<!--
-$PostgreSQL: pgsql/doc/src/sgml/release.sgml,v 1.325 2005/01/30 01:26:41 tgl Exp $
+$PostgreSQL: pgsql/doc/src/sgml/release.sgml,v 1.375 2005/09/24 20:35:21 tgl Exp $
+
+Typical markup:
+
+[A-Z][A-Z ]+[A-Z] <command>
+[A-Za-z_][A-Za-z0-9_]+() <function>
+[A-Za-z_]/[A-Za-z_]+ <filename>
+PostgreSQL <productname>
+pg_[A-Za-z0-9_] <application>
+[A-Z][A-Z] <type>, <envar>, <literal>
+&<> use &
-->
<appendix id="release">
<title>Release Notes</title>
- <sect1 id="release-8-0-1">
- <title>Release 8.0.1</title>
+ <sect1 id="release-8-1">
+ <title>Release 8.1</title>
<note>
- <title>Release date</title>
- <simpara>2005-01-31</simpara>
+ <title>Release date</title>
+ <simpara>2005-1?-??, Current as of 2005-09-15</simpara>
</note>
- <para>
- This release contains a variety of fixes from 8.0.0, including several
- security-related issues.
- </para>
-
<sect2>
- <title>Migration to version 8.0.1</title>
-
+ <title>Overview</title>
+
<para>
- A dump/restore is not required for those running 8.0.0.
+ Major changes in this release:
</para>
- </sect2>
-
- <sect2>
- <title>Changes</title>
-
-<itemizedlist>
-<listitem><para>Disallow <command>LOAD</> to non-superusers</para>
-<para>
-On platforms that will automatically execute initialization functions of a
-shared library (this includes at least Windows and ELF-based Unixen),
-<command>LOAD</> can be used to make the server execute arbitrary code.
-Thanks to NGS Software for reporting this.</para></listitem>
-<listitem><para>Check that creator of an aggregate function has the right to
-execute the specified transition functions</para>
-<para>
-This oversight made it possible to bypass denial of EXECUTE
-permission on a function.</para></listitem>
-<listitem><para>Fix security and 64-bit issues in
-contrib/intagg</para></listitem>
-<listitem><para>Add needed STRICT marking to some contrib functions (Kris
-Jurka)</para></listitem>
-<listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
-many parameters (Neil)</para></listitem>
-<listitem><para>Make <command>ALTER TABLE ADD COLUMN</> enforce domain
-constraints in all cases</para></listitem>
-<listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
-<para>
-The result of the join was mistakenly supposed to be sorted the same as the
-left input. This could not only deliver mis-sorted output to the user, but
-in case of nested merge joins could give outright wrong answers.
-</para></listitem>
-<listitem><para>Improve planning of grouped aggregate queries</para></listitem>
-<listitem><para><command>ROLLBACK TO <replaceable>savepoint</></command>
-closes cursors created since the savepoint</para></listitem>
-<listitem><para>Fix inadequate backend stack size on Windows</para></listitem>
-<listitem><para>Avoid SHGetSpecialFolderPath() on Windows
-(Magnus)</para></listitem>
-<listitem><para>Fix some problems in running pg_autovacuum as a Windows
-service (Dave Page)</para></listitem>
-<listitem><para>Multiple minor bug fixes in
-pg_dump/pg_restore</para></listitem>
-<listitem><para>Fix ecpg segfault with named structs used in
-typedefs (Michael)</para></listitem>
-</itemizedlist>
-
- </sect2>
- </sect1>
-
- <sect1 id="release-8-0">
- <title>Release 8.0</title>
-
- <note>
- <title>Release date</title>
- <simpara>2005-01-19</simpara>
- </note>
-
- <sect2>
- <title>Overview</title>
-
- <para>
- Major changes in this release:
<variablelist>
+
<varlistentry>
<term>
- Microsoft Windows Native Server
+ Improve concurrent access to the shared buffer cache (Tom)
</term>
<listitem>
<para>
- This is the first <productname>PostgreSQL</productname> release
- to run natively on <trademark class=registered>Microsoft Windows</> as
- a server. It can run as a <productname>Windows</> service. This
- release supports NT-based Windows releases like
- <productname>Windows 2000</>, <productname>Windows XP</>, and
- <productname>Windows 2003</>. Older releases like
- <productname>Windows 95</>, <productname>Windows 98</>, and
- <productname>Windows ME</> are not supported because these operating
- systems do not have the infrastructure to support
- <productname>PostgreSQL</productname>. A separate installer
- project has been created to ease installation on
- <productname>Windows</> — see <ulink
- url="http://pgfoundry.org/projects/pginstaller">
- http://pgfoundry.org/projects/pginstaller</ulink>.
+ Access to the shared buffer cache was identified as a
+ significant scalability problem, particularly on multi-CPU
+ systems. In this release, the way that locking is done in the
+ buffer manager has been overhauled to reduce lock contention
+ and improve scalability. The buffer manager has also been
+ changed to use a <quote>clock sweep</quote> replacement
+ policy.
</para>
+ </listitem>
+ </varlistentry>
- <para>
- Although tested throughout our release cycle, the Windows port
- does not have the benefit of years of use in production
- environments that <productname>PostgreSQL</productname> has on
- Unix platforms. Therefore it should be treated with the same
- level of caution as you would a new product.
- </para>
+ <varlistentry>
+ <term>
+ Allow index scans to use an intermediate in-memory bitmap (Tom)
+ </term>
+ <listitem>
<para>
- Previous releases required the Unix emulation toolkit
- <productname>Cygwin</> in order to run the server on Windows
- operating systems. <productname>PostgreSQL</productname> has
- supported native clients on Windows for many years.
+ In previous releases, only a single index could be used to do
+ lookups on a table. With this feature, if a query has
+ <command>WHERE tab.col1 = 4 and tab.col2 = 9</>, and there is
+ no multicolumn index on <literal>col1</> and <literal>col2</>,
+ but there is an index on <literal>col1</> and another on
+ <literal>col2</>, it is possible to search both indexes and
+ combine the results in memory, then do heap fetches for only
+ the rows matching both the <literal>col1</> and
+ <literal>col2</> restrictions. This is very useful in
+ environments that have a lot of unstructured queries where it
+ is impossible to create indexes that match all possible access
+ conditions. Bitmap scans are useful even with a single index,
+ as they reduce the amount of random access needed; a bitmap
+ index scan is efficient for retrieving fairly large fractions
+ of the complete table, whereas plain index scans are not.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- Savepoints
+ Add two-phase commit (Heikki Linnakangas, Alvaro, Tom)
</term>
-
+
<listitem>
<para>
- 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.
+ Two-phase commit allows transactions to be "prepared" on several
+ computers, and once all computers have successfully prepared
+ their transactions (none failed), all transactions can be
+ committed. Even if a machine crashes after a prepare, the
+ prepared transaction can be committed after the machine is
+ restarted. New syntax includes <command>PREPARE TRANSACTION</> and
+ <command>COMMIT/ROLLBACK PREPARED</>. A new system view
+ <literal>pg_prepared_xacts</> has also been added.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- Point-In-Time Recovery
+ Create a new role system that replaces users and groups
+ (Stephen Frost)
</term>
<listitem>
<para>
- 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.
+ Roles are a combination of users and groups. Like users, they
+ can have login capability, and like groups, a role can have
+ other roles as members. Roles basically remove the distinction
+ between users and groups. For example, a role can:
+ </para>
+
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Have login capability (optionally)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Own objects
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Hold access permissions for database objects
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Inherit permissions from other roles it is a member of
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ <para>
+ Once a user logs into a role, she obtains capabilities of
+ the login role plus any inherited roles, and can use
+ <command>SET ROLE</> to switch to other roles she is a member of.
+ This feature is a generalization of the SQL standard's concept of
+ roles.
+ This change also replaces <structname>pg_shadow</> and
+ <structname>pg_group</> by new role-capable catalogs
+ <structname>pg_authid</> and <structname>pg_auth_members</>. The old
+ tables are redefined as read-only views on the new role tables.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- Tablespaces
+ Automatically use indexes for <function>MIN()</> and
+ <function>MAX()</> (Tom)
</term>
-
+
<listitem>
<para>
- 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 <application>initlocation</> and
- manual symlink management for such tasks.
- </para>
+ In previous releases, the only way to use an index for
+ <function>MIN()</> or <function>MAX()</> was to rewrite the
+ query as <command>SELECT col FROM tab ORDER BY col LIMIT 1</>.
+ Index usage now happens automatically.
+ </para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- Improved Buffer Management, <command>CHECKPOINT</command>,
- <command>VACUUM</command>
+ Move <filename>/contrib/pg_autovacuum</> into the main server
+ (Alvaro)
</term>
<listitem>
<para>
- 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.
+ Integrating autovacuum into the server allows it to be
+ automatically started and stopped in sync with the database
+ server, and allows autovacuum to be configured from
+ <filename>postgresql.conf</>.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- Change Column Types
+ Add shared row level locks using <command>SELECT ... FOR SHARE</>
+ (Alvaro)
</term>
<listitem>
<para>
- A column's data type can now be changed with <command>ALTER
- TABLE</command>.
+ While <productname>PostgreSQL</productname>'s MVCC locking
+ allows <command>SELECT</> to never be blocked by writers and
+ therefore does not need shared row locks for typical operations,
+ shared locks are useful for applications that require shared row
+ locking. In particular this reduces the locking requirements
+ imposed by referential integrity checks.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- New Perl Server-Side Language
+ Add dependencies on shared objects, specifically roles
+ (Alvaro)
</term>
<listitem>
<para>
- A new version of the <application>plperl</> server-side language now
- supports a persistent shared storage area, triggers, returning records
- and arrays of records, and SPI calls to access the database.
+ This extension of the dependency mechanism prevents roles from
+ being dropped while there are still database objects they own.
+ Formerly it was possible to accidentally <quote>orphan</> objects by
+ deleting their owner. While this could be recovered from, it
+ was messy and unpleasant.
</para>
</listitem>
</varlistentry>
<varlistentry>
<term>
- Comma-separated-value (CSV) support in <command>COPY</command>
+ Improve performance for partitioned tables (Simon)
</term>
<listitem>
<para>
- <command>COPY</command> can now read and write
- comma-separated-value files. It has the flexibility to
- interpret non-standard quoting and separation characters too.
+ The new <varname>constraint_exclusion</varname> configuration
+ parameter avoids lookups on child tables where constraints indicate
+ that no matching rows exist in the child table.
+ </para>
+ <para>
+ This allows for a basic type of table partitioning. If child tables
+ store separate key ranges and this is enforced using appropriate
+ <command>CHECK</> constraints, the optimizer will skip child
+ table accesses when the constraint guarantees no matching rows
+ exist in the child table.
</para>
</listitem>
</varlistentry>
</variablelist>
- </para>
- </sect2>
+ </sect2>
+
+ <sect2>
+ <title>Migration to version 8.1</title>
+
+ <para>
+ A dump/restore using <application>pg_dump</application> is required
+ for those wishing to migrate data from any previous release.
+ </para>
+
+ <para>
+ The 8.0 release announced that the <function>to_char()</> function
+ for intervals would be removed in 8.1. However, since no better API
+ has been suggested, <function>to_char(interval)</> has been enhanced in
+ 8.1 and will remain in the server.
+ </para>
- <sect2>
- <title>Migration to version 8.0</title>
-
- <para>
- A dump/restore using <application>pg_dump</application> is
- required for those wishing to migrate data from any previous
- release.
- </para>
-
- <para>
- Observe the following incompatibilities:
+ <para>
+ Observe the following incompatibilities:
+ </para>
<itemizedlist>
<listitem>
<para>
- In <option>READ COMMITTED</> 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.
+ <varname>add_missing_from</> is now false by default (Neil)
+ </para>
+ <para>
+ By default, we now generate an error if a table is used in a query
+ without a <command>FROM</> reference. The old behavior is still
+ available, but the parameter must be set to 'true' to obtain it.
</para>
</listitem>
<listitem>
<para>
- Functions declared <option>STABLE</> or <option>IMMUTABLE</> 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
- <command>SELECT</>.
+ Cause input of a zero-length string ('') for float4/float8/oid
+ to throw an error, rather than treating it as a zero (Neil)
+ </para>
+ <para>
+ This change is consistent with the current handling of
+ zero-length strings for integers. The schedule for this change
+ was announced in 8.0.
</para>
</listitem>
<listitem>
<para>
- Non-deferred <option>AFTER</> 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.
+ <varname>default_with_oids</> is now false by default (Neil)
+ </para>
+ <para>
+ With this option set to false, user-created tables no longer
+ have an OID column unless <command>WITH OIDS</> is specified in
+ <command>CREATE TABLE</>. Though OIDs have existed in all
+ releases of <productname>PostgreSQL</>, their use is limited
+ because they are only four bytes long and the counter is shared
+ across all installed databases. The preferred way of uniquely
+ identifying rows is via sequences and the <type>SERIAL</> type,
+ which have been supported since <productname>PostgreSQL</> 6.4.
</para>
</listitem>
<listitem>
<para>
- Server configuration parameters <varname>virtual_host</> and
- <varname>tcpip_socket</> have been replaced with a more general
- parameter <varname>listen_addresses</>. Also, the server now listens on
- <literal>localhost</> by default, which eliminates the need for the
- <literal>-i</> postmaster switch in many scenarios.
+ Add <literal>E''</> syntax so eventually ordinary strings can
+ treat backslashes literally (Bruce)
</para>
- </listitem>
+ <para>
+ Currently <productname>PostgreSQL</productname> processes a
+ backslash in a string literal as introducing a special escape sequence,
+ e.g. <literal>\n</> or <literal>\010</>.
+ While this allows easy entry of special values, it is
+ non-standard and makes porting of applications from other
+ databases more difficult. For this reason, the
+ <productname>PostgreSQL</productname> project is planning to
+ remove the special meaning of backslashes in strings. For
+ backward compatibility and for users who want special backslash
+ processing, a new string syntax has been created. This new string
+ syntax is formed by writing an <literal>E</> immediately preceding the
+ single quote that starts the string, e.g. <literal>E'hi\n'</>. While
+ this release does not change the handling of backslashes in strings, it
+ does add new configuration parameters to help users migrate applications
+ for future releases:
+ </para>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ <varname>standard_conforming_strings</> — does this release
+ treat backslashes literally in ordinary strings?
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <varname>escape_string_warning</> — warn about backslashes in
+ ordinary (non-E) strings
+ </para>
+ </listitem>
+
+ </itemizedlist>
- <listitem>
<para>
- Server configuration parameters <varname>SortMem</> and
- <varname>VacuumMem</> have been renamed to <varname>work_mem</>
- and <varname>maintenance_work_mem</> to better reflect their
- use. The original names are still supported in
- <command>SET</command> and <command>SHOW</command>.
+ The <varname>standard_conforming_strings</> value is read-only.
+ Applications can retrieve the value to know how backslashes are
+ processed. (Presence of the parameter can also be taken as an
+ indication that <literal>E''</> string syntax is supported.)
+ In a future release, <varname>standard_conforming_strings</>
+ will be true, meaning backslashes will be treated literally in
+ non-E strings. To prepare for this change, use <literal>E''</>
+ strings in places that need special backslash processing, and
+ turn on <varname>escape_string_warning</> to find additional
+ strings that need to be converted to use <literal>E''</>.
+ Also, use two single-quotes (<literal>''</>) to embed a literal
+ single-quote in a string, rather than the
+ <productname>PostgreSQL</productname>-supported syntax of
+ backslash single-quote (<literal>\'</>). The former is
+ standards-conforming and does not require the use of the
+ <literal>E''</> string syntax. You can also use the
+ <literal>$$</> string syntax, which does not treat backslashes
+ specially.
</para>
</listitem>
<listitem>
<para>
- Server configuration parameters <varname>log_pid</>,
- <varname>log_timestamp</>, and <varname>log_source_port</> have been
- replaced with a more general parameter <varname>log_line_prefix</>.
+ Make <command>REINDEX DATABASE</> reindex all indexes in the
+ database (Tom)
+ </para>
+ <para>
+ Formerly, <command>REINDEX DATABASE</> reindexed only
+ system tables. This new behavior seems more intuitive. A new
+ command <command>REINDEX SYSTEM</> provides the old functionality
+ of reindexing just the system tables.
</para>
</listitem>
<listitem>
<para>
- Server configuration parameter <varname>syslog</> has been
- replaced with a more logical <varname>log_destination</> variable to
- control the log output destination.
+ Read-only large object descriptors now obey MVCC snapshot semantics
+ </para>
+ <para>
+ When a large object is opened with <literal>INV_READ</> (and not
+ <literal>INV_WRITE</>), the data read from the descriptor will now
+ reflect a <quote>snapshot</> of the large object's state at the
+ time of the transaction snapshot in use by the query that called
+ <function>lo_open()</>. To obtain the old behavior of always
+ returning the latest committed data, include <literal>INV_WRITE</>
+ in the mode flags for <function>lo_open()</>.
</para>
</listitem>
<listitem>
<para>
- Server configuration parameter <varname>log_statement</> has been
- changed so it can selectively log just database modification or
- data definition statements. Server configuration parameter
- <varname>log_duration</> now prints only when <varname>log_statement</>
- prints the query.
+ In <application>psql</application>, treat unquoted
+ <literal>\{digit}+</> sequences as octal (Bruce)
+ </para>
+ <para>
+ In previous releases, <literal>\{digit}+</> sequences were
+ treated as decimal, and only <literal>\0{digit}+</> were treated
+ as octal. This change was made for consistency.
</para>
</listitem>
<listitem>
<para>
- Server configuration parameter <varname>max_expr_depth</> parameter has
- been replaced with <varname>max_stack_depth</> 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.
+ Remove grammar productions for prefix and postfix <literal>%</>
+ and <literal>^</> operators
+ (Tom)
+ </para>
+ <para>
+ These have never been documented and complicated the use of the
+ modulus operator (<literal>%</>) with negative numbers.
</para>
</listitem>
<listitem>
<para>
- The <function>length()</> function no longer counts trailing spaces in
- <type>CHAR(n)</> values.
+ Make <literal>&<</> and <literal>&></> for polygons
+ consistent with the box "over" operators (Tom)
</para>
</listitem>
- <listitem>
- <para>
- Casting an integer to <type>BIT(N)</> selects the rightmost N bits of the
- integer, not the leftmost N bits as before.
- </para>
- </listitem>
-
- <listitem>
- <para>
- Updating an element or slice of a NULL array value now produces
- a non-NULL array result, namely an array containing
- just the assigned-to positions.
- </para>
- </listitem>
-
<listitem>
<para>
- 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 <literal>""</>, 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).
+ <command>CREATE LANGUAGE</> may ignore the provided arguments
+ in favor of information from <structname>pg_pltemplate</>
+ (Tom)
</para>
- </listitem>
-
- <listitem>
<para>
- Overflow in integer arithmetic operations is now detected and
- reported as an error.
+ A new system catalog <structname>pg_pltemplate</> has been defined
+ to carry information about the preferred definitions of procedural
+ languages (such as whether they have validator functions). When
+ an entry exists in this catalog for the language being created,
+ <command>CREATE LANGUAGE</> will ignore all its parameters except the
+ language name and instead use the catalog information. This measure
+ was taken because of increasing problems with obsolete language
+ definitions being loaded by old dump files. As of 8.1,
+ <application>pg_dump</> will dump procedural language definitions as
+ just <command>CREATE LANGUAGE <replaceable>name</></command>, relying
+ on a template entry to exist at load time. We expect this will be a
+ more future-proof representation.
</para>
</listitem>
<listitem>
<para>
- The arithmetic operators associated with the single-byte
- <type>"char"</> data type have been removed.
+ Make <function>pg_cancel_backend(int)</function> return a
+ <type>boolean</type> rather than an <type>integer</type> (Neil)
</para>
</listitem>
+ </itemizedlist>
+ </sect2>
+
+ <sect2>
+ <title>Additional Changes</title>
+
+ <para>
+ Below you will find a detailed account of the additional changes
+ between <productname>PostgreSQL</productname> 8.1 and the
+ previous major release.
+ </para>
+
+ <sect3>
+ <title>Performance Improvements</title>
+ <itemizedlist>
- <listitem>
- <para>
- The <function>extract()</> function (also called
- <function>date_part</>) 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Improve GiST and rtree index performance (Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- <type>CIDR</> values now must have their non-masked bits be zero.
- For example, we no longer allow
- <literal>204.248.199.1/31</literal> as a <type>CIDR</> value. Such
- values should never have been accepted by
- <productname>PostgreSQL</productname> and will now be rejected.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Improve the optimizer, including auto-resizing of hash joins
+ (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- <command>EXECUTE</command> now returns a completion tag that
- matches the executed statement.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Overhaul internal API in several areas
+ </para>
+ </listitem>
- <listitem>
- <para>
- <application>psql</>'s <command>\copy</> command now reads or
- writes to the query's <literal>stdin/stdout</>, rather than
- <application>psql</>'s <literal>stdin/stdout</>. The previous
- behavior can be accessed via new
- <option>pstdin</>/<option>pstdout</> parameters.
- </para>
- </listitem>
+ <listitem>
+ Change WAL record CRCs from 64-bit to 32-bit (Tom)
+ </para>
+ <para>
+ We determined that the extra cost of computing 64-bit CRCs was
+ significant, and the gain in reliability too marginal to justify it.
+ </para>
+ </listitem>
- <listitem>
- <para>
- The JDBC client interface has been removed from the core
- distribution, and is now hosted at <ulink url=
- "http://jdbc.postgresql.org">http://jdbc.postgresql.org</ulink>.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Prevent writing large empty gaps in WAL pages (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- The Tcl client interface has also been removed. There are several
- Tcl interfaces now hosted at <ulink url=
- "http://gborg.postgresql.org">http://gborg.postgresql.org</ulink>.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow non-consecutive index columns to be used in a multi-column
+ index (Tom)
+ </para>
+ <para>
+ For example, this allows an index on columns a,b,c to be used in
+ a query with <command>WHERE a = 4 and c = 10</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- 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 <command>SET</>/<command>SHOW</>
- <varname>TimeZone</> may
- be different from what your platform provides.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Skip WAL logging for <command>CREATE TABLE AS</> /
+ <command>SELECT INTO</> (Simon)
+ </para>
+ <para>
+ Since a crash during <command>CREATE TABLE AS</> would cause the
+ table to be dropped during recovery, there is no reason to WAL
+ log as the table is loaded. (Logging still happens if WAL
+ archiving is enabled, however.)
+ </para>
+ </listitem>
- <listitem>
- <para>
- <application>Configure</>'s threading option no longer requires
- users to run tests or edit configuration files; threading options
- are now detected automatically.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow concurrent GIST index access (Teodor, Oleg)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Now that tablespaces have been implemented,
- <application>initlocation</> has been removed.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add configuration parameter <varname>full_page_writes</> to
+ control writing full pages to WAL (Bruce)
+ </para>
+ <para>
+ To prevent partial disk writes from corrupting the database,
+ <productname>PostgreSQL</productname> writes a complete copy of
+ each database disk page to WAL the first time it is modified
+ after a checkpoint. This option turns off that functionality for more
+ speed. This is safe to use with battery-backed disk caches where
+ partial page writes cannot happen.
+ </para>
+ </listitem>
- </itemizedlist>
- </para>
- </sect2>
+ <listitem>
+ <para>
+ Use <literal>O_DIRECT</> if available when using
+ <literal>O_SYNC</> for <varname>wal_sync_method</varname>
+ (Itagaki Takahiro)
+ </para>
+ <para>
+ <literal>O_DIRECT</> causes disk writes to bypass the kernel
+ cache, and for WAL writes, this improves performance.
+ </para>
+ </listitem>
- <sect2>
- <title>Deprecated Features</title>
+ <listitem>
+ <para>
+ Improve <command>COPY FROM</> performance (Alon Goldshuv)
+ </para>
+ <para>
+ This was accomplished by reading <command>COPY</> input in
+ larger chunks, rather than character by character.
+ </para>
+ </listitem>
- <para>
- Some aspects of <productname>PostgreSQL</productname>'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.
- </para>
+ <listitem>
+ <para>
+ Improve the performance of <function>COUNT()</function>,
+ <function>SUM</function>, <function>AVG()</function>,
+ <function>STDDEV()</function>, and
+ <function>VARIANCE()</function> (Neil, Tom)
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
- <itemizedlist>
- <listitem>
- <para>
- The 8.1 release will remove the function
- <function>to_char(interval, text)</>.
- </para>
- </listitem>
+ <sect3>
+ <title>Server Changes</title>
+ <itemizedlist>
- <listitem>
- <para>
- The server now warns of empty strings passed to
- <type>oid</type>/<type>float4</type>/<type>float8</type> 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Prevent problems due to transaction ID (XID) wraparound (Tom)
+ </para>
+ <para>
+ The server will now warn when the transaction counter approaches
+ the wraparound point. If the counter becomes too close to wraparound,
+ the server will stop accepting queries. This ensures that data is
+ not lost before needed vacuuming is performed.
+ </para>
+ </listitem>
- <listitem>
- <para>
- By default, tables in <productname>PostgreSQL</productname> 8.0
- and earlier are created with <type>OID</>s. In the next release,
- this will <emphasis>not</emphasis> be the case: to create a table
- that contains <type>OID</>s, the <option>WITH OIDS</> clause must
- be specified or the <varname>default_with_oids</varname>
- configuration parameter must be set. Users are encouraged to
- explicitly specify <option>WITH OIDS</> if their tables
- require OIDs for compatibility with future releases of
- <productname>PostgreSQL</productname>.
- </para>
- </listitem>
- </itemizedlist>
- </sect2>
+ <listitem>
+ <para>
+ Fix problems with object IDs (OIDs) conflicting with existing system
+ objects after the OID counter has wrapped around (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add warning about the need to increase
+ <varname>max_fsm_relations</> and <varname>max_fsm_pages</>
+ during <command>VACUUM</> (Ron Mayer)
+ </para>
+ </listitem>
- <sect2>
- <title>Changes</title>
+ <listitem>
+ <para>
+ Add <varname>temp_buffers</> configuration parameter to allow
+ users to determine the size of the local buffer area for
+ temporary table access (Tom)
+ </para>
+ </listitem>
- <para>
- Below you will find a detailed account of the changes between
- release 8.0 and the previous major release.
- </para>
+ <listitem>
+ <para>
+ Add session start time and client IP address to
+ <literal>pg_stat_activity</> (Magnus)
+ </para>
+ </listitem>
- <sect3>
- <title>Performance Improvements</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ Enhance <literal>pg_locks</> display (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Support cross-data-type index usage (Tom)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Log queries for client-side <command>PREPARE</> and
+ <command>EXECUTE</> (Simon)
+ </para>
+ </listitem>
- <listitem>
- <para>
- New buffer replacement strategy that improves caching (Jan)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow Kerberos name and user name case sensitivity to be
+ specified in <filename>postgresql.conf</> (Magnus)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add subprocess to write dirty buffers periodically to reduce
- checkpoint writes (Jan)
- </para>
- <para>
- 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
- <function>sync()</> call, but instead <function>fsync()</>s just
- the files written since the last checkpoint. This should improve
- performance and minimize degradation during checkpoints.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add configuration parameter <varname>krb_server_hostname</> so
+ that the server hostname can be specified as part of service
+ principal (Todd Kover)
+ </para>
+ <para>
+ If not set, any service principal matching an entry in the
+ keytab can be used. This is new Kerberos matching behavior in
+ this release.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add ability to prolong vacuum to reduce performance impact (Jan)
- </para>
- <para>
- On busy systems, <command>VACUUM</command> performs many I/O
- requests which can hurt performance for other users. This
- release allows you to slow down <command>VACUUM</command> to
- reduce its impact on other users, though this increases the
- total duration of <command>VACUUM</command>.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <varname>log_line_prefix</> options for millisecond
+ timestamps (<literal>%m</>) and remote host (<literal>%h</>) (Ed
+ L.)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
- </para>
- <para>
- This improves the way indexes are scanned when many duplicate
- values exist in the index.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add WAL logging for GIST indexes (Teodor, Oleg)
+ </para>
+ <para>
+ GIST indexes are now safe for crash and point-in-time recovery.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Use dynamically-generated table size estimates while planning (Tom)
- </para>
- <para>
- Formerly the planner estimated table sizes using the values seen
- by the last <command>VACUUM</command> or <command>ANALYZE</command>,
- 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
- <command>VACUUM</command> or <command>ANALYZE</command>. This should
- produce more reliable estimates in cases where the table size has
- changed significantly since the last housekeeping command.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Remove old <filename>*.backup</> files when we do
+ <function>pg_stop_backup()</> (Bruce)
+ </para>
+ <para>
+ This prevents a large number of <filename>*.backup</> files from
+ existing in <filename>/pg_xlog</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Improved index usage with <literal>OR</> clauses (Tom)
- </para>
- <para>
- 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 <literal>OR</> clause.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add configuration parameters to control TCP/IP keep-alive
+ times for idle, interval, and count (Oliver Jowett)
+ </para>
- <listitem>
- <para>
- Improve matching of partial index clauses (Tom)
- </para>
- <para>
- The server is now smarter about using partial indexes in queries
- involving complex <option>WHERE</> clauses.
- </para>
- </listitem>
+ <para>
+ These values can be changed to allow more rapid detection of
+ lost client connections.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Improve performance of the GEQO optimizer (Tom)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add per-user and per-database connection limits (Petr Jelinek)
+ </para>
+ <para>
+ Using <command>ALTER USER</> and <command>ALTER DATABASE</>,
+ limits can now be enforced on the maximum number of sessions that
+ can concurrently connect as a specific user or to a specific database.
+ Setting the limit to zero disables user or database connections.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Miscellaneous optimizer improvements
- </para>
- <para>
- There is not room here to list all the minor improvements made, but
- numerous special cases work better than in prior releases.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow more than two gigabytes of shared memory and per-backend
+ work memory on 64-bit machines (Koichi Suzuki)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Improve lookup speed for C functions (Tom)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ New system catalog <structname>pg_pltemplate</> allows overriding
+ obsolete procedural-language definitions in dump files (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add type-specific <command>ANALYZE</command> statistics
- capability (Mark Cave-Ayland)
- </para>
- <para>
- This feature allows more flexibility in generating statistics
- for non-standard data types.
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Query Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add temporary views (Koju Iijima, Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- <command>ANALYZE</command> now collects statistics for
- expression indexes (Tom)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Fix <command>HAVING</> without any aggregate functions or
+ <command>GROUP BY</> so that the query returns a single group (Tom)
+ </para>
+ <para>
+ Previously, such a case would treat the <command>HAVING</>
+ clause the same as a <command>WHERE</> clause. This was not per spec.
+ </para>
+ </listitem>
- <listitem>
- <para>
- New two-stage sampling method for <command>ANALYZE</command>
- (Manfred Koizar)
- </para>
- <para>
- This gives better statistics when the density of valid rows is very
- different in different regions of a table.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <command>USING</> clause to allow additional tables to be
+ specified to <command>DELETE</> (Euler Taveira de Oliveira, Neil)
+ </para>
+ <para>
+ In prior releases, there was no clear method for specifying
+ additional tables to be used for joins in a <command>DELETE</>
+ statement. <command>UPDATE</> already has a <literal>FROM</>
+ clause for this purpose.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Speed up <command>TRUNCATE</command> (Tom)
- </para>
- <para>
- This buys back some of the performance loss observed in 7.4, while still
- keeping <command>TRUNCATE</command> transaction-safe.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add support for <literal>\x</> hex escapes in backend and ecpg
+ strings (Bruce)
+ </para>
+ <para>
+ This is just like the standard C <literal>\x</> escape syntax.
+ Octal escapes were already supported.
+ </para>
+ </listitem>
- </itemizedlist>
- </sect3>
+ <listitem>
+ <para>
+ Add <command>BETWEEN SYMMETRIC</> query syntax (Pavel Stehule)
+ </para>
+ <para>
+ This feature allows <command>BETWEEN</> comparisons without
+ requiring the first value to be less than the second. For
+ example, <command>2 BETWEEN [ASYMMETRIC] 3 AND 1</> returns
+ false, while <command>2 BETWEEN SYMMETRIC 3 AND 1</> returns
+ true. <command>BETWEEN ASYMMETRIC</> was already supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Add <command>NOWAIT</> option to <command>SELECT ... FOR
+ UPDATE/SHARE</> (Hans-Juergen Schoenig)
+ </para>
+ <para>
+ While the <varname>statement_timeout</> configuration
+ parameter allows a query taking more than a certain amount of
+ time to be cancelled, the <command>NOWAIT</> option allows a
+ query to be canceled as soon as a <command>SELECT ... FOR
+ UPDATE/SHARE</> command cannot immediately acquire a row lock.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Object Manipulation Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Track dependencies of shared objects (Alvaro)
+ </para>
+ <para>
+ <productname>PostgreSQL</productname> allows global tables
+ (users, databases, tablespaces) to reference information in
+ multiple databases. This addition adds dependency information
+ for global tables, so, for example, user ownership can be
+ tracked across databases, so a user who owns something in any
+ database can no longer be removed. Dependency tracking already
+ existed for database-local objects.
+ </para>
+ </listitem>
- <sect3>
- <title>Server Changes</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ Allow limited <command>ALTER OWNER</> commands to be performed
+ by the object owner (Stephen Frost)
+ </para>
+ <para>
+ Prior releases allowed only superusers to change object owners.
+ Now, ownership can be transferred if the user executing the command
+ owns the object and would be able to create it as the new owner
+ (that is, the user is a member of the new owning role and that role
+ has the CREATE permission that would be needed to create the object
+ afresh).
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add WAL file archiving and point-in-time recovery (Simon Riggs)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <command>ALTER</> object <command>SET SCHEMA</> capability
+ for some object types (tables, functions, types) (Bernd Helmle)
+ </para>
+ <para>
+ This allows objects to be moved to different schemas.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <command>ALTER TABLE ENABLE/DISABLE TRIGGER</command> to
+ disable triggers (Satoshi Nagayasu)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add tablespaces so admins can control disk layout (Gavin)
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Utility Command Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Allow <command>TRUNCATE</> to truncate multiple tables in a
+ single command (Alvaro)
+ </para>
+ <para>
+ Because of referential integrity checks, it is not allowed to
+ truncate a table that is part of a referential integrity
+ constraint. Using this new functionality, <command>TRUNCATE</>
+ can be used to truncate such tables, if both tables involved in
+ a referential integrity constraint are truncated in a single
+ <command>TRUNCATE</> command.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add a built-in log rotation program (Andreas Pflug)
- </para>
- <para>
- It is now possible to log server messages conveniently without
- relying on either <application>syslog</> or an external log
- rotation program.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Properly process carriage returns and line feeds in
+ <command>COPY CSV</> mode (Andrew)
+ </para>
+ <para>
+ In release 8.0, carriage returns and line feeds in <command>CSV
+ COPY TO</> were processed in an inconsistent manner. (This was
+ documented on the TODO list.)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add new read-only server configuration parameters to show server
- compile-time settings: <varname>block_size</>,
- <varname>integer_datetimes</>, <varname>max_function_args</>,
- <varname>max_identifier_length</>, <varname>max_index_keys</> (Joe)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <command>COPY WITH CSV HEADER</> to allow a header line as
+ the first line in <command>COPY</> (Andrew)
+ </para>
+ <para>
+ This allows handling of the common <command>CSV</> usage of
+ placing the column names on the first line of the data file. For
+ <command>COPY TO</>, the first line contains the column names,
+ and for <command>COPY FROM</>, the first line is ignored.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Make quoting of <literal>sameuser</>, <literal>samegroup</>, and
- <literal>all</> remove special meaning of these terms in
- <filename>pg_hba.conf</> (Andrew)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ On Windows, display better sub-second precision in
+ <command>EXPLAIN ANALYZE</> (Magnus)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Use clearer IPv6 name <literal>::1/128</> for
- <literal>localhost</> in default <filename>pg_hba.conf</> (Andrew)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add trigger duration display to <command>EXPLAIN ANALYZE</>
+ (Tom)
+ </para>
+ <para>
+ Prior releases included trigger execution time as part of the
+ total execution time, but did not show it separately. It is now
+ possible to see how much time is spent in each trigger.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Use CIDR format in <filename>pg_hba.conf</> examples (Andrew)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add support for <literal>\x</> hex escapes in <command>COPY</>
+ (Sergey Ten)
+ </para>
+ <para>
+ Previous releases only supported octal escapes.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Rename server configuration parameters <varname>SortMem</> and
- <varname>VacuumMem</> to <varname>work_mem</> and
- <varname>maintenance_work_mem</> (Old names still supported) (Tom)
- </para>
- <para>
- This change was made to clarify that bulk operations such as index and
- foreign key creation use <varname>maintenance_work_mem</>, while
- <varname>work_mem</> is for workspaces used during query execution.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Make <command>SHOW ALL</> include variable descriptions
+ (Matthias Schmidt)
+ </para>
+ <para>
+ <command>SHOW</> varname still only displays the variable's
+ value and does not include the description.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow logging of session disconnections using server configuration
- <varname>log_disconnections</> (Andrew)
- </para>
- </listitem>
-
- <listitem>
- <para>
- Add new server configuration parameter <varname>log_line_prefix</> to
- allow control of information emitted in each log line (Andrew)
- </para>
- <para>
- Available information includes user name, database name, remote IP
- address, and session start time.
- </para>
- </listitem>
-
- <listitem>
- <para>
- Remove server configuration parameters <varname>log_pid</>,
- <varname>log_timestamp</>, <varname>log_source_port</>; functionality
- superseded by <varname>log_line_prefix</> (Andrew)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Make <application>initdb</application> create a new standard
+ database called <literal>postgres</>, and convert utilities to
+ use <literal>postgres</> rather than <literal>template1</> for
+ standard lookups (Dave)
+ </para>
+ <para>
+ In prior releases, <literal>template1</> was used both as a
+ default connection for utilities like
+ <application>createuser</application>, and as a template for
+ new databases. This caused <command>CREATE DATABASE</> to
+ sometimes fail, because a new database cannot be created if
+ anyone else is in the template database. With this change, the
+ default connection database is now <literal>postgres</>,
+ meaning it is much less likely someone will be using
+ <literal>template1</> during <command>CREATE DATABASE</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Replace the <varname>virtual_host</> and <varname>tcpip_socket</>
- parameters with a unified <varname>listen_addresses</> parameter
- (Andrew, Tom)
- </para>
- <para>
- <varname>virtual_host</> could only specify a single IP address to
- listen on. <varname>listen_addresses</> allows multiple addresses
- to be specified.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Create new <application>reindexdb</application> command-line
+ utility by moving <filename>/contrib/reindexdb</> into the
+ server (Euler Taveira de Oliveira)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Listen on localhost by default, which eliminates the need for the
- <option>-i</> postmaster switch in many scenarios (Andrew)
- </para>
- <para>
- Listening on localhost (<literal>127.0.0.1</>) opens no new
- security holes but allows configurations like Windows and JDBC,
- which do not support local sockets, to work without special
- adjustments.
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Data Type and Function Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add <function>MAX()</> and <function>MIN()</> aggregates for
+ array types (Koju Iijima)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Remove <varname>syslog</> server configuration parameter, and add more
- logical <varname>log_destination</> variable to control log output
- location (Magnus)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Fix <function>to_date()</> and <function>to_timestamp()</> to
+ behave reasonably when <literal>CC</> and <literal>YY</> fields
+ are both used (Karel Zak)
+ </para>
+ <para>
+ If the format specification contains <literal>CC</> and a year
+ specification is <literal>YYY</> or longer, ignore the
+ <literal>CC</>. If the year specification is <literal>YY</> or
+ shorter, interpret <literal>CC</> as the previous century.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Change server configuration parameter <varname>log_statement</> to take
- values <varname>all</>, <varname>mod</>, <varname>ddl</>, or
- <varname>none</> to select which queries are logged (Bruce)
- </para>
- <para>
- This allows administrators to log only data definition changes or
- only data modification statements.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>md5(bytea)</> (Abhijit Menon-Sen)
+ </para>
+ <para>
+ <function>md5(text)</> already existed.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Some logging-related configuration parameters could formerly be adjusted
- by ordinary users, but only in the <quote>more verbose</> direction.
- They are now treated more strictly: only superusers can set them.
- However, a superuser may use <command>ALTER USER</> 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 <literal>PGOPTIONS</>.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Fix <type>CHAR()</> to properly pad spaces to the specified
+ length when using a multiple-byte character set (Yoshiyuki
+ Asaba)
+ </para>
+ <para>
+ In prior releases, the padding of <type>CHAR()</> was incorrect
+ because it only padded to the specified number of bytes without
+ considering how many characters were stored.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow configuration files to be placed outside the data directory (mlw)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ Add support for <command>numeric ^ numeric</> based on
+ <function>power(numeric, numeric)</>
+ </para>
+ <para>
+ The function already existed, but there was no operator assigned
+ to it.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Plan prepared queries only when first executed so constants can be
- used for statistics (Oliver Jowett)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Fix <type>NUMERIC</> modulus by properly truncating the quotient
+ during computation (Bruce)
+ </para>
+ <para>
+ In previous releases, modulus for large values sometimes
+ returned negative results due to rounding of the quotient.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>DECLARE CURSOR</command> to take parameters
- (Oliver Jowett)
- </para>
- <para>
- It is now useful to issue <command>DECLARE CURSOR</command> in a
- <function>Parse</> message with parameters. The parameter values
- sent at <function>Bind</> time will be substituted into the
- execution of the cursor's query.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add a function <function>lastval()</>(Dennis Björklund)
+ </para>
+ <para>
+ <function>lastval()</> is a simplified version of
+ <function>currval()</>. It automatically determines the proper
+ sequence name based on the most recent <function>nextval()</> or
+ <function>setval()</> call performed by the current session.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Fix hash joins and aggregates of <type>inet</type> and
- <type>cidr</type> data types (Tom)
- </para>
- <para>
- Release 7.4 handled hashing of mixed <type>inet</type> and
- <type>cidr</type> values incorrectly. (This bug did not exist
- in prior releases because they wouldn't try to hash either
- data type.)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>to_timestamp(DOUBLE PRECISION)</>
+ </para>
+ <para>
+ Converts Unix seconds since 1970 to a <type>TIMESTAMP WITH
+ TIMEZONE</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Make <varname>log_duration</> print only when <varname>log_statement</>
- prints the query (Ed L.)
- </para>
- </listitem>
+ <listitem>
+ Add <function>pg_postmaster_start_time()</> function (Euler
+ Taveira de Oliveira, Matthias Schmidt)
+ </para>
+ </listitem>
- </itemizedlist>
- </sect3>
+ <listitem>
+ <para>
+ Allow the full use of time zone names in <command>AT TIME
+ ZONE</>, not just the short list previously available (Magnus)
+ </para>
+ <para>
+ Previously, only a predefined list of time zone names were
+ supported by <command>AT TIME ZONE</>. Now any supported time
+ zone name can be used, e.g.:
+ <programlisting>
+ SELECT CURRENT_TIMESTAMP AT TIME ZONE 'Europe/London';
+ </programlisting>
+ In the above query, the time zone used is adjusted based on the
+ daylight savings time rules that were in effect on the supplied
+ date.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Add <function>GREATEST()</> and <function>LEAST()</> variadic
+ functions (Pavel Stehule)
+ </para>
+ <para>
+ These functions take a variable number of arguments and return
+ the greatest or least value among the arguments.
+ </para>
+ </listitem>
- <sect3>
- <title>Query Changes</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ Add <function>pg_column_size()</> (Mark Kirkwood)
+ </para>
+ <para>
+ This returns storage size of a column, including any
+ compression used.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add savepoints (nested transactions) (Alvaro)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>regexp_replace()</> (Atsushi Ogawa)
+ </para>
+ <para>
+ This allows regular expression replacement, like sed. An optional
+ flag argument allows selection of global (replace all) and
+ case-insensitive modes.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Unsupported isolation levels are now accepted and promoted to the
- nearest supported level (Peter)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Fix interval division and multiplication (Bruce)
+ </para>
+ <para>
+ Previous versions sometimes returned unjustified results, like
+ <command>'4 months'::interval / 5</> returning <command>'1 mon
+ -6 days'</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>BEGIN WORK</command> to specify transaction
- isolation levels like <command>START TRANSACTION</command> does
- (Bruce)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add an internal day field to <type>INTERVAL</> so a one day
+ interval can be distinguished from a 24 hour interval (Michael
+ Glaesemann)
+ </para>
+ <para>
+ Days that contain a daylight savings time adjustment are not 24
+ hours, but typically 23 or 25 hours. This change allows days
+ (not fixed 24-hour periods) to be added to dates who's result
+ includes a daylight savings time adjustment period. Therefore,
+ while in previous releases <literal>1 day</> and <literal>24
+ hours</> were interchangeable interval values, in this release
+ they are treated differently, e.g.
+ <programlisting>
+ '2005-05-03 00:00:00 EST' + '1 day' = '2005-05-04 00:00:00-04'
+ '2005-05-03 00:00:00 EST' + '24 hours' = '2005-05-04 01:00:00-04'
+ </programlisting>
+ </para>
+ </listitem>
- <listitem>
- <para>
- Fix table permission checking for cases in which rules generate
- a query type different from the originally submitted query (Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>justify_days()</> and <function>justify_hours()</>
+ (Michael Glaesemann)
+ </para>
+ <para>
+ These functions, respectively, adjust days to an appropriate
+ number of full months and days, and adjust hours to an
+ appropriate number of full days and hours.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
- David Fetter)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Move <filename>/contrib/dbsize</> into the backend, and rename
+ some of the functions (Dave Page, Andreas Pflug)
+ </para>
+ <para>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ <function>pg_tablespace_size()</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>pg_database_size()</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>pg_relation_size()</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>pg_total_relation_size()</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>pg_size_pretty()</>
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </para>
+ <para>
+ <function>pg_total_relation_size()</> includes indexes and TOAST
+ tables.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Make <literal>CASE val WHEN compval1 THEN ...</> evaluate <literal>val</> only once (Tom)
- </para>
- <para>
- <option>CASE</> no longer evaluates the tested expression multiple
- times. This has benefits when the expression is complex or is
- volatile.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add functions for read-only file access to the cluster directory
+ (Dave Page, Andreas Pflug)
+ </para>
+ <para>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ <function>pg_stat_file()</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>pg_read_file()</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>pg_ls_dir()</>
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>pg_reload_conf()</> to force reloading of the
+ configuration files (Dave Page, Andreas Pflug)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Test <option>HAVING</> before computing target list of an
- aggregate query (Tom)
- </para>
- <para>
- Fixes improper failure of cases such as <literal>SELECT SUM(win)/SUM(lose)
- ... GROUP BY ... HAVING SUM(lose) > 0</>. This should work but formerly
- could fail with divide-by-zero.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>pg_rotate_logfile()</> to force rotation of the
+ server log file (Dave Page, Andreas Pflug)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Replace <varname>max_expr_depth</> parameter with
- <varname>max_stack_depth</> parameter, measured in kilobytes of stack
- size (Tom)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Change <literal>pg_stat_*</> views to include TOAST tables (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow arbitrary row expressions (Tom)
- </para>
- <para>
- 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.
- </para>
- </listitem>
-
- <listitem>
- <para>
- Allow <option>LIKE</>/<option>ILIKE</> to be used as the operator
- in row and subselect comparisons (Fabien Coelho)
- </para>
- </listitem>
-
- <listitem>
- <para>
- Avoid locale-specific case conversion of basic ASCII letters in
- identifiers and keywords (Tom)
- </para>
- <para>
- This solves the <quote>Turkish problem</> with mangling of words
- containing <literal>I</> and <literal>i</>. Folding of characters
- outside the 7-bit-ASCII set is still locale-aware.
- </para>
- </listitem>
-
- <listitem>
- <para>
- Improve syntax error reporting (Fabien, Tom)
- </para>
- <para>
- Syntax error reports are more useful than before.
- </para>
- </listitem>
-
- <listitem>
- <para>
- Change <command>EXECUTE</command> to return a completion tag
- matching the executed statement (Kris Jurka)
- </para>
- <para>
- Previous releases return an <command>EXECUTE</command> tag for
- any <command>EXECUTE</command> call. In this release, the tag
- returned will reflect the command executed.
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Encoding and Locale Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Rename some encodings to be more consistent and to follow
+ international standards (Bruce)
+ </para>
+ <para>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ <literal>UNICODE</> is now <literal>UTF8</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <literal>ALT</> is now <literal>WIN866</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <literal>WIN</> is now <literal>WIN1251</>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <literal>TCVN</> is now <literal>WIN1258</>
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </para>
- <listitem>
- <para>
- Avoid emitting <option>NATURAL CROSS JOIN</> in rule listings (Tom)
- </para>
- <para>
- Such a clause makes no logical sense, but in some cases the rule
- decompiler formerly produced this syntax.
- </para>
- </listitem>
+ <para>
+ The original names still work.
+ </para>
+ </listitem>
- </itemizedlist>
- </sect3>
+ <listitem>
+ <para>
+ Add support for <literal>WIN1252</> encoding (Roland Volkmann)
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Add support for 3 and 4-byte <literal>UTF8</> characters (John
+ Hansen)
+ </para>
+ <para>
+ Previously only one and two-byte <literal>UTF8</> characters
+ were supported. This is particularly important for support for
+ some Chinese characters.
+ </para>
+ </listitem>
- <sect3>
- <title>Object Manipulation Changes</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ Allow direct conversion between <literal>EUC_JP</> and
+ <literal>SJIS</> to improve performance (Atsushi Ogawa)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <command>COMMENT ON</> for casts, conversions, languages,
- operator classes, and large objects (Christopher)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow the UTF8 encoding to work on Windows (Magnus)
+ </para>
+ <para>
+ This is done by mapping UTF8 to the Windows-native UTF16
+ implementation.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add new server configuration parameter <varname>default_with_oids</> to
- control whether tables are created with <type>OID</>s by default (Neil)
- </para>
- <para>
- This allows administrators to control whether <command>CREATE
- TABLE</command> commands create tables with or without <type>OID</>
- columns by default. (Note: the current factory default setting for
- <varname>default_with_oids</> is <literal>TRUE</>, but the default
- will become <literal>FALSE</> in future releases.)
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>General Server-Side Language Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Fix <command>ALTER LANGUAGE RENAME</> (Sergey Yatskevich)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>WITH</> / <option>WITHOUT OIDS</> clause to
- <command>CREATE TABLE AS</command> (Neil)
- </para>
- </listitem>
+ <listitem>
+ Allow function characteristics, like strictness and volatility,
+ to be modified via <command>ALTER FUNCTION</> (Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>ALTER TABLE DROP COLUMN</> to drop an <type>OID</>
- column (<command>ALTER TABLE SET WITHOUT OIDS</> still works)
- (Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Increase the maximum number of function arguments to 100 (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow composite types as table columns (Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow SQL and PL/PgSQL functions to use <command>OUT</> and
+ <command>INOUT</> parameters (Tom)
+ </para>
+ <para>
+ <command>OUT</> is an alternate way for a function to return
+ values. Instead of using <command>RETURN</>, values can be
+ returned by assigning to parameters declared as <command>OUT</> or
+ <command>INOUT</>. This is notationally simpler in some cases,
+ particularly so when multiple values need to be returned.
+ While returning multiple values from a function
+ was possible in previous releases, this greatly simplifies the
+ process. (The feature will be extended to other server-side
+ languages in future releases.)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>ALTER ... ADD COLUMN</> with defaults and
- <option>NOT NULL</> constraints; works per SQL spec (Rod)
- </para>
- <para>
- It is now possible for <option>ADD COLUMN</> to create a column
- that is not initially filled with NULLs, but with a specified
- default value.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Move language handler functions into the <literal>pg_catalog</> schema
+ </para>
+ <para>
+ This makes it easier to drop the public schema if desired.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <command>ALTER COLUMN TYPE</> to change column's type (Rod)
- </para>
- <para>
- It is now possible to alter a column's data type without dropping
- and re-adding the column.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>SPI_getnspname()</function> to SPI (Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow multiple <command>ALTER</> actions in a single <command>ALTER
- TABLE</command> command (Rod)
- </para>
- <para>
- This is particularly useful for <command>ALTER</> commands that
- rewrite the table (which include <option>ALTER COLUMN TYPE</> and
- <option>ADD COLUMN</> with a default). By grouping
- <command>ALTER</> commands together, the table need be rewritten
- only once.
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+ <sect3>
+ <title>PL/PgSQL Server-Side Language Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Overhaul the memory management of PL/PgSQL functions (Neil)
+ </para>
+ <para>
+ The parsetree of each function is now stored in a separate
+ memory context. This allows this memory to be easily reclaimed
+ when it is no longer needed.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>ALTER TABLE</command> to add <type>SERIAL</type>
- columns (Tom)
- </para>
- <para>
- This falls out from the new capability of specifying defaults for new
- columns.
- </para>
- </listitem>
+ <listitem>
+ Check function syntax at <command>CREATE FUNCTION</> time,
+ rather than at runtime (Neil)
+ </para>
+ Previously, most syntax errors were reported only when the
+ function was executed.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow changing the owners of aggregates, conversions, databases,
- functions, operators, operator classes, schemas, types, and tablespaces
- (Christopher, Euler Taveira de Oliveira)
- </para>
- <para>
- Previously this required modifying the system tables directly.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow <command>OPEN</> to open non-<command>SELECT</> queries
+ like <command>EXPLAIN</> and <command>SHOW</> (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow temporary object creation to be limited to <option>SECURITY
- DEFINER</> functions (Sean Chittenden)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ No longer require functions to issue a <command>RETURN</>
+ statement (Tom)
+ </para>
+ <para>
+ This is a byproduct of the newly added <command>OUT</> and
+ <command>INOUT</> functionality. <command>RETURN</> can
+ be omitted when it is not needed to provide the function's
+ return value.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>ALTER TABLE ... SET WITHOUT CLUSTER</> (Christopher)
- </para>
- <para>
- Prior to this release, there was no way to clear an auto-cluster
- specification except to modify the system tables.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add support for an optional <command>INTO</> clause to
+ PL/PgSQL's <command>EXECUTE</> statement (Pavel Stehule, Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Constraint/Index/<type>SERIAL</> names are now
- <replaceable>table_column_type</>
- with numbers appended to guarantee uniqueness within the schema
- (Tom)
- </para>
- <para>
- The SQL specification states that such names should be unique
- within a schema.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Make <command>CREATE TABLE AS</> set <command>ROW_COUNT</> (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <function>pg_get_serial_sequence()</> to return a
- <type>SERIAL</> column's sequence name (Christopher)
- </para>
- <para>
- This allows automated scripts to reliably find the <type>SERIAL</>
- sequence name.
- </para>
- </listitem>
+ <listitem>
+ Define <literal>SQLSTATE</> and <literal>SQLERRM</> to return
+ the <literal>SQLSTATE</> and error message of the current
+ exception (Pavel Stehule, Neil)
+ </para>
+ <para>
+ These variables are only defined inside exception blocks.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Warn when primary/foreign key data type mismatch requires costly lookup
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow the parameters to the <command>RAISE</> statement to be
+ expressions (Pavel Stehule, Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- New <command>ALTER INDEX</> command to allow moving of indexes
- between tablespaces (Gavin)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add a loop <command>CONTINUE</> statement (Pavel Stehule, Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Make <command>ALTER TABLE OWNER</> change dependent sequence
- ownership too (Alvaro)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow block and loop labels (Pavel Stehule)
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>PL/Perl Server-Side Language Changes</title>
+ <itemizedlist>
- </itemizedlist>
- </sect3>
-
+ <listitem>
+ <para>
+ Allow large result sets to be returned efficiently (Abhijit
+ Menon-Sen)
+ </para>
+ <para>
+ This allows functions to use <function>return_next()</> to avoid
+ building the entire result set in memory.
+ </para>
+ </listitem>
- <sect3>
- <title>Utility Command Changes</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ Allow one-row-at-a-time retrieval of query results (Abhijit Menon-Sen)
+ </para>
+ <para>
+ This allows functions to use <function>spi_query()</> and
+ <function>spi_fetchrow()</> to avoid accumulating the entire
+ result set in memory.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>CREATE SCHEMA</command> to create triggers,
- indexes, and sequences (Neil)
- </para>
- </listitem>
+ <listitem>
+ Force PL/Perl to handle strings as <literal>UTF8</> if the
+ server encoding is <literal>UTF8</> (David Kamholz)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>ALSO</> keyword to <command>CREATE RULE</command> (Fabien
- Coelho)
- </para>
- <para>
- This allows <option>ALSO</> to be added to rule creation to contrast it with
- <option>INSTEAD</> rules.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add a validator function for PL/Perl (Andrew)
+ </para>
+ <para>
+ This allows syntax errors to be reported at definition time,
+ rather than execution time.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>NOWAIT</> option to <command>LOCK</command> (Tatsuo)
- </para>
- <para>
- This allows the <command>LOCK</command> command to fail if it
- would have to wait for the requested lock.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow PL/Perl to return a Perl array when the function returns
+ an array type (Andrew)
+ </para>
+ <para>
+ This basically maps <productname>PostgreSQL</productname> arrays
+ to Perl arrays.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <command>COPY</command> to read and write
- comma-separated-value (CSV) files (Andrew, Bruce)
- </para>
- </listitem>
+ <listitem>
+ Allow Perl non-fatal warnings to generate <command>NOTICE</>
+ messages (Andrew)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Generate error if the <command>COPY</command> delimiter and NULL
- string conflict (Bruce)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow Perl's <literal>strict</> mode to be enabled (Andrew)
+ </para>
+ </listitem>
- <listitem>
- <para>
- <command>GRANT</command>/<command>REVOKE</command> behavior
- follows the SQL spec more closely
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
- <listitem>
- <para>
- Avoid locking conflict between <command>CREATE INDEX</command>
- and <command>CHECKPOINT</command> (Tom)
- </para>
- <para>
- In 7.3 and 7.4, a long-running B-tree index build could block concurrent
- <command>CHECKPOINT</>s from completing, thereby causing WAL bloat because the
- WAL log could not be recycled.
- </para>
- </listitem>
- <listitem>
- <para>
- Database-wide <command>ANALYZE</command> does not hold locks
- across tables (Tom)
- </para>
- <para>
- 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 <command>ANALYZE</command>
- inside a transaction block (<command>BEGIN</command> block); it
- must be able to commit and start a new transaction for each
- table.
- </para>
- </listitem>
+ <sect3>
+ <title><application>psql</> Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add <command>\set ON_ERROR_ROLLBACK</> to allow statements in
+ a transaction to error without affecting the rest of the
+ transaction (Greg Sabino Mullane)
+ </para>
+ <para>
+ This is basically implemented by wrapping every statement in a
+ sub-transaction.
+ </para>
+ </listitem>
- <listitem>
- <para>
- <command>REINDEX</command> does not exclusively lock the index's
- parent table anymore
- </para>
- <para>
- The index itself is still exclusively locked, but readers of the
- table can continue if they are not using the particular index
- being rebuilt.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add support for <literal>\x</> hex strings in
+ <application>psql</> variables (Bruce)
+ </para>
+ <para>
+ Octal escapes were already supported.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Erase MD5 user passwords when a user is renamed (Bruce)
- </para>
- <para>
- <productname>PostgreSQL</productname> 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add support for <command>troff -ms</> output format (Roger
+ Leigh)
+ </para>
+ </listitem>
- <listitem>
- <para>
- New <application>pg_ctl</> <option>kill</> option for Windows (Andrew)
- </para>
- <para>
- Windows does not have a <literal>kill</> command to send signals to
- backends so this capability was added to <application>pg_ctl</>.
- </para>
- </listitem>
+ <listitem>
+ Allow the history file location to be controlled by
+ <envar>HISTFILE</> (Andreas Seltenreich)
+ </para>
+ <para>
+ This allows configuration of per-database history storage.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Information schema improvements
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Prevent <command>\x</> (expanded mode) from affecting
+ backslash-command displays (Neil)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>--pwfile</> option to
- <application>initdb</application> so the initial password can be
- set by GUI tools (Magnus)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <option>-L</> option to <application>psql</application> to
+ log sessions (Lorne Sunley)
+ </para>
+ <para>
+ This option was added because some operating systems do not have
+ simple command-line activity logging functionality.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Detect locale/encoding mismatch in
- <application>initdb</application> (Peter)
- </para>
- </listitem>
+ <listitem>
+ Make <command>\d</> show the tablespaces of indexes (Qingqing
+ Zhou)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>register</> command to <application>pg_ctl</> to
- register Windows operating system service (Dave Page)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow <application>psql</application> help (<command>\h</>) to
+ make a best guess on the proper help information (Greg Sabino
+ Mullane)
+ </para>
+ <para>
+ This allows the user to just add <command>\h</> to the front of
+ the syntax error query and get help on the supported syntax.
+ Previously any additional query text beyond the command name
+ had to be removed to use <command>\h</>.
+ </para>
+ </listitem>
- </itemizedlist>
- </sect3>
+ <listitem>
+ <para>
+ Add <command>\pset numericlocale</> to allow numbers to be
+ output in a locale-aware format (Eugen Nedelcu)
+ </para>
+ <para>
+ For example, using <literal>C</> locale <literal>100000</> would
+ be output as <literal>100,000.0</> while a European locale might
+ output this value as <literal>100.000,0</>.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Make startup banner show both server version number and
+ <application>psql</>'s version number, when they are different (Bruce)
+ </para>
+ <para>
+ Also, a warning will be shown if the server and <application>psql</>
+ are from different major releases.
+ </para>
+ </listitem>
- <sect3>
- <title>Data Type and Function Changes</title>
- <itemizedlist>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title><application>pg_dump</> Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add <option>-n</> / <option>--schema</> switch to
+ <application>pg_restore</> (Richard van den Berg)
+ </para>
+ <para>
+ This allows only objects from a specified schema to be restored.
+ </para>
+ </listitem>
- <listitem>
- <para>
- More complete support for composite types (row types) (Tom)
- </para>
- <para>
- Composite values can be used in many places where only scalar values
- worked before.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow <application>pg_dump</> to dump a consistent snapshot of
+ large objects (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Reject non-rectangular array values as erroneous (Joe)
- </para>
- <para>
- Formerly, <function>array_in</> would silently build a
- surprising result.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow <application>pg_dump</> to dump large objects even in
+ text mode (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Overflow in integer arithmetic operations is now detected (Tom)
- </para>
- </listitem>
+ <listitem>
+ Dump comments for large objects (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- The arithmetic operators associated with the single-byte
- <type>"char"</> data type have been removed.
- </para>
- <para>
- Formerly, the parser would select these operators in many situations
- where an <quote>unable to select an operator</> error would be more
- appropriate, such as <literal>null * null</>. If you actually want
- to do arithmetic on a <type>"char"</> column, you can cast it to
- integer explicitly.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <option>--encoding</> to <application>pg_dump</>
+ (Magnus Hagander)
+ </para>
+ <para>
+ This allows a database to be dumped in an encoding that is
+ different from the server's encoding. This is valuable when
+ transferring the dump to a machine with a different encoding.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Syntax checking of array input values considerably tightened up (Joe)
- </para>
- <para>
- Junk that was previously allowed in odd places with odd results
- now causes an <literal>ERROR</>, for example, non-whitespace
- after the closing right brace.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Rely on <structname>pg_pltemplate</> for procedural languages (Tom)
+ </para>
+ <para>
+ If the call handler for a procedural language is in the
+ <literal>pg_catalog</> schema, <application>pg_dump</> does not
+ dump the handler. Instead, it dumps the language using just
+ <command>CREATE LANGUAGE <replaceable>name</></command>,
+ relying on the <structname>pg_pltemplate</> catalog to provide
+ the language's creation parameters at load time.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Empty-string array element values must now be written as
- <literal>""</>, rather than writing nothing (Joe)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title><application>libpq</application> Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add a <envar>PGPASSFILE</> environment variable to specify the
+ password file's filename (Andrew)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Array element trailing whitespace is now ignored (Joe)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add <function>lo_create()</>, that is similar to
+ <function>lo_creat()</> but allows the OID of the large object
+ to be specified (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Emit array values with explicit array bounds when lower bound is not one
- (Joe)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Make <application>libpq</application> consistently return an error
+ to the client application on <function>malloc()</function>
+ failure (Neil)
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Source Code Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add spinlock support for the Itanium processor using Intel
+ compiler (Vikram Kalsi)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Accept <literal>YYYY-monthname-DD</> as a date string (Tom)
- </para>
- </listitem>
+ <listitem>
+ Add Kerberos 5 support for Windows (Magnus)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Make <function>netmask</> and <function>hostmask</> functions
- return maximum-length mask length (Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Add Chinese FAQ (laser@pgsqldb.com)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Change factorial function to return <type>numeric</type> (Gavin)
- </para>
- <para>
- Returning <type>numeric</type> allows the factorial function to
- work for a wider range of input values.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Rename Rendezvous to Bonjour to match OS/X feature renaming
+ (Bruce)
+ </para>
+ </listitem>
- <listitem>
- <para>
- <function>to_char</>/<function>to_date()</> date conversion
- improvements (Kurt Roeckx, Fabien Coelho)
- </para>
- </listitem>
+ <listitem>
+ Add support for <literal>fsync_writethrough</literal> on
+ Darwin (Chris Campbell)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Make <function>length()</> disregard trailing spaces in
- <type>CHAR(n)</> (Gavin)
- </para>
- <para>
- This change was made to improve consistency: trailing spaces are
- semantically insignificant in <type>CHAR(n)</> data, so they
- should not be counted by <function>length()</>.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Streamline the passing of information within the server, the
+ optimizer, and the lock system (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Warn about empty string being passed to
- <type>OID</>/<type>float4</>/<type>float8</> data types (Neil)
- </para>
- <para>
- 8.1 will throw an error instead.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow <application>pg_config</> to be compiled using MSVC (Andrew)
+ </para>
+ <para>
+ This is required to build DBD::Pg using <application>MSVC</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow leading or trailing whitespace in
- <type>int2</>/<type>int4</>/<type>int8</>/<type>float4</>/<type>float8</>
- input routines
- (Neil)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Remove support for Kerberos V4 (Magnus)
+ </para>
+ <para>
+ Kerberos 4 had security vulnerabilities and is no longer
+ maintained.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Better support for IEEE <literal>Infinity</> and <literal>NaN</>
- values in <type>float4</type>/<type>float8</type> (Neil)
- </para>
- <para>
- These should now work on all platforms that support IEEE-compliant
- floating point arithmetic.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Code cleanups (Coverity static analysis performed by
+ EnterpriseDB)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <option>week</> option to <function>date_trunc()</> (Robert Creager)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Modify <filename>postgresql.conf</> to use documentation defaults
+ <literal>on</>/<literal>off</> rather than
+ <literal>true</>/<literal>false</> (Bruce)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Fix <function>to_char</function> for <literal>1 BC</>
- (previously it returned <literal>1 AD</>) (Bruce)
- </para>
- </listitem>
+ <listitem>
+ Enhance <application>pg_config</> to be able to report more
+ build-time values (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Fix <function>date_part(year)</> for BC dates (previously it
- returned one less than the correct year) (Bruce)
- </para>
- </listitem>
+ <listitem>
+ Allow <application>libpq</application> to be built thread-safe
+ on Windows (Dave Page)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Fix <function>date_part()</> to return the proper millennium and
- century (Fabien Coelho)
- </para>
- <para>
- 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Allow IPv6 connections to be used on Win32 (Andrew)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <function>ceiling()</> as an alias for <function>ceil()</>,
- and <function>power()</> as an alias for <function>pow()</> for
- standards compliance (Neil)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Move private declarations from <filename>gist.h</filename> to
+ <filename>gist_private.h</filename> (Neil)
+ </para>
- <listitem>
- <para>
- Change <function>ln()</>, <function>log()</>,
- <function>power()</>, and <function>sqrt()</> to emit the correct
- <literal>SQLSTATE</> error codes for certain error conditions, as
- specified by SQL:2003 (Neil)
- </para>
- </listitem>
+ <para>
+ In previous releases, <filename>gist.h</> contained both the
+ public GiST API (intended for use by authors of GiST index
+ implementations) as well as some private declarations used by
+ the implementation of GiST itself. The latter have been moved
+ to a separate file, <filename>gist_private.h</>. Most GiST
+ index implementations should be unaffected.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <function>width_bucket()</> function as defined by SQL:2003 (Neil)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Overhaul GiST memory management (Neil)
+ </para>
- <listitem>
- <para>
- Add <function>generate_series()</> functions to simplify working
- with numeric sets (Joe)
- </para>
- </listitem>
+ <para>
+ GiST methods are now always invoked in a short-lived memory
+ context. Therefore, memory allocated via <function>palloc()</>
+ will be reclaimed automatically, so GiST index implementations
+ do not need to manually release allocated memory via
+ <function>pfree()</>.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
- <listitem>
- <para>
- Fix <function>upper/lower/initcap()</> functions to work with
- multibyte encodings (Tom)
- </para>
- </listitem>
- <listitem>
- <para>
- Add boolean and bitwise integer <option>AND</>/<option>OR</>
- aggregates (Fabien Coelho)
- </para>
- </listitem>
+ <sect3>
+ <title>Contrib Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add <filename>/contrib/pg_buffercache</> contrib module (Mark
+ Kirkwood)
+ </para>
+ <para>
+ This displays the contents of the buffer cache, for debugging and
+ performance tuning purposes.
+ </para>
+ </listitem>
- <listitem>
- <para>
- New session information functions to return network addresses for client
- and server (Sean Chittenden)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Remove <filename>/contrib/array</> because it is obsolete (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add function to determine the area of a closed path (Sean Chittenden)
- </para>
- </listitem>
+ <listitem>
+ Clean up the <filename>/contrib/lo</> module (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add function to send cancel request to other backends (Magnus)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Move <filename>/contrib/findoidjoins</> to
+ <filename>/src/tools</> (Tom)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <type>interval</> plus <type>datetime</> operators (Tom)
- </para>
- <para>
- The reverse ordering, <type>datetime</> plus <type>interval</>,
- was already supported, but both are required by the SQL standard.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Remove the <literal><<</>, <literal>>></>,
+ <literal>&<</>, and <literal>&></> operators from
+ <filename>/contrib/cube</>
+ </para>
+ <para>
+ These operators were not useful.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Casting an integer to <type>BIT(N)</> selects the rightmost N bits
- of the integer
- (Tom)
- </para>
- <para>
- In prior releases, the leftmost N bits were selected, but this was
- deemed unhelpful, not to mention inconsistent with casting from bit
- to int.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Improve <filename>/contrib/btree_gist</> (Janko Richter)
+ </para>
+ </listitem>
- <listitem>
- <para>
- Require <type>CIDR</> values to have all non-masked bits be zero
- (Kevin Brintnall)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Improve <filename>/contrib/pgcrypto</> (Marko Kreen)
+ </para>
- </itemizedlist>
- </sect3>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Implementation of OpenPGP symmetric-key and public-key encryption
+ </para>
+ <para>
+ Both RSA and Elgamal public-key algorithms are supported.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Stand alone build: include SHA256/384/512 hashes, Fortuna PRNG
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ OpenSSL build: support 3DES, use internal AES with OpenSSL < 0.9.7
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Take build parameters (OpenSSL, zlib) from <filename>configure</> result
+ </para>
+ <para>
+ There is no need to edit the <filename>Makefile</> anymore.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove support for <filename>libmhash</> and <filename>libmcrypt</>
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+
+ </sect2>
+ </sect1>
+
+ <sect1 id="release-8-0-3">
+ <title>Release 8.0.3</title>
- <sect3>
- <title>Server-Side Language Changes</title>
- <itemizedlist>
+ <note>
+ <title>Release date</title>
+ <simpara>2005-05-09</simpara>
+ </note>
- <listitem>
- <para>
- In <literal>READ COMMITTED</> 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.
- </para>
- </listitem>
+ <para>
+ This release contains a variety of fixes from 8.0.2, including several
+ security-related issues.
+ </para>
- <listitem>
- <para>
- Functions declared <literal>STABLE</> or <literal>IMMUTABLE</> 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
- <command>SELECT</>. There is a considerable performance gain from
- declaring a function <literal>STABLE</> or <literal>IMMUTABLE</>
- rather than <literal>VOLATILE</>.
- </para>
- </listitem>
+ <sect2>
+ <title>Migration to version 8.0.3</title>
- <listitem>
- <para>
- Non-deferred <option>AFTER</> 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
- non-deferred foreign key checks occur before proceeding with the
- function.
- </para>
- </listitem>
+ <para>
+ A dump/restore is not required for those running 8.0.X. However,
+ it is one possible way of handling two significant security problems
+ that have been found in the initial contents of 8.0.X system
+ catalogs. A dump/initdb/reload sequence using 8.0.3's initdb will
+ automatically correct these problems.
+ </para>
- <listitem>
- <para>
- Allow function parameters to be declared with names (Dennis Bjorklund)
- </para>
- <para>
- This allows better documentation of functions. Whether the names
- actually do anything depends on the specific function language
- being used.
- </para>
- </listitem>
+ <para>
+ The larger security problem is that the built-in character set encoding
+ conversion functions can be invoked from SQL commands by unprivileged
+ users, but the functions were not designed for such use and are not
+ secure against malicious choices of arguments. The fix involves changing
+ the declared parameter list of these functions so that they can no longer
+ be invoked from SQL commands. (This does not affect their normal use
+ by the encoding conversion machinery.)
+ </para>
- <listitem>
- <para>
- Allow PL/pgSQL parameter names to be referenced in the function (Dennis Bjorklund)
- </para>
- <para>
- This basically creates an automatic alias for each named parameter.
- </para>
- </listitem>
+ <para>
+ The lesser problem is that the <filename>contrib/tsearch2</> module
+ creates several functions that are improperly declared to return
+ <type>internal</> when they do not accept <type>internal</> arguments.
+ This breaks type safety for all functions using <type>internal</>
+ arguments.
+ </para>
- <listitem>
- <para>
- Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
- </para>
- <para>
- This allows us to catch simple syntax errors sooner.
- </para>
- </listitem>
+ <para>
+ It is strongly recommended that all installations repair these errors,
+ either by initdb or by following the manual repair procedure given
+ below. The errors at least allow unprivileged database users to crash
+ their server process, and may allow unprivileged users to gain the
+ privileges of a database superuser.
+ </para>
- <listitem>
- <para>
- More support for composite types (row and record variables) in PL/pgSQL
- </para>
- <para>
- For example, it now works to pass a rowtype variable to another function
- as a single variable.
- </para>
- </listitem>
+ <para>
+ If you wish not to do an initdb, perform the same manual repair
+ procedures shown in the <link linkend="release-7-4-8">7.4.8 release
+ notes</link>.
+ </para>
+ </sect2>
- <listitem>
- <para>
- Default values for PL/pgSQL variables can now reference previously
- declared variables
- </para>
- </listitem>
+ <sect2>
+ <title>Changes</title>
- <listitem>
- <para>
- Improve parsing of PL/pgSQL FOR loops (Tom)
- </para>
- <para>
- Parsing is now driven by presence of <literal>".."</> rather than
- data type of <option>FOR</> variable. This makes no difference for
- correct functions, but should result in more understandable error
- messages when a mistake is made.
- </para>
- </listitem>
+<itemizedlist>
+<listitem><para>Change encoding function signature to prevent
+misuse</para></listitem>
+<listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
+<type>INTERNAL</> function results</para></listitem>
+<listitem><para>Guard against incorrect second parameter to
+<function>record_out</></para></listitem>
+<listitem><para>Repair ancient race condition that allowed a transaction to be
+seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
+than for other purposes</para>
+<para>This is an extremely serious bug since it could lead to apparent
+data inconsistencies being briefly visible to applications.</para></listitem>
+<listitem><para>Repair race condition between relation extension and
+VACUUM</para>
+<para>This could theoretically have caused loss of a page's worth of
+freshly-inserted data, although the scenario seems of very low probability.
+There are no known cases of it having caused more than an Assert failure.
+</para></listitem>
+<listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
+<para>
+The comparison code was wrong in the case where the
+<literal>--enable-integer-datetimes</> configuration switch had been used.
+NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
+it will need to be <command>REINDEX</>ed after installing this update, because
+the fix corrects the sort order of column values.
+</para></listitem>
+<listitem><para>Fix <function>EXTRACT(EPOCH)</> for
+<type>TIME WITH TIME ZONE</> values</para></listitem>
+<listitem><para>Fix mis-display of negative fractional seconds in
+<type>INTERVAL</> values</para>
+<para>
+This error only occurred when the
+<literal>--enable-integer-datetimes</> configuration switch had been used.
+</para></listitem>
+<listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
+correctly (Neil)</para></listitem>
+<listitem><para>Still more 64-bit fixes for
+<filename>contrib/intagg</></para></listitem>
+<listitem><para>Prevent incorrect optimization of functions returning
+<type>RECORD</></para></listitem>
+<listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
+<listitem><para>Fix Borland makefile for libpq</para></listitem>
+<listitem><para>Fix <filename>contrib/btree_gist</> for <type>timetz</> type
+(Teodor)</para></listitem>
+<listitem><para>Make <command>pg_ctl</> check the PID found in
+<filename>postmaster.pid</> to see if it is still a live
+process</para></listitem>
+<listitem><para>Fix <command>pg_dump</>/<command>pg_restore</> problems caused
+by addition of dump timestamps</para></listitem>
+<listitem><para>Fix interaction between materializing holdable cursors and
+firing deferred triggers during transaction commit</para></listitem>
+<listitem><para>Fix memory leak in SQL functions returning pass-by-reference
+data types</para></listitem>
+</itemizedlist>
- <listitem>
- <para>
- Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
- </para>
- </listitem>
+ </sect2>
+ </sect1>
- <listitem>
- <para>
- 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 <literal>catch</>.
- Formerly, it was not possible to catch such errors.
- </para>
- </listitem>
+ <sect1 id="release-8-0-2">
+ <title>Release 8.0.2</title>
- <listitem>
- <para>
- Accept <command>ELSEIF</> in PL/pgSQL (Neil)
- </para>
- <para>
- Previously PL/pgSQL only allowed <command>ELSIF</>, but many people
- are accustomed to spelling this keyword <command>ELSEIF</>.
- </para>
- </listitem>
+ <note>
+ <title>Release date</title>
+ <simpara>2005-04-07</simpara>
+ </note>
- </itemizedlist>
- </sect3>
+ <para>
+ This release contains a variety of fixes from 8.0.1.
+ </para>
+ <sect2>
+ <title>Migration to version 8.0.2</title>
- <sect3>
- <title><application>psql</> Changes</title>
- <itemizedlist>
+ <para>
+ A dump/restore is not required for those running 8.0.*.
+ This release updates the major version number of the
+ <productname>PostgreSQL</productname> libraries, so it might be
+ necessary to re-link some user applications if they cannot
+ find the properly-numbered shared library.
+ </para>
+ </sect2>
- <listitem>
- <para>
- Improve <application>psql</> information display about database
- objects (Christopher)
- </para>
- </listitem>
+ <sect2>
+ <title>Changes</title>
- <listitem>
- <para>
- Allow <application>psql</> to display group membership in
- <command>\du</> and <command>\dg</> (Markus Bertheau)
- </para>
- </listitem>
+<itemizedlist>
+<listitem><para>Increment the major version number of all interface
+ libraries (Bruce)</para>
+<para>
+ This should have been done in 8.0.0. It is required so 7.4.X versions
+ of PostgreSQL client applications, like <application>psql</>,
+ can be used on the same machine as 8.0.X applications. This might require
+ re-linking user applications that use these libraries.
+</para></listitem>
+<listitem><para>Add Windows-only <varname>wal_sync_method</> setting of
+ <option>fsync_writethrough</> (Magnus, Bruce)</para>
+<para>
+ This setting causes <productname>PostgreSQL</productname> to write through
+ any disk-drive write cache when writing to WAL.
+ This behavior was formerly called <option>fsync</>, but was
+ renamed because it acts quite differently from <option>fsync</> on other
+ platforms.
+</para>
+</listitem>
+<listitem><para>Enable the <varname>wal_sync_method</> setting of
+ <option>open_datasync</> on Windows, and make it the default for that
+ platform (Magnus, Bruce)</para>
+<para>
+ Because the default is no longer <option>fsync_writethrough</>,
+ data loss is possible during a power failure if the disk drive has
+ write caching enabled. To turn off the write cache on Windows,
+ from the <application>Device Manager</>, choose the drive properties,
+ then <literal>Policies</>.
+</para>
+</listitem>
+<listitem><para>New cache management algorithm <acronym>2Q</> replaces
+ <acronym>ARC</> (Tom)</para>
+<para>
+ This was done to avoid a pending US patent on <acronym>ARC</>. The
+ <acronym>2Q</> code might be a few percentage points slower than
+ <acronym>ARC</> for some work loads. A better cache management algorithm
+ will appear in 8.1.
+</para></listitem>
+<listitem><para>Planner adjustments to improve behavior on freshly-created
+tables (Tom)</para></listitem>
+<listitem><para>Allow plpgsql to assign to an element of an array that is
+initially <literal>NULL</> (Tom)</para>
+<para>
+ Formerly the array would remain <literal>NULL</>, but now it becomes a
+ single-element array. The main SQL engine was changed to handle
+ <command>UPDATE</> of a null array value this way in 8.0, but the similar
+ case in plpgsql was overlooked.
+</para>
+</listitem>
+<listitem><para>Convert <literal>\r\n</> and <literal>\r</> to <literal>\n</>
+in plpython function bodies (Michael Fuhr)</para>
+<para>
+ This prevents syntax errors when plpython code is written on a Windows or
+ Mac client.
+</para>
+</listitem>
+<listitem><para>Allow SPI cursors to handle utility commands that return rows,
+such as <command>EXPLAIN</> (Tom)</para></listitem>
+<listitem><para>Fix <command>CLUSTER</> failure after <command>ALTER TABLE
+ SET WITHOUT OIDS</> (Tom)</para></listitem>
+<listitem><para>Reduce memory usage of <command>ALTER TABLE ADD COLUMN</>
+ (Neil)</para></listitem>
+<listitem><para>Fix <command>ALTER LANGUAGE RENAME</> (Tom)</para></listitem>
+<listitem><para>Document the Windows-only <literal>register</> and
+<literal>unregister</> options of <application>pg_ctl</> (Magnus)</para></listitem>
+<listitem><para>Ensure operations done during backend shutdown are counted by
+statistics collector</para>
+<para>
+ This is expected to resolve reports of <application>pg_autovacuum</>
+ not vacuuming the system catalogs often enough — it was not being
+ told about catalog deletions caused by temporary table removal during
+ backend exit.
+</para></listitem>
+<listitem><para>Change the Windows default for configuration parameter
+ <varname>log_destination</> to <option>eventlog</> (Magnus)</para>
+<para>
+ By default, a server running on Windows will now send log output to the
+ Windows event logger rather than standard error.
+</para></listitem>
+<listitem><para>Make Kerberos authentication work on Windows (Magnus)</para></listitem>
+<listitem><para>Allow <command>ALTER DATABASE RENAME</> by superusers
+ who aren't flagged as having CREATEDB privilege (Tom)</para></listitem>
+<listitem><para>Modify WAL log entries for <command>CREATE</> and
+<command>DROP DATABASE</> to not specify absolute paths (Tom)</para>
+ <para>This allows point-in-time recovery on a different machine with possibly
+ different database location. Note that <command>CREATE TABLESPACE</> still
+ poses a hazard in such situations.
+</para></listitem>
+<listitem><para>Fix crash from a backend exiting with an open transaction
+ that created a table and opened a cursor on it (Tom)</para></listitem>
+<listitem><para>Fix <function>array_map()</> so it can call PL functions
+ (Tom)</para></listitem>
+<listitem><para>Several <filename>contrib/tsearch2</> and
+<filename>contrib/btree_gist</> fixes (Teodor)
+ </para></listitem>
+<listitem><para>Fix crash of some <filename>contrib/pgcrypto</>
+ functions on some platforms (Marko Kreen)</para></listitem>
+<listitem><para>Fix <filename>contrib/intagg</> for 64-bit platforms
+ (Tom)</para></listitem>
+<listitem><para>Fix ecpg bugs in parsing of <command>CREATE</> statement
+ (Michael)</para></listitem>
+<listitem><para>Work around gcc bug on powerpc and amd64 causing problems in
+ ecpg (Christof Petig)</para></listitem>
+<listitem><para>Do not use locale-aware versions of <function>upper()</>,
+ <function>lower()</>, and <function>initcap()</> when the locale is
+ <literal>C</> (Bruce)</para>
+<para>
+ This allows these functions to work on platforms that generate errors
+ for non-7-bit data when the locale is <literal>C</>.
+</para></listitem>
+<listitem><para>Fix <function>quote_ident()</> to quote names that match keywords (Tom)</para></listitem>
+<listitem><para>Fix <function>to_date()</> to behave reasonably when
+ <literal>CC</> and <literal>YY</> fields are both used (Karel)</para></listitem>
+<listitem><para>Prevent <function>to_char(interval)</> from failing
+ when given a zero-month interval (Tom)</para></listitem>
+<listitem><para>Fix wrong week returned by <function>date_trunc('week')</>
+(Bruce)</para>
+<para>
+ <function>date_trunc('week')</>
+ returned the wrong year for the first few days of January in some years.
+</para></listitem>
+<listitem><para>Use the correct default mask length for class <literal>D</>
+ addresses in <type>INET</> data types (Tom)</para></listitem>
+</itemizedlist>
- <listitem>
- <para>
- Prevent <application>psql</> <command>\dn</command> from showing
- temporary schemas (Bruce)
- </para>
- </listitem>
+ </sect2>
+ </sect1>
- <listitem>
- <para>
- Allow <application>psql</> to handle tilde user expansion for file
- names (Zach Irmen)
- </para>
- </listitem>
+ <sect1 id="release-8-0-1">
+ <title>Release 8.0.1</title>
- <listitem>
- <para>
- Allow <application>psql</> to display fancy prompts, including
- color, via <application>readline</> (Reece Hart, Chet Ramey)
- </para>
- </listitem>
+ <note>
+ <title>Release date</title>
+ <simpara>2005-01-31</simpara>
+ </note>
- <listitem>
- <para>
- Make <application>psql</> <command>\copy</> match <command>COPY</command> command syntax
- fully (Tom)
- </para>
- </listitem>
+ <para>
+ This release contains a variety of fixes from 8.0.0, including several
+ security-related issues.
+ </para>
- <listitem>
- <para>
- Show the location of syntax errors (Fabien Coelho, Tom)
- </para>
- </listitem>
+ <sect2>
+ <title>Migration to version 8.0.1</title>
- <listitem>
- <para>
- Add <command>CLUSTER</command> information to <application>psql</>
- <command>\d</> display
- (Bruce)
- </para>
- </listitem>
+ <para>
+ A dump/restore is not required for those running 8.0.0.
+ </para>
+ </sect2>
- <listitem>
- <para>
- Change <application>psql</> <command>\copy stdin/stdout</> to read
- from command input/output (Bruce)
- </para>
- </listitem>
+ <sect2>
+ <title>Changes</title>
- <listitem>
- <para>
- Add <option>pstdin</>/<option>pstdout</> to read from
- <application>psql</>'s <literal>stdin</>/<literal>stdout</> (Mark
- Feit)
- </para>
- </listitem>
+<itemizedlist>
+<listitem><para>Disallow <command>LOAD</> to non-superusers</para>
+<para>
+On platforms that will automatically execute initialization functions of a
+shared library (this includes at least Windows and ELF-based Unixen),
+<command>LOAD</> can be used to make the server execute arbitrary code.
+Thanks to NGS Software for reporting this.</para></listitem>
+<listitem><para>Check that creator of an aggregate function has the right to
+execute the specified transition functions</para>
+<para>
+This oversight made it possible to bypass denial of EXECUTE
+permission on a function.</para></listitem>
+<listitem><para>Fix security and 64-bit issues in
+contrib/intagg</para></listitem>
+<listitem><para>Add needed STRICT marking to some contrib functions (Kris
+Jurka)</para></listitem>
+<listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
+many parameters (Neil)</para></listitem>
+<listitem><para>Make <command>ALTER TABLE ADD COLUMN</> enforce domain
+constraints in all cases</para></listitem>
+<listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
+<para>
+The result of the join was mistakenly supposed to be sorted the same as the
+left input. This could not only deliver mis-sorted output to the user, but
+in case of nested merge joins could give outright wrong answers.
+</para></listitem>
+<listitem><para>Improve planning of grouped aggregate queries</para></listitem>
+<listitem><para><command>ROLLBACK TO <replaceable>savepoint</></command>
+closes cursors created since the savepoint</para></listitem>
+<listitem><para>Fix inadequate backend stack size on Windows</para></listitem>
+<listitem><para>Avoid SHGetSpecialFolderPath() on Windows
+(Magnus)</para></listitem>
+<listitem><para>Fix some problems in running pg_autovacuum as a Windows
+service (Dave Page)</para></listitem>
+<listitem><para>Multiple minor bug fixes in
+pg_dump/pg_restore</para></listitem>
+<listitem><para>Fix ecpg segfault with named structs used in
+typedefs (Michael)</para></listitem>
+</itemizedlist>
- <listitem>
- <para>
- Add global <application>psql</> configuration file, <filename>psqlrc.sample</filename>
- (Bruce)
- </para>
- <para>
- This allows a central file where global <application>psql</> startup commands can
- be stored.
- </para>
- </listitem>
+ </sect2>
+ </sect1>
- <listitem>
- <para>
- Have <application>psql</> <command>\d+</> indicate if the table
- has an <type>OID</> column (Neil)
- </para>
- </listitem>
+ <sect1 id="release-8-0">
+ <title>Release 8.0</title>
- <listitem>
- <para>
- On Windows, use binary mode in <application>psql</> when reading files so control-Z
- is not seen as end-of-file
- </para>
- </listitem>
+ <note>
+ <title>Release date</title>
+ <simpara>2005-01-19</simpara>
+ </note>
+
+ <sect2>
+ <title>Overview</title>
+
+ <para>
+ Major changes in this release:
+ </para>
- <listitem>
- <para>
- Have <command>\dn+</> show permissions and description for schemas (Dennis
- Bjorklund)
- </para>
- </listitem>
+ <variablelist>
+ <varlistentry>
+ <term>
+ Microsoft Windows Native Server
+ </term>
- <listitem>
- <para>
- Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ This is the first <productname>PostgreSQL</productname> release
+ to run natively on <trademark class=registered>Microsoft Windows</> as
+ a server. It can run as a <productname>Windows</> service. This
+ release supports NT-based Windows releases like
+ <productname>Windows 2000</>, <productname>Windows XP</>, and
+ <productname>Windows 2003</>. Older releases like
+ <productname>Windows 95</>, <productname>Windows 98</>, and
+ <productname>Windows ME</> are not supported because these operating
+ systems do not have the infrastructure to support
+ <productname>PostgreSQL</productname>. A separate installer
+ project has been created to ease installation on
+ <productname>Windows</> — see <ulink
+ url="http://www.postgresql.org/ftp/win32/"></ulink>.
+ </para>
- <listitem>
- <para>
- Allow boolean settings to be set using upper or lower case (Michael Paesold)
- </para>
- </listitem>
+ <para>
+ Although tested throughout our release cycle, the Windows port
+ does not have the benefit of years of use in production
+ environments that <productname>PostgreSQL</productname> has on
+ Unix platforms. Therefore it should be treated with the same
+ level of caution as you would a new product.
+ </para>
- </itemizedlist>
- </sect3>
+ <para>
+ Previous releases required the Unix emulation toolkit
+ <productname>Cygwin</> in order to run the server on Windows
+ operating systems. <productname>PostgreSQL</productname> has
+ supported native clients on Windows for many years.
+ </para>
+ </listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>
+ Savepoints
+ </term>
- <sect3>
- <title><application>pg_dump</> Changes</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ 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.
+ </para>
+ </listitem>
+ </varlistentry>
- <listitem>
- <para>
- Use dependency information to improve the reliability of
- <application>pg_dump</> (Tom)
- </para>
- <para>
- This should solve the longstanding problems with related objects
- sometimes being dumped in the wrong order.
- </para>
- </listitem>
+ <varlistentry>
+ <term>
+ Point-In-Time Recovery
+ </term>
- <listitem>
- <para>
- Have <application>pg_dump</> output objects in alphabetical order if possible (Tom)
- </para>
- <para>
- This should make it easier to identify changes between
- dump files.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ 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.
+ </para>
+ </listitem>
+ </varlistentry>
- <listitem>
- <para>
- Allow <application>pg_restore</> to ignore some SQL errors (Fabien Coelho)
- </para>
- <para>
- This makes <application>pg_restore</>'s behavior similar to the
- results of feeding a <application>pg_dump</> output script to
- <application>psql</>. 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.
- </para>
- </listitem>
+ <varlistentry>
+ <term>
+ Tablespaces
+ </term>
- <listitem>
- <para>
- <application>pg_restore</> <option>-l</> display now includes
- objects' schema names
- </para>
- </listitem>
+ <listitem>
+ <para>
+ 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 <application>initlocation</> and
+ manual symlink management for such tasks.
+ </para>
+ </listitem>
+ </varlistentry>
- <listitem>
- <para>
- New begin/end markers in <application>pg_dump</> text output (Bruce)
- </para>
- </listitem>
+ <varlistentry>
+ <term>
+ Improved Buffer Management, <command>CHECKPOINT</command>,
+ <command>VACUUM</command>
+ </term>
- <listitem>
- <para>
- Add start/stop times for
- <application>pg_dump</>/<application>pg_dumpall</> in verbose mode
- (Bruce)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ 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.
+ </para>
+ </listitem>
+ </varlistentry>
- <listitem>
- <para>
- Allow most <application>pg_dump</> options in
- <application>pg_dumpall</> (Christopher)
- </para>
- </listitem>
+ <varlistentry>
+ <term>
+ Change Column Types
+ </term>
- <listitem>
- <para>
- Have <application>pg_dump</> use <command>ALTER OWNER</> rather
- than <command>SET SESSION AUTHORIZATION</> by default
- (Christopher)
- </para>
- </listitem>
+ <listitem>
+ A column's data type can now be changed with <command>ALTER
+ TABLE</command>.
+ </para>
+ </listitem>
+ </varlistentry>
- </itemizedlist>
- </sect3>
+ <varlistentry>
+ <term>
+ New Perl Server-Side Language
+ </term>
+ <listitem>
+ <para>
+ A new version of the <application>plperl</> server-side language now
+ supports a persistent shared storage area, triggers, returning records
+ and arrays of records, and SPI calls to access the database.
+ </para>
+ </listitem>
+ </varlistentry>
- <sect3>
- <title>libpq Changes</title>
- <itemizedlist>
+ <varlistentry>
+ <term>
+ Comma-separated-value (CSV) support in <command>COPY</command>
+ </term>
- <listitem>
- <para>
- Make libpq's <option>SIGPIPE</> handling thread-safe (Bruce)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ <command>COPY</command> can now read and write
+ comma-separated-value files. It has the flexibility to
+ interpret non-standard quoting and separation characters too.
+ </para>
+ </listitem>
+ </varlistentry>
- <listitem>
- <para>
- Add <function>PQmbdsplen()</> which returns the display length
- of a character (Tatsuo)
- </para>
- </listitem>
+ </variablelist>
+ </sect2>
+
+ <sect2>
+ <title>Migration to version 8.0</title>
+
+ <para>
+ A dump/restore using <application>pg_dump</application> is
+ required for those wishing to migrate data from any previous
+ release.
+ </para>
+
+ <para>
+ Observe the following incompatibilities:
+ </para>
+
+ <itemizedlist>
- <listitem>
- <para>
- Add thread locking to <application>SSL</> and
- <application>Kerberos</> connections (Manfred Spraul)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ In <option>READ COMMITTED</> 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.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow <function>PQoidValue()</>, <function>PQcmdTuples()</>, and
- <function>PQoidStatus()</> to work on <command>EXECUTE</command>
- commands (Neil)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Functions declared <option>STABLE</> or <option>IMMUTABLE</> 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
+ <command>SELECT</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Add <function>PQserverVersion()</> to provide more convenient
- access to the server version number (Greg Sabino Mullane)
- </para>
- </listitem>
-
- <listitem>
- <para>
- Add <function>PQprepare/PQsendPrepared()</> functions to support
- preparing statements without necessarily specifying the data types
- of their parameters (Abhijit Menon-Sen)
- </para>
- </listitem>
-
- <listitem>
- <para>
- Many ECPG improvements, including <command>SET DESCRIPTOR</> (Michael)
- </para>
- </listitem>
-
- </itemizedlist>
- </sect3>
+ <listitem>
+ <para>
+ Non-deferred <option>AFTER</> 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.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Server configuration parameters <varname>virtual_host</> and
+ <varname>tcpip_socket</> have been replaced with a more general
+ parameter <varname>listen_addresses</>. Also, the server now listens on
+ <literal>localhost</> by default, which eliminates the need for the
+ <literal>-i</> postmaster switch in many scenarios.
+ </para>
+ </listitem>
- <sect3>
- <title>Source Code Changes</title>
- <itemizedlist>
+ <listitem>
+ <para>
+ Server configuration parameters <varname>SortMem</> and
+ <varname>VacuumMem</> have been renamed to <varname>work_mem</>
+ and <varname>maintenance_work_mem</> to better reflect their
+ use. The original names are still supported in
+ <command>SET</command> and <command>SHOW</command>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Server configuration parameters <varname>log_pid</>,
+ <varname>log_timestamp</>, and <varname>log_source_port</> have been
+ replaced with a more general parameter <varname>log_line_prefix</>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Shell script commands converted to C versions for Windows support (Andrew)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Server configuration parameter <varname>syslog</> has been
+ replaced with a more logical <varname>log_destination</> variable to
+ control the log output destination.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Create an extension makefile framework (Fabien Coelho, Peter)
- </para>
- <para>
- This simplifies the task of building extensions outside the original
- source tree.
- </para>
- </listitem>
+ <listitem>
+ Server configuration parameter <varname>log_statement</> has been
+ changed so it can selectively log just database modification or
+ data definition statements. Server configuration parameter
+ <varname>log_duration</> now prints only when <varname>log_statement</>
+ prints the query.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Support relocatable installations (Bruce)
- </para>
- <para>
- Directory paths for installed files (such as the
- <filename>/share</> 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.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Server configuration parameter <varname>max_expr_depth</> parameter has
+ been replaced with <varname>max_stack_depth</> 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.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Use <option>--with-docdir</> to choose installation location of documentation; also
- allow <option>--infodir</> (Peter)
- </para>
- </listitem>
+ <listitem>
+ The <function>length()</> function no longer counts trailing spaces in
+ <type>CHAR(n)</> values.
+ </para>
+ </listitem>
<listitem>
<para>
- Add <option>--without-docdir</> to prevent installation of documentation (Peter)
+ Casting an integer to <type>BIT(N)</> selects the rightmost N bits of the
+ integer, not the leftmost N bits as before.
</para>
</listitem>
<listitem>
<para>
- Upgrade to <application>DocBook</> V4.2 SGML (Peter)
+ Updating an element or slice of a NULL array value now produces
+ a non-NULL array result, namely an array containing
+ just the assigned-to positions.
</para>
</listitem>
- <listitem>
- <para>
- New <literal>PostgreSQL</> <application>CVS</> tag (Marc)
- </para>
- <para>
- This was done to make it easier for organizations to manage their
- own copies of the <productname>PostgreSQL</productname>
- <application>CVS</> repository. File version stamps from the master
- repository will not get munged by checking into or out of a copied
- repository.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ 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 <literal>""</>, 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).
+ </para>
+ </listitem>
- <listitem>
- <para>
- Clarify locking code (Manfred Koizar)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Overflow in integer arithmetic operations is now detected and
+ reported as an error.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Buffer manager cleanup (Neil)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The arithmetic operators associated with the single-byte
+ <type>"char"</> data type have been removed.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Decouple platform tests from CPU spinlock code (Bruce, Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The <function>extract()</> function (also called
+ <function>date_part</>) 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.
+ </para>
+ </listitem>
<listitem>
<para>
- Add inlined test-and-set code on PA-RISC for <application>gcc</>
- (ViSolve, Tom)
+ <type>CIDR</> values now must have their non-masked bits be zero.
+ For example, we no longer allow
+ <literal>204.248.199.1/31</literal> as a <type>CIDR</> value. Such
+ values should never have been accepted by
+ <productname>PostgreSQL</productname> and will now be rejected.
</para>
</listitem>
- <listitem>
- <para>
- Improve i386 spinlock code (Manfred Spraul)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ <command>EXECUTE</command> now returns a completion tag that
+ matches the executed statement.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Clean up spinlock assembly code to avoid warnings from newer
- <application>gcc</> releases (Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ <application>psql</>'s <command>\copy</> command now reads or
+ writes to the query's <literal>stdin/stdout</>, rather than
+ <application>psql</>'s <literal>stdin/stdout</>. The previous
+ behavior can be accessed via new
+ <option>pstdin</>/<option>pstdout</> parameters.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Remove JDBC from source tree; now a separate project
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The JDBC client interface has been removed from the core
+ distribution, and is now hosted at <ulink url=
+ "http://jdbc.postgresql.org"></ulink>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Remove the libpgtcl client interface; now a separate project
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The Tcl client interface has also been removed. There are several
+ Tcl interfaces now hosted at <ulink url=
+ "http://gborg.postgresql.org"></ulink>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- More accurately estimate memory and file descriptor usage (Tom)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ 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 <command>SET</>/<command>SHOW</>
+ <varname>TimeZone</> may
+ be different from what your platform provides.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Improvements to the Mac OS X startup scripts (Ray A.)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ <application>Configure</>'s threading option no longer requires
+ users to run tests or edit configuration files; threading options
+ are now detected automatically.
+ </para>
+ </listitem>
- <listitem>
- <para>
- New <function>fsync()</> test program (Bruce)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Now that tablespaces have been implemented,
+ <application>initlocation</> has been removed.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Major documentation improvements (Neil, Peter)
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The API for user-defined GiST indexes has been changed. The
+ Union and PickSplit methods are now passed a pointer to a
+ special <structname>GistEntryVector</structname> structure,
+ rather than a <type>bytea</type>.
+ </para>
+ </listitem>
- <listitem>
- <para>
- Remove <application>pg_encoding</application>; not needed
- anymore
- </para>
- </listitem>
+ </itemizedlist>
+ </sect2>
- <listitem>
- <para>
- Remove <application>pg_id</application>; not needed anymore
- </para>
- </listitem>
+ <sect2>
+ <title>Deprecated Features</title>
- <listitem>
- <para>
- Remove <application>initlocation</application>; not needed
- anymore
+ <para>
+ Some aspects of <productname>PostgreSQL</productname>'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.
</para>
- </listitem>
-
- <listitem>
- <para>
- Auto-detect thread flags (no more manual testing) (Bruce)
- </para>
- </listitem>
-
- <listitem>
- <para>
- Use Olson's public domain <application>timezone</> library (Magnus)
- </para>
- </listitem>
-
- <listitem>
- <para>
- With threading enabled, use thread flags on Unixware for
- backend executables too (Bruce)
- </para>
- <para>
- Unixware can not mix threaded and non-threaded object files in the
- same executable, so everything must be compiled as threaded.
- </para>
- </listitem>
-
- <listitem>
- <para>
- <application>psql</> now uses a <application>flex</>-generated
- lexical analyzer to process command strings
- </para>
- </listitem>
-
- <listitem>
- <para>
- Reimplement the linked list data structure used throughout the
- backend (Neil)
- </para>
- <para>
- This improves performance by allowing list append and length
- operations to be more efficient.
- </para>
- </listitem>
-
- <listitem>
- <para>
- Allow dynamically loaded modules to create their own server configuration
- parameters (Thomas Hallgren)
- </para>
- </listitem>
-
- <listitem>
- <para>
- New Brazilian version of FAQ (Euler Taveira de Oliveira)
- </para>
- </listitem>
-
- <listitem>
- <para>
- Add French FAQ (Guillaume Lelarge)
- </para>
- </listitem>
-
- <listitem>
- <para>
- New <application>pgevent</> for Windows logging
- </para>
- </listitem>
-
- <listitem>
- <para>
- Make libpq and ECPG build as proper shared libraries on OS X (Tom)
- </para>
- </listitem>
+
+ <itemizedlist>
+ <listitem>
+ <para>
+ The 8.1 release will remove the <function>to_char()</> function
+ for intervals.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The server now warns of empty strings passed to
+ <type>oid</type>/<type>float4</type>/<type>float8</type> 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ By default, tables in <productname>PostgreSQL</productname> 8.0
+ and earlier are created with <type>OID</>s. In the next release,
+ this will <emphasis>not</emphasis> be the case: to create a table
+ that contains <type>OID</>s, the <option>WITH OIDS</> clause must
+ be specified or the <varname>default_with_oids</varname>
+ configuration parameter must be set. Users are encouraged to
+ explicitly specify <option>WITH OIDS</> if their tables
+ require OIDs for compatibility with future releases of
+ <productname>PostgreSQL</productname>.
+ </para>
+ </listitem>
- </itemizedlist>
- </sect3>
+ </itemizedlist>
+ </sect2>
+
+ <sect2>
+ <title>Changes</title>
+
+ <para>
+ Below you will find a detailed account of the changes between
+ release 8.0 and the previous major release.
+ </para>
+
+ <sect3>
+ <title>Performance Improvements</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Support cross-data-type index usage (Tom)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New buffer replacement strategy that improves caching (Jan)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add subprocess to write dirty buffers periodically to reduce
+ checkpoint writes (Jan)
+ </para>
+ <para>
+ 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
+ <function>sync()</> call, but instead <function>fsync()</>s just
+ the files written since the last checkpoint. This should improve
+ performance and minimize degradation during checkpoints.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add ability to prolong vacuum to reduce performance impact (Jan)
+ </para>
+ <para>
+ On busy systems, <command>VACUUM</command> performs many I/O
+ requests which can hurt performance for other users. This
+ release allows you to slow down <command>VACUUM</command> to
+ reduce its impact on other users, though this increases the
+ total duration of <command>VACUUM</command>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
+ </para>
+ <para>
+ This improves the way indexes are scanned when many duplicate
+ values exist in the index.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Use dynamically-generated table size estimates while planning (Tom)
+ </para>
+ <para>
+ Formerly the planner estimated table sizes using the values seen
+ by the last <command>VACUUM</command> or <command>ANALYZE</command>,
+ 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
+ <command>VACUUM</command> or <command>ANALYZE</command>. This should
+ produce more reliable estimates in cases where the table size has
+ changed significantly since the last housekeeping command.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improved index usage with <literal>OR</> clauses (Tom)
+ </para>
+ <para>
+ 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 <literal>OR</> clause.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve matching of partial index clauses (Tom)
+ </para>
+ <para>
+ The server is now smarter about using partial indexes in queries
+ involving complex <option>WHERE</> clauses.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve performance of the GEQO optimizer (Tom)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Miscellaneous optimizer improvements
+ </para>
+ <para>
+ There is not room here to list all the minor improvements made, but
+ numerous special cases work better than in prior releases.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve lookup speed for C functions (Tom)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add type-specific <command>ANALYZE</command> statistics
+ capability (Mark Cave-Ayland)
+ </para>
+ <para>
+ This feature allows more flexibility in generating statistics
+ for non-standard data types.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <command>ANALYZE</command> now collects statistics for
+ expression indexes (Tom)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New two-stage sampling method for <command>ANALYZE</command>
+ (Manfred Koizar)
+ </para>
+ <para>
+ This gives better statistics when the density of valid rows is very
+ different in different regions of a table.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Speed up <command>TRUNCATE</command> (Tom)
+ </para>
+ <para>
+ This buys back some of the performance loss observed in 7.4, while still
+ keeping <command>TRUNCATE</command> transaction-safe.
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Server Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add WAL file archiving and point-in-time recovery (Simon Riggs)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add tablespaces so admins can control disk layout (Gavin)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add a built-in log rotation program (Andreas Pflug)
+ </para>
+ <para>
+ It is now possible to log server messages conveniently without
+ relying on either <application>syslog</> or an external log
+ rotation program.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add new read-only server configuration parameters to show server
+ compile-time settings: <varname>block_size</>,
+ <varname>integer_datetimes</>, <varname>max_function_args</>,
+ <varname>max_identifier_length</>, <varname>max_index_keys</> (Joe)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make quoting of <literal>sameuser</>, <literal>samegroup</>, and
+ <literal>all</> remove special meaning of these terms in
+ <filename>pg_hba.conf</> (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Use clearer IPv6 name <literal>::1/128</> for
+ <literal>localhost</> in default <filename>pg_hba.conf</> (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Use CIDR format in <filename>pg_hba.conf</> examples (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Rename server configuration parameters <varname>SortMem</> and
+ <varname>VacuumMem</> to <varname>work_mem</> and
+ <varname>maintenance_work_mem</> (Old names still supported) (Tom)
+ </para>
+ <para>
+ This change was made to clarify that bulk operations such as index and
+ foreign key creation use <varname>maintenance_work_mem</>, while
+ <varname>work_mem</> is for workspaces used during query execution.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow logging of session disconnections using server configuration
+ <varname>log_disconnections</> (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add new server configuration parameter <varname>log_line_prefix</> to
+ allow control of information emitted in each log line (Andrew)
+ </para>
+ <para>
+ Available information includes user name, database name, remote IP
+ address, and session start time.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove server configuration parameters <varname>log_pid</>,
+ <varname>log_timestamp</>, <varname>log_source_port</>; functionality
+ superseded by <varname>log_line_prefix</> (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Replace the <varname>virtual_host</> and <varname>tcpip_socket</>
+ parameters with a unified <varname>listen_addresses</> parameter
+ (Andrew, Tom)
+ </para>
+ <para>
+ <varname>virtual_host</> could only specify a single IP address to
+ listen on. <varname>listen_addresses</> allows multiple addresses
+ to be specified.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Listen on localhost by default, which eliminates the need for the
+ <option>-i</> postmaster switch in many scenarios (Andrew)
+ </para>
+ <para>
+ Listening on localhost (<literal>127.0.0.1</>) opens no new
+ security holes but allows configurations like Windows and JDBC,
+ which do not support local sockets, to work without special
+ adjustments.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove <varname>syslog</> server configuration parameter, and add more
+ logical <varname>log_destination</> variable to control log output
+ location (Magnus)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Change server configuration parameter <varname>log_statement</> to take
+ values <varname>all</>, <varname>mod</>, <varname>ddl</>, or
+ <varname>none</> to select which queries are logged (Bruce)
+ </para>
+ <para>
+ This allows administrators to log only data definition changes or
+ only data modification statements.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Some logging-related configuration parameters could formerly be adjusted
+ by ordinary users, but only in the <quote>more verbose</> direction.
+ They are now treated more strictly: only superusers can set them.
+ However, a superuser may use <command>ALTER USER</> 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 <literal>PGOPTIONS</>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow configuration files to be placed outside the data directory (mlw)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Plan prepared queries only when first executed so constants can be
+ used for statistics (Oliver Jowett)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <command>DECLARE CURSOR</command> to take parameters
+ (Oliver Jowett)
+ </para>
+ <para>
+ It is now useful to issue <command>DECLARE CURSOR</command> in a
+ <function>Parse</> message with parameters. The parameter values
+ sent at <function>Bind</> time will be substituted into the
+ execution of the cursor's query.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Fix hash joins and aggregates of <type>inet</type> and
+ <type>cidr</type> data types (Tom)
+ </para>
+ <para>
+ Release 7.4 handled hashing of mixed <type>inet</type> and
+ <type>cidr</type> values incorrectly. (This bug did not exist
+ in prior releases because they wouldn't try to hash either
+ data type.)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make <varname>log_duration</> print only when <varname>log_statement</>
+ prints the query (Ed L.)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Query Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add savepoints (nested transactions) (Alvaro)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Unsupported isolation levels are now accepted and promoted to the
+ nearest supported level (Peter)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <command>BEGIN WORK</command> to specify transaction
+ isolation levels like <command>START TRANSACTION</command> does
+ (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Fix table permission checking for cases in which rules generate
+ a query type different from the originally submitted query (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
+ David Fetter)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make <literal>CASE val WHEN compval1 THEN ...</> evaluate <literal>val</> only once (Tom)
+ </para>
+ <para>
+ <option>CASE</> no longer evaluates the tested expression multiple
+ times. This has benefits when the expression is complex or is
+ volatile.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Test <option>HAVING</> before computing target list of an
+ aggregate query (Tom)
+ </para>
+ <para>
+ Fixes improper failure of cases such as <literal>SELECT SUM(win)/SUM(lose)
+ ... GROUP BY ... HAVING SUM(lose) > 0</>. This should work but formerly
+ could fail with divide-by-zero.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Replace <varname>max_expr_depth</> parameter with
+ <varname>max_stack_depth</> parameter, measured in kilobytes of stack
+ size (Tom)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow arbitrary row expressions (Tom)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <option>LIKE</>/<option>ILIKE</> to be used as the operator
+ in row and subselect comparisons (Fabien Coelho)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Avoid locale-specific case conversion of basic ASCII letters in
+ identifiers and keywords (Tom)
+ </para>
+ <para>
+ This solves the <quote>Turkish problem</> with mangling of words
+ containing <literal>I</> and <literal>i</>. Folding of characters
+ outside the 7-bit-ASCII set is still locale-aware.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve syntax error reporting (Fabien, Tom)
+ </para>
+ <para>
+ Syntax error reports are more useful than before.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Change <command>EXECUTE</command> to return a completion tag
+ matching the executed statement (Kris Jurka)
+ </para>
+ <para>
+ Previous releases return an <command>EXECUTE</command> tag for
+ any <command>EXECUTE</command> call. In this release, the tag
+ returned will reflect the command executed.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Avoid emitting <option>NATURAL CROSS JOIN</> in rule listings (Tom)
+ </para>
+ <para>
+ Such a clause makes no logical sense, but in some cases the rule
+ decompiler formerly produced this syntax.
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Object Manipulation Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Add <command>COMMENT ON</> for casts, conversions, languages,
+ operator classes, and large objects (Christopher)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add new server configuration parameter <varname>default_with_oids</> to
+ control whether tables are created with <type>OID</>s by default (Neil)
+ </para>
+ <para>
+ This allows administrators to control whether <command>CREATE
+ TABLE</command> commands create tables with or without <type>OID</>
+ columns by default. (Note: the current factory default setting for
+ <varname>default_with_oids</> is <literal>TRUE</>, but the default
+ will become <literal>FALSE</> in future releases.)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>WITH</> / <option>WITHOUT OIDS</> clause to
+ <command>CREATE TABLE AS</command> (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <command>ALTER TABLE DROP COLUMN</> to drop an <type>OID</>
+ column (<command>ALTER TABLE SET WITHOUT OIDS</> still works)
+ (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow composite types as table columns (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <command>ALTER ... ADD COLUMN</> with defaults and
+ <option>NOT NULL</> constraints; works per SQL spec (Rod)
+ </para>
+ <para>
+ It is now possible for <option>ADD COLUMN</> to create a column
+ that is not initially filled with NULLs, but with a specified
+ default value.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <command>ALTER COLUMN TYPE</> to change column's type (Rod)
+ </para>
+ <para>
+ It is now possible to alter a column's data type without dropping
+ and re-adding the column.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow multiple <command>ALTER</> actions in a single <command>ALTER
+ TABLE</command> command (Rod)
+ </para>
+ <para>
+ This is particularly useful for <command>ALTER</> commands that
+ rewrite the table (which include <option>ALTER COLUMN TYPE</> and
+ <option>ADD COLUMN</> with a default). By grouping
+ <command>ALTER</> commands together, the table need be rewritten
+ only once.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <command>ALTER TABLE</command> to add <type>SERIAL</type>
+ columns (Tom)
+ </para>
+ <para>
+ This falls out from the new capability of specifying defaults for new
+ columns.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow changing the owners of aggregates, conversions, databases,
+ functions, operators, operator classes, schemas, types, and tablespaces
+ (Christopher, Euler Taveira de Oliveira)
+ </para>
+ <para>
+ Previously this required modifying the system tables directly.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow temporary object creation to be limited to <option>SECURITY
+ DEFINER</> functions (Sean Chittenden)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>ALTER TABLE ... SET WITHOUT CLUSTER</> (Christopher)
+ </para>
+ <para>
+ Prior to this release, there was no way to clear an auto-cluster
+ specification except to modify the system tables.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Constraint/Index/<type>SERIAL</> names are now
+ <replaceable>table_column_type</>
+ with numbers appended to guarantee uniqueness within the schema
+ (Tom)
+ </para>
+ <para>
+ The SQL specification states that such names should be unique
+ within a schema.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>pg_get_serial_sequence()</> to return a
+ <type>SERIAL</> column's sequence name (Christopher)
+ </para>
+ <para>
+ This allows automated scripts to reliably find the <type>SERIAL</>
+ sequence name.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Warn when primary/foreign key data type mismatch requires costly lookup
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <command>ALTER INDEX</> command to allow moving of indexes
+ between tablespaces (Gavin)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make <command>ALTER TABLE OWNER</> change dependent sequence
+ ownership too (Alvaro)
+ </para>
+ </listitem>
+
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Utility Command Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Allow <command>CREATE SCHEMA</command> to create triggers,
+ indexes, and sequences (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>ALSO</> keyword to <command>CREATE RULE</command> (Fabien
+ Coelho)
+ </para>
+ <para>
+ This allows <option>ALSO</> to be added to rule creation to contrast it with
+ <option>INSTEAD</> rules.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>NOWAIT</> option to <command>LOCK</command> (Tatsuo)
+ </para>
+ <para>
+ This allows the <command>LOCK</command> command to fail if it
+ would have to wait for the requested lock.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <command>COPY</command> to read and write
+ comma-separated-value (CSV) files (Andrew, Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Generate error if the <command>COPY</command> delimiter and NULL
+ string conflict (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <command>GRANT</command>/<command>REVOKE</command> behavior
+ follows the SQL spec more closely
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Avoid locking conflict between <command>CREATE INDEX</command>
+ and <command>CHECKPOINT</command> (Tom)
+ </para>
+ <para>
+ In 7.3 and 7.4, a long-running B-tree index build could block concurrent
+ <command>CHECKPOINT</>s from completing, thereby causing WAL bloat because the
+ WAL log could not be recycled.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Database-wide <command>ANALYZE</command> does not hold locks
+ across tables (Tom)
+ </para>
+ <para>
+ 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 <command>ANALYZE</command>
+ inside a transaction block (<command>BEGIN</command> block); it
+ must be able to commit and start a new transaction for each
+ table.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <command>REINDEX</command> does not exclusively lock the index's
+ parent table anymore
+ </para>
+ <para>
+ The index itself is still exclusively locked, but readers of the
+ table can continue if they are not using the particular index
+ being rebuilt.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Erase MD5 user passwords when a user is renamed (Bruce)
+ </para>
+ <para>
+ <productname>PostgreSQL</productname> 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <application>pg_ctl</> <option>kill</> option for Windows (Andrew)
+ </para>
+ <para>
+ Windows does not have a <literal>kill</> command to send signals to
+ backends so this capability was added to <application>pg_ctl</>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Information schema improvements
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>--pwfile</> option to
+ <application>initdb</application> so the initial password can be
+ set by GUI tools (Magnus)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Detect locale/encoding mismatch in
+ <application>initdb</application> (Peter)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>register</> command to <application>pg_ctl</> to
+ register Windows operating system service (Dave Page)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Data Type and Function Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ More complete support for composite types (row types) (Tom)
+ </para>
+ <para>
+ Composite values can be used in many places where only scalar values
+ worked before.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Reject non-rectangular array values as erroneous (Joe)
+ </para>
+ <para>
+ Formerly, <function>array_in</> would silently build a
+ surprising result.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Overflow in integer arithmetic operations is now detected (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The arithmetic operators associated with the single-byte
+ <type>"char"</> data type have been removed.
+ </para>
+ <para>
+ Formerly, the parser would select these operators in many situations
+ where an <quote>unable to select an operator</> error would be more
+ appropriate, such as <literal>null * null</>. If you actually want
+ to do arithmetic on a <type>"char"</> column, you can cast it to
+ integer explicitly.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Syntax checking of array input values considerably tightened up (Joe)
+ </para>
+ <para>
+ Junk that was previously allowed in odd places with odd results
+ now causes an <literal>ERROR</>, for example, non-whitespace
+ after the closing right brace.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Empty-string array element values must now be written as
+ <literal>""</>, rather than writing nothing (Joe)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Array element trailing whitespace is now ignored (Joe)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Emit array values with explicit array bounds when lower bound is not one
+ (Joe)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Accept <literal>YYYY-monthname-DD</> as a date string (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make <function>netmask</> and <function>hostmask</> functions
+ return maximum-length mask length (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Change factorial function to return <type>numeric</type> (Gavin)
+ </para>
+ <para>
+ Returning <type>numeric</type> allows the factorial function to
+ work for a wider range of input values.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <function>to_char</>/<function>to_date()</> date conversion
+ improvements (Kurt Roeckx, Fabien Coelho)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make <function>length()</> disregard trailing spaces in
+ <type>CHAR(n)</> (Gavin)
+ </para>
+ <para>
+ This change was made to improve consistency: trailing spaces are
+ semantically insignificant in <type>CHAR(n)</> data, so they
+ should not be counted by <function>length()</>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Warn about empty string being passed to
+ <type>OID</>/<type>float4</>/<type>float8</> data types (Neil)
+ </para>
+ <para>
+ 8.1 will throw an error instead.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow leading or trailing whitespace in
+ <type>int2</>/<type>int4</>/<type>int8</>/<type>float4</>/<type>float8</>
+ input routines
+ (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Better support for IEEE <literal>Infinity</> and <literal>NaN</>
+ values in <type>float4</type>/<type>float8</type> (Neil)
+ </para>
+ <para>
+ These should now work on all platforms that support IEEE-compliant
+ floating point arithmetic.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>week</> option to <function>date_trunc()</> (Robert Creager)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Fix <function>to_char</function> for <literal>1 BC</>
+ (previously it returned <literal>1 AD</>) (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Fix <function>date_part(year)</> for BC dates (previously it
+ returned one less than the correct year) (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Fix <function>date_part()</> to return the proper millennium and
+ century (Fabien Coelho)
+ </para>
+ <para>
+ 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>ceiling()</> as an alias for <function>ceil()</>,
+ and <function>power()</> as an alias for <function>pow()</> for
+ standards compliance (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Change <function>ln()</>, <function>log()</>,
+ <function>power()</>, and <function>sqrt()</> to emit the correct
+ <literal>SQLSTATE</> error codes for certain error conditions, as
+ specified by SQL:2003 (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>width_bucket()</> function as defined by SQL:2003 (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>generate_series()</> functions to simplify working
+ with numeric sets (Joe)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Fix <function>upper/lower/initcap()</> functions to work with
+ multibyte encodings (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add boolean and bitwise integer <option>AND</>/<option>OR</>
+ aggregates (Fabien Coelho)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New session information functions to return network addresses for client
+ and server (Sean Chittenden)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add function to determine the area of a closed path (Sean Chittenden)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add function to send cancel request to other backends (Magnus)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <type>interval</> plus <type>datetime</> operators (Tom)
+ </para>
+ <para>
+ The reverse ordering, <type>datetime</> plus <type>interval</>,
+ was already supported, but both are required by the SQL standard.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Casting an integer to <type>BIT(N)</> selects the rightmost N bits
+ of the integer
+ (Tom)
+ </para>
+ <para>
+ In prior releases, the leftmost N bits were selected, but this was
+ deemed unhelpful, not to mention inconsistent with casting from bit
+ to int.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Require <type>CIDR</> values to have all non-masked bits be zero
+ (Kevin Brintnall)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Server-Side Language Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ In <literal>READ COMMITTED</> 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Functions declared <literal>STABLE</> or <literal>IMMUTABLE</> 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
+ <command>SELECT</>. There is a considerable performance gain from
+ declaring a function <literal>STABLE</> or <literal>IMMUTABLE</>
+ rather than <literal>VOLATILE</>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Non-deferred <option>AFTER</> 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
+ non-deferred foreign key checks occur before proceeding with the
+ function.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow function parameters to be declared with names (Dennis Bjorklund)
+ </para>
+ <para>
+ This allows better documentation of functions. Whether the names
+ actually do anything depends on the specific function language
+ being used.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow PL/pgSQL parameter names to be referenced in the function (Dennis Bjorklund)
+ </para>
+ <para>
+ This basically creates an automatic alias for each named parameter.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
+ </para>
+ <para>
+ This allows us to catch simple syntax errors sooner.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ More support for composite types (row and record variables) in PL/pgSQL
+ </para>
+ <para>
+ For example, it now works to pass a rowtype variable to another function
+ as a single variable.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Default values for PL/pgSQL variables can now reference previously
+ declared variables
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve parsing of PL/pgSQL FOR loops (Tom)
+ </para>
+ <para>
+ Parsing is now driven by presence of <literal>".."</> rather than
+ data type of <option>FOR</> variable. This makes no difference for
+ correct functions, but should result in more understandable error
+ messages when a mistake is made.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ 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 <literal>catch</>.
+ Formerly, it was not possible to catch such errors.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Accept <command>ELSEIF</> in PL/pgSQL (Neil)
+ </para>
+ <para>
+ Previously PL/pgSQL only allowed <command>ELSIF</>, but many people
+ are accustomed to spelling this keyword <command>ELSEIF</>.
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title><application>psql</> Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Improve <application>psql</> information display about database
+ objects (Christopher)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <application>psql</> to display group membership in
+ <command>\du</> and <command>\dg</> (Markus Bertheau)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Prevent <application>psql</> <command>\dn</command> from showing
+ temporary schemas (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <application>psql</> to handle tilde user expansion for file
+ names (Zach Irmen)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <application>psql</> to display fancy prompts, including
+ color, via <application>readline</> (Reece Hart, Chet Ramey)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make <application>psql</> <command>\copy</> match <command>COPY</command> command syntax
+ fully (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Show the location of syntax errors (Fabien Coelho, Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <command>CLUSTER</command> information to <application>psql</>
+ <command>\d</> display
+ (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Change <application>psql</> <command>\copy stdin/stdout</> to read
+ from command input/output (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>pstdin</>/<option>pstdout</> to read from
+ <application>psql</>'s <literal>stdin</>/<literal>stdout</> (Mark
+ Feit)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add global <application>psql</> configuration file, <filename>psqlrc.sample</filename>
+ (Bruce)
+ </para>
+ <para>
+ This allows a central file where global <application>psql</> startup commands can
+ be stored.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Have <application>psql</> <command>\d+</> indicate if the table
+ has an <type>OID</> column (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ On Windows, use binary mode in <application>psql</> when reading files so control-Z
+ is not seen as end-of-file
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Have <command>\dn+</> show permissions and description for schemas (Dennis
+ Bjorklund)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow boolean settings to be set using upper or lower case (Michael Paesold)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title><application>pg_dump</> Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Use dependency information to improve the reliability of
+ <application>pg_dump</> (Tom)
+ </para>
+ <para>
+ This should solve the longstanding problems with related objects
+ sometimes being dumped in the wrong order.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Have <application>pg_dump</> output objects in alphabetical order if possible (Tom)
+ </para>
+ <para>
+ This should make it easier to identify changes between
+ dump files.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <application>pg_restore</> to ignore some SQL errors (Fabien Coelho)
+ </para>
+ <para>
+ This makes <application>pg_restore</>'s behavior similar to the
+ results of feeding a <application>pg_dump</> output script to
+ <application>psql</>. 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <application>pg_restore</> <option>-l</> display now includes
+ objects' schema names
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New begin/end markers in <application>pg_dump</> text output (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add start/stop times for
+ <application>pg_dump</>/<application>pg_dumpall</> in verbose mode
+ (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow most <application>pg_dump</> options in
+ <application>pg_dumpall</> (Christopher)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Have <application>pg_dump</> use <command>ALTER OWNER</> rather
+ than <command>SET SESSION AUTHORIZATION</> by default
+ (Christopher)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>libpq Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Make libpq's <option>SIGPIPE</> handling thread-safe (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>PQmbdsplen()</> which returns the display length
+ of a character (Tatsuo)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add thread locking to <application>SSL</> and
+ <application>Kerberos</> connections (Manfred Spraul)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <function>PQoidValue()</>, <function>PQcmdTuples()</>, and
+ <function>PQoidStatus()</> to work on <command>EXECUTE</command>
+ commands (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>PQserverVersion()</> to provide more convenient
+ access to the server version number (Greg Sabino Mullane)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <function>PQprepare/PQsendPrepared()</> functions to support
+ preparing statements without necessarily specifying the data types
+ of their parameters (Abhijit Menon-Sen)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Many ECPG improvements, including <command>SET DESCRIPTOR</> (Michael)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Source Code Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Shell script commands converted to C versions for Windows support (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Create an extension makefile framework (Fabien Coelho, Peter)
+ </para>
+ <para>
+ This simplifies the task of building extensions outside the original
+ source tree.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Support relocatable installations (Bruce)
+ </para>
+ <para>
+ Directory paths for installed files (such as the
+ <filename>/share</> 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.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Use <option>--with-docdir</> to choose installation location of documentation; also
+ allow <option>--infodir</> (Peter)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add <option>--without-docdir</> to prevent installation of documentation (Peter)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Upgrade to <application>DocBook</> V4.2 SGML (Peter)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <literal>PostgreSQL</> <application>CVS</> tag (Marc)
+ </para>
+ <para>
+ This was done to make it easier for organizations to manage their
+ own copies of the <productname>PostgreSQL</productname>
+ <application>CVS</> repository. File version stamps from the master
+ repository will not get munged by checking into or out of a copied
+ repository.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Clarify locking code (Manfred Koizar)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Buffer manager cleanup (Neil)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Decouple platform tests from CPU spinlock code (Bruce, Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add inlined test-and-set code on PA-RISC for <application>gcc</>
+ (ViSolve, Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improve i386 spinlock code (Manfred Spraul)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Clean up spinlock assembly code to avoid warnings from newer
+ <application>gcc</> releases (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove JDBC from source tree; now a separate project
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove the libpgtcl client interface; now a separate project
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ More accurately estimate memory and file descriptor usage (Tom)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Improvements to the Mac OS X startup scripts (Ray A.)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <function>fsync()</> test program (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Major documentation improvements (Neil, Peter)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove <application>pg_encoding</application>; not needed
+ anymore
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove <application>pg_id</application>; not needed anymore
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Remove <application>initlocation</application>; not needed
+ anymore
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Auto-detect thread flags (no more manual testing) (Bruce)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Use Olson's public domain <application>timezone</> library (Magnus)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ With threading enabled, use thread flags on Unixware for
+ backend executables too (Bruce)
+ </para>
+ <para>
+ Unixware cannot mix threaded and non-threaded object files in the
+ same executable, so everything must be compiled as threaded.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <application>psql</> now uses a <application>flex</>-generated
+ lexical analyzer to process command strings
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Reimplement the linked list data structure used throughout the
+ backend (Neil)
+ </para>
+ <para>
+ This improves performance by allowing list append and length
+ operations to be more efficient.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow dynamically loaded modules to create their own server configuration
+ parameters (Thomas Hallgren)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New Brazilian version of FAQ (Euler Taveira de Oliveira)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add French FAQ (Guillaume Lelarge)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <application>pgevent</> for Windows logging
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Make libpq and ECPG build as proper shared libraries on OS X (Tom)
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+
+ <sect3>
+ <title>Contrib Changes</title>
+ <itemizedlist>
+
+ <listitem>
+ <para>
+ Overhaul of <filename>contrib/dblink</> (Joe)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <filename>contrib/dbmirror</> improvements (Steven Singer)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <filename>contrib/xml2</> (John Gray, Torchbox)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Updated <filename>contrib/mysql</filename>
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New version of <filename>contrib/btree_gist</> (Teodor)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ New <filename>contrib/trgm</>, trigram matching for
+ <productname>PostgreSQL</productname> (Teodor)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Many <filename>contrib/tsearch2</> improvements (Teodor)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add double metaphone to <filename>contrib/fuzzystrmatch</> (Andrew)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Allow <filename>contrib/pg_autovacuum</> to run as a Windows service (Dave Page)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Add functions to <filename>contrib/dbsize</> (Andreas Pflug)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Removed <filename>contrib/pg_logger</>: obsoleted by integrated logging
+ subprocess
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Removed <filename>contrib/rserv</>: obsoleted by various separate projects
+ </para>
+ </listitem>
+
+ </itemizedlist>
+ </sect3>
+
+ </sect2>
+ </sect1>
+
+ <sect1 id="release-7-4-8">
+ <title>Release 7.4.8</title>
+ <note>
+ <title>Release date</title>
+ <simpara>2005-05-09</simpara>
+ </note>
- <sect3>
- <title>Contrib Changes</title>
- <itemizedlist>
+ <para>
+ This release contains a variety of fixes from 7.4.7, including several
+ security-related issues.
+ </para>
- <listitem>
- <para>
- Overhaul of <filename>contrib/dblink</> (Joe)
- </para>
- </listitem>
+ <sect2>
+ <title>Migration to version 7.4.8</title>
- <listitem>
- <para>
- <filename>contrib/dbmirror</> improvements (Steven Singer)
- </para>
- </listitem>
+ <para>
+ A dump/restore is not required for those running 7.4.X. However,
+ it is one possible way of handling two significant security problems
+ that have been found in the initial contents of 7.4.X system
+ catalogs. A dump/initdb/reload sequence using 7.4.8's initdb will
+ automatically correct these problems.
+ </para>
- <listitem>
- <para>
- New <filename>contrib/xml2</> (John Gray, Torchbox)
- </para>
- </listitem>
+ <para>
+ The larger security problem is that the built-in character set encoding
+ conversion functions can be invoked from SQL commands by unprivileged
+ users, but the functions were not designed for such use and are not
+ secure against malicious choices of arguments. The fix involves changing
+ the declared parameter list of these functions so that they can no longer
+ be invoked from SQL commands. (This does not affect their normal use
+ by the encoding conversion machinery.)
+ </para>
- <listitem>
- <para>
- Updated <filename>contrib/mysql</filename>
- </para>
- </listitem>
+ <para>
+ The lesser problem is that the <filename>contrib/tsearch2</> module
+ creates several functions that are misdeclared to return
+ <type>internal</> when they do not accept <type>internal</> arguments.
+ This breaks type safety for all functions using <type>internal</>
+ arguments.
+ </para>
- <listitem>
- <para>
- New version of <filename>contrib/btree_gist</> (Teodor)
- </para>
- </listitem>
+ <para>
+ It is strongly recommended that all installations repair these errors,
+ either by initdb or by following the manual repair procedures given
+ below. The errors at least allow unprivileged database users to crash
+ their server process, and may allow unprivileged users to gain the
+ privileges of a database superuser.
+ </para>
- <listitem>
- <para>
- New <filename>contrib/trgm</>, trigram matching for
- <productname>PostgreSQL</productname> (Teodor)
- </para>
- </listitem>
+ <para>
+ If you wish not to do an initdb, perform the following procedures instead.
+ As the database superuser, do:
- <listitem>
- <para>
- Many <filename>contrib/tsearch2</> improvements (Teodor)
- </para>
- </listitem>
+<programlisting>
+BEGIN;
+UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
+WHERE pronamespace = 11 AND pronargs = 5
+ AND proargtypes[2] = 'cstring'::regtype;
+-- The command should report having updated 90 rows;
+-- if not, rollback and investigate instead of committing!
+COMMIT;
+</programlisting>
- <listitem>
- <para>
- Add double metaphone to <filename>contrib/fuzzystrmatch</> (Andrew)
- </para>
- </listitem>
+ Next, if you have installed <filename>contrib/tsearch2</>, do
- <listitem>
- <para>
- Allow <filename>contrib/pg_autovacuum</> to run as a Windows service (Dave Page)
- </para>
- </listitem>
+<programlisting>
+BEGIN;
+UPDATE pg_proc SET proargtypes[0] = 'internal'::regtype
+WHERE oid IN (
+ 'dex_init(text)'::regprocedure,
+ 'snb_en_init(text)'::regprocedure,
+ 'snb_ru_init(text)'::regprocedure,
+ 'spell_init(text)'::regprocedure,
+ 'syn_init(text)'::regprocedure
+);
+-- The command should report having updated 5 rows;
+-- if not, rollback and investigate instead of committing!
+COMMIT;
+</programlisting>
- <listitem>
- <para>
- Add functions to <filename>contrib/dbsize</> (Andreas Pflug)
- </para>
- </listitem>
+ If this command fails with a message like <quote>function
+ "dex_init(text)" does not exist</>, then either <filename>tsearch2</>
+ is not installed in this database, or you already did the update.
+ </para>
- <listitem>
- <para>
- Removed <filename>contrib/pg_logger</>: obsoleted by integrated logging
- subprocess
- </para>
- </listitem>
+ <para>
+ The above procedures must be carried out in <emphasis>each</> database
+ of an installation, including <literal>template1</>, and ideally
+ including <literal>template0</> as well. If you do not fix the
+ template databases then any subsequently created databases will contain
+ the same errors. <literal>template1</> can be fixed in the same way
+ as any other database, but fixing <literal>template0</> requires
+ additional steps. First, from any database issue
+<programlisting>
+UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
+</programlisting>
+ Next connect to <literal>template0</> and perform the above repair
+ procedures. Finally, do
+<programlisting>
+-- re-freeze template0:
+VACUUM FREEZE;
+-- and protect it against future alterations:
+UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
+</programlisting>
+ </para>
+ </sect2>
- <listitem>
- <para>
- Removed <filename>contrib/rserv</>: obsoleted by various separate projects
- </para>
- </listitem>
+ <sect2>
+ <title>Changes</title>
- </itemizedlist>
- </sect3>
+<itemizedlist>
+<listitem><para>Change encoding function signature to prevent
+misuse</para></listitem>
+<listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
+<type>INTERNAL</> function results</para></listitem>
+<listitem><para>Repair ancient race condition that allowed a transaction to be
+seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
+than for other purposes</para>
+<para>This is an extremely serious bug since it could lead to apparent
+data inconsistencies being briefly visible to applications.</para></listitem>
+<listitem><para>Repair race condition between relation extension and
+VACUUM</para>
+<para>This could theoretically have caused loss of a page's worth of
+freshly-inserted data, although the scenario seems of very low probability.
+There are no known cases of it having caused more than an Assert failure.
+</para></listitem>
+<listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
+<para>
+The comparison code was wrong in the case where the
+<literal>--enable-integer-datetimes</> configuration switch had been used.
+NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
+it will need to be <command>REINDEX</>ed after installing this update, because
+the fix corrects the sort order of column values.
+</para></listitem>
+<listitem><para>Fix <function>EXTRACT(EPOCH)</> for
+<type>TIME WITH TIME ZONE</> values</para></listitem>
+<listitem><para>Fix mis-display of negative fractional seconds in
+<type>INTERVAL</> values</para>
+<para>
+This error only occurred when the
+<literal>--enable-integer-datetimes</> configuration switch had been used.
+</para></listitem>
+<listitem><para>Ensure operations done during backend shutdown are counted by
+statistics collector</para>
+<para>
+ This is expected to resolve reports of <application>pg_autovacuum</>
+ not vacuuming the system catalogs often enough — it was not being
+ told about catalog deletions caused by temporary table removal during
+ backend exit.
+</para></listitem>
+<listitem><para>Additional buffer overrun checks in plpgsql
+(Neil)</para></listitem>
+<listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
+correctly (Neil)</para></listitem>
+<listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
+(Marko Kreen)</para></listitem>
+<listitem><para>Still more 64-bit fixes for
+<filename>contrib/intagg</></para></listitem>
+<listitem><para>Prevent incorrect optimization of functions returning
+<type>RECORD</></para></listitem>
+<listitem><para>Prevent <function>to_char(interval)</> from dumping core for
+month-related formats</para></listitem>
+<listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
+<listitem><para>Fix <function>array_map</> to call PL functions correctly</para></listitem>
+<listitem><para>Fix permission checking in <command>ALTER DATABASE RENAME</></para></listitem>
+<listitem><para>Fix <command>ALTER LANGUAGE RENAME</></para></listitem>
+<listitem><para>Make <function>RemoveFromWaitQueue</> clean up after itself</para>
+<para>
+This fixes a lock management error that would only be visible if a transaction
+was kicked out of a wait for a lock (typically by query cancel) and then the
+holder of the lock released it within a very narrow window.
+</para></listitem>
+<listitem><para>Fix problem with untyped parameter appearing in
+<command>INSERT ... SELECT</></para></listitem>
+<listitem><para>Fix <command>CLUSTER</> failure after
+<command>ALTER TABLE SET WITHOUT OIDS</></para></listitem>
+</itemizedlist>
- </sect2>
- </sect1>
+ </sect2>
+ </sect1>
<sect1 id="release-7-4-7">
<title>Release 7.4.7</title>
GMT</para></listitem>
</itemizedlist>
- </sect2>
- </sect1>
+ </sect2>
+ </sect1>
<sect1 id="release-7-4-6">
<title>Release 7.4.6</title>
ECPG prepare statement</para></listitem>
</itemizedlist>
- </sect2>
- </sect1>
+ </sect2>
+ </sect1>
<sect1 id="release-7-4-5">
<title>Release 7.4.5</title>
</para></listitem>
</itemizedlist>
- </sect2>
- </sect1>
+ </sect2>
+ </sect1>
<sect1 id="release-7-4-4">
<title>Release 7.4.4</title>
<listitem><para>Translation updates (various contributors)</para></listitem>
</itemizedlist>
- </sect2>
- </sect1>
+ </sect2>
+ </sect1>
<sect1 id="release-7-4-3">
<title>Release 7.4.3</title>
<listitem><para>Numerous translation updates (various contributors)</para></listitem>
</itemizedlist>
- </sect2>
- </sect1>
-
+ </sect2>
+ </sect1>
+
<sect1 id="release-7-4-2">
<title>Release 7.4.2</title>
<listitem><para>Various JDBC fixes</para></listitem>
</itemizedlist>
- </sect2>
- </sect1>
+ </sect2>
+ </sect1>
<sect1 id="release-7-4-1">
<title>Release 7.4.1</title>
<para>
Major changes in this release:
+ </para>
- <variablelist>
- <varlistentry>
- <term>
- <literal>IN</literal> / <literal>NOT IN</literal> subqueries are
- now much more efficient
- </term>
+ <variablelist>
+ <varlistentry>
+ <term>
+ <literal>IN</literal> / <literal>NOT IN</literal> subqueries are
+ now much more efficient
+ </term>
- <listitem>
- In previous releases, <literal>IN</literal>/<literal>NOT
- IN</literal> subqueries were joined to the upper query by
- sequentially scanning the subquery looking for a match. The
- 7.4 code uses the same sophisticated techniques used by
- ordinary joins and so is much faster. An
- <literal>IN</literal> will now usually be as fast as or faster
- than an equivalent <literal>EXISTS</literal> subquery; this
- reverses the conventional wisdom that applied to previous
- releases.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ In previous releases, <literal>IN</literal>/<literal>NOT
+ IN</literal> subqueries were joined to the upper query by
+ sequentially scanning the subquery looking for a match. The
+ 7.4 code uses the same sophisticated techniques used by
+ ordinary joins and so is much faster. An
+ <literal>IN</literal> will now usually be as fast as or faster
+ than an equivalent <literal>EXISTS</literal> subquery; this
+ reverses the conventional wisdom that applied to previous
+ releases.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Improved <literal>GROUP BY</literal> processing by using hash buckets
- </term>
+ <varlistentry>
+ <term>
+ Improved <literal>GROUP BY</literal> processing by using hash buckets
+ </term>
- <listitem>
- In previous releases, rows to be grouped had to be sorted
- first. The 7.4 code can do <literal>GROUP BY</literal>
- without sorting, by accumulating results into a hash table
- with one entry per group. It will still use the sort
- technique, however, if the hash table is estimated to be too
- large to fit in <varname>sort_mem</>.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ In previous releases, rows to be grouped had to be sorted
+ first. The 7.4 code can do <literal>GROUP BY</literal>
+ without sorting, by accumulating results into a hash table
+ with one entry per group. It will still use the sort
+ technique, however, if the hash table is estimated to be too
+ large to fit in <varname>sort_mem</>.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- New multikey hash join capability
- </term>
+ <varlistentry>
+ <term>
+ New multikey hash join capability
+ </term>
- <listitem>
- In previous releases, hash joins could only occur on single
- keys. This release allows multicolumn hash joins.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ In previous releases, hash joins could only occur on single
+ keys. This release allows multicolumn hash joins.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Queries using the explicit <literal>JOIN</literal> syntax are
- now better optimized
- </term>
+ <varlistentry>
+ <term>
+ Queries using the explicit <literal>JOIN</literal> syntax are
+ now better optimized
+ </term>
- <listitem>
- Prior releases evaluated queries using the explicit
- <literal>JOIN</literal> syntax only in the order implied by
- the syntax. 7.4 allows full optimization of these queries,
- meaning the optimizer considers all possible join orderings
- and chooses the most efficient. Outer joins, however, must
- still follow the declared ordering.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ Prior releases evaluated queries using the explicit
+ <literal>JOIN</literal> syntax only in the order implied by
+ the syntax. 7.4 allows full optimization of these queries,
+ meaning the optimizer considers all possible join orderings
+ and chooses the most efficient. Outer joins, however, must
+ still follow the declared ordering.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Faster and more powerful regular expression code
- </term>
+ <varlistentry>
+ <term>
+ Faster and more powerful regular expression code
+ </term>
- <listitem>
- The entire regular expression module has been replaced with a
- new version by Henry Spencer, originally written for Tcl. The
- code greatly improves performance and supports several flavors
- of regular expressions.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ The entire regular expression module has been replaced with a
+ new version by Henry Spencer, originally written for Tcl. The
+ code greatly improves performance and supports several flavors
+ of regular expressions.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Function-inlining for simple SQL functions
- </term>
+ <varlistentry>
+ <term>
+ Function-inlining for simple SQL functions
+ </term>
- <listitem>
- Simple SQL functions can now be inlined by including their SQL
- in the main query. This improves performance by eliminating
- per-call overhead. That means simple SQL functions now
- behave like macros.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ Simple SQL functions can now be inlined by including their SQL
+ in the main query. This improves performance by eliminating
+ per-call overhead. That means simple SQL functions now
+ behave like macros.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Full support for IPv6 connections and IPv6 address data types
- </term>
+ <varlistentry>
+ <term>
+ Full support for IPv6 connections and IPv6 address data types
+ </term>
- <listitem>
- Previous releases allowed only IPv4 connections, and the IP
- data types only supported IPv4 addresses. This release adds
- full IPv6 support in both of these areas.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ Previous releases allowed only IPv4 connections, and the IP
+ data types only supported IPv4 addresses. This release adds
+ full IPv6 support in both of these areas.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Major improvements in SSL performance and reliability
- </term>
+ <varlistentry>
+ <term>
+ Major improvements in SSL performance and reliability
+ </term>
- <listitem>
- Several people very familiar with the SSL API have overhauled
- our SSL code to improve SSL key negotiation and error
- recovery.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ Several people very familiar with the SSL API have overhauled
+ our SSL code to improve SSL key negotiation and error
+ recovery.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Make free space map efficiently reuse empty index pages,
- and other free space management improvements
- </term>
+ <varlistentry>
+ <term>
+ Make free space map efficiently reuse empty index pages,
+ and other free space management improvements
+ </term>
- <listitem>
- In previous releases, B-tree index pages that were left empty
- because of deleted rows could only be reused by rows with
- index values similar to the rows originally indexed on that
- page. In 7.4, <command>VACUUM</command> records empty index
- pages and allows them to be reused for any future index rows.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ In previous releases, B-tree index pages that were left empty
+ because of deleted rows could only be reused by rows with
+ index values similar to the rows originally indexed on that
+ page. In 7.4, <command>VACUUM</command> records empty index
+ pages and allows them to be reused for any future index rows.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- SQL-standard information schema
- </term>
+ <varlistentry>
+ <term>
+ SQL-standard information schema
+ </term>
- <listitem>
- The information schema provides a standardized and stable way
- to access information about the schema objects defined in a
- database.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ The information schema provides a standardized and stable way
+ to access information about the schema objects defined in a
+ database.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Cursors conform more closely to the SQL standard
- </term>
+ <varlistentry>
+ <term>
+ Cursors conform more closely to the SQL standard
+ </term>
- <listitem>
- The commands <command>FETCH</command> and
- <command>MOVE</command> have been overhauled to conform more
- closely to the SQL standard.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ The commands <command>FETCH</command> and
+ <command>MOVE</command> have been overhauled to conform more
+ closely to the SQL standard.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Cursors can exist outside transactions
- </term>
+ <varlistentry>
+ <term>
+ Cursors can exist outside transactions
+ </term>
- <listitem>
- These cursors are also called holdable cursors.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ These cursors are also called holdable cursors.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- New client-to-server protocol
- </term>
+ <varlistentry>
+ <term>
+ New client-to-server protocol
+ </term>
- <listitem>
- The new protocol adds error codes, more status information,
- faster startup, better support for binary data transmission,
- parameter values separated from SQL commands, prepared
- statements available at the protocol level, and cleaner
- recovery from <command>COPY</command> failures. The older
- protocol is still supported by both server and clients.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ The new protocol adds error codes, more status information,
+ faster startup, better support for binary data transmission,
+ parameter values separated from SQL commands, prepared
+ statements available at the protocol level, and cleaner
+ recovery from <command>COPY</command> failures. The older
+ protocol is still supported by both server and clients.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
-
<application>libpq</application> and
-
<application>ECPG</application> applications are now fully
- thread-safe
- </term>
+ <varlistentry>
+ <term>
+ <application>libpq</application> and
+ <application>ECPG</application> applications are now fully
+ thread-safe
+ </term>
- <listitem>
-
While previous <application>libpq</application> releases
- already supported threads, this release improves thread safety
- by fixing some non-thread-safe code that was used during
- database connection startup. The <command>configure</command>
- option <option>--enable-thread-safety</option> must be used to
- enable this feature.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ While previous <application>libpq</application> releases
+ already supported threads, this release improves thread safety
+ by fixing some non-thread-safe code that was used during
+ database connection startup. The <command>configure</command>
+ option <option>--enable-thread-safety</option> must be used to
+ enable this feature.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- New version of full-text indexing
- </term>
+ <varlistentry>
+ <term>
+ New version of full-text indexing
+ </term>
- <listitem>
- A new full-text indexing suite is available in
- <filename>contrib/tsearch2</filename>.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ A new full-text indexing suite is available in
+ <filename>contrib/tsearch2</filename>.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- New autovacuum tool
- </term>
+ <varlistentry>
+ <term>
+ New autovacuum tool
+ </term>
- <listitem>
- The new autovacuum tool in
- <filename>contrib/autovacuum</filename> monitors the database
- statistics tables for
- <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
- activity and automatically vacuums tables when needed.
- </para>
- </listitem>
- </varlistentry>
+ <listitem>
+ <para>
+ The new autovacuum tool in
+ <filename>contrib/autovacuum</filename> monitors the database
+ statistics tables for
+ <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
+ activity and automatically vacuums tables when needed.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Array handling has been improved and moved into the server core
- </term>
+ <varlistentry>
+ <term>
+ Array handling has been improved and moved into the server core
+ </term>
- <listitem>
- Many array limitations have been removed, and arrays behave
- more like fully-supported data types.
- </para>
- </listitem>
- </varlistentry>
- </variablelist>
- </para>
+ <listitem>
+ <para>
+ Many array limitations have been removed, and arrays behave
+ more like fully-supported data types.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ </variablelist>
</sect2>
<sect2>
<para>
Observe the following incompatibilities:
+ </para>
- <itemizedlist>
- <listitem>
- The server-side autocommit setting was removed and
- reimplemented in client applications and languages.
- Server-side autocommit was causing too many problems with
- languages and applications that wanted to control their own
- autocommit behavior, so autocommit was removed from the server
- and added to individual client APIs as appropriate.
- </para>
- </listitem>
+ <itemizedlist>
+ <listitem>
+ <para>
+ The server-side autocommit setting was removed and
+ reimplemented in client applications and languages.
+ Server-side autocommit was causing too many problems with
+ languages and applications that wanted to control their own
+ autocommit behavior, so autocommit was removed from the server
+ and added to individual client APIs as appropriate.
+ </para>
+ </listitem>
- <listitem>
- Error message wording has changed substantially in this
- release. Significant effort was invested to make the messages
- more consistent and user-oriented. If your applications try to
- detect different error conditions by parsing the error message,
- you are strongly encouraged to use the new error code facility instead.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Error message wording has changed substantially in this
+ release. Significant effort was invested to make the messages
+ more consistent and user-oriented. If your applications try to
+ detect different error conditions by parsing the error message,
+ you are strongly encouraged to use the new error code facility instead.
+ </para>
+ </listitem>
- <listitem>
- Inner joins using the explicit <literal>JOIN</literal> syntax
- may behave differently because they are now better
- optimized.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Inner joins using the explicit <literal>JOIN</literal> syntax
+ may behave differently because they are now better
+ optimized.
+ </para>
+ </listitem>
- <listitem>
- <para>
- A number of server configuration parameters have been renamed
- for clarity, primarily those related to
- logging.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ A number of server configuration parameters have been renamed
+ for clarity, primarily those related to
+ logging.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <literal>FETCH 0</literal> or <literal>MOVE 0</literal> now
+ does nothing. In prior releases, <literal>FETCH 0</literal>
+ would fetch all remaining rows, and <literal>MOVE 0</literal>
+ would move to the end of the cursor.
+ </para>
+ </listitem>
- <listitem>
- <para>
- <literal>FETCH 0</literal> or <literal>MOVE 0</literal> now
- does nothing. In prior releases, <literal>FETCH 0</literal>
- would fetch all remaining rows, and <literal>MOVE 0</literal>
- would move to the end of the cursor.
- </para>
- </listitem>
-
- <listitem>
- <para>
- <command>FETCH</command> and <command>MOVE</command> now return
- the actual number of rows fetched/moved, or zero if at the
- beginning/end of the cursor. Prior releases would return the
- row count passed to the command, not the number of rows
- actually fetched or moved.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ <command>FETCH</command> and <command>MOVE</command> now return
+ the actual number of rows fetched/moved, or zero if at the
+ beginning/end of the cursor. Prior releases would return the
+ row count passed to the command, not the number of rows
+ actually fetched or moved.
+ </para>
+ </listitem>
- <listitem>
- <command>COPY</command> now can process files that use
- carriage-return or carriage-return/line-feed end-of-line
- sequences. Literal carriage-returns and line-feeds are no
- longer accepted in data values; use <literal>\r</literal> and
- <literal>\n</literal> instead.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ <command>COPY</command> now can process files that use
+ carriage-return or carriage-return/line-feed end-of-line
+ sequences. Literal carriage-returns and line-feeds are no
+ longer accepted in data values; use <literal>\r</literal> and
+ <literal>\n</literal> instead.
+ </para>
+ </listitem>
- <listitem>
- Trailing spaces are now trimmed when converting from type
- <type>char(<replaceable>n</>)</type> to
- <type>varchar(<replaceable>n</>)</type> or <type>text</type>.
- This is what most people always expected to happen anyway.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Trailing spaces are now trimmed when converting from type
+ <type>char(<replaceable>n</>)</type> to
+ <type>varchar(<replaceable>n</>)</type> or <type>text</type>.
+ This is what most people always expected to happen anyway.
+ </para>
+ </listitem>
- <listitem>
- The data type <type>float(<replaceable>p</>)</type> now
- measures <replaceable>p</> in binary digits, not decimal
- digits. The new behavior follows the SQL standard.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The data type <type>float(<replaceable>p</>)</type> now
+ measures <replaceable>p</> in binary digits, not decimal
+ digits. The new behavior follows the SQL standard.
+ </para>
+ </listitem>
- <listitem>
- Ambiguous date values now must match the ordering specified by
- the <varname>datestyle</varname> setting. In prior releases, a
- date specification of <literal>10/20/03</> was interpreted as a
- date in October even if <varname>datestyle</> specified that
- the day should be first. 7.4 will throw an error if a date
- specification is invalid for the current setting of
- <varname>datestyle</>.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ Ambiguous date values now must match the ordering specified by
+ the <varname>datestyle</varname> setting. In prior releases, a
+ date specification of <literal>10/20/03</> was interpreted as a
+ date in October even if <varname>datestyle</> specified that
+ the day should be first. 7.4 will throw an error if a date
+ specification is invalid for the current setting of
+ <varname>datestyle</>.
+ </para>
+ </listitem>
- <listitem>
- The functions <function>oidrand</function>,
- <function>oidsrand</function>, and
- <function>userfntest</function> have been removed. These
- functions were determined to be no longer useful.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The functions <function>oidrand</function>,
+ <function>oidsrand</function>, and
+ <function>userfntest</function> have been removed. These
+ functions were determined to be no longer useful.
+ </para>
+ </listitem>
- <listitem>
- String literals specifying time-varying date/time values, such
- as <literal>'now'</literal> or <literal>'today'</literal> will
- no longer work as expected in column default expressions; they
- now cause the time of the table creation to be the default, not
- the time of the insertion. Functions such as
- <function>now()</>, <function>current_timestamp</>, or
- <function>current_date</function> should be used instead.
- </para>
+ <listitem>
+ <para>
+ String literals specifying time-varying date/time values, such
+ as <literal>'now'</literal> or <literal>'today'</literal> will
+ no longer work as expected in column default expressions; they
+ now cause the time of the table creation to be the default, not
+ the time of the insertion. Functions such as
+ <function>now()</>, <function>current_timestamp</>, or
+ <function>current_date</function> should be used instead.
+ </para>
- In previous releases, there was special code so that strings
- such as <literal>'now'</literal> were interpreted at
- <command>INSERT</> time and not at table creation time, but
- this work around didn't cover all cases. Release 7.4 now
- requires that defaults be defined properly using functions such
- as <function>now()</> or <function>current_timestamp</>. These
- will work in all situations.
- </para>
- </listitem>
+ <para>
+ In previous releases, there was special code so that strings
+ such as <literal>'now'</literal> were interpreted at
+ <command>INSERT</> time and not at table creation time, but
+ this work around didn't cover all cases. Release 7.4 now
+ requires that defaults be defined properly using functions such
+ as <function>now()</> or <function>current_timestamp</>. These
+ will work in all situations.
+ </para>
+ </listitem>
- <listitem>
- The dollar sign (<literal>$</>) is no longer allowed in
- operator names. It can instead be a non-first character in
- identifiers. This was done to improve compatibility with other
- database systems, and to avoid syntax problems when parameter
- placeholders (<literal>$<replaceable>n</></>) are written
- adjacent to operators.
- </para>
- </listitem>
- </itemizedlist>
- </para>
+ <listitem>
+ <para>
+ The dollar sign (<literal>$</>) is no longer allowed in
+ operator names. It can instead be a non-first character in
+ identifiers. This was done to improve compatibility with other
+ database systems, and to avoid syntax problems when parameter
+ placeholders (<literal>$<replaceable>n</></>) are written
+ adjacent to operators.
+ </para>
+ </listitem>
+
+ </itemizedlist>
</sect2>
<sect2>
</sect2>
</sect1>
+ <sect1 id="release-7-3-10">
+ <title>Release 7.3.10</title>
+
+ <note>
+ <title>Release date</title>
+ <simpara>2005-05-09</simpara>
+ </note>
+
+ <para>
+ This release contains a variety of fixes from 7.3.9, including several
+ security-related issues.
+ </para>
+
+ <sect2>
+ <title>Migration to version 7.3.10</title>
+
+ <para>
+ A dump/restore is not required for those running 7.3.X. However,
+ it is one possible way of handling a significant security problem
+ that has been found in the initial contents of 7.3.X system
+ catalogs. A dump/initdb/reload sequence using 7.3.10's initdb will
+ automatically correct this problem.
+ </para>
+
+ <para>
+ The security problem is that the built-in character set encoding
+ conversion functions can be invoked from SQL commands by unprivileged
+ users, but the functions were not designed for such use and are not
+ secure against malicious choices of arguments. The fix involves changing
+ the declared parameter list of these functions so that they can no longer
+ be invoked from SQL commands. (This does not affect their normal use
+ by the encoding conversion machinery.)
+ It is strongly recommended that all installations repair this error,
+ either by initdb or by following the manual repair procedure given
+ below. The error at least allows unprivileged database users to crash
+ their server process, and may allow unprivileged users to gain the
+ privileges of a database superuser.
+ </para>
+
+ <para>
+ If you wish not to do an initdb, perform the following procedure instead.
+ As the database superuser, do:
+
+<programlisting>
+BEGIN;
+UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
+WHERE pronamespace = 11 AND pronargs = 5
+ AND proargtypes[2] = 'cstring'::regtype;
+-- The command should report having updated 90 rows;
+-- if not, rollback and investigate instead of committing!
+COMMIT;
+</programlisting>
+ </para>
+
+ <para>
+ The above procedure must be carried out in <emphasis>each</> database
+ of an installation, including <literal>template1</>, and ideally
+ including <literal>template0</> as well. If you do not fix the
+ template databases then any subsequently created databases will contain
+ the same error. <literal>template1</> can be fixed in the same way
+ as any other database, but fixing <literal>template0</> requires
+ additional steps. First, from any database issue
+<programlisting>
+UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
+</programlisting>
+ Next connect to <literal>template0</> and perform the above repair
+ procedure. Finally, do
+<programlisting>
+-- re-freeze template0:
+VACUUM FREEZE;
+-- and protect it against future alterations:
+UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
+</programlisting>
+ </para>
+ </sect2>
+
+ <sect2>
+ <title>Changes</title>
+
+<itemizedlist>
+<listitem><para>Change encoding function signature to prevent
+misuse</para></listitem>
+<listitem><para>Repair ancient race condition that allowed a transaction to be
+seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
+than for other purposes</para>
+<para>This is an extremely serious bug since it could lead to apparent
+data inconsistencies being briefly visible to applications.</para></listitem>
+<listitem><para>Repair race condition between relation extension and
+VACUUM</para>
+<para>This could theoretically have caused loss of a page's worth of
+freshly-inserted data, although the scenario seems of very low probability.
+There are no known cases of it having caused more than an Assert failure.
+</para></listitem>
+<listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
+<para>
+The comparison code was wrong in the case where the
+<literal>--enable-integer-datetimes</> configuration switch had been used.
+NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
+it will need to be <command>REINDEX</>ed after installing this update, because
+the fix corrects the sort order of column values.
+</para></listitem>
+<listitem><para>Fix <function>EXTRACT(EPOCH)</> for
+<type>TIME WITH TIME ZONE</> values</para></listitem>
+<listitem><para>Fix mis-display of negative fractional seconds in
+<type>INTERVAL</> values</para>
+<para>
+This error only occurred when the
+<literal>--enable-integer-datetimes</> configuration switch had been used.
+</para></listitem>
+<listitem><para>Additional buffer overrun checks in plpgsql
+(Neil)</para></listitem>
+<listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
+correctly (Neil)</para></listitem>
+<listitem><para>Prevent <function>to_char(interval)</> from dumping core for
+month-related formats</para></listitem>
+<listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
+(Marko Kreen)</para></listitem>
+<listitem><para>Still more 64-bit fixes for
+<filename>contrib/intagg</></para></listitem>
+<listitem><para>Prevent incorrect optimization of functions returning
+<type>RECORD</></para></listitem>
+</itemizedlist>
+
+ </sect2>
+ </sect1>
+
<sect1 id="release-7-3-9">
<title>Release 7.3.9</title>
</para>
- <sect2>
- <title>Migration to version 7.3.1</title>
-
- <para>
- A dump/restore is <emphasis>not</emphasis> required for those
- running version 7.3. However, it should be noted that the main
- <productname>PostgreSQL</productname> interface library, libpq,
- has a new major version number for this release, which may require
- recompilation of client code in certain cases.
- </para>
- </sect2>
-
- <sect2>
- <title>Changes</title>
-
-<itemizedlist>
-<listitem><para>Fix a core dump of COPY TO when client/server encodings don't match (Tom)</para></listitem>
-<listitem><para>Allow <application>pg_dump</> to work with pre-7.2 servers (Philip)</para></listitem>
-<listitem><para>contrib/adddepend fixes (Tom)</para></listitem>
-<listitem><para>Fix problem with deletion of per-user/per-database config settings (Tom)</para></listitem>
-<listitem><para>contrib/vacuumlo fix (Tom)</para></listitem>
-<listitem><para>Allow 'password' encryption even when pg_shadow contains MD5 passwords (Bruce)</para></listitem>
-<listitem><para>contrib/dbmirror fix (Steven Singer)</para></listitem>
-<listitem><para>Optimizer fixes (Tom)</para></listitem>
-<listitem><para>contrib/tsearch fixes (Teodor Sigaev, Magnus)</para></listitem>
-<listitem><para>Allow locale names to be mixed case (Nicolai Tufar)</para></listitem>
-<listitem><para>Increment libpq library's major version number (Bruce)</para></listitem>
-<listitem><para>pg_hba.conf error reporting fixes (Bruce, Neil)</para></listitem>
-<listitem><para>Add SCO Openserver 5.0.4 as a supported platform (Bruce)</para></listitem>
-<listitem><para>Prevent EXPLAIN from crashing server (Tom)</para></listitem>
-<listitem><para>SSL fixes (Nathan Mueller)</para></listitem>
-<listitem><para>Prevent composite column creation via ALTER TABLE (Tom)</para></listitem>
-</itemizedlist>
- </sect2>
- </sect1>
-
-
- <sect1 id="release-7-3">
- <title>Release 7.3</title>
-
- <note>
- <title>Release date</title>
- <simpara>2002-11-27</simpara>
- </note>
-
- <sect2>
- <title>Overview</title>
-
- <para>
- Major changes in this release:
-
- <variablelist>
- <varlistentry>
- <term>Schemas</term>
- <listitem>
- <para>
- Schemas allow users to create objects in separate namespaces,
- so two people or applications can have tables with the same
- name. There is also a public schema for shared tables.
- Table/index creation can be restricted by removing privileges
- on the public schema.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Drop Column</term>
- <listitem>
- <para>
- <productname>PostgreSQL</productname> now supports the
- <literal>ALTER TABLE ... DROP COLUMN</literal> functionality.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Table Functions</term>
- <listitem>
- <para>
- Functions returning multiple rows and/or multiple columns are
- now much easier to use than before. You can call such a
- <quote>table function</quote> in the <literal>SELECT</literal>
- <literal>FROM</literal> clause, treating its output like a
- table. Also, <application>PL/pgSQL</application> functions can
- now return sets.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Prepared Queries</term>
- <listitem>
- <para>
- <productname>PostgreSQL</productname> now supports prepared
- queries, for improved performance.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Dependency Tracking</term>
- <listitem>
- <para>
- <productname>PostgreSQL</productname> now records object
- dependencies, which allows improvements in many areas.
- <command>DROP</command> statements now take either
- <literal>CASCADE</> or <literal>RESTRICT</> to control whether
- dependent objects are also dropped.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Privileges</term>
- <listitem>
- <para>
- Functions and procedural languages now have privileges, and
- functions can be defined to run with the privileges of their
- creator.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Internationalization</term>
- <listitem>
- <para>
- Both multibyte and locale support are now always enabled.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Logging</term>
- <listitem>
- <para>
- A variety of logging options have been enhanced.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Interfaces</term>
- <listitem>
- <para>
- A large number of interfaces have been moved to <ulink
- url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
- where they can be developed and released independently.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Functions/Identifiers</term>
- <listitem>
- <para>
- By default, functions can now take up to 32 parameters, and
- identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
- is now deprecated: there are specific <quote>pseudo-datatypes</>
- to represent each of the former meanings of <literal>OPAQUE</>
- in function argument and result types.
- </para>
- </listitem>
- </varlistentry>
+ <sect2>
+ <title>Migration to version 7.3.1</title>
- </variablelist>
+ <para>
+ A dump/restore is <emphasis>not</emphasis> required for those
+ running version 7.3. However, it should be noted that the main
+ <productname>PostgreSQL</productname> interface library, libpq,
+ has a new major version number for this release, which may require
+ recompilation of client code in certain cases.
</para>
</sect2>
<sect2>
- <title>Migration to version 7.3</title>
+ <title>Changes</title>
- <para>
- A dump/restore using <application>pg_dump</> is required for those
- wishing to migrate data from any previous release. If your
- application examines the system catalogs, additional changes will
- be required due to the introduction of schemas in 7.3; for more
- information, see: <ulink
- url="http://developer.postgresql.org/~momjian/upgrade_tips_7.3">
- http://developer.postgresql.org/~momjian/upgrade_tips_7.3</>.
- </para>
+<itemizedlist>
+<listitem><para>Fix a core dump of COPY TO when client/server encodings don't match (Tom)</para></listitem>
+<listitem><para>Allow <application>pg_dump</> to work with pre-7.2 servers (Philip)</para></listitem>
+<listitem><para>contrib/adddepend fixes (Tom)</para></listitem>
+<listitem><para>Fix problem with deletion of per-user/per-database config settings (Tom)</para></listitem>
+<listitem><para>contrib/vacuumlo fix (Tom)</para></listitem>
+<listitem><para>Allow 'password' encryption even when pg_shadow contains MD5 passwords (Bruce)</para></listitem>
+<listitem><para>contrib/dbmirror fix (Steven Singer)</para></listitem>
+<listitem><para>Optimizer fixes (Tom)</para></listitem>
+<listitem><para>contrib/tsearch fixes (Teodor Sigaev, Magnus)</para></listitem>
+<listitem><para>Allow locale names to be mixed case (Nicolai Tufar)</para></listitem>
+<listitem><para>Increment libpq library's major version number (Bruce)</para></listitem>
+<listitem><para>pg_hba.conf error reporting fixes (Bruce, Neil)</para></listitem>
+<listitem><para>Add SCO Openserver 5.0.4 as a supported platform (Bruce)</para></listitem>
+<listitem><para>Prevent EXPLAIN from crashing server (Tom)</para></listitem>
+<listitem><para>SSL fixes (Nathan Mueller)</para></listitem>
+<listitem><para>Prevent composite column creation via ALTER TABLE (Tom)</para></listitem>
+</itemizedlist>
+ </sect2>
+ </sect1>
+
+
+ <sect1 id="release-7-3">
+ <title>Release 7.3</title>
+
+ <note>
+ <title>Release date</title>
+ <simpara>2002-11-27</simpara>
+ </note>
+
+ <sect2>
+ <title>Overview</title>
<para>
- Observe the following incompatibilities:
+ Major changes in this release:
+ </para>
- <itemizedlist>
+ <variablelist>
+ <varlistentry>
+ <term>Schemas</term>
<listitem>
<para>
- Pre-6.3 clients are no longer supported.
+ Schemas allow users to create objects in separate namespaces,
+ so two people or applications can have tables with the same
+ name. There is also a public schema for shared tables.
+ Table/index creation can be restricted by removing privileges
+ on the public schema.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Drop Column</term>
<listitem>
<para>
- <filename>pg_hba.conf</filename> now has a column for the user
- name and additional features. Existing files need to be
- adjusted.
+ <productname>PostgreSQL</productname> now supports the
+ <literal>ALTER TABLE ... DROP COLUMN</literal> functionality.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Table Functions</term>
<listitem>
<para>
- Several <filename>postgresql.conf</filename> logging parameters
- have been renamed.
+ Functions returning multiple rows and/or multiple columns are
+ now much easier to use than before. You can call such a
+ <quote>table function</quote> in the <literal>SELECT</literal>
+ <literal>FROM</literal> clause, treating its output like a
+ table. Also, <application>PL/pgSQL</application> functions can
+ now return sets.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Prepared Queries</term>
<listitem>
<para>
- <literal>LIMIT #,#</literal> has been disabled; use
- <literal>LIMIT # OFFSET #</literal>.
+ <productname>PostgreSQL</productname> now supports prepared
+ queries, for improved performance.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Dependency Tracking</term>
<listitem>
<para>
- <command>INSERT</command> statements with column lists must
- specify a value for each specified column. For example,
- <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
- is now invalid. It's still allowed to supply fewer columns than
- expected if the <command>INSERT</command> does not have a column list.
+ <productname>PostgreSQL</productname> now records object
+ dependencies, which allows improvements in many areas.
+ <command>DROP</command> statements now take either
+ <literal>CASCADE</> or <literal>RESTRICT</> to control whether
+ dependent objects are also dropped.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Privileges</term>
<listitem>
<para>
- <type>serial</type> columns are no longer automatically
- <literal>UNIQUE</>; thus, an index will not automatically be
- created.
+ Functions and procedural languages now have privileges, and
+ functions can be defined to run with the privileges of their
+ creator.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Internationalization</term>
<listitem>
<para>
- A <command>SET</command> command inside an aborted transaction
- is now rolled back.
+ Both multibyte and locale support are now always enabled.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Logging</term>
<listitem>
<para>
- <command>COPY</command> no longer considers missing trailing
- columns to be null. All columns need to be specified.
- (However, one may achieve a similar effect by specifying a
- column list in the <command>COPY</command> command.)
+ A variety of logging options have been enhanced.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Interfaces</term>
<listitem>
<para>
- The data type <type>timestamp</type> is now equivalent to
- <type>timestamp without time zone</type>, instead of
- <type>timestamp with time zone</type>.
+ A large number of interfaces have been moved to <ulink
+ url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
+ where they can be developed and released independently.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Functions/Identifiers</term>
<listitem>
<para>
- Pre-7.3 databases loaded into 7.3 will not have the new object
- dependencies for <type>serial</type> columns, unique
- constraints, and foreign keys. See the directory
- <filename>contrib/adddepend/</filename> for a detailed
- description and a script that will add such dependencies.
+ By default, functions can now take up to 32 parameters, and
+ identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
+ is now deprecated: there are specific <quote>pseudo-datatypes</>
+ to represent each of the former meanings of <literal>OPAQUE</>
+ in function argument and result types.
</para>
</listitem>
+ </varlistentry>
- <listitem>
- <para>
- An empty string (<literal>''</literal>) is no longer allowed as
- the input into an integer field. Formerly, it was silently
- interpreted as 0.
- </para>
- </listitem>
+ </variablelist>
+ </sect2>
- </itemizedlist>
+ <sect2>
+ <title>Migration to version 7.3</title>
+
+ <para>
+ A dump/restore using <application>pg_dump</> is required for those
+ wishing to migrate data from any previous release. If your
+ application examines the system catalogs, additional changes will
+ be required due to the introduction of schemas in 7.3; for more
+ information, see: <ulink
+ url="http://developer.postgresql.org/~momjian/upgrade_tips_7.3"></>.
+ </para>
+
+ <para>
+ Observe the following incompatibilities:
</para>
+ <itemizedlist>
+ <listitem>
+ <para>
+ Pre-6.3 clients are no longer supported.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <filename>pg_hba.conf</filename> now has a column for the user
+ name and additional features. Existing files need to be
+ adjusted.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Several <filename>postgresql.conf</filename> logging parameters
+ have been renamed.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <literal>LIMIT #,#</literal> has been disabled; use
+ <literal>LIMIT # OFFSET #</literal>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <command>INSERT</command> statements with column lists must
+ specify a value for each specified column. For example,
+ <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
+ is now invalid. It's still allowed to supply fewer columns than
+ expected if the <command>INSERT</command> does not have a column list.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <type>serial</type> columns are no longer automatically
+ <literal>UNIQUE</>; thus, an index will not automatically be
+ created.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ A <command>SET</command> command inside an aborted transaction
+ is now rolled back.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ <command>COPY</command> no longer considers missing trailing
+ columns to be null. All columns need to be specified.
+ (However, one may achieve a similar effect by specifying a
+ column list in the <command>COPY</command> command.)
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The data type <type>timestamp</type> is now equivalent to
+ <type>timestamp without time zone</type>, instead of
+ <type>timestamp with time zone</type>.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ Pre-7.3 databases loaded into 7.3 will not have the new object
+ dependencies for <type>serial</type> columns, unique
+ constraints, and foreign keys. See the directory
+ <filename>contrib/adddepend/</filename> for a detailed
+ description and a script that will add such dependencies.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ An empty string (<literal>''</literal>) is no longer allowed as
+ the input into an integer field. Formerly, it was silently
+ interpreted as 0.
+ </para>
+ </listitem>
+
+ </itemizedlist>
</sect2>
<sect2>
</sect2>
</sect1>
+ <sect1 id="release-7-2-8">
+ <title>Release 7.2.8</title>
+
+ <note>
+ <title>Release date</title>
+ <simpara>2005-05-09</simpara>
+ </note>
+
+ <para>
+ This release contains a variety of fixes from 7.2.7, including one
+ security-related issue.
+ </para>
+
+ <sect2>
+ <title>Migration to version 7.2.8</title>
+
+ <para>
+ A dump/restore is not required for those running 7.2.X.
+ </para>
+ </sect2>
+
+ <sect2>
+ <title>Changes</title>
+
+<itemizedlist>
+<listitem><para>Repair ancient race condition that allowed a transaction to be
+seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
+than for other purposes</para>
+<para>This is an extremely serious bug since it could lead to apparent
+data inconsistencies being briefly visible to applications.</para></listitem>
+<listitem><para>Repair race condition between relation extension and
+VACUUM</para>
+<para>This could theoretically have caused loss of a page's worth of
+freshly-inserted data, although the scenario seems of very low probability.
+There are no known cases of it having caused more than an Assert failure.
+</para></listitem>
+<listitem><para>Fix <function>EXTRACT(EPOCH)</> for
+<type>TIME WITH TIME ZONE</> values</para></listitem>
+<listitem><para>Additional buffer overrun checks in plpgsql
+(Neil)</para></listitem>
+<listitem><para>Fix pg_dump to dump index names and trigger names containing
+<literal>%</> correctly (Neil)</para></listitem>
+<listitem><para>Prevent <function>to_char(interval)</> from dumping core for
+month-related formats</para></listitem>
+<listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
+(Marko Kreen)</para></listitem>
+</itemizedlist>
+
+ </sect2>
+ </sect1>
+
<sect1 id="release-7-2-7">
<title>Release 7.2.7</title>
<sect2>
<title>Migration to version 7.2.1</title>
-
- <para>
- A dump/restore is <emphasis>not</emphasis> required for those
- running version 7.2.
- </para>
- </sect2>
-
- <sect2>
- <title>Changes</title>
-
-<itemizedlist>
-<listitem><para>Ensure that sequence counters do not go backwards after a crash (Tom)</para></listitem>
-<listitem><para>Fix pgaccess kanji-conversion key binding (Tatsuo)</para></listitem>
-<listitem><para>Optimizer improvements (Tom)</para></listitem>
-<listitem><para>Cash I/O improvements (Tom)</para></listitem>
-<listitem><para>New Russian FAQ</para></listitem>
-<listitem><para>Compile fix for missing AuthBlockSig (Heiko)</para></listitem>
-<listitem><para>Additional time zones and time zone fixes (Thomas)</para></listitem>
-<listitem><para>Allow psql \connect to handle mixed case database and user names (Tom)</para></listitem>
-<listitem><para>Return proper OID on command completion even with ON INSERT rules (Tom)</para></listitem>
-<listitem><para>Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)</para></listitem>
-<listitem><para>Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)</para></listitem>
-<listitem><para>Improve handling of multiple UNIONs with different lengths (Tom)</para></listitem>
-<listitem><para>contrib/btree_gist improvements (Teodor Sigaev)</para></listitem>
-<listitem><para>contrib/tsearch dictionary improvements, see README.tsearch for an additional installation step (Thomas T. Thai, Teodor Sigaev)</para></listitem>
-<listitem><para>Fix for array subscripts handling (Tom)</para></listitem>
-<listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
-</itemizedlist>
- </sect2>
- </sect1>
-
-
- <sect1 id="release-7-2">
- <title>Release 7.2</title>
-
- <note>
- <title>Release date</title>
- <simpara>2002-02-04</simpara>
- </note>
-
- <sect2>
- <title>Overview</title>
-
- <para>
- This release improves <productname>PostgreSQL</> for use in
- high-volume applications.
- </para>
-
- <para>
- Major changes in this release:
-
- <variablelist>
- <varlistentry>
- <term>VACUUM</term>
- <listitem>
- <para>
- Vacuuming no longer locks tables, thus allowing normal user
- access during the vacuum. A new <command>VACUUM FULL</>
- command does old-style vacuum by locking the table and
- shrinking the on-disk copy of the table.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Transactions</term>
- <listitem>
- <para>
- There is no longer a problem with installations that exceed
- four billion transactions.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>OIDs</term>
- <listitem>
- <para>
- OIDs are now optional. Users can now create tables without
- OIDs for cases where OID usage is excessive.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Optimizer</term>
- <listitem>
- <para>
- The system now computes histogram column statistics during
- <command>ANALYZE</>, allowing much better optimizer choices.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Security</term>
- <listitem>
- <para>
- A new MD5 encryption option allows more secure storage and
- transfer of passwords. A new Unix-domain socket
- authentication option is available on Linux and BSD systems.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Statistics</term>
- <listitem>
- <para>
- Administrators can use the new table access statistics module
- to get fine-grained information about table and index usage.
- </para>
- </listitem>
- </varlistentry>
-
- <varlistentry>
- <term>Internationalization</term>
- <listitem>
- <para>
- Program and library messages can now be displayed in several
- languages.
- </para>
- </listitem>
- </varlistentry>
- </variablelist>
+
+ <para>
+ A dump/restore is <emphasis>not</emphasis> required for those
+ running version 7.2.
</para>
</sect2>
<sect2>
- <title>Migration to version 7.2</title>
+ <title>Changes</title>
+
+<itemizedlist>
+<listitem><para>Ensure that sequence counters do not go backwards after a crash (Tom)</para></listitem>
+<listitem><para>Fix pgaccess kanji-conversion key binding (Tatsuo)</para></listitem>
+<listitem><para>Optimizer improvements (Tom)</para></listitem>
+<listitem><para>Cash I/O improvements (Tom)</para></listitem>
+<listitem><para>New Russian FAQ</para></listitem>
+<listitem><para>Compile fix for missing AuthBlockSig (Heiko)</para></listitem>
+<listitem><para>Additional time zones and time zone fixes (Thomas)</para></listitem>
+<listitem><para>Allow psql \connect to handle mixed case database and user names (Tom)</para></listitem>
+<listitem><para>Return proper OID on command completion even with ON INSERT rules (Tom)</para></listitem>
+<listitem><para>Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)</para></listitem>
+<listitem><para>Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)</para></listitem>
+<listitem><para>Improve handling of multiple UNIONs with different lengths (Tom)</para></listitem>
+<listitem><para>contrib/btree_gist improvements (Teodor Sigaev)</para></listitem>
+<listitem><para>contrib/tsearch dictionary improvements, see README.tsearch for an additional installation step (Thomas T. Thai, Teodor Sigaev)</para></listitem>
+<listitem><para>Fix for array subscripts handling (Tom)</para></listitem>
+<listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
+</itemizedlist>
+ </sect2>
+ </sect1>
+
+
+ <sect1 id="release-7-2">
+ <title>Release 7.2</title>
+
+ <note>
+ <title>Release date</title>
+ <simpara>2002-02-04</simpara>
+ </note>
+
+ <sect2>
+ <title>Overview</title>
<para>
- A dump/restore using <command>pg_dump</command> is required for
- those wishing to migrate data from any previous release.
+ This release improves <productname>PostgreSQL</> for use in
+ high-volume applications.
</para>
<para>
- Observe the following incompatibilities:
+ Major changes in this release:
+ </para>
- <itemizedlist>
+ <variablelist>
+ <varlistentry>
+ <term>VACUUM</term>
<listitem>
<para>
- The semantics of the <command>VACUUM</command> command have
- changed in this release. You may wish to update your
- maintenance procedures accordingly.
+ Vacuuming no longer locks tables, thus allowing normal user
+ access during the vacuum. A new <command>VACUUM FULL</>
+ command does old-style vacuum by locking the table and
+ shrinking the on-disk copy of the table.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Transactions</term>
<listitem>
<para>
- In this release, comparisons using <literal>= NULL</literal>
- will always return false (or NULL, more precisely). Previous
- releases automatically transformed this syntax to <literal>IS
- NULL</literal>. The old behavior can be re-enabled using a
- <filename>postgresql.conf</filename> parameter.
+ There is no longer a problem with installations that exceed
+ four billion transactions.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>OIDs</term>
<listitem>
<para>
- The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
- configuration is now only reloaded after receiving a
- <systemitem>SIGHUP</> signal, not with each connection.
+ OIDs are now optional. Users can now create tables without
+ OIDs for cases where OID usage is excessive.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Optimizer</term>
<listitem>
<para>
- The function <filename>octet_length()</> now returns the uncompressed data length.
+ The system now computes histogram column statistics during
+ <command>ANALYZE</>, allowing much better optimizer choices.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Security</term>
<listitem>
<para>
- The date/time value <literal>'current'</literal> is no longer
- available. You will need to rewrite your applications.
+ A new MD5 encryption option allows more secure storage and
+ transfer of passwords. A new Unix-domain socket
+ authentication option is available on Linux and BSD systems.
</para>
</listitem>
+ </varlistentry>
+ <varlistentry>
+ <term>Statistics</term>
<listitem>
<para>
- The <literal>timestamp()</literal>, <literal>time()</literal>,
- and <literal>interval()</literal> functions are no longer
- available. Instead of <literal>timestamp()</literal>, use
- <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
+ Administrators can use the new table access statistics module
+ to get fine-grained information about table and index usage.
</para>
</listitem>
- </itemizedlist>
+ </varlistentry>
+
+ <varlistentry>
+ <term>Internationalization</term>
+ <listitem>
+ <para>
+ Program and library messages can now be displayed in several
+ languages.
+ </para>
+ </listitem>
+ </varlistentry>
+
+ </variablelist>
+ </sect2>
+
+ <sect2>
+ <title>Migration to version 7.2</title>
+
+ <para>
+ A dump/restore using <command>pg_dump</command> is required for
+ those wishing to migrate data from any previous release.
+ </para>
+
+ <para>
+ Observe the following incompatibilities:
</para>
+ <itemizedlist>
+ <listitem>
+ <para>
+ The semantics of the <command>VACUUM</command> command have
+ changed in this release. You may wish to update your
+ maintenance procedures accordingly.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ In this release, comparisons using <literal>= NULL</literal>
+ will always return false (or NULL, more precisely). Previous
+ releases automatically transformed this syntax to <literal>IS
+ NULL</literal>. The old behavior can be re-enabled using a
+ <filename>postgresql.conf</filename> parameter.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
+ configuration is now only reloaded after receiving a
+ <systemitem>SIGHUP</> signal, not with each connection.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The function <filename>octet_length()</> now returns the uncompressed data length.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The date/time value <literal>'current'</literal> is no longer
+ available. You will need to rewrite your applications.
+ </para>
+ </listitem>
+
+ <listitem>
+ <para>
+ The <literal>timestamp()</literal>, <literal>time()</literal>,
+ and <literal>interval()</literal> functions are no longer
+ available. Instead of <literal>timestamp()</literal>, use
+ <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
+ </para>
+ </listitem>
+
+ </itemizedlist>
+
<para>
The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
in the next release. You should change your queries to use
<para>
Major changes in this release:
+ </para>
- <variablelist>
- <varlistentry>
- <term>
- Write-ahead Log (WAL)
- </term>
- <listitem>
+ <variablelist>
+ <varlistentry>
+ <term>
+ Write-ahead Log (WAL)
+ </term>
+ <listitem>
+ <para>
To maintain database consistency in case of an operating system crash,
previous releases of <productname>PostgreSQL</productname> have forced
all data modifications to disk before each transaction commit. With
WAL, only one log file must be flushed to disk, greatly improving
performance. If you have been using -F in previous releases to
disable disk flushes, you may want to consider discontinuing its use.
- </para>
- </listitem>
- </varlistentry>
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- TOAST
- </term>
- <listitem>
- TOAST - Previous releases had a compiled-in row length limit,
+ <varlistentry>
+ <term>
+ TOAST
+ </term>
+ <listitem>
+ <para>
+ TOAST - Previous releases had a compiled-in row length limit,
typically 8k - 32k. This limit made storage of long text fields
difficult. With TOAST, long rows of any length can be stored with good
performance.
- </para>
- </listitem>
- </varlistentry>
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Outer Joins
- </term>
- <listitem>
+ <varlistentry>
+ <term>
+ Outer Joins
+ </term>
+ <listitem>
+ <para>
We now support outer joins. The UNION/NOT IN
workaround for outer joins is no longer required. We use the SQL92
outer join syntax.
- </para>
- </listitem>
- </varlistentry>
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Function Manager
- </term>
- <listitem>
+ <varlistentry>
+ <term>
+ Function Manager
+ </term>
+ <listitem>
+ <para>
The previous C function manager did not
handle null values properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
function manager does. You can continue using your old custom
functions, but you may want to rewrite them in the future to use the new
function manager call interface.
- </para>
- </listitem>
- </varlistentry>
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Complex Queries
- </term>
- <listitem>
+ <varlistentry>
+ <term>
+ Complex Queries
+ </term>
+ <listitem>
+ <para>
A large number of complex queries that were
unsupported in previous releases now work. Many combinations of views,
aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
now work properly. Inherited tables are now accessed by default.
Subqueries in FROM are now supported.
- </para>
- </listitem>
- </varlistentry>
+ </para>
+ </listitem>
+ </varlistentry>
- </variablelist>
- </para>
+ </variablelist>
<sect2>
<title>Migration to version 7.1</title>
<para>
Major changes in this release:
+ </para>
- <variablelist>
- <varlistentry>
- <term>
- Foreign Keys
- </term>
- <listitem>
- Foreign keys are now implemented, with the exception of PARTIAL MATCH
- foreign keys. Many users have been asking for this feature, and we are
- pleased to offer it.
- </para>
- </listitem>
- </varlistentry>
+ <variablelist>
+ <varlistentry>
+ <term>
+ Foreign Keys
+ </term>
+ <listitem>
+ <para>
+ Foreign keys are now implemented, with the exception of PARTIAL MATCH
+ foreign keys. Many users have been asking for this feature, and we are
+ pleased to offer it.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Optimizer Overhaul
- </term>
- <listitem>
- Continuing on work started a year ago, the optimizer has been
- improved, allowing better query plan selection and faster performance
- with less memory usage.
- </para>
- </listitem>
- </varlistentry>
+ <varlistentry>
+ <term>
+ Optimizer Overhaul
+ </term>
+ <listitem>
+ <para>
+ Continuing on work started a year ago, the optimizer has been
+ improved, allowing better query plan selection and faster performance
+ with less memory usage.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
-
Updated <application>psql</application>
- </term>
- <listitem>
-
<application>psql</application>, our interactive terminal monitor, has been
-
updated with a variety of new features. See the <application>psql</application> manual page for details.
- </para>
- </listitem>
- </varlistentry>
+ <varlistentry>
+ <term>
+ Updated <application>psql</application>
+ </term>
+ <listitem>
+ <para>
+ <application>psql</application>, our interactive terminal monitor, has been
+ updated with a variety of new features. See the <application>psql</application> manual page for details.
+ </para>
+ </listitem>
+ </varlistentry>
- <varlistentry>
- <term>
- Join Syntax
- </term>
- <listitem>
- SQL92 join syntax is now supported, though only as
- <literal>INNER JOIN</> for this release. <literal>JOIN</>,
- <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
- and <literal>JOIN</>/<literal>ON</> are available, as are
- column correlation names.
- </para>
- </listitem>
- </varlistentry>
- </variablelist>
- </para>
+ <varlistentry>
+ <term>
+ Join Syntax
+ </term>
+ <listitem>
+ <para>
+ SQL92 join syntax is now supported, though only as
+ <literal>INNER JOIN</> for this release. <literal>JOIN</>,
+ <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
+ and <literal>JOIN</>/<literal>ON</> are available, as are
+ column correlation names.
+ </para>
+ </listitem>
+
+ </varlistentry>
+ </variablelist>
<sect2>
<title>Migration to version 7.0</title>
<para>
Interface and compatibility issues to consider for the new
release include:
+ </para>
- <itemizedlist>
- <listitem>
- <para>
- The date/time types <type>datetime</type> and
- <type>timespan</type> have been superseded by the
- SQL92-defined types <type>timestamp</type> and
- <type>interval</type>. Although there has been some effort to
- ease the transition by allowing
- <productname>PostgreSQL</productname> to recognize
- the deprecated type names and translate them to the new type
- names, this mechanism may not be completely transparent to
- your existing application.
- </para>
- </listitem>
-
- <listitem>
- <para>
- The optimizer has been substantially improved in the area of
- query cost estimation. In some cases, this will result in
- decreased query times as the optimizer makes a better choice
- for the preferred plan. However, in a small number of cases,
- usually involving pathological distributions of data, your
- query times may go up. If you are dealing with large amounts
- of data, you may want to check your queries to verify
- performance.
- </para>
- </listitem>
+ <itemizedlist>
+ <listitem>
+ <para>
+ The date/time types <type>datetime</type> and
+ <type>timespan</type> have been superseded by the
+ SQL92-defined types <type>timestamp</type> and
+ <type>interval</type>. Although there has been some effort to
+ ease the transition by allowing
+ <productname>PostgreSQL</productname> to recognize
+ the deprecated type names and translate them to the new type
+ names, this mechanism may not be completely transparent to
+ your existing application.
+ </para>
+ </listitem>
- <listitem>
- <para>
- The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
- interfaces have been upgraded and extended.
- </para>
- </listitem>
+ <listitem>
+ <para>
+ The optimizer has been substantially improved in the area of
+ query cost estimation. In some cases, this will result in
+ decreased query times as the optimizer makes a better choice
+ for the preferred plan. However, in a small number of cases,
+ usually involving pathological distributions of data, your
+ query times may go up. If you are dealing with large amounts
+ of data, you may want to check your queries to verify
+ performance.
+ </para>
+ </listitem>
- <listitem>
- <para>
- The string function <function>CHAR_LENGTH</function> is now a
- native function. Previous versions translated this into a call
- to <function>LENGTH</function>, which could result in
- ambiguity with other types implementing
- <function>LENGTH</function> such as the geometric types.
- </para>
- </listitem>
- </itemizedlist>
- </para>
+ <listitem>
+ <para>
+ The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
+ interfaces have been upgraded and extended.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ The string function <function>CHAR_LENGTH</function> is now a
+ native function. Previous versions translated this into a call
+ to <function>LENGTH</function>, which could result in
+ ambiguity with other types implementing
+ <function>LENGTH</function> such as the geometric types.
+ </para>
+ </listitem>
+ </itemizedlist>
</sect2>
<sect2>
<para>
There are some general 6.3 issues that I want to mention. These are
- only the big items that can not be described in one sentence. A review
+ only the big items that cannot be described in one sentence. A review
of the detailed changes list is still needed.
</para>
<para>
sgml-local-ecat-files:nil
End:
-->
-
projects / postgresql / blobdiff