"><p> + In <span class="application">psql</span>, avoid repeated <span class="quote">“<span class="quote">could not + print result table</span>”</span> failures after the first such error + (Álvaro Herrera) + </p></li><li class="listitem"><p> + Fix race condition in detection of file modification by + <span class="application">psql</span>'s <code class="literal">\e</code> and related + commands (Laurenz Albe) + </p><p> + A very fast typist could fool the code's file-timestamp-based + detection of whether the temporary edit file was changed. + </p></li><li class="listitem"><p> + Fix <span class="application">pg_dump</span>'s dumping of + generated columns in partitioned tables (Peter Eisentraut) + </p><p> + A fix introduced in the previous minor release should not be applied + to partitioned tables, only traditionally-inherited tables. + </p></li><li class="listitem"><p> + Fix missed file version check + in <span class="application">pg_restore</span> (Tom Lane) + </p><p> + When reading a custom-format archive from a non-seekable source, + <span class="application">pg_restore</span> neglected to check the + archive version. If it was fed a newer archive version than it + can support, it would fail messily later on. + </p></li><li class="listitem"><p> + Add some more checks to <span class="application">pg_upgrade</span> for + user tables containing non-upgradable data types (Tom Lane) + </p><p> + Fix detection of some cases where a non-upgradable data type is + embedded within a container type (such as an array or range). + Also disallow upgrading when user tables contain columns of + system-defined composite types, since those types' OIDs are not + stable across versions. + </p></li><li class="listitem"><p> + Fix incorrect progress-reporting calculation + in <span class="application">pg_checksums</span> (Shinya Kato) + </p></li><li class="listitem"><p> + Fix <span class="application">pg_waldump</span> to + count <code class="literal">XACT</code> records correctly when generating + per-record statistics (Kyotaro Horiguchi) + </p></li><li class="listitem"><p> + Fix <code class="filename">contrib/amcheck</code> to not complain about the + tuple flags <code class="literal">HEAP_XMAX_LOCK_ONLY</code> + and <code class="literal">HEAP_KEYS_UPDATED</code> both being set + (Julien Rouhaud) + </p><p> + This is a valid state after <code class="literal">SELECT FOR UPDATE</code>. + </p></li><li class="listitem"><p> + Adjust VPATH build rules to support recent Oracle Developer Studio + compiler versions (Noah Misch) + </p></li><li class="listitem"><p> + Fix testing of PL/Python for Python 3 on Solaris (Noah Misch) + </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-13-4.html" title="E.1. Release 13.4">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-13-2.html" title="E.3. Release 13.2">Next</a></td></tr><tr><td width="40%" align="left" valign="top">E.1. Release 13.4 </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.3. Release 13.2</td></tr></table></div></body></html>
\ No newline at end of file |