Current File : //opt/alt/postgresql11/usr/share/doc/alt-postgresql11-9.2.24/html/release-8-0.html
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<HTML
><HEAD
><TITLE
>Release 8.0</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 8.0.1"
HREF="release-8-0-1.html"><LINK
REL="NEXT"
TITLE="Release 7.4.30"
HREF="release-7-4-30.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 8.0.1"
HREF="release-8-0-1.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 7.4.30"
HREF="release-7-4-30.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="SECT1"
><H1
CLASS="SECT1"
><A
NAME="RELEASE-8-0"
>E.196. Release 8.0</A
></H1
><DIV
CLASS="FORMALPARA"
><P
><B
>Release date: </B
>2005-01-19</P
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN139436"
>E.196.1. Overview</A
></H2
><P
> Major changes in this release:
</P
><P
></P
><DIV
CLASS="VARIABLELIST"
><DL
><DT
>Microsoft Windows Native Server</DT
><DD
><P
> This is the first <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> release
to run natively on <SPAN
CLASS="TRADEMARK"
>Microsoft Windows</SPAN
>® as
a server. It can run as a <SPAN
CLASS="PRODUCTNAME"
>Windows</SPAN
> service. This
release supports NT-based Windows releases like
<SPAN
CLASS="PRODUCTNAME"
>Windows 2000 SP4</SPAN
>, <SPAN
CLASS="PRODUCTNAME"
>Windows XP</SPAN
>, and
<SPAN
CLASS="PRODUCTNAME"
>Windows 2003</SPAN
>. Older releases like
<SPAN
CLASS="PRODUCTNAME"
>Windows 95</SPAN
>, <SPAN
CLASS="PRODUCTNAME"
>Windows 98</SPAN
>, and
<SPAN
CLASS="PRODUCTNAME"
>Windows ME</SPAN
> are not supported because these operating
systems do not have the infrastructure to support
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>. A separate installer
project has been created to ease installation on
<SPAN
CLASS="PRODUCTNAME"
>Windows</SPAN
> — see <A
HREF="http://www.postgresql.org/ftp/win32/"
TARGET="_top"
>http://www.postgresql.org/ftp/win32/</A
>.
</P
><P
> Although tested throughout our release cycle, the Windows port
does not have the benefit of years of use in production
environments that <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> has on
Unix platforms. Therefore it should be treated with the same
level of caution as you would a new product.
</P
><P
> Previous releases required the Unix emulation toolkit
<SPAN
CLASS="PRODUCTNAME"
>Cygwin</SPAN
> in order to run the server on Windows
operating systems. <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> has
supported native clients on Windows for many years.
</P
></DD
><DT
>Savepoints</DT
><DD
><P
> Savepoints allow specific parts of a transaction to be aborted
without affecting the remainder of the transaction. Prior
releases had no such capability; there was no way to recover
from a statement failure within a transaction except by
aborting the whole transaction. This feature is valuable for
application writers who require error recovery within a
complex transaction.
</P
></DD
><DT
>Point-In-Time Recovery</DT
><DD
><P
> In previous releases there was no way to recover from disk
drive failure except to restore from a previous backup or use
a standby replication server. Point-in-time recovery allows
continuous backup of the server. You can recover either to
the point of failure or to some transaction in the past.
</P
></DD
><DT
>Tablespaces</DT
><DD
><P
> Tablespaces allow administrators to select different file systems
for storage of individual tables, indexes, and databases.
This improves performance and control over disk space
usage. Prior releases used <SPAN
CLASS="APPLICATION"
>initlocation</SPAN
> and
manual symlink management for such tasks.
</P
></DD
><DT
>Improved Buffer Management, <TT
CLASS="COMMAND"
>CHECKPOINT</TT
>,
<TT
CLASS="COMMAND"
>VACUUM</TT
></DT
><DD
><P
> This release has a more intelligent buffer replacement strategy,
which will make better use of available shared buffers and
improve performance. The performance impact of vacuum and
checkpoints is also lessened.
</P
></DD
><DT
>Change Column Types</DT
><DD
><P
> A column's data type can now be changed with <TT
CLASS="COMMAND"
>ALTER
TABLE</TT
>.
</P
></DD
><DT
>New Perl Server-Side Language</DT
><DD
><P
> A new version of the <SPAN
CLASS="APPLICATION"
>plperl</SPAN
> server-side language now
supports a persistent shared storage area, triggers, returning records
and arrays of records, and SPI calls to access the database.
</P
></DD
><DT
>Comma-separated-value (CSV) support in <TT
CLASS="COMMAND"
>COPY</TT
></DT
><DD
><P
> <TT
CLASS="COMMAND"
>COPY</TT
> can now read and write
comma-separated-value files. It has the flexibility to
interpret nonstandard quoting and separation characters too.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN139496"
>E.196.2. Migration to Version 8.0</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
><P
></P
><UL
><LI
><P
> In <TT
CLASS="OPTION"
>READ COMMITTED</TT
> serialization mode, volatile functions
now see the results of concurrent transactions committed up to the
beginning of each statement within the function, rather than up to the
beginning of the interactive command that called the function.
</P
></LI
><LI
><P
> Functions declared <TT
CLASS="OPTION"
>STABLE</TT
> or <TT
CLASS="OPTION"
>IMMUTABLE</TT
> always
use the snapshot of the calling query, and therefore do not see the
effects of actions taken after the calling query starts, whether in
their own transaction or other transactions. Such a function must be
read-only, too, meaning that it cannot use any SQL commands other than
<TT
CLASS="COMMAND"
>SELECT</TT
>.
</P
></LI
><LI
><P
> Nondeferred <TT
CLASS="OPTION"
>AFTER</TT
> triggers are now fired immediately
after completion of the triggering query, rather than upon
finishing the current interactive command. This makes a
difference when the triggering query occurred within a function:
the trigger is invoked before the function proceeds to its next
operation.
</P
></LI
><LI
><P
> Server configuration parameters <TT
CLASS="VARNAME"
>virtual_host</TT
> and
<TT
CLASS="VARNAME"
>tcpip_socket</TT
> have been replaced with a more general
parameter <TT
CLASS="VARNAME"
>listen_addresses</TT
>. Also, the server now listens on
<TT
CLASS="LITERAL"
>localhost</TT
> by default, which eliminates the need for the
<TT
CLASS="LITERAL"
>-i</TT
> postmaster switch in many scenarios.
</P
></LI
><LI
><P
> Server configuration parameters <TT
CLASS="VARNAME"
>SortMem</TT
> and
<TT
CLASS="VARNAME"
>VacuumMem</TT
> have been renamed to <TT
CLASS="VARNAME"
>work_mem</TT
>
and <TT
CLASS="VARNAME"
>maintenance_work_mem</TT
> to better reflect their
use. The original names are still supported in
<TT
CLASS="COMMAND"
>SET</TT
> and <TT
CLASS="COMMAND"
>SHOW</TT
>.
</P
></LI
><LI
><P
> Server configuration parameters <TT
CLASS="VARNAME"
>log_pid</TT
>,
<TT
CLASS="VARNAME"
>log_timestamp</TT
>, and <TT
CLASS="VARNAME"
>log_source_port</TT
> have been
replaced with a more general parameter <TT
CLASS="VARNAME"
>log_line_prefix</TT
>.
</P
></LI
><LI
><P
> Server configuration parameter <TT
CLASS="VARNAME"
>syslog</TT
> has been
replaced with a more logical <TT
CLASS="VARNAME"
>log_destination</TT
> variable to
control the log output destination.
</P
></LI
><LI
><P
> Server configuration parameter <TT
CLASS="VARNAME"
>log_statement</TT
> has been
changed so it can selectively log just database modification or
data definition statements. Server configuration parameter
<TT
CLASS="VARNAME"
>log_duration</TT
> now prints only when <TT
CLASS="VARNAME"
>log_statement</TT
>
prints the query.
</P
></LI
><LI
><P
> Server configuration parameter <TT
CLASS="VARNAME"
>max_expr_depth</TT
> parameter has
been replaced with <TT
CLASS="VARNAME"
>max_stack_depth</TT
> which measures the
physical stack size rather than the expression nesting depth. This
helps prevent session termination due to stack overflow caused by
recursive functions.
</P
></LI
><LI
><P
> The <CODE
CLASS="FUNCTION"
>length()</CODE
> function no longer counts trailing spaces in
<TT
CLASS="TYPE"
>CHAR(n)</TT
> values.
</P
></LI
><LI
><P
> Casting an integer to <TT
CLASS="TYPE"
>BIT(N)</TT
> selects the rightmost N bits of the
integer, not the leftmost N bits as before.
</P
></LI
><LI
><P
> Updating an element or slice of a NULL array value now produces
a nonnull array result, namely an array containing
just the assigned-to positions.
</P
></LI
><LI
><P
> Syntax checking of array input values has been tightened up
considerably. Junk that was previously allowed in odd places with
odd results now causes an error. Empty-string element values
must now be written as <TT
CLASS="LITERAL"
>""</TT
>, rather than writing nothing.
Also changed behavior with respect to whitespace surrounding
array elements: trailing whitespace is now ignored, for symmetry
with leading whitespace (which has always been ignored).
</P
></LI
><LI
><P
> Overflow in integer arithmetic operations is now detected and
reported as an error.
</P
></LI
><LI
><P
> The arithmetic operators associated with the single-byte
<TT
CLASS="TYPE"
>"char"</TT
> data type have been removed.
</P
></LI
><LI
><P
> The <CODE
CLASS="FUNCTION"
>extract()</CODE
> function (also called
<CODE
CLASS="FUNCTION"
>date_part</CODE
>) now returns the proper year for BC dates.
It previously returned one less than the correct year. The
function now also returns the proper values for millennium and
century.
</P
></LI
><LI
><P
> <TT
CLASS="TYPE"
>CIDR</TT
> values now must have their nonmasked bits be zero.
For example, we no longer allow
<TT
CLASS="LITERAL"
>204.248.199.1/31</TT
> as a <TT
CLASS="TYPE"
>CIDR</TT
> value. Such
values should never have been accepted by
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> and will now be rejected.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>EXECUTE</TT
> now returns a completion tag that
matches the executed statement.
</P
></LI
><LI
><P
> <SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="COMMAND"
>\copy</TT
> command now reads or
writes to the query's <TT
CLASS="LITERAL"
>stdin/stdout</TT
>, rather than
<SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="LITERAL"
>stdin/stdout</TT
>. The previous
behavior can be accessed via new
<TT
CLASS="OPTION"
>pstdin</TT
>/<TT
CLASS="OPTION"
>pstdout</TT
> parameters.
</P
></LI
><LI
><P
> The JDBC client interface has been removed from the core
distribution, and is now hosted at <A
HREF="http://jdbc.postgresql.org"
TARGET="_top"
>http://jdbc.postgresql.org</A
>.
</P
></LI
><LI
><P
> The Tcl client interface has also been removed. There are several
Tcl interfaces now hosted at <A
HREF="http://gborg.postgresql.org"
TARGET="_top"
>http://gborg.postgresql.org</A
>.
</P
></LI
><LI
><P
> The server now uses its own time zone database, rather than the
one supplied by the operating system. This will provide consistent
behavior across all platforms. In most cases, there should be
little noticeable difference in time zone behavior, except that
the time zone names used by <TT
CLASS="COMMAND"
>SET</TT
>/<TT
CLASS="COMMAND"
>SHOW</TT
>
<TT
CLASS="VARNAME"
>TimeZone</TT
> might be different from what your platform provides.
</P
></LI
><LI
><P
> <SPAN
CLASS="APPLICATION"
>Configure</SPAN
>'s threading option no longer requires
users to run tests or edit configuration files; threading options
are now detected automatically.
</P
></LI
><LI
><P
> Now that tablespaces have been implemented,
<SPAN
CLASS="APPLICATION"
>initlocation</SPAN
> has been removed.
</P
></LI
><LI
><P
> The API for user-defined GiST indexes has been changed. The
Union and PickSplit methods are now passed a pointer to a
special <TT
CLASS="STRUCTNAME"
>GistEntryVector</TT
> structure,
rather than a <TT
CLASS="TYPE"
>bytea</TT
>.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN139607"
>E.196.3. Deprecated Features</A
></H2
><P
> Some aspects of <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>'s behavior
have been determined to be suboptimal. For the sake of backward
compatibility these have not been removed in 8.0, but they are
considered deprecated and will be removed in the next major
release.
</P
><P
></P
><UL
><LI
><P
> The 8.1 release will remove the <CODE
CLASS="FUNCTION"
>to_char()</CODE
> function
for intervals.
</P
></LI
><LI
><P
> The server now warns of empty strings passed to
<TT
CLASS="TYPE"
>oid</TT
>/<TT
CLASS="TYPE"
>float4</TT
>/<TT
CLASS="TYPE"
>float8</TT
> data
types, but continues to interpret them as zeroes as before.
In the next major release, empty strings will be considered
invalid input for these data types.
</P
></LI
><LI
><P
> By default, tables in <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> 8.0
and earlier are created with <TT
CLASS="TYPE"
>OID</TT
>s. In the next release,
this will <SPAN
CLASS="emphasis"
><I
CLASS="EMPHASIS"
>not</I
></SPAN
> be the case: to create a table
that contains <TT
CLASS="TYPE"
>OID</TT
>s, the <TT
CLASS="OPTION"
>WITH OIDS</TT
> clause must
be specified or the <TT
CLASS="VARNAME"
>default_with_oids</TT
>
configuration parameter must be set. Users are encouraged to
explicitly specify <TT
CLASS="OPTION"
>WITH OIDS</TT
> if their tables
require OIDs for compatibility with future releases of
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT2"
><H2
CLASS="SECT2"
><A
NAME="AEN139630"
>E.196.4. Changes</A
></H2
><P
> Below you will find a detailed account of the changes between
release 8.0 and the previous major release.
</P
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN139633"
>E.196.4.1. Performance Improvements</A
></H3
><P
></P
><UL
><LI
><P
> Support cross-data-type index usage (Tom)
</P
><P
> Before this change, many queries would not use an index if the data
types did not match exactly. This improvement makes index usage more
intuitive and consistent.
</P
></LI
><LI
><P
> New buffer replacement strategy that improves caching (Jan)
</P
><P
> Prior releases used a least-recently-used (LRU) cache to keep
recently referenced pages in memory. The LRU algorithm
did not consider the number of times a specific cache entry was
accessed, so large table scans could force out useful cache pages.
The new cache algorithm uses four separate lists to track most
recently used and most frequently used cache pages and dynamically
optimize their replacement based on the work load. This should
lead to much more efficient use of the shared buffer cache.
Administrators who have tested shared buffer sizes in the past
should retest with this new cache replacement policy.
</P
></LI
><LI
><P
> Add subprocess to write dirty buffers periodically to reduce
checkpoint writes (Jan)
</P
><P
> In previous releases, the checkpoint process, which runs every few
minutes, would write all dirty buffers to the operating system's
buffer cache then flush all dirty operating system buffers to
disk. This resulted in a periodic spike in disk usage that often
hurt performance. The new code uses a background writer to trickle
disk writes at a steady pace so checkpoints have far fewer dirty
pages to write to disk. Also, the new code does not issue a global
<CODE
CLASS="FUNCTION"
>sync()</CODE
> call, but instead <CODE
CLASS="FUNCTION"
>fsync()</CODE
>s just
the files written since the last checkpoint. This should improve
performance and minimize degradation during checkpoints.
</P
></LI
><LI
><P
> Add ability to prolong vacuum to reduce performance impact (Jan)
</P
><P
> On busy systems, <TT
CLASS="COMMAND"
>VACUUM</TT
> performs many I/O
requests which can hurt performance for other users. This
release allows you to slow down <TT
CLASS="COMMAND"
>VACUUM</TT
> to
reduce its impact on other users, though this increases the
total duration of <TT
CLASS="COMMAND"
>VACUUM</TT
>.
</P
></LI
><LI
><P
> Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
</P
><P
> This improves the way indexes are scanned when many duplicate
values exist in the index.
</P
></LI
><LI
><P
> Use dynamically-generated table size estimates while planning (Tom)
</P
><P
> Formerly the planner estimated table sizes using the values seen
by the last <TT
CLASS="COMMAND"
>VACUUM</TT
> or <TT
CLASS="COMMAND"
>ANALYZE</TT
>,
both as to physical table size (number of pages) and number of rows.
Now, the current physical table size is obtained from the kernel,
and the number of rows is estimated by multiplying the table size
by the row density (rows per page) seen by the last
<TT
CLASS="COMMAND"
>VACUUM</TT
> or <TT
CLASS="COMMAND"
>ANALYZE</TT
>. This should
produce more reliable estimates in cases where the table size has
changed significantly since the last housekeeping command.
</P
></LI
><LI
><P
> Improved index usage with <TT
CLASS="LITERAL"
>OR</TT
> clauses (Tom)
</P
><P
> This allows the optimizer to use indexes in statements with many OR
clauses that would not have been indexed in the past. It can also use
multi-column indexes where the first column is specified and the second
column is part of an <TT
CLASS="LITERAL"
>OR</TT
> clause.
</P
></LI
><LI
><P
> Improve matching of partial index clauses (Tom)
</P
><P
> The server is now smarter about using partial indexes in queries
involving complex <TT
CLASS="OPTION"
>WHERE</TT
> clauses.
</P
></LI
><LI
><P
> Improve performance of the GEQO optimizer (Tom)
</P
><P
> The GEQO optimizer is used to plan queries involving many tables (by
default, twelve or more). This release speeds up the way queries are
analyzed to decrease time spent in optimization.
</P
></LI
><LI
><P
> Miscellaneous optimizer improvements
</P
><P
> There is not room here to list all the minor improvements made, but
numerous special cases work better than in prior releases.
</P
></LI
><LI
><P
> Improve lookup speed for C functions (Tom)
</P
><P
> This release uses a hash table to lookup information for dynamically
loaded C functions. This improves their speed so they perform nearly as
quickly as functions that are built into the server executable.
</P
></LI
><LI
><P
> Add type-specific <TT
CLASS="COMMAND"
>ANALYZE</TT
> statistics
capability (Mark Cave-Ayland)
</P
><P
> This feature allows more flexibility in generating statistics
for nonstandard data types.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>ANALYZE</TT
> now collects statistics for
expression indexes (Tom)
</P
><P
> Expression indexes (also called functional indexes) allow users to
index not just columns but the results of expressions and function
calls. With this release, the optimizer can gather and use statistics
about the contents of expression indexes. This will greatly improve
the quality of planning for queries in which an expression index is
relevant.
</P
></LI
><LI
><P
> New two-stage sampling method for <TT
CLASS="COMMAND"
>ANALYZE</TT
>
(Manfred Koizar)
</P
><P
> This gives better statistics when the density of valid rows is very
different in different regions of a table.
</P
></LI
><LI
><P
> Speed up <TT
CLASS="COMMAND"
>TRUNCATE</TT
> (Tom)
</P
><P
> This buys back some of the performance loss observed in 7.4, while still
keeping <TT
CLASS="COMMAND"
>TRUNCATE</TT
> transaction-safe.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN139698"
>E.196.4.2. Server Changes</A
></H3
><P
></P
><UL
><LI
><P
> Add WAL file archiving and point-in-time recovery (Simon Riggs)
</P
></LI
><LI
><P
> Add tablespaces so admins can control disk layout (Gavin)
</P
></LI
><LI
><P
> Add a built-in log rotation program (Andreas Pflug)
</P
><P
> It is now possible to log server messages conveniently without
relying on either <SPAN
CLASS="APPLICATION"
>syslog</SPAN
> or an external log
rotation program.
</P
></LI
><LI
><P
> Add new read-only server configuration parameters to show server
compile-time settings: <TT
CLASS="VARNAME"
>block_size</TT
>,
<TT
CLASS="VARNAME"
>integer_datetimes</TT
>, <TT
CLASS="VARNAME"
>max_function_args</TT
>,
<TT
CLASS="VARNAME"
>max_identifier_length</TT
>, <TT
CLASS="VARNAME"
>max_index_keys</TT
> (Joe)
</P
></LI
><LI
><P
> Make quoting of <TT
CLASS="LITERAL"
>sameuser</TT
>, <TT
CLASS="LITERAL"
>samegroup</TT
>, and
<TT
CLASS="LITERAL"
>all</TT
> remove special meaning of these terms in
<TT
CLASS="FILENAME"
>pg_hba.conf</TT
> (Andrew)
</P
></LI
><LI
><P
> Use clearer IPv6 name <TT
CLASS="LITERAL"
>::1/128</TT
> for
<TT
CLASS="LITERAL"
>localhost</TT
> in default <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> (Andrew)
</P
></LI
><LI
><P
> Use CIDR format in <TT
CLASS="FILENAME"
>pg_hba.conf</TT
> examples (Andrew)
</P
></LI
><LI
><P
> Rename server configuration parameters <TT
CLASS="VARNAME"
>SortMem</TT
> and
<TT
CLASS="VARNAME"
>VacuumMem</TT
> to <TT
CLASS="VARNAME"
>work_mem</TT
> and
<TT
CLASS="VARNAME"
>maintenance_work_mem</TT
> (Old names still supported) (Tom)
</P
><P
> This change was made to clarify that bulk operations such as index and
foreign key creation use <TT
CLASS="VARNAME"
>maintenance_work_mem</TT
>, while
<TT
CLASS="VARNAME"
>work_mem</TT
> is for workspaces used during query execution.
</P
></LI
><LI
><P
> Allow logging of session disconnections using server configuration
<TT
CLASS="VARNAME"
>log_disconnections</TT
> (Andrew)
</P
></LI
><LI
><P
> Add new server configuration parameter <TT
CLASS="VARNAME"
>log_line_prefix</TT
> to
allow control of information emitted in each log line (Andrew)
</P
><P
> Available information includes user name, database name, remote IP
address, and session start time.
</P
></LI
><LI
><P
> Remove server configuration parameters <TT
CLASS="VARNAME"
>log_pid</TT
>,
<TT
CLASS="VARNAME"
>log_timestamp</TT
>, <TT
CLASS="VARNAME"
>log_source_port</TT
>; functionality
superseded by <TT
CLASS="VARNAME"
>log_line_prefix</TT
> (Andrew)
</P
></LI
><LI
><P
> Replace the <TT
CLASS="VARNAME"
>virtual_host</TT
> and <TT
CLASS="VARNAME"
>tcpip_socket</TT
>
parameters with a unified <TT
CLASS="VARNAME"
>listen_addresses</TT
> parameter
(Andrew, Tom)
</P
><P
> <TT
CLASS="VARNAME"
>virtual_host</TT
> could only specify a single IP address to
listen on. <TT
CLASS="VARNAME"
>listen_addresses</TT
> allows multiple addresses
to be specified.
</P
></LI
><LI
><P
> Listen on localhost by default, which eliminates the need for the
<TT
CLASS="OPTION"
>-i</TT
> postmaster switch in many scenarios (Andrew)
</P
><P
> Listening on localhost (<TT
CLASS="LITERAL"
>127.0.0.1</TT
>) opens no new
security holes but allows configurations like Windows and JDBC,
which do not support local sockets, to work without special
adjustments.
</P
></LI
><LI
><P
> Remove <TT
CLASS="VARNAME"
>syslog</TT
> server configuration parameter, and add more
logical <TT
CLASS="VARNAME"
>log_destination</TT
> variable to control log output
location (Magnus)
</P
></LI
><LI
><P
> Change server configuration parameter <TT
CLASS="VARNAME"
>log_statement</TT
> to take
values <TT
CLASS="VARNAME"
>all</TT
>, <TT
CLASS="VARNAME"
>mod</TT
>, <TT
CLASS="VARNAME"
>ddl</TT
>, or
<TT
CLASS="VARNAME"
>none</TT
> to select which queries are logged (Bruce)
</P
><P
> This allows administrators to log only data definition changes or
only data modification statements.
</P
></LI
><LI
><P
> Some logging-related configuration parameters could formerly be adjusted
by ordinary users, but only in the <SPAN
CLASS="QUOTE"
>"more verbose"</SPAN
> direction.
They are now treated more strictly: only superusers can set them.
However, a superuser can use <TT
CLASS="COMMAND"
>ALTER USER</TT
> to provide per-user
settings of these values for non-superusers. Also, it is now possible
for superusers to set values of superuser-only configuration parameters
via <TT
CLASS="LITERAL"
>PGOPTIONS</TT
>.
</P
></LI
><LI
><P
> Allow configuration files to be placed outside the data directory (mlw)
</P
><P
> By default, configuration files are kept in the cluster's top directory.
With this addition, configuration files can be placed outside the
data directory, easing administration.
</P
></LI
><LI
><P
> Plan prepared queries only when first executed so constants can be
used for statistics (Oliver Jowett)
</P
><P
> Prepared statements plan queries once and execute them many
times. While prepared queries avoid the overhead of re-planning
on each use, the quality of the plan suffers from not knowing the exact
parameters to be used in the query. In this release, planning of
unnamed prepared statements is delayed until the first execution,
and the actual parameter values of that execution are used as
optimization hints. This allows use of out-of-line parameter passing
without incurring a performance penalty.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>DECLARE CURSOR</TT
> to take parameters
(Oliver Jowett)
</P
><P
> It is now useful to issue <TT
CLASS="COMMAND"
>DECLARE CURSOR</TT
> in a
<CODE
CLASS="FUNCTION"
>Parse</CODE
> message with parameters. The parameter values
sent at <CODE
CLASS="FUNCTION"
>Bind</CODE
> time will be substituted into the
execution of the cursor's query.
</P
></LI
><LI
><P
> Fix hash joins and aggregates of <TT
CLASS="TYPE"
>inet</TT
> and
<TT
CLASS="TYPE"
>cidr</TT
> data types (Tom)
</P
><P
> Release 7.4 handled hashing of mixed <TT
CLASS="TYPE"
>inet</TT
> and
<TT
CLASS="TYPE"
>cidr</TT
> values incorrectly. (This bug did not exist
in prior releases because they wouldn't try to hash either
data type.)
</P
></LI
><LI
><P
> Make <TT
CLASS="VARNAME"
>log_duration</TT
> print only when <TT
CLASS="VARNAME"
>log_statement</TT
>
prints the query (Ed L.)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN139806"
>E.196.4.3. Query Changes</A
></H3
><P
></P
><UL
><LI
><P
> Add savepoints (nested transactions) (Alvaro)
</P
></LI
><LI
><P
> Unsupported isolation levels are now accepted and promoted to the
nearest supported level (Peter)
</P
><P
> The SQL specification states that if a database doesn't support a
specific isolation level, it should use the next more restrictive level.
This change complies with that recommendation.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>BEGIN WORK</TT
> to specify transaction
isolation levels like <TT
CLASS="COMMAND"
>START TRANSACTION</TT
> does
(Bruce)
</P
></LI
><LI
><P
> Fix table permission checking for cases in which rules generate
a query type different from the originally submitted query (Tom)
</P
></LI
><LI
><P
> Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
David Fetter)
</P
><P
> In previous releases, because single quotes had to be used to
quote a function's body, the use of single quotes inside the
function text required use of two single quotes or other error-prone
notations. With this release we add the ability to use "dollar
quoting" to quote a block of text. The ability to use different
quoting delimiters at different nesting levels greatly simplifies
the task of quoting correctly, especially in complex functions.
Dollar quoting can be used anywhere quoted text is needed.
</P
></LI
><LI
><P
> Make <TT
CLASS="LITERAL"
>CASE val WHEN compval1 THEN ...</TT
> evaluate <TT
CLASS="LITERAL"
>val</TT
> only once (Tom)
</P
><P
> <TT
CLASS="OPTION"
>CASE</TT
> no longer evaluates the tested expression multiple
times. This has benefits when the expression is complex or is
volatile.
</P
></LI
><LI
><P
> Test <TT
CLASS="OPTION"
>HAVING</TT
> before computing target list of an
aggregate query (Tom)
</P
><P
> Fixes improper failure of cases such as <TT
CLASS="LITERAL"
>SELECT SUM(win)/SUM(lose)
... GROUP BY ... HAVING SUM(lose) > 0</TT
>. This should work but formerly
could fail with divide-by-zero.
</P
></LI
><LI
><P
> Replace <TT
CLASS="VARNAME"
>max_expr_depth</TT
> parameter with
<TT
CLASS="VARNAME"
>max_stack_depth</TT
> parameter, measured in kilobytes of stack
size (Tom)
</P
><P
> This gives us a fairly bulletproof defense against crashing due to
runaway recursive functions. Instead of measuring the depth of expression
nesting, we now directly measure the size of the execution stack.
</P
></LI
><LI
><P
> Allow arbitrary row expressions (Tom)
</P
><P
> This release allows SQL expressions to contain arbitrary composite
types, that is, row values. It also allows functions to more easily
take rows as arguments and return row values.
</P
></LI
><LI
><P
> Allow <TT
CLASS="OPTION"
>LIKE</TT
>/<TT
CLASS="OPTION"
>ILIKE</TT
> to be used as the operator
in row and subselect comparisons (Fabien Coelho)
</P
></LI
><LI
><P
> Avoid locale-specific case conversion of basic ASCII letters in
identifiers and keywords (Tom)
</P
><P
> This solves the <SPAN
CLASS="QUOTE"
>"Turkish problem"</SPAN
> with mangling of words
containing <TT
CLASS="LITERAL"
>I</TT
> and <TT
CLASS="LITERAL"
>i</TT
>. Folding of characters
outside the 7-bit-ASCII set is still locale-aware.
</P
></LI
><LI
><P
> Improve syntax error reporting (Fabien, Tom)
</P
><P
> Syntax error reports are more useful than before.
</P
></LI
><LI
><P
> Change <TT
CLASS="COMMAND"
>EXECUTE</TT
> to return a completion tag
matching the executed statement (Kris Jurka)
</P
><P
> Previous releases return an <TT
CLASS="COMMAND"
>EXECUTE</TT
> tag for
any <TT
CLASS="COMMAND"
>EXECUTE</TT
> call. In this release, the tag
returned will reflect the command executed.
</P
></LI
><LI
><P
> Avoid emitting <TT
CLASS="OPTION"
>NATURAL CROSS JOIN</TT
> in rule listings (Tom)
</P
><P
> Such a clause makes no logical sense, but in some cases the rule
decompiler formerly produced this syntax.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN139865"
>E.196.4.4. Object Manipulation Changes</A
></H3
><P
></P
><UL
><LI
><P
> Add <TT
CLASS="COMMAND"
>COMMENT ON</TT
> for casts, conversions, languages,
operator classes, and large objects (Christopher)
</P
></LI
><LI
><P
> Add new server configuration parameter <TT
CLASS="VARNAME"
>default_with_oids</TT
> to
control whether tables are created with <TT
CLASS="TYPE"
>OID</TT
>s by default (Neil)
</P
><P
> This allows administrators to control whether <TT
CLASS="COMMAND"
>CREATE
TABLE</TT
> commands create tables with or without <TT
CLASS="TYPE"
>OID</TT
>
columns by default. (Note: the current factory default setting for
<TT
CLASS="VARNAME"
>default_with_oids</TT
> is <TT
CLASS="LITERAL"
>TRUE</TT
>, but the default
will become <TT
CLASS="LITERAL"
>FALSE</TT
> in future releases.)
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>WITH</TT
> / <TT
CLASS="OPTION"
>WITHOUT OIDS</TT
> clause to
<TT
CLASS="COMMAND"
>CREATE TABLE AS</TT
> (Neil)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>ALTER TABLE DROP COLUMN</TT
> to drop an <TT
CLASS="TYPE"
>OID</TT
>
column (<TT
CLASS="COMMAND"
>ALTER TABLE SET WITHOUT OIDS</TT
> still works)
(Tom)
</P
></LI
><LI
><P
> Allow composite types as table columns (Tom)
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>ALTER ... ADD COLUMN</TT
> with defaults and
<TT
CLASS="OPTION"
>NOT NULL</TT
> constraints; works per SQL spec (Rod)
</P
><P
> It is now possible for <TT
CLASS="OPTION"
>ADD COLUMN</TT
> to create a column
that is not initially filled with NULLs, but with a specified
default value.
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>ALTER COLUMN TYPE</TT
> to change column's type (Rod)
</P
><P
> It is now possible to alter a column's data type without dropping
and re-adding the column.
</P
></LI
><LI
><P
> Allow multiple <TT
CLASS="COMMAND"
>ALTER</TT
> actions in a single <TT
CLASS="COMMAND"
>ALTER
TABLE</TT
> command (Rod)
</P
><P
> This is particularly useful for <TT
CLASS="COMMAND"
>ALTER</TT
> commands that
rewrite the table (which include <TT
CLASS="OPTION"
>ALTER COLUMN TYPE</TT
> and
<TT
CLASS="OPTION"
>ADD COLUMN</TT
> with a default). By grouping
<TT
CLASS="COMMAND"
>ALTER</TT
> commands together, the table need be rewritten
only once.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>ALTER TABLE</TT
> to add <TT
CLASS="TYPE"
>SERIAL</TT
>
columns (Tom)
</P
><P
> This falls out from the new capability of specifying defaults for new
columns.
</P
></LI
><LI
><P
> Allow changing the owners of aggregates, conversions, databases,
functions, operators, operator classes, schemas, types, and tablespaces
(Christopher, Euler Taveira de Oliveira)
</P
><P
> Previously this required modifying the system tables directly.
</P
></LI
><LI
><P
> Allow temporary object creation to be limited to <TT
CLASS="OPTION"
>SECURITY
DEFINER</TT
> functions (Sean Chittenden)
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>ALTER TABLE ... SET WITHOUT CLUSTER</TT
> (Christopher)
</P
><P
> Prior to this release, there was no way to clear an auto-cluster
specification except to modify the system tables.
</P
></LI
><LI
><P
> Constraint/Index/<TT
CLASS="TYPE"
>SERIAL</TT
> names are now
<TT
CLASS="REPLACEABLE"
><I
>table_column_type</I
></TT
>
with numbers appended to guarantee uniqueness within the schema
(Tom)
</P
><P
> The SQL specification states that such names should be unique
within a schema.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>pg_get_serial_sequence()</CODE
> to return a
<TT
CLASS="TYPE"
>SERIAL</TT
> column's sequence name (Christopher)
</P
><P
> This allows automated scripts to reliably find the <TT
CLASS="TYPE"
>SERIAL</TT
>
sequence name.
</P
></LI
><LI
><P
> Warn when primary/foreign key data type mismatch requires costly lookup
</P
></LI
><LI
><P
> New <TT
CLASS="COMMAND"
>ALTER INDEX</TT
> command to allow moving of indexes
between tablespaces (Gavin)
</P
></LI
><LI
><P
> Make <TT
CLASS="COMMAND"
>ALTER TABLE OWNER</TT
> change dependent sequence
ownership too (Alvaro)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN139946"
>E.196.4.5. Utility Command Changes</A
></H3
><P
></P
><UL
><LI
><P
> Allow <TT
CLASS="COMMAND"
>CREATE SCHEMA</TT
> to create triggers,
indexes, and sequences (Neil)
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>ALSO</TT
> keyword to <TT
CLASS="COMMAND"
>CREATE RULE</TT
> (Fabien
Coelho)
</P
><P
> This allows <TT
CLASS="OPTION"
>ALSO</TT
> to be added to rule creation to contrast it with
<TT
CLASS="OPTION"
>INSTEAD</TT
> rules.
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>NOWAIT</TT
> option to <TT
CLASS="COMMAND"
>LOCK</TT
> (Tatsuo)
</P
><P
> This allows the <TT
CLASS="COMMAND"
>LOCK</TT
> command to fail if it
would have to wait for the requested lock.
</P
></LI
><LI
><P
> Allow <TT
CLASS="COMMAND"
>COPY</TT
> to read and write
comma-separated-value (CSV) files (Andrew, Bruce)
</P
></LI
><LI
><P
> Generate error if the <TT
CLASS="COMMAND"
>COPY</TT
> delimiter and NULL
string conflict (Bruce)
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>GRANT</TT
>/<TT
CLASS="COMMAND"
>REVOKE</TT
> behavior
follows the SQL spec more closely
</P
></LI
><LI
><P
> Avoid locking conflict between <TT
CLASS="COMMAND"
>CREATE INDEX</TT
>
and <TT
CLASS="COMMAND"
>CHECKPOINT</TT
> (Tom)
</P
><P
> In 7.3 and 7.4, a long-running B-tree index build could block concurrent
<TT
CLASS="COMMAND"
>CHECKPOINT</TT
>s from completing, thereby causing WAL bloat because the
WAL log could not be recycled.
</P
></LI
><LI
><P
> Database-wide <TT
CLASS="COMMAND"
>ANALYZE</TT
> does not hold locks
across tables (Tom)
</P
><P
> This reduces the potential for deadlocks against other backends
that want exclusive locks on tables. To get the benefit of this
change, do not execute database-wide <TT
CLASS="COMMAND"
>ANALYZE</TT
>
inside a transaction block (<TT
CLASS="COMMAND"
>BEGIN</TT
> block); it
must be able to commit and start a new transaction for each
table.
</P
></LI
><LI
><P
> <TT
CLASS="COMMAND"
>REINDEX</TT
> does not exclusively lock the index's
parent table anymore
</P
><P
> The index itself is still exclusively locked, but readers of the
table can continue if they are not using the particular index
being rebuilt.
</P
></LI
><LI
><P
> Erase MD5 user passwords when a user is renamed (Bruce)
</P
><P
> <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> uses the user name as salt
when encrypting passwords via MD5. When a user's name is changed,
the salt will no longer match the stored MD5 password, so the
stored password becomes useless. In this release a notice is
generated and the password is cleared. A new password must then
be assigned if the user is to be able to log in with a password.
</P
></LI
><LI
><P
> New <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
> <TT
CLASS="OPTION"
>kill</TT
> option for Windows (Andrew)
</P
><P
> Windows does not have a <TT
CLASS="LITERAL"
>kill</TT
> command to send signals to
backends so this capability was added to <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
>.
</P
></LI
><LI
><P
> Information schema improvements
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>--pwfile</TT
> option to
<SPAN
CLASS="APPLICATION"
>initdb</SPAN
> so the initial password can be
set by GUI tools (Magnus)
</P
></LI
><LI
><P
> Detect locale/encoding mismatch in
<SPAN
CLASS="APPLICATION"
>initdb</SPAN
> (Peter)
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>register</TT
> command to <SPAN
CLASS="APPLICATION"
>pg_ctl</SPAN
> to
register Windows operating system service (Dave Page)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140015"
>E.196.4.6. Data Type and Function Changes</A
></H3
><P
></P
><UL
><LI
><P
> More complete support for composite types (row types) (Tom)
</P
><P
> Composite values can be used in many places where only scalar values
worked before.
</P
></LI
><LI
><P
> Reject nonrectangular array values as erroneous (Joe)
</P
><P
> Formerly, <CODE
CLASS="FUNCTION"
>array_in</CODE
> would silently build a
surprising result.
</P
></LI
><LI
><P
> Overflow in integer arithmetic operations is now detected (Tom)
</P
></LI
><LI
><P
> The arithmetic operators associated with the single-byte
<TT
CLASS="TYPE"
>"char"</TT
> data type have been removed.
</P
><P
> Formerly, the parser would select these operators in many situations
where an <SPAN
CLASS="QUOTE"
>"unable to select an operator"</SPAN
> error would be more
appropriate, such as <TT
CLASS="LITERAL"
>null * null</TT
>. If you actually want
to do arithmetic on a <TT
CLASS="TYPE"
>"char"</TT
> column, you can cast it to
integer explicitly.
</P
></LI
><LI
><P
> Syntax checking of array input values considerably tightened up (Joe)
</P
><P
> Junk that was previously allowed in odd places with odd results
now causes an <TT
CLASS="LITERAL"
>ERROR</TT
>, for example, non-whitespace
after the closing right brace.
</P
></LI
><LI
><P
> Empty-string array element values must now be written as
<TT
CLASS="LITERAL"
>""</TT
>, rather than writing nothing (Joe)
</P
><P
> Formerly, both ways of writing an empty-string element value were
allowed, but now a quoted empty string is required. The case where
nothing at all appears will probably be considered to be a NULL
element value in some future release.
</P
></LI
><LI
><P
> Array element trailing whitespace is now ignored (Joe)
</P
><P
> Formerly leading whitespace was ignored, but trailing whitespace
between an element value and the delimiter or right brace was
significant. Now trailing whitespace is also ignored.
</P
></LI
><LI
><P
> Emit array values with explicit array bounds when lower bound is not one
(Joe)
</P
></LI
><LI
><P
> Accept <TT
CLASS="LITERAL"
>YYYY-monthname-DD</TT
> as a date string (Tom)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>netmask</CODE
> and <CODE
CLASS="FUNCTION"
>hostmask</CODE
> functions
return maximum-length mask length (Tom)
</P
></LI
><LI
><P
> Change factorial function to return <TT
CLASS="TYPE"
>numeric</TT
> (Gavin)
</P
><P
> Returning <TT
CLASS="TYPE"
>numeric</TT
> allows the factorial function to
work for a wider range of input values.
</P
></LI
><LI
><P
> <CODE
CLASS="FUNCTION"
>to_char</CODE
>/<CODE
CLASS="FUNCTION"
>to_date()</CODE
> date conversion
improvements (Kurt Roeckx, Fabien Coelho)
</P
></LI
><LI
><P
> Make <CODE
CLASS="FUNCTION"
>length()</CODE
> disregard trailing spaces in
<TT
CLASS="TYPE"
>CHAR(n)</TT
> (Gavin)
</P
><P
> This change was made to improve consistency: trailing spaces are
semantically insignificant in <TT
CLASS="TYPE"
>CHAR(n)</TT
> data, so they
should not be counted by <CODE
CLASS="FUNCTION"
>length()</CODE
>.
</P
></LI
><LI
><P
> Warn about empty string being passed to
<TT
CLASS="TYPE"
>OID</TT
>/<TT
CLASS="TYPE"
>float4</TT
>/<TT
CLASS="TYPE"
>float8</TT
> data types (Neil)
</P
><P
> 8.1 will throw an error instead.
</P
></LI
><LI
><P
> Allow leading or trailing whitespace in
<TT
CLASS="TYPE"
>int2</TT
>/<TT
CLASS="TYPE"
>int4</TT
>/<TT
CLASS="TYPE"
>int8</TT
>/<TT
CLASS="TYPE"
>float4</TT
>/<TT
CLASS="TYPE"
>float8</TT
>
input routines
(Neil)
</P
></LI
><LI
><P
> Better support for IEEE <TT
CLASS="LITERAL"
>Infinity</TT
> and <TT
CLASS="LITERAL"
>NaN</TT
>
values in <TT
CLASS="TYPE"
>float4</TT
>/<TT
CLASS="TYPE"
>float8</TT
> (Neil)
</P
><P
> These should now work on all platforms that support IEEE-compliant
floating point arithmetic.
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>week</TT
> option to <CODE
CLASS="FUNCTION"
>date_trunc()</CODE
> (Robert Creager)
</P
></LI
><LI
><P
> Fix <CODE
CLASS="FUNCTION"
>to_char</CODE
> for <TT
CLASS="LITERAL"
>1 BC</TT
>
(previously it returned <TT
CLASS="LITERAL"
>1 AD</TT
>) (Bruce)
</P
></LI
><LI
><P
> Fix <CODE
CLASS="FUNCTION"
>date_part(year)</CODE
> for BC dates (previously it
returned one less than the correct year) (Bruce)
</P
></LI
><LI
><P
> Fix <CODE
CLASS="FUNCTION"
>date_part()</CODE
> to return the proper millennium and
century (Fabien Coelho)
</P
><P
> In previous versions, the century and millennium results had a wrong
number and started in the wrong year, as compared to standard
reckoning of such things.
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>ceiling()</CODE
> as an alias for <CODE
CLASS="FUNCTION"
>ceil()</CODE
>,
and <CODE
CLASS="FUNCTION"
>power()</CODE
> as an alias for <CODE
CLASS="FUNCTION"
>pow()</CODE
> for
standards compliance (Neil)
</P
></LI
><LI
><P
> Change <CODE
CLASS="FUNCTION"
>ln()</CODE
>, <CODE
CLASS="FUNCTION"
>log()</CODE
>,
<CODE
CLASS="FUNCTION"
>power()</CODE
>, and <CODE
CLASS="FUNCTION"
>sqrt()</CODE
> to emit the correct
<TT
CLASS="LITERAL"
>SQLSTATE</TT
> error codes for certain error conditions, as
specified by SQL:2003 (Neil)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>width_bucket()</CODE
> function as defined by SQL:2003 (Neil)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>generate_series()</CODE
> functions to simplify working
with numeric sets (Joe)
</P
></LI
><LI
><P
> Fix <CODE
CLASS="FUNCTION"
>upper/lower/initcap()</CODE
> functions to work with
multibyte encodings (Tom)
</P
></LI
><LI
><P
> Add boolean and bitwise integer <TT
CLASS="OPTION"
>AND</TT
>/<TT
CLASS="OPTION"
>OR</TT
>
aggregates (Fabien Coelho)
</P
></LI
><LI
><P
> New session information functions to return network addresses for client
and server (Sean Chittenden)
</P
></LI
><LI
><P
> Add function to determine the area of a closed path (Sean Chittenden)
</P
></LI
><LI
><P
> Add function to send cancel request to other backends (Magnus)
</P
></LI
><LI
><P
> Add <TT
CLASS="TYPE"
>interval</TT
> plus <TT
CLASS="TYPE"
>datetime</TT
> operators (Tom)
</P
><P
> The reverse ordering, <TT
CLASS="TYPE"
>datetime</TT
> plus <TT
CLASS="TYPE"
>interval</TT
>,
was already supported, but both are required by the SQL standard.
</P
></LI
><LI
><P
> Casting an integer to <TT
CLASS="TYPE"
>BIT(N)</TT
> selects the rightmost N bits
of the integer
(Tom)
</P
><P
> In prior releases, the leftmost N bits were selected, but this was
deemed unhelpful, not to mention inconsistent with casting from bit
to int.
</P
></LI
><LI
><P
> Require <TT
CLASS="TYPE"
>CIDR</TT
> values to have all nonmasked bits be zero
(Kevin Brintnall)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140152"
>E.196.4.7. Server-Side Language Changes</A
></H3
><P
></P
><UL
><LI
><P
> In <TT
CLASS="LITERAL"
>READ COMMITTED</TT
> serialization mode, volatile functions
now see the results of concurrent transactions committed up to the
beginning of each statement within the function, rather than up to the
beginning of the interactive command that called the function.
</P
></LI
><LI
><P
> Functions declared <TT
CLASS="LITERAL"
>STABLE</TT
> or <TT
CLASS="LITERAL"
>IMMUTABLE</TT
> always
use the snapshot of the calling query, and therefore do not see the
effects of actions taken after the calling query starts, whether in
their own transaction or other transactions. Such a function must be
read-only, too, meaning that it cannot use any SQL commands other than
<TT
CLASS="COMMAND"
>SELECT</TT
>. There is a considerable performance gain from
declaring a function <TT
CLASS="LITERAL"
>STABLE</TT
> or <TT
CLASS="LITERAL"
>IMMUTABLE</TT
>
rather than <TT
CLASS="LITERAL"
>VOLATILE</TT
>.
</P
></LI
><LI
><P
> Nondeferred <TT
CLASS="OPTION"
>AFTER</TT
> triggers are now fired immediately
after completion of the triggering query, rather than upon
finishing the current interactive command. This makes a difference
when the triggering query occurred within a function: the trigger
is invoked before the function proceeds to its next operation. For
example, if a function inserts a new row into a table, any
nondeferred foreign key checks occur before proceeding with the
function.
</P
></LI
><LI
><P
> Allow function parameters to be declared with names (Dennis Björklund)
</P
><P
> This allows better documentation of functions. Whether the names
actually do anything depends on the specific function language
being used.
</P
></LI
><LI
><P
> Allow PL/pgSQL parameter names to be referenced in the function (Dennis Björklund)
</P
><P
> This basically creates an automatic alias for each named parameter.
</P
></LI
><LI
><P
> Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
</P
><P
> This allows us to catch simple syntax errors sooner.
</P
></LI
><LI
><P
> More support for composite types (row and record variables) in PL/pgSQL
</P
><P
> For example, it now works to pass a rowtype variable to another function
as a single variable.
</P
></LI
><LI
><P
> Default values for PL/pgSQL variables can now reference previously
declared variables
</P
></LI
><LI
><P
> Improve parsing of PL/pgSQL FOR loops (Tom)
</P
><P
> Parsing is now driven by presence of <TT
CLASS="LITERAL"
>".."</TT
> rather than
data type of <TT
CLASS="OPTION"
>FOR</TT
> variable. This makes no difference for
correct functions, but should result in more understandable error
messages when a mistake is made.
</P
></LI
><LI
><P
> Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
</P
></LI
><LI
><P
> In PL/Tcl, SPI commands are now run in subtransactions. If an error
occurs, the subtransaction is cleaned up and the error is reported
as an ordinary Tcl error, which can be trapped with <TT
CLASS="LITERAL"
>catch</TT
>.
Formerly, it was not possible to catch such errors.
</P
></LI
><LI
><P
> Accept <TT
CLASS="COMMAND"
>ELSEIF</TT
> in PL/pgSQL (Neil)
</P
><P
> Previously PL/pgSQL only allowed <TT
CLASS="COMMAND"
>ELSIF</TT
>, but many people
are accustomed to spelling this keyword <TT
CLASS="COMMAND"
>ELSEIF</TT
>.
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140199"
>E.196.4.8. <SPAN
CLASS="APPLICATION"
>psql</SPAN
> Changes</A
></H3
><P
></P
><UL
><LI
><P
> Improve <SPAN
CLASS="APPLICATION"
>psql</SPAN
> information display about database
objects (Christopher)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>psql</SPAN
> to display group membership in
<TT
CLASS="COMMAND"
>\du</TT
> and <TT
CLASS="COMMAND"
>\dg</TT
> (Markus Bertheau)
</P
></LI
><LI
><P
> Prevent <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <TT
CLASS="COMMAND"
>\dn</TT
> from showing
temporary schemas (Bruce)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>psql</SPAN
> to handle tilde user expansion for file
names (Zach Irmen)
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>psql</SPAN
> to display fancy prompts, including
color, via <SPAN
CLASS="APPLICATION"
>readline</SPAN
> (Reece Hart, Chet Ramey)
</P
></LI
><LI
><P
> Make <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <TT
CLASS="COMMAND"
>\copy</TT
> match <TT
CLASS="COMMAND"
>COPY</TT
> command syntax
fully (Tom)
</P
></LI
><LI
><P
> Show the location of syntax errors (Fabien Coelho, Tom)
</P
></LI
><LI
><P
> Add <TT
CLASS="COMMAND"
>CLUSTER</TT
> information to <SPAN
CLASS="APPLICATION"
>psql</SPAN
>
<TT
CLASS="COMMAND"
>\d</TT
> display
(Bruce)
</P
></LI
><LI
><P
> Change <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <TT
CLASS="COMMAND"
>\copy stdin/stdout</TT
> to read
from command input/output (Bruce)
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>pstdin</TT
>/<TT
CLASS="OPTION"
>pstdout</TT
> to read from
<SPAN
CLASS="APPLICATION"
>psql</SPAN
>'s <TT
CLASS="LITERAL"
>stdin</TT
>/<TT
CLASS="LITERAL"
>stdout</TT
> (Mark
Feit)
</P
></LI
><LI
><P
> Add global <SPAN
CLASS="APPLICATION"
>psql</SPAN
> configuration file, <TT
CLASS="FILENAME"
>psqlrc.sample</TT
>
(Bruce)
</P
><P
> This allows a central file where global <SPAN
CLASS="APPLICATION"
>psql</SPAN
> startup commands can
be stored.
</P
></LI
><LI
><P
> Have <SPAN
CLASS="APPLICATION"
>psql</SPAN
> <TT
CLASS="COMMAND"
>\d+</TT
> indicate if the table
has an <TT
CLASS="TYPE"
>OID</TT
> column (Neil)
</P
></LI
><LI
><P
> On Windows, use binary mode in <SPAN
CLASS="APPLICATION"
>psql</SPAN
> when reading files so control-Z
is not seen as end-of-file
</P
></LI
><LI
><P
> Have <TT
CLASS="COMMAND"
>\dn+</TT
> show permissions and description for schemas (Dennis
Björklund)
</P
></LI
><LI
><P
> Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
</P
></LI
><LI
><P
> Allow boolean settings to be set using upper or lower case (Michael Paesold)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140266"
>E.196.4.9. <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> Changes</A
></H3
><P
></P
><UL
><LI
><P
> Use dependency information to improve the reliability of
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> (Tom)
</P
><P
> This should solve the longstanding problems with related objects
sometimes being dumped in the wrong order.
</P
></LI
><LI
><P
> Have <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> output objects in alphabetical order if possible (Tom)
</P
><P
> This should make it easier to identify changes between
dump files.
</P
></LI
><LI
><P
> Allow <SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
> to ignore some SQL errors (Fabien Coelho)
</P
><P
> This makes <SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
>'s behavior similar to the
results of feeding a <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> output script to
<SPAN
CLASS="APPLICATION"
>psql</SPAN
>. In most cases, ignoring errors and plowing
ahead is the most useful thing to do. Also added was a pg_restore
option to give the old behavior of exiting on an error.
</P
></LI
><LI
><P
> <SPAN
CLASS="APPLICATION"
>pg_restore</SPAN
> <TT
CLASS="OPTION"
>-l</TT
> display now includes
objects' schema names
</P
></LI
><LI
><P
> New begin/end markers in <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> text output (Bruce)
</P
></LI
><LI
><P
> Add start/stop times for
<SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
>/<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> in verbose mode
(Bruce)
</P
></LI
><LI
><P
> Allow most <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> options in
<SPAN
CLASS="APPLICATION"
>pg_dumpall</SPAN
> (Christopher)
</P
></LI
><LI
><P
> Have <SPAN
CLASS="APPLICATION"
>pg_dump</SPAN
> use <TT
CLASS="COMMAND"
>ALTER OWNER</TT
> rather
than <TT
CLASS="COMMAND"
>SET SESSION AUTHORIZATION</TT
> by default
(Christopher)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140305"
>E.196.4.10. libpq Changes</A
></H3
><P
></P
><UL
><LI
><P
> Make libpq's <TT
CLASS="OPTION"
>SIGPIPE</TT
> handling thread-safe (Bruce)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>PQmbdsplen()</CODE
> which returns the display length
of a character (Tatsuo)
</P
></LI
><LI
><P
> Add thread locking to <SPAN
CLASS="APPLICATION"
>SSL</SPAN
> and
<SPAN
CLASS="APPLICATION"
>Kerberos</SPAN
> connections (Manfred Spraul)
</P
></LI
><LI
><P
> Allow <CODE
CLASS="FUNCTION"
>PQoidValue()</CODE
>, <CODE
CLASS="FUNCTION"
>PQcmdTuples()</CODE
>, and
<CODE
CLASS="FUNCTION"
>PQoidStatus()</CODE
> to work on <TT
CLASS="COMMAND"
>EXECUTE</TT
>
commands (Neil)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>PQserverVersion()</CODE
> to provide more convenient
access to the server version number (Greg Sabino Mullane)
</P
></LI
><LI
><P
> Add <CODE
CLASS="FUNCTION"
>PQprepare/PQsendPrepared()</CODE
> functions to support
preparing statements without necessarily specifying the data types
of their parameters (Abhijit Menon-Sen)
</P
></LI
><LI
><P
> Many ECPG improvements, including <TT
CLASS="COMMAND"
>SET DESCRIPTOR</TT
> (Michael)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140333"
>E.196.4.11. Source Code Changes</A
></H3
><P
></P
><UL
><LI
><P
> Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
</P
></LI
><LI
><P
> Shell script commands converted to C versions for Windows support (Andrew)
</P
></LI
><LI
><P
> Create an extension makefile framework (Fabien Coelho, Peter)
</P
><P
> This simplifies the task of building extensions outside the original
source tree.
</P
></LI
><LI
><P
> Support relocatable installations (Bruce)
</P
><P
> Directory paths for installed files (such as the
<TT
CLASS="FILENAME"
>/share</TT
> directory) are now computed relative to the
actual location of the executables, so that an installation tree
can be moved to another place without reconfiguring and
rebuilding.
</P
></LI
><LI
><P
> Use <TT
CLASS="OPTION"
>--with-docdir</TT
> to choose installation location of documentation; also
allow <TT
CLASS="OPTION"
>--infodir</TT
> (Peter)
</P
></LI
><LI
><P
> Add <TT
CLASS="OPTION"
>--without-docdir</TT
> to prevent installation of documentation (Peter)
</P
></LI
><LI
><P
> Upgrade to <SPAN
CLASS="APPLICATION"
>DocBook</SPAN
> V4.2 SGML (Peter)
</P
></LI
><LI
><P
> New <TT
CLASS="LITERAL"
>PostgreSQL</TT
> <SPAN
CLASS="APPLICATION"
>CVS</SPAN
> tag (Marc)
</P
><P
> This was done to make it easier for organizations to manage their
own copies of the <SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
>
<SPAN
CLASS="APPLICATION"
>CVS</SPAN
> repository. File version stamps from the master
repository will not get munged by checking into or out of a copied
repository.
</P
></LI
><LI
><P
> Clarify locking code (Manfred Koizar)
</P
></LI
><LI
><P
> Buffer manager cleanup (Neil)
</P
></LI
><LI
><P
> Decouple platform tests from CPU spinlock code (Bruce, Tom)
</P
></LI
><LI
><P
> Add inlined test-and-set code on PA-RISC for <SPAN
CLASS="APPLICATION"
>gcc</SPAN
>
(ViSolve, Tom)
</P
></LI
><LI
><P
> Improve i386 spinlock code (Manfred Spraul)
</P
></LI
><LI
><P
> Clean up spinlock assembly code to avoid warnings from newer
<SPAN
CLASS="APPLICATION"
>gcc</SPAN
> releases (Tom)
</P
></LI
><LI
><P
> Remove JDBC from source tree; now a separate project
</P
></LI
><LI
><P
> Remove the libpgtcl client interface; now a separate project
</P
></LI
><LI
><P
> More accurately estimate memory and file descriptor usage (Tom)
</P
></LI
><LI
><P
> Improvements to the Mac OS X startup scripts (Ray A.)
</P
></LI
><LI
><P
> New <CODE
CLASS="FUNCTION"
>fsync()</CODE
> test program (Bruce)
</P
></LI
><LI
><P
> Major documentation improvements (Neil, Peter)
</P
></LI
><LI
><P
> Remove <SPAN
CLASS="APPLICATION"
>pg_encoding</SPAN
>; not needed
anymore
</P
></LI
><LI
><P
> Remove <SPAN
CLASS="APPLICATION"
>pg_id</SPAN
>; not needed anymore
</P
></LI
><LI
><P
> Remove <SPAN
CLASS="APPLICATION"
>initlocation</SPAN
>; not needed
anymore
</P
></LI
><LI
><P
> Auto-detect thread flags (no more manual testing) (Bruce)
</P
></LI
><LI
><P
> Use Olson's public domain <SPAN
CLASS="APPLICATION"
>timezone</SPAN
> library (Magnus)
</P
></LI
><LI
><P
> With threading enabled, use thread flags on Unixware for
backend executables too (Bruce)
</P
><P
> Unixware cannot mix threaded and nonthreaded object files in the
same executable, so everything must be compiled as threaded.
</P
></LI
><LI
><P
> <SPAN
CLASS="APPLICATION"
>psql</SPAN
> now uses a <SPAN
CLASS="APPLICATION"
>flex</SPAN
>-generated
lexical analyzer to process command strings
</P
></LI
><LI
><P
> Reimplement the linked list data structure used throughout the
backend (Neil)
</P
><P
> This improves performance by allowing list append and length
operations to be more efficient.
</P
></LI
><LI
><P
> Allow dynamically loaded modules to create their own server configuration
parameters (Thomas Hallgren)
</P
></LI
><LI
><P
> New Brazilian version of FAQ (Euler Taveira de Oliveira)
</P
></LI
><LI
><P
> Add French FAQ (Guillaume Lelarge)
</P
></LI
><LI
><P
> New <SPAN
CLASS="APPLICATION"
>pgevent</SPAN
> for Windows logging
</P
></LI
><LI
><P
> Make libpq and ECPG build as proper shared libraries on OS X (Tom)
</P
></LI
></UL
></DIV
><DIV
CLASS="SECT3"
><H3
CLASS="SECT3"
><A
NAME="AEN140426"
>E.196.4.12. Contrib Changes</A
></H3
><P
></P
><UL
><LI
><P
> Overhaul of <TT
CLASS="FILENAME"
>contrib/dblink</TT
> (Joe)
</P
></LI
><LI
><P
> <TT
CLASS="FILENAME"
>contrib/dbmirror</TT
> improvements (Steven Singer)
</P
></LI
><LI
><P
> New <TT
CLASS="FILENAME"
>contrib/xml2</TT
> (John Gray, Torchbox)
</P
></LI
><LI
><P
> Updated <TT
CLASS="FILENAME"
>contrib/mysql</TT
>
</P
></LI
><LI
><P
> New version of <TT
CLASS="FILENAME"
>contrib/btree_gist</TT
> (Teodor)
</P
></LI
><LI
><P
> New <TT
CLASS="FILENAME"
>contrib/trgm</TT
>, trigram matching for
<SPAN
CLASS="PRODUCTNAME"
>PostgreSQL</SPAN
> (Teodor)
</P
></LI
><LI
><P
> Many <TT
CLASS="FILENAME"
>contrib/tsearch2</TT
> improvements (Teodor)
</P
></LI
><LI
><P
> Add double metaphone to <TT
CLASS="FILENAME"
>contrib/fuzzystrmatch</TT
> (Andrew)
</P
></LI
><LI
><P
> Allow <TT
CLASS="FILENAME"
>contrib/pg_autovacuum</TT
> to run as a Windows service (Dave Page)
</P
></LI
><LI
><P
> Add functions to <TT
CLASS="FILENAME"
>contrib/dbsize</TT
> (Andreas Pflug)
</P
></LI
><LI
><P
> Removed <TT
CLASS="FILENAME"
>contrib/pg_logger</TT
>: obsoleted by integrated logging
subprocess
</P
></LI
><LI
><P
> Removed <TT
CLASS="FILENAME"
>contrib/rserv</TT
>: obsoleted by various separate projects
</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-8-0-1.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-7-4-30.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Release 8.0.1</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 7.4.30</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
Mini Shell