diff options
Diffstat (limited to 'doc/src/sgml/html/release-15-3.html')
| -rw-r--r-- | doc/src/sgml/html/release-15-3.html | 603 |
1 files changed, 603 insertions, 0 deletions
diff --git a/doc/src/sgml/html/release-15-3.html b/doc/src/sgml/html/release-15-3.html new file mode 100644 index 0000000..f4f0e9d --- /dev/null +++ b/doc/src/sgml/html/release-15-3.html @@ -0,0 +1,603 @@ +<?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 15.3</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-15-4.html" title="E.1. Release 15.4" /><link rel="next" href="release-15-2.html" title="E.3. Release 15.2" /></head><body id="docContent" class="container-fluid col-10"><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">E.2. Release 15.3</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="release-15-4.html" title="E.1. Release 15.4">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 15.4 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="release-15-2.html" title="E.3. Release 15.2">Next</a></td></tr></table><hr /></div><div class="sect1" id="RELEASE-15-3"><div class="titlepage"><div><div><h2 class="title" style="clear: both">E.2. Release 15.3</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="release-15-3.html#id-1.11.6.6.4">E.2.1. Migration to Version 15.3</a></span></dt><dt><span class="sect2"><a href="release-15-3.html#id-1.11.6.6.5">E.2.2. Changes</a></span></dt></dl></div><p><strong>Release date: </strong>2023-05-11</p><p> + This release contains a variety of fixes from 15.2. + For information about new features in major release 15, see + <a class="xref" href="release-15.html" title="E.5. Release 15">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 15.3</h3></div></div></div><p> + A dump/restore is not required for those running 15.X. + </p><p> + However, if you are upgrading from a version earlier than 15.1, + see <a class="xref" href="release-15-1.html" title="E.4. Release 15.1">Section E.4</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 <code class="command">CREATE SCHEMA</code> from defeating changes + in <code class="varname">search_path</code> (Alexander Lakhin) + </p><p> + Within a <code class="command">CREATE SCHEMA</code> command, objects in the + prevailing <code class="varname">search_path</code>, as well as those in the + newly-created schema, would be visible even within a called + function or script that attempted to set a + secure <code class="varname">search_path</code>. This could allow any user + having permission to create a schema to hijack the privileges of a + security definer function or extension script. + </p><p> + The <span class="productname">PostgreSQL</span> Project thanks + Alexander Lakhin for reporting this problem. + (CVE-2023-2454) + </p></li><li class="listitem"><p> + Enforce row-level security policies correctly after inlining a + set-returning function (Stephen Frost, Tom Lane) + </p><p> + If a set-returning SQL-language function refers to a table having + row-level security policies, and it can be inlined into a calling + query, those RLS policies would not get enforced properly in some + cases involving re-using a cached plan under a different role. + This could allow a user to see or modify rows that should have been + invisible. + </p><p> + The <span class="productname">PostgreSQL</span> Project thanks + Wolfgang Walther for reporting this problem. + (CVE-2023-2455) + </p></li><li class="listitem"><p> + Fix potential corruption of the template (source) database after + <code class="command">CREATE DATABASE</code> with the <code class="literal">STRATEGY + WAL_LOG</code> option (Nathan Bossart, Ryo Matsumura) + </p><p> + Improper buffer handling created a risk that any later modification + of the template's <code class="structname">pg_class</code> catalog would be + lost. + </p></li><li class="listitem"><p> + Fix memory leakage and unnecessary disk reads + during <code class="command">CREATE DATABASE</code> with the <code class="literal">STRATEGY + WAL_LOG</code> option (Andres Freund) + </p></li><li class="listitem"><p> + Avoid crash when the new schema name is omitted + in <code class="command">CREATE SCHEMA</code> (Michael Paquier) + </p><p> + The SQL standard allows writing <code class="literal">CREATE SCHEMA AUTHORIZATION + <em class="replaceable"><code>owner_name</code></em></code>, with the schema + name defaulting to <em class="replaceable"><code>owner_name</code></em>. However + some code paths expected the schema name to be present and would + fail. + </p></li><li class="listitem"><p> + Fix various planner failures with <code class="command">MERGE</code> + commands (Tom Lane) + </p><p> + Planning could fail with errors like <span class="quote">“<span class="quote">variable not found in + subplan target list</span>”</span> or <span class="quote">“<span class="quote">PlaceHolderVar found where not + expected</span>”</span>. + </p></li><li class="listitem"><p> + Fix the row count reported by <code class="command">MERGE</code> for some + corner cases (Dean Rasheed) + </p><p> + The row count reported in the command tag counted rows that actually + hadn't been modified due to a <code class="literal">BEFORE ROW</code> trigger + returning NULL. This is inconsistent with what happens in + plain <code class="command">UPDATE</code> or <code class="command">DELETE</code>, so + change it to not count such rows. Also, avoid counting a row twice + when <code class="command">MERGE</code> moves it into a different partition of + a partitioned table. + </p></li><li class="listitem"><p> + Fix <code class="command">MERGE</code> problems with concurrent updates + (Dean Rasheed, Álvaro Herrera) + </p><p> + Some cases misbehaved if a row to be updated or deleted + by <code class="command">MERGE</code> had just been updated by a concurrent + transaction. This could lead to a crash, or the wrong merge action + being executed, or no action at all. + </p></li><li class="listitem"><p> + Add support for decompiling <code class="command">MERGE</code> + commands (Álvaro Herrera) + </p><p> + This was overlooked when <code class="command">MERGE</code> was added, but + it's essential support for <code class="command">MERGE</code> in new-style SQL + functions. + </p></li><li class="listitem"><p> + Fix enabling/disabling of foreign-key triggers in partitioned tables + (Tom Lane) + </p><p> + <code class="command">ALTER TABLE ... ENABLE/DISABLE TRIGGER</code> failed if + applied to a partitioned table's foreign-key enforcement triggers, + because it tried to locate the clone triggers for the partitions by + name, and they do not have the same name. Locate them by + parent-trigger OID instead. + </p></li><li class="listitem"><p> + Disallow altering composite types that are stored in indexes + (Tom Lane) + </p><p> + <code class="command">ALTER TYPE</code> disallows non-binary-compatible + modifications of composite types if they are stored in any table + columns. (Perhaps that will be allowed someday, but it hasn't + happened yet; the locking implications of rewriting many tables are + daunting.) We overlooked the possibility that an index might + contain a composite type that doesn't also appear in its table. + </p></li><li class="listitem"><p> + Disallow system columns as elements of foreign keys (Tom Lane) + </p><p> + Since the removal of OID as a system column, there is no plausible + use-case for this, and various bits of code no longer support it. + Disallow it rather than trying to fix all the cases. + </p></li><li class="listitem"><p> + Ensure that <code class="command">COPY TO</code> from an RLS-enabled parent + table does not copy any rows from child tables (Antonin Houska) + </p><p> + The documentation is quite clear that <code class="command">COPY TO</code> + copies rows from only the named table, not any inheritance children + it may have. However, if row-level security was enabled on the table + then this stopped being true. + </p></li><li class="listitem"><p> + Avoid possible crash when <code class="function">array_position()</code> + or <code class="function">array_positions()</code> is passed an empty array + (Tom Lane) + </p></li><li class="listitem"><p> + Fix possible out-of-bounds fetch in <code class="function">to_char()</code> + (Tom Lane) + </p><p> + With bad luck this could have resulted in a server crash. + </p></li><li class="listitem"><p> + Avoid buffer overread in <code class="function">translate()</code> function + (Daniil Anisimov) + </p><p> + When using the deletion feature, the function might fetch the byte + just after the input string, creating a small risk of crash. + </p></li><li class="listitem"><p> + Adjust text-search-related character classification logic to + correctly detect whether the prevailing locale + is <code class="literal">C</code> (Jeff Davis) + </p><p> + This code got confused if the database's default collation uses ICU. + </p></li><li class="listitem"><p> + Avoid possible crash on empty input for type <code class="type">interval</code> + (Tom Lane) + </p></li><li class="listitem"><p> + Re-allow exponential notation in ISO-8601 interval fields + (Tom Lane) + </p><p> + Interval input like <code class="literal">P0.1e10D</code> isn't officially + sanctioned by ISO-8601, but we accepted it for a long time before + version 15, so re-allow it. + </p></li><li class="listitem"><p> + Fix error cursor setting for parse errors in JSON string literals + (Tom Lane) + </p><p> + Most cases in which a syntax error is detected in a string literal + within a JSON value failed to set the error cursor appropriately. + This led at least to an unhelpful error message (pointing to the + token before the string, rather than the actual trouble spot), and + could even result in a crash in v14 and later. + </p></li><li class="listitem"><p> + Fix data corruption due to <code class="varname">vacuum_defer_cleanup_age</code> + being larger than the current 64-bit xid (Andres Freund) + </p><p> + In v14 and later with non-default settings + of <code class="varname">vacuum_defer_cleanup_age</code>, it was possible to + compute a very large vacuum cleanup horizon xid, leading to vacuum + removing rows that are still live. v12 and v13 have a lesser form + of the same problem affecting only GiST indexes, which could lead to + index pages getting recycled too early. + </p></li><li class="listitem"><p> + Fix parser's failure to detect some cases of improperly-nested + aggregates (Tom Lane) + </p><p> + This oversight could lead to executor failures for queries that + should have been rejected as invalid. + </p></li><li class="listitem"><p> + Fix data structure corruption during parsing of + serial <code class="literal">SEQUENCE NAME</code> options (David Rowley) + </p><p> + This can lead to trouble if an event trigger captures the corrupted + parse tree. + </p></li><li class="listitem"><p> + Correctly update plan nodes' parallel-safety markings when moving + initplans from one node to another (Tom Lane) + </p><p> + This planner oversight could lead to <span class="quote">“<span class="quote">subplan was not + initialized</span>”</span> errors at runtime. + </p></li><li class="listitem"><p> + Avoid failure with PlaceHolderVars in extended-statistics code + (Tom Lane) + </p><p> + Use of dependency-type extended statistics could fail with + <span class="quote">“<span class="quote">PlaceHolderVar found where not expected</span>”</span>. + </p></li><li class="listitem"><p> + Fix incorrect tests for whether a qual clause applied to a subquery + can be transformed into a window aggregate <span class="quote">“<span class="quote">run + condition</span>”</span> within the subquery (David Rowley) + </p><p> + A SubPlan within such a clause would cause assertion failures or + incorrect answers, as would some other unusual cases. + </p></li><li class="listitem"><p> + Disable the inverse-transition optimization for window aggregates + when the call contains sub-SELECTs (David Rowley) + </p><p> + This optimization requires that the aggregate's argument expressions + have repeatable results, which might not hold for a sub-SELECT. + </p></li><li class="listitem"><p> + Fix oversights in execution of nested <code class="literal">ARRAY[]</code> + constructs (Alexander Lakhin, Tom Lane) + </p><p> + Correctly detect overflow of the total space needed for the result + array, avoiding a possible crash due to undersized output + allocation. Also ensure that any trailing padding space in the + result array is zeroed; while leaving garbage there is harmless for + most purposes, it can result in odd behavior later. + </p></li><li class="listitem"><p> + Prevent crash when updating a field within an + array-of-domain-over-composite-type column (Dmitry Dolgov) + </p></li><li class="listitem"><p> + Fix partition pruning logic for partitioning on boolean columns + (David Rowley) + </p><p> + Pruning with a condition like <code class="literal">boolcol IS NOT TRUE</code> + was done incorrectly, leading to possibly not returning rows in + which <code class="literal">boolcol</code> is NULL. Also, the rather unlikely + case of partitioning on <code class="literal">NOT boolcol</code> was handled + incorrectly. + </p></li><li class="listitem"><p> + Fix race condition in per-batch cleanup during parallel hash join + (Thomas Munro, Melanie Plageman) + </p><p> + A crash was possible given unlucky timing and + <code class="varname">parallel_leader_participation</code> + = <code class="literal">off</code> (which is not the default). + </p></li><li class="listitem"><p> + Recalculate <code class="literal">GENERATED</code> columns after an + EvalPlanQual check (Tom Lane) + </p><p> + In <code class="literal">READ COMMITTED</code> isolation mode, the effects of + a row update might need to get reapplied to a newer version of the + row than the query found originally. If so, we need to recompute + any <code class="literal">GENERATED</code> columns, in case they depend on + columns that were changed by the concurrent update. + </p></li><li class="listitem"><p> + Fix memory leak in Memoize plan execution (David Rowley) + </p></li><li class="listitem"><p> + Fix buffer refcount leak when using batched inserts for a foreign + table included in a partitioned tree (Alexander Pyhalov) + </p></li><li class="listitem"><p> + Restore support for + sub-millisecond <code class="varname">vacuum_cost_delay</code> settings + (Thomas Munro) + </p></li><li class="listitem"><p> + Don't balance vacuum cost delay when a table has a + per-relation <code class="varname">vacuum_cost_delay</code> setting of zero + (Masahiko Sawada) + </p><p> + Delay balancing is supposed to be disabled whenever autovacuum is + processing a table with a + per-relation <code class="varname">vacuum_cost_delay</code> setting, but this + was done only for positive settings, not zero. + </p></li><li class="listitem"><p> + Fix corner-case crashes when columns have been added to the end of a + view (Tom Lane) + </p></li><li class="listitem"><p> + Repair rare failure of MULTIEXPR_SUBLINK subplans in partitioned + updates (Andres Freund, Tom Lane) + </p><p> + Use of the syntax <code class="literal">INSERT ... ON CONFLICT DO UPDATE SET (c1, + ...) = (SELECT ...)</code> with a partitioned target table could + result in failure if any child table is dissimilar from the parent + (for example, different physical column order). + This typically manifested as failure of consistency checks in the + executor; but a crash or incorrect data updates are also possible. + </p></li><li class="listitem"><p> + Fix handling of <code class="literal">DEFAULT</code> markers within a + multi-row <code class="literal">INSERT ... VALUES</code> query on a view that + has a <code class="literal">DO ALSO INSERT ... SELECT</code> rule (Dean + Rasheed) + </p><p> + Such cases typically failed with <span class="quote">“<span class="quote">unrecognized node + type</span>”</span> errors or assertion failures. + </p></li><li class="listitem"><p> + Support references to <code class="literal">OLD</code> + and <code class="literal">NEW</code> within subqueries in rule actions + (Dean Rasheed, Tom Lane) + </p><p> + Such references are really lateral references, but the server could + crash if the subquery wasn't explicitly marked + with <code class="literal">LATERAL</code>. Arrange to do that implicitly when + necessary. + </p></li><li class="listitem"><p> + When decompiling a rule or SQL function body + containing <code class="command">INSERT</code>/<code class="command">UPDATE</code>/<code class="command">DELETE</code> + within <code class="command">WITH</code>, take care to print the correct alias + for the target table (Tom Lane) + </p></li><li class="listitem"><p> + Fix glitches in <code class="literal">SERIALIZABLE READ ONLY</code> + optimization (Thomas Munro) + </p><p> + Transactions already marked as <span class="quote">“<span class="quote">doomed</span>”</span> confused the + safe-snapshot optimization for <code class="literal">SERIALIZABLE READ + ONLY</code> transactions. The optimization was unnecessarily + skipped in some cases. In other cases an assertion failure occurred + (but there was no problem in non-assert builds). + </p></li><li class="listitem"><p> + Avoid leaking cache callback slots in + the <code class="literal">pgoutput</code> logical decoding plugin (Shi Yu) + </p><p> + Multiple cycles of starting up and shutting down the plugin within a + single session would eventually lead to an <span class="quote">“<span class="quote">out of + relcache_callback_list slots</span>”</span> error. + </p></li><li class="listitem"><p> + Avoid unnecessary calls to custom validators for index operator + class options (Alexander Korotkov) + </p><p> + This change fixes some cases where an unexpected error was thrown. + </p></li><li class="listitem"><p> + Avoid useless work while scanning a multi-column BRIN index with + multiple scan keys (Tomas Vondra) + </p><p> + The existing code effectively considered only the last scan key + while deciding whether a range matched, thus usually scanning more + of the index than it needed to. + </p></li><li class="listitem"><p> + Fix netmask handling in BRIN inet_minmax_multi_ops opclass + (Tomas Vondra) + </p><p> + This error triggered an assertion failure in assert-enabled builds, + but is mostly harmless in production builds. + </p></li><li class="listitem"><p> + Fix dereference of dangling pointer during buffering build of a GiST + index (Alexander Lakhin) + </p><p> + This error seems to usually be harmless in production builds, as the + fetched value is noncritical; but in principle it could cause a + server crash. + </p></li><li class="listitem"><p> + Ignore dropped columns and generated columns during logical + replication of an update or delete action (Onder Kalaci, Shi Yu) + </p><p> + Replication with the <code class="literal">REPLICA IDENTITY FULL</code> option + failed if the table contained such columns. + </p></li><li class="listitem"><p> + Correct the name of the wait event for SLRU buffer I/O for commit + timestamps (Alexander Lakhin) + </p><p> + This wait event is named <code class="literal">CommitTsBuffer</code> according + to the documentation, but the code had it + as <code class="literal">CommitTSBuffer</code>. Change the code to match the + documentation, as that way is more consistent with the naming of + related wait events. + </p></li><li class="listitem"><p> + Re-activate reporting of wait event <code class="literal">SLRUFlushSync</code> + (Thomas Munro) + </p><p> + Reporting of this type of wait was accidentally removed in code + refactoring. + </p></li><li class="listitem"><p> + Avoid possible underflow when calculating how many WAL segments to + keep (Kyotaro Horiguchi) + </p><p> + This could result in not honoring <code class="varname">wal_keep_size</code> + accurately. + </p></li><li class="listitem"><p> + Disable startup progress reporting overhead in standby mode + (Bharath Rupireddy) + </p><p> + In standby mode, we don't actually report progress of recovery, + but we were doing work to track it anyway. + </p></li><li class="listitem"><p> + Support RSA-PSS certificates with SCRAM-SHA-256 channel binding + (Jacob Champion, Heikki Linnakangas) + </p><p> + This feature requires building with OpenSSL 1.1.1 or newer. Both + the server and <span class="application">libpq</span> are affected. + </p></li><li class="listitem"><p> + Avoid race condition with process ID tracking on Windows (Thomas Munro) + </p><p> + The operating system could recycle a PID before the postmaster + observed that that child process was gone. This could lead to + tracking more than one child with the same PID, resulting in + confusion. + </p></li><li class="listitem"><p> + Fix <code class="function">list_copy_head()</code> to work correctly on an + empty List (David Rowley) + </p><p> + This case is not known to be reached by any + core <span class="productname">PostgreSQL</span> code, but extensions + might rely on it working. + </p></li><li class="listitem"><p> + Add missing cases to <code class="function">SPI_result_code_string()</code> + (Dean Rasheed) + </p></li><li class="listitem"><p> + Fix erroneous Valgrind markings + in <code class="function">AllocSetRealloc()</code> (Karina Litskevich) + </p><p> + In the unusual case where the size of a large (>8kB) palloc chunk + is decreased, a Valgrind-aware build would mismark the defined-ness + state of the memory released from the chunk, possibly causing + incorrect results during Valgrind testing. + </p></li><li class="listitem"><p> + Fix assertion failure for <code class="command">MERGE</code> into a + partitioned table with row-level security enabled (Dean Rasheed) + </p></li><li class="listitem"><p> + Avoid assertion failure when decoding a transactional logical + replication message (Tomas Vondra) + </p></li><li class="listitem"><p> + Avoid locale sensitivity when processing regular expression escapes + (Jeff Davis) + </p><p> + A backslash followed by a non-ASCII character could sometimes cause + an assertion failure, depending on the prevailing locale. + </p></li><li class="listitem"><p> + Avoid trying to write an empty WAL record + in <code class="function">log_newpage_range()</code> when the last few pages + in the specified range are empty (Matthias van de Meent) + </p><p> + It is not entirely clear whether this case is reachable in released + branches, but if it is then an assertion failure could occur. + </p></li><li class="listitem"><p> + Fix session-lifespan memory leakage in <span class="application">plpgsql</span> + <code class="literal">DO</code> blocks that use cast expressions + (Ajit Awekar, Tom Lane) + </p></li><li class="listitem"><p> + Tighten array dimensionality checks when converting Perl + list structures to multi-dimensional SQL arrays (Tom Lane) + </p><p> + <span class="application">plperl</span> could misbehave when the nesting + of sub-lists is inconsistent so that the data does not represent a + rectangular array of values. Such cases now produce errors, but + previously they could result in a crash or garbage output. + </p></li><li class="listitem"><p> + Tighten array dimensionality checks when converting Python + list structures to multi-dimensional SQL arrays (Tom Lane) + </p><p> + <span class="application">plpython</span> could misbehave when dealing + with empty sub-lists, or when the nesting of sub-lists is + inconsistent so that the data does not represent a rectangular array + of values. The former should result in an empty output array, and + the latter in an error. But some cases resulted in a crash, and + others in unexpected output. + </p></li><li class="listitem"><p> + Fix unwinding of exception stack + in <span class="application">plpython</span> (Xing Guo) + </p><p> + Some rare failure cases could return without cleaning up the PG_TRY + exception stack, risking a crash if another error was raised before + the next stack level was unwound. + </p></li><li class="listitem"><p> + Fix inconsistent GSS-encryption error handling + in <span class="application">libpq</span>'s + <code class="function">PQconnectPoll()</code> + (Michael Paquier) + </p><p> + With <code class="option">gssencmode</code> set to <code class="literal">require</code>, + the connection was not marked dead after a GSS initialization + failure. Make it fail immediately, as the equivalent case for TLS + encryption has long done. + </p></li><li class="listitem"><p> + Fix possible data corruption in <span class="application">ecpg</span> + programs built with the <code class="option">-C ORACLE</code> option + (Kyotaro Horiguchi) + </p><p> + When <code class="function">ecpg_get_data()</code> is called + with <code class="varname">varcharsize</code> set to zero, it could write a + terminating zero character into the last byte of the preceding + field, truncating the data in that field. + </p></li><li class="listitem"><p> + Fix <span class="application">pg_dump</span> so that partitioned tables + that are hash-partitioned on an enum-type column can be restored + successfully (Tom Lane) + </p><p> + Since the hash codes for enum values depend on the OIDs assigned to + the enum, they are typically different after a dump and restore, + meaning that rows often need to go into a different partition than + they were in originally. Users can work around that by specifying + the <code class="option">--load-via-partition-root</code> option; but since + there is very little chance of success without that, + teach <span class="application">pg_dump</span> to apply it automatically + to such tables. + </p><p> + Also, fix <span class="application">pg_restore</span> to not try + to <code class="command">TRUNCATE</code> target tables before restoring into + them when <code class="option">--load-via-partition-root</code> mode is used. + This avoids a hazard of deadlocks and lost data. + </p></li><li class="listitem"><p> + Correctly detect non-seekable files on Windows + (Juan José Santamaría Flecha, Michael Paquier, Daniel Watzinger) + </p><p> + This bug led to misbehavior when <span class="application">pg_dump</span> + writes to a pipe or <span class="application">pg_restore</span> reads from + one. + </p></li><li class="listitem"><p> + In <span class="application">pgbench</span>'s <span class="quote">“<span class="quote">prepared</span>”</span> + mode, prepare all the commands in a pipeline before starting the + pipeline (Álvaro Herrera) + </p><p> + This avoids a failure when a pgbench script tries to + start a serializable transaction inside a pipeline. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/amcheck</code>'s heap checking code, deal + correctly with tuples having zero xmin or xmax (Robert Haas) + </p></li><li class="listitem"><p> + In <code class="filename">contrib/amcheck</code>, deal sanely with xids that + appear to be before epoch zero (Andres Freund) + </p><p> + In cases of corruption we might see a wrapped-around 32-bit xid that + appears to be before the first xid epoch. Promoting such a value to + 64-bit form produced a value far in the future, resulting in wrong + reports. Return FirstNormalFullTransactionId in such cases so that + things work reasonably sanely. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/basebackup_to_shell</code>, properly detect + failure to open a pipe (Robert Haas) + </p></li><li class="listitem"><p> + In <code class="filename">contrib/hstore_plpython</code>, avoid crashing if + the Python value to be transformed isn't a mapping (Dmitry Dolgov, + Tom Lane) + </p><p> + This should give an error, but Python 3 changed some APIs in a way + that caused the check to misbehave, allowing a crash to ensue. + </p></li><li class="listitem"><p> + Require the <code class="literal">siglen</code> option of a GiST index on + an <code class="type">ltree</code> column, if specified, to be a multiple of 4 + (Alexander Korotkov) + </p><p> + Other values result in misaligned accesses to index content, which + is harmless on Intel-compatible hardware but can cause a crash on + some other architectures. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/pageinspect</code>, add defenses against + incorrect input for the <code class="function">gist_page_items()</code> function + (Dmitry Koval) + </p></li><li class="listitem"><p> + Fix misbehavior in <code class="filename">contrib/pg_trgm</code> with an + unsatisfiable regular expression (Tom Lane) + </p><p> + A regex such as <code class="literal">$foo</code> is legal but unsatisfiable; + the regex compiler recognizes that and produces an empty NFA graph. + Attempting to optimize such a graph into a pg_trgm GIN or GiST index + qualification resulted in accessing off the end of a work array, + possibly leading to crashes. + </p></li><li class="listitem"><p> + Fix handling of escape sequences + in <code class="filename">contrib/postgres_fdw</code>'s + <code class="varname">application_name</code> parameter (Kyotaro Horiguchi, + Michael Paquier) + </p><p> + The code to expand these could fail if executed in a background + process, as for example during auto-analyze of a foreign table. + </p></li><li class="listitem"><p> + In <code class="filename">contrib/pg_walinspect</code>, limit memory usage + of <code class="function">pg_get_wal_records_info()</code> (Bharath Rupireddy) + </p></li><li class="listitem"><p> + Use the <code class="option">--strip-unneeded</code> option when stripping + static libraries with + GNU-compatible <span class="application">strip</span> (Tom Lane) + </p><p> + Previously, <code class="literal">make install-strip</code> used + the <code class="option">-x</code> option in this case. This change avoids + misbehavior of <span class="application">llvm-strip</span>, and gives + slightly smaller output as well. + </p></li><li class="listitem"><p> + Stop recommending auto-download of DTD files for building the + documentation, and indeed disable it (Aleksander Alekseev, Peter + Eisentraut, Tom Lane) + </p><p> + It appears no longer possible to build the SGML documentation + without a local installation of the DocBook DTD files. + Formerly <span class="application">xsltproc</span> could download those + files on-the-fly from sourceforge.net; but sourceforge.net now + permits only HTTPS access, and no common version + of <span class="application">xsltproc</span> supports that. Hence, remove + the bits of our documentation suggesting that that's possible or + useful, and instead + add <span class="application">xsltproc</span>'s <code class="option">--nonet</code> + option to the build recipes. + </p></li><li class="listitem"><p> + When running TAP tests in PGXS builds, use a saner location for the + temporary <code class="filename">portlock</code> directory (Peter Eisentraut) + </p><p> + Place it under <code class="filename">tmp_check</code> in the build + directory. With the previous coding, a PGXS build would try to place + it in the installation directory, which is not necessarily writable. + </p></li><li class="listitem"><p> + Update time zone data files to <span class="application">tzdata</span> + release 2023c for DST law changes in Egypt, Greenland, Morocco, and + Palestine. + </p><p> + When observing Moscow time, Europe/Kirov and Europe/Volgograd now + use the abbreviations MSK/MSD instead of numeric abbreviations, + for consistency with other timezones observing Moscow time. + Also, America/Yellowknife is no longer distinct from America/Edmonton; + this affects some pre-1948 timestamps in that area. + </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-15-4.html" title="E.1. Release 15.4">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-15-2.html" title="E.3. Release 15.2">Next</a></td></tr><tr><td width="40%" align="left" valign="top">E.1. Release 15.4 </td><td width="20%" align="center"><a accesskey="h" href="index.html" title="PostgreSQL 15.4 Documentation">Home</a></td><td width="40%" align="right" valign="top"> E.3. Release 15.2</td></tr></table></div></body></html>
\ No newline at end of file |