<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Release 9.4</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 9.6.4 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Release Notes"
HREF="release.html"><LINK
REL="PREVIOUS"
TITLE="Release 9.4.1"
HREF="release-9-4-1.html"><LINK
REL="NEXT"
TITLE="Release 9.3.18"
HREF="release-9-3-18.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="2017-08-11T02:27:18"></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 9.6.4 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="Release 9.4.1"
HREF="release-9-4-1.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 9.3.18"
HREF="release-9-3-18.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RELEASE-9-4"
>E.28. Release 9.4</A
></H1
><DIV
CLASS="FORMALPARA"
><P
><B
>Release date: </B
>2014-12-18</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN137913"
>E.28.1. Overview</A
></H2
><P
> Major enhancements in <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 9.4 include:
</P
><P
></P
><UL
><LI
><P
> Add <A
HREF="datatype-json.html"
><TT
CLASS="TYPE"
>jsonb</TT
></A
>, a more
capable and efficient data type for storing <ACRONYM
CLASS="ACRONYM"
>JSON</ACRONYM
> data
</P
></LI
><LI
><P
> Add new <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> command <A
HREF="sql-altersystem.html"
>ALTER SYSTEM</A
>
for changing <TT
CLASS="FILENAME"
>postgresql.conf</TT
> configuration file entries
</P
></LI
><LI
><P
> Reduce lock strength for some <A
HREF="sql-altertable.html"
>ALTER TABLE</A
>
commands
</P
></LI
><LI
><P
> Allow <A
HREF="rules-materializedviews.html"
>materialized views</A
>
to be refreshed without blocking concurrent reads
</P
></LI
><LI
><P
> Add support for <A
HREF="logicaldecoding.html"
>logical decoding</A
>
of WAL data, to allow database changes to be streamed out in a
customizable format
</P
></LI
><LI
><P
> Allow <A
HREF="bgworker.html"
>background worker processes</A
>
to be dynamically registered, started and terminated
</P
></LI
></UL
><P
> The above items are explained in more detail in the sections below.
</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN137941"
>E.28.2. Migration to Version 9.4</A
></H2
><P
> A dump/restore using <A
HREF="app-pg-dumpall.html"
><SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
></A
>, or use
of <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
>, is required for those wishing to migrate
data from any previous release.
</P
><P
> Version 9.4 contains a number of changes that may affect compatibility
with previous releases. Observe the following incompatibilities:
</P
><P
></P
><UL
><LI
><P
> Tighten checks for multidimensional <A
HREF="arrays.html"
>array</A
> input (Bruce Momjian)
</P
><P
> Previously, an input array string that started with a single-element
sub-array could later contain multi-element sub-arrays,
e.g. <TT
CLASS="LITERAL"
>'{{1}, {2,3}}'::int[]</TT
> would be accepted.
</P
></LI
><LI
><P
> When converting values of type <TT
CLASS="TYPE"
>date</TT
>, <TT
CLASS="TYPE"
>timestamp</TT
>
or <TT
CLASS="TYPE"
>timestamptz</TT
>
to <A
HREF="datatype-json.html"
><TT
CLASS="TYPE"
>JSON</TT
></A
>, render the
values in a format compliant with ISO 8601 (Andrew Dunstan)
</P
><P
> Previously such values were rendered according to the current
<A
HREF="runtime-config-client.html#GUC-DATESTYLE"
>DateStyle</A
> setting; but many JSON processors
require timestamps to be in ISO 8601 format. If necessary, the
previous behavior can be obtained by explicitly casting the datetime
value to <TT
CLASS="TYPE"
>text</TT
> before passing it to the JSON conversion
function.
</P
></LI
><LI
><P
> The <A
HREF="functions-json.html#FUNCTIONS-JSON-OP-TABLE"
><TT
CLASS="TYPE"
>json</TT
>
<TT
CLASS="LITERAL"
>#></TT
> <TT
CLASS="TYPE"
>text[]</TT
></A
> path extraction operator now
returns its lefthand input, not NULL, if the array is empty (Tom Lane)
</P
><P
> This is consistent with the notion that this represents zero
applications of the simple field/element extraction
operator <TT
CLASS="LITERAL"
>-></TT
>. Similarly, <TT
CLASS="TYPE"
>json</TT
>
<TT
CLASS="LITERAL"
>#>></TT
> <TT
CLASS="TYPE"
>text[]</TT
> with an empty array merely
coerces its lefthand input to text.
</P
></LI
><LI
><P
> Corner cases in
the <A
HREF="functions-json.html#FUNCTIONS-JSON-OP-TABLE"
><TT
CLASS="TYPE"
>JSON</TT
>
field/element/path extraction operators</A
> now return NULL rather
than raising an error (Tom Lane)
</P
><P
> For example, applying field extraction to a JSON array now yields NULL
not an error. This is more consistent (since some comparable cases such
as no-such-field already returned NULL), and it makes it safe to create
expression indexes that use these operators, since they will now not
throw errors for any valid JSON input.
</P
></LI
><LI
><P
> Cause consecutive whitespace in <A
HREF="functions-formatting.html#FUNCTIONS-FORMATTING-TABLE"
><CODE
CLASS="FUNCTION"
>to_timestamp()</CODE
></A
>
and <CODE
CLASS="FUNCTION"
>to_date()</CODE
> format strings to consume a corresponding
number of characters in the input string (whitespace or not), then
conditionally consume adjacent whitespace, if not in <TT
CLASS="LITERAL"
>FX</TT
>
mode (Jeevan Chalke)
</P
><P
> Previously, consecutive whitespace characters in a non-<TT
CLASS="LITERAL"
>FX</TT
>
format string behaved like a single whitespace character and consumed
all adjacent whitespace in the input string. For example, previously
a format string of three spaces would consume only the first space in
<TT
CLASS="LITERAL"
>' 12'</TT
>, but it will now consume all three characters.
</P
></LI
><LI
><P
> Fix <A
HREF="functions-textsearch.html#TEXTSEARCH-FUNCTIONS-TABLE"
><CODE
CLASS="FUNCTION"
>ts_rank_cd()</CODE
></A
>
to ignore stripped lexemes (Alex Hill)
</P
><P
> Previously, stripped lexemes were treated as if they had a default
location, producing a rank of dubious usefulness.
</P
></LI
><LI
><P
> For functions declared to
take <A
HREF="xfunc-sql.html#XFUNC-SQL-VARIADIC-FUNCTIONS"
><TT
CLASS="LITERAL"
>VARIADIC
"any"</TT
></A
>, an actual parameter marked as <TT
CLASS="LITERAL"
>VARIADIC</TT
>
must be of a determinable array type (Pavel Stehule)
</P
><P
> Such parameters can no longer be written as an undecorated string
literal or <TT
CLASS="LITERAL"
>NULL</TT
>; a cast to an appropriate array data type
will now be required. Note that this does not affect parameters not
marked <TT
CLASS="LITERAL"
>VARIADIC</TT
>.
</P
></LI
><LI
><P
> Ensure that whole-row variables expose the expected column names
to functions that pay attention to column names within composite
arguments (Tom Lane)
</P
><P
> Constructs like <TT
CLASS="LITERAL"
>row_to_json(tab.*)</TT
> now always emit column
names that match the column aliases visible for table <TT
CLASS="LITERAL"
>tab</TT
>
at the point of the call. In previous releases the emitted column
names would sometimes be the table's actual column names regardless
of any aliases assigned in the query.
</P
></LI
><LI
><P
> <A
HREF="sql-discard.html"
>DISCARD</A
> now also discards sequence-related state
(Fabrízio de Royes Mello, Robert Haas)
</P
></LI
><LI
><P
> Rename <A
HREF="sql-explain.html"
><TT
CLASS="COMMAND"
>EXPLAIN
ANALYZE</TT
></A
>'s <SPAN
CLASS="QUOTE"
>"total runtime"</SPAN
> output
to <SPAN
CLASS="QUOTE"
>"execution time"</SPAN
> (Tom Lane)
</P
><P
> Now that planning time is also reported, the previous name was
confusing.
</P
></LI
><LI
><P
> <A
HREF="sql-show.html"
><TT
CLASS="COMMAND"
>SHOW TIME ZONE</TT
></A
> now
outputs simple numeric UTC offsets in <ACRONYM
CLASS="ACRONYM"
>POSIX</ACRONYM
> timezone
format (Tom Lane)
</P
><P
> Previously, such timezone settings were displayed as <A
HREF="datatype-datetime.html#DATATYPE-INTERVAL-OUTPUT"
><TT
CLASS="TYPE"
>interval</TT
></A
> values.
The new output is properly interpreted by <TT
CLASS="COMMAND"
>SET TIME ZONE</TT
>
when passed as a simple string, whereas the old output required
special treatment to be re-parsed correctly.
</P
></LI
><LI
><P
> Foreign data wrappers that support updating foreign tables must
consider the possible presence of <TT
CLASS="LITERAL"
>AFTER ROW</TT
> triggers
(Noah Misch)
</P
><P
> When an <TT
CLASS="LITERAL"
>AFTER ROW</TT
> trigger is present, all columns of the
table must be returned by updating actions, since the trigger might
inspect any or all of them. Previously, foreign tables never had
triggers, so the FDW might optimize away fetching columns not mentioned
in the <TT
CLASS="LITERAL"
>RETURNING</TT
> clause (if any).
</P
></LI
><LI
><P
> Prevent <A
HREF="ddl-constraints.html#DDL-CONSTRAINTS-CHECK-CONSTRAINTS"
><TT
CLASS="LITERAL"
>CHECK</TT
></A
>
constraints from referencing system columns, except
<TT
CLASS="STRUCTFIELD"
>tableoid</TT
> (Amit Kapila)
</P
><P
> Previously such check constraints were allowed, but they would often
cause errors during restores.
</P
></LI
><LI
><P
> Use the last specified <A
HREF="recovery-target-settings.html"
>recovery
target parameter</A
> if multiple target parameters are specified
(Heikki Linnakangas)
</P
><P
> Previously, there was an undocumented precedence order among
the <TT
CLASS="LITERAL"
>recovery_target_<TT
CLASS="REPLACEABLE"
><I
>xxx</I
></TT
></TT
> parameters.
</P
></LI
><LI
><P
> On Windows, automatically preserve quotes in command strings supplied
by the user (Heikki Linnakangas)
</P
><P
> User commands that did their own quote preservation might need
adjustment. This is likely to be an issue for commands used in
<A
HREF="runtime-config-wal.html#GUC-ARCHIVE-COMMAND"
>archive_command</A
>, <A
HREF="archive-recovery-settings.html#RESTORE-COMMAND"
>restore_command</A
>,
and <A
HREF="sql-copy.html"
><TT
CLASS="COMMAND"
>COPY TO/FROM PROGRAM</TT
></A
>.
</P
></LI
><LI
><P
> Remove catalog column <A
HREF="catalog-pg-class.html"
><TT
CLASS="STRUCTFIELD"
>pg_class.reltoastidxid</TT
></A
>
(Michael Paquier)
</P
></LI
><LI
><P
> Remove catalog column <A
HREF="catalog-pg-rewrite.html"
><TT
CLASS="STRUCTFIELD"
>pg_rewrite.ev_attr</TT
></A
>
(Kevin Grittner)
</P
><P
> Per-column rules have not been supported since
<SPAN
CLASS="APPLICATION"
>PostgreSQL</SPAN
> 7.3.
</P
></LI
><LI
><P
> Remove native support for <SPAN
CLASS="APPLICATION"
>Kerberos</SPAN
> authentication
(<TT
CLASS="OPTION"
>--with-krb5</TT
>, etc)
(Magnus Hagander)
</P
><P
> The supported way to use <SPAN
CLASS="APPLICATION"
>Kerberos</SPAN
> authentication is
with <ACRONYM
CLASS="ACRONYM"
>GSSAPI</ACRONYM
>. The native code has been deprecated since
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 8.3.
</P
></LI
><LI
><P
> In <SPAN
CLASS="APPLICATION"
>PL/Python</SPAN
>, handle domains over arrays like the
underlying array type (Rodolfo Campero)
</P
><P
> Previously such values were treated as strings.
</P
></LI
><LI
><P
> Make libpq's <A
HREF="libpq-connect.html#LIBPQ-PQCONNECTDBPARAMS"
><CODE
CLASS="FUNCTION"
>PQconnectdbParams()</CODE
></A
>
and <A
HREF="libpq-connect.html#LIBPQ-PQPINGPARAMS"
><CODE
CLASS="FUNCTION"
>PQpingParams()</CODE
></A
>
functions process zero-length strings as defaults (Adrian
Vondendriesch)
</P
><P
> Previously, these functions treated zero-length string values as
selecting the default in only some cases.
</P
></LI
><LI
><P
> Change empty arrays returned by the <A
HREF="intarray.html"
>intarray</A
> module
to be zero-dimensional arrays (Bruce Momjian)
</P
><P
> Previously, empty arrays were returned as zero-length one-dimensional
arrays, whose text representation looked the same as zero-dimensional
arrays (<TT
CLASS="LITERAL"
>{}</TT
>), but they acted differently in array
operations. <SPAN
CLASS="APPLICATION"
>intarray</SPAN
>'s behavior in this area now
matches the built-in array operators.
</P
></LI
><LI
><P
> <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
> now uses <TT
CLASS="OPTION"
>-U</TT
>
or <TT
CLASS="OPTION"
>--username</TT
> to specify the user name (Bruce Momjian)
</P
><P
> Previously this option was spelled <TT
CLASS="OPTION"
>-u</TT
> or <TT
CLASS="OPTION"
>--user</TT
>,
but that was inconsistent with other tools.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN138093"
>E.28.3. Changes</A
></H2
><P
> Below you will find a detailed account of the changes between
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 9.4 and the previous major
release.
</P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138097"
>E.28.3.1. Server</A
></H3
><P
></P
><UL
><LI
><P
> Allow <A
HREF="bgworker.html"
>background worker processes</A
>
to be dynamically registered, started and terminated (Robert Haas)
</P
><P
> The new <TT
CLASS="FILENAME"
>worker_spi</TT
> module shows an example of use
of this feature.
</P
></LI
><LI
><P
> Allow dynamic allocation of shared memory segments (Robert Haas,
Amit Kapila)
</P
><P
> This feature is illustrated in the <TT
CLASS="FILENAME"
>test_shm_mq</TT
>
module.
</P
></LI
><LI
><P
> During crash recovery or immediate shutdown, send uncatchable
termination signals (<SPAN
CLASS="SYSTEMITEM"
>SIGKILL</SPAN
>) to child processes
that do not shut down promptly (MauMau, Álvaro Herrera)
</P
><P
> This reduces the likelihood of leaving orphaned child processes
behind after <A
HREF="app-postmaster.html"
><SPAN
CLASS="APPLICATION"
>postmaster</SPAN
></A
> shutdown, as well
as ensuring that crash recovery can proceed if some child processes
have become <SPAN
CLASS="QUOTE"
>"stuck"</SPAN
>.
</P
></LI
><LI
><P
> Improve randomness of the database system identifier (Tom Lane)
</P
></LI
><LI
><P
> Make <A
HREF="sql-vacuum.html"
>VACUUM</A
> properly report dead but
not-yet-removable rows to the statistics collector (Hari Babu)
</P
><P
> Previously these were reported as live rows.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138121"
>E.28.3.1.1. Indexes</A
></H4
><P
></P
><UL
><LI
><P
> Reduce <A
HREF="gin.html"
><ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
></A
> index size
(Alexander Korotkov, Heikki Linnakangas)
</P
><P
> Indexes upgraded via <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
> will work fine
but will still be in the old, larger <ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
> format.
Use <A
HREF="sql-reindex.html"
>REINDEX</A
> to recreate old GIN indexes in the
new format.
</P
></LI
><LI
><P
> Improve speed of multi-key <A
HREF="gin.html"
><ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
></A
> lookups (Alexander Korotkov,
Heikki Linnakangas)
</P
></LI
><LI
><P
> Add <A
HREF="gist.html"
><ACRONYM
CLASS="ACRONYM"
>GiST</ACRONYM
></A
> index support
for <A
HREF="datatype-net-types.html#DATATYPE-INET"
><TT
CLASS="TYPE"
>inet</TT
></A
> and
<A
HREF="datatype-net-types.html#DATATYPE-CIDR"
><TT
CLASS="TYPE"
>cidr</TT
></A
> data types
(Emre Hasegeli)
</P
><P
> Such indexes improve <A
HREF="functions-net.html#CIDR-INET-OPERATORS-TABLE"
>subnet and supernet</A
>
lookups and ordering comparisons.
</P
></LI
><LI
><P
> Fix rare race condition in B-tree page deletion (Heikki Linnakangas)
</P
></LI
><LI
><P
> Make the handling of interrupted B-tree page splits more robust
(Heikki Linnakangas)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138150"
>E.28.3.1.2. General Performance</A
></H4
><P
></P
><UL
><LI
><P
> Allow multiple backends to insert
into <A
HREF="wal.html"
><ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
></A
> buffers
concurrently (Heikki Linnakangas)
</P
><P
> This improves parallel write performance.
</P
></LI
><LI
><P
> Conditionally write only the modified portion of updated rows to
<A
HREF="wal.html"
><ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
></A
> (Amit Kapila)
</P
></LI
><LI
><P
> Improve performance of aggregate functions used as <A
HREF="sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS"
>window functions</A
>
(David Rowley, Florian Pflug, Tom Lane)
</P
></LI
><LI
><P
> Improve speed of aggregates that
use <A
HREF="datatype-numeric.html"
><TT
CLASS="TYPE"
>numeric</TT
></A
> state
values (Hadi Moshayedi)
</P
></LI
><LI
><P
> Attempt to <A
HREF="routine-vacuuming.html#VACUUM-FOR-WRAPAROUND"
>freeze</A
>
tuples when tables are rewritten with <A
HREF="sql-cluster.html"
>CLUSTER</A
> or <A
HREF="sql-vacuum.html"
><TT
CLASS="COMMAND"
>VACUUM FULL</TT
></A
> (Robert Haas,
Andres Freund)
</P
><P
> This can avoid the need to freeze the tuples in the future.
</P
></LI
><LI
><P
> Improve speed of <A
HREF="sql-copy.html"
>COPY</A
> with default <A
HREF="functions-sequence.html#FUNCTIONS-SEQUENCE-TABLE"
><CODE
CLASS="FUNCTION"
>nextval()</CODE
></A
>
columns (Simon Riggs)
</P
></LI
><LI
><P
> Improve speed of accessing many different <A
HREF="sql-createsequence.html"
>sequences</A
> in the same session
(David Rowley)
</P
></LI
><LI
><P
> Raise hard limit on the number of tuples held in memory during sorting
and B-tree index builds (Noah Misch)
</P
></LI
><LI
><P
> Reduce memory allocated by <SPAN
CLASS="APPLICATION"
>PL/pgSQL</SPAN
>
<A
HREF="sql-do.html"
>DO</A
> blocks (Tom Lane)
</P
></LI
><LI
><P
> Make the planner more aggressive about extracting restriction clauses
from mixed <TT
CLASS="LITERAL"
>AND</TT
>/<TT
CLASS="LITERAL"
>OR</TT
> clauses (Tom Lane)
</P
></LI
><LI
><P
> Disallow pushing volatile <TT
CLASS="LITERAL"
>WHERE</TT
> clauses down
into <TT
CLASS="LITERAL"
>DISTINCT</TT
> subqueries (Tom Lane)
</P
><P
> Pushing down a <TT
CLASS="LITERAL"
>WHERE</TT
> clause can produce a more
efficient plan overall, but at the cost of evaluating the clause
more often than is implied by the text of the query; so don't do it
if the clause contains any volatile functions.
</P
></LI
><LI
><P
> Auto-resize the catalog caches (Heikki Linnakangas)
</P
><P
> This reduces memory consumption for sessions accessing only a few
tables, and improves performance for sessions accessing many tables.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138203"
>E.28.3.1.3. Monitoring</A
></H4
><P
></P
><UL
><LI
><P
> Add <A
HREF="monitoring-stats.html#PG-STAT-ARCHIVER-VIEW"
>pg_stat_archiver</A
> system view to
report <A
HREF="wal.html"
><ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
></A
> archiver activity
(Gabriele Bartolini)
</P
></LI
><LI
><P
> Add <TT
CLASS="STRUCTFIELD"
>n_mod_since_analyze</TT
> columns to
<A
HREF="monitoring-stats.html#PG-STAT-ALL-TABLES-VIEW"
>pg_stat_all_tables</A
> and related system views
(Mark Kirkwood)
</P
><P
> These columns expose the system's estimate of the number of changed
tuples since the table's last <A
HREF="sql-analyze.html"
>ANALYZE</A
>. This
estimate drives decisions about when to auto-analyze.
</P
></LI
><LI
><P
> Add <TT
CLASS="STRUCTFIELD"
>backend_xid</TT
> and <TT
CLASS="STRUCTFIELD"
>backend_xmin</TT
>
columns to the system view <A
HREF="monitoring-stats.html#PG-STAT-ACTIVITY-VIEW"
>pg_stat_activity</A
>,
and a <TT
CLASS="STRUCTFIELD"
>backend_xmin</TT
> column to
<A
HREF="monitoring-stats.html#PG-STAT-REPLICATION-VIEW"
>pg_stat_replication</A
> (Christian Kruse)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138224"
>E.28.3.1.4. <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
></A
></H4
><P
></P
><UL
><LI
><P
> Add support for <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> <ACRONYM
CLASS="ACRONYM"
>ECDH</ACRONYM
> key exchange
(Marko Kreen)
</P
><P
> This allows use of Elliptic Curve keys for server authentication.
Such keys are faster and have better security than <ACRONYM
CLASS="ACRONYM"
>RSA</ACRONYM
>
keys. The new configuration parameter
<A
HREF="runtime-config-connection.html#GUC-SSL-ECDH-CURVE"
>ssl_ecdh_curve</A
>
controls which curve is used for <ACRONYM
CLASS="ACRONYM"
>ECDH</ACRONYM
>.
</P
></LI
><LI
><P
> Improve the default <A
HREF="runtime-config-connection.html#GUC-SSL-CIPHERS"
>ssl_ciphers</A
> setting
(Marko Kreen)
</P
></LI
><LI
><P
> By default, the server not the client now controls the preference
order of <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> ciphers
(Marko Kreen)
</P
><P
> Previously, the order specified by <A
HREF="runtime-config-connection.html#GUC-SSL-CIPHERS"
>ssl_ciphers</A
>
was usually ignored in favor of client-side defaults, which are not
configurable in most <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> clients. If
desired, the old behavior can be restored via the new configuration
parameter <A
HREF="runtime-config-connection.html#GUC-SSL-PREFER-SERVER-CIPHERS"
>ssl_prefer_server_ciphers</A
>.
</P
></LI
><LI
><P
> Make <A
HREF="runtime-config-logging.html#GUC-LOG-CONNECTIONS"
>log_connections</A
> show <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
>
encryption information (Andreas Kunert)
</P
></LI
><LI
><P
> Improve <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> renegotiation handling (Álvaro
Herrera)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138253"
>E.28.3.1.5. Server Settings</A
></H4
><P
></P
><UL
><LI
><P
> Add new <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> command <A
HREF="sql-altersystem.html"
>ALTER SYSTEM</A
>
for changing <TT
CLASS="FILENAME"
>postgresql.conf</TT
> configuration file entries
(Amit Kapila)
</P
><P
> Previously such settings could only be changed by manually
editing <TT
CLASS="FILENAME"
>postgresql.conf</TT
>.
</P
></LI
><LI
><P
> Add <A
HREF="runtime-config-resource.html#GUC-AUTOVACUUM-WORK-MEM"
>autovacuum_work_mem</A
> configuration parameter
to control the amount of memory used by autovacuum workers
(Peter Geoghegan)
</P
></LI
><LI
><P
> Add <A
HREF="runtime-config-resource.html#GUC-HUGE-PAGES"
>huge_pages</A
> parameter to allow using huge
memory pages on Linux (Christian Kruse, Richard Poole, Abhijit
Menon-Sen)
</P
><P
> This can improve performance on large-memory systems.
</P
></LI
><LI
><P
> Add <A
HREF="runtime-config-resource.html#GUC-MAX-WORKER-PROCESSES"
>max_worker_processes</A
> parameter
to limit the number of background workers (Robert Haas)
</P
><P
> This is helpful in configuring a standby server to have the
required number of worker processes (the same as the primary).
</P
></LI
><LI
><P
> Add superuser-only <A
HREF="runtime-config-client.html#GUC-SESSION-PRELOAD-LIBRARIES"
>session_preload_libraries</A
>
parameter to load libraries at session start (Peter Eisentraut)
</P
><P
> In contrast to <A
HREF="runtime-config-client.html#GUC-LOCAL-PRELOAD-LIBRARIES"
>local_preload_libraries</A
>, this
parameter can load any shared library, not just those in
the <TT
CLASS="FILENAME"
>$libdir/plugins</TT
> directory.
</P
></LI
><LI
><P
> Add <A
HREF="runtime-config-wal.html#GUC-WAL-LOG-HINTS"
>wal_log_hints</A
> parameter to enable WAL
logging of hint-bit changes (Sawada Masahiko)
</P
><P
> Hint bit changes are not normally logged, except when checksums are
enabled. This is useful for external tools
like <SPAN
CLASS="APPLICATION"
>pg_rewind</SPAN
>.
</P
></LI
><LI
><P
> Increase the default settings of <A
HREF="runtime-config-resource.html#GUC-WORK-MEM"
>work_mem</A
>
and <A
HREF="runtime-config-resource.html#GUC-MAINTENANCE-WORK-MEM"
>maintenance_work_mem</A
> by four times (Bruce
Momjian)
</P
><P
> The new defaults are 4MB and 64MB respectively.
</P
></LI
><LI
><P
> Increase the default setting of <A
HREF="runtime-config-query.html#GUC-EFFECTIVE-CACHE-SIZE"
>effective_cache_size</A
>
to 4GB (Bruce Momjian, Tom Lane)
</P
></LI
><LI
><P
> Allow <CODE
CLASS="FUNCTION"
>printf</CODE
>-style space padding to be
specified in <A
HREF="runtime-config-logging.html#GUC-LOG-LINE-PREFIX"
>log_line_prefix</A
> (David Rowley)
</P
></LI
><LI
><P
> Allow terabyte units (<TT
CLASS="LITERAL"
>TB</TT
>) to be used when specifying
configuration variable values (Simon Riggs)
</P
></LI
><LI
><P
> Show <ACRONYM
CLASS="ACRONYM"
>PID</ACRONYM
>s of lock holders and waiters and improve
information about relations in <A
HREF="runtime-config-logging.html#GUC-LOG-LOCK-WAITS"
>log_lock_waits</A
>
log messages (Christian Kruse)
</P
></LI
><LI
><P
> Reduce server logging level when loading shared libraries (Peter
Geoghegan)
</P
><P
> The previous level was <TT
CLASS="LITERAL"
>LOG</TT
>, which was too verbose
for libraries loaded per-session.
</P
></LI
><LI
><P
> On Windows, make <TT
CLASS="LITERAL"
>SQL_ASCII</TT
>-encoded databases and server
processes (e.g., <A
HREF="app-postmaster.html"
><SPAN
CLASS="APPLICATION"
>postmaster</SPAN
></A
>) emit messages in
the character encoding of the server's Windows user locale
(Alexander Law, Noah Misch)
</P
><P
> Previously these messages were output in the Windows
<ACRONYM
CLASS="ACRONYM"
>ANSI</ACRONYM
> code page.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138314"
>E.28.3.2. Replication and Recovery</A
></H3
><P
></P
><UL
><LI
><P
> Add <A
HREF="warm-standby.html#STREAMING-REPLICATION-SLOTS"
>replication
slots</A
> to coordinate activity on streaming standbys with the
node they are streaming from (Andres Freund, Robert Haas)
</P
><P
> Replication slots allow preservation of resources like
<ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> files on the primary until they are no longer
needed by standby servers.
</P
></LI
><LI
><P
> Add recovery parameter <A
HREF="standby-settings.html#RECOVERY-MIN-APPLY-DELAY"
>recovery_min_apply_delay</A
>
to delay replication (Robert Haas, Fabrízio de Royes Mello,
Simon Riggs)
</P
><P
> Delaying replay on standby servers can be useful for recovering
from user errors.
</P
></LI
><LI
><P
> Add <A
HREF="recovery-target-settings.html#RECOVERY-TARGET"
>recovery_target</A
>
option <TT
CLASS="OPTION"
>immediate</TT
> to stop <A
HREF="wal.html"
><ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
></A
> recovery as soon as a
consistent state is reached (MauMau, Heikki Linnakangas)
</P
></LI
><LI
><P
> Improve recovery target processing (Heikki Linnakangas)
</P
><P
> The timestamp reported
by <A
HREF="functions-admin.html#FUNCTIONS-RECOVERY-INFO-TABLE"
><CODE
CLASS="FUNCTION"
>pg_last_xact_replay_timestamp()</CODE
></A
>
now reflects already-committed records, not transactions about to
be committed. Recovering to a restore point now replays the restore
point, rather than stopping just before the restore point.
</P
></LI
><LI
><P
> <A
HREF="functions-admin.html#FUNCTIONS-ADMIN-BACKUP-TABLE"
><CODE
CLASS="FUNCTION"
>pg_switch_xlog()</CODE
></A
>
now clears any unused trailing space in the old <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> file
(Heikki Linnakangas)
</P
><P
> This improves the compression ratio for <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> files.
</P
></LI
><LI
><P
> Report failure return codes from <A
HREF="archive-recovery-settings.html"
>external recovery commands</A
>
(Peter Eisentraut)
</P
></LI
><LI
><P
> Reduce spinlock contention during <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> replay (Heikki
Linnakangas)
</P
></LI
><LI
><P
> Write <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> records of running transactions more
frequently (Andres Freund)
</P
><P
> This allows standby servers to start faster and clean up resources
more aggressively.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138354"
>E.28.3.2.1. <A
HREF="logicaldecoding.html"
>Logical Decoding</A
></A
></H4
><P
> Logical decoding allows database changes to be streamed in a
configurable format. The data is read from
the <A
HREF="wal.html"
><ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
></A
> and transformed into the
desired target format. To implement this feature, the following changes
were made:
</P
><P
></P
><UL
><LI
><P
> Add support for <A
HREF="logicaldecoding.html"
>logical decoding</A
>
of WAL data, to allow database changes to be streamed out in a
customizable format
(Andres Freund)
</P
></LI
><LI
><P
> Add new <A
HREF="runtime-config-wal.html#GUC-WAL-LEVEL"
>wal_level</A
> setting <TT
CLASS="OPTION"
>logical</TT
>
to enable logical change-set encoding in <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> (Andres
Freund)
</P
></LI
><LI
><P
> Add table-level parameter <A
HREF="catalog-pg-class.html"
><TT
CLASS="LITERAL"
>REPLICA IDENTITY</TT
></A
>
to control logical replication (Andres Freund)
</P
></LI
><LI
><P
> Add relation option <A
HREF="sql-createtable.html#SQL-CREATETABLE-STORAGE-PARAMETERS"
><TT
CLASS="OPTION"
>user_catalog_table</TT
></A
>
to identify user-created tables involved in logical change-set
encoding (Andres Freund)
</P
></LI
><LI
><P
> Add <A
HREF="app-pgrecvlogical.html"
><SPAN
CLASS="APPLICATION"
>pg_recvlogical</SPAN
></A
> application to receive
logical-decoding data (Andres Freund)
</P
></LI
><LI
><P
> Add <A
HREF="test-decoding.html"
>test_decoding</A
> module to illustrate logical
decoding at the <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> level (Andres Freund)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138384"
>E.28.3.3. Queries</A
></H3
><P
></P
><UL
><LI
><P
> Add <A
HREF="queries-table-expressions.html#QUERIES-TABLEFUNCTIONS"
><TT
CLASS="LITERAL"
>WITH
ORDINALITY</TT
></A
> syntax to number the rows returned from a
set-returning function in the <TT
CLASS="LITERAL"
>FROM</TT
> clause
(Andrew Gierth, David Fetter)
</P
><P
> This is particularly useful for functions like
<CODE
CLASS="FUNCTION"
>unnest()</CODE
>.
</P
></LI
><LI
><P
> Add <A
HREF="queries-table-expressions.html#QUERIES-TABLEFUNCTIONS"
><TT
CLASS="LITERAL"
>ROWS
FROM()</TT
></A
> syntax to allow horizontal concatenation of
set-returning functions in the <TT
CLASS="LITERAL"
>FROM</TT
> clause (Andrew Gierth)
</P
></LI
><LI
><P
> Allow <A
HREF="sql-select.html"
>SELECT</A
> to have
an empty target list (Tom Lane)
</P
><P
> This was added so that views that select from a table with zero
columns can be dumped and restored correctly.
</P
></LI
><LI
><P
> Ensure that <A
HREF="sql-select.html"
><TT
CLASS="LITERAL"
>SELECT ... FOR UPDATE
NOWAIT</TT
></A
> does not wait in corner cases involving
already-concurrently-updated tuples (Craig Ringer and Thomas Munro)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138407"
>E.28.3.4. Utility Commands</A
></H3
><P
></P
><UL
><LI
><P
> Add <A
HREF="sql-discard.html"
><TT
CLASS="COMMAND"
>DISCARD
SEQUENCES</TT
></A
> command to discard cached sequence-related state
(Fabrízio de Royes Mello, Robert Haas)
</P
><P
> <TT
CLASS="COMMAND"
>DISCARD ALL</TT
> will now also discard such information.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>FORCE NULL</TT
> option
to <A
HREF="sql-copy.html"
><TT
CLASS="COMMAND"
>COPY FROM</TT
></A
>, which
causes quoted strings matching the specified null string to be
converted to NULLs in <TT
CLASS="LITERAL"
>CSV</TT
> mode (Ian Barwick, Michael
Paquier)
</P
><P
> Without this option, only unquoted matching strings will be imported
as null values.
</P
></LI
><LI
><P
> Issue warnings for commands used outside of transaction blocks
when they can have no effect (Bruce Momjian)
</P
><P
> New warnings are issued for <TT
CLASS="COMMAND"
>SET
LOCAL</TT
>, <TT
CLASS="COMMAND"
>SET CONSTRAINTS</TT
>, <TT
CLASS="COMMAND"
>SET TRANSACTION</TT
> and
<TT
CLASS="COMMAND"
>ABORT</TT
> when used outside a transaction block.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138430"
>E.28.3.4.1. <A
HREF="sql-explain.html"
>EXPLAIN</A
></A
></H4
><P
></P
><UL
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXPLAIN ANALYZE</TT
> show planning time (Andreas
Karlsson)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXPLAIN</TT
> show the grouping columns in Agg and
Group nodes (Tom Lane)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXPLAIN ANALYZE</TT
> show exact and lossy
block counts in bitmap heap scans (Etsuro Fujita)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138443"
>E.28.3.4.2. Views</A
></H4
><P
></P
><UL
><LI
><P
> Allow a <A
HREF="rules-materializedviews.html"
>materialized view</A
>
to be refreshed without blocking other sessions from reading the view
meanwhile (Kevin Grittner)
</P
><P
> This is done with <A
HREF="sql-refreshmaterializedview.html"
><TT
CLASS="COMMAND"
>REFRESH MATERIALIZED
VIEW CONCURRENTLY</TT
></A
>.
</P
></LI
><LI
><P
> Allow views to be <A
HREF="sql-createview.html#SQL-CREATEVIEW-UPDATABLE-VIEWS"
>automatically
updated</A
> even if they contain some non-updatable columns
(Dean Rasheed)
</P
><P
> Previously the presence of non-updatable output columns such as
expressions, literals, and function calls prevented automatic
updates. Now <TT
CLASS="COMMAND"
>INSERT</TT
>s, <TT
CLASS="COMMAND"
>UPDATE</TT
>s and
<TT
CLASS="COMMAND"
>DELETE</TT
>s are supported, provided that they do not
attempt to assign new values to any of the non-updatable columns.
</P
></LI
><LI
><P
> Allow control over whether <TT
CLASS="COMMAND"
>INSERT</TT
>s and
<TT
CLASS="COMMAND"
>UPDATE</TT
>s can add rows to an auto-updatable view that
would not appear in the view (Dean Rasheed)
</P
><P
> This is controlled with the new <A
HREF="sql-createview.html"
>CREATE VIEW</A
>
clause <TT
CLASS="LITERAL"
>WITH CHECK OPTION</TT
>.
</P
></LI
><LI
><P
> Allow <A
HREF="rules-privileges.html"
>security barrier views</A
>
to be automatically updatable (Dean Rasheed)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138469"
>E.28.3.5. Object Manipulation</A
></H3
><P
></P
><UL
><LI
><P
> Support triggers on <A
HREF="sql-createforeigntable.html"
>foreign
tables</A
> (Ronan Dunklau)
</P
></LI
><LI
><P
> Allow moving groups of objects from one tablespace to another
using the <TT
CLASS="LITERAL"
>ALL IN TABLESPACE ... SET TABLESPACE</TT
> form of
<A
HREF="sql-altertable.html"
>ALTER TABLE</A
>, <A
HREF="sql-alterindex.html"
>ALTER INDEX</A
>, or
<A
HREF="sql-altermaterializedview.html"
>ALTER MATERIALIZED VIEW</A
> (Stephen Frost)
</P
></LI
><LI
><P
> Allow changing foreign key constraint deferrability
via <A
HREF="sql-altertable.html"
>ALTER TABLE</A
> ... <TT
CLASS="LITERAL"
>ALTER
CONSTRAINT</TT
> (Simon Riggs)
</P
></LI
><LI
><P
> Reduce lock strength for some <A
HREF="sql-altertable.html"
>ALTER TABLE</A
>
commands
(Simon Riggs, Noah Misch, Robert Haas)
</P
><P
> Specifically, <TT
CLASS="LITERAL"
>VALIDATE CONSTRAINT</TT
>, <TT
CLASS="LITERAL"
>CLUSTER
ON</TT
>, <TT
CLASS="LITERAL"
>SET WITHOUT CLUSTER</TT
>, <TT
CLASS="LITERAL"
>ALTER COLUMN
SET STATISTICS</TT
>, <TT
CLASS="LITERAL"
>ALTER COLUMN</TT
> <TT
CLASS="LITERAL"
>SET</TT
>
<TT
CLASS="OPTION"
>(attribute_option)</TT
>, <TT
CLASS="LITERAL"
>ALTER COLUMN RESET</TT
>
<TT
CLASS="OPTION"
>(attribute_option)</TT
> no longer require <TT
CLASS="LITERAL"
>ACCESS
EXCLUSIVE</TT
> locks.
</P
></LI
><LI
><P
> Allow tablespace options to be set
in <A
HREF="sql-createtablespace.html"
>CREATE TABLESPACE</A
> (Vik Fearing)
</P
><P
> Formerly these options could only be set
via <A
HREF="sql-altertablespace.html"
>ALTER TABLESPACE</A
>.
</P
></LI
><LI
><P
> Allow <A
HREF="sql-createaggregate.html"
>CREATE AGGREGATE</A
> to define the estimated
size of the aggregate's transition state data (Hadi Moshayedi)
</P
><P
> Proper use of this feature allows the planner to better estimate
how much memory will be used by aggregates.
</P
></LI
><LI
><P
> Fix <TT
CLASS="COMMAND"
>DROP IF EXISTS</TT
> to avoid errors for non-existent
objects in more cases (Pavel Stehule, Dean Rasheed)
</P
></LI
><LI
><P
> Improve how system relations are identified (Andres Freund,
Robert Haas)
</P
><P
> Previously, relations once moved into the <TT
CLASS="LITERAL"
>pg_catalog</TT
>
schema could no longer be modified or dropped.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138515"
>E.28.3.6. Data Types</A
></H3
><P
></P
><UL
><LI
><P
> Fully implement the <A
HREF="datatype-geometric.html#DATATYPE-LINE"
><TT
CLASS="TYPE"
>line</TT
></A
> data type (Peter
Eisentraut)
</P
><P
> The line <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>segment</I
></SPAN
> data type (<A
HREF="datatype-geometric.html#DATATYPE-LSEG"
><TT
CLASS="TYPE"
>lseg</TT
></A
>) has always been
fully supported. The previous <TT
CLASS="TYPE"
>line</TT
> data type (which was
enabled only via a compile-time option) is not binary or
dump-compatible with the new implementation.
</P
></LI
><LI
><P
> Add <A
HREF="datatype-pg-lsn.html"
><TT
CLASS="TYPE"
>pg_lsn</TT
></A
>
data type to represent a <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> log sequence number
(<ACRONYM
CLASS="ACRONYM"
>LSN</ACRONYM
>) (Robert Haas, Michael Paquier)
</P
></LI
><LI
><P
> Allow single-point <A
HREF="datatype-geometric.html#DATATYPE-POLYGON"
><TT
CLASS="TYPE"
>polygon</TT
></A
>s to be converted
to <A
HREF="datatype-geometric.html#DATATYPE-CIRCLE"
><TT
CLASS="TYPE"
>circle</TT
></A
>s
(Bruce Momjian)
</P
></LI
><LI
><P
> Support time zone abbreviations that change UTC offset from time to
time (Tom Lane)
</P
><P
> Previously, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> assumed that the UTC offset
associated with a time zone abbreviation (such as <TT
CLASS="LITERAL"
>EST</TT
>)
never changes in the usage of any particular locale. However this
assumption fails in the real world, so introduce the ability for a
zone abbreviation to represent a UTC offset that sometimes changes.
Update the zone abbreviation definition files to make use of this
feature in timezone locales that have changed the UTC offset of their
abbreviations since 1970 (according to the IANA timezone database).
In such timezones, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> will now associate the
correct UTC offset with the abbreviation depending on the given date.
</P
></LI
><LI
><P
> Allow 5+ digit years for non-<ACRONYM
CLASS="ACRONYM"
>ISO</ACRONYM
> <A
HREF="datatype-datetime.html"
><TT
CLASS="TYPE"
>timestamp</TT
></A
> and
<TT
CLASS="TYPE"
>date</TT
> strings, where appropriate (Bruce Momjian)
</P
></LI
><LI
><P
> Add checks for overflow/underflow of <A
HREF="datatype-datetime.html"
><TT
CLASS="TYPE"
>interval</TT
></A
> values
(Bruce Momjian)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138555"
>E.28.3.6.1. <A
HREF="datatype-json.html"
><ACRONYM
CLASS="ACRONYM"
>JSON</ACRONYM
></A
></A
></H4
><P
></P
><UL
><LI
><P
> Add <A
HREF="datatype-json.html"
><TT
CLASS="TYPE"
>jsonb</TT
></A
>, a more
capable and efficient data type for storing <ACRONYM
CLASS="ACRONYM"
>JSON</ACRONYM
> data
(Oleg Bartunov, Teodor Sigaev, Alexander
Korotkov, Peter Geoghegan, Andrew Dunstan)
</P
><P
> This new type allows faster access to values within a JSON
document, and faster and more useful indexing of JSON columns.
Scalar values in <TT
CLASS="TYPE"
>jsonb</TT
> documents are stored as appropriate
scalar SQL types, and the JSON document structure is pre-parsed
rather than being stored as text as in the original <TT
CLASS="TYPE"
>json</TT
>
data type.
</P
></LI
><LI
><P
> Add new JSON functions to allow for the construction
of arbitrarily complex JSON trees (Andrew Dunstan, Laurence Rowe)
</P
><P
> New functions include <A
HREF="functions-json.html#FUNCTIONS-JSON-PROCESSING-TABLE"
><CODE
CLASS="FUNCTION"
>json_array_elements_text()</CODE
></A
>,
<CODE
CLASS="FUNCTION"
>json_build_array()</CODE
>, <CODE
CLASS="FUNCTION"
>json_object()</CODE
>,
<CODE
CLASS="FUNCTION"
>json_object_agg()</CODE
>, <CODE
CLASS="FUNCTION"
>json_to_record()</CODE
>,
and <CODE
CLASS="FUNCTION"
>json_to_recordset()</CODE
>.
</P
></LI
><LI
><P
> Add <A
HREF="functions-json.html#FUNCTIONS-JSON-PROCESSING-TABLE"
><CODE
CLASS="FUNCTION"
>json_typeof()</CODE
></A
>
to return the data type of a <TT
CLASS="TYPE"
>json</TT
> value (Andrew Tipton)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138583"
>E.28.3.7. Functions</A
></H3
><P
></P
><UL
><LI
><P
> Add <A
HREF="functions-datetime.html#FUNCTIONS-DATETIME-DELAY"
><CODE
CLASS="FUNCTION"
>pg_sleep_for(interval)</CODE
></A
>
and <CODE
CLASS="FUNCTION"
>pg_sleep_until(timestamp)</CODE
> to specify
delays more flexibly (Vik Fearing, Julien Rouhaud)
</P
><P
> The existing <CODE
CLASS="FUNCTION"
>pg_sleep()</CODE
> function only supports delays
specified in seconds.
</P
></LI
><LI
><P
> Add <A
HREF="functions-array.html#ARRAY-FUNCTIONS-TABLE"
><CODE
CLASS="FUNCTION"
>cardinality()</CODE
></A
>
function for arrays (Marko Tiikkaja)
</P
><P
> This returns the total number of elements in the array, or zero
for an array with no elements.
</P
></LI
><LI
><P
> Add <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> functions to allow <A
HREF="lo-funcs.html"
>large
object reads/writes</A
> at arbitrary offsets (Pavel Stehule)
</P
></LI
><LI
><P
> Allow <A
HREF="functions-array.html#ARRAY-FUNCTIONS-TABLE"
><CODE
CLASS="FUNCTION"
>unnest()</CODE
></A
>
to take multiple arguments, which are individually unnested then
horizontally concatenated (Andrew Gierth)
</P
></LI
><LI
><P
> Add functions to construct <TT
CLASS="TYPE"
>time</TT
>s, <TT
CLASS="TYPE"
>date</TT
>s,
<TT
CLASS="TYPE"
>timestamp</TT
>s, <TT
CLASS="TYPE"
>timestamptz</TT
>s, and <TT
CLASS="TYPE"
>interval</TT
>s
from individual values, rather than strings (Pavel Stehule)
</P
><P
> These functions' names are prefixed with <TT
CLASS="LITERAL"
>make_</TT
>,
e.g. <A
HREF="functions-datetime.html#FUNCTIONS-DATETIME-TABLE"
><CODE
CLASS="FUNCTION"
>make_date()</CODE
></A
>.
</P
></LI
><LI
><P
> Make <A
HREF="functions-formatting.html#FUNCTIONS-FORMATTING-TABLE"
><CODE
CLASS="FUNCTION"
>to_char()</CODE
></A
>'s
<TT
CLASS="LITERAL"
>TZ</TT
> format specifier return a useful value for simple
numeric time zone offsets (Tom Lane)
</P
><P
> Previously, <TT
CLASS="LITERAL"
>to_char(CURRENT_TIMESTAMP, 'TZ')</TT
> returned
an empty string if the <TT
CLASS="LITERAL"
>timezone</TT
> was set to a constant
like <TT
CLASS="LITERAL"
>-4</TT
>.
</P
></LI
><LI
><P
> Add timezone offset format specifier <TT
CLASS="LITERAL"
>OF</TT
> to <A
HREF="functions-formatting.html#FUNCTIONS-FORMATTING-TABLE"
><CODE
CLASS="FUNCTION"
>to_char()</CODE
></A
>
(Bruce Momjian)
</P
></LI
><LI
><P
> Improve the random seed used for <A
HREF="functions-math.html#FUNCTIONS-MATH-RANDOM-TABLE"
><CODE
CLASS="FUNCTION"
>random()</CODE
></A
>
(Honza Horak)
</P
></LI
><LI
><P
> Tighten validity checking for Unicode code points in <A
HREF="functions-string.html#FUNCTIONS-STRING-OTHER"
><CODE
CLASS="FUNCTION"
>chr(int)</CODE
></A
>
(Tom Lane)
</P
><P
> This function now only accepts values that are valid UTF8 characters
according to RFC 3629.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138640"
>E.28.3.7.1. System Information Functions</A
></H4
><P
></P
><UL
><LI
><P
> Add functions for looking up objects in <TT
CLASS="STRUCTNAME"
>pg_class</TT
>,
<TT
CLASS="STRUCTNAME"
>pg_proc</TT
>, <TT
CLASS="STRUCTNAME"
>pg_type</TT
>, and
<TT
CLASS="STRUCTNAME"
>pg_operator</TT
> that do not generate errors for
non-existent objects (Yugo Nagata, Nozomi Anzai,
Robert Haas)
</P
><P
> For example, <A
HREF="functions-info.html#FUNCTIONS-INFO-CATALOG-TABLE"
><CODE
CLASS="FUNCTION"
>to_regclass()</CODE
></A
>
does a lookup in <TT
CLASS="STRUCTNAME"
>pg_class</TT
> similarly to
the <TT
CLASS="TYPE"
>regclass</TT
> input function, but it returns NULL for a
non-existent object instead of failing.
</P
></LI
><LI
><P
> Add function <A
HREF="functions-admin.html#FUNCTIONS-ADMIN-DBLOCATION"
><CODE
CLASS="FUNCTION"
>pg_filenode_relation()</CODE
></A
>
to allow for more efficient lookup of relation names from filenodes
(Andres Freund)
</P
></LI
><LI
><P
> Add <TT
CLASS="STRUCTFIELD"
>parameter_default</TT
> column to <A
HREF="infoschema-parameters.html"
><TT
CLASS="STRUCTNAME"
>information_schema.parameters</TT
></A
>
view (Peter Eisentraut)
</P
></LI
><LI
><P
> Make <A
HREF="infoschema-schemata.html"
><TT
CLASS="STRUCTNAME"
>information_schema.schemata</TT
></A
>
show all accessible schemas (Peter Eisentraut)
</P
><P
> Previously it only showed schemas owned by the current user.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138668"
>E.28.3.7.2. Aggregates</A
></H4
><P
></P
><UL
><LI
><P
> Add control over which rows are passed
into aggregate functions via the <A
HREF="sql-expressions.html#SYNTAX-AGGREGATES"
><TT
CLASS="LITERAL"
>FILTER</TT
></A
> clause
(David Fetter)
</P
></LI
><LI
><P
> Support ordered-set (<A
HREF="sql-expressions.html#SYNTAX-AGGREGATES"
><TT
CLASS="LITERAL"
>WITHIN GROUP</TT
></A
>)
aggregates (Atri Sharma, Andrew Gierth, Tom Lane)
</P
></LI
><LI
><P
> Add standard ordered-set aggregates <A
HREF="functions-aggregate.html#FUNCTIONS-ORDEREDSET-TABLE"
><CODE
CLASS="FUNCTION"
>percentile_cont()</CODE
></A
>,
<CODE
CLASS="FUNCTION"
>percentile_disc()</CODE
>, <CODE
CLASS="FUNCTION"
>mode()</CODE
>, <A
HREF="functions-aggregate.html#FUNCTIONS-HYPOTHETICAL-TABLE"
><CODE
CLASS="FUNCTION"
>rank()</CODE
></A
>,
<CODE
CLASS="FUNCTION"
>dense_rank()</CODE
>, <CODE
CLASS="FUNCTION"
>percent_rank()</CODE
>, and
<CODE
CLASS="FUNCTION"
>cume_dist()</CODE
>
(Atri Sharma, Andrew Gierth)
</P
></LI
><LI
><P
> Support <A
HREF="xfunc-sql.html#XFUNC-SQL-VARIADIC-FUNCTIONS"
><TT
CLASS="LITERAL"
>VARIADIC</TT
></A
>
aggregate functions (Tom Lane)
</P
></LI
><LI
><P
> Allow polymorphic aggregates to have non-polymorphic state data
types (Tom Lane)
</P
><P
> This allows proper declaration in SQL of aggregates like the built-in
aggregate <CODE
CLASS="FUNCTION"
>array_agg()</CODE
>.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138698"
>E.28.3.8. Server-Side Languages</A
></H3
><P
></P
><UL
><LI
><P
> Add event trigger support to <A
HREF="plperl.html"
>PL/Perl</A
>
and <A
HREF="pltcl.html"
>PL/Tcl</A
> (Dimitri Fontaine)
</P
></LI
><LI
><P
> Convert <A
HREF="datatype-numeric.html"
><TT
CLASS="TYPE"
>numeric</TT
></A
>
values to <TT
CLASS="TYPE"
>decimal</TT
> in <A
HREF="plpython.html"
>PL/Python</A
>
(Szymon Guz, Ronan Dunklau)
</P
><P
> Previously such values were converted to Python <TT
CLASS="TYPE"
>float</TT
> values,
risking loss of precision.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138713"
>E.28.3.8.1. <A
HREF="plpgsql.html"
>PL/pgSQL</A
> Server-Side Language</A
></H4
><P
></P
><UL
><LI
><P
> Add ability to retrieve the current PL/pgSQL call stack
using <A
HREF="plpgsql-control-structures.html#PLPGSQL-CALL-STACK"
><TT
CLASS="COMMAND"
>GET
DIAGNOSTICS</TT
></A
>
(Pavel Stehule, Stephen Frost)
</P
></LI
><LI
><P
> Add option <A
HREF="plpgsql-statements.html#PLPGSQL-STATEMENTS-SQL-ONEROW"
><TT
CLASS="OPTION"
>print_strict_params</TT
></A
>
to display the parameters passed to a query that violated a
<TT
CLASS="LITERAL"
>STRICT</TT
> constraint (Marko Tiikkaja)
</P
></LI
><LI
><P
> Add variables <A
HREF="plpgsql-development-tips.html#PLPGSQL-EXTRA-CHECKS"
><TT
CLASS="VARNAME"
>plpgsql.extra_warnings</TT
></A
>
and <TT
CLASS="VARNAME"
>plpgsql.extra_errors</TT
> to enable additional PL/pgSQL
warnings and errors (Marko Tiikkaja, Petr Jelinek)
</P
><P
> Currently only warnings/errors about shadowed variables are available.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138732"
>E.28.3.9. <A
HREF="libpq.html"
><SPAN
CLASS="APPLICATION"
>libpq</SPAN
></A
></A
></H3
><P
></P
><UL
><LI
><P
> Make libpq's <A
HREF="libpq-connect.html#LIBPQ-PQCONNDEFAULTS"
><CODE
CLASS="FUNCTION"
>PQconndefaults()</CODE
></A
>
function ignore invalid service files (Steve Singer, Bruce Momjian)
</P
><P
> Previously it returned NULL if an incorrect service file was
encountered.
</P
></LI
><LI
><P
> Accept <ACRONYM
CLASS="ACRONYM"
>TLS</ACRONYM
> protocol versions beyond <TT
CLASS="LITERAL"
>TLSv1</TT
>
in libpq (Marko Kreen)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138746"
>E.28.3.10. Client Applications</A
></H3
><P
></P
><UL
><LI
><P
> Add <A
HREF="app-createuser.html"
><SPAN
CLASS="APPLICATION"
>createuser</SPAN
></A
> option <TT
CLASS="OPTION"
>-g</TT
>
to specify role membership (Christopher Browne)
</P
></LI
><LI
><P
> Add <A
HREF="app-vacuumdb.html"
><SPAN
CLASS="APPLICATION"
>vacuumdb</SPAN
></A
>
option <TT
CLASS="OPTION"
>--analyze-in-stages</TT
> to analyze in stages of
increasing granularity (Peter Eisentraut)
</P
><P
> This allows minimal statistics to be created quickly.
</P
></LI
><LI
><P
> Make <A
HREF="app-pgresetxlog.html"
><SPAN
CLASS="APPLICATION"
>pg_resetxlog</SPAN
></A
> with option <TT
CLASS="OPTION"
>-n</TT
>
output current and potentially changed values (Rajeev Rastogi)
</P
></LI
><LI
><P
> Make <A
HREF="app-initdb.html"
>initdb</A
> throw error for incorrect locale
settings, rather than silently falling back to a default choice
(Tom Lane)
</P
></LI
><LI
><P
> Make <A
HREF="app-pg-ctl.html"
><SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
></A
> return exit code <TT
CLASS="LITERAL"
>4</TT
> for
an inaccessible data directory (Amit Kapila, Bruce Momjian)
</P
><P
> This behavior more closely matches the Linux Standard Base
(<ACRONYM
CLASS="ACRONYM"
>LSB</ACRONYM
>) Core Specification.
</P
></LI
><LI
><P
> On Windows, ensure that a non-absolute <TT
CLASS="OPTION"
>-D</TT
> path
specification is interpreted relative
to <A
HREF="app-pg-ctl.html"
><SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
></A
>'s current directory
(Kumar Rajeev Rastogi)
</P
><P
> Previously it would be interpreted relative to whichever directory
the underlying Windows service was started in.
</P
></LI
><LI
><P
> Allow <CODE
CLASS="FUNCTION"
>sizeof()</CODE
> in <A
HREF="ecpg.html"
>ECPG</A
>
C array definitions (Michael Meskes)
</P
></LI
><LI
><P
> Make <A
HREF="ecpg.html"
>ECPG</A
> properly handle nesting
of C-style comments in both C and <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> text
(Michael Meskes)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138784"
>E.28.3.10.1. <A
HREF="app-psql.html"
><SPAN
CLASS="APPLICATION"
>psql</SPAN
></A
></A
></H4
><P
></P
><UL
><LI
><P
> Suppress <SPAN
CLASS="QUOTE"
>"No rows"</SPAN
> output in <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <A
HREF="app-psql.html#APP-PSQL-META-COMMANDS"
><TT
CLASS="OPTION"
>expanded</TT
></A
>
mode when the footer is disabled (Bruce Momjian)
</P
></LI
><LI
><P
> Allow Control-C to abort <SPAN
CLASS="APPLICATION"
>psql</SPAN
> when it's hung at
connection startup (Peter Eisentraut)
</P
></LI
></UL
><DIV
CLASS="SECT5"
><H5
CLASS="SECT5"
><A
NAME="AEN138797"
>E.28.3.10.1.1. <A
HREF="app-psql.html#APP-PSQL-META-COMMANDS"
>Backslash Commands</A
></A
></H5
><P
></P
><UL
><LI
><P
> Make <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="COMMAND"
>\db+</TT
> show tablespace options
(Magnus Hagander)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\do+</TT
> display the functions
that implement the operators (Marko Tiikkaja)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\d+</TT
> output an
<TT
CLASS="LITERAL"
>OID</TT
> line only if an <TT
CLASS="LITERAL"
>oid</TT
> column
exists in the table (Bruce Momjian)
</P
><P
> Previously, the presence or absence of an <TT
CLASS="LITERAL"
>oid</TT
>
column was always reported.
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\d</TT
> show disabled system triggers (Bruce
Momjian)
</P
><P
> Previously, if you disabled all triggers, only user triggers
would show as disabled.
</P
></LI
><LI
><P
> Fix <TT
CLASS="COMMAND"
>\copy</TT
> to no longer require
a space between <TT
CLASS="LITERAL"
>stdin</TT
> and a semicolon (Etsuro Fujita)
</P
></LI
><LI
><P
> Output the row count at the end of <TT
CLASS="COMMAND"
>\copy</TT
>, just
like <TT
CLASS="COMMAND"
>COPY</TT
> already did (Kumar Rajeev Rastogi)
</P
></LI
><LI
><P
> Fix <TT
CLASS="COMMAND"
>\conninfo</TT
> to display the
server's <ACRONYM
CLASS="ACRONYM"
>IP</ACRONYM
> address for connections using
<TT
CLASS="LITERAL"
>hostaddr</TT
> (Fujii Masao)
</P
><P
> Previously <TT
CLASS="COMMAND"
>\conninfo</TT
> could not display the server's
<ACRONYM
CLASS="ACRONYM"
>IP</ACRONYM
> address in such cases.
</P
></LI
><LI
><P
> Show the <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> protocol version in
<TT
CLASS="COMMAND"
>\conninfo</TT
> (Marko Kreen)
</P
></LI
><LI
><P
> Add tab completion for <TT
CLASS="COMMAND"
>\pset</TT
>
(Pavel Stehule)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>\pset</TT
> with no arguments
to show all settings (Gilles Darold)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\s</TT
> display the name of the history file it wrote
without converting it to an absolute path (Tom Lane)
</P
><P
> The code previously attempted to convert a relative file name to
an absolute path for display, but frequently got it wrong.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138849"
>E.28.3.10.2. <A
HREF="app-pgdump.html"
>pg_dump</A
></A
></H4
><P
></P
><UL
><LI
><P
> Allow <A
HREF="app-pgrestore.html"
>pg_restore</A
> options
<TT
CLASS="OPTION"
>-I</TT
>, <TT
CLASS="OPTION"
>-P</TT
>, <TT
CLASS="OPTION"
>-T</TT
> and <TT
CLASS="OPTION"
>-n</TT
>
to be specified multiple times (Heikki Linnakangas)
</P
><P
> This allows multiple objects to be restored in one operation.
</P
></LI
><LI
><P
> Optionally add <TT
CLASS="LITERAL"
>IF EXISTS</TT
> clauses to the <TT
CLASS="COMMAND"
>DROP</TT
>
commands emitted when removing old objects during a restore (Pavel
Stehule)
</P
><P
> This change prevents unnecessary errors when removing old objects.
The new <TT
CLASS="OPTION"
>--if-exists</TT
> option
for <A
HREF="app-pgdump.html"
>pg_dump</A
>, <A
HREF="app-pg-dumpall.html"
><SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
></A
>,
and <A
HREF="app-pgrestore.html"
>pg_restore</A
> is only available
when <TT
CLASS="OPTION"
>--clean</TT
> is also specified.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN138871"
>E.28.3.10.3. <A
HREF="app-pgbasebackup.html"
>pg_basebackup</A
></A
></H4
><P
></P
><UL
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>pg_basebackup</SPAN
> option <TT
CLASS="OPTION"
>--xlogdir</TT
>
to specify the <TT
CLASS="FILENAME"
>pg_xlog</TT
> directory location (Haribabu
Kommi)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>pg_basebackup</SPAN
> to relocate tablespaces in
the backup copy (Steeve Lennmark)
</P
><P
> This is particularly useful for using <SPAN
CLASS="APPLICATION"
>pg_basebackup</SPAN
>
on the same machine as the primary.
</P
></LI
><LI
><P
> Allow network-stream base backups to be throttled (Antonin Houska)
</P
><P
> This can be controlled with the <SPAN
CLASS="APPLICATION"
>pg_basebackup</SPAN
>
<TT
CLASS="OPTION"
>--max-rate</TT
> parameter.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN138890"
>E.28.3.11. Source Code</A
></H3
><P
></P
><UL
><LI
><P
> Improve the way tuples are frozen to preserve forensic information
(Robert Haas, Andres Freund)
</P
><P
> This change removes the main objection to freezing tuples as soon
as possible. Code that inspects tuple flag bits will need to be
modified.
</P
></LI
><LI
><P
> No longer require function prototypes for functions marked with the
<A
HREF="xfunc-c.html"
><CODE
CLASS="FUNCTION"
>PG_FUNCTION_INFO_V1</CODE
></A
>
macro (Peter Eisentraut)
</P
><P
> This change eliminates the need to write boilerplate prototypes.
Note that the <CODE
CLASS="FUNCTION"
>PG_FUNCTION_INFO_V1</CODE
> macro must appear
before the corresponding function definition to avoid compiler
warnings.
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>SnapshotNow</TT
> and
<CODE
CLASS="FUNCTION"
>HeapTupleSatisfiesNow()</CODE
> (Robert Haas)
</P
><P
> All existing uses have been switched to more appropriate snapshot
types. Catalog scans now use <ACRONYM
CLASS="ACRONYM"
>MVCC</ACRONYM
> snapshots.
</P
></LI
><LI
><P
> Add an <ACRONYM
CLASS="ACRONYM"
>API</ACRONYM
> to allow memory allocations over one gigabyte
(Noah Misch)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>psprintf()</CODE
> to simplify memory allocation during
string composition (Peter Eisentraut, Tom Lane)
</P
></LI
><LI
><P
> Support <CODE
CLASS="FUNCTION"
>printf()</CODE
> size modifier <TT
CLASS="LITERAL"
>z</TT
> to
print <TT
CLASS="TYPE"
>size_t</TT
> values (Andres Freund)
</P
></LI
><LI
><P
> Change <ACRONYM
CLASS="ACRONYM"
>API</ACRONYM
> of <CODE
CLASS="FUNCTION"
>appendStringInfoVA()</CODE
>
to better use <CODE
CLASS="FUNCTION"
>vsnprintf()</CODE
> (David Rowley, Tom Lane)
</P
></LI
><LI
><P
> Allow new types of external toast datums to be created (Andres
Freund)
</P
></LI
><LI
><P
> Add single-reader, single-writer, lightweight shared message queue
(Robert Haas)
</P
></LI
><LI
><P
> Improve spinlock speed on x86_64 <ACRONYM
CLASS="ACRONYM"
>CPU</ACRONYM
>s (Heikki
Linnakangas)
</P
></LI
><LI
><P
> Remove spinlock support for unsupported platforms
<SPAN
CLASS="PRODUCTNAME"
>SINIX</SPAN
>, <SPAN
CLASS="PRODUCTNAME"
>Sun3</SPAN
>, and
<SPAN
CLASS="PRODUCTNAME"
>NS32K</SPAN
> (Robert Haas)
</P
></LI
><LI
><P
> Remove <ACRONYM
CLASS="ACRONYM"
>IRIX</ACRONYM
> port (Robert Haas)
</P
></LI
><LI
><P
> Reduce the number of semaphores required by
<TT
CLASS="OPTION"
>--disable-spinlocks</TT
> builds (Robert Haas)
</P
></LI
><LI
><P
> Rewrite <SPAN
CLASS="APPLICATION"
>duplicate_oids</SPAN
> Unix shell script in
<SPAN
CLASS="APPLICATION"
>Perl</SPAN
> (Andrew Dunstan)
</P
></LI
><LI
><P
> Add Test Anything Protocol (<ACRONYM
CLASS="ACRONYM"
>TAP</ACRONYM
>) tests for client
programs (Peter Eisentraut)
</P
><P
> Currently, these tests are run by <TT
CLASS="LITERAL"
>make check-world</TT
>
only if the <TT
CLASS="OPTION"
>--enable-tap-tests</TT
> option was given
to <SPAN
CLASS="APPLICATION"
>configure</SPAN
>.
This might become the default behavior in some future release.
</P
></LI
><LI
><P
> Add make targets <TT
CLASS="OPTION"
>check-tests</TT
> and
<TT
CLASS="OPTION"
>installcheck-tests</TT
>, which allow selection of individual
tests to be run (Andrew Dunstan)
</P
></LI
><LI
><P
> Remove <TT
CLASS="OPTION"
>maintainer-check</TT
> makefile rule (Peter Eisentraut)
</P
><P
> The default build rules now include all the formerly-optional tests.
</P
></LI
><LI
><P
> Improve support for <TT
CLASS="ENVAR"
>VPATH</TT
> builds of <ACRONYM
CLASS="ACRONYM"
>PGXS</ACRONYM
>
modules (Cédric Villemain, Andrew Dunstan, Peter Eisentraut)
</P
></LI
><LI
><P
> Upgrade to Autoconf 2.69 (Peter Eisentraut)
</P
></LI
><LI
><P
> Add a <SPAN
CLASS="APPLICATION"
>configure</SPAN
> flag that appends custom text to the
<TT
CLASS="ENVAR"
>PG_VERSION</TT
> string (Oskari Saarenmaa)
</P
><P
> This is useful for packagers building custom binaries.
</P
></LI
><LI
><P
> Improve DocBook <ACRONYM
CLASS="ACRONYM"
>XML</ACRONYM
> validity (Peter Eisentraut)
</P
></LI
><LI
><P
> Fix various minor security and sanity issues reported by the
<SPAN
CLASS="PRODUCTNAME"
>Coverity</SPAN
> scanner (Stephen Frost)
</P
></LI
><LI
><P
> Improve detection of invalid memory usage when testing
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> with <SPAN
CLASS="APPLICATION"
>Valgrind</SPAN
>
(Noah Misch)
</P
></LI
><LI
><P
> Improve sample <SPAN
CLASS="APPLICATION"
>Emacs</SPAN
> configuration file
<TT
CLASS="FILENAME"
>emacs.samples</TT
> (Peter Eisentraut)
</P
><P
> Also add <TT
CLASS="FILENAME"
>.dir-locals.el</TT
> to the top of the source tree.
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>pgindent</SPAN
> to accept a command-line list
of typedefs (Bruce Momjian)
</P
></LI
><LI
><P
> Make <SPAN
CLASS="APPLICATION"
>pgindent</SPAN
> smarter about blank lines
around preprocessor conditionals (Bruce Momjian)
</P
></LI
><LI
><P
> Avoid most uses of <TT
CLASS="COMMAND"
>dlltool</TT
>
in <SPAN
CLASS="PRODUCTNAME"
>Cygwin</SPAN
> and
<SPAN
CLASS="PRODUCTNAME"
>Mingw</SPAN
> builds (Marco Atzeri, Hiroshi Inoue)
</P
></LI
><LI
><P
> Support client-only installs in <ACRONYM
CLASS="ACRONYM"
>MSVC</ACRONYM
> (Windows) builds
(MauMau)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN139002"
>E.28.3.12. Additional Modules</A
></H3
><P
></P
><UL
><LI
><P
> Add <A
HREF="pgprewarm.html"
>pg_prewarm</A
> extension to preload relation data
into the shared buffer cache at server start (Robert Haas)
</P
><P
> This allows reaching full operating performance more quickly.
</P
></LI
><LI
><P
> Add <ACRONYM
CLASS="ACRONYM"
>UUID</ACRONYM
> random number generator
<CODE
CLASS="FUNCTION"
>gen_random_uuid()</CODE
> to <A
HREF="pgcrypto.html"
>pgcrypto</A
>
(Oskari Saarenmaa)
</P
><P
> This allows creation of version 4 <ACRONYM
CLASS="ACRONYM"
>UUID</ACRONYM
>s without
requiring installation of <A
HREF="uuid-ossp.html"
>uuid-ossp</A
>.
</P
></LI
><LI
><P
> Allow <A
HREF="uuid-ossp.html"
>uuid-ossp</A
> to work with
the <SPAN
CLASS="SYSTEMITEM"
>BSD</SPAN
> or <SPAN
CLASS="SYSTEMITEM"
>e2fsprogs</SPAN
> UUID libraries,
not only the <SPAN
CLASS="SYSTEMITEM"
>OSSP</SPAN
> UUID library (Matteo Beccati)
</P
><P
> This improves the <SPAN
CLASS="APPLICATION"
>uuid-ossp</SPAN
> module's portability
since it no longer has to have the increasingly-obsolete OSSP
library. The module's name is now rather a misnomer, but we won't
change it.
</P
></LI
><LI
><P
> Add option to <A
HREF="auto-explain.html"
>auto_explain</A
> to include trigger
execution time (Horiguchi Kyotaro)
</P
></LI
><LI
><P
> Fix <A
HREF="pgstattuple.html"
>pgstattuple</A
> to not report rows from
uncommitted transactions as dead (Robert Haas)
</P
></LI
><LI
><P
> Make <A
HREF="pgstattuple.html"
>pgstattuple</A
> functions
use <TT
CLASS="TYPE"
>regclass</TT
>-type arguments (Satoshi Nagayasu)
</P
><P
> While <TT
CLASS="TYPE"
>text</TT
>-type arguments are still supported, they
may be removed in a future major release.
</P
></LI
><LI
><P
> Improve consistency of <A
HREF="pgrowlocks.html"
>pgrowlocks</A
> output to honor
snapshot rules more consistently (Robert Haas)
</P
></LI
><LI
><P
> Improve <A
HREF="pgtrgm.html"
>pg_trgm</A
>'s choice of trigrams for indexed
regular expression searches (Alexander Korotkov)
</P
><P
> This change discourages use of trigrams containing whitespace, which
are usually less selective.
</P
></LI
><LI
><P
> Allow <A
HREF="pgxlogdump.html"
><SPAN
CLASS="APPLICATION"
>pg_xlogdump</SPAN
></A
> to report a live log stream
with <TT
CLASS="OPTION"
>--follow</TT
> (Heikki Linnakangas)
</P
></LI
><LI
><P
> Store <A
HREF="cube.html"
>cube</A
> data more compactly (Stas Kelvich)
</P
><P
> Existing data must be dumped/restored to use the new format.
The old format can still be read.
</P
></LI
><LI
><P
> Reduce <A
HREF="vacuumlo.html"
><SPAN
CLASS="APPLICATION"
>vacuumlo</SPAN
></A
> client-side memory usage by using
a cursor (Andrew Dunstan)
</P
></LI
><LI
><P
> Dramatically reduce memory consumption
in <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
> (Bruce Momjian)
</P
></LI
><LI
><P
> Pass <A
HREF="pgupgrade.html"
><SPAN
CLASS="APPLICATION"
>pg_upgrade</SPAN
></A
>'s user name (<TT
CLASS="OPTION"
>-U</TT
>) option to
generated analyze scripts (Bruce Momjian)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN139062"
>E.28.3.12.1. <A
HREF="pgbench.html"
><SPAN
CLASS="APPLICATION"
>pgbench</SPAN
></A
></A
></H4
><P
></P
><UL
><LI
><P
> Remove line length limit for <SPAN
CLASS="APPLICATION"
>pgbench</SPAN
> scripts (Sawada
Masahiko)
</P
><P
> The previous line limit was <TT
CLASS="ENVAR"
>BUFSIZ</TT
>.
</P
></LI
><LI
><P
> Add long option names to <SPAN
CLASS="APPLICATION"
>pgbench</SPAN
> (Fabien Coelho)
</P
></LI
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>pgbench</SPAN
> option <TT
CLASS="OPTION"
>--rate</TT
> to control
the transaction rate (Fabien Coelho)
</P
></LI
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>pgbench</SPAN
> option <TT
CLASS="OPTION"
>--progress</TT
> to
print periodic progress reports
(Fabien Coelho)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN139082"
>E.28.3.12.2. <A
HREF="pgstatstatements.html"
>pg_stat_statements</A
></A
></H4
><P
></P
><UL
><LI
><P
> Make <SPAN
CLASS="APPLICATION"
>pg_stat_statements</SPAN
> use a file, rather than
shared memory, for query text storage (Peter Geoghegan)
</P
><P
> This removes the previous limitation on query text length, and
allows a higher number of unique statements to be tracked by default.
</P
></LI
><LI
><P
> Allow reporting of <SPAN
CLASS="APPLICATION"
>pg_stat_statements</SPAN
>'s internal
query hash identifier (Daniel Farina, Sameer Thakur, Peter
Geoghegan)
</P
></LI
><LI
><P
> Add the ability to retrieve all <SPAN
CLASS="APPLICATION"
>pg_stat_statements</SPAN
>
information except the query text (Peter Geoghegan)
</P
><P
> This allows monitoring tools to fetch query text only for
just-created entries, improving performance during repeated querying
of the statistics.
</P
></LI
><LI
><P
> Make <SPAN
CLASS="APPLICATION"
>pg_stat_statements</SPAN
> ignore <TT
CLASS="COMMAND"
>DEALLOCATE</TT
>
commands (Fabien Coelho)
</P
><P
> It already ignored <TT
CLASS="COMMAND"
>PREPARE</TT
>, as well as planning time in
general, so this seems more consistent.
</P
></LI
><LI
><P
> Save the statistics file into <TT
CLASS="FILENAME"
>$PGDATA/pg_stat</TT
> at server
shutdown, rather than <TT
CLASS="FILENAME"
>$PGDATA/global</TT
> (Fujii Masao)
</P
></LI
></UL
></DIV
></DIV
></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-9-4-1.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-9-3-18.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Release 9.4.1</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 9.3.18</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
