| Current File : //usr/share/doc/postgresql-9.2.24/html/release-8-3.html |
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Release 8.3</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 9.2.24 Documentation"
HREF="index.html"><LINK
REL="UP"
TITLE="Release Notes"
HREF="release.html"><LINK
REL="PREVIOUS"
TITLE="Release 8.3.1"
HREF="release-8-3-1.html"><LINK
REL="NEXT"
TITLE="Release 8.2.23"
HREF="release-8-2-23.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="2017-11-06T22:43:11"></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"
><A
HREF="index.html"
>PostgreSQL 9.2.24 Documentation</A
></TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="top"
><A
TITLE="Release 8.3.1"
HREF="release-8-3-1.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 8.2.23"
HREF="release-8-2-23.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RELEASE-8-3"
>E.121. Release 8.3</A
></H1
><DIV
CLASS="FORMALPARA"
><P
><B
>Release date: </B
>2008-02-04</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN131333"
>E.121.1. Overview</A
></H2
><P
> With significant new functionality and performance enhancements,
this release represents a major leap forward for
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>. This was made possible by a growing
community that has dramatically accelerated the pace of
development. This release adds the following major features:
</P
><P
></P
><UL
><LI
><P
> Full text search is integrated into the core database system
</P
></LI
><LI
><P
> Support for the SQL/XML standard, including new operators and an
<TT
CLASS="TYPE"
>XML</TT
> data type
</P
></LI
><LI
><P
> Enumerated data types (<TT
CLASS="TYPE"
>ENUM</TT
>)
</P
></LI
><LI
><P
> Arrays of composite types
</P
></LI
><LI
><P
> Universally Unique Identifier (<TT
CLASS="TYPE"
>UUID</TT
>) data type
</P
></LI
><LI
><P
> Add control over whether <TT
CLASS="LITERAL"
>NULL</TT
>s sort first or last
</P
></LI
><LI
><P
> Updatable cursors
</P
></LI
><LI
><P
> Server configuration parameters can now be set on a per-function
basis
</P
></LI
><LI
><P
> User-defined types can now have type modifiers
</P
></LI
><LI
><P
> Automatically re-plan cached queries when table
definitions change or statistics are updated
</P
></LI
><LI
><P
> Numerous improvements in logging and statistics collection
</P
></LI
><LI
><P
> Support Security Service Provider Interface (<ACRONYM
CLASS="ACRONYM"
>SSPI</ACRONYM
>) for
authentication on Windows
</P
></LI
><LI
><P
> Support multiple concurrent autovacuum processes, and other
autovacuum improvements
</P
></LI
><LI
><P
> Allow the whole <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> distribution to be compiled
with <SPAN
CLASS="PRODUCTNAME"
>Microsoft Visual C++</SPAN
>
</P
></LI
></UL
><P
> Major performance improvements are listed below. Most of
these enhancements are automatic and do not require user changes or
tuning:
</P
><P
></P
><UL
><LI
><P
> Asynchronous commit delays writes to WAL during transaction commit
</P
></LI
><LI
><P
> Checkpoint writes can be spread over a longer time period to smooth
the I/O spike during each checkpoint
</P
></LI
><LI
><P
> Heap-Only Tuples (<ACRONYM
CLASS="ACRONYM"
>HOT</ACRONYM
>) accelerate space reuse for
most <TT
CLASS="COMMAND"
>UPDATE</TT
>s and <TT
CLASS="COMMAND"
>DELETE</TT
>s
</P
></LI
><LI
><P
> Just-in-time background writer strategy improves disk write
efficiency
</P
></LI
><LI
><P
> Using non-persistent transaction IDs for read-only transactions
reduces overhead and <TT
CLASS="COMMAND"
>VACUUM</TT
> requirements
</P
></LI
><LI
><P
> Per-field and per-row storage overhead has been reduced
</P
></LI
><LI
><P
> Large sequential scans no longer force out frequently used
cached pages
</P
></LI
><LI
><P
> Concurrent large sequential scans can now share disk reads
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>ORDER BY ... LIMIT</TT
> can be done without sorting
</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="AEN131399"
>E.121.2. Migration to Version 8.3</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="AEN131404"
>E.121.2.1. General</A
></H3
><P
></P
><UL
><LI
><P
> Non-character data types are no longer automatically cast to
<TT
CLASS="TYPE"
>TEXT</TT
> (Peter, Tom)
</P
><P
> Previously, if a non-character value was supplied to an operator or
function that requires <TT
CLASS="TYPE"
>text</TT
> input, it was automatically
cast to <TT
CLASS="TYPE"
>text</TT
>, for most (though not all) built-in data types.
This no longer happens: an explicit cast to <TT
CLASS="TYPE"
>text</TT
> is now
required for all non-character-string types. For example, these
expressions formerly worked:
</P><PRE
CLASS="PROGRAMLISTING"
>substr(current_date, 1, 4)
23 LIKE '2%'</PRE
><P>
but will now draw <SPAN
CLASS="QUOTE"
>"function does not exist"</SPAN
> and <SPAN
CLASS="QUOTE"
>"operator
does not exist"</SPAN
> errors respectively. Use an explicit cast instead:
</P><PRE
CLASS="PROGRAMLISTING"
>substr(current_date::text, 1, 4)
23::text LIKE '2%'</PRE
><P>
(Of course, you can use the more verbose <TT
CLASS="LITERAL"
>CAST()</TT
> syntax too.)
The reason for the change is that these automatic casts too often caused
surprising behavior. An example is that in previous releases, this
expression was accepted but did not do what was expected:
</P><PRE
CLASS="PROGRAMLISTING"
>current_date < 2017-11-17</PRE
><P>
This is actually comparing a date to an integer, which should be
(and now is) rejected — but in the presence of automatic
casts both sides were cast to <TT
CLASS="TYPE"
>text</TT
> and a textual comparison
was done, because the <TT
CLASS="LITERAL"
>text < text</TT
> operator was able
to match the expression when no other <TT
CLASS="LITERAL"
><</TT
> operator could.
</P
><P
> Types <TT
CLASS="TYPE"
>char(<TT
CLASS="REPLACEABLE"
><I
>n</I
></TT
>)</TT
> and
<TT
CLASS="TYPE"
>varchar(<TT
CLASS="REPLACEABLE"
><I
>n</I
></TT
>)</TT
> still cast to <TT
CLASS="TYPE"
>text</TT
>
automatically. Also, automatic casting to <TT
CLASS="TYPE"
>text</TT
> still works for
inputs to the concatenation (<TT
CLASS="LITERAL"
>||</TT
>) operator, so long as least
one input is a character-string type.
</P
></LI
><LI
><P
> Full text search features from <TT
CLASS="FILENAME"
>contrib/tsearch2</TT
> have
been moved into the core server, with some minor syntax changes
</P
><P
> <TT
CLASS="FILENAME"
>contrib/tsearch2</TT
> now contains a compatibility
interface.
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>ARRAY(SELECT ...)</TT
>, where the <TT
CLASS="COMMAND"
>SELECT</TT
>
returns no rows, now returns an empty array, rather than NULL
(Tom)
</P
></LI
><LI
><P
> The array type name for a base data type is no longer always the base
type's name with an underscore prefix
</P
><P
> The old naming convention is still honored when possible, but
application code should no longer depend on it. Instead
use the new <TT
CLASS="LITERAL"
>pg_type.typarray</TT
> column to
identify the array data type associated with a given type.
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>ORDER BY ... USING</TT
> <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> must now
use a less-than or greater-than <TT
CLASS="REPLACEABLE"
><I
>operator</I
></TT
> that is
defined in a btree operator class
</P
><P
> This restriction was added to prevent inconsistent results.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>SET LOCAL</TT
> changes now persist until
the end of the outermost transaction, unless rolled back (Tom)
</P
><P
> Previously <TT
CLASS="COMMAND"
>SET LOCAL</TT
>'s effects were lost
after subtransaction commit (<TT
CLASS="COMMAND"
>RELEASE SAVEPOINT</TT
>
or exit from a PL/pgSQL exception block).
</P
></LI
><LI
><P
> Commands rejected in transaction blocks are now also rejected in
multiple-statement query strings (Tom)
</P
><P
> For example, <TT
CLASS="LITERAL"
>"BEGIN; DROP DATABASE; COMMIT"</TT
> will now be
rejected even if submitted as a single query message.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>ROLLBACK</TT
> outside a transaction block now
issues <TT
CLASS="LITERAL"
>NOTICE</TT
> instead of <TT
CLASS="LITERAL"
>WARNING</TT
> (Bruce)
</P
></LI
><LI
><P
> Prevent <TT
CLASS="COMMAND"
>NOTIFY</TT
>/<TT
CLASS="COMMAND"
>LISTEN</TT
>/<TT
CLASS="COMMAND"
>UNLISTEN</TT
>
from accepting schema-qualified names (Bruce)
</P
><P
> Formerly, these commands accepted <TT
CLASS="LITERAL"
>schema.relation</TT
> but
ignored the schema part, which was confusing.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>ALTER SEQUENCE</TT
> no longer affects the sequence's
<CODE
CLASS="FUNCTION"
>currval()</CODE
> state (Tom)
</P
></LI
><LI
><P
> Foreign keys now must match indexable conditions for
cross-data-type references (Tom)
</P
><P
> This improves semantic consistency and helps avoid
performance problems.
</P
></LI
><LI
><P
> Restrict object size functions to users who have reasonable
permissions to view such information (Tom)
</P
><P
> For example, <CODE
CLASS="FUNCTION"
>pg_database_size()</CODE
> now requires
<TT
CLASS="LITERAL"
>CONNECT</TT
> permission, which is granted to everyone by
default. <CODE
CLASS="FUNCTION"
>pg_tablespace_size()</CODE
> requires
<TT
CLASS="LITERAL"
>CREATE</TT
> permission in the tablespace, or is allowed if
the tablespace is the default tablespace for the database.
</P
></LI
><LI
><P
> Remove the undocumented <TT
CLASS="LITERAL"
>!!=</TT
> (not in) operator (Tom)
</P
><P
> <TT
CLASS="LITERAL"
>NOT IN (SELECT ...)</TT
> is the proper way to
perform this operation.
</P
></LI
><LI
><P
> Internal hashing functions are now more uniformly-distributed (Tom)
</P
><P
> If application code was computing and storing hash values using
internal <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> hashing functions, the hash
values must be regenerated.
</P
></LI
><LI
><P
> C-code conventions for handling variable-length data values
have changed (Greg Stark, Tom)
</P
><P
> The new <CODE
CLASS="FUNCTION"
>SET_VARSIZE()</CODE
> macro <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>must</I
></SPAN
> be used
to set the length of generated <TT
CLASS="TYPE"
>varlena</TT
> values. Also, it
might be necessary to expand (<SPAN
CLASS="QUOTE"
>"de-TOAST"</SPAN
>) input values
in more cases.
</P
></LI
><LI
><P
> Continuous archiving no longer reports each successful archive
operation to the server logs unless <TT
CLASS="LITERAL"
>DEBUG</TT
> level is used
(Simon)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131505"
>E.121.2.2. Configuration Parameters</A
></H3
><P
></P
><UL
><LI
><P
> Numerous changes in administrative server parameters
</P
><P
> <TT
CLASS="VARNAME"
>bgwriter_lru_percent</TT
>,
<TT
CLASS="VARNAME"
>bgwriter_all_percent</TT
>,
<TT
CLASS="VARNAME"
>bgwriter_all_maxpages</TT
>,
<TT
CLASS="VARNAME"
>stats_start_collector</TT
>, and
<TT
CLASS="VARNAME"
>stats_reset_on_server_start</TT
> are removed.
<TT
CLASS="VARNAME"
>redirect_stderr</TT
> is renamed to
<TT
CLASS="VARNAME"
>logging_collector</TT
>.
<TT
CLASS="VARNAME"
>stats_command_string</TT
> is renamed to
<TT
CLASS="VARNAME"
>track_activities</TT
>.
<TT
CLASS="VARNAME"
>stats_block_level</TT
> and <TT
CLASS="VARNAME"
>stats_row_level</TT
>
are merged into <TT
CLASS="VARNAME"
>track_counts</TT
>.
A new boolean configuration parameter, <TT
CLASS="VARNAME"
>archive_mode</TT
>,
controls archiving. Autovacuum's default settings have changed.
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>stats_start_collector</TT
> parameter (Tom)
</P
><P
> We now always start the collector process, unless <ACRONYM
CLASS="ACRONYM"
>UDP</ACRONYM
>
socket creation fails.
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>stats_reset_on_server_start</TT
> parameter (Tom)
</P
><P
> This was removed because <CODE
CLASS="FUNCTION"
>pg_stat_reset()</CODE
>
can be used for this purpose.
</P
></LI
><LI
><P
> Commenting out a parameter in <TT
CLASS="FILENAME"
>postgresql.conf</TT
> now
causes it to revert to its default value (Joachim Wieland)
</P
><P
> Previously, commenting out an entry left the parameter's value unchanged
until the next server restart.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131538"
>E.121.2.3. Character Encodings</A
></H3
><P
></P
><UL
><LI
><P
> Add more checks for invalidly-encoded data (Andrew)
</P
><P
> This change plugs some holes that existed in literal backslash
escape string processing and <TT
CLASS="COMMAND"
>COPY</TT
> escape
processing. Now the de-escaped string is rechecked to see if the
result created an invalid multi-byte character.
</P
></LI
><LI
><P
> Disallow database encodings that are inconsistent with the server's
locale setting (Tom)
</P
><P
> On most platforms, <TT
CLASS="LITERAL"
>C</TT
> locale is the only locale that
will work with any database encoding. Other locale settings imply
a specific encoding and will misbehave if the database encoding
is something different. (Typical symptoms include bogus textual
sort order and wrong results from <CODE
CLASS="FUNCTION"
>upper()</CODE
> or
<CODE
CLASS="FUNCTION"
>lower()</CODE
>.) The server now rejects attempts to create
databases that have an incompatible encoding.
</P
></LI
><LI
><P
> Ensure that <CODE
CLASS="FUNCTION"
>chr()</CODE
> cannot create
invalidly-encoded values (Andrew)
</P
><P
> In UTF8-encoded databases the argument of <CODE
CLASS="FUNCTION"
>chr()</CODE
> is
now treated as a Unicode code point. In other multi-byte encodings
<CODE
CLASS="FUNCTION"
>chr()</CODE
>'s argument must designate a 7-bit ASCII
character. Zero is no longer accepted.
<CODE
CLASS="FUNCTION"
>ascii()</CODE
> has been adjusted to match.
</P
></LI
><LI
><P
> Adjust <CODE
CLASS="FUNCTION"
>convert()</CODE
> behavior to ensure encoding
validity (Andrew)
</P
><P
> The two argument form of <CODE
CLASS="FUNCTION"
>convert()</CODE
> has been
removed. The three argument form now takes a <TT
CLASS="TYPE"
>bytea</TT
>
first argument and returns a <TT
CLASS="TYPE"
>bytea</TT
>. To cover the
loss of functionality, three new functions have been added:
</P
><P
></P
><UL
><LI
><P
> <CODE
CLASS="FUNCTION"
>convert_from(bytea, name)</CODE
> returns
<TT
CLASS="TYPE"
>text</TT
> — converts the first argument from the named
encoding to the database encoding
</P
></LI
><LI
><P
> <CODE
CLASS="FUNCTION"
>convert_to(text, name)</CODE
> returns
<TT
CLASS="TYPE"
>bytea</TT
> — converts the first argument from the
database encoding to the named encoding
</P
></LI
><LI
><P
> <CODE
CLASS="FUNCTION"
>length(bytea, name)</CODE
> returns
<TT
CLASS="TYPE"
>integer</TT
> — gives the length of the first
argument in characters in the named encoding
</P
></LI
></UL
></LI
><LI
><P
> Remove <TT
CLASS="LITERAL"
>convert(argument USING conversion_name)</TT
>
(Andrew)
</P
><P
> Its behavior did not match the SQL standard.
</P
></LI
><LI
><P
> Make JOHAB encoding client-only (Tatsuo)
</P
><P
> JOHAB is not safe as a server-side encoding.
</P
></LI
></UL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN131585"
>E.121.3. Changes</A
></H2
><P
> Below you will find a detailed account of the
changes between <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 8.3 and
the previous major release.
</P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131589"
>E.121.3.1. Performance</A
></H3
><P
></P
><UL
><LI
><P
> Asynchronous commit delays writes to WAL during transaction commit
(Simon)
</P
><P
> This feature dramatically increases performance for short data-modifying
transactions. The disadvantage is that because disk writes are delayed,
if the database or operating system crashes before data is written to
the disk, committed data will be lost. This feature is useful for
applications that can accept some data loss. Unlike turning off
<TT
CLASS="VARNAME"
>fsync</TT
>, using asynchronous commit does not put
database consistency at risk; the worst case is that after a crash the
last few reportedly-committed transactions might not be committed after
all.
This feature is enabled by turning off <TT
CLASS="VARNAME"
>synchronous_commit</TT
>
(which can be done per-session or per-transaction, if some transactions
are critical and others are not).
<TT
CLASS="VARNAME"
>wal_writer_delay</TT
> can be adjusted to control the maximum
delay before transactions actually reach disk.
</P
></LI
><LI
><P
> Checkpoint writes can be spread over a longer time period to smooth
the I/O spike during each checkpoint (Itagaki Takahiro and Heikki
Linnakangas)
</P
><P
> Previously all modified buffers were forced to disk as quickly as
possible during a
checkpoint, causing an I/O spike that decreased server performance.
This new approach spreads out disk writes during checkpoints,
reducing peak I/O usage. (User-requested and shutdown checkpoints
are still written as quickly as possible.)
</P
></LI
><LI
><P
> Heap-Only Tuples (<ACRONYM
CLASS="ACRONYM"
>HOT</ACRONYM
>) accelerate space reuse for most
<TT
CLASS="COMMAND"
>UPDATE</TT
>s and <TT
CLASS="COMMAND"
>DELETE</TT
>s (Pavan Deolasee, with
ideas from many others)
</P
><P
> <TT
CLASS="COMMAND"
>UPDATE</TT
>s and <TT
CLASS="COMMAND"
>DELETE</TT
>s leave dead tuples
behind, as do failed <TT
CLASS="COMMAND"
>INSERT</TT
>s. Previously only
<TT
CLASS="COMMAND"
>VACUUM</TT
> could reclaim space taken by dead tuples. With
<ACRONYM
CLASS="ACRONYM"
>HOT</ACRONYM
> dead tuple space can be automatically reclaimed at
the time of <TT
CLASS="COMMAND"
>INSERT</TT
> or <TT
CLASS="COMMAND"
>UPDATE</TT
> if no changes
are made to indexed columns. This allows for more consistent
performance. Also, <ACRONYM
CLASS="ACRONYM"
>HOT</ACRONYM
> avoids adding duplicate index
entries.
</P
></LI
><LI
><P
> Just-in-time background writer strategy improves disk write
efficiency (Greg Smith, Itagaki Takahiro)
</P
><P
> This greatly reduces the need for manual tuning of the background
writer.
</P
></LI
><LI
><P
> Per-field and per-row storage overhead have been reduced
(Greg Stark, Heikki Linnakangas)
</P
><P
> Variable-length data types with data values less than 128 bytes long
will see a storage decrease of 3 to 6 bytes. For example, two adjacent
<TT
CLASS="TYPE"
>char(1)</TT
> fields now use 4 bytes instead of 16. Row headers
are also 4 bytes shorter than before.
</P
></LI
><LI
><P
> Using non-persistent transaction IDs for read-only transactions
reduces overhead and <TT
CLASS="COMMAND"
>VACUUM</TT
> requirements (Florian Pflug)
</P
><P
> Non-persistent transaction IDs do not increment the global
transaction counter. Therefore, they reduce the load on
<TT
CLASS="STRUCTNAME"
>pg_clog</TT
> and increase the time between forced
vacuums to prevent transaction ID wraparound.
Other performance
improvements were also made that should improve concurrency.
</P
></LI
><LI
><P
> Avoid incrementing the command counter after a read-only command (Tom)
</P
><P
> There was formerly a hard limit of 2<SUP
>32</SUP
>
(4 billion) commands per transaction. Now only commands that
actually changed the database count, so while this limit still
exists, it should be significantly less annoying.
</P
></LI
><LI
><P
> Create a dedicated <ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
> writer process to off-load
work from backends (Simon)
</P
></LI
><LI
><P
> Skip unnecessary WAL writes for <TT
CLASS="COMMAND"
>CLUSTER</TT
> and
<TT
CLASS="COMMAND"
>COPY</TT
> (Simon)
</P
><P
> Unless WAL archiving is enabled, the system now avoids WAL writes
for <TT
CLASS="COMMAND"
>CLUSTER</TT
> and just <CODE
CLASS="FUNCTION"
>fsync()</CODE
>s the
table at the end of the command. It also does the same for
<TT
CLASS="COMMAND"
>COPY</TT
> if the table was created in the same
transaction.
</P
></LI
><LI
><P
> Large sequential scans no longer force out frequently used
cached pages (Simon, Heikki, Tom)
</P
></LI
><LI
><P
> Concurrent large sequential scans can now share disk reads (Jeff Davis)
</P
><P
> This is accomplished by starting the new sequential scan in the
middle of the table (where another sequential scan is already
in-progress) and wrapping around to the beginning to finish. This
can affect the order of returned rows in a query that does not
specify <TT
CLASS="LITERAL"
>ORDER BY</TT
>. The <TT
CLASS="VARNAME"
>synchronize_seqscans</TT
>
configuration parameter can be used to disable this if necessary.
</P
></LI
><LI
><P
> <TT
CLASS="LITERAL"
>ORDER BY ... LIMIT</TT
> can be done without sorting
(Greg Stark)
</P
><P
> This is done by sequentially scanning the table and tracking just
the <SPAN
CLASS="QUOTE"
>"top N"</SPAN
> candidate rows, rather than performing a
full sort of the entire table. This is useful when there is no
matching index and the <TT
CLASS="LITERAL"
>LIMIT</TT
> is not large.
</P
></LI
><LI
><P
> Put a rate limit on messages sent to the statistics
collector by backends
(Tom)
</P
><P
> This reduces overhead for short transactions, but might sometimes
increase the delay before statistics are tallied.
</P
></LI
><LI
><P
> Improve hash join performance for cases with many NULLs (Tom)
</P
></LI
><LI
><P
> Speed up operator lookup for cases with non-exact datatype matches (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131662"
>E.121.3.2. Server</A
></H3
><P
></P
><UL
><LI
><P
> Autovacuum is now enabled by default (Alvaro)
</P
><P
> Several changes were made to eliminate disadvantages of having
autovacuum enabled, thereby justifying the change in default.
Several other autovacuum parameter defaults were also modified.
</P
></LI
><LI
><P
> Support multiple concurrent autovacuum processes (Alvaro, Itagaki
Takahiro)
</P
><P
> This allows multiple vacuums to run concurrently. This prevents
vacuuming of a large table from delaying vacuuming of smaller tables.
</P
></LI
><LI
><P
> Automatically re-plan cached queries when table
definitions change or statistics are updated (Tom)
</P
><P
> Previously PL/pgSQL functions that referenced temporary tables
would fail if the temporary table was dropped and recreated
between function invocations, unless <TT
CLASS="LITERAL"
>EXECUTE</TT
> was
used. This improvement fixes that problem and many related issues.
</P
></LI
><LI
><P
> Add a <TT
CLASS="VARNAME"
>temp_tablespaces</TT
> parameter to control
the tablespaces for temporary tables and files (Jaime Casanova,
Albert Cervera, Bernd Helmle)
</P
><P
> This parameter defines a list of tablespaces to be used. This
enables spreading the I/O load across multiple tablespaces. A random
tablespace is chosen each time a temporary object is created.
Temporary files are no longer stored in per-database
<TT
CLASS="FILENAME"
>pgsql_tmp/</TT
> directories but in per-tablespace
directories.
</P
></LI
><LI
><P
> Place temporary tables' TOAST tables in special schemas named
<TT
CLASS="LITERAL"
>pg_toast_temp_<TT
CLASS="REPLACEABLE"
><I
>nnn</I
></TT
></TT
> (Tom)
</P
><P
> This allows low-level code to recognize these tables as temporary,
which enables various optimizations such as not WAL-logging changes
and using local rather than shared buffers for access. This also
fixes a bug wherein backends unexpectedly held open file references
to temporary TOAST tables.
</P
></LI
><LI
><P
> Fix problem that a constant flow of new connection requests could
indefinitely delay the postmaster from completing a shutdown or
a crash restart (Tom)
</P
></LI
><LI
><P
> Guard against a very-low-probability data loss scenario by preventing
re-use of a deleted table's relfilenode until after the next
checkpoint (Heikki)
</P
></LI
><LI
><P
> Fix <TT
CLASS="COMMAND"
>CREATE CONSTRAINT TRIGGER</TT
>
to convert old-style foreign key trigger definitions into regular
foreign key constraints (Tom)
</P
><P
> This will ease porting of foreign key constraints carried forward from
pre-7.3 databases, if they were never converted using
<TT
CLASS="FILENAME"
>contrib/adddepend</TT
>.
</P
></LI
><LI
><P
> Fix <TT
CLASS="LITERAL"
>DEFAULT NULL</TT
> to override inherited defaults (Tom)
</P
><P
> <TT
CLASS="LITERAL"
>DEFAULT NULL</TT
> was formerly considered a noise phrase, but it
should (and now does) override non-null defaults that would otherwise
be inherited from a parent table or domain.
</P
></LI
><LI
><P
> Add new encodings EUC_JIS_2004 and SHIFT_JIS_2004 (Tatsuo)
</P
><P
> These new encodings can be converted to and from UTF-8.
</P
></LI
><LI
><P
> Change server startup log message from <SPAN
CLASS="QUOTE"
>"database system is
ready"</SPAN
> to <SPAN
CLASS="QUOTE"
>"database system is ready to accept
connections"</SPAN
>, and adjust its timing
</P
><P
> The message now appears only when the postmaster is really ready
to accept connections.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131707"
>E.121.3.3. Monitoring</A
></H3
><P
></P
><UL
><LI
><P
> Add <TT
CLASS="VARNAME"
>log_autovacuum_min_duration</TT
> parameter to
support configurable logging of autovacuum activity (Simon, Alvaro)
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>log_lock_waits</TT
> parameter to log lock waiting
(Simon)
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>log_temp_files</TT
> parameter to log temporary
file usage (Bill Moran)
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>log_checkpoints</TT
> parameter to improve logging
of checkpoints (Greg Smith, Heikki)
</P
></LI
><LI
><P
> <TT
CLASS="VARNAME"
>log_line_prefix</TT
> now supports
<TT
CLASS="LITERAL"
>%s</TT
> and <TT
CLASS="LITERAL"
>%c</TT
> escapes in all
processes (Andrew)
</P
><P
> Previously these escapes worked only for user sessions, not for
background database processes.
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>log_restartpoints</TT
> to control logging of
point-in-time recovery restart points (Simon)
</P
></LI
><LI
><P
> Last transaction end time is now logged at end of recovery and at
each logged restart point (Simon)
</P
></LI
><LI
><P
> Autovacuum now reports its activity start time in
<TT
CLASS="LITERAL"
>pg_stat_activity</TT
> (Tom)
</P
></LI
><LI
><P
> Allow server log output in comma-separated value (CSV) format (Arul
Shaji, Greg Smith, Andrew Dunstan)
</P
><P
> CSV-format log files can easily be loaded into a database table for
subsequent analysis.
</P
></LI
><LI
><P
> Use PostgreSQL-supplied timezone support for formatting timestamps
displayed in the server log (Tom)
</P
><P
> This avoids Windows-specific problems with localized time zone
names that are in the wrong encoding. There is a new
<TT
CLASS="VARNAME"
>log_timezone</TT
> parameter that controls the timezone
used in log messages, independently of the client-visible
<TT
CLASS="VARNAME"
>timezone</TT
> parameter.
</P
></LI
><LI
><P
> New system view <TT
CLASS="LITERAL"
>pg_stat_bgwriter</TT
> displays
statistics about background writer activity (Magnus)
</P
></LI
><LI
><P
> Add new columns for database-wide tuple statistics to
<TT
CLASS="LITERAL"
>pg_stat_database</TT
> (Magnus)
</P
></LI
><LI
><P
> Add an <TT
CLASS="LITERAL"
>xact_start</TT
> (transaction start time) column to
<TT
CLASS="LITERAL"
>pg_stat_activity</TT
> (Neil)
</P
><P
> This makes it easier to identify long-running transactions.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>n_live_tuples</TT
> and <TT
CLASS="LITERAL"
>n_dead_tuples</TT
> columns
to <TT
CLASS="LITERAL"
>pg_stat_all_tables</TT
> and related views (Glen
Parker)
</P
></LI
><LI
><P
> Merge <TT
CLASS="VARNAME"
>stats_block_level</TT
> and <TT
CLASS="VARNAME"
>stats_row_level</TT
>
parameters into a single parameter <TT
CLASS="VARNAME"
>track_counts</TT
>, which
controls all messages sent to the statistics collector process
(Tom)
</P
></LI
><LI
><P
> Rename <TT
CLASS="VARNAME"
>stats_command_string</TT
> parameter to
<TT
CLASS="VARNAME"
>track_activities</TT
> (Tom)
</P
></LI
><LI
><P
> Fix statistical counting of live and dead tuples to recognize that
committed and aborted transactions have different effects (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131771"
>E.121.3.4. Authentication</A
></H3
><P
></P
><UL
><LI
><P
> Support Security Service Provider Interface (<ACRONYM
CLASS="ACRONYM"
>SSPI</ACRONYM
>) for
authentication on Windows (Magnus)
</P
></LI
><LI
><P
> Support GSSAPI authentication (Henry Hotz, Magnus)
</P
><P
> This should be preferred to native Kerberos authentication because
GSSAPI is an industry standard.
</P
></LI
><LI
><P
> Support a global SSL configuration file (Victor Wagner)
</P
></LI
><LI
><P
> Add <TT
CLASS="VARNAME"
>ssl_ciphers</TT
> parameter to control accepted SSL ciphers
(Victor Wagner)
</P
></LI
><LI
><P
> Add a Kerberos realm parameter, <TT
CLASS="VARNAME"
>krb_realm</TT
> (Magnus)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131788"
>E.121.3.5. Write-Ahead Log (<ACRONYM
CLASS="ACRONYM"
>WAL</ACRONYM
>) and Continuous Archiving</A
></H3
><P
></P
><UL
><LI
><P
> Change the timestamps recorded in transaction WAL records from
time_t to TimestampTz representation (Tom)
</P
><P
> This provides sub-second resolution in WAL, which can be useful for
point-in-time recovery.
</P
></LI
><LI
><P
> Reduce WAL disk space needed by warm standby servers (Simon)
</P
><P
> This change allows a warm standby server to pass the name of the earliest
still-needed WAL file to the recovery script, allowing automatic removal
of no-longer-needed WAL files. This is done using <TT
CLASS="LITERAL"
>%r</TT
> in
the <TT
CLASS="VARNAME"
>restore_command</TT
> parameter of
<TT
CLASS="FILENAME"
>recovery.conf</TT
>.
</P
></LI
><LI
><P
> New boolean configuration parameter, <TT
CLASS="VARNAME"
>archive_mode</TT
>,
controls archiving (Simon)
</P
><P
> Previously setting <TT
CLASS="VARNAME"
>archive_command</TT
> to an empty string
turned off archiving. Now <TT
CLASS="VARNAME"
>archive_mode</TT
> turns archiving
on and off, independently of <TT
CLASS="VARNAME"
>archive_command</TT
>. This is
useful for stopping archiving temporarily.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131808"
>E.121.3.6. Queries</A
></H3
><P
></P
><UL
><LI
><P
> Full text search is integrated into the core database
system (Teodor, Oleg)
</P
><P
> Text search has been improved, moved into the core code, and is now
installed by default. <TT
CLASS="FILENAME"
>contrib/tsearch2</TT
> now contains
a compatibility interface.
</P
></LI
><LI
><P
> Add control over whether <TT
CLASS="LITERAL"
>NULL</TT
>s sort first or last (Teodor, Tom)
</P
><P
> The syntax is <TT
CLASS="LITERAL"
>ORDER BY ... NULLS FIRST/LAST</TT
>.
</P
></LI
><LI
><P
> Allow per-column ascending/descending (<TT
CLASS="LITERAL"
>ASC</TT
>/<TT
CLASS="LITERAL"
>DESC</TT
>)
ordering options for indexes (Teodor, Tom)
</P
><P
> Previously a query using <TT
CLASS="LITERAL"
>ORDER BY</TT
> with mixed
<TT
CLASS="LITERAL"
>ASC</TT
>/<TT
CLASS="LITERAL"
>DESC</TT
> specifiers could not fully use
an index. Now an index can be fully used in such cases if the
index was created with matching
<TT
CLASS="LITERAL"
>ASC</TT
>/<TT
CLASS="LITERAL"
>DESC</TT
> specifications.
<TT
CLASS="LITERAL"
>NULL</TT
> sort order within an index can be controlled, too.
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>col IS NULL</TT
> to use an index (Teodor)
</P
></LI
><LI
><P
> Updatable cursors (Arul Shaji, Tom)
</P
><P
> This eliminates the need to reference a primary key to
<TT
CLASS="COMMAND"
>UPDATE</TT
> or <TT
CLASS="COMMAND"
>DELETE</TT
> rows returned by a cursor.
The syntax is <TT
CLASS="LITERAL"
>UPDATE/DELETE WHERE CURRENT OF</TT
>.
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>FOR UPDATE</TT
> in cursors (Arul Shaji, Tom)
</P
></LI
><LI
><P
> Create a general mechanism that supports casts to and from the
standard string types (<TT
CLASS="TYPE"
>TEXT</TT
>, <TT
CLASS="TYPE"
>VARCHAR</TT
>,
<TT
CLASS="TYPE"
>CHAR</TT
>) for <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>every</I
></SPAN
> datatype, by
invoking the datatype's I/O functions (Tom)
</P
><P
> Previously, such casts were available only for types that had
specialized function(s) for the purpose.
These new casts are assignment-only in the to-string direction,
explicit-only in the other direction, and therefore should create no
surprising behavior.
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>UNION</TT
> and related constructs to return a domain
type, when all inputs are of that domain type (Tom)
</P
><P
> Formerly, the output would be considered to be of the domain's base
type.
</P
></LI
><LI
><P
> Allow limited hashing when using two different data types (Tom)
</P
><P
> This allows hash joins, hash indexes, hashed subplans, and hash
aggregation to be used in situations involving cross-data-type
comparisons, if the data types have compatible hash functions.
Currently, cross-data-type hashing support exists for
<TT
CLASS="TYPE"
>smallint</TT
>/<TT
CLASS="TYPE"
>integer</TT
>/<TT
CLASS="TYPE"
>bigint</TT
>,
and for <TT
CLASS="TYPE"
>float4</TT
>/<TT
CLASS="TYPE"
>float8</TT
>.
</P
></LI
><LI
><P
> Improve optimizer logic for detecting when variables are equal
in a <TT
CLASS="LITERAL"
>WHERE</TT
> clause (Tom)
</P
><P
> This allows mergejoins to work with descending sort orders, and
improves recognition of redundant sort columns.
</P
></LI
><LI
><P
> Improve performance when planning large inheritance trees in
cases where most tables are excluded by constraints (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131868"
>E.121.3.7. Object Manipulation</A
></H3
><P
></P
><UL
><LI
><P
> Arrays of composite types (David Fetter, Andrew, Tom)
</P
><P
> In addition to arrays of explicitly-declared composite types,
arrays of the rowtypes of regular tables and views are now
supported, except for rowtypes of system catalogs, sequences, and TOAST
tables.
</P
></LI
><LI
><P
> Server configuration parameters can now be set on a per-function
basis (Tom)
</P
><P
> For example, functions can now set their own
<TT
CLASS="VARNAME"
>search_path</TT
> to prevent unexpected behavior if a
different <TT
CLASS="VARNAME"
>search_path</TT
> exists at run-time. Security
definer functions should set <TT
CLASS="VARNAME"
>search_path</TT
> to
avoid security loopholes.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>CREATE/ALTER FUNCTION</TT
> now supports
<TT
CLASS="LITERAL"
>COST</TT
> and <TT
CLASS="LITERAL"
>ROWS</TT
> options (Tom)
</P
><P
> <TT
CLASS="LITERAL"
>COST</TT
> allows specification of the cost of a
function call. <TT
CLASS="LITERAL"
>ROWS</TT
> allows specification of
the average number or rows returned by a set-returning function.
These values are used by the optimizer in choosing the best plan.
</P
></LI
><LI
><P
> Implement <TT
CLASS="COMMAND"
>CREATE TABLE LIKE ... INCLUDING
INDEXES</TT
> (Trevor Hardcastle, Nikhil Sontakke, Neil)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>CREATE INDEX CONCURRENTLY</TT
> to ignore
transactions in other databases (Simon)
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>ALTER VIEW ... RENAME TO</TT
> and <TT
CLASS="COMMAND"
>ALTER
SEQUENCE ... RENAME TO</TT
> (David Fetter, Neil)
</P
><P
> Previously this could only be done via <TT
CLASS="COMMAND"
>ALTER TABLE ...
RENAME TO</TT
>.
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>CREATE/DROP/RENAME DATABASE</TT
> wait briefly for
conflicting backends to exit before failing (Tom)
</P
><P
> This increases the likelihood that these commands will succeed.
</P
></LI
><LI
><P
> Allow triggers and rules to be deactivated in groups using a
configuration parameter, for replication purposes (Jan)
</P
><P
> This allows replication systems to disable triggers and rewrite
rules as a group without modifying the system catalogs directly.
The behavior is controlled by <TT
CLASS="COMMAND"
>ALTER TABLE</TT
> and a new
parameter <TT
CLASS="VARNAME"
>session_replication_role</TT
>.
</P
></LI
><LI
><P
> User-defined types can now have type modifiers (Teodor, Tom)
</P
><P
> This allows a user-defined type to take a modifier, like
<TT
CLASS="TYPE"
>ssnum(7)</TT
>. Previously only built-in
data types could have modifiers.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131913"
>E.121.3.8. Utility Commands</A
></H3
><P
></P
><UL
><LI
><P
> Non-superuser database owners now are able to add trusted procedural
languages to their databases by default (Jeremy Drake)
</P
><P
> While this is reasonably safe, some administrators might wish to
revoke the privilege. It is controlled by
<TT
CLASS="STRUCTNAME"
>pg_pltemplate</TT
>.<TT
CLASS="STRUCTFIELD"
>tmpldbacreate</TT
>.
</P
></LI
><LI
><P
> Allow a session's current parameter setting to be used as the
default for future sessions (Tom)
</P
><P
> This is done with <TT
CLASS="LITERAL"
>SET ... FROM CURRENT</TT
> in
<TT
CLASS="COMMAND"
>CREATE/ALTER FUNCTION</TT
>, <TT
CLASS="COMMAND"
>ALTER
DATABASE</TT
>, or <TT
CLASS="COMMAND"
>ALTER ROLE</TT
>.
</P
></LI
><LI
><P
> Implement new commands <TT
CLASS="COMMAND"
>DISCARD ALL</TT
>,
<TT
CLASS="COMMAND"
>DISCARD PLANS</TT
>, <TT
CLASS="COMMAND"
>DISCARD
TEMPORARY</TT
>, <TT
CLASS="COMMAND"
>CLOSE ALL</TT
>, and
<TT
CLASS="COMMAND"
>DEALLOCATE ALL</TT
> (Marko Kreen, Neil)
</P
><P
> These commands simplify resetting a database session to its initial
state, and are particularly useful for connection-pooling software.
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>CLUSTER</TT
> MVCC-safe (Heikki Linnakangas)
</P
><P
> Formerly, <TT
CLASS="COMMAND"
>CLUSTER</TT
> would discard all tuples
that were committed dead, even if there were still transactions
that should be able to see them under MVCC visibility rules.
</P
></LI
><LI
><P
> Add new <TT
CLASS="COMMAND"
>CLUSTER</TT
> syntax: <TT
CLASS="LITERAL"
>CLUSTER
<TT
CLASS="REPLACEABLE"
><I
>table</I
></TT
> USING <TT
CLASS="REPLACEABLE"
><I
>index</I
></TT
></TT
>
(Holger Schurig)
</P
><P
> The old <TT
CLASS="COMMAND"
>CLUSTER</TT
> syntax is still supported, but
the new form is considered more logical.
</P
></LI
><LI
><P
> Fix <TT
CLASS="COMMAND"
>EXPLAIN</TT
> so it can show complex plans
more accurately (Tom)
</P
><P
> References to subplan outputs are now always shown correctly,
instead of using <TT
CLASS="LITERAL"
>?column<TT
CLASS="REPLACEABLE"
><I
>N</I
></TT
>?</TT
>
for complicated cases.
</P
></LI
><LI
><P
> Limit the amount of information reported when a user is dropped
(Alvaro)
</P
><P
> Previously, dropping (or attempting to drop) a user who owned many
objects could result in large <TT
CLASS="LITERAL"
>NOTICE</TT
> or
<TT
CLASS="LITERAL"
>ERROR</TT
> messages listing all these objects; this
caused problems for some client applications. The length of the
message is now limited, although a full list is still sent to the
server log.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131960"
>E.121.3.9. Data Types</A
></H3
><P
></P
><UL
><LI
><P
> Support for the SQL/XML standard, including new operators and an
<TT
CLASS="TYPE"
>XML</TT
> data type (Nikolay Samokhvalov, Pavel Stehule, Peter)
</P
></LI
><LI
><P
> Enumerated data types (<TT
CLASS="TYPE"
>ENUM</TT
>) (Tom Dunstan)
</P
><P
> This feature provides convenient support for fields that have a
small, fixed set of allowed values. An example of creating an
<TT
CLASS="LITERAL"
>ENUM</TT
> type is
<TT
CLASS="LITERAL"
>CREATE TYPE mood AS ENUM ('sad', 'ok', 'happy')</TT
>.
</P
></LI
><LI
><P
> Universally Unique Identifier (<TT
CLASS="TYPE"
>UUID</TT
>) data type (Gevik
Babakhani, Neil)
</P
><P
> This closely matches <ACRONYM
CLASS="ACRONYM"
>RFC</ACRONYM
> 4122.
</P
></LI
><LI
><P
> Widen the <TT
CLASS="TYPE"
>MONEY</TT
> data type to 64 bits (D'Arcy Cain)
</P
><P
> This greatly increases the range of supported <TT
CLASS="TYPE"
>MONEY</TT
>
values.
</P
></LI
><LI
><P
> Fix <TT
CLASS="TYPE"
>float4</TT
>/<TT
CLASS="TYPE"
>float8</TT
> to handle
<TT
CLASS="LITERAL"
>Infinity</TT
> and <TT
CLASS="LITERAL"
>NAN</TT
> (Not A Number)
consistently (Bruce)
</P
><P
> The code formerly was not consistent about distinguishing
<TT
CLASS="LITERAL"
>Infinity</TT
> from overflow conditions.
</P
></LI
><LI
><P
> Allow leading and trailing whitespace during input of
<TT
CLASS="TYPE"
>boolean</TT
> values (Neil)
</P
></LI
><LI
><P
> Prevent <TT
CLASS="COMMAND"
>COPY</TT
> from using digits and lowercase letters as
delimiters (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN131996"
>E.121.3.10. Functions</A
></H3
><P
></P
><UL
><LI
><P
> Add new regular expression functions
<CODE
CLASS="FUNCTION"
>regexp_matches()</CODE
>,
<CODE
CLASS="FUNCTION"
>regexp_split_to_array()</CODE
>, and
<CODE
CLASS="FUNCTION"
>regexp_split_to_table()</CODE
> (Jeremy Drake, Neil)
</P
><P
> These functions provide extraction of regular expression
subexpressions and allow splitting a string using a POSIX regular
expression.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>lo_truncate()</CODE
> for large object truncation
(Kris Jurka)
</P
></LI
><LI
><P
> Implement <CODE
CLASS="FUNCTION"
>width_bucket()</CODE
> for the <TT
CLASS="TYPE"
>float8</TT
>
data type (Neil)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_stat_clear_snapshot()</CODE
> to discard
statistics snapshots collected during the current transaction
(Tom)
</P
><P
> The first request for statistics in a transaction takes a statistics
snapshot that does not change during the transaction. This function
allows the snapshot to be discarded and a new snapshot loaded during
the next statistics query. This is particularly useful for PL/pgSQL
functions, which are confined to a single transaction.
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>isodow</TT
> option to <CODE
CLASS="FUNCTION"
>EXTRACT()</CODE
> and
<CODE
CLASS="FUNCTION"
>date_part()</CODE
> (Bruce)
</P
><P
> This returns the day of the week, with Sunday as seven.
(<TT
CLASS="LITERAL"
>dow</TT
> returns Sunday as zero.)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>ID</TT
> (ISO day of week) and <TT
CLASS="LITERAL"
>IDDD</TT
> (ISO
day of year) format codes for <CODE
CLASS="FUNCTION"
>to_char()</CODE
>,
<CODE
CLASS="FUNCTION"
>to_date()</CODE
>, and <CODE
CLASS="FUNCTION"
>to_timestamp()</CODE
> (Brendan
Jurd)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>to_timestamp()</CODE
> and <CODE
CLASS="FUNCTION"
>to_date()</CODE
>
assume <TT
CLASS="LITERAL"
>TM</TT
> (trim) option for potentially
variable-width fields (Bruce)
</P
><P
> This matches <SPAN
CLASS="PRODUCTNAME"
>Oracle</SPAN
>'s behavior.
</P
></LI
><LI
><P
> Fix off-by-one conversion error in
<CODE
CLASS="FUNCTION"
>to_date()</CODE
>/<CODE
CLASS="FUNCTION"
>to_timestamp()</CODE
>
<TT
CLASS="LITERAL"
>D</TT
> (non-ISO day of week) fields (Bruce)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>setseed()</CODE
> return void, rather than a
useless integer value (Neil)
</P
></LI
><LI
><P
> Add a hash function for <TT
CLASS="TYPE"
>NUMERIC</TT
> (Neil)
</P
><P
> This allows hash indexes and hash-based plans to be used with
<TT
CLASS="TYPE"
>NUMERIC</TT
> columns.
</P
></LI
><LI
><P
> Improve efficiency of
<TT
CLASS="LITERAL"
>LIKE</TT
>/<TT
CLASS="LITERAL"
>ILIKE</TT
>, especially for
multi-byte character sets like UTF-8 (Andrew, Itagaki Takahiro)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>currtid()</CODE
> functions require
<TT
CLASS="LITERAL"
>SELECT</TT
> privileges on the target table (Tom)
</P
></LI
><LI
><P
> Add several <CODE
CLASS="FUNCTION"
>txid_*()</CODE
> functions to query
active transaction IDs (Jan)
</P
><P
> This is useful for various replication solutions.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132062"
>E.121.3.11. PL/pgSQL Server-Side Language</A
></H3
><P
></P
><UL
><LI
><P
> Add scrollable cursor support, including directional control in
<TT
CLASS="COMMAND"
>FETCH</TT
> (Pavel Stehule)
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>IN</TT
> as an alternative to
<TT
CLASS="LITERAL"
>FROM</TT
> in PL/pgSQL's <TT
CLASS="COMMAND"
>FETCH</TT
>
statement, for consistency with the backend's
<TT
CLASS="COMMAND"
>FETCH</TT
> command (Pavel Stehule)
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>MOVE</TT
> to PL/pgSQL (Magnus, Pavel Stehule,
Neil)
</P
></LI
><LI
><P
> Implement <TT
CLASS="COMMAND"
>RETURN QUERY</TT
> (Pavel Stehule, Neil)
</P
><P
> This adds convenient syntax for PL/pgSQL set-returning functions
that want to return the result of a query. <TT
CLASS="COMMAND"
>RETURN QUERY</TT
>
is easier and more efficient than a loop
around <TT
CLASS="COMMAND"
>RETURN NEXT</TT
>.
</P
></LI
><LI
><P
> Allow function parameter names to be qualified with the
function's name (Tom)
</P
><P
> For example, <TT
CLASS="LITERAL"
>myfunc.myvar</TT
>. This is particularly
useful for specifying variables in a query where the variable
name might match a column name.
</P
></LI
><LI
><P
> Make qualification of variables with block labels work properly (Tom)
</P
><P
> Formerly, outer-level block labels could unexpectedly interfere with
recognition of inner-level record or row references.
</P
></LI
><LI
><P
> Tighten requirements for <TT
CLASS="LITERAL"
>FOR</TT
> loop
<TT
CLASS="LITERAL"
>STEP</TT
> values (Tom)
</P
><P
> Prevent non-positive <TT
CLASS="LITERAL"
>STEP</TT
> values, and handle
loop overflows.
</P
></LI
><LI
><P
> Improve accuracy when reporting syntax error locations (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132098"
>E.121.3.12. Other Server-Side Languages</A
></H3
><P
></P
><UL
><LI
><P
> Allow type-name arguments to PL/Perl
<CODE
CLASS="FUNCTION"
>spi_prepare()</CODE
> to be data type aliases in
addition to names found in <TT
CLASS="LITERAL"
>pg_type</TT
> (Andrew)
</P
></LI
><LI
><P
> Allow type-name arguments to PL/Python
<CODE
CLASS="FUNCTION"
>plpy.prepare()</CODE
> to be data type aliases in
addition to names found in <TT
CLASS="LITERAL"
>pg_type</TT
> (Andrew)
</P
></LI
><LI
><P
> Allow type-name arguments to PL/Tcl <CODE
CLASS="FUNCTION"
>spi_prepare</CODE
> to
be data type aliases in addition to names found in
<TT
CLASS="LITERAL"
>pg_type</TT
> (Andrew)
</P
></LI
><LI
><P
> Enable PL/PythonU to compile on Python 2.5 (Marko Kreen)
</P
></LI
><LI
><P
> Support a true PL/Python boolean type in compatible Python versions
(Python 2.3 and later) (Marko Kreen)
</P
></LI
><LI
><P
> Fix PL/Tcl problems with thread-enabled <TT
CLASS="FILENAME"
>libtcl</TT
> spawning
multiple threads within the backend (Steve Marshall, Paul Bayer,
Doug Knight)
</P
><P
> This caused all sorts of unpleasantness.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132121"
>E.121.3.13. <A
HREF="app-psql.html"
><SPAN
CLASS="APPLICATION"
>psql</SPAN
></A
></A
></H3
><P
></P
><UL
><LI
><P
> List disabled triggers separately in <TT
CLASS="LITERAL"
>\d</TT
> output
(Brendan Jurd)
</P
></LI
><LI
><P
> In <TT
CLASS="LITERAL"
>\d</TT
> patterns, always match <TT
CLASS="LITERAL"
>$</TT
>
literally (Tom)
</P
></LI
><LI
><P
> Show aggregate return types in <TT
CLASS="LITERAL"
>\da</TT
> output
(Greg Sabino Mullane)
</P
></LI
><LI
><P
> Add the function's volatility status to the output of
<TT
CLASS="LITERAL"
>\df+</TT
> (Neil)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>\prompt</TT
> capability (Chad Wagner)
</P
></LI
><LI
><P
> Allow <TT
CLASS="LITERAL"
>\pset</TT
>, <TT
CLASS="LITERAL"
>\t</TT
>, and
<TT
CLASS="LITERAL"
>\x</TT
> to specify <TT
CLASS="LITERAL"
>on</TT
> or <TT
CLASS="LITERAL"
>off</TT
>,
rather than just toggling (Chad Wagner)
</P
></LI
><LI
><P
> Add <TT
CLASS="LITERAL"
>\sleep</TT
> capability (Jan)
</P
></LI
><LI
><P
> Enable <TT
CLASS="LITERAL"
>\timing</TT
> output for <TT
CLASS="LITERAL"
>\copy</TT
> (Andrew)
</P
></LI
><LI
><P
> Improve <TT
CLASS="LITERAL"
>\timing</TT
> resolution on Windows
(Itagaki Takahiro)
</P
></LI
><LI
><P
> Flush <TT
CLASS="LITERAL"
>\o</TT
> output after each backslash command (Tom)
</P
></LI
><LI
><P
> Correctly detect and report errors while reading a <TT
CLASS="LITERAL"
>-f</TT
>
input file (Peter)
</P
></LI
><LI
><P
> Remove <TT
CLASS="LITERAL"
>-u</TT
> option (this option has long been deprecated)
(Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132168"
>E.121.3.14. <A
HREF="app-pgdump.html"
><SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
></A
></A
></H3
><P
></P
><UL
><LI
><P
> Add <TT
CLASS="LITERAL"
>--tablespaces-only</TT
> and <TT
CLASS="LITERAL"
>--roles-only</TT
>
options to <SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> (Dave Page)
</P
></LI
><LI
><P
> Add an output file option to
<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> (Dave Page)
</P
><P
> This is primarily useful on Windows, where output redirection of
child <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> processes does not work.
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> to accept an initial-connection
database name rather than the default
<TT
CLASS="LITERAL"
>template1</TT
> (Dave Page)
</P
></LI
><LI
><P
> In <TT
CLASS="LITERAL"
>-n</TT
> and <TT
CLASS="LITERAL"
>-t</TT
> switches, always match
<TT
CLASS="LITERAL"
>$</TT
> literally (Tom)
</P
></LI
><LI
><P
> Improve performance when a database has thousands of objects (Tom)
</P
></LI
><LI
><P
> Remove <TT
CLASS="LITERAL"
>-u</TT
> option (this option has long been deprecated)
(Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132197"
>E.121.3.15. Other Client Applications</A
></H3
><P
></P
><UL
><LI
><P
> In <SPAN
CLASS="APPLICATION"
>initdb</SPAN
>, allow the location of the
<TT
CLASS="FILENAME"
>pg_xlog</TT
> directory to be specified
(Euler Taveira de Oliveira)
</P
></LI
><LI
><P
> Enable server core dump generation in <SPAN
CLASS="APPLICATION"
>pg_regress</SPAN
>
on supported operating systems (Andrew)
</P
></LI
><LI
><P
> Add a <TT
CLASS="LITERAL"
>-t</TT
> (timeout) parameter to <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
>
(Bruce)
</P
><P
> This controls how long <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
> will wait when waiting
for server startup or shutdown. Formerly the timeout was hard-wired
as 60 seconds.
</P
></LI
><LI
><P
> Add a <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
> option to control generation
of server core dumps (Andrew)
</P
></LI
><LI
><P
> Allow Control-C to cancel <SPAN
CLASS="APPLICATION"
>clusterdb</SPAN
>,
<SPAN
CLASS="APPLICATION"
>reindexdb</SPAN
>, and <SPAN
CLASS="APPLICATION"
>vacuumdb</SPAN
> (Itagaki
Takahiro, Magnus)
</P
></LI
><LI
><P
> Suppress command tag output for <SPAN
CLASS="APPLICATION"
>createdb</SPAN
>,
<SPAN
CLASS="APPLICATION"
>createuser</SPAN
>, <SPAN
CLASS="APPLICATION"
>dropdb</SPAN
>, and
<SPAN
CLASS="APPLICATION"
>dropuser</SPAN
> (Peter)
</P
><P
> The <TT
CLASS="LITERAL"
>--quiet</TT
> option is ignored and will be removed in 8.4.
Progress messages when acting on all databases now go to stdout
instead of stderr because they are not actually errors.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132229"
>E.121.3.16. <A
HREF="libpq.html"
><SPAN
CLASS="APPLICATION"
>libpq</SPAN
></A
></A
></H3
><P
></P
><UL
><LI
><P
> Interpret the <TT
CLASS="LITERAL"
>dbName</TT
> parameter of
<CODE
CLASS="FUNCTION"
>PQsetdbLogin()</CODE
> as a <TT
CLASS="LITERAL"
>conninfo</TT
> string if
it contains an equals sign (Andrew)
</P
><P
> This allows use of <TT
CLASS="LITERAL"
>conninfo</TT
> strings in client
programs that still use <TT
CLASS="LITERAL"
>PQsetdbLogin()</TT
>.
</P
></LI
><LI
><P
> Support a global <ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> configuration file (Victor
Wagner)
</P
></LI
><LI
><P
> Add environment variable <TT
CLASS="VARNAME"
>PGSSLKEY</TT
> to control
<ACRONYM
CLASS="ACRONYM"
>SSL</ACRONYM
> hardware keys (Victor Wagner)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>lo_truncate()</CODE
> for large object
truncation (Kris Jurka)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>PQconnectionNeedsPassword()</CODE
> that returns
true if the server required a password but none was supplied
(Joe Conway, Tom)
</P
><P
> If this returns true after a failed connection attempt, a client
application should prompt the user for a password. In the past
applications have had to check for a specific error message string to
decide whether a password is needed; that approach is now
deprecated.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>PQconnectionUsedPassword()</CODE
> that returns
true if the supplied password was actually used
(Joe Conway, Tom)
</P
><P
> This is useful in some security contexts where it is important
to know whether a user-supplied password is actually valid.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132260"
>E.121.3.17. <A
HREF="ecpg.html"
><SPAN
CLASS="APPLICATION"
>ecpg</SPAN
></A
></A
></H3
><P
></P
><UL
><LI
><P
> Use V3 frontend/backend protocol (Michael)
</P
><P
> This adds support for server-side prepared statements.
</P
></LI
><LI
><P
> Use native threads, instead of pthreads, on Windows (Magnus)
</P
></LI
><LI
><P
> Improve thread-safety of ecpglib (Itagaki Takahiro)
</P
></LI
><LI
><P
> Make the ecpg libraries export only necessary API symbols (Michael)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132274"
>E.121.3.18. <SPAN
CLASS="APPLICATION"
>Windows</SPAN
> Port</A
></H3
><P
></P
><UL
><LI
><P
> Allow the whole <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> distribution to be compiled
with <SPAN
CLASS="PRODUCTNAME"
>Microsoft Visual C++</SPAN
> (Magnus and others)
</P
><P
> This allows Windows-based developers to use familiar development
and debugging tools.
Windows executables made with Visual C++ might also have better
stability and performance than those made with other tool sets.
The client-only Visual C++ build scripts have been removed.
</P
></LI
><LI
><P
> Drastically reduce postmaster's memory usage when it has many child
processes (Magnus)
</P
></LI
><LI
><P
> Allow regression tests to be started by an administrative
user (Magnus)
</P
></LI
><LI
><P
> Add native shared memory implementation (Magnus)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132289"
>E.121.3.19. Server Programming Interface (<ACRONYM
CLASS="ACRONYM"
>SPI</ACRONYM
>)</A
></H3
><P
></P
><UL
><LI
><P
> Add cursor-related functionality in SPI (Pavel Stehule)
</P
><P
> Allow access to the cursor-related planning options, and add
<TT
CLASS="COMMAND"
>FETCH</TT
>/<TT
CLASS="COMMAND"
>MOVE</TT
> routines.
</P
></LI
><LI
><P
> Allow execution of cursor commands through
<CODE
CLASS="FUNCTION"
>SPI_execute</CODE
> (Tom)
</P
><P
> The macro <TT
CLASS="LITERAL"
>SPI_ERROR_CURSOR</TT
> still exists but will
never be returned.
</P
></LI
><LI
><P
> SPI plan pointers are now declared as <TT
CLASS="LITERAL"
>SPIPlanPtr</TT
> instead of
<TT
CLASS="LITERAL"
>void *</TT
> (Tom)
</P
><P
> This does not break application code, but switching is
recommended to help catch simple programming mistakes.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132308"
>E.121.3.20. Build Options</A
></H3
><P
></P
><UL
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>configure</SPAN
> option <TT
CLASS="LITERAL"
>--enable-profiling</TT
>
to enable code profiling (works only with <SPAN
CLASS="APPLICATION"
>gcc</SPAN
>)
(Korry Douglas and Nikhil Sontakke)
</P
></LI
><LI
><P
> Add <SPAN
CLASS="APPLICATION"
>configure</SPAN
> option <TT
CLASS="LITERAL"
>--with-system-tzdata</TT
>
to use the operating system's time zone database (Peter)
</P
></LI
><LI
><P
> Fix <ACRONYM
CLASS="ACRONYM"
>PGXS</ACRONYM
> so extensions can be built against PostgreSQL
installations whose <SPAN
CLASS="APPLICATION"
>pg_config</SPAN
> program does not
appear first in the <TT
CLASS="VARNAME"
>PATH</TT
> (Tom)
</P
></LI
><LI
><P
> Support <TT
CLASS="COMMAND"
>gmake draft</TT
> when building the
<ACRONYM
CLASS="ACRONYM"
>SGML</ACRONYM
> documentation (Bruce)
</P
><P
> Unless <TT
CLASS="LITERAL"
>draft</TT
> is used, the documentation build will
now be repeated if necessary to ensure the index is up-to-date.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132331"
>E.121.3.21. Source Code</A
></H3
><P
></P
><UL
><LI
><P
> Rename macro <TT
CLASS="LITERAL"
>DLLIMPORT</TT
> to <TT
CLASS="LITERAL"
>PGDLLIMPORT</TT
> to
avoid conflicting with third party includes (like Tcl) that
define <TT
CLASS="LITERAL"
>DLLIMPORT</TT
> (Magnus)
</P
></LI
><LI
><P
> Create <SPAN
CLASS="QUOTE"
>"operator families"</SPAN
> to improve planning of
queries involving cross-data-type comparisons (Tom)
</P
></LI
><LI
><P
> Update GIN <CODE
CLASS="FUNCTION"
>extractQuery()</CODE
> API to allow signalling
that nothing can satisfy the query (Teodor)
</P
></LI
><LI
><P
> Move <TT
CLASS="LITERAL"
>NAMEDATALEN</TT
> definition from
<TT
CLASS="FILENAME"
>postgres_ext.h</TT
> to <TT
CLASS="FILENAME"
>pg_config_manual.h</TT
>
(Peter)
</P
></LI
><LI
><P
> Provide <CODE
CLASS="FUNCTION"
>strlcpy()</CODE
> and
<CODE
CLASS="FUNCTION"
>strlcat()</CODE
> on all platforms, and replace
error-prone uses of <CODE
CLASS="FUNCTION"
>strncpy()</CODE
>,
<CODE
CLASS="FUNCTION"
>strncat()</CODE
>, etc (Peter)
</P
></LI
><LI
><P
> Create hooks to let an external plugin monitor (or even replace) the
planner and create plans for hypothetical situations (Gurjeet
Singh, Tom)
</P
></LI
><LI
><P
> Create a function variable <TT
CLASS="LITERAL"
>join_search_hook</TT
> to let plugins
override the join search order portion of the planner (Julius
Stroffek)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>tas()</CODE
> support for Renesas' M32R processor
(Kazuhiro Inaoka)
</P
></LI
><LI
><P
> <CODE
CLASS="FUNCTION"
>quote_identifier()</CODE
> and
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> no longer quote keywords that are
unreserved according to the grammar (Tom)
</P
></LI
><LI
><P
> Change the on-disk representation of the <TT
CLASS="TYPE"
>NUMERIC</TT
>
data type so that the <TT
CLASS="STRUCTFIELD"
>sign_dscale</TT
> word comes
before the weight (Tom)
</P
></LI
><LI
><P
> Use <ACRONYM
CLASS="ACRONYM"
>SYSV</ACRONYM
> semaphores rather than POSIX on Darwin
>= 6.0, i.e., OS X 10.2 and up (Chris Marcellino)
</P
></LI
><LI
><P
> Add <A
HREF="acronyms.html"
>acronym</A
> and <A
HREF="creating-cluster.html#CREATING-CLUSTER-NFS"
>NFS</A
> documentation
sections (Bruce)
</P
></LI
><LI
><P
> "Postgres" is now documented as an accepted alias for
"PostgreSQL" (Peter)
</P
></LI
><LI
><P
> Add documentation about preventing database server spoofing when
the server is down (Bruce)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN132383"
>E.121.3.22. Contrib</A
></H3
><P
></P
><UL
><LI
><P
> Move <TT
CLASS="FILENAME"
>contrib</TT
> <TT
CLASS="FILENAME"
>README</TT
> content into the
main <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> documentation (Albert Cervera i
Areny)
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/pageinspect</TT
> module for low-level
page inspection (Simon, Heikki)
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/pg_standby</TT
> module for controlling
warm standby operation (Simon)
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/uuid-ossp</TT
> module for generating
<TT
CLASS="TYPE"
>UUID</TT
> values using the OSSP UUID library (Peter)
</P
><P
> Use <SPAN
CLASS="APPLICATION"
>configure</SPAN
>
<TT
CLASS="LITERAL"
>--with-ossp-uuid</TT
> to activate. This takes
advantage of the new <TT
CLASS="TYPE"
>UUID</TT
> builtin type.
</P
></LI
><LI
><P
> Add <TT
CLASS="FILENAME"
>contrib/dict_int</TT
>,
<TT
CLASS="FILENAME"
>contrib/dict_xsyn</TT
>, and
<TT
CLASS="FILENAME"
>contrib/test_parser</TT
> modules to provide
sample add-on text search dictionary templates and parsers
(Sergey Karpov)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>contrib/pgbench</SPAN
> to set the fillfactor (Pavan
Deolasee)
</P
></LI
><LI
><P
> Add timestamps to <SPAN
CLASS="APPLICATION"
>contrib/pgbench</SPAN
> <TT
CLASS="LITERAL"
>-l</TT
>
(Greg Smith)
</P
></LI
><LI
><P
> Add usage count statistics to
<TT
CLASS="FILENAME"
>contrib/pgbuffercache</TT
> (Greg Smith)
</P
></LI
><LI
><P
> Add GIN support for <TT
CLASS="FILENAME"
>contrib/hstore</TT
> (Teodor)
</P
></LI
><LI
><P
> Add GIN support for <TT
CLASS="FILENAME"
>contrib/pg_trgm</TT
> (Guillaume Smet, Teodor)
</P
></LI
><LI
><P
> Update OS/X startup scripts in
<TT
CLASS="FILENAME"
>contrib/start-scripts</TT
> (Mark Cotner, David
Fetter)
</P
></LI
><LI
><P
> Restrict <CODE
CLASS="FUNCTION"
>pgrowlocks()</CODE
> and
<CODE
CLASS="FUNCTION"
>dblink_get_pkey()</CODE
> to users who have
<TT
CLASS="LITERAL"
>SELECT</TT
> privilege on the target table (Tom)
</P
></LI
><LI
><P
> Restrict <TT
CLASS="FILENAME"
>contrib/pgstattuple</TT
> functions to
superusers (Tom)
</P
></LI
><LI
><P
> <TT
CLASS="FILENAME"
>contrib/xml2</TT
> is deprecated and planned for
removal in 8.4 (Peter)
</P
><P
> The new XML support in core PostgreSQL supersedes this module.
</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-8-3-1.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-8-2-23.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Release 8.3.1</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 8.2.23</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>