diff options
Diffstat (limited to 'doc/src/sgml/html/release-14-3.html')
| -rw-r--r-- | doc/src/sgml/html/release-14-3.html | 429 |
1 files changed, 429 insertions, 0 deletions
diff --git a/doc/src/sgml/html/release-14-3.html b/doc/src/sgml/html/release-14-3.html new file mode 100644 index 0000000..a64dd99 --- /dev/null +++ b/doc/src/sgml/html/release-14-3.html @@ -0,0 +1,429 @@ +<?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 14.3</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 Vsnapshot" /><link rel="prev" href="release-14-4.html" title="E.2. Release 14.4" /><link rel="next" href="release-14-2.html" title="E.4. Release 14.2" /></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.3. Release 14.3</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="release-14-4.html" title="E.2. Release 14.4">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 14.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="release-14-2.html" title="E.4. Release 14.2">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="RELEASE-14-3"><div class="titlepage"><div><div><h2 class="title" style="clear: both">E.3. Release 14.3</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="release-14-3.html#id-1.11.6.7.4">E.3.1. Migration to Version 14.3</a></span></dt><dt><span class="sect2"><a href="release-14-3.html#id-1.11.6.7.5">E.3.2. Changes</a></span></dt></dl></div><p><strong>Release date: </strong>2022-05-12</p><p> + This release contains a variety of fixes from 14.2. + For information about new features in major release 14, see + <a class="xref" href="release-14.html" title="E.6. Release 14">Section E.6</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 14.3</h3></div></div></div><p> + A dump/restore is not required for those running 14.X. + </p><p> + However, if you have any GiST indexes on columns of type + <code class="type">ltree</code> (supplied by the <code class="filename">contrib/ltree</code> + extension), you should re-index them after updating. + See the second changelog entry below. + </p><p> + Also, if you are upgrading from a version earlier than 14.2, + see <a class="xref" href="release-14-2.html" title="E.4. Release 14.2">Section E.4</a>. + </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> + Confine additional operations within <span class="quote">“<span class="quote">security restricted + operation</span>”</span> sandboxes (Sergey Shinderuk, Noah Misch) + </p><p> + Autovacuum, <code class="command">CLUSTER</code>, <code class="command">CREATE + INDEX</code>, <code class="command">REINDEX</code>, <code class="command">REFRESH + MATERIALIZED VIEW</code>, + and <span class="application">pg_amcheck</span> activated + the <span class="quote">“<span class="quote">security restricted operation</span>”</span> protection + mechanism too late, or even not at all in some code paths. + A user having permission to create non-temporary objects within a + database could define an object that would execute arbitrary SQL + code with superuser permissions the next time that autovacuum + processed the object, or that some superuser ran one of the affected + commands against it. + </p><p> + The <span class="productname">PostgreSQL</span> Project thanks + Alexander Lakhin for reporting this problem. + (CVE-2022-1552) + </p></li><li class="listitem"><p> + Fix default signature length for <code class="literal">gist_ltree_ops</code> + indexes (Tomas Vondra, Alexander Korotkov) + </p><p> + The default signature length (hash size) for GiST indexes + on <code class="type">ltree</code> columns was accidentally changed while + upgrading that operator class to support operator class parameters. + If any operations had been done on such an index without first + upgrading the <code class="literal">ltree</code> extension to version 1.2, + they were done assuming that the signature length was 28 bytes + rather than the intended 8. This means it is very likely that such + indexes are now corrupt. For safety we recommend re-indexing all + GiST indexes on <code class="type">ltree</code> columns after installing this + update. (Note that GiST indexes on <code class="type">ltree[]</code> + columns, that is arrays of <code class="type">ltree</code>, are not affected.) + </p></li><li class="listitem"><p> + Stop using query-provided column aliases for the columns of + whole-row variables that refer to plain tables (Tom Lane) + </p><p> + The column names in tuples produced by a whole-row variable (such + as <code class="literal">tbl.*</code> in contexts other than the top level of + a <code class="literal">SELECT</code> list) are now always those of the + associated named composite type, if there is one. We'd previously + attempted to make them track any column aliases that had been + applied to the <code class="literal">FROM</code> entry the variable refers to. + But that's semantically dubious, because really then the output of + the variable is not at all of the composite type it claims to be. + Previous attempts to deal with that inconsistency had bad results up + to and including storing unreadable data on disk, so just give up on + the whole idea. + </p><p> + In cases where it's important to be able to relabel such columns, + a workaround is to introduce an extra level of + sub-<code class="literal">SELECT</code>, so that the whole-row variable is + referring to the sub-<code class="literal">SELECT</code>'s output and not to a + plain table. Then the variable is of type <code class="type">record</code> + to begin with and there's no issue. + </p></li><li class="listitem"><p> + Fix incorrect roundoff when extracting epoch values from intervals + (Peter Eisentraut) + </p><p> + The new <code class="type">numeric</code>-based code + for <code class="function">EXTRACT()</code> failed to yield results + equivalent to the old <code class="type">float</code>-based code, as a result of + accidentally truncating the <code class="literal">DAYS_PER_YEAR</code> value + to an integer. + </p></li><li class="listitem"><p> + Defend against <code class="literal">pg_stat_get_replication_slot(NULL)</code> + (Andres Freund) + </p><p> + This function should be marked strict in the catalog data, but it + was not in v14, so add a run-time check instead. + </p></li><li class="listitem"><p> + Fix incorrect output for types <code class="type">timestamptz</code> + and <code class="type">timetz</code> in <code class="function">table_to_xmlschema()</code> + and allied functions (Renan Soares Lopes) + </p><p> + The xmlschema output for these types included a malformed regular + expression. + </p></li><li class="listitem"><p> + Avoid core dump in parser for a <code class="literal">VALUES</code> clause with + zero columns (Tom Lane) + </p></li><li class="listitem"><p> + Fix planner failure when a Result plan node appears immediately + underneath an Append node (Etsuro Fujita) + </p><p> + Recently-added code to support asynchronous remote queries failed to + handle this case, leading to crashes or errors about unrecognized + node types. + </p></li><li class="listitem"><p> + Fix planner failure if a query using <code class="literal">SEARCH</code> + or <code class="literal">CYCLE</code> features contains a duplicate CTE name + (Tom Lane, Kyotaro Horiguchi) + </p><p> + When the name of the recursive <code class="literal">WITH</code> query is + re-used within itself, the planner could crash or report odd errors + such as <span class="quote">“<span class="quote">could not find attribute 2 in subquery + targetlist</span>”</span>. + </p></li><li class="listitem"><p> + Fix planner errors for <code class="function">GROUPING()</code> constructs + that reference outer query levels (Richard Guo, Tom Lane) + </p></li><li class="listitem"><p> + Fix plan generation for index-only scans on indexes with + both returnable and non-returnable columns (Tom Lane) + </p><p> + The previous coding could try to read non-returnable columns + in addition to the returnable ones. This was fairly harmless + because it didn't actually do anything with the bogus values, + but it fell foul of a recently-added error check that rejected + such a plan. + </p></li><li class="listitem"><p> + Avoid accessing a no-longer-pinned shared buffer while attempting + to lock an outdated tuple during EvalPlanQual (Tom Lane) + </p><p> + The code would touch the buffer a couple more times after releasing + its pin. In theory another process could recycle the buffer (or + more likely, try to defragment its free space) as soon as the pin is + gone, probably leading to failure to find the newer version of the + tuple. + </p></li><li class="listitem"><p> + Fix query-lifespan memory leak in an IndexScan node that is + performing reordering (Aliaksandr Kalenik) + </p></li><li class="listitem"><p> + Fix <code class="command">ALTER FUNCTION</code> to support changing a + function's parallelism property and + its <code class="literal">SET</code>-variable list in the same command (Tom + Lane) + </p><p> + The parallelism property change was lost if the same command also + updated the function's <code class="literal">SET</code> clause. + </p></li><li class="listitem"><p> + Tighten lookup of the index <span class="quote">“<span class="quote">owned by</span>”</span> a constraint + (Tom Lane, Japin Li) + </p><p> + Some code paths mistook the index depended on by a foreign key + constraint for one owned by a unique or primary key constraint, + resulting in odd errors during certain <code class="command">ALTER + TABLE</code> operations on tables having foreign key constraints. + </p></li><li class="listitem"><p> + Fix bogus errors from attempts to alter system columns of tables + (Tom Lane) + </p><p> + The system should just tell you that you can't do it, but sometimes + it would report <span class="quote">“<span class="quote">no owned sequence found</span>”</span> instead. + </p></li><li class="listitem"><p> + Fix mis-sorting of table rows when <code class="command">CLUSTER</code>ing + using an index whose leading key is an expression (Peter Geoghegan, + Thomas Munro) + </p><p> + The table would be rebuilt with the correct data, but in an order + having little to do with the index order. + </p></li><li class="listitem"><p> + Prevent data loss if a system crash occurs shortly after a sorted + GiST index build (Heikki Linnakangas) + </p><p> + The code path for building GiST indexes using sorting neglected + to <code class="function">fsync</code> the file upon completion. This could + result in a corrupted index if the operating system crashed shortly + later. + </p></li><li class="listitem"><p> + Fix risk of deadlock failures while dropping a partitioned index + (Jimmy Yih, Gaurab Dey, Tom Lane) + </p><p> + Ensure that the required table and index locks are taken in the + standard order (parents before children, tables before indexes). + The previous coding for <code class="command">DROP INDEX</code> did it + differently, and so could deadlock against concurrent queries taking + these locks in the standard order. + </p></li><li class="listitem"><p> + Fix race condition between <code class="command">DROP TABLESPACE</code> and + checkpointing (Nathan Bossart) + </p><p> + The checkpoint forced by <code class="command">DROP TABLESPACE</code> could + sometimes fail to remove all dead files from the tablespace's + directory, leading to a bogus <span class="quote">“<span class="quote">tablespace is not empty</span>”</span> + error. + </p></li><li class="listitem"><p> + Fix possible trouble in crash recovery after + a <code class="command">TRUNCATE</code> command that overlaps a checkpoint + (Kyotaro Horiguchi, Heikki Linnakangas, Robert Haas) + </p><p> + <code class="command">TRUNCATE</code> must ensure that the table's disk file + is truncated before the checkpoint is allowed to complete. + Otherwise, replay starting from that checkpoint might find + unexpected data in the supposedly-removed pages, possibly causing + replay failure. + </p></li><li class="listitem"><p> + Fix unsafe toast-data accesses during temporary object cleanup + (Andres Freund) + </p><p> + Temporary-object deletion during server process exit could fail + with <span class="quote">“<span class="quote">FATAL: cannot fetch toast data without an active + snapshot</span>”</span>. This was usually harmless since the next use of + that temporary schema would clean up successfully. + </p></li><li class="listitem"><p> + Re-allow underscore as the first character in a custom parameter + name (Japin Li) + </p><p> + Such names were unintentionally disallowed in v14. + </p></li><li class="listitem"><p> + Add <code class="literal">regress</code> option for + the <code class="varname">compute_query_id</code> parameter (Michael Paquier) + </p><p> + This is intended to facilitate testing, by allowing query IDs to be + computed but not shown in <code class="command">EXPLAIN</code> output. + </p></li><li class="listitem"><p> + Improve wait logic in RegisterSyncRequest (Thomas Munro) + </p><p> + If we run out of space in the checkpointer sync request queue (which + is hopefully rare on real systems, but is common when testing with a + very small buffer pool), we wait for it to drain. While waiting, we + should report that as a wait event so that users know what is going + on, and also watch for postmaster death, since otherwise the loop might + never terminate if the checkpointer has already exited. + </p></li><li class="listitem"><p> + Wake up for latch events when the checkpointer is waiting between + writes (Thomas Munro) + </p><p> + This improves responsiveness to backends sending sync requests. + The change also creates a proper wait event class for these waits. + </p></li><li class="listitem"><p> + Fix <span class="quote">“<span class="quote">PANIC: xlog flush request is not satisfied</span>”</span> + failure during standby promotion when there is a missing WAL + continuation record (Sami Imseih) + </p></li><li class="listitem"><p> + Fix possibility of self-deadlock in hot standby conflict handling + (Andres Freund) + </p><p> + With unlucky timing, the WAL-applying process could get stuck + while waiting for some other process to release a buffer lock. + </p></li><li class="listitem"><p> + Fix possible mis-identification of the correct ancestor relation + to publish logical replication changes through (Tomas Vondra, Hou + zj, Amit Kapila) + </p><p> + If <code class="varname">publish_via_partition_root</code> is enabled, and + there are multiple publications naming different ancestors of the + currently-modified relation, the wrong ancestor might be chosen for + reporting the change. + </p></li><li class="listitem"><p> + Ensure that logical replication apply workers can be restarted even + when we're up against + the <code class="varname">max_sync_workers_per_subscription</code> limit + (Amit Kapila) + </p><p> + Faulty coding of the limit check caused a restarted worker to exit + immediately, leaving fewer workers than there should be. + </p></li><li class="listitem"><p> + Include unchanged replica identity key columns in the WAL log for an + update, if they are stored out-of-line (Dilip Kumar, Amit Kapila) + </p><p> + Otherwise subscribers cannot see the values and will fail to + replicate the update. + </p></li><li class="listitem"><p> + Cope correctly with platforms that have no support for altering the + server process's display in <span class="application">ps(1)</span> (Andrew + Dunstan) + </p><p> + Few platforms are like this (the only supported one is Cygwin), so + we'd managed not to notice that refactoring introduced a potential + memory clobber. + </p></li><li class="listitem"><p> + Make the server more robust against missed timer interrupts (Michael + Harris, Tom Lane) + </p><p> + An optimization added in v14 meant that if a server process somehow + missed a timer interrupt, it would never again ask the kernel for + another one, thus breaking timeout detection for the remainder of the + session. This seems unduly fragile, so add a recovery path. + </p></li><li class="listitem"><p> + Disallow execution of SPI functions during PL/Perl function + compilation (Tom Lane) + </p><p> + Perl can be convinced to execute user-defined code during compilation + of a PL/Perl function. However, it's not okay for such code to try + to invoke SQL operations via SPI. That results in a crash, and if + it didn't crash it would be a security hazard, because we really + don't want code execution during function validation. Put in a + check to give a friendlier error message instead. + </p></li><li class="listitem"><p> + Make <span class="application">libpq</span> accept root-owned SSL private + key files (David Steele) + </p><p> + This change synchronizes <span class="application">libpq</span>'s rules + for safe ownership and permissions of SSL key files with the rules + the server has used since release 9.6. Namely, in addition to the + current rules, allow the case where the key file is owned by root + and has permissions <code class="literal">rw-r-----</code> or less. This is + helpful for system-wide management of key files. + </p></li><li class="listitem"><p> + Fix behavior of <span class="application">libpq</span>'s + <code class="function">PQisBusy()</code> function after a connection failure + (Tom Lane) + </p><p> + If we'd detected a write failure, <code class="function">PQisBusy()</code> + would always return true, which is the wrong thing: we want input + processing to carry on normally until we've read whatever is + available from the server. The practical effect of this error is + that applications using <span class="application">libpq</span>'s + async-query API would typically detect connection loss only + when <code class="function">PQconsumeInput()</code> returns a hard failure. + With this fix, a connection loss will normally be reported via an + error <code class="structname">PGresult</code> object, which is a much + cleaner behavior for most applications. + </p></li><li class="listitem"><p> + Re-allow <em class="replaceable"><code>database</code></em>.<em class="replaceable"><code>schema</code></em>.<em class="replaceable"><code>table</code></em> + patterns + in <span class="application">psql</span>, <span class="application">pg_dump</span>, + and <span class="application">pg_amcheck</span> (Mark Dilger) + </p><p> + Versions before v14 silently ignored all but + the <em class="replaceable"><code>schema</code></em> + and <em class="replaceable"><code>table</code></em> fragments of a pattern + containing more than one dot. Refactoring in v14 accidentally broke + that use-case. Reinstate it, but now complain if the first fragment + is not the name of the current database. + </p></li><li class="listitem"><p> + Make <span class="application">pg_ctl</span> recheck postmaster aliveness + while waiting for stop/restart/promote actions (Tom Lane) + </p><p> + <span class="application">pg_ctl</span> would verify that the postmaster + is alive as a side-effect of sending the stop or promote signal, but + then it just naively waited to see the on-disk state change. If the + postmaster died uncleanly without having removed its PID file or + updated the control file, <span class="application">pg_ctl</span> would + wait until timeout. Instead make it recheck every so often that the + postmaster process is still there. + </p></li><li class="listitem"><p> + Fix error handling in <span class="application">pg_waldump</span> (Kyotaro + Horiguchi, Andres Freund) + </p><p> + While trying to read a WAL file to determine the WAL segment size, + <span class="application">pg_waldump</span> would report an incorrect + error for the case of a too-short file. In addition, the file name + reported in this and related error messages could be garbage. + </p></li><li class="listitem"><p> + Ensure that <code class="filename">contrib/pageinspect</code> functions cope + with all-zero pages (Michael Paquier) + </p><p> + This is a legitimate edge case, but the module was mostly unprepared + for it. Arrange to return nulls, or no rows, as appropriate; that + seems more useful than raising an error. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/pageinspect</code>, add defenses against + incorrect page <span class="quote">“<span class="quote">special space</span>”</span> contents, tighten checks + for correct page size, and add some missing checks that an index is + of the expected type (Michael Paquier, Justin Pryzby, Julien + Rouhaud) + </p><p> + These changes make it less likely that the module will crash on bad + data. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/postgres_fdw</code>, disable batch + insertion when <code class="literal">BEFORE INSERT ... FOR EACH ROW</code> + triggers exist on the foreign table (Etsuro Fujita) + </p><p> + Such a trigger might query the table it's on and expect to see + previously-inserted rows. With batch insertion, those rows might + not be visible yet, so disable the feature to avoid unexpected + behavior. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/postgres_fdw</code>, verify + that <code class="literal">ORDER BY</code> clauses are safe to ship before + requesting a remotely-ordered query, and include + a <code class="literal">USING</code> clause if necessary (Ronan Dunklau) + </p><p> + This fix prevents situations where the remote server might sort in a + different order than we intend. While sometimes that would be only + cosmetic, it could produce thoroughly wrong results if the remote + data is used as input for a locally-performed merge join. + </p></li><li class="listitem"><p> + Fix <span class="application">configure</span> to handle platforms that + have <code class="filename">sys/epoll.h</code> but + not <code class="filename">sys/signalfd.h</code> (Tom Lane) + </p></li><li class="listitem"><p> + Update <acronym class="acronym">JIT</acronym> code to work with LLVM 14 + (Thomas Munro) + </p></li><li class="listitem"><p> + Clean up assorted failures under <span class="productname">clang</span>'s + <code class="literal">-fsanitize=undefined</code> checks (Tom Lane, Andres + Freund, Zhihong Yu) + </p><p> + Most of these changes are just for pro-forma compliance with the + letter of the C and POSIX standards, and are unlikely to have any + effect on production builds. + </p></li><li class="listitem"><p> + Do not add OpenSSL dependencies to <span class="application">libpq</span>'s + <code class="filename">pkg-config</code> file when building without OpenSSL + (Fabrice Fontaine) + </p></li><li class="listitem"><p> + Fix PL/Perl so it builds on C compilers that don't support statements + nested within expressions (Tom Lane) + </p><p> + </p></li><li class="listitem"><p> + Fix possible build failure of <span class="application">pg_dumpall</span> + on Windows, when not using MSVC to build (Andres Freund) + </p></li><li class="listitem"><p> + In Windows builds, use <span class="application">gendef</span> instead + of <span class="application">pexports</span> to build DEF files (Andrew + Dunstan) + </p><p> + This adapts the build process to work on recent MSys tool chains. + </p></li><li class="listitem"><p> + Prevent extra expansion of shell wildcard patterns in programs built + under MinGW (Andrew Dunstan) + </p><p> + For some reason the C library provided by MinGW will expand shell + wildcard characters in a program's command-line arguments by + default. This is confusing, not least because it doesn't happen + under MSVC, so turn it off. + </p></li><li class="listitem"><p> + Update time zone data files to <span class="application">tzdata</span> + release 2022a for DST law changes in Palestine, plus historical + corrections for Chile and Ukraine. + </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-14-4.html" title="E.2. Release 14.4">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-14-2.html" title="E.4. Release 14.2">Next</a></td></tr><tr><td width="40%" align="left" valign="top">E.2. Release 14.4 </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 14.5 Documentation">Home</a></td><td width="40%" align="right" valign="top"> E.4. Release 14.2</td></tr></table></div></body></html>
\ No newline at end of file |