<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Release 8.4</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 8.4.12 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Release Notes"
HREF="release.html"><LINK
REL="PREVIOUS"
TITLE="Release 8.4.1"
HREF="release-8-4-1.html"><LINK
REL="NEXT"
TITLE="Release 8.3.19"
HREF="release-8-3-19.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="2012-05-31T23:30:11"></HEAD
><BODY
CLASS="SECT1"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="5"
ALIGN="center"
VALIGN="bottom"
>PostgreSQL 8.4.12 Documentation</TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="release-8-4-1.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
HREF="release.html"
>Fast Backward</A
></TD
><TD
WIDTH="60%"
ALIGN="center"
VALIGN="bottom"
>Appendix E. Release Notes</TD
><TD
WIDTH="10%"
ALIGN="right"
VALIGN="top"
><A
HREF="release.html"
>Fast Forward</A
></TD
><TD
WIDTH="10%"
ALIGN="right"
VALIGN="top"
><A
HREF="release-8-3-19.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RELEASE-8-4"
>E.13. Release 8.4</A
></H1
><DIV
CLASS="NOTE"
><BLOCKQUOTE
CLASS="NOTE"
><P
><B
>Release Date: </B
>2009-07-01</P
></BLOCKQUOTE
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN95837"
>E.13.1. Overview</A
></H2
><P
> After many years of development, <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> has
become feature-complete in many areas. This release shows a
targeted approach to adding features (e.g., authentication,
monitoring, space reuse), and adds capabilities defined in the
later SQL standards. The major areas of enhancement are:
</P
><P
></P
><UL
><LI
><P
> Windowing Functions
</P
></LI
><LI
><P
> Common Table Expressions and Recursive Queries
</P
></LI
><LI
><P
> Default and variadic parameters for functions
</P
></LI
><LI
><P
> Parallel Restore
</P
></LI
><LI
><P
> Column Permissions
</P
></LI
><LI
><P
> Per-database locale settings
</P
></LI
><LI
><P
> Improved hash indexes
</P
></LI
><LI
><P
> Improved join performance for <TT
CLASS="LITERAL"
>EXISTS</TT
> and <TT
CLASS="LITERAL"
>NOT EXISTS</TT
> queries
</P
></LI
><LI
><P
> Easier-to-use Warm Standby
</P
></LI
><LI
><P
> Automatic sizing of the Free Space Map
</P
></LI
><LI
><P
> Visibility Map (greatly reduces vacuum overhead for slowly-changing tables)
</P
></LI
><LI
><P
> Version-aware psql (backslash commands work against older servers)
</P
></LI
><LI
><P
> Support SSL certificates for user authentication
</P
></LI
><LI
><P
> Per-function runtime statistics
</P
></LI
><LI
><P
> Easy editing of functions in psql
</P
></LI
><LI
><P
> New contrib modules: pg_stat_statements, auto_explain, citext, btree_gin
</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="AEN95877"
>E.13.2. Migration to Version 8.4</A
></H2
><P
> A dump/restore using <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> is
required for those wishing to migrate data from any previous
release.
</P
><P
> Observe the following incompatibilities:
</P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN95882"
>E.13.2.1. General</A
></H3
><P
></P
><UL
><LI
><P
> Use 64-bit integer datetimes by default (Neil Conway)
</P
><P
> Previously this was selected by <SPAN
CLASS="APPLICATION"
>configure</SPAN
>'s
<TT
CLASS="OPTION"
>--enable-integer-datetimes</TT
> option. To retain
the old behavior, build with <TT
CLASS="OPTION"
>--disable-integer-datetimes</TT
>.
</P
></LI
><LI
><P
> Remove <SPAN
CLASS="APPLICATION"
>ipcclean</SPAN
> utility command (Bruce)
</P
><P
> The utility only worked on a few platforms. Users should use
their operating system tools instead.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN95895"
>E.13.2.2. Server Settings</A
></H3
><P
></P
><UL
><LI
><P
> Change default setting for
<TT
CLASS="LITERAL"
>log_min_messages</TT
> to <TT
CLASS="LITERAL"
>warning</TT
> (previously
it was <TT
CLASS="LITERAL"
>notice</TT
>) to reduce log file volume (Tom)
</P
></LI
><LI
><P
> Change default setting for <TT
CLASS="LITERAL"
>max_prepared_transactions</TT
> to
zero (previously it was 5) (Tom)
</P
></LI
><LI
><P
> Make <TT
CLASS="LITERAL"
>debug_print_parse</TT
>, <TT
CLASS="LITERAL"
>debug_print_rewritten</TT
>,
and <TT
CLASS="LITERAL"
>debug_print_plan</TT
>
output appear at <TT
CLASS="LITERAL"
>LOG</TT
> message level, not
<TT
CLASS="LITERAL"
>DEBUG1</TT
> as formerly (Tom)
</P
></LI
><LI
><P
> Make <TT
CLASS="LITERAL"
>debug_pretty_print</TT
> default to <TT
CLASS="LITERAL"
>on</TT
> (Tom)
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>explain_pretty_print</TT
> parameter (no longer needed) (Tom)
</P
></LI
><LI
><P
> Make <TT
CLASS="VARNAME"
>log_temp_files</TT
> settable by superusers only, like other
logging options (Simon Riggs)
</P
></LI
><LI
><P
> Remove automatic appending of the epoch timestamp when no <TT
CLASS="LITERAL"
>%</TT
>
escapes are present in <TT
CLASS="LITERAL"
>log_filename</TT
> (Robert Haas)
</P
><P
> This change was made because some users wanted a fixed log filename,
for use with an external log rotation tool.
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>log_restartpoints</TT
> from <TT
CLASS="FILENAME"
>recovery.conf</TT
>;
instead use <TT
CLASS="VARNAME"
>log_checkpoints</TT
> (Simon)
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>krb_realm</TT
> and <TT
CLASS="VARNAME"
>krb_server_hostname</TT
>;
these are now set in <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> instead (Magnus)
</P
></LI
><LI
><P
> There are also significant changes in <A
HREF="release-8-4.html#RELEASE-8-4-PG-HBA-CONF"
><TT
CLASS="FILENAME"
>pg_hba.conf</TT
></A
>,
as described below.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN95942"
>E.13.2.3. Queries</A
></H3
><P
></P
><UL
><LI
><P
> Change <TT
CLASS="COMMAND"
>TRUNCATE</TT
> and <TT
CLASS="COMMAND"
>LOCK</TT
> to
apply to child tables of the specified table(s) (Peter)
</P
><P
> These commands now accept an <TT
CLASS="LITERAL"
>ONLY</TT
> option that prevents
processing child tables; this option must be used if the old
behavior is needed.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>SELECT DISTINCT</TT
> and
<TT
CLASS="LITERAL"
>UNION</TT
>/<TT
CLASS="LITERAL"
>INTERSECT</TT
>/<TT
CLASS="LITERAL"
>EXCEPT</TT
>
no longer always produce sorted output (Tom)
</P
><P
> Previously, these types of queries always removed duplicate rows
by means of Sort/Unique processing (i.e., sort then remove adjacent
duplicates). Now they can be implemented by hashing, which will not
produce sorted output. If an application relied on the output being
in sorted order, the recommended fix is to add an <TT
CLASS="LITERAL"
>ORDER BY</TT
>
clause. As a short-term workaround, the previous behavior can be
restored by disabling <TT
CLASS="LITERAL"
>enable_hashagg</TT
>, but that is a very
performance-expensive fix. <TT
CLASS="LITERAL"
>SELECT DISTINCT ON</TT
> never uses
hashing, however, so its behavior is unchanged.
</P
></LI
><LI
><P
> Force child tables to inherit <TT
CLASS="LITERAL"
>CHECK</TT
> constraints from parents
(Alex Hunsaker, Nikhil Sontakke, Tom)
</P
><P
> Formerly it was possible to drop such a constraint from a child
table, allowing rows that violate the constraint to be visible
when scanning the parent table. This was deemed inconsistent,
as well as contrary to SQL standard.
</P
></LI
><LI
><P
> Disallow negative <TT
CLASS="LITERAL"
>LIMIT</TT
> or <TT
CLASS="LITERAL"
>OFFSET</TT
>
values, rather than treating them as zero (Simon)
</P
></LI
><LI
><P
> Disallow <TT
CLASS="COMMAND"
>LOCK TABLE</TT
> outside a transaction block
(Tom)
</P
><P
> Such an operation is useless because the lock would be released
immediately.
</P
></LI
><LI
><P
> Sequences now contain an additional <TT
CLASS="STRUCTFIELD"
>start_value</TT
> column
(Zoltan Boszormenyi)
</P
><P
> This supports <TT
CLASS="COMMAND"
>ALTER SEQUENCE ... RESTART</TT
>.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN95978"
>E.13.2.4. Functions and Operators</A
></H3
><P
></P
><UL
><LI
><P
> Make <TT
CLASS="TYPE"
>numeric</TT
> zero raised to a fractional power return
<TT
CLASS="LITERAL"
>0</TT
>, rather than throwing an error, and make
<TT
CLASS="TYPE"
>numeric</TT
> zero raised to the zero power return <TT
CLASS="LITERAL"
>1</TT
>,
rather than error (Bruce)
</P
><P
> This matches the longstanding <TT
CLASS="TYPE"
>float8</TT
> behavior.
</P
></LI
><LI
><P
> Allow unary minus of floating-point values to produce minus zero (Tom)
</P
><P
> The changed behavior is more <ACRONYM
CLASS="ACRONYM"
>IEEE</ACRONYM
>-standard
compliant.
</P
></LI
><LI
><P
> Throw an error if an escape character is the last character in
a <TT
CLASS="LITERAL"
>LIKE</TT
> pattern (i.e., it has nothing to escape) (Tom)
</P
><P
> Previously, such an escape character was silently ignored,
thus possibly masking application logic errors.
</P
></LI
><LI
><P
> Remove <TT
CLASS="LITERAL"
>~=~</TT
> and <TT
CLASS="LITERAL"
>~<>~</TT
> operators
formerly used for <TT
CLASS="LITERAL"
>LIKE</TT
> index comparisons (Tom)
</P
><P
> Pattern indexes now use the regular equality operator.
</P
></LI
><LI
><P
> <CODE
CLASS="FUNCTION"
>xpath()</CODE
> now passes its arguments to <SPAN
CLASS="APPLICATION"
>libxml</SPAN
>
without any changes (Andrew)
</P
><P
> This means that the XML argument must be a well-formed XML document.
The previous coding attempted to allow XML fragments, but it did not
work well.
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>xmlelement()</CODE
> format attribute values just like
content values (Peter)
</P
><P
> Previously, attribute values were formatted according to the
normal SQL output behavior, which is sometimes at odds with
XML rules.
</P
></LI
><LI
><P
> Rewrite memory management for <SPAN
CLASS="APPLICATION"
>libxml</SPAN
>-using functions
(Tom)
</P
><P
> This change should avoid some compatibility problems with use of
<SPAN
CLASS="APPLICATION"
>libxml</SPAN
> in PL/Perl and other add-on code.
</P
></LI
><LI
><P
> Adopt a faster algorithm for hash functions (Kenneth Marshall,
based on work of Bob Jenkins)
</P
><P
> Many of the built-in hash functions now deliver different results on
little-endian and big-endian platforms.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96020"
>E.13.2.4.1. Temporal Functions and Operators</A
></H4
><P
></P
><UL
><LI
><P
> <TT
CLASS="VARNAME"
>DateStyle</TT
> no longer controls <TT
CLASS="TYPE"
>interval</TT
> output
formatting; instead there is a new variable <TT
CLASS="VARNAME"
>IntervalStyle</TT
>
(Ron Mayer)
</P
></LI
><LI
><P
> Improve consistency of handling of fractional seconds in
<TT
CLASS="TYPE"
>timestamp</TT
> and <TT
CLASS="TYPE"
>interval</TT
> output (Ron Mayer)
</P
><P
> This may result in displaying a different number of fractional
digits than before, or rounding instead of truncating.
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>to_char()</CODE
>'s localized month/day names depend
on <TT
CLASS="VARNAME"
>LC_TIME</TT
>, not <TT
CLASS="VARNAME"
>LC_MESSAGES</TT
> (Euler
Taveira de Oliveira)
</P
></LI
><LI
><P
> Cause <CODE
CLASS="FUNCTION"
>to_date()</CODE
> and <CODE
CLASS="FUNCTION"
>to_timestamp()</CODE
>
to more consistently report errors for invalid input (Brendan
Jurd)
</P
><P
> Previous versions would often ignore or silently misread input
that did not match the format string. Such cases will now
result in an error.
</P
></LI
><LI
><P
> Fix <CODE
CLASS="FUNCTION"
>to_timestamp()</CODE
> to not require upper/lower case
matching for meridian (<TT
CLASS="LITERAL"
>AM</TT
>/<TT
CLASS="LITERAL"
>PM</TT
>) and era
(<TT
CLASS="LITERAL"
>BC</TT
>/<TT
CLASS="LITERAL"
>AD</TT
>) format designations (Brendan
Jurd)
</P
><P
> For example, input value <TT
CLASS="LITERAL"
>ad</TT
> now matches the format
string <TT
CLASS="LITERAL"
>AD</TT
>.
</P
></LI
></UL
></DIV
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN96053"
>E.13.3. Changes</A
></H2
><P
> Below you will find a detailed account of the changes between
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 8.4 and the previous major
release.
</P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96057"
>E.13.3.1. Performance</A
></H3
><P
></P
><UL
><LI
><P
> Improve optimizer statistics calculations (Jan Urbanski, Tom)
</P
><P
> In particular, estimates for full-text-search operators are
greatly improved.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>SELECT DISTINCT</TT
> and
<TT
CLASS="LITERAL"
>UNION</TT
>/<TT
CLASS="LITERAL"
>INTERSECT</TT
>/<TT
CLASS="LITERAL"
>EXCEPT</TT
> to
use hashing (Tom)
</P
><P
> This means that these types of queries no longer automatically
produce sorted output.
</P
></LI
><LI
><P
> Create explicit concepts of semi-joins and anti-joins (Tom)
</P
><P
> This work formalizes our previous ad-hoc treatment of <TT
CLASS="LITERAL"
>IN
(SELECT ...)</TT
> clauses, and extends it to <TT
CLASS="LITERAL"
>EXISTS</TT
> and
<TT
CLASS="LITERAL"
>NOT EXISTS</TT
> clauses. It should result in significantly
better planning of <TT
CLASS="LITERAL"
>EXISTS</TT
> and <TT
CLASS="LITERAL"
>NOT EXISTS</TT
>
queries. In general, logically equivalent <TT
CLASS="LITERAL"
>IN</TT
> and
<TT
CLASS="LITERAL"
>EXISTS</TT
> clauses should now have similar performance,
whereas previously <TT
CLASS="LITERAL"
>IN</TT
> often won.
</P
></LI
><LI
><P
> Improve optimization of sub-selects beneath outer joins (Tom)
</P
><P
> Formerly, a sub-select or view could not be optimized very well if it
appeared within the nullable side of an outer join and contained
non-strict expressions (for instance, constants) in its result list.
</P
></LI
><LI
><P
> Improve the performance of <CODE
CLASS="FUNCTION"
>text_position()</CODE
> and
related functions by using Boyer-Moore-Horspool searching (David
Rowley)
</P
><P
> This is particularly helpful for long search patterns.
</P
></LI
><LI
><P
> Reduce I/O load of writing the statistics collection file
by writing the file only when requested (Martin Pihlak)
</P
></LI
><LI
><P
> Improve performance for bulk inserts (Robert Haas, Simon)
</P
></LI
><LI
><P
> Increase the default value of <TT
CLASS="VARNAME"
>default_statistics_target</TT
>
from <TT
CLASS="LITERAL"
>10</TT
> to <TT
CLASS="LITERAL"
>100</TT
> (Greg Sabino Mullane,
Tom)
</P
><P
> The maximum value was also increased from <TT
CLASS="LITERAL"
>1000</TT
> to
<TT
CLASS="LITERAL"
>10000</TT
>.
</P
></LI
><LI
><P
> Perform <TT
CLASS="VARNAME"
>constraint_exclusion</TT
> checking by default
in queries involving inheritance or <TT
CLASS="LITERAL"
>UNION ALL</TT
> (Tom)
</P
><P
> A new <TT
CLASS="VARNAME"
>constraint_exclusion</TT
> setting,
<TT
CLASS="LITERAL"
>partition</TT
>, was added to specify this behavior.
</P
></LI
><LI
><P
> Allow I/O read-ahead for bitmap index scans (Greg Stark)
</P
><P
> The amount of read-ahead is controlled by
<TT
CLASS="VARNAME"
>effective_io_concurrency</TT
>. This feature is available only
if the kernel has <CODE
CLASS="FUNCTION"
>posix_fadvise()</CODE
> support.
</P
></LI
><LI
><P
> Inline simple set-returning <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> functions in
<TT
CLASS="LITERAL"
>FROM</TT
> clauses (Richard Rowell)
</P
></LI
><LI
><P
> Improve performance of multi-batch hash joins by providing a special
case for join key values that are especially common in the outer
relation (Bryce Cutt, Ramon Lawrence)
</P
></LI
><LI
><P
> Reduce volume of temporary data in multi-batch hash joins
by suppressing <SPAN
CLASS="QUOTE"
>"physical tlist"</SPAN
> optimization (Michael
Henderson, Ramon Lawrence)
</P
></LI
><LI
><P
> Avoid waiting for idle-in-transaction sessions during
<TT
CLASS="COMMAND"
>CREATE INDEX CONCURRENTLY</TT
> (Simon)
</P
></LI
><LI
><P
> Improve performance of shared cache invalidation (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96126"
>E.13.3.2. Server</A
></H3
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96128"
>E.13.3.2.1. Settings</A
></H4
><P
></P
><UL
><LI
><P
> Convert many <TT
CLASS="FILENAME"
>postgresql.conf</TT
> settings to enumerated
values so that <TT
CLASS="LITERAL"
>pg_settings</TT
> can display the valid
values (Magnus)
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>cursor_tuple_fraction</TT
> parameter to control the
fraction of a cursor's rows that the planner assumes will be
fetched (Robert Hell)
</P
></LI
><LI
><P
> Allow underscores in the names of custom variable
classes in <TT
CLASS="FILENAME"
>postgresql.conf</TT
> (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96141"
>E.13.3.2.2. Authentication and security</A
></H4
><P
></P
><UL
><LI
><P
> Remove support for the (insecure) <TT
CLASS="LITERAL"
>crypt</TT
> authentication method
(Magnus)
</P
><P
> This effectively obsoletes pre-<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 7.2 client
libraries, as there is no longer any non-plaintext password method that
they can use.
</P
></LI
><LI
><P
> Support regular expressions in <TT
CLASS="FILENAME"
>pg_ident.conf</TT
>
(Magnus)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="PRODUCTNAME"
>Kerberos</SPAN
>/<ACRONYM
CLASS="ACRONYM"
>GSSAPI</ACRONYM
> parameters
to be changed without restarting the postmaster (Magnus)
</P
></LI
><LI
><P
> Support <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> certificate chains in server certificate
file (Andrew Gierth)
</P
><P
> Including the full certificate chain makes the client able
to verify the certificate without having all intermediate CA
certificates present in the local store, which is often the case for
commercial CAs.
</P
></LI
><LI
><P
> Report appropriate error message for combination of <TT
CLASS="LITERAL"
>MD5</TT
>
authentication and <TT
CLASS="VARNAME"
>db_user_namespace</TT
> enabled (Bruce)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="RELEASE-8-4-PG-HBA-CONF"
>E.13.3.2.3. <TT
CLASS="FILENAME"
>pg_hba.conf</TT
></A
></H4
><P
></P
><UL
><LI
><P
> Change all authentication options to use <TT
CLASS="LITERAL"
>name=value</TT
>
syntax (Magnus)
</P
><P
> This makes incompatible changes to the <TT
CLASS="LITERAL"
>ldap</TT
>,
<TT
CLASS="LITERAL"
>pam</TT
> and <TT
CLASS="LITERAL"
>ident</TT
> authentication methods. All
<TT
CLASS="FILENAME"
>pg_hba.conf</TT
> entries with these methods need to be
rewritten using the new format.
</P
></LI
><LI
><P
> Remove the <TT
CLASS="LITERAL"
>ident sameuser</TT
> option, instead making that
behavior the default if no usermap is specified (Magnus)
</P
></LI
><LI
><P
> Allow a usermap parameter for all external authentication methods
(Magnus)
</P
><P
> Previously a usermap was only supported for <TT
CLASS="LITERAL"
>ident</TT
>
authentication.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>clientcert</TT
> option to control requesting of a
client certificate (Magnus)
</P
><P
> Previously this was controlled by the presence of a root
certificate file in the server's data directory.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>cert</TT
> authentication method to allow
<SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>user</I
></SPAN
> authentication via <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> certificates
(Magnus)
</P
><P
> Previously <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> certificates could only verify that
the client had access to a certificate, not authenticate a
user.
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>krb5</TT
>, <TT
CLASS="LITERAL"
>gssapi</TT
> and <TT
CLASS="LITERAL"
>sspi</TT
>
realm and <TT
CLASS="LITERAL"
>krb5</TT
> host settings to be specified in
<TT
CLASS="FILENAME"
>pg_hba.conf</TT
> (Magnus)
</P
><P
> These override the settings in <TT
CLASS="FILENAME"
>postgresql.conf</TT
>.
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>include_realm</TT
> parameter for <TT
CLASS="LITERAL"
>krb5</TT
>,
<TT
CLASS="LITERAL"
>gssapi</TT
>, and <TT
CLASS="LITERAL"
>sspi</TT
> methods (Magnus)
</P
><P
> This allows identical usernames from different realms to be
authenticated as different database users using usermaps.
</P
></LI
><LI
><P
> Parse <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> fully when it is loaded,
so that errors are reported immediately (Magnus)
</P
><P
> Previously, most errors in the file wouldn't be detected until clients
tried to connect, so an erroneous file could render the system
unusable. With the new behavior, if an error is detected during
reload then the bad file is rejected and the postmaster continues
to use its old copy.
</P
></LI
><LI
><P
> Show all parsing errors in <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> instead of
aborting after the first one (Selena Deckelmann)
</P
></LI
><LI
><P
> Support <TT
CLASS="LITERAL"
>ident</TT
> authentication over Unix-domain sockets
on <SPAN
CLASS="PRODUCTNAME"
>Solaris</SPAN
> (Garick Hamlin)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96221"
>E.13.3.2.4. Continuous Archiving</A
></H4
><P
></P
><UL
><LI
><P
> Provide an option to <CODE
CLASS="FUNCTION"
>pg_start_backup()</CODE
> to force its
implied checkpoint to finish as quickly as possible (Tom)
</P
><P
> The default behavior avoids excess I/O consumption, but that is
pointless if no concurrent query activity is going on.
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>pg_stop_backup()</CODE
> wait for modified <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
>
files to be archived (Simon)
</P
><P
> This guarantees that the backup is valid at the time
<CODE
CLASS="FUNCTION"
>pg_stop_backup()</CODE
> completes.
</P
></LI
><LI
><P
> When archiving is enabled, rotate the last WAL segment at shutdown
so that all transactions can be archived immediately
(Guillaume Smet, Heikki)
</P
></LI
><LI
><P
> Delay <SPAN
CLASS="QUOTE"
>"smart"</SPAN
> shutdown while a continuous archiving base backup
is in progress (Laurenz Albe)
</P
></LI
><LI
><P
> Cancel a continuous archiving base backup if <SPAN
CLASS="QUOTE"
>"fast"</SPAN
> shutdown
is requested (Laurenz Albe)
</P
></LI
><LI
><P
> Allow <TT
CLASS="FILENAME"
>recovery.conf</TT
> boolean variables to take the
same range of string values as <TT
CLASS="FILENAME"
>postgresql.conf</TT
>
boolean variables
(Bruce)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96246"
>E.13.3.2.5. Monitoring</A
></H4
><P
></P
><UL
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_conf_load_time()</CODE
> to report when
the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> configuration files were last loaded
(George Gensure)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_terminate_backend()</CODE
> to safely terminate a
backend (the <TT
CLASS="LITERAL"
>SIGTERM</TT
> signal works also) (Tom, Bruce)
</P
><P
> While it's always been possible to <TT
CLASS="LITERAL"
>SIGTERM</TT
> a single
backend, this was previously considered unsupported; and testing
of the case found some bugs that are now fixed.
</P
></LI
><LI
><P
> Add ability to track user-defined functions' call counts and
runtimes (Martin Pihlak)
</P
><P
> Function statistics appear in a new system view,
<TT
CLASS="LITERAL"
>pg_stat_user_functions</TT
>. Tracking is controlled
by the new parameter <TT
CLASS="VARNAME"
>track_functions</TT
>.
</P
></LI
><LI
><P
> Allow specification of the maximum query string size in
<TT
CLASS="LITERAL"
>pg_stat_activity</TT
> via new
<TT
CLASS="VARNAME"
>track_activity_query_size</TT
> parameter (Thomas Lee)
</P
></LI
><LI
><P
> Increase the maximum line length sent to <SPAN
CLASS="APPLICATION"
>syslog</SPAN
>, in
hopes of improving performance (Tom)
</P
></LI
><LI
><P
> Add read-only configuration variables <TT
CLASS="VARNAME"
>segment_size</TT
>,
<TT
CLASS="VARNAME"
>wal_block_size</TT
>, and <TT
CLASS="VARNAME"
>wal_segment_size</TT
>
(Bernd Helmle)
</P
></LI
><LI
><P
> When reporting a deadlock, report the text of all queries involved
in the deadlock to the server log (Itagaki Takahiro)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_stat_get_activity(pid)</CODE
> function to return
information about a specific process id (Magnus)
</P
></LI
><LI
><P
> Allow the location of the server's statistics file to be specified
via <TT
CLASS="VARNAME"
>stats_temp_directory</TT
> (Magnus)
</P
><P
> This allows the statistics file to be placed in a
<ACRONYM
CLASS="ACRONYM"
>RAM</ACRONYM
>-resident directory to reduce I/O requirements.
On startup/shutdown, the file is copied to its traditional location
(<TT
CLASS="LITERAL"
>$PGDATA/global/</TT
>) so it is preserved across restarts.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96287"
>E.13.3.3. Queries</A
></H3
><P
></P
><UL
><LI
><P
> Add support for <TT
CLASS="LITERAL"
>WINDOW</TT
> functions (Hitoshi Harada)
</P
></LI
><LI
><P
> Add support for <TT
CLASS="LITERAL"
>WITH</TT
> clauses (CTEs), including <TT
CLASS="LITERAL"
>WITH
RECURSIVE</TT
> (Yoshiyuki Asaba, Tatsuo Ishii, Tom)
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>TABLE</TT
> command (Peter)
</P
><P
> <TT
CLASS="LITERAL"
>TABLE tablename</TT
> is a SQL standard short-hand for
<TT
CLASS="LITERAL"
>SELECT * FROM tablename</TT
>.
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>AS</TT
> to be optional when specifying a
<TT
CLASS="COMMAND"
>SELECT</TT
> (or <TT
CLASS="LITERAL"
>RETURNING</TT
>) column output
label (Hiroshi Saito)
</P
><P
> This works so long as the column label is not any
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> keyword; otherwise <TT
CLASS="LITERAL"
>AS</TT
> is still
needed.
</P
></LI
><LI
><P
> Support set-returning functions in <TT
CLASS="COMMAND"
>SELECT</TT
> result lists
even for functions that return their result via a tuplestore (Tom)
</P
><P
> In particular, this means that functions written in PL/pgSQL
and other PL languages can now be called this way.
</P
></LI
><LI
><P
> Support set-returning functions in the output of aggregation
and grouping queries (Tom)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>SELECT FOR UPDATE</TT
>/<TT
CLASS="LITERAL"
>SHARE</TT
> to work
on inheritance trees (Tom)
</P
></LI
><LI
><P
> Add infrastructure for <ACRONYM
CLASS="ACRONYM"
>SQL/MED</ACRONYM
> (Martin Pihlak,
Peter)
</P
><P
> There are no remote or external <ACRONYM
CLASS="ACRONYM"
>SQL/MED</ACRONYM
> capabilities
yet, but this change provides a standardized and future-proof
system for managing connection information for modules like
<TT
CLASS="FILENAME"
>dblink</TT
> and <TT
CLASS="FILENAME"
>plproxy</TT
>.
</P
></LI
><LI
><P
> Invalidate cached plans when referenced schemas, functions, operators,
or operator classes are modified (Martin Pihlak, Tom)
</P
><P
> This improves the system's ability to respond to on-the-fly
DDL changes.
</P
></LI
><LI
><P
> Allow comparison of composite types and allow arrays of
anonymous composite types (Tom)
</P
><P
> This allows constructs such as
<TT
CLASS="LITERAL"
>row(1, 1.1) = any (array[row(7, 7.7), row(1, 1.0)])</TT
>.
This is particularly useful in recursive queries.
</P
></LI
><LI
><P
> Add support for Unicode string literal and identifier specifications
using code points, e.g. <TT
CLASS="LITERAL"
>U&'d\0061t\+000061'</TT
>
(Peter)
</P
></LI
><LI
><P
> Reject <TT
CLASS="LITERAL"
>\000</TT
> in string literals and <TT
CLASS="COMMAND"
>COPY</TT
> data
(Tom)
</P
><P
> Previously, this was accepted but had the effect of terminating
the string contents.
</P
></LI
><LI
><P
> Improve the parser's ability to report error locations (Tom)
</P
><P
> An error location is now reported for many semantic errors,
such as mismatched datatypes, that previously could not be localized.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96346"
>E.13.3.3.1. <TT
CLASS="COMMAND"
>TRUNCATE</TT
></A
></H4
><P
></P
><UL
><LI
><P
> Support statement-level <TT
CLASS="LITERAL"
>ON TRUNCATE</TT
> triggers (Simon)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>RESTART</TT
>/<TT
CLASS="LITERAL"
>CONTINUE IDENTITY</TT
> options
for <TT
CLASS="COMMAND"
>TRUNCATE TABLE</TT
>
(Zoltan Boszormenyi)
</P
><P
> The start value of a sequence can be changed by <TT
CLASS="COMMAND"
>ALTER
SEQUENCE START WITH</TT
>.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>TRUNCATE tab1, tab1</TT
> to succeed (Bruce)
</P
></LI
><LI
><P
> Add a separate <TT
CLASS="COMMAND"
>TRUNCATE</TT
> permission (Robert Haas)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96366"
>E.13.3.3.2. <TT
CLASS="COMMAND"
>EXPLAIN</TT
></A
></H4
><P
></P
><UL
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXPLAIN VERBOSE</TT
> show the output columns of each
plan node (Tom)
</P
><P
> Previously <TT
CLASS="COMMAND"
>EXPLAIN VERBOSE</TT
> output an internal
representation of the query plan. (That behavior is now
available via <TT
CLASS="VARNAME"
>debug_print_plan</TT
>.)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXPLAIN</TT
> identify subplans and initplans with
individual labels (Tom)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXPLAIN</TT
> honor <TT
CLASS="VARNAME"
>debug_print_plan</TT
> (Tom)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>EXPLAIN</TT
> on <TT
CLASS="COMMAND"
>CREATE TABLE AS</TT
> (Peter)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96387"
>E.13.3.3.3. <TT
CLASS="LITERAL"
>LIMIT</TT
>/<TT
CLASS="LITERAL"
>OFFSET</TT
></A
></H4
><P
></P
><UL
><LI
><P
> Allow sub-selects in <TT
CLASS="LITERAL"
>LIMIT</TT
> and <TT
CLASS="LITERAL"
>OFFSET</TT
> (Tom)
</P
></LI
><LI
><P
> Add <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>-standard syntax for
<TT
CLASS="LITERAL"
>LIMIT</TT
>/<TT
CLASS="LITERAL"
>OFFSET</TT
> capabilities (Peter)
</P
><P
> To wit,
<TT
CLASS="LITERAL"
>OFFSET num {ROW|ROWS} FETCH {FIRST|NEXT} [num] {ROW|ROWS}
ONLY</TT
>.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96403"
>E.13.3.4. Object Manipulation</A
></H3
><P
></P
><UL
><LI
><P
> Add support for column-level privileges (Stephen Frost, KaiGai
Kohei)
</P
></LI
><LI
><P
> Refactor multi-object <TT
CLASS="COMMAND"
>DROP</TT
> operations to reduce the
need for <TT
CLASS="LITERAL"
>CASCADE</TT
> (Alex Hunsaker)
</P
><P
> For example, if table <TT
CLASS="LITERAL"
>B</TT
> has a dependency on table
<TT
CLASS="LITERAL"
>A</TT
>, the command <TT
CLASS="LITERAL"
>DROP TABLE A, B</TT
> no longer
requires the <TT
CLASS="LITERAL"
>CASCADE</TT
> option.
</P
></LI
><LI
><P
> Fix various problems with concurrent <TT
CLASS="COMMAND"
>DROP</TT
> commands
by ensuring that locks are taken before we begin to drop dependencies
of an object (Tom)
</P
></LI
><LI
><P
> Improve reporting of dependencies during <TT
CLASS="COMMAND"
>DROP</TT
>
commands (Tom)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>WITH [NO] DATA</TT
> clause to <TT
CLASS="COMMAND"
>CREATE TABLE
AS</TT
>, per the <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> standard (Peter, Tom)
</P
></LI
><LI
><P
> Add support for user-defined I/O conversion casts (Heikki)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>CREATE AGGREGATE</TT
> to use an <TT
CLASS="TYPE"
>internal</TT
>
transition datatype (Tom)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>LIKE</TT
> clause to <TT
CLASS="COMMAND"
>CREATE TYPE</TT
> (Tom)
</P
><P
> This simplifies creation of data types that use the same internal
representation as an existing type.
</P
></LI
><LI
><P
> Allow specification of the type category and <SPAN
CLASS="QUOTE"
>"preferred"</SPAN
>
status for user-defined base types (Tom)
</P
><P
> This allows more control over the coercion behavior of user-defined
types.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>CREATE OR REPLACE VIEW</TT
> to add columns to the
end of a view (Robert Haas)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96446"
>E.13.3.4.1. <TT
CLASS="COMMAND"
>ALTER</TT
></A
></H4
><P
></P
><UL
><LI
><P
> Add <TT
CLASS="COMMAND"
>ALTER TYPE RENAME</TT
> (Petr Jelinek)
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>ALTER SEQUENCE ... RESTART</TT
> (with no parameter) to
reset a sequence to its initial value (Zoltan Boszormenyi)
</P
></LI
><LI
><P
> Modify the <TT
CLASS="COMMAND"
>ALTER TABLE</TT
> syntax to allow all reasonable
combinations for tables, indexes, sequences, and views (Tom)
</P
><P
> This change allows the following new syntaxes:
<P
></P
></P><UL
><LI
><P
> <TT
CLASS="COMMAND"
>ALTER SEQUENCE OWNER TO</TT
>
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>ALTER VIEW ALTER COLUMN SET/DROP DEFAULT</TT
>
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>ALTER VIEW OWNER TO</TT
>
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>ALTER VIEW SET SCHEMA</TT
>
</P
></LI
></UL
><P>
There is no actual new functionality here, but formerly
you had to say <TT
CLASS="COMMAND"
>ALTER TABLE</TT
> to do these things,
which was confusing.
</P
></LI
><LI
><P
> Add support for the syntax <TT
CLASS="COMMAND"
>ALTER TABLE ... ALTER COLUMN
... SET DATA TYPE</TT
> (Peter)
</P
><P
> This is <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>-standard syntax for functionality that
was already supported.
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>ALTER TABLE SET WITHOUT OIDS</TT
> rewrite the table
to physically remove <TT
CLASS="TYPE"
>OID</TT
> values (Tom)
</P
><P
> Also, add <TT
CLASS="COMMAND"
>ALTER TABLE SET WITH OIDS</TT
> to rewrite the
table to add <TT
CLASS="TYPE"
>OID</TT
>s.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96486"
>E.13.3.4.2. Database Manipulation</A
></H4
><P
></P
><UL
><LI
><P
> Improve reporting of
<TT
CLASS="COMMAND"
>CREATE</TT
>/<TT
CLASS="COMMAND"
>DROP</TT
>/<TT
CLASS="COMMAND"
>RENAME DATABASE</TT
>
failure when uncommitted prepared transactions are the cause
(Tom)
</P
></LI
><LI
><P
> Make <TT
CLASS="VARNAME"
>LC_COLLATE</TT
> and <TT
CLASS="VARNAME"
>LC_CTYPE</TT
> into
per-database settings (Radek Strnad, Heikki)
</P
><P
> This makes collation similar to encoding, which was always
configurable per database.
</P
></LI
><LI
><P
> Improve checks that the database encoding, collation
(<TT
CLASS="VARNAME"
>LC_COLLATE</TT
>), and character classes
(<TT
CLASS="VARNAME"
>LC_CTYPE</TT
>) match (Heikki, Tom)
</P
><P
> Note in particular that a new database's encoding and locale
settings can be changed only when copying from <TT
CLASS="LITERAL"
>template0</TT
>.
This prevents possibly copying data that doesn't match the settings.
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>ALTER DATABASE SET TABLESPACE</TT
> to move a database
to a new tablespace (Guillaume Lelarge, Bernd Helmle)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96508"
>E.13.3.5. Utility Operations</A
></H3
><P
></P
><UL
><LI
><P
> Add a <TT
CLASS="LITERAL"
>VERBOSE</TT
> option to the <TT
CLASS="COMMAND"
>CLUSTER</TT
> command and
<SPAN
CLASS="APPLICATION"
>clusterdb</SPAN
> (Jim Cox)
</P
></LI
><LI
><P
> Decrease memory requirements for recording pending trigger
events (Tom)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96518"
>E.13.3.5.1. Indexes</A
></H4
><P
></P
><UL
><LI
><P
> Dramatically improve the speed of building and accessing hash
indexes (Tom Raney, Shreya Bhargava)
</P
><P
> This allows hash indexes to be sometimes faster than btree
indexes. However, hash indexes are still not crash-safe.
</P
></LI
><LI
><P
> Make hash indexes store only the hash code, not the full value of
the indexed column (Xiao Meng)
</P
><P
> This greatly reduces the size of hash indexes for long indexed
values, improving performance.
</P
></LI
><LI
><P
> Implement fast update option for GIN indexes (Teodor, Oleg)
</P
><P
> This option greatly improves update speed at a small penalty in search
speed.
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>xxx_pattern_ops</TT
> indexes can now be used for simple
equality comparisons, not only for <TT
CLASS="LITERAL"
>LIKE</TT
> (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96534"
>E.13.3.5.2. Full Text Indexes</A
></H4
><P
></P
><UL
><LI
><P
> Remove the requirement to use <TT
CLASS="LITERAL"
>@@@</TT
> when doing
<ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
> weighted lookups on full text indexes (Tom, Teodor)
</P
><P
> The normal <TT
CLASS="LITERAL"
>@@</TT
> text search operator can be used
instead.
</P
></LI
><LI
><P
> Add an optimizer selectivity function for <TT
CLASS="LITERAL"
>@@</TT
> text
search operations (Jan Urbanski)
</P
></LI
><LI
><P
> Allow prefix matching in full text searches (Teodor Sigaev,
Oleg Bartunov)
</P
></LI
><LI
><P
> Support multi-column <ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
> indexes (Teodor Sigaev)
</P
></LI
><LI
><P
> Improve support for Nepali language and Devanagari alphabet (Teodor)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96553"
>E.13.3.5.3. <TT
CLASS="COMMAND"
>VACUUM</TT
></A
></H4
><P
></P
><UL
><LI
><P
> Track free space in separate per-relation <SPAN
CLASS="QUOTE"
>"fork"</SPAN
> files (Heikki)
</P
><P
> Free space discovered by <TT
CLASS="COMMAND"
>VACUUM</TT
> is now recorded in
<TT
CLASS="FILENAME"
>*_fsm</TT
> files, rather than in a fixed-sized shared memory
area. The <TT
CLASS="VARNAME"
>max_fsm_pages</TT
> and <TT
CLASS="VARNAME"
>max_fsm_relations</TT
>
settings have been removed, greatly simplifying administration of
free space management.
</P
></LI
><LI
><P
> Add a visibility map to track pages that do not require
vacuuming (Heikki)
</P
><P
> This allows <TT
CLASS="COMMAND"
>VACUUM</TT
> to avoid scanning all of
a table when only a portion of the table needs vacuuming.
The visibility map is stored in per-relation <SPAN
CLASS="QUOTE"
>"fork"</SPAN
> files.
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>vacuum_freeze_table_age</TT
> parameter to control
when <TT
CLASS="COMMAND"
>VACUUM</TT
> should ignore the visibility map and
do a full table scan to freeze tuples (Heikki)
</P
></LI
><LI
><P
> Track transaction snapshots more carefully (Alvaro)
</P
><P
> This improves <TT
CLASS="COMMAND"
>VACUUM</TT
>'s ability to reclaim space
in the presence of long-running transactions.
</P
></LI
><LI
><P
> Add ability to specify per-relation autovacuum and <ACRONYM
CLASS="ACRONYM"
>TOAST</ACRONYM
>
parameters in <TT
CLASS="COMMAND"
>CREATE TABLE</TT
> (Alvaro, Euler Taveira de
Oliveira)
</P
><P
> Autovacuum options used to be stored in a system table.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>--freeze</TT
> option to <SPAN
CLASS="APPLICATION"
>vacuumdb</SPAN
>
(Bruce)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96587"
>E.13.3.6. Data Types</A
></H3
><P
></P
><UL
><LI
><P
> Add a <TT
CLASS="LITERAL"
>CaseSensitive</TT
> option for text search synonym
dictionaries (Simon)
</P
></LI
><LI
><P
> Improve the precision of <TT
CLASS="TYPE"
>NUMERIC</TT
> division (Tom)
</P
></LI
><LI
><P
> Add basic arithmetic operators for <TT
CLASS="TYPE"
>int2</TT
> with <TT
CLASS="TYPE"
>int8</TT
>
(Tom)
</P
><P
> This eliminates the need for explicit casting in some situations.
</P
></LI
><LI
><P
> Allow <TT
CLASS="TYPE"
>UUID</TT
> input to accept an optional hyphen after
every fourth digit (Robert Haas)
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>on</TT
>/<TT
CLASS="LITERAL"
>off</TT
> as input for the boolean data type
(Itagaki Takahiro)
</P
></LI
><LI
><P
> Allow spaces around <TT
CLASS="LITERAL"
>NaN</TT
> in the input string for
type <TT
CLASS="TYPE"
>numeric</TT
> (Sam Mason)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96612"
>E.13.3.6.1. Temporal Data Types</A
></H4
><P
></P
><UL
><LI
><P
> Reject year <TT
CLASS="LITERAL"
>0 BC</TT
> and years <TT
CLASS="LITERAL"
>000</TT
> and
<TT
CLASS="LITERAL"
>0000</TT
> (Tom)
</P
><P
> Previously these were interpreted as <TT
CLASS="LITERAL"
>1 BC</TT
>.
(Note: years <TT
CLASS="LITERAL"
>0</TT
> and <TT
CLASS="LITERAL"
>00</TT
> are still assumed to be
the year 2000.)
</P
></LI
><LI
><P
> Include <TT
CLASS="LITERAL"
>SGT</TT
> (Singapore time) in the default list of
known time zone abbreviations (Tom)
</P
></LI
><LI
><P
> Support <TT
CLASS="LITERAL"
>infinity</TT
> and <TT
CLASS="LITERAL"
>-infinity</TT
> as
values of type <TT
CLASS="TYPE"
>date</TT
> (Tom)
</P
></LI
><LI
><P
> Make parsing of <TT
CLASS="TYPE"
>interval</TT
> literals more standard-compliant
(Tom, Ron Mayer)
</P
><P
> For example, <TT
CLASS="LITERAL"
>INTERVAL '1' YEAR</TT
> now does what it's
supposed to.
</P
></LI
><LI
><P
> Allow <TT
CLASS="TYPE"
>interval</TT
> fractional-seconds precision to be specified
after the <TT
CLASS="LITERAL"
>second</TT
> keyword, for <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> standard
compliance (Tom)
</P
><P
> Formerly the precision had to be specified after the keyword
<TT
CLASS="TYPE"
>interval</TT
>. (For backwards compatibility, this syntax is still
supported, though deprecated.) Data type definitions will now be
output using the standard format.
</P
></LI
><LI
><P
> Support the <ACRONYM
CLASS="ACRONYM"
>IS0 8601</ACRONYM
> <TT
CLASS="TYPE"
>interval</TT
> syntax (Ron
Mayer, Kevin Grittner)
</P
><P
> For example, <TT
CLASS="LITERAL"
>INTERVAL 'P1Y2M3DT4H5M6.7S'</TT
> is now
supported.
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>IntervalStyle</TT
> parameter
which controls how <TT
CLASS="TYPE"
>interval</TT
> values are output (Ron Mayer)
</P
><P
> Valid values are: <TT
CLASS="LITERAL"
>postgres</TT
>, <TT
CLASS="LITERAL"
>postgres_verbose</TT
>,
<TT
CLASS="LITERAL"
>sql_standard</TT
>, <TT
CLASS="LITERAL"
>iso_8601</TT
>. This setting also
controls the handling of negative <TT
CLASS="TYPE"
>interval</TT
> input when only
some fields have positive/negative designations.
</P
></LI
><LI
><P
> Improve consistency of handling of fractional seconds in
<TT
CLASS="TYPE"
>timestamp</TT
> and <TT
CLASS="TYPE"
>interval</TT
> output (Ron Mayer)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96664"
>E.13.3.6.2. Arrays</A
></H4
><P
></P
><UL
><LI
><P
> Improve the handling of casts applied to <TT
CLASS="LITERAL"
>ARRAY[]</TT
>
constructs, such as <TT
CLASS="LITERAL"
>ARRAY[...]::integer[]</TT
>
(Brendan Jurd)
</P
><P
> Formerly <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> attempted to determine a data type
for the <TT
CLASS="LITERAL"
>ARRAY[]</TT
> construct without reference to the ensuing
cast. This could fail unnecessarily in many cases, in particular when
the <TT
CLASS="LITERAL"
>ARRAY[]</TT
> construct was empty or contained only
ambiguous entries such as <TT
CLASS="LITERAL"
>NULL</TT
>. Now the cast is consulted
to determine the type that the array elements must be.
</P
></LI
><LI
><P
> Make <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>-syntax <TT
CLASS="TYPE"
>ARRAY</TT
> dimensions optional
to match the <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
> standard (Peter)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>array_ndims()</CODE
> to return the number
of dimensions of an array (Robert Haas)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>array_length()</CODE
> to return the length
of an array for a specified dimension (Jim Nasby, Robert
Haas, Peter Eisentraut)
</P
></LI
><LI
><P
> Add aggregate function <CODE
CLASS="FUNCTION"
>array_agg()</CODE
>, which
returns all aggregated values as a single array (Robert Haas,
Jeff Davis, Peter)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>unnest()</CODE
>, which converts an array to
individual row values (Tom)
</P
><P
> This is the opposite of <CODE
CLASS="FUNCTION"
>array_agg()</CODE
>.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>array_fill()</CODE
> to create arrays initialized with
a value (Pavel Stehule)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>generate_subscripts()</CODE
> to simplify generating
the range of an array's subscripts (Pavel Stehule)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96701"
>E.13.3.6.3. Wide-Value Storage (<ACRONYM
CLASS="ACRONYM"
>TOAST</ACRONYM
>)</A
></H4
><P
></P
><UL
><LI
><P
> Consider <ACRONYM
CLASS="ACRONYM"
>TOAST</ACRONYM
> compression on values as short as
32 bytes (previously 256 bytes) (Greg Stark)
</P
></LI
><LI
><P
> Require 25% minimum space savings before using <ACRONYM
CLASS="ACRONYM"
>TOAST</ACRONYM
>
compression (previously 20% for small values and any-savings-at-all
for large values) (Greg)
</P
></LI
><LI
><P
> Improve <ACRONYM
CLASS="ACRONYM"
>TOAST</ACRONYM
> heuristics for rows that have a mix of large
and small toastable fields, so that we prefer to push large values out
of line and don't compress small values unnecessarily (Greg, Tom)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96714"
>E.13.3.7. Functions</A
></H3
><P
></P
><UL
><LI
><P
> Document that <CODE
CLASS="FUNCTION"
>setseed()</CODE
> allows values from
<TT
CLASS="LITERAL"
>-1</TT
> to <TT
CLASS="LITERAL"
>1</TT
> (not just <TT
CLASS="LITERAL"
>0</TT
> to
<TT
CLASS="LITERAL"
>1</TT
>), and enforce the valid range (Kris Jurka)
</P
></LI
><LI
><P
> Add server-side function <CODE
CLASS="FUNCTION"
>lo_import(filename, oid)</CODE
>
(Tatsuo)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>quote_nullable()</CODE
>, which behaves like
<CODE
CLASS="FUNCTION"
>quote_literal()</CODE
> but returns the string <TT
CLASS="LITERAL"
>NULL</TT
> for
a null argument (Brendan Jurd)
</P
></LI
><LI
><P
> Improve full text search <CODE
CLASS="FUNCTION"
>headline()</CODE
> function to
allow extracting several fragments of text (Sushant Sinha)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>suppress_redundant_updates_trigger()</CODE
> trigger
function to avoid overhead for non-data-changing updates (Andrew)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>div(numeric, numeric)</CODE
> to perform <TT
CLASS="TYPE"
>numeric</TT
>
division without rounding (Tom)
</P
></LI
><LI
><P
> Add <TT
CLASS="TYPE"
>timestamp</TT
> and <TT
CLASS="TYPE"
>timestamptz</TT
> versions of
<CODE
CLASS="FUNCTION"
>generate_series()</CODE
> (Hitoshi Harada)
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96747"
>E.13.3.7.1. Object Information Functions</A
></H4
><P
></P
><UL
><LI
><P
> Implement <CODE
CLASS="FUNCTION"
>current_query()</CODE
> for use by functions
that need to know the currently running query (Tomas Doran)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_get_keywords()</CODE
> to return a list of the
parser keywords (Dave Page)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_get_functiondef()</CODE
> to see a function's
definition (Abhijit Menon-Sen)
</P
></LI
><LI
><P
> Allow the second argument of <CODE
CLASS="FUNCTION"
>pg_get_expr()</CODE
> to be zero
when deparsing an expression that does not contain variables (Tom)
</P
></LI
><LI
><P
> Modify <CODE
CLASS="FUNCTION"
>pg_relation_size()</CODE
> to use <TT
CLASS="LITERAL"
>regclass</TT
>
(Heikki)
</P
><P
> <CODE
CLASS="FUNCTION"
>pg_relation_size(data_type_name)</CODE
> no longer works.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>boot_val</TT
> and <TT
CLASS="LITERAL"
>reset_val</TT
> columns to
<TT
CLASS="LITERAL"
>pg_settings</TT
> output (Greg Smith)
</P
></LI
><LI
><P
> Add source file name and line number columns to
<TT
CLASS="LITERAL"
>pg_settings</TT
> output for variables set in a configuration
file (Magnus, Alvaro)
</P
><P
> For security reasons, these columns are only visible to superusers.
</P
></LI
><LI
><P
> Add support for <TT
CLASS="VARNAME"
>CURRENT_CATALOG</TT
>,
<TT
CLASS="VARNAME"
>CURRENT_SCHEMA</TT
>, <TT
CLASS="VARNAME"
>SET CATALOG</TT
>, <TT
CLASS="VARNAME"
>SET
SCHEMA</TT
> (Peter)
</P
><P
> These provide <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>-standard syntax for existing features.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_typeof()</CODE
> which returns the data type
of any value (Brendan Jurd)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>version()</CODE
> return information about whether
the server is a 32- or 64-bit binary (Bruce)
</P
></LI
><LI
><P
> Fix the behavior of information schema columns
<TT
CLASS="STRUCTFIELD"
>is_insertable_into</TT
> and <TT
CLASS="STRUCTFIELD"
>is_updatable</TT
> to
be consistent (Peter)
</P
></LI
><LI
><P
> Improve the behavior of information schema
<TT
CLASS="STRUCTFIELD"
>datetime_precision</TT
> columns (Peter)
</P
><P
> These columns now show zero for <TT
CLASS="TYPE"
>date</TT
> columns, and 6
(the default precision) for <TT
CLASS="TYPE"
>time</TT
>, <TT
CLASS="TYPE"
>timestamp</TT
>, and
<TT
CLASS="TYPE"
>interval</TT
> without a declared precision, rather than showing
null as formerly.
</P
></LI
><LI
><P
> Convert remaining builtin set-returning functions to use
<TT
CLASS="LITERAL"
>OUT</TT
> parameters (Jaime Casanova)
</P
><P
> This makes it possible to call these functions without specifying
a column list: <CODE
CLASS="FUNCTION"
>pg_show_all_settings()</CODE
>,
<CODE
CLASS="FUNCTION"
>pg_lock_status()</CODE
>, <CODE
CLASS="FUNCTION"
>pg_prepared_xact()</CODE
>,
<CODE
CLASS="FUNCTION"
>pg_prepared_statement()</CODE
>, <CODE
CLASS="FUNCTION"
>pg_cursor()</CODE
>
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>pg_*_is_visible()</CODE
> and
<CODE
CLASS="FUNCTION"
>has_*_privilege()</CODE
> functions return <TT
CLASS="LITERAL"
>NULL</TT
>
for invalid OIDs, rather than reporting an error (Tom)
</P
></LI
><LI
><P
> Extend <CODE
CLASS="FUNCTION"
>has_*_privilege()</CODE
> functions to allow inquiring
about the OR of multiple privileges in one call (Stephen
Frost, Tom)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>has_column_privilege()</CODE
> and
<CODE
CLASS="FUNCTION"
>has_any_column_privilege()</CODE
> functions (Stephen
Frost, Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96824"
>E.13.3.7.2. Function Creation</A
></H4
><P
></P
><UL
><LI
><P
> Support variadic functions (functions with a variable number
of arguments) (Pavel Stehule)
</P
><P
> Only trailing arguments can be optional, and they all must be
of the same data type.
</P
></LI
><LI
><P
> Support default values for function arguments (Pavel Stehule)
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>CREATE FUNCTION ... RETURNS TABLE</TT
> clause (Pavel
Stehule)
</P
></LI
><LI
><P
> Allow <ACRONYM
CLASS="ACRONYM"
>SQL</ACRONYM
>-language functions to return the output
of an <TT
CLASS="COMMAND"
>INSERT</TT
>/<TT
CLASS="COMMAND"
>UPDATE</TT
>/<TT
CLASS="COMMAND"
>DELETE</TT
>
<TT
CLASS="LITERAL"
>RETURNING</TT
> clause (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96842"
>E.13.3.7.3. PL/pgSQL Server-Side Language</A
></H4
><P
></P
><UL
><LI
><P
> Support <TT
CLASS="LITERAL"
>EXECUTE USING</TT
> for easier insertion of data
values into a dynamic query string (Pavel Stehule)
</P
></LI
><LI
><P
> Allow looping over the results of a cursor using a <TT
CLASS="LITERAL"
>FOR</TT
>
loop (Pavel Stehule)
</P
></LI
><LI
><P
> Support <TT
CLASS="LITERAL"
>RETURN QUERY EXECUTE</TT
> (Pavel
Stehule)
</P
></LI
><LI
><P
> Improve the <TT
CLASS="LITERAL"
>RAISE</TT
> command (Pavel Stehule)
<P
></P
></P><UL
><LI
><P
> Support <TT
CLASS="LITERAL"
>DETAIL</TT
> and <TT
CLASS="LITERAL"
>HINT</TT
> fields
</P
></LI
><LI
><P
> Support specification of the <TT
CLASS="LITERAL"
>SQLSTATE</TT
> error code
</P
></LI
><LI
><P
> Support an exception name parameter
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>RAISE</TT
> without parameters in an exception
block to re-throw the current error
</P
></LI
></UL
><P>
</P
></LI
><LI
><P
> Allow specification of <TT
CLASS="VARNAME"
>SQLSTATE</TT
> codes
in <TT
CLASS="LITERAL"
>EXCEPTION</TT
> lists (Pavel Stehule)
</P
><P
> This is useful for handling custom <TT
CLASS="VARNAME"
>SQLSTATE</TT
> codes.
</P
></LI
><LI
><P
> Support the <TT
CLASS="LITERAL"
>CASE</TT
> statement (Pavel Stehule)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>RETURN QUERY</TT
> set the special <TT
CLASS="LITERAL"
>FOUND</TT
> and
<TT
CLASS="COMMAND"
>GET DIAGNOSTICS</TT
> <TT
CLASS="LITERAL"
>ROW_COUNT</TT
> variables
(Pavel Stehule)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>FETCH</TT
> and <TT
CLASS="COMMAND"
>MOVE</TT
> set the
<TT
CLASS="COMMAND"
>GET DIAGNOSTICS</TT
> <TT
CLASS="LITERAL"
>ROW_COUNT</TT
> variable
(Andrew Gierth)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>EXIT</TT
> without a label always exit the innermost
loop (Tom)
</P
><P
> Formerly, if there were a <TT
CLASS="LITERAL"
>BEGIN</TT
> block more closely nested
than any loop, it would exit that block instead. The new behavior
matches Oracle(TM) and is also what was previously stated by our own
documentation.
</P
></LI
><LI
><P
> Make processing of string literals and nested block comments
match the main SQL parser's processing (Tom)
</P
><P
> In particular, the format string in <TT
CLASS="COMMAND"
>RAISE</TT
> now works
the same as any other string literal, including being subject
to <TT
CLASS="VARNAME"
>standard_conforming_strings</TT
>. This change also
fixes other cases in which valid commands would fail when
<TT
CLASS="VARNAME"
>standard_conforming_strings</TT
> is on.
</P
></LI
><LI
><P
> Avoid memory leakage when the same function is called at varying
exception-block nesting depths (Tom)
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN96904"
>E.13.3.8. Client Applications</A
></H3
><P
></P
><UL
><LI
><P
> Fix <TT
CLASS="LITERAL"
>pg_ctl restart</TT
> to preserve command-line arguments
(Bruce)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>-w</TT
>/<TT
CLASS="LITERAL"
>--no-password</TT
> option that
prevents password prompting in all utilities that have a
<TT
CLASS="LITERAL"
>-W</TT
>/<TT
CLASS="LITERAL"
>--password</TT
> option (Peter)
</P
></LI
><LI
><P
> Remove <TT
CLASS="OPTION"
>-q</TT
> (quiet) option of <SPAN
CLASS="APPLICATION"
>createdb</SPAN
>,
<SPAN
CLASS="APPLICATION"
>createuser</SPAN
>, <SPAN
CLASS="APPLICATION"
>dropdb</SPAN
>,
<SPAN
CLASS="APPLICATION"
>dropuser</SPAN
> (Peter)
</P
><P
> These options have had no effect since <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
8.3.
</P
></LI
></UL
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96925"
>E.13.3.8.1. <SPAN
CLASS="APPLICATION"
>psql</SPAN
></A
></H4
><P
></P
><UL
><LI
><P
> Remove verbose startup banner; now just suggest <TT
CLASS="LITERAL"
>help</TT
>
(Joshua Drake)
</P
></LI
><LI
><P
> Make <TT
CLASS="LITERAL"
>help</TT
> show common backslash commands (Greg
Sabino Mullane)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>\pset format wrapped</TT
> mode to wrap output to the
screen width, or file/pipe output too if <TT
CLASS="LITERAL"
>\pset columns</TT
>
is set (Bryce Nesbitt)
</P
></LI
><LI
><P
> Allow all supported spellings of boolean values in <TT
CLASS="COMMAND"
>\pset</TT
>,
rather than just <TT
CLASS="LITERAL"
>on</TT
> and <TT
CLASS="LITERAL"
>off</TT
> (Bruce)
</P
><P
> Formerly, any string other than <SPAN
CLASS="QUOTE"
>"off"</SPAN
> was silently taken
to mean <TT
CLASS="LITERAL"
>true</TT
>. <SPAN
CLASS="APPLICATION"
>psql</SPAN
> will now complain
about unrecognized spellings (but still take them as <TT
CLASS="LITERAL"
>true</TT
>).
</P
></LI
><LI
><P
> Use the pager for wide output (Bruce)
</P
></LI
><LI
><P
> Require a space between a one-letter backslash command and its first
argument (Bernd Helmle)
</P
><P
> This removes a historical source of ambiguity.
</P
></LI
><LI
><P
> Improve tab completion support for schema-qualified and
quoted identifiers (Greg Sabino Mullane)
</P
></LI
><LI
><P
> Add optional <TT
CLASS="LITERAL"
>on</TT
>/<TT
CLASS="LITERAL"
>off</TT
> argument for
<TT
CLASS="COMMAND"
>\timing</TT
> (David Fetter)
</P
></LI
><LI
><P
> Display access control rights on multiple lines (Brendan
Jurd, Andreas Scherbaum)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\l</TT
> show database access privileges (Andrew Gilligan)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\l+</TT
> show database sizes, if permissions
allow (Andrew Gilligan)
</P
></LI
><LI
><P
> Add the <TT
CLASS="COMMAND"
>\ef</TT
> command to edit function definitions
(Abhijit Menon-Sen)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN96972"
>E.13.3.8.2. <SPAN
CLASS="APPLICATION"
>psql</SPAN
> \d* commands</A
></H4
><P
></P
><UL
><LI
><P
> Make <TT
CLASS="COMMAND"
>\d*</TT
> commands that do not have a pattern argument
show system objects only if the <TT
CLASS="LITERAL"
>S</TT
> modifier is specified
(Greg Sabino Mullane, Bruce)
</P
><P
> The former behavior was inconsistent across different variants
of <TT
CLASS="COMMAND"
>\d</TT
>, and in most cases it provided no easy way to see
just user objects.
</P
></LI
><LI
><P
> Improve <TT
CLASS="COMMAND"
>\d*</TT
> commands to work with older
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> server versions (back to 7.4),
not only the current server version
(Guillaume Lelarge)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\d</TT
> show foreign-key constraints that reference
the selected table (Kenneth D'Souza)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\d</TT
> on a sequence show its column values
(Euler Taveira de Oliveira)
</P
></LI
><LI
><P
> Add column storage type and other relation options to the
<TT
CLASS="COMMAND"
>\d+</TT
> display (Gregory Stark, Euler Taveira de
Oliveira)
</P
></LI
><LI
><P
> Show relation size in <TT
CLASS="COMMAND"
>\dt+</TT
> output (Dickson S.
Guedes)
</P
></LI
><LI
><P
> Show the possible values of <TT
CLASS="LITERAL"
>enum</TT
> types in <TT
CLASS="COMMAND"
>\dT+</TT
>
(David Fetter)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>\dC</TT
> to accept a wildcard pattern, which matches
either datatype involved in the cast (Tom)
</P
></LI
><LI
><P
> Add a function type column to <TT
CLASS="COMMAND"
>\df</TT
>'s output, and add
options to list only selected types of functions (David Fetter)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>\df</TT
> not hide functions that take or return
type <TT
CLASS="TYPE"
>cstring</TT
> (Tom)
</P
><P
> Previously, such functions were hidden because most of them are
datatype I/O functions, which were deemed uninteresting. The new
policy about hiding system functions by default makes this wart
unnecessary.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN97013"
>E.13.3.8.3. <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
></A
></H4
><P
></P
><UL
><LI
><P
> Add a <TT
CLASS="LITERAL"
>--no-tablespaces</TT
> option to
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
>/<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
>/<SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
>
so that dumps can be restored to clusters that have non-matching
tablespace layouts (Gavin Roy)
</P
></LI
><LI
><P
> Remove <TT
CLASS="OPTION"
>-d</TT
> and <TT
CLASS="OPTION"
>-D</TT
> options from
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> and <SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> (Tom)
</P
><P
> These options were too frequently confused with the option to
select a database name in other <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
client applications. The functionality is still available,
but you must now spell out the long option name
<TT
CLASS="OPTION"
>--inserts</TT
> or <TT
CLASS="OPTION"
>--column-inserts</TT
>.
</P
></LI
><LI
><P
> Remove <TT
CLASS="OPTION"
>-i</TT
>/<TT
CLASS="OPTION"
>--ignore-version</TT
> option from
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> and <SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> (Tom)
</P
><P
> Use of this option does not throw an error, but it has no
effect. This option was removed because the version checks
are necessary for safety.
</P
></LI
><LI
><P
> Disable <TT
CLASS="VARNAME"
>statement_timeout</TT
> during dump and restore
(Joshua Drake)
</P
></LI
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
>/<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> option
<TT
CLASS="OPTION"
>--lock-wait-timeout</TT
> (David Gould)
</P
><P
> This allows dumps to fail if unable to acquire a shared lock
within the specified amount of time.
</P
></LI
><LI
><P
> Reorder <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> <TT
CLASS="LITERAL"
>--data-only</TT
> output
to dump tables referenced by foreign keys before
the referencing tables (Tom)
</P
><P
> This allows data loads when foreign keys are already present.
If circular references make a safe ordering impossible, a
<TT
CLASS="LITERAL"
>NOTICE</TT
> is issued.
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
>, <SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
>, and
<SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
> to use a specified role (Benedek
László)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
> to use multiple concurrent
connections to do the restore (Andrew)
</P
><P
> The number of concurrent connections is controlled by the option
<TT
CLASS="LITERAL"
>--jobs</TT
>. This is supported only for custom-format archives.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN97065"
>E.13.3.9. Programming Tools</A
></H3
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN97067"
>E.13.3.9.1. <SPAN
CLASS="APPLICATION"
>libpq</SPAN
></A
></H4
><P
></P
><UL
><LI
><P
> Allow the <TT
CLASS="TYPE"
>OID</TT
> to be specified when importing a large
object, via new function <CODE
CLASS="FUNCTION"
>lo_import_with_oid()</CODE
> (Tatsuo)
</P
></LI
><LI
><P
> Add <SPAN
CLASS="QUOTE"
>"events"</SPAN
> support (Andrew Chernow, Merlin Moncure)
</P
><P
> This adds the ability to register callbacks to manage private
data associated with <TT
CLASS="STRUCTNAME"
>PGconn</TT
> and <TT
CLASS="STRUCTNAME"
>PGresult</TT
>
objects.
</P
></LI
><LI
><P
> Improve error handling to allow the return of multiple
error messages as multi-line error reports (Magnus)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>PQexecParams()</CODE
> and related functions return
<TT
CLASS="VARNAME"
>PGRES_EMPTY_QUERY</TT
> for an empty query (Tom)
</P
><P
> They previously returned <TT
CLASS="VARNAME"
>PGRES_COMMAND_OK</TT
>.
</P
></LI
><LI
><P
> Document how to avoid the overhead of <CODE
CLASS="FUNCTION"
>WSACleanup()</CODE
>
on Windows (Andrew Chernow)
</P
></LI
><LI
><P
> Do not rely on Kerberos tickets to determine the default database
username (Magnus)
</P
><P
> Previously, a Kerberos-capable build of libpq would use the
principal name from any available Kerberos ticket as default
database username, even if the connection wasn't using Kerberos
authentication. This was deemed inconsistent and confusing.
The default username is now determined the same way with or
without Kerberos. Note however that the database username must still
match the ticket when Kerberos authentication is used.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN97095"
>E.13.3.9.2. <SPAN
CLASS="APPLICATION"
>libpq</SPAN
> <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> (Secure Sockets Layer)
support</A
></H4
><P
></P
><UL
><LI
><P
> Fix certificate validation for <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> connections
(Magnus)
</P
><P
> <SPAN
CLASS="APPLICATION"
>libpq</SPAN
> now supports verifying both the certificate
and the name of the server when making <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
>
connections. If a root certificate is not available to use for
verification, <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> connections will fail. The
<TT
CLASS="LITERAL"
>sslmode</TT
> parameter is used to enable certificate
verification and set the level of checking.
The default is still not to do any verification, allowing connections
to SSL-enabled servers without requiring a root certificate on the
client.
</P
></LI
><LI
><P
> Support wildcard server certificates (Magnus)
</P
><P
> If a certificate <ACRONYM
CLASS="ACRONYM"
>CN</ACRONYM
> starts with <TT
CLASS="LITERAL"
>*</TT
>, it will
be treated as a wildcard when matching the hostname, allowing the
use of the same certificate for multiple servers.
</P
></LI
><LI
><P
> Allow the file locations for client certificates to be specified
(Mark Woodward, Alvaro, Magnus)
</P
></LI
><LI
><P
> Add a <CODE
CLASS="FUNCTION"
>PQinitOpenSSL</CODE
> function to allow greater control
over OpenSSL/libcrypto initialization (Andrew Chernow)
</P
></LI
><LI
><P
> Make <SPAN
CLASS="APPLICATION"
>libpq</SPAN
> unregister its <SPAN
CLASS="APPLICATION"
>OpenSSL</SPAN
>
callbacks when no database connections remain open
(Bruce, Magnus, Russell Smith)
</P
><P
> This is required for applications that unload the libpq library,
otherwise invalid <SPAN
CLASS="APPLICATION"
>OpenSSL</SPAN
> callbacks will remain.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN97124"
>E.13.3.9.3. <SPAN
CLASS="APPLICATION"
>ecpg</SPAN
></A
></H4
><P
></P
><UL
><LI
><P
> Add localization support for messages (Euler Taveira de
Oliveira)
</P
></LI
><LI
><P
> ecpg parser is now automatically generated from the server
parser (Michael)
</P
><P
> Previously the ecpg parser was hand-maintained.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT4"
><H4
CLASS="SECT4"
><A
NAME="AEN97133"
>E.13.3.9.4. Server Programming Interface (<ACRONYM
CLASS="ACRONYM"
>SPI</ACRONYM
>)</A
></H4
><P
></P
><UL
><LI
><P
> Add support for single-use plans with out-of-line
parameters (Tom)
</P
></LI
><LI
><P
> Add new <TT
CLASS="VARNAME"
>SPI_OK_REWRITTEN</TT
> return code for
<CODE
CLASS="FUNCTION"
>SPI_execute()</CODE
> (Heikki)
</P
><P
> This is used when a command is rewritten to another type of
command.
</P
></LI
><LI
><P
> Remove unnecessary inclusions from <TT
CLASS="FILENAME"
>executor/spi.h</TT
> (Tom)
</P
><P
> SPI-using modules might need to add some <TT
CLASS="LITERAL"
>#include</TT
>
lines if they were depending on <TT
CLASS="FILENAME"
>spi.h</TT
> to include
things for them.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN97150"
>E.13.3.10. Build Options</A
></H3
><P
></P
><UL
><LI
><P
> Update build system to use <SPAN
CLASS="PRODUCTNAME"
>Autoconf</SPAN
> 2.61 (Peter)
</P
></LI
><LI
><P
> Require <SPAN
CLASS="PRODUCTNAME"
>GNU bison</SPAN
> for source code builds (Peter)
</P
><P
> This has effectively been required for several years, but now there
is no infrastructure claiming to support other parser tools.
</P
></LI
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>pg_config</SPAN
> <TT
CLASS="LITERAL"
>--htmldir</TT
> option
(Peter)
</P
></LI
><LI
><P
> Pass <TT
CLASS="TYPE"
>float4</TT
> by value inside the server (Zoltan
Boszormenyi)
</P
><P
> Add <SPAN
CLASS="APPLICATION"
>configure</SPAN
> option
<TT
CLASS="LITERAL"
>--disable-float4-byval</TT
> to use the old behavior.
External C functions that use old-style (version 0) call convention
and pass or return <TT
CLASS="TYPE"
>float4</TT
> values will be broken by this
change, so you may need the <SPAN
CLASS="APPLICATION"
>configure</SPAN
> option if you
have such functions and don't want to update them.
</P
></LI
><LI
><P
> Pass <TT
CLASS="TYPE"
>float8</TT
>, <TT
CLASS="TYPE"
>int8</TT
>, and related datatypes
by value inside the server on 64-bit platforms (Zoltan Boszormenyi)
</P
><P
> Add <SPAN
CLASS="APPLICATION"
>configure</SPAN
> option
<TT
CLASS="LITERAL"
>--disable-float8-byval</TT
> to use the old behavior.
As above, this change might break old-style external C functions.
</P
></LI
><LI
><P
> Add configure options <TT
CLASS="LITERAL"
>--with-segsize</TT
>,
<TT
CLASS="LITERAL"
>--with-blocksize</TT
>, <TT
CLASS="LITERAL"
>--with-wal-blocksize</TT
>,
<TT
CLASS="LITERAL"
>--with-wal-segsize</TT
> (Zdenek Kotala, Tom)
</P
><P
> This simplifies build-time control over several constants that
previously could only be changed by editing
<TT
CLASS="FILENAME"
>pg_config_manual.h</TT
>.
</P
></LI
><LI
><P
> Allow threaded builds on <SPAN
CLASS="PRODUCTNAME"
>Solaris</SPAN
> 2.5 (Bruce)
</P
></LI
><LI
><P
> Use the system's <CODE
CLASS="FUNCTION"
>getopt_long()</CODE
> on <SPAN
CLASS="PRODUCTNAME"
>Solaris</SPAN
>
(Zdenek Kotala, Tom)
</P
><P
> This makes option processing more consistent with what Solaris users
expect.
</P
></LI
><LI
><P
> Add support for the <SPAN
CLASS="PRODUCTNAME"
>Sun Studio</SPAN
> compiler on
<SPAN
CLASS="PRODUCTNAME"
>Linux</SPAN
> (Julius Stroffek)
</P
></LI
><LI
><P
> Append the major version number to the backend <SPAN
CLASS="APPLICATION"
>gettext</SPAN
>
domain, and the <TT
CLASS="LITERAL"
>soname</TT
> major version number to
libraries' <SPAN
CLASS="APPLICATION"
>gettext</SPAN
> domain (Peter)
</P
><P
> This simplifies parallel installations of multiple versions.
</P
></LI
><LI
><P
> Add support for code coverage testing with <SPAN
CLASS="APPLICATION"
>gcov</SPAN
>
(Michelle Caisse)
</P
></LI
><LI
><P
> Allow out-of-tree builds on <SPAN
CLASS="PRODUCTNAME"
>Mingw</SPAN
> and
<SPAN
CLASS="PRODUCTNAME"
>Cygwin</SPAN
> (Richard Evans)
</P
></LI
><LI
><P
> Fix the use of <SPAN
CLASS="PRODUCTNAME"
>Mingw</SPAN
> as a cross-compiling source
platform (Peter)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN97215"
>E.13.3.11. Source Code</A
></H3
><P
></P
><UL
><LI
><P
> Support 64-bit time zone data files (Heikki)
</P
><P
> This adds support for daylight saving time (<ACRONYM
CLASS="ACRONYM"
>DST</ACRONYM
>)
calculations beyond the year 2038.
</P
></LI
><LI
><P
> Deprecate use of platform's <TT
CLASS="TYPE"
>time_t</TT
> data type (Tom)
</P
><P
> Some platforms have migrated to 64-bit <TT
CLASS="TYPE"
>time_t</TT
>, some have
not, and Windows can't make up its mind what it's doing. Define
<TT
CLASS="TYPE"
>pg_time_t</TT
> to have the same meaning as <TT
CLASS="TYPE"
>time_t</TT
>,
but always be 64 bits (unless the platform has no 64-bit integer type),
and use that type in all module APIs and on-disk data formats.
</P
></LI
><LI
><P
> Fix bug in handling of the time zone database when cross-compiling
(Richard Evans)
</P
></LI
><LI
><P
> Link backend object files in one step, rather than in stages
(Peter)
</P
></LI
><LI
><P
> Improve <SPAN
CLASS="APPLICATION"
>gettext</SPAN
> support to allow better translation
of plurals (Peter)
</P
></LI
><LI
><P
> Add message translation support to the PL languages (Alvaro, Peter)
</P
></LI
><LI
><P
> Add more <SPAN
CLASS="APPLICATION"
>DTrace</SPAN
> probes (Robert Lor)
</P
></LI
><LI
><P
> Enable <SPAN
CLASS="APPLICATION"
>DTrace</SPAN
> support on <SPAN
CLASS="APPLICATION"
>Mac OS X
Leopard</SPAN
> and other non-Solaris platforms (Robert Lor)
</P
></LI
><LI
><P
> Simplify and standardize conversions between C strings and
<TT
CLASS="TYPE"
>text</TT
> datums, by providing common functions for the purpose
(Brendan Jurd, Tom)
</P
></LI
><LI
><P
> Clean up the <TT
CLASS="FILENAME"
>include/catalog/</TT
> header files so that
frontend programs can include them without including
<TT
CLASS="FILENAME"
>postgres.h</TT
>
(Zdenek Kotala)
</P
></LI
><LI
><P
> Make <TT
CLASS="TYPE"
>name</TT
> char-aligned, and suppress zero-padding of
<TT
CLASS="TYPE"
>name</TT
> entries in indexes (Tom)
</P
></LI
><LI
><P
> Recover better if dynamically-loaded code executes <CODE
CLASS="FUNCTION"
>exit()</CODE
>
(Tom)
</P
></LI
><LI
><P
> Add a hook to let plug-ins monitor the executor (Itagaki
Takahiro)
</P
></LI
><LI
><P
> Add a hook to allow the planner's statistics lookup behavior to
be overridden (Simon Riggs)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>shmem_startup_hook()</CODE
> for custom shared memory
requirements (Tom)
</P
></LI
><LI
><P
> Replace the index access method <CODE
CLASS="FUNCTION"
>amgetmulti</CODE
> entry point
with <CODE
CLASS="FUNCTION"
>amgetbitmap</CODE
>, and extend the API for
<CODE
CLASS="FUNCTION"
>amgettuple</CODE
> to support run-time determination of
operator lossiness (Heikki, Tom, Teodor)
</P
><P
> The API for GIN and GiST opclass <CODE
CLASS="FUNCTION"
>consistent</CODE
> functions
has been extended as well.
</P
></LI
><LI
><P
> Add support for partial-match searches in <ACRONYM
CLASS="ACRONYM"
>GIN</ACRONYM
> indexes
(Teodor Sigaev, Oleg Bartunov)
</P
></LI
><LI
><P
> Replace <TT
CLASS="STRUCTNAME"
>pg_class</TT
> column <TT
CLASS="STRUCTFIELD"
>reltriggers</TT
>
with boolean <TT
CLASS="STRUCTFIELD"
>relhastriggers</TT
> (Simon)
</P
><P
> Also remove unused <TT
CLASS="STRUCTNAME"
>pg_class</TT
> columns
<TT
CLASS="STRUCTFIELD"
>relukeys</TT
>, <TT
CLASS="STRUCTFIELD"
>relfkeys</TT
>, and
<TT
CLASS="STRUCTFIELD"
>relrefs</TT
>.
</P
></LI
><LI
><P
> Add a <TT
CLASS="STRUCTFIELD"
>relistemp</TT
> column to <TT
CLASS="STRUCTNAME"
>pg_class</TT
>
to ease identification of temporary tables (Tom)
</P
></LI
><LI
><P
> Move platform <ACRONYM
CLASS="ACRONYM"
>FAQ</ACRONYM
>s into the main documentation
(Peter)
</P
></LI
><LI
><P
> Prevent parser input files from being built with any conflicts
(Peter)
</P
></LI
><LI
><P
> Add support for the <TT
CLASS="LITERAL"
>KOI8U</TT
> (Ukrainian) encoding
(Peter)
</P
></LI
><LI
><P
> Add Japanese message translations (Japan PostgreSQL Users Group)
</P
><P
> This used to be maintained as a separate project.
</P
></LI
><LI
><P
> Fix problem when setting <TT
CLASS="VARNAME"
>LC_MESSAGES</TT
> on
<SPAN
CLASS="APPLICATION"
>MSVC</SPAN
>-built systems (Hiroshi Inoue, Hiroshi
Saito, Magnus)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN97305"
>E.13.3.12. Contrib</A
></H3
><P
></P
><UL
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/auto_explain</TT
> to automatically run
<TT
CLASS="COMMAND"
>EXPLAIN</TT
> on queries exceeding a specified duration
(Itagaki Takahiro, Tom)
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/btree_gin</TT
> to allow GIN indexes to
handle more datatypes (Oleg, Teodor)
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/citext</TT
> to provide a case-insensitive,
multibyte-aware text data type (David Wheeler)
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/pg_stat_statements</TT
> for server-wide
tracking of statement execution statistics (Itagaki Takahiro)
</P
></LI
><LI
><P
> Add duration and query mode options to <TT
CLASS="FILENAME"
>contrib/pgbench</TT
>
(Itagaki Takahiro)
</P
></LI
><LI
><P
> Make <TT
CLASS="FILENAME"
>contrib/pgbench</TT
> use table names
<TT
CLASS="STRUCTNAME"
>pgbench_accounts</TT
>, <TT
CLASS="STRUCTNAME"
>pgbench_branches</TT
>,
<TT
CLASS="STRUCTNAME"
>pgbench_history</TT
>, and <TT
CLASS="STRUCTNAME"
>pgbench_tellers</TT
>,
rather than just <TT
CLASS="STRUCTNAME"
>accounts</TT
>, <TT
CLASS="STRUCTNAME"
>branches</TT
>,
<TT
CLASS="STRUCTNAME"
>history</TT
>, and <TT
CLASS="STRUCTNAME"
>tellers</TT
> (Tom)
</P
><P
> This is to reduce the risk of accidentally destroying real data
by running <SPAN
CLASS="APPLICATION"
>pgbench</SPAN
>.
</P
></LI
><LI
><P
> Fix <TT
CLASS="FILENAME"
>contrib/pgstattuple</TT
> to handle tables and
indexes with over 2 billion pages (Tatsuhito Kasahara)
</P
></LI
><LI
><P
> In <TT
CLASS="FILENAME"
>contrib/fuzzystrmatch</TT
>, add a version of the
Levenshtein string-distance function that allows the user to
specify the costs of insertion, deletion, and substitution
(Volkan Yazici)
</P
></LI
><LI
><P
> Make <TT
CLASS="FILENAME"
>contrib/ltree</TT
> support multibyte encodings
(laser)
</P
></LI
><LI
><P
> Enable <TT
CLASS="FILENAME"
>contrib/dblink</TT
> to use connection information
stored in the SQL/MED catalogs (Joe Conway)
</P
></LI
><LI
><P
> Improve <TT
CLASS="FILENAME"
>contrib/dblink</TT
>'s reporting of errors from
the remote server (Joe Conway)
</P
></LI
><LI
><P
> Make <TT
CLASS="FILENAME"
>contrib/dblink</TT
> set <TT
CLASS="VARNAME"
>client_encoding</TT
>
to match the local database's encoding (Joe Conway)
</P
><P
> This prevents encoding problems when communicating with a remote
database that uses a different encoding.
</P
></LI
><LI
><P
> Make sure <TT
CLASS="FILENAME"
>contrib/dblink</TT
> uses a password supplied
by the user, and not accidentally taken from the server's
<TT
CLASS="FILENAME"
>.pgpass</TT
> file (Joe Conway)
</P
><P
> This is a minor security enhancement.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>fsm_page_contents()</CODE
>
to <TT
CLASS="FILENAME"
>contrib/pageinspect</TT
> (Heikki)
</P
></LI
><LI
><P
> Modify <CODE
CLASS="FUNCTION"
>get_raw_page()</CODE
> to support free space map
(<TT
CLASS="FILENAME"
>*_fsm</TT
>) files. Also update
<TT
CLASS="FILENAME"
>contrib/pg_freespacemap</TT
>.
</P
></LI
><LI
><P
> Add support for multibyte encodings to <TT
CLASS="FILENAME"
>contrib/pg_trgm</TT
>
(Teodor)
</P
></LI
><LI
><P
> Rewrite <TT
CLASS="FILENAME"
>contrib/intagg</TT
> to use new
functions <CODE
CLASS="FUNCTION"
>array_agg()</CODE
> and <CODE
CLASS="FUNCTION"
>unnest()</CODE
>
(Tom)
</P
></LI
><LI
><P
> Make <TT
CLASS="FILENAME"
>contrib/pg_standby</TT
> recover all available WAL before
failover (Fujii Masao, Simon, Heikki)
</P
><P
> To make this work safely, you now need to set the new
<TT
CLASS="LITERAL"
>recovery_end_command</TT
> option in <TT
CLASS="FILENAME"
>recovery.conf</TT
>
to clean up the trigger file after failover. <SPAN
CLASS="APPLICATION"
>pg_standby</SPAN
>
will no longer remove the trigger file itself.
</P
></LI
><LI
><P
> <TT
CLASS="FILENAME"
>contrib/pg_standby</TT
>'s <TT
CLASS="OPTION"
>-l</TT
> option is now a no-op,
because it is unsafe to use a symlink (Simon)
</P
></LI
></UL
></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-8-4-1.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-8-3-19.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Release 8.4.1</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 8.3.19</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
distrib
>
Mandriva
>
current
>
i586
>
media
>
main-updates
>
by-pkgid
>
fc62ce67f262cdcd253dc7f849ce3223
>
files
>
680
