<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> <HTML ><HEAD ><TITLE >Release 9.3.5</TITLE ><META NAME="GENERATOR" CONTENT="Modular DocBook HTML Stylesheet Version 1.79"><LINK REV="MADE" HREF="mailto:pgsql-docs@postgresql.org"><LINK REL="HOME" TITLE="PostgreSQL 9.6.10 Documentation" HREF="index.html"><LINK REL="UP" TITLE="Release Notes" HREF="release.html"><LINK REL="PREVIOUS" TITLE="Release 9.3.6" HREF="release-9-3-6.html"><LINK REL="NEXT" TITLE="Release 9.3.4" HREF="release-9-3-4.html"><LINK REL="STYLESHEET" TYPE="text/css" HREF="stylesheet.css"><META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=ISO-8859-1"><META NAME="creation" CONTENT="2018-10-16T23:40:43"></HEAD ><BODY CLASS="SECT1" ><DIV CLASS="NAVHEADER" ><TABLE SUMMARY="Header navigation table" WIDTH="100%" BORDER="0" CELLPADDING="0" CELLSPACING="0" ><TR ><TH COLSPAN="4" ALIGN="center" VALIGN="bottom" ><A HREF="index.html" >PostgreSQL 9.6.10 Documentation</A ></TH ></TR ><TR ><TD WIDTH="10%" ALIGN="left" VALIGN="top" ><A TITLE="Release 9.3.6" HREF="release-9-3-6.html" ACCESSKEY="P" >Prev</A ></TD ><TD WIDTH="10%" ALIGN="left" VALIGN="top" ><A HREF="release.html" ACCESSKEY="U" >Up</A ></TD ><TD WIDTH="60%" ALIGN="center" VALIGN="bottom" >Appendix E. Release Notes</TD ><TD WIDTH="20%" ALIGN="right" VALIGN="top" ><A TITLE="Release 9.3.4" HREF="release-9-3-4.html" ACCESSKEY="N" >Next</A ></TD ></TR ></TABLE ><HR ALIGN="LEFT" WIDTH="100%"></DIV ><DIV CLASS="SECT1" ><H1 CLASS="SECT1" ><A NAME="RELEASE-9-3-5" >E.66. Release 9.3.5</A ></H1 ><DIV CLASS="FORMALPARA" ><P ><B >Release date: </B >2014-07-24</P ></DIV ><P > This release contains a variety of fixes from 9.3.4. For information about new features in the 9.3 major release, see <A HREF="release-9-3.html" >Section E.71</A >. </P ><DIV CLASS="SECT2" ><H2 CLASS="SECT2" ><A NAME="AEN144530" >E.66.1. Migration to Version 9.3.5</A ></H2 ><P > A dump/restore is not required for those running 9.3.X. </P ><P > However, this release corrects a logic error in <SPAN CLASS="APPLICATION" >pg_upgrade</SPAN >, as well as an index corruption problem in some GiST indexes. See the first two changelog entries below to find out whether your installation has been affected and what steps you should take if so. </P ><P > Also, if you are upgrading from a version earlier than 9.3.4, see <A HREF="release-9-3-4.html" >Section E.67</A >. </P ></DIV ><DIV CLASS="SECT2" ><H2 CLASS="SECT2" ><A NAME="AEN144537" >E.66.2. Changes</A ></H2 ><P ></P ><UL ><LI ><P > In <SPAN CLASS="APPLICATION" >pg_upgrade</SPAN >, remove <TT CLASS="FILENAME" >pg_multixact</TT > files left behind by <SPAN CLASS="APPLICATION" >initdb</SPAN > (Bruce Momjian) </P ><P > If you used a pre-9.3.5 version of <SPAN CLASS="APPLICATION" >pg_upgrade</SPAN > to upgrade a database cluster to 9.3, it might have left behind a file <TT CLASS="FILENAME" >$PGDATA/pg_multixact/offsets/0000</TT > that should not be there and will eventually cause problems in <TT CLASS="COMMAND" >VACUUM</TT >. <SPAN CLASS="emphasis" ><I CLASS="EMPHASIS" >However, in common cases this file is actually valid and must not be removed.</I ></SPAN > To determine whether your installation has this problem, run this query as superuser, in any database of the cluster: </P><PRE CLASS="PROGRAMLISTING" >WITH list(file) AS (SELECT * FROM pg_ls_dir('pg_multixact/offsets')) SELECT EXISTS (SELECT * FROM list WHERE file = '0000') AND NOT EXISTS (SELECT * FROM list WHERE file = '0001') AND NOT EXISTS (SELECT * FROM list WHERE file = 'FFFF') AND EXISTS (SELECT * FROM list WHERE file != '0000') AS file_0000_removal_required;</PRE ><P> If this query returns <TT CLASS="LITERAL" >t</TT >, manually remove the file <TT CLASS="FILENAME" >$PGDATA/pg_multixact/offsets/0000</TT >. Do nothing if the query returns <TT CLASS="LITERAL" >f</TT >. </P ></LI ><LI ><P > Correctly initialize padding bytes in <TT CLASS="FILENAME" >contrib/btree_gist</TT > indexes on <TT CLASS="TYPE" >bit</TT > columns (Heikki Linnakangas) </P ><P > This error could result in incorrect query results due to values that should compare equal not being seen as equal. Users with GiST indexes on <TT CLASS="TYPE" >bit</TT > or <TT CLASS="TYPE" >bit varying</TT > columns should <TT CLASS="COMMAND" >REINDEX</TT > those indexes after installing this update. </P ></LI ><LI ><P > Protect against torn pages when deleting GIN list pages (Heikki Linnakangas) </P ><P > This fix prevents possible index corruption if a system crash occurs while the page update is being written to disk. </P ></LI ><LI ><P > Don't clear the right-link of a GiST index page while replaying updates from WAL (Heikki Linnakangas) </P ><P > This error could lead to transiently wrong answers from GiST index scans performed in Hot Standby. </P ></LI ><LI ><P > Fix corner-case infinite loop during insertion into an SP-GiST text index (Tom Lane) </P ></LI ><LI ><P > Fix incorrect answers from SP-GiST index searches with <TT CLASS="LITERAL" >-|-</TT > (range adjacency) operator (Heikki Linnakangas) </P ></LI ><LI ><P > Fix wraparound handling for <TT CLASS="FILENAME" >pg_multixact/members</TT > (Álvaro Herrera) </P ></LI ><LI ><P > Truncate <TT CLASS="STRUCTNAME" >pg_multixact</TT > during checkpoints, not during <TT CLASS="COMMAND" >VACUUM</TT > (Álvaro Herrera) </P ><P > This change ensures that <TT CLASS="STRUCTNAME" >pg_multixact</TT > segments can't be removed if they'd still be needed during WAL replay after a crash. </P ></LI ><LI ><P > Fix possible inconsistency of all-visible flags after WAL recovery (Heikki Linnakangas) </P ></LI ><LI ><P > Fix possibly-incorrect cache invalidation during nested calls to <CODE CLASS="FUNCTION" >ReceiveSharedInvalidMessages</CODE > (Andres Freund) </P ></LI ><LI ><P > Fix race condition when updating a tuple concurrently locked by another process (Andres Freund, Álvaro Herrera) </P ></LI ><LI ><P > Fix <SPAN CLASS="QUOTE" >"could not find pathkey item to sort"</SPAN > planner failures with <TT CLASS="LITERAL" >UNION ALL</TT > over subqueries reading from tables with inheritance children (Tom Lane) </P ></LI ><LI ><P > Don't assume a subquery's output is unique if there's a set-returning function in its targetlist (David Rowley) </P ><P > This oversight could lead to misoptimization of constructs like <TT CLASS="LITERAL" >WHERE x IN (SELECT y, generate_series(1,10) FROM t GROUP BY y)</TT >. </P ></LI ><LI ><P > Improve planner to drop constant-NULL inputs of <TT CLASS="LITERAL" >AND</TT >/<TT CLASS="LITERAL" >OR</TT > when possible (Tom Lane) </P ><P > This change fixes some cases where the more aggressive parameter substitution done by 9.2 and later can lead to a worse plan than older versions produced. </P ></LI ><LI ><P > Ensure that the planner sees equivalent <TT CLASS="LITERAL" >VARIADIC</TT > and non-<TT CLASS="LITERAL" >VARIADIC</TT > function calls as equivalent (Tom Lane) </P ><P > This bug could for example result in failure to use expression indexes involving variadic functions. It might be necessary to re-create such indexes, and/or re-create views including variadic function calls that should match the indexes, for the fix to be effective for existing 9.3 installations. </P ></LI ><LI ><P > Fix handling of nested <TT CLASS="TYPE" >JSON</TT > objects in <CODE CLASS="FUNCTION" >json_populate_recordset()</CODE > and friends (Michael Paquier, Tom Lane) </P ><P > A nested <TT CLASS="TYPE" >JSON</TT > object could result in previous fields of the parent object not being shown in the output. </P ></LI ><LI ><P > Fix identification of input type category in <CODE CLASS="FUNCTION" >to_json()</CODE > and friends (Tom Lane) </P ><P > This is known to have led to inadequate quoting of <TT CLASS="TYPE" >money</TT > fields in the <TT CLASS="TYPE" >JSON</TT > result, and there may have been wrong results for other data types as well. </P ></LI ><LI ><P > Fix failure to detoast fields in composite elements of structured types (Tom Lane) </P ><P > This corrects cases where TOAST pointers could be copied into other tables without being dereferenced. If the original data is later deleted, it would lead to errors like <SPAN CLASS="QUOTE" >"missing chunk number 0 for toast value ..."</SPAN > when the now-dangling pointer is used. </P ></LI ><LI ><P > Fix <SPAN CLASS="QUOTE" >"record type has not been registered"</SPAN > failures with whole-row references to the output of Append plan nodes (Tom Lane) </P ></LI ><LI ><P > Fix possible crash when invoking a user-defined function while rewinding a cursor (Tom Lane) </P ></LI ><LI ><P > Fix query-lifespan memory leak while evaluating the arguments for a function in <TT CLASS="LITERAL" >FROM</TT > (Tom Lane) </P ></LI ><LI ><P > Fix session-lifespan memory leaks in regular-expression processing (Tom Lane, Arthur O'Dwyer, Greg Stark) </P ></LI ><LI ><P > Fix data encoding error in <TT CLASS="FILENAME" >hungarian.stop</TT > (Tom Lane) </P ></LI ><LI ><P > Prevent foreign tables from being created with OIDS when <A HREF="runtime-config-compatible.html#GUC-DEFAULT-WITH-OIDS" >default_with_oids</A > is true (Etsuro Fujita) </P ></LI ><LI ><P > Fix liveness checks for rows that were inserted in the current transaction and then deleted by a now-rolled-back subtransaction (Andres Freund) </P ><P > This could cause problems (at least spurious warnings, and at worst an infinite loop) if <TT CLASS="COMMAND" >CREATE INDEX</TT > or <TT CLASS="COMMAND" >CLUSTER</TT > were done later in the same transaction. </P ></LI ><LI ><P > Clear <TT CLASS="STRUCTNAME" >pg_stat_activity</TT >.<TT CLASS="STRUCTFIELD" >xact_start</TT > during <TT CLASS="COMMAND" >PREPARE TRANSACTION</TT > (Andres Freund) </P ><P > After the <TT CLASS="COMMAND" >PREPARE</TT >, the originating session is no longer in a transaction, so it should not continue to display a transaction start time. </P ></LI ><LI ><P > Fix <TT CLASS="COMMAND" >REASSIGN OWNED</TT > to not fail for text search objects (Álvaro Herrera) </P ></LI ><LI ><P > Prevent <TT CLASS="STRUCTNAME" >pg_class</TT >.<TT CLASS="STRUCTFIELD" >relminmxid</TT > values from going backwards during <TT CLASS="COMMAND" >VACUUM FULL</TT > (Álvaro Herrera) </P ></LI ><LI ><P > Reduce indentation in rule/view dumps to improve readability and avoid excessive whitespace (Greg Stark, Tom Lane) </P ><P > This change reduces the amount of indentation applied to nested constructs, including some cases that the user probably doesn't think of as nested, such as UNION lists. Previously, deeply nested constructs were printed with an amount of whitespace growing as O(N^2), which created a performance problem and even risk of out-of-memory failures. Now the indentation is reduced modulo 40, which is initially odd to look at but seems to preserve readability better than simply limiting the indentation would do. Redundant parenthesization of UNION lists has been reduced as well. </P ></LI ><LI ><P > Fix dumping of rules/views when subsequent addition of a column has resulted in multiple input columns matching a <TT CLASS="LITERAL" >USING</TT > specification (Tom Lane) </P ></LI ><LI ><P > Repair view printing for some cases involving functions in <TT CLASS="LITERAL" >FROM</TT > that return a composite type containing dropped columns (Tom Lane) </P ></LI ><LI ><P > Block signals during postmaster startup (Tom Lane) </P ><P > This ensures that the postmaster will properly clean up after itself if, for example, it receives <SPAN CLASS="SYSTEMITEM" >SIGINT</SPAN > while still starting up. </P ></LI ><LI ><P > Fix client host name lookup when processing <TT CLASS="FILENAME" >pg_hba.conf</TT > entries that specify host names instead of IP addresses (Tom Lane) </P ><P > Ensure that reverse-DNS lookup failures are reported, instead of just silently not matching such entries. Also ensure that we make only one reverse-DNS lookup attempt per connection, not one per host name entry, which is what previously happened if the lookup attempts failed. </P ></LI ><LI ><P > Allow the root user to use <TT CLASS="LITERAL" >postgres -C variable</TT > and <TT CLASS="LITERAL" >postgres --describe-config</TT > (MauMau) </P ><P > The prohibition on starting the server as root does not need to extend to these operations, and relaxing it prevents failure of <SPAN CLASS="APPLICATION" >pg_ctl</SPAN > in some scenarios. </P ></LI ><LI ><P > Secure Unix-domain sockets of temporary postmasters started during <TT CLASS="LITERAL" >make check</TT > (Noah Misch) </P ><P > Any local user able to access the socket file could connect as the server's bootstrap superuser, then proceed to execute arbitrary code as the operating-system user running the test, as we previously noted in CVE-2014-0067. This change defends against that risk by placing the server's socket in a temporary, mode 0700 subdirectory of <TT CLASS="FILENAME" >/tmp</TT >. The hazard remains however on platforms where Unix sockets are not supported, notably Windows, because then the temporary postmaster must accept local TCP connections. </P ><P > A useful side effect of this change is to simplify <TT CLASS="LITERAL" >make check</TT > testing in builds that override <TT CLASS="LITERAL" >DEFAULT_PGSOCKET_DIR</TT >. Popular non-default values like <TT CLASS="FILENAME" >/var/run/postgresql</TT > are often not writable by the build user, requiring workarounds that will no longer be necessary. </P ></LI ><LI ><P > Fix tablespace creation WAL replay to work on Windows (MauMau) </P ></LI ><LI ><P > Fix detection of socket creation failures on Windows (Bruce Momjian) </P ></LI ><LI ><P > On Windows, allow new sessions to absorb values of PGC_BACKEND parameters (such as <A HREF="runtime-config-logging.html#GUC-LOG-CONNECTIONS" >log_connections</A >) from the configuration file (Amit Kapila) </P ><P > Previously, if such a parameter were changed in the file post-startup, the change would have no effect. </P ></LI ><LI ><P > Properly quote executable path names on Windows (Nikhil Deshpande) </P ><P > This oversight could cause <SPAN CLASS="APPLICATION" >initdb</SPAN > and <SPAN CLASS="APPLICATION" >pg_upgrade</SPAN > to fail on Windows, if the installation path contained both spaces and <TT CLASS="LITERAL" >@</TT > signs. </P ></LI ><LI ><P > Fix linking of <SPAN CLASS="APPLICATION" >libpython</SPAN > on macOS (Tom Lane) </P ><P > The method we previously used can fail with the Python library supplied by Xcode 5.0 and later. </P ></LI ><LI ><P > Avoid buffer bloat in <SPAN CLASS="APPLICATION" >libpq</SPAN > when the server consistently sends data faster than the client can absorb it (Shin-ichi Morita, Tom Lane) </P ><P > <SPAN CLASS="APPLICATION" >libpq</SPAN > could be coerced into enlarging its input buffer until it runs out of memory (which would be reported misleadingly as <SPAN CLASS="QUOTE" >"lost synchronization with server"</SPAN >). Under ordinary circumstances it's quite far-fetched that data could be continuously transmitted more quickly than the <CODE CLASS="FUNCTION" >recv()</CODE > loop can absorb it, but this has been observed when the client is artificially slowed by scheduler constraints. </P ></LI ><LI ><P > Ensure that LDAP lookup attempts in <SPAN CLASS="APPLICATION" >libpq</SPAN > time out as intended (Laurenz Albe) </P ></LI ><LI ><P > Fix <SPAN CLASS="APPLICATION" >ecpg</SPAN > to do the right thing when an array of <TT CLASS="TYPE" >char *</TT > is the target for a FETCH statement returning more than one row, as well as some other array-handling fixes (Ashutosh Bapat) </P ></LI ><LI ><P > Fix <SPAN CLASS="APPLICATION" >pg_dump</SPAN > to cope with a materialized view that depends on a table's primary key (Tom Lane) </P ><P > This occurs if the view's query relies on functional dependency to abbreviate a <TT CLASS="LITERAL" >GROUP BY</TT > list. <SPAN CLASS="APPLICATION" >pg_dump</SPAN > got sufficiently confused that it dumped the materialized view as a regular view. </P ></LI ><LI ><P > Fix parsing of <SPAN CLASS="APPLICATION" >pg_dumpall</SPAN >'s <TT CLASS="OPTION" >-i</TT > switch (Tom Lane) </P ></LI ><LI ><P > Fix <SPAN CLASS="APPLICATION" >pg_restore</SPAN >'s processing of old-style large object comments (Tom Lane) </P ><P > A direct-to-database restore from an archive file generated by a pre-9.0 version of <SPAN CLASS="APPLICATION" >pg_dump</SPAN > would usually fail if the archive contained more than a few comments for large objects. </P ></LI ><LI ><P > Fix <SPAN CLASS="APPLICATION" >pg_upgrade</SPAN > for cases where the new server creates a TOAST table but the old version did not (Bruce Momjian) </P ><P > This rare situation would manifest as <SPAN CLASS="QUOTE" >"relation OID mismatch"</SPAN > errors. </P ></LI ><LI ><P > In <SPAN CLASS="APPLICATION" >pg_upgrade</SPAN >, preserve <TT CLASS="STRUCTNAME" >pg_database</TT >.<TT CLASS="STRUCTFIELD" >datminmxid</TT > and <TT CLASS="STRUCTNAME" >pg_class</TT >.<TT CLASS="STRUCTFIELD" >relminmxid</TT > values from the old cluster, or insert reasonable values when upgrading from pre-9.3; also defend against unreasonable values in the core server (Bruce Momjian, Álvaro Herrera, Tom Lane) </P ><P > These changes prevent scenarios in which autovacuum might insist on scanning the entire cluster's contents immediately upon starting the new cluster, or in which tracking of unfrozen MXID values might be disabled completely. </P ></LI ><LI ><P > Prevent <TT CLASS="FILENAME" >contrib/auto_explain</TT > from changing the output of a user's <TT CLASS="COMMAND" >EXPLAIN</TT > (Tom Lane) </P ><P > If <TT CLASS="FILENAME" >auto_explain</TT > is active, it could cause an <TT CLASS="LITERAL" >EXPLAIN (ANALYZE, TIMING OFF)</TT > command to nonetheless print timing information. </P ></LI ><LI ><P > Fix query-lifespan memory leak in <TT CLASS="FILENAME" >contrib/dblink</TT > (MauMau, Joe Conway) </P ></LI ><LI ><P > In <TT CLASS="FILENAME" >contrib/pgcrypto</TT > functions, ensure sensitive information is cleared from stack variables before returning (Marko Kreen) </P ></LI ><LI ><P > Prevent use of already-freed memory in <TT CLASS="FILENAME" >contrib/pgstattuple</TT >'s <CODE CLASS="FUNCTION" >pgstat_heap()</CODE > (Noah Misch) </P ></LI ><LI ><P > In <TT CLASS="FILENAME" >contrib/uuid-ossp</TT >, cache the state of the OSSP UUID library across calls (Tom Lane) </P ><P > This improves the efficiency of UUID generation and reduces the amount of entropy drawn from <TT CLASS="FILENAME" >/dev/urandom</TT >, on platforms that have that. </P ></LI ><LI ><P > Update time zone data files to <SPAN CLASS="APPLICATION" >tzdata</SPAN > release 2014e for DST law changes in Crimea, Egypt, and Morocco. </P ></LI ></UL ></DIV ></DIV ><DIV CLASS="NAVFOOTER" ><HR ALIGN="LEFT" WIDTH="100%"><TABLE SUMMARY="Footer navigation table" WIDTH="100%" BORDER="0" CELLPADDING="0" CELLSPACING="0" ><TR ><TD WIDTH="33%" ALIGN="left" VALIGN="top" ><A HREF="release-9-3-6.html" ACCESSKEY="P" >Prev</A ></TD ><TD WIDTH="34%" ALIGN="center" VALIGN="top" ><A HREF="index.html" ACCESSKEY="H" >Home</A ></TD ><TD WIDTH="33%" ALIGN="right" VALIGN="top" ><A HREF="release-9-3-4.html" ACCESSKEY="N" >Next</A ></TD ></TR ><TR ><TD WIDTH="33%" ALIGN="left" VALIGN="top" >Release 9.3.6</TD ><TD WIDTH="34%" ALIGN="center" VALIGN="top" ><A HREF="release.html" ACCESSKEY="U" >Up</A ></TD ><TD WIDTH="33%" ALIGN="right" VALIGN="top" >Release 9.3.4</TD ></TR ></TABLE ></DIV ></BODY ></HTML >