2 $PostgreSQL: pgsql/doc/src/sgml/release.sgml,v 1.340 2005/08/23 11:53:33 momjian Exp $
5 <appendix id="release">
6 <title>Release Notes</title>
8 <sect1 id="release-8-1">
9 <title>Release 8.1</title>
12 <title>Release date</title>
13 <simpara>2005-1?-??, Current as of 2005-08-16</simpara>
17 <title>Overview</title>
20 Major changes in this release:
27 Improve concurrent access to the shared buffer cache (Tom)
32 This was accomplished by eliminating global locks and using a clock
33 sweep algorithm to find free buffers.
40 Allow indexes to be used for MIN/MAX (Tom)
45 In previous releases, the only way to use index for MIN/MAX was to rewrite
46 the query as SELECT col FROM tab ORDER BY col LIMIT 1. This now happens
96 <title>Migration to version 8.1</title>
99 A dump/restore using <application>pg_dump</application> is
100 required for those wishing to migrate data from any previous
105 The 8.0 release announced that the <function>to_char()</> function
106 for intervals would be removed in 8.1. However, since no better API has been
107 suggested, to_char(interval) has been enhanced in 8.1 and will remain in the
112 Observe the following incompatibilities:
118 Change add_missing_from to 'false'
121 Generate an error if a table used in a query without a FROM reference (Neil)
122 No more SELECT pg_class.*;
128 Cause input of a zero-length string ('') for float4/float8/oid to throw
129 an error, rather than treat it as a zero (Neil)
132 This change is consistent with the current handling of zero-length
133 strings for integers. The schedule for this change was announced in 8.0.
139 Make default_with_oids default to false (Neil)
142 With this option set to false, user-created tables no
143 have an the usually-invisible OID column unless WITH OIDS
144 is specified in CREATE TABLE. Though OIDs have existed in all previous
145 releases of PostgreSQL, their use is limited because they are only four
146 bytes long and the counter is unique across all installed databases.
147 The preferred way of uniquely identifying rows is via sequences and
148 SERIAL, which has been supported since PostgreSQL 6.4.
154 In psql, treat unquoted \{digit}+ sequences as octal (Bruce)
157 In previous releases, \{digit}+ sequences were treated as
158 decimal, and only \0{digit}+ were treated as octal. This
159 change was made for consistency.
165 Remove grammar productions for prefix and postfix % and ^ operators
169 These have never been documented and complicated the use of modulus
170 (%) with negative numbers.
176 Make "&<" and "&>" for polygons consistent with the box
177 "over" operators (Tom)
186 <title>Changes</title>
189 Below you will find a detailed account of the changes between
190 release 8.1 and the previous major release.
194 <title>Performance Improvements</title>
199 Improve rtree index capabilities and performance (Neil)
206 Improve the optimizer, including auto-resizing of hash joins (Tom)
212 Overhaul internal API in several areas to improve performance
218 Add in-memory bitmaps which allows multiple indexes to be merged in a
222 This allows multiple indexes to be combined to access a single
229 Change WAL CRC records from 64bit to 32bit to improve performance
236 Prevent writing large empty gaps in WAL pages (Tom)
242 Allow non-consecutive index columns to be used in a multi-column index
246 For example, this allows an index on columns a,b,c to be used in a
247 query with WHERE a = 4 and c = 10
253 Skip WAL logging for CREATE TABLE AS / SELECT INTO (Simon)
256 Since a crash during CREATE TABLE would cause the table to be dropped
257 during recovery, there is no reason to WAL log as the table is loaded.
263 Allow concurrent GIST index access, greatly improving performance
270 Add GUC full_page_writes to control writing full pages to WAL (Bruce)
273 To prevent partial disk writes from corrupting the database,
274 PostgreSQL writes a complete copy of each database disk page to WAL
275 the first time it is modified after a checkpoint. This turns off that
276 functionality for users with battery-backed disk caches where partial
277 page writes can not happen.
283 Add constraint_exclusion to restrict child table lookups based on
284 table constraints (Simon)
287 This allows for a type of table partitioning. If child table placed
288 in a different tablespaces using appropriate CHECK constraints, the
289 optimizer will skip child table accesses if the constraint guarantees
290 no matching rows exist in the child table.
296 Use O_DIRECT if available when using O_SYNC for wal_sync_method (ITAGAKI
300 O_DIRECT causes disk writes to bypass the kernel cache, and for WAL
301 writes, this improves performance.
307 Improve COPY FROM performance (Alon Goldshuv)
310 This was accomplished by reading COPY input in larger chunks, rather
311 than character by character.
320 <title>Server Changes</title>
325 Prevent problems due to transaction ID wraparound (Tom)
328 This was accomplished by warning the transaction counter is
329 near the the earliest transaction id determined by the
330 database whose last vacuum is the oldest. If the limit is
331 reached, the server will no longer accept queries.
337 Add warning about the need to increase "max_fsm_relations" and
338 "max_fsm_pages" during VACUUM (Ron Mayer)
344 Add temp_buffers GUC variable to allow users to determine the size
345 of the local buffer area for temporary table access (Tom)
351 Add session start time and client IP address to pg_stat_activity (Magnus)
357 Enhance pg_locks display (Tom)
363 Log queries for client-side PREPARE and EXECUTE (Simon)
369 Allow Kerberos name and user name case sensitivity to be specified from
376 Add log_line_prefix options for millisecond timestamps (%m) and
377 remote host (%h) (Ed L.)
383 Add WAL logging for GIST indexes (Teodor, Oleg)
386 GIST indexes now work for online backup and crash recovery
392 Add GUC krb_server_hostname so the server hostname can be specified as
393 part of service principal (Todd Kover)
396 If not set, any service principal matching an entry in the keytab
397 can be used. This is new Kerberos matching behavior in this release.
403 Remove old *.backup files when we do pg_stop_backup() (Bruce)
406 This prevents a large number of *.backup files from existing in
413 Replace pg_shadow and pg_group by new role-capable catalogs pg_authid
420 Add SET ROLE (Stephen Frost)
426 Move /contrib/pgautovacuum into the main server (Alvaro Herrera)
432 Add GUC variables to control keep-alive times for idle, interval, and
433 count (Oliver Jowett)
439 Add per-user and per-database connection limits (Petr Jelinek)
445 Fix problem of OID wraparound conflicting with existing system objects
455 <title>Query Changes</title>
460 Add temporary views (Koju Iijima)
466 Fix HAVING without aggregate functions and no GROUP BY to behave
467 as if the main query returns a single group (Tom)
473 Add USING clause to allow additional tables to be specified to DELETE
474 (Euler Taveira de Oliveira)
480 Add support for \x hex escapes in backend and ecpg strings (Bruce)
483 This supports the typical C standard \x escape. Octal was already
490 Add shared row level locks using SELECT ... FOR SHARE (Alvaro)
493 While PostgreSQL's MVCC locking allows SELECT to never be blocked by writers
494 and therefore does not need shared row locks for typical operations,
495 shared locks are useful for applications that require shared row locking,
496 and to reduce the locking requirements to maintain referential integrity.
502 Add BETWEEN SYMMETRIC query syntax (Pavel Stehule)
505 This feature allows BETWEEN comparisons without requiring the first
506 value to be less than the second. For example, 2 BETWEEN [ASYMMETRIC] 3 AND 1
507 returns false, while 2 BETWEEN SYMMETRIC 3 AND 1 returns true. BETWEEN
508 ASYMMETRIC was already supported.
514 Add E'' syntax so eventually normal strings can treat backslashes
518 Currently PostgreSQL considers a backslash to be a special character
519 so the character after the backslash is treated specially. While this
520 allows easy entry of special values, it is non-standard and makes
521 porting of application from other databases more difficult. For this
522 reason, the PostgreSQL project is moving to remove the special meaning
523 of backslashes in strings, and allow only an E preceeding a string to
524 turn on the special handling of backslashes. For this reason, this
525 release adds several new GUC variables related to backslash
528 escape_string_warning - warn about backslashes in non-E strings
529 escape_string_syntax - does this release support the E'' syntax?
530 standard_conforming_strings - does this release treat backslashes
531 literally in non-E strings
534 The last two values are read-only and should assist in the porting of
535 applications. Applications can retrieve these values to know how
536 backslashes are processed. In a later release,
537 standard_conforming_strings will be true, meaning backslashes will be
538 treated literally in non-E strings. To prepare for this change, use
539 E'' strings in places that need special backslash processing, and turn
540 on escape_string_warning to find additional strings that need to be
541 converted to use E''.
547 Add NOWAIT option to SELECT ... FOR UPDATE/SHARE (Hans-Juergen Schoenig)
550 While SET statement_timeout allows a query taking over a certain
551 amount of time to be cancelled, the NO WAIT option allows a query to
552 be canceled as soon as a SELECT ... FOR UPDATE/SHARE can not
553 immediately acquire a row lock.
562 <title>Object Manipulation Changes</title>
567 Track dependencies of shared objects (Alvaro)
570 PostgreSQL allows global tables (users, databases, tablespaces) to
571 reference information in multiple databases. This addition adds
572 dependency information for global tables, so, for example, user
573 ownership can be tracked across databases, so a user who owns
574 something in any database can no longer be removed. Dependency
575 tracking already existed for database-local objects.
581 Allow ALTER OWNER commands to be performed by the object owner as
582 well as the superuser (Stephen Frost)
585 Prior releases only allowed super-user to change object owners, even
586 if the current owner was executing the command and the new owner
587 have permission to create objects of that type.
593 Add ALTER object SET SCHEMA capability for some object types
594 (tables, functions, types) (Bernd Helmle)
597 This allows objects to be moved to different schemas.
606 <title>Utility Command Changes</title>
611 Allow TRUNCATE to truncate multiple files in a single command (Alvaro)
614 Because of referential integrity checks, it is impossible to truncate
615 a table that is part of a referential integrity constraint. Using this
616 new functionality, TRUNCATE can be used to truncate all tables
617 involved in referential integrity if they are truncated in a
618 single TRUNCATE command.
624 Properly process carriage returns and line feeds in COPY CSV mode
628 In release 8.0, carriage returns and line feeds in CSV COPY TO were
629 processed in an inconsistent manner. (This was documented on the TODO
637 Add COPY WITH CSV HEADER to allow a heading line as the first line in
641 This allows handling of the common CSV usage of placing the column
642 names on the first line of the data file. For COPY TO, the first line
643 contains the column names, and for COPY FROM, the first line is
650 On Win32, display better sub-second precision in EXPLAIN ANALYZE (Magnus)
656 Add trigger duration display to EXPLAIN ANALYZE (Tom)
659 Prior releases lumped trigger execution time into the total execution
666 Add support for \x hex escapes in COPY (Sergey Ten)
669 Previous releases only supported octal escapes.
675 Have SHOW ALL include variable descriptions (Matthias Schmidt)
678 SHOW varname still only displays the variable's value and does not
679 include the description.
685 Add two-phase commit (Heikki Linnakangas, Alvaro, Tom)
687 Two-phase commit allows transactions to be "prepared" on several
688 computers, and once all computers have successfully prepared their
689 transactions (and can not be rolled back) all transactions can be
690 committed. Even if a machine crashes after a prepare, the prepared
691 transaction can be committed after it is restarted. New syntax
692 includes PREPARE TRANSACTION and COMMIT/ROLLBACK PREPARED. A new
693 system view pg_prepared_xacts has also been added.
699 Have initdb create new standard database called "postgres" and convert utilities
700 to use "postgres" rather than "template1" for standard lookups (Dave)
703 In prior releases, template1 was used both as a default
704 connection for utilities like createuser, and as a template for
705 new databases. This caused CREATE DATABASE to sometimes fail
706 because a new database cannot be created if anyone else is in
707 the template database. With this change, the default connection
708 database is now 'postgres', meaning it is much less likely
709 someone will be using template1 during CREATE DATABASE.
715 Make REINDEX DATABASE reindex all indexes in the database (Tom)
718 The old behavior of REINDEX database reindexed only system tables.
719 This new behavior seems more intuitive. A new command REINDEX SYSTEM
720 allows for reindexing just the system tables.
726 Create new reindexdb command-line utility by moving /contrib/reindexdb
727 into the server (Euler Taveira de Oliveira)
736 <title>Data Type and Function Changes</title>
741 Add MAX() and MIN() aggregates for array types (Koju Iijima)
750 Fix to_date() and to_timestamp() to behave reasonably when
751 CC and YY fields are both used (Karel Zak)
754 If the format specification contains CC and a year specification is
755 YYY or longer, ignore the CC. If the year specification is
756 YY or shorter, interpret CC as the previous century. ?
762 Add md5(bytea) (Abhijit Menon-Sen)
765 md5(text) already existed.
771 Fix CHAR() to properly pad out to the specified length when
772 using a multiple-byte character set (Yoshiyuki Asaba)
775 In prior releases, the padding of CHAR() was incorrect because it only
776 padded to the specified number of bytes without considering how many
777 characters were stored.
783 Add support for NUMERIC ^ NUMERIC based on power(numeric, numeric)
786 The function already existed, but there was no operator assigned to
793 Fix NUMERIC modulus by properly truncating the quotient during
797 In previous releases, modulus for large values sometimes returned
798 negative results due to the rounding of the quotient.
804 Add a function lastval(), which returns the value returned by the
805 last nextval() or setval() performed by the current session. (Dennis
809 lastval() is a simplified version of currval(). It automatically
810 determines the proper sequence name based on the most recent
817 Add to_timestamp(double precision)
820 Converts Unix seconds since 1970 to a timestamp with timezone.
826 Add pg_postmaster_start_time() function (Euler Taveira de Oliveira,
833 Allow the full use of time zone names in "AT TIME ZONE", not just
834 the short list previously available (Magnus)
837 Previously, only a predefined list of time zone names were supported
838 by AT TIME ZONE. Now any supported time zone name can be used, e.g.:
841 SELECT CURRENT_TIMESTAMP AT TIME ZONE 'Europe/London';
843 In the above query, the daylight savings time rules were in effect on
851 Add Oracle-compatible GREATEST and LEAST functions (Pavel Stehule)
854 These functions take a variable number of arguments and return the
855 greatest or least value.
861 Add pg_column_size() to return storage size of a column, including
862 possible compression (Mark Kirkwood)
868 Add regexp_replace() (Atsushi Ogawa)
871 This allows regular expression replacement, like sed. A four-argument
872 version also allows for global (replace all) and case-insensitive
879 Fix interval division and multiplication (Bruce)
882 Previous versions sometimes returned unjustified results, like
883 '4 months'::interval / 5 returning '1 mon -6 days'.
889 Add internal 'day' field to INTERVAL so a 1 day interval can be
890 distinguished from a 24 hour interval (Michael Glaesemann)
893 Days that contain a daylight savings time adjustment are not 24 hours,
894 but typically 23 or 25 hours. This change allows days (not fixed
895 24-hour periods) to be added to dates who's result includes a daylight
896 savings time adjustment period. Therefore, while in previous releases
897 '1 day' and '24 hours' where interchangeable interval periods, in this
898 release they are treated differently, e.g.
904 '2005-05-03 00:00:00 EST' + '1 day' = '2005-05-04 00:00:00-04'
910 '2005-05-03 00:00:00 EST' + '24 hours' = '2005-05-04 01:00:00-04'
916 Move /contrib/dbsize into the backend, and rename some of the functions
917 (Dave Page, Andreas Pflug)
920 The new functions are:
924 pg_complete_relation_size()
928 complete_relation_size() includes indexes and TOAST tables.
934 Add justify_days() and justify_hours (Michael Glaesemann)
937 These functions, respectively, adjust days to an appropriate number of
938 full months and dayss, and adjust hours to an appropriate number
939 of full days and hours.
945 Add files to do read-only I/O on the cluster directory (Dave Page,
956 Add pg_reload_conf() to force reloading of the configuration files (Dave Page,
963 Add pg_rotate_logfile() to force rotation of the server log file (Dave Page,
969 Change pg_stat_* views to show TOAST tables (Tom)
976 <title>Encoding and Locale Changes</title>
981 Rename some encodings to be more consistent and to follow international
985 Encoding name changes were:
992 The original names still work.
998 Add support for Win1252 encoding (Roland Volkmann)
1004 Add support for 3 and 4-byte UTF8 characters (John Hansen)
1007 Previously only one and two-byte UTF8 characters were supported.
1008 This is particularly important for support for some Chinese character.
1014 Allow direct conversion between EUC_JP and SJIS to improve performance (Atsushi
1024 <title>General Server-Side Language Changes</title>
1029 Fix ALTER LANGUAGE RENAME (Sergey Yatskevich)
1035 Allow function characteristics, like strictness and volatility,
1036 to be modified via ALTER FUNCTION (Neil)
1042 Increase the maximum number of function arguments to 100 (Tom)
1048 Allow SQL, plperl, PL/PgSQL functions to use OUT and INOUT parameters (Tom)
1051 OUT is an alternate way for a function to return values. Instead of
1052 using RETURNS, the function's parameters can be specified as OUT or
1053 INOUT, allowing multiple values to be returned by the function. While
1054 returning multiple values from a function was possible in previous
1055 releases, this greatly simplifies the process.
1063 <title>PL/PgSQL Server-Side Language Changes</title>
1068 Reduce memory usage of PL/PgSQL functions (Neil)
1071 Each function now has its own memory context that can be freed when
1078 Check function syntax as CREATE FUNCTION time, rather than at runtime
1082 Previously, syntax errors were reported only when the function was
1089 Allow OPEN to open non-SELECT queries like EXPLAIN and SHOW (Tom)
1095 No longer require functions to issue a RETURN statement (Tom)
1098 This is a byproduct of the newly added OUT and INOUT functionality.
1104 Add support for an optional INTO clause to PL/PgSQL's EXECUTE command
1111 Make CREATE TABLE AS set ROW_COUNT (Tom)
1117 Define SQLSTATE and SQLERRM to return the SQLSTATE and
1118 error message of the current exception (Pavel Stehule)
1121 These variables are only accessable inside exception blocks.
1127 Allow the parameters to the RAISE statement to be expressions
1134 Add a loop CONTINUE statement (Pavel Stehule)
1140 Allow block and loop labels (Pavel Stehule)
1149 <title>PL/Perl Server-Side Language Changes</title>
1154 Allow large result sets to be returned efficiently (Abhijit Menon-Sen)
1157 This allows functions to use return_next() to avoid building
1158 the entire result set in memory.
1164 Allow one-row-at-a-time retrieval of query results (Abhijit)
1167 This allows functions to use spi_query() and spi_fetchrow() to
1168 avoid accumulating the entire result set in memory.
1174 Force PL/Perl to handle strings as UTF8 if the server encoding is UTF8
1181 Add a validator function for PL/Perl (Andrew)
1184 This allows syntax errors to be reported at definition time, rather
1185 than execution time.
1191 Allow PL/Perl to return a Perl array when the function returns an array
1195 This basically maps PostgreSQL arrays to Perl arrays.
1204 <title><application>psql</> Changes</title>
1209 Add psql \set ON_ERROR_ROLLBACK to allow statements in a transaction to
1210 error without affecting the rest of the transaction (Greg Sabino Mullane)
1213 This is basically implemented by wrapping every statement in a subtransaction.
1219 Add support for \x hex strings in psql variables (Bruce)
1222 Octal escapes were already supported.
1228 Add psql support for troff "-ms" output format (Roger Leigh)
1234 Allow psql's history file location to be controlled by HISTFILE (Andreas
1238 This allows configuration of per-database history storage.
1244 Prevent psql \x (expanded mode) from affecting other backslash output
1251 Add -L option to psql to log sessions (Lorne Sunley)
1254 This option was added because some operating systems do not have
1255 simple command-line activity logging functionality.
1261 Have psql \d show tablespace of indexes (Qingqing Zhou)
1267 Allow psql \h to make a best guess on the proper help information (Greg
1271 This allows the user to just add \h to the front of the syntax error
1272 query and get help on the supported syntax. Previously any additional
1273 query text beyond the help topics supported had to be removed to use
1280 Add psql \pset numericlocale to allow numbers to be output in a
1281 locale-aware format (Eugen Nedelcu)
1284 For example, using C locale 100000 would be output as 100,000.0 and
1285 European locale might output this value as 100.000,0.
1294 <title><application>pg_dump</> Changes</title>
1299 Add -n / --schema switch to pg_restore (Richard van den Bergg)
1302 This allows only objects from a specified schema to be restored.
1308 Allow pg_dump to dump a consistent snapshot of large objects (Tom)
1314 Allow pg_dump to dump large objects even in text mode (Tom)
1320 Dump comments for large objects (Tom)
1326 Add --encoding to pg_dump (Magnus Hagander)
1329 This allows a database to be dumped in an encoding that is different
1330 from the server's encoding. This is valuable when transferring the dump
1331 to a machine with a different encoding.
1340 <title>libpq Changes</title>
1345 Add a "PGPASSFILE" environment variable to specify the password
1346 file's filename (Andrew Dunstan)
1352 Add lo_create(), that is similar to lo_creat() but allows the OID of the large
1353 object to be specified (Tom)
1362 <title>Source Code Changes</title>
1367 Add spinlock support for the Itanium processor using Intel compiler (Vikram
1374 Add Kerberos 5 support for Win32 (Magnus)
1380 Add Chinese FAQ (laser@pgsqldb.com)
1386 Rename Rendezvous to Bonjour to match OS/X feature renaming (Bruce)
1392 Add support for wal_fsync_writethrough for Darwin (Chris Campbell)
1398 Streamline the passing information within the server, the optimizer,
1399 and the lock system (Tom)
1405 Allow pg_config to be compiled using MSVC (Andrew Dunstan)
1408 This is required to build DBD::Pg using MSVC.
1414 Remove support for Kerberos V4 (Magnus)
1417 Kerberos 4 had security vulnerabilities and is no longer being
1418 maintained by the authors.
1424 Code cleanups (Coverity static analysis performed by EnterpriseDB)
1430 Modify postgresql.conf to use documention defaults on/off rather
1431 than true/false (Bruce)
1437 Enhance pg_config to be able to report more build-time values (Tom)
1443 Replace rtree index code with code from /contrib/rtree_gist (Tom)
1446 The improved capabilities of GIST indexes made this possible.
1455 <title>Contrib Changes</title>
1460 Add /contrib/pg_buffercache contrib module (Mark Kirkwood)
1463 This displays the contents of the buffer cache, for debugging and
1464 performance tuning purposes.
1470 Remove /contrib/array because it is obsolete (Tom)
1476 Cleanup the contrib/lo module (Tom)
1482 Move /contrib/findoidjoins to src/tools (Tom)
1488 Remove the <<, >>, &<, and &> operators for contrib/cube
1491 These operators were not useful.
1497 Improve /contrib/btree_gist (Janko Richter)
1503 /contrib/pgcrypto - Remove support for libmhash/libmcrypt (Marko Kreen)
1509 /contrib/pgcrypto - Add support for new encryption methods (Marko Kreen)
1513 SHA2 (SHA256, SHA384, SHA512)
1526 <sect1 id="release-8-0-3">
1527 <title>Release 8.0.3</title>
1530 <title>Release date</title>
1531 <simpara>2005-05-09</simpara>
1535 This release contains a variety of fixes from 8.0.2, including several
1536 security-related issues.
1540 <title>Migration to version 8.0.3</title>
1543 A dump/restore is not required for those running 8.0.X. However,
1544 it is one possible way of handling two significant security problems
1545 that have been found in the initial contents of 8.0.X system
1546 catalogs. A dump/initdb/reload sequence using 8.0.3's initdb will
1547 automatically correct these problems.
1551 The larger security problem is that the built-in character set encoding
1552 conversion functions can be invoked from SQL commands by unprivileged
1553 users, but the functions were not designed for such use and are not
1554 secure against malicious choices of arguments. The fix involves changing
1555 the declared parameter list of these functions so that they can no longer
1556 be invoked from SQL commands. (This does not affect their normal use
1557 by the encoding conversion machinery.)
1561 The lesser problem is that the <filename>contrib/tsearch2</> module
1562 creates several functions that are misdeclared to return
1563 <type>internal</> when they do not accept <type>internal</> arguments.
1564 This breaks type safety for all functions using <type>internal</>
1569 It is strongly recommended that all installations repair these errors,
1570 either by initdb or by following the manual repair procedure given
1571 below. The errors at least allow unprivileged database users to crash
1572 their server process, and may allow unprivileged users to gain the
1573 privileges of a database superuser.
1577 If you wish not to do an initdb, perform the same manual repair
1578 procedures shown in the <link linkend="release-7-4-8">7.4.8 release
1584 <title>Changes</title>
1587 <listitem><para>Change encoding function signature to prevent
1588 misuse</para></listitem>
1589 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
1590 <type>INTERNAL</> function results</para></listitem>
1591 <listitem><para>Guard against incorrect second parameter to
1592 <function>record_out</></para></listitem>
1593 <listitem><para>Repair ancient race condition that allowed a transaction to be
1594 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
1595 than for other purposes</para>
1596 <para>This is an extremely serious bug since it could lead to apparent
1597 data inconsistencies being briefly visible to applications.</para></listitem>
1598 <listitem><para>Repair race condition between relation extension and
1600 <para>This could theoretically have caused loss of a page's worth of
1601 freshly-inserted data, although the scenario seems of very low probability.
1602 There are no known cases of it having caused more than an Assert failure.
1604 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
1606 The comparison code was wrong in the case where the
1607 <literal>--enable-integer-datetimes</> configuration switch had been used.
1608 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
1609 it will need to be <command>REINDEX</>ed after installing this update, because
1610 the fix corrects the sort order of column values.
1612 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
1613 <type>TIME WITH TIME ZONE</> values</para></listitem>
1614 <listitem><para>Fix mis-display of negative fractional seconds in
1615 <type>INTERVAL</> values</para>
1617 This error only occurred when the
1618 <literal>--enable-integer-datetimes</> configuration switch had been used.
1620 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
1621 correctly (Neil)</para></listitem>
1622 <listitem><para>Still more 64-bit fixes for
1623 <filename>contrib/intagg</></para></listitem>
1624 <listitem><para>Prevent incorrect optimization of functions returning
1625 <type>RECORD</></para></listitem>
1626 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
1627 <listitem><para>Fix Borland makefile for libpq</para></listitem>
1628 <listitem><para>Fix <filename>contrib/btree_gist</> for <type>timetz</> type
1629 (Teodor)</para></listitem>
1630 <listitem><para>Make <command>pg_ctl</> check the PID found in
1631 <filename>postmaster.pid</> to see if it is still a live
1632 process</para></listitem>
1633 <listitem><para>Fix <command>pg_dump</>/<command>pg_restore</> problems caused
1634 by addition of dump timestamps</para></listitem>
1635 <listitem><para>Fix interaction between materializing holdable cursors and
1636 firing deferred triggers during transaction commit</para></listitem>
1637 <listitem><para>Fix memory leak in SQL functions returning pass-by-reference
1638 data types</para></listitem>
1644 <sect1 id="release-8-0-2">
1645 <title>Release 8.0.2</title>
1648 <title>Release date</title>
1649 <simpara>2005-04-07</simpara>
1653 This release contains a variety of fixes from 8.0.1.
1657 <title>Migration to version 8.0.2</title>
1660 A dump/restore is not required for those running 8.0.*.
1661 This release updates the major version number of the
1662 <productname>PostgreSQL</productname> libraries, so it might be
1663 necessary to re-link some user applications if they cannot
1664 find the properly-numbered shared library.
1669 <title>Changes</title>
1672 <listitem><para>Increment the major version number of all interface
1673 libraries (Bruce)</para>
1675 This should have been done in 8.0.0. It is required so 7.4.X versions
1676 of PostgreSQL client applications, like <application>psql</>,
1677 can be used on the same machine as 8.0.X applications. This might require
1678 re-linking user applications that use these libraries.
1680 <listitem><para>Add Windows-only <varname>wal_sync_method</> setting of
1681 <option>fsync_writethrough</> (Magnus, Bruce)</para>
1683 This setting causes <productname>PostgreSQL</productname> to write through
1684 any disk-drive write cache when writing to WAL.
1685 This behavior was formerly called <option>fsync</>, but was
1686 renamed because it acts quite differently from <option>fsync</> on other
1690 <listitem><para>Enable the <varname>wal_sync_method</> setting of
1691 <option>open_datasync</> on Windows, and make it the default for that
1692 platform (Magnus, Bruce)</para>
1694 Because the default is no longer <option>fsync_writethrough</>,
1695 data loss is possible during a power failure if the disk drive has
1696 write caching enabled. To turn off the write cache on Windows,
1697 from the <application>Device Manager</>, choose the drive properties,
1698 then <literal>Policies</>.
1701 <listitem><para>New cache management algorithm <acronym>2Q</> replaces
1702 <acronym>ARC</> (Tom)</para>
1704 This was done to avoid a pending US patent on <acronym>ARC</>. The
1705 <acronym>2Q</> code might be a few percentage points slower than
1706 <acronym>ARC</> for some work loads. A better cache management algorithm
1709 <listitem><para>Planner adjustments to improve behavior on freshly-created
1710 tables (Tom)</para></listitem>
1711 <listitem><para>Allow plpgsql to assign to an element of an array that is
1712 initially <literal>NULL</> (Tom)</para>
1714 Formerly the array would remain <literal>NULL</>, but now it becomes a
1715 single-element array. The main SQL engine was changed to handle
1716 <command>UPDATE</> of a null array value this way in 8.0, but the similar
1717 case in plpgsql was overlooked.
1720 <listitem><para>Convert <literal>\r\n</> and <literal>\r</> to <literal>\n</>
1721 in plpython function bodies (Michael Fuhr)</para>
1723 This prevents syntax errors when plpython code is written on a Windows or
1727 <listitem><para>Allow SPI cursors to handle utility commands that return rows,
1728 such as <command>EXPLAIN</> (Tom)</para></listitem>
1729 <listitem><para>Fix <command>CLUSTER</> failure after <command>ALTER TABLE
1730 SET WITHOUT OIDS</> (Tom)</para></listitem>
1731 <listitem><para>Reduce memory usage of <command>ALTER TABLE ADD COLUMN</>
1732 (Neil)</para></listitem>
1733 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</> (Tom)</para></listitem>
1734 <listitem><para>Document the Windows-only <literal>register</> and
1735 <literal>unregister</> options of <application>pg_ctl</> (Magnus)</para></listitem>
1736 <listitem><para>Ensure operations done during backend shutdown are counted by
1737 statistics collector</para>
1739 This is expected to resolve reports of <application>pg_autovacuum</>
1740 not vacuuming the system catalogs often enough — it was not being
1741 told about catalog deletions caused by temporary table removal during
1744 <listitem><para>Change the Windows default for configuration parameter
1745 <varname>log_destination</> to <option>eventlog</> (Magnus)</para>
1747 By default, a server running on Windows will now send log output to the
1748 Windows event logger rather than standard error.
1750 <listitem><para>Make Kerberos authentication work on Windows (Magnus)</para></listitem>
1751 <listitem><para>Allow <command>ALTER DATABASE RENAME</> by superusers
1752 who aren't flagged as having CREATEDB privilege (Tom)</para></listitem>
1753 <listitem><para>Modify WAL log entries for <command>CREATE</> and
1754 <command>DROP DATABASE</> to not specify absolute paths (Tom)</para>
1755 <para>This allows point-in-time recovery on a different machine with possibly
1756 different database location. Note that <command>CREATE TABLESPACE</> still
1757 poses a hazard in such situations.
1759 <listitem><para>Fix crash from a backend exiting with an open transaction
1760 that created a table and opened a cursor on it (Tom)</para></listitem>
1761 <listitem><para>Fix <function>array_map()</> so it can call PL functions
1762 (Tom)</para></listitem>
1763 <listitem><para>Several <filename>contrib/tsearch2</> and
1764 <filename>contrib/btree_gist</> fixes (Teodor)
1766 <listitem><para>Fix crash of some <filename>contrib/pgcrypto</>
1767 functions on some platforms (Marko Kreen)</para></listitem>
1768 <listitem><para>Fix <filename>contrib/intagg</> for 64-bit platforms
1769 (Tom)</para></listitem>
1770 <listitem><para>Fix ecpg bugs in parsing of <command>CREATE</> statement
1771 (Michael)</para></listitem>
1772 <listitem><para>Work around gcc bug on powerpc and amd64 causing problems in
1773 ecpg (Christof Petig)</para></listitem>
1774 <listitem><para>Do not use locale-aware versions of <function>upper()</>,
1775 <function>lower()</>, and <function>initcap()</> when the locale is
1776 <literal>C</> (Bruce)</para>
1778 This allows these functions to work on platforms that generate errors
1779 for non-7-bit data when the locale is <literal>C</>.
1781 <listitem><para>Fix <function>quote_ident()</> to quote names that match keywords (Tom)</para></listitem>
1782 <listitem><para>Fix <function>to_date()</> to behave reasonably when
1783 <literal>CC</> and <literal>YY</> fields are both used (Karel)</para></listitem>
1784 <listitem><para>Prevent <function>to_char(interval)</> from failing
1785 when given a zero-month interval (Tom)</para></listitem>
1786 <listitem><para>Fix wrong week returned by <function>date_trunc('week')</>
1789 <function>date_trunc('week')</>
1790 returned the wrong year for the first few days of January in some years.
1792 <listitem><para>Use the correct default mask length for class <literal>D</>
1793 addresses in <type>INET</> data types (Tom)</para></listitem>
1799 <sect1 id="release-8-0-1">
1800 <title>Release 8.0.1</title>
1803 <title>Release date</title>
1804 <simpara>2005-01-31</simpara>
1808 This release contains a variety of fixes from 8.0.0, including several
1809 security-related issues.
1813 <title>Migration to version 8.0.1</title>
1816 A dump/restore is not required for those running 8.0.0.
1821 <title>Changes</title>
1824 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
1826 On platforms that will automatically execute initialization functions of a
1827 shared library (this includes at least Windows and ELF-based Unixen),
1828 <command>LOAD</> can be used to make the server execute arbitrary code.
1829 Thanks to NGS Software for reporting this.</para></listitem>
1830 <listitem><para>Check that creator of an aggregate function has the right to
1831 execute the specified transition functions</para>
1833 This oversight made it possible to bypass denial of EXECUTE
1834 permission on a function.</para></listitem>
1835 <listitem><para>Fix security and 64-bit issues in
1836 contrib/intagg</para></listitem>
1837 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
1838 Jurka)</para></listitem>
1839 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
1840 many parameters (Neil)</para></listitem>
1841 <listitem><para>Make <command>ALTER TABLE ADD COLUMN</> enforce domain
1842 constraints in all cases</para></listitem>
1843 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
1845 The result of the join was mistakenly supposed to be sorted the same as the
1846 left input. This could not only deliver mis-sorted output to the user, but
1847 in case of nested merge joins could give outright wrong answers.
1849 <listitem><para>Improve planning of grouped aggregate queries</para></listitem>
1850 <listitem><para><command>ROLLBACK TO <replaceable>savepoint</></command>
1851 closes cursors created since the savepoint</para></listitem>
1852 <listitem><para>Fix inadequate backend stack size on Windows</para></listitem>
1853 <listitem><para>Avoid SHGetSpecialFolderPath() on Windows
1854 (Magnus)</para></listitem>
1855 <listitem><para>Fix some problems in running pg_autovacuum as a Windows
1856 service (Dave Page)</para></listitem>
1857 <listitem><para>Multiple minor bug fixes in
1858 pg_dump/pg_restore</para></listitem>
1859 <listitem><para>Fix ecpg segfault with named structs used in
1860 typedefs (Michael)</para></listitem>
1866 <sect1 id="release-8-0">
1867 <title>Release 8.0</title>
1870 <title>Release date</title>
1871 <simpara>2005-01-19</simpara>
1875 <title>Overview</title>
1878 Major changes in this release:
1884 Microsoft Windows Native Server
1889 This is the first <productname>PostgreSQL</productname> release
1890 to run natively on <trademark class=registered>Microsoft Windows</> as
1891 a server. It can run as a <productname>Windows</> service. This
1892 release supports NT-based Windows releases like
1893 <productname>Windows 2000</>, <productname>Windows XP</>, and
1894 <productname>Windows 2003</>. Older releases like
1895 <productname>Windows 95</>, <productname>Windows 98</>, and
1896 <productname>Windows ME</> are not supported because these operating
1897 systems do not have the infrastructure to support
1898 <productname>PostgreSQL</productname>. A separate installer
1899 project has been created to ease installation on
1900 <productname>Windows</> — see <ulink
1901 url="http://www.postgresql.org/ftp/win32/"></ulink>.
1905 Although tested throughout our release cycle, the Windows port
1906 does not have the benefit of years of use in production
1907 environments that <productname>PostgreSQL</productname> has on
1908 Unix platforms. Therefore it should be treated with the same
1909 level of caution as you would a new product.
1913 Previous releases required the Unix emulation toolkit
1914 <productname>Cygwin</> in order to run the server on Windows
1915 operating systems. <productname>PostgreSQL</productname> has
1916 supported native clients on Windows for many years.
1928 Savepoints allow specific parts of a transaction to be aborted
1929 without affecting the remainder of the transaction. Prior
1930 releases had no such capability; there was no way to recover
1931 from a statement failure within a transaction except by
1932 aborting the whole transaction. This feature is valuable for
1933 application writers who require error recovery within a
1934 complex transaction.
1941 Point-In-Time Recovery
1946 In previous releases there was no way to recover from disk
1947 drive failure except to restore from a previous backup or use
1948 a standby replication server. Point-in-time recovery allows
1949 continuous backup of the server. You can recover either to
1950 the point of failure or to some transaction in the past.
1962 Tablespaces allow administrators to select different file systems
1963 for storage of individual tables, indexes, and databases.
1964 This improves performance and control over disk space
1965 usage. Prior releases used <application>initlocation</> and
1966 manual symlink management for such tasks.
1973 Improved Buffer Management, <command>CHECKPOINT</command>,
1974 <command>VACUUM</command>
1979 This release has a more intelligent buffer replacement strategy,
1980 which will make better use of available shared buffers and
1981 improve performance. The performance impact of vacuum and
1982 checkpoints is also lessened.
1994 A column's data type can now be changed with <command>ALTER
2002 New Perl Server-Side Language
2007 A new version of the <application>plperl</> server-side language now
2008 supports a persistent shared storage area, triggers, returning records
2009 and arrays of records, and SPI calls to access the database.
2016 Comma-separated-value (CSV) support in <command>COPY</command>
2021 <command>COPY</command> can now read and write
2022 comma-separated-value files. It has the flexibility to
2023 interpret non-standard quoting and separation characters too.
2032 <title>Migration to version 8.0</title>
2035 A dump/restore using <application>pg_dump</application> is
2036 required for those wishing to migrate data from any previous
2041 Observe the following incompatibilities:
2047 In <option>READ COMMITTED</> serialization mode, volatile functions
2048 now see the results of concurrent transactions committed up to the
2049 beginning of each statement within the function, rather than up to the
2050 beginning of the interactive command that called the function.
2056 Functions declared <option>STABLE</> or <option>IMMUTABLE</> always
2057 use the snapshot of the calling query, and therefore do not see the
2058 effects of actions taken after the calling query starts, whether in
2059 their own transaction or other transactions. Such a function must be
2060 read-only, too, meaning that it cannot use any SQL commands other than
2067 Non-deferred <option>AFTER</> triggers are now fired immediately
2068 after completion of the triggering query, rather than upon
2069 finishing the current interactive command. This makes a
2070 difference when the triggering query occurred within a function:
2071 the trigger is invoked before the function proceeds to its next
2078 Server configuration parameters <varname>virtual_host</> and
2079 <varname>tcpip_socket</> have been replaced with a more general
2080 parameter <varname>listen_addresses</>. Also, the server now listens on
2081 <literal>localhost</> by default, which eliminates the need for the
2082 <literal>-i</> postmaster switch in many scenarios.
2088 Server configuration parameters <varname>SortMem</> and
2089 <varname>VacuumMem</> have been renamed to <varname>work_mem</>
2090 and <varname>maintenance_work_mem</> to better reflect their
2091 use. The original names are still supported in
2092 <command>SET</command> and <command>SHOW</command>.
2098 Server configuration parameters <varname>log_pid</>,
2099 <varname>log_timestamp</>, and <varname>log_source_port</> have been
2100 replaced with a more general parameter <varname>log_line_prefix</>.
2106 Server configuration parameter <varname>syslog</> has been
2107 replaced with a more logical <varname>log_destination</> variable to
2108 control the log output destination.
2114 Server configuration parameter <varname>log_statement</> has been
2115 changed so it can selectively log just database modification or
2116 data definition statements. Server configuration parameter
2117 <varname>log_duration</> now prints only when <varname>log_statement</>
2124 Server configuration parameter <varname>max_expr_depth</> parameter has
2125 been replaced with <varname>max_stack_depth</> which measures the
2126 physical stack size rather than the expression nesting depth. This
2127 helps prevent session termination due to stack overflow caused by
2128 recursive functions.
2134 The <function>length()</> function no longer counts trailing spaces in
2135 <type>CHAR(n)</> values.
2141 Casting an integer to <type>BIT(N)</> selects the rightmost N bits of the
2142 integer, not the leftmost N bits as before.
2148 Updating an element or slice of a NULL array value now produces
2149 a non-NULL array result, namely an array containing
2150 just the assigned-to positions.
2156 Syntax checking of array input values has been tightened up
2157 considerably. Junk that was previously allowed in odd places with
2158 odd results now causes an error. Empty-string element values
2159 must now be written as <literal>""</>, rather than writing nothing.
2160 Also changed behavior with respect to whitespace surrounding
2161 array elements: trailing whitespace is now ignored, for symmetry
2162 with leading whitespace (which has always been ignored).
2168 Overflow in integer arithmetic operations is now detected and
2169 reported as an error.
2175 The arithmetic operators associated with the single-byte
2176 <type>"char"</> data type have been removed.
2182 The <function>extract()</> function (also called
2183 <function>date_part</>) now returns the proper year for BC dates.
2184 It previously returned one less than the correct year. The
2185 function now also returns the proper values for millennium and
2192 <type>CIDR</> values now must have their non-masked bits be zero.
2193 For example, we no longer allow
2194 <literal>204.248.199.1/31</literal> as a <type>CIDR</> value. Such
2195 values should never have been accepted by
2196 <productname>PostgreSQL</productname> and will now be rejected.
2202 <command>EXECUTE</command> now returns a completion tag that
2203 matches the executed statement.
2209 <application>psql</>'s <command>\copy</> command now reads or
2210 writes to the query's <literal>stdin/stdout</>, rather than
2211 <application>psql</>'s <literal>stdin/stdout</>. The previous
2212 behavior can be accessed via new
2213 <option>pstdin</>/<option>pstdout</> parameters.
2219 The JDBC client interface has been removed from the core
2220 distribution, and is now hosted at <ulink url=
2221 "http://jdbc.postgresql.org"></ulink>.
2227 The Tcl client interface has also been removed. There are several
2228 Tcl interfaces now hosted at <ulink url=
2229 "http://gborg.postgresql.org"></ulink>.
2235 The server now uses its own time zone database, rather than the
2236 one supplied by the operating system. This will provide consistent
2237 behavior across all platforms. In most cases, there should be
2238 little noticeable difference in time zone behavior, except that
2239 the time zone names used by <command>SET</>/<command>SHOW</>
2240 <varname>TimeZone</> may
2241 be different from what your platform provides.
2247 <application>Configure</>'s threading option no longer requires
2248 users to run tests or edit configuration files; threading options
2249 are now detected automatically.
2255 Now that tablespaces have been implemented,
2256 <application>initlocation</> has been removed.
2262 The API for user-defined GiST indexes has been changed. The
2263 Union and PickSplit methods are now passed a pointer to a
2264 special <structname>GistEntryVector</structname> structure,
2265 rather than a <type>bytea</type>.
2274 <title>Deprecated Features</title>
2277 Some aspects of <productname>PostgreSQL</productname>'s behavior
2278 have been determined to be suboptimal. For the sake of backward
2279 compatibility these have not been removed in 8.0, but they are
2280 considered deprecated and will be removed in the next major
2287 The 8.1 release will remove the <function>to_char()</> function
2294 The server now warns of empty strings passed to
2295 <type>oid</type>/<type>float4</type>/<type>float8</type> data
2296 types, but continues to interpret them as zeroes as before.
2297 In the next major release, empty strings will be considered
2298 invalid input for these data types.
2304 By default, tables in <productname>PostgreSQL</productname> 8.0
2305 and earlier are created with <type>OID</>s. In the next release,
2306 this will <emphasis>not</emphasis> be the case: to create a table
2307 that contains <type>OID</>s, the <option>WITH OIDS</> clause must
2308 be specified or the <varname>default_with_oids</varname>
2309 configuration parameter must be set. Users are encouraged to
2310 explicitly specify <option>WITH OIDS</> if their tables
2311 require OIDs for compatibility with future releases of
2312 <productname>PostgreSQL</productname>.
2319 <title>Changes</title>
2322 Below you will find a detailed account of the changes between
2323 release 8.0 and the previous major release.
2327 <title>Performance Improvements</title>
2332 Support cross-data-type index usage (Tom)
2335 Before this change, many queries would not use an index if the data
2336 types did not match exactly. This improvement makes index usage more
2337 intuitive and consistent.
2343 New buffer replacement strategy that improves caching (Jan)
2346 Prior releases used a least-recently-used (LRU) cache to keep
2347 recently referenced pages in memory. The LRU algorithm
2348 did not consider the number of times a specific cache entry was
2349 accessed, so large table scans could force out useful cache pages.
2350 The new cache algorithm uses four separate lists to track most
2351 recently used and most frequently used cache pages and dynamically
2352 optimize their replacement based on the work load. This should
2353 lead to much more efficient use of the shared buffer cache.
2354 Administrators who have tested shared buffer sizes in the past
2355 should retest with this new cache replacement policy.
2361 Add subprocess to write dirty buffers periodically to reduce
2362 checkpoint writes (Jan)
2365 In previous releases, the checkpoint process, which runs every few
2366 minutes, would write all dirty buffers to the operating system's
2367 buffer cache then flush all dirty operating system buffers to
2368 disk. This resulted in a periodic spike in disk usage that often
2369 hurt performance. The new code uses a background writer to trickle
2370 disk writes at a steady pace so checkpoints have far fewer dirty
2371 pages to write to disk. Also, the new code does not issue a global
2372 <function>sync()</> call, but instead <function>fsync()</>s just
2373 the files written since the last checkpoint. This should improve
2374 performance and minimize degradation during checkpoints.
2380 Add ability to prolong vacuum to reduce performance impact (Jan)
2383 On busy systems, <command>VACUUM</command> performs many I/O
2384 requests which can hurt performance for other users. This
2385 release allows you to slow down <command>VACUUM</command> to
2386 reduce its impact on other users, though this increases the
2387 total duration of <command>VACUUM</command>.
2393 Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
2396 This improves the way indexes are scanned when many duplicate
2397 values exist in the index.
2403 Use dynamically-generated table size estimates while planning (Tom)
2406 Formerly the planner estimated table sizes using the values seen
2407 by the last <command>VACUUM</command> or <command>ANALYZE</command>,
2408 both as to physical table size (number of pages) and number of rows.
2409 Now, the current physical table size is obtained from the kernel,
2410 and the number of rows is estimated by multiplying the table size
2411 by the row density (rows per page) seen by the last
2412 <command>VACUUM</command> or <command>ANALYZE</command>. This should
2413 produce more reliable estimates in cases where the table size has
2414 changed significantly since the last housekeeping command.
2420 Improved index usage with <literal>OR</> clauses (Tom)
2423 This allows the optimizer to use indexes in statements with many OR
2424 clauses that would not have been indexed in the past. It can also use
2425 multi-column indexes where the first column is specified and the second
2426 column is part of an <literal>OR</> clause.
2432 Improve matching of partial index clauses (Tom)
2435 The server is now smarter about using partial indexes in queries
2436 involving complex <option>WHERE</> clauses.
2442 Improve performance of the GEQO optimizer (Tom)
2445 The GEQO optimizer is used to plan queries involving many tables (by
2446 default, twelve or more). This release speeds up the way queries are
2447 analyzed to decrease time spent in optimization.
2453 Miscellaneous optimizer improvements
2456 There is not room here to list all the minor improvements made, but
2457 numerous special cases work better than in prior releases.
2463 Improve lookup speed for C functions (Tom)
2466 This release uses a hash table to lookup information for dynamically
2467 loaded C functions. This improves their speed so they perform nearly as
2468 quickly as functions that are built into the server executable.
2474 Add type-specific <command>ANALYZE</command> statistics
2475 capability (Mark Cave-Ayland)
2478 This feature allows more flexibility in generating statistics
2479 for non-standard data types.
2485 <command>ANALYZE</command> now collects statistics for
2486 expression indexes (Tom)
2489 Expression indexes (also called functional indexes) allow users to
2490 index not just columns but the results of expressions and function
2491 calls. With this release, the optimizer can gather and use statistics
2492 about the contents of expression indexes. This will greatly improve
2493 the quality of planning for queries in which an expression index is
2500 New two-stage sampling method for <command>ANALYZE</command>
2504 This gives better statistics when the density of valid rows is very
2505 different in different regions of a table.
2511 Speed up <command>TRUNCATE</command> (Tom)
2514 This buys back some of the performance loss observed in 7.4, while still
2515 keeping <command>TRUNCATE</command> transaction-safe.
2524 <title>Server Changes</title>
2529 Add WAL file archiving and point-in-time recovery (Simon Riggs)
2535 Add tablespaces so admins can control disk layout (Gavin)
2541 Add a built-in log rotation program (Andreas Pflug)
2544 It is now possible to log server messages conveniently without
2545 relying on either <application>syslog</> or an external log
2552 Add new read-only server configuration parameters to show server
2553 compile-time settings: <varname>block_size</>,
2554 <varname>integer_datetimes</>, <varname>max_function_args</>,
2555 <varname>max_identifier_length</>, <varname>max_index_keys</> (Joe)
2561 Make quoting of <literal>sameuser</>, <literal>samegroup</>, and
2562 <literal>all</> remove special meaning of these terms in
2563 <filename>pg_hba.conf</> (Andrew)
2569 Use clearer IPv6 name <literal>::1/128</> for
2570 <literal>localhost</> in default <filename>pg_hba.conf</> (Andrew)
2576 Use CIDR format in <filename>pg_hba.conf</> examples (Andrew)
2582 Rename server configuration parameters <varname>SortMem</> and
2583 <varname>VacuumMem</> to <varname>work_mem</> and
2584 <varname>maintenance_work_mem</> (Old names still supported) (Tom)
2587 This change was made to clarify that bulk operations such as index and
2588 foreign key creation use <varname>maintenance_work_mem</>, while
2589 <varname>work_mem</> is for workspaces used during query execution.
2595 Allow logging of session disconnections using server configuration
2596 <varname>log_disconnections</> (Andrew)
2602 Add new server configuration parameter <varname>log_line_prefix</> to
2603 allow control of information emitted in each log line (Andrew)
2606 Available information includes user name, database name, remote IP
2607 address, and session start time.
2613 Remove server configuration parameters <varname>log_pid</>,
2614 <varname>log_timestamp</>, <varname>log_source_port</>; functionality
2615 superseded by <varname>log_line_prefix</> (Andrew)
2621 Replace the <varname>virtual_host</> and <varname>tcpip_socket</>
2622 parameters with a unified <varname>listen_addresses</> parameter
2626 <varname>virtual_host</> could only specify a single IP address to
2627 listen on. <varname>listen_addresses</> allows multiple addresses
2634 Listen on localhost by default, which eliminates the need for the
2635 <option>-i</> postmaster switch in many scenarios (Andrew)
2638 Listening on localhost (<literal>127.0.0.1</>) opens no new
2639 security holes but allows configurations like Windows and JDBC,
2640 which do not support local sockets, to work without special
2647 Remove <varname>syslog</> server configuration parameter, and add more
2648 logical <varname>log_destination</> variable to control log output
2655 Change server configuration parameter <varname>log_statement</> to take
2656 values <varname>all</>, <varname>mod</>, <varname>ddl</>, or
2657 <varname>none</> to select which queries are logged (Bruce)
2660 This allows administrators to log only data definition changes or
2661 only data modification statements.
2667 Some logging-related configuration parameters could formerly be adjusted
2668 by ordinary users, but only in the <quote>more verbose</> direction.
2669 They are now treated more strictly: only superusers can set them.
2670 However, a superuser may use <command>ALTER USER</> to provide per-user
2671 settings of these values for non-superusers. Also, it is now possible
2672 for superusers to set values of superuser-only configuration parameters
2673 via <literal>PGOPTIONS</>.
2679 Allow configuration files to be placed outside the data directory (mlw)
2682 By default, configuration files are kept in the cluster's top directory.
2683 With this addition, configuration files can be placed outside the
2684 data directory, easing administration.
2690 Plan prepared queries only when first executed so constants can be
2691 used for statistics (Oliver Jowett)
2694 Prepared statements plan queries once and execute them many
2695 times. While prepared queries avoid the overhead of re-planning
2696 on each use, the quality of the plan suffers from not knowing the exact
2697 parameters to be used in the query. In this release, planning of
2698 unnamed prepared statements is delayed until the first execution,
2699 and the actual parameter values of that execution are used as
2700 optimization hints. This allows use of out-of-line parameter passing
2701 without incurring a performance penalty.
2707 Allow <command>DECLARE CURSOR</command> to take parameters
2711 It is now useful to issue <command>DECLARE CURSOR</command> in a
2712 <function>Parse</> message with parameters. The parameter values
2713 sent at <function>Bind</> time will be substituted into the
2714 execution of the cursor's query.
2720 Fix hash joins and aggregates of <type>inet</type> and
2721 <type>cidr</type> data types (Tom)
2724 Release 7.4 handled hashing of mixed <type>inet</type> and
2725 <type>cidr</type> values incorrectly. (This bug did not exist
2726 in prior releases because they wouldn't try to hash either
2733 Make <varname>log_duration</> print only when <varname>log_statement</>
2734 prints the query (Ed L.)
2743 <title>Query Changes</title>
2748 Add savepoints (nested transactions) (Alvaro)
2754 Unsupported isolation levels are now accepted and promoted to the
2755 nearest supported level (Peter)
2758 The SQL specification states that if a database doesn't support a
2759 specific isolation level, it should use the next more restrictive level.
2760 This change complies with that recommendation.
2766 Allow <command>BEGIN WORK</command> to specify transaction
2767 isolation levels like <command>START TRANSACTION</command> does
2774 Fix table permission checking for cases in which rules generate
2775 a query type different from the originally submitted query (Tom)
2781 Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
2785 In previous releases, because single quotes had to be used to
2786 quote a function's body, the use of single quotes inside the
2787 function text required use of two single quotes or other error-prone
2788 notations. With this release we add the ability to use "dollar
2789 quoting" to quote a block of text. The ability to use different
2790 quoting delimiters at different nesting levels greatly simplifies
2791 the task of quoting correctly, especially in complex functions.
2792 Dollar quoting can be used anywhere quoted text is needed.
2798 Make <literal>CASE val WHEN compval1 THEN ...</> evaluate <literal>val</> only once (Tom)
2801 <option>CASE</> no longer evaluates the tested expression multiple
2802 times. This has benefits when the expression is complex or is
2809 Test <option>HAVING</> before computing target list of an
2810 aggregate query (Tom)
2813 Fixes improper failure of cases such as <literal>SELECT SUM(win)/SUM(lose)
2814 ... GROUP BY ... HAVING SUM(lose) > 0</>. This should work but formerly
2815 could fail with divide-by-zero.
2821 Replace <varname>max_expr_depth</> parameter with
2822 <varname>max_stack_depth</> parameter, measured in kilobytes of stack
2826 This gives us a fairly bulletproof defense against crashing due to
2827 runaway recursive functions. Instead of measuring the depth of expression
2828 nesting, we now directly measure the size of the execution stack.
2834 Allow arbitrary row expressions (Tom)
2837 This release allows SQL expressions to contain arbitrary composite
2838 types, that is, row values. It also allows functions to more easily
2839 take rows as arguments and return row values.
2845 Allow <option>LIKE</>/<option>ILIKE</> to be used as the operator
2846 in row and subselect comparisons (Fabien Coelho)
2852 Avoid locale-specific case conversion of basic ASCII letters in
2853 identifiers and keywords (Tom)
2856 This solves the <quote>Turkish problem</> with mangling of words
2857 containing <literal>I</> and <literal>i</>. Folding of characters
2858 outside the 7-bit-ASCII set is still locale-aware.
2864 Improve syntax error reporting (Fabien, Tom)
2867 Syntax error reports are more useful than before.
2873 Change <command>EXECUTE</command> to return a completion tag
2874 matching the executed statement (Kris Jurka)
2877 Previous releases return an <command>EXECUTE</command> tag for
2878 any <command>EXECUTE</command> call. In this release, the tag
2879 returned will reflect the command executed.
2885 Avoid emitting <option>NATURAL CROSS JOIN</> in rule listings (Tom)
2888 Such a clause makes no logical sense, but in some cases the rule
2889 decompiler formerly produced this syntax.
2898 <title>Object Manipulation Changes</title>
2903 Add <command>COMMENT ON</> for casts, conversions, languages,
2904 operator classes, and large objects (Christopher)
2910 Add new server configuration parameter <varname>default_with_oids</> to
2911 control whether tables are created with <type>OID</>s by default (Neil)
2914 This allows administrators to control whether <command>CREATE
2915 TABLE</command> commands create tables with or without <type>OID</>
2916 columns by default. (Note: the current factory default setting for
2917 <varname>default_with_oids</> is <literal>TRUE</>, but the default
2918 will become <literal>FALSE</> in future releases.)
2924 Add <option>WITH</> / <option>WITHOUT OIDS</> clause to
2925 <command>CREATE TABLE AS</command> (Neil)
2931 Allow <command>ALTER TABLE DROP COLUMN</> to drop an <type>OID</>
2932 column (<command>ALTER TABLE SET WITHOUT OIDS</> still works)
2939 Allow composite types as table columns (Tom)
2945 Allow <command>ALTER ... ADD COLUMN</> with defaults and
2946 <option>NOT NULL</> constraints; works per SQL spec (Rod)
2949 It is now possible for <option>ADD COLUMN</> to create a column
2950 that is not initially filled with NULLs, but with a specified
2957 Add <command>ALTER COLUMN TYPE</> to change column's type (Rod)
2960 It is now possible to alter a column's data type without dropping
2961 and re-adding the column.
2967 Allow multiple <command>ALTER</> actions in a single <command>ALTER
2968 TABLE</command> command (Rod)
2971 This is particularly useful for <command>ALTER</> commands that
2972 rewrite the table (which include <option>ALTER COLUMN TYPE</> and
2973 <option>ADD COLUMN</> with a default). By grouping
2974 <command>ALTER</> commands together, the table need be rewritten
2981 Allow <command>ALTER TABLE</command> to add <type>SERIAL</type>
2985 This falls out from the new capability of specifying defaults for new
2992 Allow changing the owners of aggregates, conversions, databases,
2993 functions, operators, operator classes, schemas, types, and tablespaces
2994 (Christopher, Euler Taveira de Oliveira)
2997 Previously this required modifying the system tables directly.
3003 Allow temporary object creation to be limited to <option>SECURITY
3004 DEFINER</> functions (Sean Chittenden)
3010 Add <option>ALTER TABLE ... SET WITHOUT CLUSTER</> (Christopher)
3013 Prior to this release, there was no way to clear an auto-cluster
3014 specification except to modify the system tables.
3020 Constraint/Index/<type>SERIAL</> names are now
3021 <replaceable>table_column_type</>
3022 with numbers appended to guarantee uniqueness within the schema
3026 The SQL specification states that such names should be unique
3033 Add <function>pg_get_serial_sequence()</> to return a
3034 <type>SERIAL</> column's sequence name (Christopher)
3037 This allows automated scripts to reliably find the <type>SERIAL</>
3044 Warn when primary/foreign key data type mismatch requires costly lookup
3050 New <command>ALTER INDEX</> command to allow moving of indexes
3051 between tablespaces (Gavin)
3057 Make <command>ALTER TABLE OWNER</> change dependent sequence
3058 ownership too (Alvaro)
3068 <title>Utility Command Changes</title>
3073 Allow <command>CREATE SCHEMA</command> to create triggers,
3074 indexes, and sequences (Neil)
3080 Add <option>ALSO</> keyword to <command>CREATE RULE</command> (Fabien
3084 This allows <option>ALSO</> to be added to rule creation to contrast it with
3085 <option>INSTEAD</> rules.
3091 Add <option>NOWAIT</> option to <command>LOCK</command> (Tatsuo)
3094 This allows the <command>LOCK</command> command to fail if it
3095 would have to wait for the requested lock.
3101 Allow <command>COPY</command> to read and write
3102 comma-separated-value (CSV) files (Andrew, Bruce)
3108 Generate error if the <command>COPY</command> delimiter and NULL
3109 string conflict (Bruce)
3115 <command>GRANT</command>/<command>REVOKE</command> behavior
3116 follows the SQL spec more closely
3122 Avoid locking conflict between <command>CREATE INDEX</command>
3123 and <command>CHECKPOINT</command> (Tom)
3126 In 7.3 and 7.4, a long-running B-tree index build could block concurrent
3127 <command>CHECKPOINT</>s from completing, thereby causing WAL bloat because the
3128 WAL log could not be recycled.
3134 Database-wide <command>ANALYZE</command> does not hold locks
3138 This reduces the potential for deadlocks against other backends
3139 that want exclusive locks on tables. To get the benefit of this
3140 change, do not execute database-wide <command>ANALYZE</command>
3141 inside a transaction block (<command>BEGIN</command> block); it
3142 must be able to commit and start a new transaction for each
3149 <command>REINDEX</command> does not exclusively lock the index's
3150 parent table anymore
3153 The index itself is still exclusively locked, but readers of the
3154 table can continue if they are not using the particular index
3161 Erase MD5 user passwords when a user is renamed (Bruce)
3164 <productname>PostgreSQL</productname> uses the user name as salt
3165 when encrypting passwords via MD5. When a user's name is changed,
3166 the salt will no longer match the stored MD5 password, so the
3167 stored password becomes useless. In this release a notice is
3168 generated and the password is cleared. A new password must then
3169 be assigned if the user is to be able to log in with a password.
3175 New <application>pg_ctl</> <option>kill</> option for Windows (Andrew)
3178 Windows does not have a <literal>kill</> command to send signals to
3179 backends so this capability was added to <application>pg_ctl</>.
3185 Information schema improvements
3191 Add <option>--pwfile</> option to
3192 <application>initdb</application> so the initial password can be
3193 set by GUI tools (Magnus)
3199 Detect locale/encoding mismatch in
3200 <application>initdb</application> (Peter)
3206 Add <option>register</> command to <application>pg_ctl</> to
3207 register Windows operating system service (Dave Page)
3216 <title>Data Type and Function Changes</title>
3221 More complete support for composite types (row types) (Tom)
3224 Composite values can be used in many places where only scalar values
3231 Reject non-rectangular array values as erroneous (Joe)
3234 Formerly, <function>array_in</> would silently build a
3241 Overflow in integer arithmetic operations is now detected (Tom)
3247 The arithmetic operators associated with the single-byte
3248 <type>"char"</> data type have been removed.
3251 Formerly, the parser would select these operators in many situations
3252 where an <quote>unable to select an operator</> error would be more
3253 appropriate, such as <literal>null * null</>. If you actually want
3254 to do arithmetic on a <type>"char"</> column, you can cast it to
3261 Syntax checking of array input values considerably tightened up (Joe)
3264 Junk that was previously allowed in odd places with odd results
3265 now causes an <literal>ERROR</>, for example, non-whitespace
3266 after the closing right brace.
3272 Empty-string array element values must now be written as
3273 <literal>""</>, rather than writing nothing (Joe)
3276 Formerly, both ways of writing an empty-string element value were
3277 allowed, but now a quoted empty string is required. The case where
3278 nothing at all appears will probably be considered to be a NULL
3279 element value in some future release.
3285 Array element trailing whitespace is now ignored (Joe)
3288 Formerly leading whitespace was ignored, but trailing whitespace
3289 between an element value and the delimiter or right brace was
3290 significant. Now trailing whitespace is also ignored.
3296 Emit array values with explicit array bounds when lower bound is not one
3303 Accept <literal>YYYY-monthname-DD</> as a date string (Tom)
3309 Make <function>netmask</> and <function>hostmask</> functions
3310 return maximum-length mask length (Tom)
3316 Change factorial function to return <type>numeric</type> (Gavin)
3319 Returning <type>numeric</type> allows the factorial function to
3320 work for a wider range of input values.
3326 <function>to_char</>/<function>to_date()</> date conversion
3327 improvements (Kurt Roeckx, Fabien Coelho)
3333 Make <function>length()</> disregard trailing spaces in
3334 <type>CHAR(n)</> (Gavin)
3337 This change was made to improve consistency: trailing spaces are
3338 semantically insignificant in <type>CHAR(n)</> data, so they
3339 should not be counted by <function>length()</>.
3345 Warn about empty string being passed to
3346 <type>OID</>/<type>float4</>/<type>float8</> data types (Neil)
3349 8.1 will throw an error instead.
3355 Allow leading or trailing whitespace in
3356 <type>int2</>/<type>int4</>/<type>int8</>/<type>float4</>/<type>float8</>
3364 Better support for IEEE <literal>Infinity</> and <literal>NaN</>
3365 values in <type>float4</type>/<type>float8</type> (Neil)
3368 These should now work on all platforms that support IEEE-compliant
3369 floating point arithmetic.
3375 Add <option>week</> option to <function>date_trunc()</> (Robert Creager)
3381 Fix <function>to_char</function> for <literal>1 BC</>
3382 (previously it returned <literal>1 AD</>) (Bruce)
3388 Fix <function>date_part(year)</> for BC dates (previously it
3389 returned one less than the correct year) (Bruce)
3395 Fix <function>date_part()</> to return the proper millennium and
3396 century (Fabien Coelho)
3399 In previous versions, the century and millennium results had a wrong
3400 number and started in the wrong year, as compared to standard
3401 reckoning of such things.
3407 Add <function>ceiling()</> as an alias for <function>ceil()</>,
3408 and <function>power()</> as an alias for <function>pow()</> for
3409 standards compliance (Neil)
3415 Change <function>ln()</>, <function>log()</>,
3416 <function>power()</>, and <function>sqrt()</> to emit the correct
3417 <literal>SQLSTATE</> error codes for certain error conditions, as
3418 specified by SQL:2003 (Neil)
3424 Add <function>width_bucket()</> function as defined by SQL:2003 (Neil)
3430 Add <function>generate_series()</> functions to simplify working
3431 with numeric sets (Joe)
3437 Fix <function>upper/lower/initcap()</> functions to work with
3438 multibyte encodings (Tom)
3444 Add boolean and bitwise integer <option>AND</>/<option>OR</>
3445 aggregates (Fabien Coelho)
3451 New session information functions to return network addresses for client
3452 and server (Sean Chittenden)
3458 Add function to determine the area of a closed path (Sean Chittenden)
3464 Add function to send cancel request to other backends (Magnus)
3470 Add <type>interval</> plus <type>datetime</> operators (Tom)
3473 The reverse ordering, <type>datetime</> plus <type>interval</>,
3474 was already supported, but both are required by the SQL standard.
3480 Casting an integer to <type>BIT(N)</> selects the rightmost N bits
3485 In prior releases, the leftmost N bits were selected, but this was
3486 deemed unhelpful, not to mention inconsistent with casting from bit
3493 Require <type>CIDR</> values to have all non-masked bits be zero
3503 <title>Server-Side Language Changes</title>
3508 In <literal>READ COMMITTED</> serialization mode, volatile functions
3509 now see the results of concurrent transactions committed up to the
3510 beginning of each statement within the function, rather than up to the
3511 beginning of the interactive command that called the function.
3517 Functions declared <literal>STABLE</> or <literal>IMMUTABLE</> always
3518 use the snapshot of the calling query, and therefore do not see the
3519 effects of actions taken after the calling query starts, whether in
3520 their own transaction or other transactions. Such a function must be
3521 read-only, too, meaning that it cannot use any SQL commands other than
3522 <command>SELECT</>. There is a considerable performance gain from
3523 declaring a function <literal>STABLE</> or <literal>IMMUTABLE</>
3524 rather than <literal>VOLATILE</>.
3530 Non-deferred <option>AFTER</> triggers are now fired immediately
3531 after completion of the triggering query, rather than upon
3532 finishing the current interactive command. This makes a difference
3533 when the triggering query occurred within a function: the trigger
3534 is invoked before the function proceeds to its next operation. For
3535 example, if a function inserts a new row into a table, any
3536 non-deferred foreign key checks occur before proceeding with the
3543 Allow function parameters to be declared with names (Dennis Bjorklund)
3546 This allows better documentation of functions. Whether the names
3547 actually do anything depends on the specific function language
3554 Allow PL/pgSQL parameter names to be referenced in the function (Dennis Bjorklund)
3557 This basically creates an automatic alias for each named parameter.
3563 Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
3566 This allows us to catch simple syntax errors sooner.
3572 More support for composite types (row and record variables) in PL/pgSQL
3575 For example, it now works to pass a rowtype variable to another function
3576 as a single variable.
3582 Default values for PL/pgSQL variables can now reference previously
3589 Improve parsing of PL/pgSQL FOR loops (Tom)
3592 Parsing is now driven by presence of <literal>".."</> rather than
3593 data type of <option>FOR</> variable. This makes no difference for
3594 correct functions, but should result in more understandable error
3595 messages when a mistake is made.
3601 Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
3607 In PL/Tcl, SPI commands are now run in subtransactions. If an error
3608 occurs, the subtransaction is cleaned up and the error is reported
3609 as an ordinary Tcl error, which can be trapped with <literal>catch</>.
3610 Formerly, it was not possible to catch such errors.
3616 Accept <command>ELSEIF</> in PL/pgSQL (Neil)
3619 Previously PL/pgSQL only allowed <command>ELSIF</>, but many people
3620 are accustomed to spelling this keyword <command>ELSEIF</>.
3629 <title><application>psql</> Changes</title>
3634 Improve <application>psql</> information display about database
3635 objects (Christopher)
3641 Allow <application>psql</> to display group membership in
3642 <command>\du</> and <command>\dg</> (Markus Bertheau)
3648 Prevent <application>psql</> <command>\dn</command> from showing
3649 temporary schemas (Bruce)
3655 Allow <application>psql</> to handle tilde user expansion for file
3662 Allow <application>psql</> to display fancy prompts, including
3663 color, via <application>readline</> (Reece Hart, Chet Ramey)
3669 Make <application>psql</> <command>\copy</> match <command>COPY</command> command syntax
3676 Show the location of syntax errors (Fabien Coelho, Tom)
3682 Add <command>CLUSTER</command> information to <application>psql</>
3683 <command>\d</> display
3690 Change <application>psql</> <command>\copy stdin/stdout</> to read
3691 from command input/output (Bruce)
3697 Add <option>pstdin</>/<option>pstdout</> to read from
3698 <application>psql</>'s <literal>stdin</>/<literal>stdout</> (Mark
3705 Add global <application>psql</> configuration file, <filename>psqlrc.sample</filename>
3709 This allows a central file where global <application>psql</> startup commands can
3716 Have <application>psql</> <command>\d+</> indicate if the table
3717 has an <type>OID</> column (Neil)
3723 On Windows, use binary mode in <application>psql</> when reading files so control-Z
3724 is not seen as end-of-file
3730 Have <command>\dn+</> show permissions and description for schemas (Dennis
3737 Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
3743 Allow boolean settings to be set using upper or lower case (Michael Paesold)
3752 <title><application>pg_dump</> Changes</title>
3757 Use dependency information to improve the reliability of
3758 <application>pg_dump</> (Tom)
3761 This should solve the longstanding problems with related objects
3762 sometimes being dumped in the wrong order.
3768 Have <application>pg_dump</> output objects in alphabetical order if possible (Tom)
3771 This should make it easier to identify changes between
3778 Allow <application>pg_restore</> to ignore some SQL errors (Fabien Coelho)
3781 This makes <application>pg_restore</>'s behavior similar to the
3782 results of feeding a <application>pg_dump</> output script to
3783 <application>psql</>. In most cases, ignoring errors and plowing
3784 ahead is the most useful thing to do. Also added was a pg_restore
3785 option to give the old behavior of exiting on an error.
3791 <application>pg_restore</> <option>-l</> display now includes
3792 objects' schema names
3798 New begin/end markers in <application>pg_dump</> text output (Bruce)
3804 Add start/stop times for
3805 <application>pg_dump</>/<application>pg_dumpall</> in verbose mode
3812 Allow most <application>pg_dump</> options in
3813 <application>pg_dumpall</> (Christopher)
3819 Have <application>pg_dump</> use <command>ALTER OWNER</> rather
3820 than <command>SET SESSION AUTHORIZATION</> by default
3830 <title>libpq Changes</title>
3835 Make libpq's <option>SIGPIPE</> handling thread-safe (Bruce)
3841 Add <function>PQmbdsplen()</> which returns the display length
3842 of a character (Tatsuo)
3848 Add thread locking to <application>SSL</> and
3849 <application>Kerberos</> connections (Manfred Spraul)
3855 Allow <function>PQoidValue()</>, <function>PQcmdTuples()</>, and
3856 <function>PQoidStatus()</> to work on <command>EXECUTE</command>
3863 Add <function>PQserverVersion()</> to provide more convenient
3864 access to the server version number (Greg Sabino Mullane)
3870 Add <function>PQprepare/PQsendPrepared()</> functions to support
3871 preparing statements without necessarily specifying the data types
3872 of their parameters (Abhijit Menon-Sen)
3878 Many ECPG improvements, including <command>SET DESCRIPTOR</> (Michael)
3887 <title>Source Code Changes</title>
3892 Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
3898 Shell script commands converted to C versions for Windows support (Andrew)
3904 Create an extension makefile framework (Fabien Coelho, Peter)
3907 This simplifies the task of building extensions outside the original
3914 Support relocatable installations (Bruce)
3917 Directory paths for installed files (such as the
3918 <filename>/share</> directory) are now computed relative to the
3919 actual location of the executables, so that an installation tree
3920 can be moved to another place without reconfiguring and
3927 Use <option>--with-docdir</> to choose installation location of documentation; also
3928 allow <option>--infodir</> (Peter)
3934 Add <option>--without-docdir</> to prevent installation of documentation (Peter)
3940 Upgrade to <application>DocBook</> V4.2 SGML (Peter)
3946 New <literal>PostgreSQL</> <application>CVS</> tag (Marc)
3949 This was done to make it easier for organizations to manage their
3950 own copies of the <productname>PostgreSQL</productname>
3951 <application>CVS</> repository. File version stamps from the master
3952 repository will not get munged by checking into or out of a copied
3959 Clarify locking code (Manfred Koizar)
3965 Buffer manager cleanup (Neil)
3971 Decouple platform tests from CPU spinlock code (Bruce, Tom)
3977 Add inlined test-and-set code on PA-RISC for <application>gcc</>
3984 Improve i386 spinlock code (Manfred Spraul)
3990 Clean up spinlock assembly code to avoid warnings from newer
3991 <application>gcc</> releases (Tom)
3997 Remove JDBC from source tree; now a separate project
4003 Remove the libpgtcl client interface; now a separate project
4009 More accurately estimate memory and file descriptor usage (Tom)
4015 Improvements to the Mac OS X startup scripts (Ray A.)
4021 New <function>fsync()</> test program (Bruce)
4027 Major documentation improvements (Neil, Peter)
4033 Remove <application>pg_encoding</application>; not needed
4040 Remove <application>pg_id</application>; not needed anymore
4046 Remove <application>initlocation</application>; not needed
4053 Auto-detect thread flags (no more manual testing) (Bruce)
4059 Use Olson's public domain <application>timezone</> library (Magnus)
4065 With threading enabled, use thread flags on Unixware for
4066 backend executables too (Bruce)
4069 Unixware cannot mix threaded and non-threaded object files in the
4070 same executable, so everything must be compiled as threaded.
4076 <application>psql</> now uses a <application>flex</>-generated
4077 lexical analyzer to process command strings
4083 Reimplement the linked list data structure used throughout the
4087 This improves performance by allowing list append and length
4088 operations to be more efficient.
4094 Allow dynamically loaded modules to create their own server configuration
4095 parameters (Thomas Hallgren)
4101 New Brazilian version of FAQ (Euler Taveira de Oliveira)
4107 Add French FAQ (Guillaume Lelarge)
4113 New <application>pgevent</> for Windows logging
4119 Make libpq and ECPG build as proper shared libraries on OS X (Tom)
4128 <title>Contrib Changes</title>
4133 Overhaul of <filename>contrib/dblink</> (Joe)
4139 <filename>contrib/dbmirror</> improvements (Steven Singer)
4145 New <filename>contrib/xml2</> (John Gray, Torchbox)
4151 Updated <filename>contrib/mysql</filename>
4157 New version of <filename>contrib/btree_gist</> (Teodor)
4163 New <filename>contrib/trgm</>, trigram matching for
4164 <productname>PostgreSQL</productname> (Teodor)
4170 Many <filename>contrib/tsearch2</> improvements (Teodor)
4176 Add double metaphone to <filename>contrib/fuzzystrmatch</> (Andrew)
4182 Allow <filename>contrib/pg_autovacuum</> to run as a Windows service (Dave Page)
4188 Add functions to <filename>contrib/dbsize</> (Andreas Pflug)
4194 Removed <filename>contrib/pg_logger</>: obsoleted by integrated logging
4201 Removed <filename>contrib/rserv</>: obsoleted by various separate projects
4211 <sect1 id="release-7-4-8">
4212 <title>Release 7.4.8</title>
4215 <title>Release date</title>
4216 <simpara>2005-05-09</simpara>
4220 This release contains a variety of fixes from 7.4.7, including several
4221 security-related issues.
4225 <title>Migration to version 7.4.8</title>
4228 A dump/restore is not required for those running 7.4.X. However,
4229 it is one possible way of handling two significant security problems
4230 that have been found in the initial contents of 7.4.X system
4231 catalogs. A dump/initdb/reload sequence using 7.4.8's initdb will
4232 automatically correct these problems.
4236 The larger security problem is that the built-in character set encoding
4237 conversion functions can be invoked from SQL commands by unprivileged
4238 users, but the functions were not designed for such use and are not
4239 secure against malicious choices of arguments. The fix involves changing
4240 the declared parameter list of these functions so that they can no longer
4241 be invoked from SQL commands. (This does not affect their normal use
4242 by the encoding conversion machinery.)
4246 The lesser problem is that the <filename>contrib/tsearch2</> module
4247 creates several functions that are misdeclared to return
4248 <type>internal</> when they do not accept <type>internal</> arguments.
4249 This breaks type safety for all functions using <type>internal</>
4254 It is strongly recommended that all installations repair these errors,
4255 either by initdb or by following the manual repair procedures given
4256 below. The errors at least allow unprivileged database users to crash
4257 their server process, and may allow unprivileged users to gain the
4258 privileges of a database superuser.
4262 If you wish not to do an initdb, perform the following procedures instead.
4263 As the database superuser, do:
4267 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
4268 WHERE pronamespace = 11 AND pronargs = 5
4269 AND proargtypes[2] = 'cstring'::regtype;
4270 -- The command should report having updated 90 rows;
4271 -- if not, rollback and investigate instead of committing!
4275 Next, if you have installed <filename>contrib/tsearch2</>, do
4279 UPDATE pg_proc SET proargtypes[0] = 'internal'::regtype
4281 'dex_init(text)'::regprocedure,
4282 'snb_en_init(text)'::regprocedure,
4283 'snb_ru_init(text)'::regprocedure,
4284 'spell_init(text)'::regprocedure,
4285 'syn_init(text)'::regprocedure
4287 -- The command should report having updated 5 rows;
4288 -- if not, rollback and investigate instead of committing!
4292 If this command fails with a message like <quote>function
4293 "dex_init(text)" does not exist</>, then either <filename>tsearch2</>
4294 is not installed in this database, or you already did the update.
4298 The above procedures must be carried out in <emphasis>each</> database
4299 of an installation, including <literal>template1</>, and ideally
4300 including <literal>template0</> as well. If you do not fix the
4301 template databases then any subsequently created databases will contain
4302 the same errors. <literal>template1</> can be fixed in the same way
4303 as any other database, but fixing <literal>template0</> requires
4304 additional steps. First, from any database issue
4306 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
4308 Next connect to <literal>template0</> and perform the above repair
4309 procedures. Finally, do
4311 -- re-freeze template0:
4313 -- and protect it against future alterations:
4314 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
4320 <title>Changes</title>
4323 <listitem><para>Change encoding function signature to prevent
4324 misuse</para></listitem>
4325 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
4326 <type>INTERNAL</> function results</para></listitem>
4327 <listitem><para>Repair ancient race condition that allowed a transaction to be
4328 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
4329 than for other purposes</para>
4330 <para>This is an extremely serious bug since it could lead to apparent
4331 data inconsistencies being briefly visible to applications.</para></listitem>
4332 <listitem><para>Repair race condition between relation extension and
4334 <para>This could theoretically have caused loss of a page's worth of
4335 freshly-inserted data, although the scenario seems of very low probability.
4336 There are no known cases of it having caused more than an Assert failure.
4338 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
4340 The comparison code was wrong in the case where the
4341 <literal>--enable-integer-datetimes</> configuration switch had been used.
4342 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
4343 it will need to be <command>REINDEX</>ed after installing this update, because
4344 the fix corrects the sort order of column values.
4346 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
4347 <type>TIME WITH TIME ZONE</> values</para></listitem>
4348 <listitem><para>Fix mis-display of negative fractional seconds in
4349 <type>INTERVAL</> values</para>
4351 This error only occurred when the
4352 <literal>--enable-integer-datetimes</> configuration switch had been used.
4354 <listitem><para>Ensure operations done during backend shutdown are counted by
4355 statistics collector</para>
4357 This is expected to resolve reports of <application>pg_autovacuum</>
4358 not vacuuming the system catalogs often enough — it was not being
4359 told about catalog deletions caused by temporary table removal during
4362 <listitem><para>Additional buffer overrun checks in plpgsql
4363 (Neil)</para></listitem>
4364 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
4365 correctly (Neil)</para></listitem>
4366 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
4367 (Marko Kreen)</para></listitem>
4368 <listitem><para>Still more 64-bit fixes for
4369 <filename>contrib/intagg</></para></listitem>
4370 <listitem><para>Prevent incorrect optimization of functions returning
4371 <type>RECORD</></para></listitem>
4372 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
4373 month-related formats</para></listitem>
4374 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
4375 <listitem><para>Fix <function>array_map</> to call PL functions correctly</para></listitem>
4376 <listitem><para>Fix permission checking in <command>ALTER DATABASE RENAME</></para></listitem>
4377 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</></para></listitem>
4378 <listitem><para>Make <function>RemoveFromWaitQueue</> clean up after itself</para>
4380 This fixes a lock management error that would only be visible if a transaction
4381 was kicked out of a wait for a lock (typically by query cancel) and then the
4382 holder of the lock released it within a very narrow window.
4384 <listitem><para>Fix problem with untyped parameter appearing in
4385 <command>INSERT ... SELECT</></para></listitem>
4386 <listitem><para>Fix <command>CLUSTER</> failure after
4387 <command>ALTER TABLE SET WITHOUT OIDS</></para></listitem>
4393 <sect1 id="release-7-4-7">
4394 <title>Release 7.4.7</title>
4397 <title>Release date</title>
4398 <simpara>2005-01-31</simpara>
4402 This release contains a variety of fixes from 7.4.6, including several
4403 security-related issues.
4407 <title>Migration to version 7.4.7</title>
4410 A dump/restore is not required for those running 7.4.X.
4415 <title>Changes</title>
4418 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
4420 On platforms that will automatically execute initialization functions of a
4421 shared library (this includes at least Windows and ELF-based Unixen),
4422 <command>LOAD</> can be used to make the server execute arbitrary code.
4423 Thanks to NGS Software for reporting this.</para></listitem>
4424 <listitem><para>Check that creator of an aggregate function has the right to
4425 execute the specified transition functions</para>
4427 This oversight made it possible to bypass denial of EXECUTE
4428 permission on a function.</para></listitem>
4429 <listitem><para>Fix security and 64-bit issues in
4430 contrib/intagg</para></listitem>
4431 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
4432 Jurka)</para></listitem>
4433 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
4434 many parameters (Neil)</para></listitem>
4435 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
4437 The result of the join was mistakenly supposed to be sorted the same as the
4438 left input. This could not only deliver mis-sorted output to the user, but
4439 in case of nested merge joins could give outright wrong answers.
4441 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
4442 <listitem><para>Fix display of negative intervals in SQL and GERMAN
4443 datestyles</para></listitem>
4444 <listitem><para>Make age(timestamptz) do calculation in local timezone not
4445 GMT</para></listitem>
4451 <sect1 id="release-7-4-6">
4452 <title>Release 7.4.6</title>
4455 <title>Release date</title>
4456 <simpara>2004-10-22</simpara>
4460 This release contains a variety of fixes from 7.4.5.
4465 <title>Migration to version 7.4.6</title>
4468 A dump/restore is not required for those running 7.4.X.
4473 <title>Changes</title>
4476 <listitem><para>Repair possible failure to update hint bits on disk</para>
4478 Under rare circumstances this oversight could lead to
4479 <quote>could not access transaction status</> failures, which qualifies
4480 it as a potential-data-loss bug.
4482 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
4484 Very large left joins using a hash join plan could fail to output unmatched
4485 left-side rows given just the right data distribution.
4487 <listitem><para>Disallow running <application>pg_ctl</> as root</para>
4489 This is to guard against any possible security issues.
4491 <listitem><para>Avoid using temp files in <filename>/tmp</> in <command>make_oidjoins_check</command></para>
4493 This has been reported as a security issue, though it's hardly worthy of
4494 concern since there is no reason for non-developers to use this script anyway.
4496 <listitem><para>Prevent forced backend shutdown from re-emitting prior command
4499 In rare cases, a client might think that its last command had succeeded when
4500 it really had been aborted by forced database shutdown.
4502 <listitem><para>Repair bug in <function>pg_stat_get_backend_idset</function></para>
4504 This could lead to misbehavior in some of the system-statistics views.
4506 <listitem><para>Fix small memory leak in postmaster</para></listitem>
4507 <listitem><para>Fix <quote>expected both swapped tables to have TOAST
4508 tables</> bug</para>
4510 This could arise in cases such as CLUSTER after ALTER TABLE DROP COLUMN.
4512 <listitem><para>Prevent <literal>pg_ctl restart</> from adding <literal>-D</> multiple times</para></listitem>
4513 <listitem><para>Fix problem with NULL values in GiST indexes</para></listitem>
4514 <listitem><para><literal>::</> is no longer interpreted as a variable in an
4515 ECPG prepare statement</para></listitem>
4521 <sect1 id="release-7-4-5">
4522 <title>Release 7.4.5</title>
4525 <title>Release date</title>
4526 <simpara>2004-08-18</simpara>
4530 This release contains one serious bug fix over 7.4.4.
4535 <title>Migration to version 7.4.5</title>
4538 A dump/restore is not required for those running 7.4.X.
4543 <title>Changes</title>
4546 <listitem><para>Repair possible crash during concurrent B-tree index insertions</para>
4548 This patch fixes a rare case in which concurrent insertions into a B-tree index
4549 could result in a server panic. No permanent damage would result, but it's
4550 still worth a re-release. The bug does not exist in pre-7.4 releases.
4557 <sect1 id="release-7-4-4">
4558 <title>Release 7.4.4</title>
4561 <title>Release date</title>
4562 <simpara>2004-08-16</simpara>
4566 This release contains a variety of fixes from 7.4.3.
4571 <title>Migration to version 7.4.4</title>
4574 A dump/restore is not required for those running 7.4.X.
4579 <title>Changes</title>
4582 <listitem><para>Prevent possible loss of committed transactions during crash</para>
4584 Due to insufficient interlocking between transaction commit and checkpointing,
4585 it was possible for transactions committed just before the most recent
4586 checkpoint to be lost, in whole or in part, following a database crash and
4587 restart. This is a serious bug that has existed
4588 since <productname>PostgreSQL</productname> 7.1.
4590 <listitem><para>Check HAVING restriction before evaluating result list of an
4591 aggregate plan</para></listitem>
4592 <listitem><para>Avoid crash when session's current user ID is deleted</para></listitem>
4593 <listitem><para>Fix hashed crosstab for zero-rows case (Joe)</para></listitem>
4594 <listitem><para>Force cache update after renaming a column in a foreign key</para></listitem>
4595 <listitem><para>Pretty-print UNION queries correctly</para></listitem>
4596 <listitem><para>Make psql handle <literal>\r\n</> newlines properly in COPY IN</para></listitem>
4597 <listitem><para><application>pg_dump</> handled ACLs with grant options incorrectly</para></listitem>
4598 <listitem><para>Fix thread support for OS X and Solaris</para></listitem>
4599 <listitem><para>Updated JDBC driver (build 215) with various fixes</para></listitem>
4600 <listitem><para>ECPG fixes</para></listitem>
4601 <listitem><para>Translation updates (various contributors)</para></listitem>
4607 <sect1 id="release-7-4-3">
4608 <title>Release 7.4.3</title>
4611 <title>Release date</title>
4612 <simpara>2004-06-14</simpara>
4616 This release contains a variety of fixes from 7.4.2.
4621 <title>Migration to version 7.4.3</title>
4624 A dump/restore is not required for those running 7.4.X.
4629 <title>Changes</title>
4632 <listitem><para>Fix temporary memory leak when using non-hashed aggregates (Tom)</para></listitem>
4633 <listitem><para>ECPG fixes, including some for Informix compatibility (Michael)</para></listitem>
4634 <listitem><para>Fixes for compiling with thread-safety, particularly Solaris (Bruce)</para></listitem>
4635 <listitem><para>Fix error in COPY IN termination when using the old network protocol (ljb)</para></listitem>
4636 <listitem><para>Several important fixes in pg_autovacuum, including fixes for
4637 large tables, unsigned oids, stability, temp tables, and debug mode
4638 (Matthew T. O'Connor)</para></listitem>
4639 <listitem><para>Fix problem with reading tar-format dumps on NetBSD and BSD/OS (Bruce)</para></listitem>
4640 <listitem><para>Several JDBC fixes</para></listitem>
4641 <listitem><para>Fix ALTER SEQUENCE RESTART where last_value equals the restart value (Tom)</para></listitem>
4642 <listitem><para>Repair failure to recalculate nested sub-selects (Tom)</para></listitem>
4643 <listitem><para>Fix problems with non-constant expressions in LIMIT/OFFSET</para></listitem>
4644 <listitem><para>Support FULL JOIN with no join clause, such as X FULL JOIN Y ON TRUE (Tom)</para></listitem>
4645 <listitem><para>Fix another zero-column table bug (Tom)</para></listitem>
4646 <listitem><para>Improve handling of non-qualified identifiers in GROUP BY clauses in sub-selects (Tom)</para>
4648 Select-list aliases within the sub-select will now take precedence over
4649 names from outer query levels.
4651 <listitem><para>Do not generate <quote>NATURAL CROSS JOIN</> when decompiling rules (Tom)</para></listitem>
4652 <listitem><para>Add checks for invalid field length in binary COPY (Tom)</para>
4654 This fixes a difficult-to-exploit security hole.
4656 <listitem><para>Avoid locking conflict between <command>ANALYZE</command> and <command>LISTEN</command>/<command>NOTIFY</command></para></listitem>
4657 <listitem><para>Numerous translation updates (various contributors)</para></listitem>
4663 <sect1 id="release-7-4-2">
4664 <title>Release 7.4.2</title>
4667 <title>Release date</title>
4668 <simpara>2004-03-08</simpara>
4672 This release contains a variety of fixes from 7.4.1.
4677 <title>Migration to version 7.4.2</title>
4680 A dump/restore is not required for those running 7.4.X. However,
4681 it may be advisable as the easiest method of incorporating fixes for
4682 two errors that have been found in the initial contents of 7.4.X system
4683 catalogs. A dump/initdb/reload sequence using 7.4.2's initdb will
4684 automatically correct these problems.
4688 The more severe of the two errors is that data type <type>anyarray</>
4689 has the wrong alignment label; this is a problem because the
4690 <structname>pg_statistic</> system catalog uses <type>anyarray</>
4691 columns. The mislabeling can cause planner misestimations and even
4692 crashes when planning queries that involve <literal>WHERE</> clauses on
4693 double-aligned columns (such as <type>float8</> and <type>timestamp</>).
4694 It is strongly recommended that all installations repair this error,
4695 either by initdb or by following the manual repair procedure given
4700 The lesser error is that the system view <structname>pg_settings</>
4701 ought to be marked as having public update access, to allow
4702 <literal>UPDATE pg_settings</> to be used as a substitute for
4703 <command>SET</>. This can also be fixed either by initdb or manually,
4704 but it is not necessary to fix unless you want to use <literal>UPDATE
4709 If you wish not to do an initdb, the following procedure will work
4710 for fixing <structname>pg_statistic</>. As the database superuser,
4714 -- clear out old data in pg_statistic:
4715 DELETE FROM pg_statistic;
4716 VACUUM pg_statistic;
4717 -- this should update 1 row:
4718 UPDATE pg_type SET typalign = 'd' WHERE oid = 2277;
4719 -- this should update 6 rows:
4720 UPDATE pg_attribute SET attalign = 'd' WHERE atttypid = 2277;
4722 -- At this point you MUST start a fresh backend to avoid a crash!
4724 -- repopulate pg_statistic:
4728 This can be done in a live database, but beware that all backends
4729 running in the altered database must be restarted before it is safe to
4730 repopulate <structname>pg_statistic</>.
4734 To repair the <structname>pg_settings</> error, simply do:
4736 GRANT SELECT, UPDATE ON pg_settings TO PUBLIC;
4741 The above procedures must be carried out in <emphasis>each</> database
4742 of an installation, including <literal>template1</>, and ideally
4743 including <literal>template0</> as well. If you do not fix the
4744 template databases then any subsequently created databases will contain
4745 the same errors. <literal>template1</> can be fixed in the same way
4746 as any other database, but fixing <literal>template0</> requires
4747 additional steps. First, from any database issue
4749 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
4751 Next connect to <literal>template0</> and perform the above repair
4752 procedures. Finally, do
4754 -- re-freeze template0:
4756 -- and protect it against future alterations:
4757 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
4763 <title>Changes</title>
4766 Release 7.4.2 incorporates all the fixes included in release 7.3.6,
4767 plus the following fixes:
4771 <listitem><para>Fix <structname>pg_statistics</> alignment bug that could crash optimizer</para>
4772 <para>See above for details about this problem.</para></listitem>
4773 <listitem><para>Allow non-super users to update <structname>pg_settings</></para></listitem>
4774 <listitem><para>Fix several optimizer bugs, most of which led to
4775 <quote>variable not found in subplan target lists</> errors</para></listitem>
4776 <listitem><para>Avoid out-of-memory failure during startup of large multiple
4777 index scan</para></listitem>
4778 <listitem><para>Fix multibyte problem that could lead to <quote>out of
4779 memory</> error during <command>COPY IN</></para></listitem>
4780 <listitem><para>Fix problems with <command>SELECT INTO</> / <command>CREATE
4781 TABLE AS</> from tables without OIDs</para></listitem>
4782 <listitem><para>Fix problems with <filename>alter_table</> regression test
4783 during parallel testing</para></listitem>
4784 <listitem><para>Fix problems with hitting open file limit, especially on OS X (Tom)</para></listitem>
4785 <listitem><para>Partial fix for Turkish-locale issues</para>
4786 <para>initdb will succeed now in Turkish locale, but there are still some
4787 inconveniences associated with the <literal>i/I</> problem.</para></listitem>
4788 <listitem><para>Make pg_dump set client encoding on restore</para></listitem>
4789 <listitem><para>Other minor pg_dump fixes</para></listitem>
4790 <listitem><para>Allow ecpg to again use C keywords as column names (Michael)</para></listitem>
4791 <listitem><para>Added ecpg <literal>WHENEVER NOT_FOUND</> to
4792 <literal>SELECT/INSERT/UPDATE/DELETE</> (Michael)</para></listitem>
4793 <listitem><para>Fix ecpg crash for queries calling set-returning functions (Michael)</para></listitem>
4794 <listitem><para>Various other ecpg fixes (Michael)</para></listitem>
4795 <listitem><para>Fixes for Borland compiler</para></listitem>
4796 <listitem><para>Thread build improvements (Bruce)</para></listitem>
4797 <listitem><para>Various other build fixes</para></listitem>
4798 <listitem><para>Various JDBC fixes</para></listitem>
4804 <sect1 id="release-7-4-1">
4805 <title>Release 7.4.1</title>
4808 <title>Release date</title>
4809 <simpara>2003-12-22</simpara>
4813 This release contains a variety of fixes from 7.4.
4818 <title>Migration to version 7.4.1</title>
4821 A dump/restore is <emphasis>not</emphasis> required for those
4826 If you want to install the fixes in the information schema
4827 you need to reload it into the database.
4828 This is either accomplished by initializing a new cluster
4829 by running <command>initdb</command>, or by running the following
4830 sequence of SQL commands in each database (ideally including
4831 <literal>template1</literal>) as a superuser in
4832 <application>psql</application>, after installing the new release:
4834 DROP SCHEMA information_schema CASCADE;
4835 \i /usr/local/pgsql/share/information_schema.sql
4837 Substitute your installation path in the second command.
4843 <title>Changes</title>
4846 <listitem><para>Fixed bug in <command>CREATE SCHEMA</command> parsing in ECPG (Michael)</para></listitem>
4847 <listitem><para>Fix compile error when <option>--enable-thread-safety</option> and <option>--with-perl</option> are used together (Peter)</para></listitem>
4848 <listitem><para>Fix for subqueries that used hash joins (Tom)</para>
4850 Certain subqueries that used hash joins would crash because of
4851 improperly shared structures.
4853 <listitem><para>Fix free space map compaction bug (Tom)</para>
4855 This fixes a bug where compaction of the free space map could lead
4856 to a database server shutdown.
4859 <listitem><para>Fix for Borland compiler build of libpq (Bruce)</para></listitem>
4860 <listitem><para>Fix <function>netmask()</function> and <function>hostmask()</function> to return the maximum-length masklen (Tom)</para>
4862 Fix these functions to return values consistent with pre-7.4
4866 <listitem><para>Several <filename>contrib/pg_autovacuum</filename> fixes</para>
4868 Fixes include improper variable initialization, missing vacuum after
4869 <command>TRUNCATE</command>, and duration computation overflow for long vacuums.
4872 <listitem><para>Allow compile of <filename>contrib/cube</filename> under Cygwin (Jason Tishler)</para></listitem>
4873 <listitem><para>Fix Solaris use of password file when no passwords are defined (Tom)</para>
4875 Fix crash on Solaris caused by use of any type of password
4876 authentication when no passwords were defined.
4879 <listitem><para>JDBC fix for thread problems, other fixes</para></listitem>
4880 <listitem><para>Fix for <type>bytea</type> index lookups (Joe)</para></listitem>
4881 <listitem><para>Fix information schema for bit data types (Peter)</para></listitem>
4882 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
4883 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
4884 <listitem><para>Make <function>PQescapeBytea</function> and <function>byteaout</function> consistent with each other (Joe)</para></listitem>
4885 <listitem><para>Escape <type>bytea</type> output for bytes > 0x7e(Joe)</para>
4887 If different client encodings are used for <type>bytea</type> output and input, it
4888 is possible for <type>bytea</type> values to be corrupted by the differing
4889 encodings. This fix escapes all bytes that might be affected.
4892 <listitem><para>Added missing <function>SPI_finish()</function> calls to dblink's <function>get_tuple_of_interest()</function> (Joe)</para></listitem>
4893 <listitem><para>New Czech FAQ</para></listitem>
4894 <listitem><para>Fix information schema view <literal>constraint_column_usage</literal> for foreign keys (Peter)</para></listitem>
4895 <listitem><para>ECPG fixes (Michael)</para></listitem>
4896 <listitem><para>Fix bug with multiple <literal>IN</literal> subqueries and joins in the subqueries (Tom)</para></listitem>
4897 <listitem><para>Alllow <literal>COUNT('x')</literal> to work (Tom)</para></listitem>
4898 <listitem><para>Install ECPG include files for Informix compatibility into separate directory (Peter)</para>
4900 Some names of ECPG include files for Informix compatibility conflicted with operating system include files.
4901 By installing them in their own directory, name conflicts have been reduced.
4904 <listitem><para>Fix SSL memory leak (Neil)</para>
4906 This release fixes a bug in 7.4 where SSL didn't free all memory it allocated.
4909 <listitem><para>Prevent <filename>pg_service.conf</filename> from using service name as default dbname (Bruce)</para></listitem>
4910 <listitem><para>Fix local ident authentication on FreeBSD (Tom)</para></listitem>
4916 <sect1 id="release-7-4">
4917 <title>Release 7.4</title>
4920 <title>Release date</title>
4921 <simpara>2003-11-17</simpara>
4925 <title>Overview</title>
4928 Major changes in this release:
4934 <literal>IN</literal> / <literal>NOT IN</literal> subqueries are
4935 now much more efficient
4940 In previous releases, <literal>IN</literal>/<literal>NOT
4941 IN</literal> subqueries were joined to the upper query by
4942 sequentially scanning the subquery looking for a match. The
4943 7.4 code uses the same sophisticated techniques used by
4944 ordinary joins and so is much faster. An
4945 <literal>IN</literal> will now usually be as fast as or faster
4946 than an equivalent <literal>EXISTS</literal> subquery; this
4947 reverses the conventional wisdom that applied to previous
4955 Improved <literal>GROUP BY</literal> processing by using hash buckets
4960 In previous releases, rows to be grouped had to be sorted
4961 first. The 7.4 code can do <literal>GROUP BY</literal>
4962 without sorting, by accumulating results into a hash table
4963 with one entry per group. It will still use the sort
4964 technique, however, if the hash table is estimated to be too
4965 large to fit in <varname>sort_mem</>.
4972 New multikey hash join capability
4977 In previous releases, hash joins could only occur on single
4978 keys. This release allows multicolumn hash joins.
4985 Queries using the explicit <literal>JOIN</literal> syntax are
4986 now better optimized
4991 Prior releases evaluated queries using the explicit
4992 <literal>JOIN</literal> syntax only in the order implied by
4993 the syntax. 7.4 allows full optimization of these queries,
4994 meaning the optimizer considers all possible join orderings
4995 and chooses the most efficient. Outer joins, however, must
4996 still follow the declared ordering.
5003 Faster and more powerful regular expression code
5008 The entire regular expression module has been replaced with a
5009 new version by Henry Spencer, originally written for Tcl. The
5010 code greatly improves performance and supports several flavors
5011 of regular expressions.
5018 Function-inlining for simple SQL functions
5023 Simple SQL functions can now be inlined by including their SQL
5024 in the main query. This improves performance by eliminating
5025 per-call overhead. That means simple SQL functions now
5033 Full support for IPv6 connections and IPv6 address data types
5038 Previous releases allowed only IPv4 connections, and the IP
5039 data types only supported IPv4 addresses. This release adds
5040 full IPv6 support in both of these areas.
5047 Major improvements in SSL performance and reliability
5052 Several people very familiar with the SSL API have overhauled
5053 our SSL code to improve SSL key negotiation and error
5061 Make free space map efficiently reuse empty index pages,
5062 and other free space management improvements
5067 In previous releases, B-tree index pages that were left empty
5068 because of deleted rows could only be reused by rows with
5069 index values similar to the rows originally indexed on that
5070 page. In 7.4, <command>VACUUM</command> records empty index
5071 pages and allows them to be reused for any future index rows.
5078 SQL-standard information schema
5083 The information schema provides a standardized and stable way
5084 to access information about the schema objects defined in a
5092 Cursors conform more closely to the SQL standard
5097 The commands <command>FETCH</command> and
5098 <command>MOVE</command> have been overhauled to conform more
5099 closely to the SQL standard.
5106 Cursors can exist outside transactions
5111 These cursors are also called holdable cursors.
5118 New client-to-server protocol
5123 The new protocol adds error codes, more status information,
5124 faster startup, better support for binary data transmission,
5125 parameter values separated from SQL commands, prepared
5126 statements available at the protocol level, and cleaner
5127 recovery from <command>COPY</command> failures. The older
5128 protocol is still supported by both server and clients.
5135 <application>libpq</application> and
5136 <application>ECPG</application> applications are now fully
5142 While previous <application>libpq</application> releases
5143 already supported threads, this release improves thread safety
5144 by fixing some non-thread-safe code that was used during
5145 database connection startup. The <command>configure</command>
5146 option <option>--enable-thread-safety</option> must be used to
5147 enable this feature.
5154 New version of full-text indexing
5159 A new full-text indexing suite is available in
5160 <filename>contrib/tsearch2</filename>.
5172 The new autovacuum tool in
5173 <filename>contrib/autovacuum</filename> monitors the database
5174 statistics tables for
5175 <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
5176 activity and automatically vacuums tables when needed.
5183 Array handling has been improved and moved into the server core
5188 Many array limitations have been removed, and arrays behave
5189 more like fully-supported data types.
5198 <title>Migration to version 7.4</title>
5201 A dump/restore using <application>pg_dump</application> is
5202 required for those wishing to migrate data from any previous
5207 Observe the following incompatibilities:
5212 The server-side autocommit setting was removed and
5213 reimplemented in client applications and languages.
5214 Server-side autocommit was causing too many problems with
5215 languages and applications that wanted to control their own
5216 autocommit behavior, so autocommit was removed from the server
5217 and added to individual client APIs as appropriate.
5223 Error message wording has changed substantially in this
5224 release. Significant effort was invested to make the messages
5225 more consistent and user-oriented. If your applications try to
5226 detect different error conditions by parsing the error message,
5227 you are strongly encouraged to use the new error code facility instead.
5233 Inner joins using the explicit <literal>JOIN</literal> syntax
5234 may behave differently because they are now better
5241 A number of server configuration parameters have been renamed
5242 for clarity, primarily those related to
5249 <literal>FETCH 0</literal> or <literal>MOVE 0</literal> now
5250 does nothing. In prior releases, <literal>FETCH 0</literal>
5251 would fetch all remaining rows, and <literal>MOVE 0</literal>
5252 would move to the end of the cursor.
5258 <command>FETCH</command> and <command>MOVE</command> now return
5259 the actual number of rows fetched/moved, or zero if at the
5260 beginning/end of the cursor. Prior releases would return the
5261 row count passed to the command, not the number of rows
5262 actually fetched or moved.
5268 <command>COPY</command> now can process files that use
5269 carriage-return or carriage-return/line-feed end-of-line
5270 sequences. Literal carriage-returns and line-feeds are no
5271 longer accepted in data values; use <literal>\r</literal> and
5272 <literal>\n</literal> instead.
5278 Trailing spaces are now trimmed when converting from type
5279 <type>char(<replaceable>n</>)</type> to
5280 <type>varchar(<replaceable>n</>)</type> or <type>text</type>.
5281 This is what most people always expected to happen anyway.
5287 The data type <type>float(<replaceable>p</>)</type> now
5288 measures <replaceable>p</> in binary digits, not decimal
5289 digits. The new behavior follows the SQL standard.
5295 Ambiguous date values now must match the ordering specified by
5296 the <varname>datestyle</varname> setting. In prior releases, a
5297 date specification of <literal>10/20/03</> was interpreted as a
5298 date in October even if <varname>datestyle</> specified that
5299 the day should be first. 7.4 will throw an error if a date
5300 specification is invalid for the current setting of
5301 <varname>datestyle</>.
5307 The functions <function>oidrand</function>,
5308 <function>oidsrand</function>, and
5309 <function>userfntest</function> have been removed. These
5310 functions were determined to be no longer useful.
5316 String literals specifying time-varying date/time values, such
5317 as <literal>'now'</literal> or <literal>'today'</literal> will
5318 no longer work as expected in column default expressions; they
5319 now cause the time of the table creation to be the default, not
5320 the time of the insertion. Functions such as
5321 <function>now()</>, <function>current_timestamp</>, or
5322 <function>current_date</function> should be used instead.
5326 In previous releases, there was special code so that strings
5327 such as <literal>'now'</literal> were interpreted at
5328 <command>INSERT</> time and not at table creation time, but
5329 this work around didn't cover all cases. Release 7.4 now
5330 requires that defaults be defined properly using functions such
5331 as <function>now()</> or <function>current_timestamp</>. These
5332 will work in all situations.
5338 The dollar sign (<literal>$</>) is no longer allowed in
5339 operator names. It can instead be a non-first character in
5340 identifiers. This was done to improve compatibility with other
5341 database systems, and to avoid syntax problems when parameter
5342 placeholders (<literal>$<replaceable>n</></>) are written
5343 adjacent to operators.
5351 <title>Changes</title>
5354 Below you will find a detailed account of the changes between
5355 release 7.4 and the previous major release.
5359 <title>Server Operation Changes</title>
5364 Allow IPv6 server connections (Nigel Kukard, Johan Jordaan,
5365 Bruce, Tom, Kurt Roeckx, Andrew Dunstan)
5371 Fix SSL to handle errors cleanly (Nathan Mueller)
5374 In prior releases, certain SSL API error reports were not
5375 handled correctly. This release fixes those problems.
5381 SSL protocol security and performance improvements (Sean Chittenden)
5384 SSL key renegotiation was happening too frequently, causing poor
5385 SSL performance. Also, initial key handling was improved.
5391 Print lock information when a deadlock is detected (Tom)
5394 This allows easier debugging of deadlock situations.
5400 Update <filename>/tmp</filename> socket modification times
5401 regularly to avoid their removal (Tom)
5404 This should help prevent <filename>/tmp</filename> directory
5405 cleaner administration scripts from removing server socket
5410 <listitem><para>Enable PAM for Mac OS X (Aaron Hillegass)</para></listitem>
5413 <para>Make B-tree indexes fully WAL-safe (Tom)</para>
5415 In prior releases, under certain rare cases, a server crash
5416 could cause B-tree indexes to become corrupt. This release
5417 removes those last few rare cases.
5421 <listitem><para>Allow B-tree index compaction and empty page reuse (Tom)</para></listitem>
5425 Fix inconsistent index lookups during split of first root page (Tom)
5428 In prior releases, when a single-page index split into two
5429 pages, there was a brief period when another database session
5430 could miss seeing an index entry. This release fixes that rare
5435 <listitem><para>Improve free space map allocation logic (Tom)</para></listitem>
5438 <para>Preserve free space information between server restarts (Tom)</para>
5440 In prior releases, the free space map was not saved when the
5441 postmaster was stopped, so newly started servers had no free
5442 space information. This release saves the free space map, and
5443 reloads it when the server is restarted.
5447 <listitem><para>Add start time to <literal>pg_stat_activity</literal> (Neil)</para></listitem>
5448 <listitem><para>New code to detect corrupt disk pages; erase with <varname>zero_damaged_pages</varname> (Tom)</para></listitem>
5449 <listitem><para>New client/server protocol: faster, no username length limit, allow clean exit from <command>COPY</command> (Tom)</para></listitem>
5450 <listitem><para>Add transaction status, table ID, column ID to client/server protocol (Tom)</para></listitem>
5451 <listitem><para>Add binary I/O to client/server protocol (Tom)</para></listitem>
5452 <listitem><para>Remove autocommit server setting; move to client applications (Tom)</para></listitem>
5453 <listitem><para>New error message wording, error codes, and three levels of error detail (Tom, Joe, Peter)</para></listitem>
5458 <title>Performance Improvements</title>
5461 <listitem><para>Add hashing for <literal>GROUP BY</literal> aggregates (Tom)</para></listitem>
5462 <listitem><para>Make nested-loop joins be smarter about multicolumn indexes (Tom)</para></listitem>
5463 <listitem><para>Allow multikey hash joins (Tom)</para></listitem>
5464 <listitem><para>Improve constant folding (Tom)</para></listitem>
5465 <listitem><para>Add ability to inline simple SQL functions (Tom)</para></listitem>
5468 <para>Reduce memory usage for queries using complex functions (Tom)</para>
5470 In prior releases, functions returning allocated memory would
5471 not free it until the query completed. This release allows the
5472 freeing of function-allocated memory when the function call
5473 completes, reducing the total memory used by functions.
5478 <para>Improve GEQO optimizer performance (Tom)</para>
5480 This release fixes several inefficiencies in the way the GEQO optimizer
5481 manages potential query paths.
5487 Allow <literal>IN</>/<literal>NOT IN</> to be handled via hash
5494 Improve <literal>NOT IN (<replaceable>subquery</>)</literal>
5501 Allow most <literal>IN</literal> subqueries to be processed as
5508 Pattern matching operations can use indexes regardless of
5512 There is no way for non-ASCII locales to use the standard
5513 indexes for <literal>LIKE</literal> comparisons. This release
5514 adds a way to create a special index for
5515 <literal>LIKE</literal>.
5520 <para>Allow the postmaster to preload libraries using <varname>preload_libraries</varname> (Joe)</para>
5522 For shared libraries that require a long time to load, this
5523 option is available so the library can be preloaded in the
5524 postmaster and inherited by all database sessions.
5530 Improve optimizer cost computations, particularly for subqueries (Tom)
5536 Avoid sort when subquery <literal>ORDER BY</literal> matches upper query (Tom)
5542 Deduce that <literal>WHERE a.x = b.y AND b.y = 42</literal> also
5543 means <literal>a.x = 42</literal> (Tom)
5549 Allow hash/merge joins on complex joins (Tom)
5555 Allow hash joins for more data types (Tom)
5561 Allow join optimization of explicit inner joins, disable with
5562 <varname>join_collapse_limit</varname> (Tom)
5568 Add parameter <varname>from_collapse_limit</varname> to control
5569 conversion of subqueries to joins (Tom)
5575 Use faster and more powerful regular expression code from Tcl
5576 (Henry Spencer, Tom)
5582 Use bit-mapped relation sets in the optimizer (Tom)
5587 <para>Improve connection startup time (Tom)</para>
5589 The new client/server protocol requires fewer network packets to
5590 start a database session.
5596 Improve trigger/constraint performance (Stephan)
5602 Improve speed of <literal>col IN (const, const, const, ...)</literal> (Tom)
5608 Fix hash indexes which were broken in rare cases (Tom)
5612 <listitem><para>Improve hash index concurrency and speed (Tom)</para>
5614 Prior releases suffered from poor hash index performance,
5615 particularly for high concurrency situations. This release fixes
5616 that, and the development group is interested in reports
5617 comparing B-tree and hash index performance.
5622 <para>Align shared buffers on 32-byte boundary for copy speed improvement (Manfred Spraul)</para>
5624 Certain CPU's perform faster data copies when addresses are
5630 <para>Data type <type>numeric</type> reimplemented for better performance (Tom)</para>
5632 <type>numeric</type> used to be stored in base 100. The new code
5633 uses base 10000, for significantly better performance.
5640 <title>Server Configuration Changes</title>
5644 <para>Rename server parameter <varname>server_min_messages</> to <varname>log_min_messages</> (Bruce)</para>
5646 This was done so most parameters that control the server logs
5647 begin with <literal>log_</>.
5651 <listitem><para>Rename <varname>show_*_stats</> to <varname>log_*_stats</> (Bruce)</para></listitem>
5652 <listitem><para>Rename <varname>show_source_port</> to <varname>log_source_port</> (Bruce)</para></listitem>
5653 <listitem><para>Rename <varname>hostname_lookup</> to <varname>log_hostname</> (Bruce)</para></listitem>
5656 <para>Add <varname>checkpoint_warning</> to warn of excessive checkpointing (Bruce)</para>
5658 In prior releases, it was difficult to determine if checkpoint
5659 was happening too frequently. This feature adds a warning to the
5660 server logs when excessive checkpointing happens.
5664 <listitem><para>New read-only server parameters for localization (Tom)</para></listitem>
5668 Change debug server log messages to output as <literal>DEBUG</>
5669 rather than <literal>LOG</> (Bruce)
5674 <para>Prevent server log variables from being turned off by non-superusers (Bruce)</para>
5676 This is a security feature so non-superusers cannot disable
5677 logging that was enabled by the administrator.
5683 <varname>log_min_messages</>/<varname>client_min_messages</> now
5684 controls <varname>debug_*</> output (Bruce)
5687 This centralizes client debug information so all debug output
5688 can be sent to either the client or server logs.
5693 <para>Add Mac OS X Rendezvous server support (Chris Campbell)</para>
5695 This allows Mac OS X hosts to query the network for available
5696 <productname>PostgreSQL</productname> servers.
5702 Add ability to print only slow statements using
5703 <varname>log_min_duration_statement</varname>
5707 This is an often requested debugging feature that allows
5708 administrators to see only slow queries in their server logs.
5713 <para>Allow <filename>pg_hba.conf</filename> to accept netmasks in CIDR format (Andrew Dunstan)</para>
5715 This allows administrators to merge the host IP address and
5716 netmask fields into a single CIDR field in <filename>pg_hba.conf</filename>.
5720 <listitem><para>New read-only parameter <varname>is_superuser</varname> (Tom)</para></listitem>
5723 <para>New parameter <varname>log_error_verbosity</varname> to control error detail (Tom)</para>
5725 This works with the new error reporting feature to supply
5726 additional error information like hints, file names and line
5732 <para><literal>postgres --describe-config</literal> now dumps server config variables (Aizaz Ahmed, Peter)</para>
5734 This option is useful for administration tools that need to know
5735 the configuration variable names and their minimums, maximums,
5736 defaults, and descriptions.
5742 Add new columns in <literal>pg_settings</literal>:
5743 <literal>context</>, <literal>type</>, <literal>source</>,
5744 <literal>min_val</>, <literal>max_val</> (Joe)
5750 Make default <varname>shared_buffers</> 1000 and
5751 <varname>max_connections</> 100, if possible (Tom)
5754 Prior versions defaulted to 64 shared buffers so <productname>PostgreSQL</productname>
5755 would start on even very old systems. This release tests the
5756 amount of shared memory allowed by the platform and selects more
5757 reasonable default values if possible. Of course, users are
5758 still encouraged to evaluate their resource load and size
5759 <varname>shared_buffers</varname> accordingly.
5765 New <filename>pg_hba.conf</filename> record type
5766 <literal>hostnossl</> to prevent SSL connections (Jon
5770 In prior releases, there was no way to prevent SSL connections
5771 if both the client and server supported SSL. This option allows
5778 Remove parameter <varname>geqo_random_seed</varname>
5785 Add server parameter <varname>regex_flavor</varname> to control regular expression processing (Tom)
5791 Make <command>pg_ctl</command> better handle nonstandard ports (Greg)
5798 <title>Query Changes</title>
5801 <listitem><para>New SQL-standard information schema (Peter)</para></listitem>
5802 <listitem><para>Add read-only transactions (Peter)</para></listitem>
5803 <listitem><para>Print key name and value in foreign-key violation messages (Dmitry Tkach)</para></listitem>
5806 <para>Allow users to see their own queries in <literal>pg_stat_activity</literal> (Kevin Brown)</para>
5808 In prior releases, only the superuser could see query strings
5809 using <literal>pg_stat_activity</literal>. Now ordinary users
5810 can see their own query strings.
5815 <para>Fix aggregates in subqueries to match SQL standard (Tom)</para>
5817 The SQL standard says that an aggregate function appearing
5818 within a nested subquery belongs to the outer query if its
5819 argument contains only outer-query variables. Prior
5820 <productname>PostgreSQL</productname> releases did not handle
5821 this fine point correctly.
5826 <para>Add option to prevent auto-addition of tables referenced in query (Nigel J. Andrews)</para>
5828 By default, tables mentioned in the query are automatically
5829 added to the <literal>FROM</> clause if they are not already
5830 there. This is compatible with historic
5831 <productname>POSTGRES</productname> behavior but is contrary to
5832 the SQL standard. This option allows selecting
5833 standard-compatible behavior.
5838 <para>Allow <literal>UPDATE ... SET col = DEFAULT</literal> (Rod)</para>
5840 This allows <command>UPDATE</command> to set a column to its
5841 declared default value.
5846 <para>Allow expressions to be used in <literal>LIMIT</>/<literal>OFFSET</> (Tom)</para>
5848 In prior releases, <literal>LIMIT</>/<literal>OFFSET</> could
5849 only use constants, not expressions.
5854 <para>Implement <literal>CREATE TABLE AS EXECUTE</literal> (Neil, Peter)</para>
5860 <title>Object Manipulation Changes</title>
5864 <para>Make <command>CREATE SEQUENCE</command> grammar more conforming to SQL:2003 (Neil)</para>
5868 <para>Add statement-level triggers (Neil)</para>
5870 While this allows a trigger to fire at the end of a statement,
5871 it does not allow the trigger to access all rows modified by the
5872 statement. This capability is planned for a future release.
5877 <para>Add check constraints for domains (Rod)</para>
5879 This greatly increases the usefulness of domains by allowing
5880 them to use check constraints.
5885 <para>Add <command>ALTER DOMAIN</command> (Rod)</para>
5887 This allows manipulation of existing domains.
5892 <para>Fix several zero-column table bugs (Tom)</para>
5894 <productname>PostgreSQL</productname> supports zero-column tables. This fixes various bugs
5895 that occur when using such tables.
5900 <para>Have <literal>ALTER TABLE ... ADD PRIMARY KEY</literal> add not-null constraint (Rod)</para>
5902 In prior releases, <literal>ALTER TABLE ... ADD
5903 PRIMARY</literal> would add a unique index, but not a not-null
5904 constraint. That is fixed in this release.
5908 <listitem><para>Add <literal>ALTER TABLE ... WITHOUT OIDS</literal> (Rod)</para>
5910 This allows control over whether new and updated rows will have
5911 an OID column. This is most useful for saving storage space.
5917 Add <literal>ALTER SEQUENCE</literal> to modify minimum, maximum,
5918 increment, cache, cycle values (Rod)
5923 <para>Add <literal>ALTER TABLE ... CLUSTER ON</literal> (Alvaro Herrera)</para>
5925 This command is used by <command>pg_dump</command> to record the
5926 cluster column for each table previously clustered. This
5927 information is used by database-wide cluster to cluster all
5928 previously clustered tables.
5932 <listitem><para>Improve automatic type casting for domains (Rod, Tom)</para></listitem>
5933 <listitem><para>Allow dollar signs in identifiers, except as first character (Tom)</para></listitem>
5934 <listitem><para>Disallow dollar signs in operator names, so <literal>x=$1</> works (Tom)</para></listitem>
5938 Allow copying table schema using <literal>LIKE
5939 <replaceable>subtable</replaceable></literal>, also SQL:2003
5940 feature <literal>INCLUDING DEFAULTS</literal> (Rod)
5946 Add <literal>WITH GRANT OPTION</literal> clause to
5947 <command>GRANT</command> (Peter)
5950 This enabled <command>GRANT</command> to give other users the
5951 ability to grant privileges on a object.
5958 <title>Utility Command Changes</title>
5962 <para>Add <literal>ON COMMIT</literal> clause to <command>CREATE TABLE</command> for temporary tables (Gavin)</para>
5964 This adds the ability for a table to be dropped or all rows
5965 deleted on transaction commit.
5970 <para>Allow cursors outside transactions using <literal>WITH HOLD</literal> (Neil)</para>
5972 In previous releases, cursors were removed at the end of the
5973 transaction that created them. Cursors can now be created with
5974 the <literal>WITH HOLD</literal> option, which allows them to
5975 continue to be accessed after the creating transaction has
5981 <para><literal>FETCH 0</literal> and <literal>MOVE 0 </literal> now do nothing (Bruce)</para>
5983 In previous releases, <literal>FETCH 0</literal> fetched all
5984 remaining rows, and <literal>MOVE 0</literal> moved to the end
5991 Cause <command>FETCH</command> and <command>MOVE</command> to
5992 return the number of rows fetched/moved, or zero if at the
5993 beginning/end of cursor, per SQL standard (Bruce)
5996 In prior releases, the row count returned by
5997 <command>FETCH</command> and <command>MOVE</command> did not
5998 accurately reflect the number of rows processed.
6003 <para>Properly handle <literal>SCROLL</literal> with cursors, or
6004 report an error (Neil)</para>
6006 Allowing random access (both forward and backward scrolling) to
6007 some kinds of queries cannot be done without some additional
6008 work. If <literal>SCROLL</literal> is specified when the cursor
6009 is created, this additional work will be performed. Furthermore,
6010 if the cursor has been created with <literal>NO SCROLL</literal>,
6011 no random access is allowed.
6017 Implement SQL-compatible options <literal>FIRST</>,
6018 <literal>LAST</>, <literal>ABSOLUTE <replaceable>n</></>,
6019 <literal>RELATIVE <replaceable>n</></> for
6020 <command>FETCH</command> and <command>MOVE</command> (Tom)
6025 <para>Allow <command>EXPLAIN</command> on <command>DECLARE CURSOR</command> (Tom)</para>
6029 <para>Allow <command>CLUSTER</command> to use index marked as pre-clustered by default (Alvaro Herrera)</para>
6033 <para>Allow <command>CLUSTER</command> to cluster all tables (Alvaro Herrera)</para>
6035 This allows all previously clustered tables in a database to be
6036 reclustered with a single command.
6040 <listitem><para>Prevent <command>CLUSTER</command> on partial indexes (Tom)</para></listitem>
6042 <listitem><para>Allow DOS and Mac line-endings in <command>COPY</> files (Bruce)</para></listitem>
6046 Disallow literal carriage return as a data value,
6047 backslash-carriage-return and <literal>\r</> are still allowed
6053 <para><command>COPY</> changes (binary, <literal>\.</>) (Tom)</para>
6057 <para>Recover from <command>COPY</command> failure cleanly (Tom)</para>
6061 <para>Prevent possible memory leaks in <command>COPY</command> (Tom)</para>
6065 <para>Make <command>TRUNCATE</command> transaction-safe (Rod)</para>
6067 <command>TRUNCATE</command> can now be used inside a
6068 transaction. If the transaction aborts, the changes made by the
6069 <command>TRUNCATE</command> are automatically rolled back.
6075 Allow prepare/bind of utility commands like
6076 <command>FETCH</command> and <command>EXPLAIN</command> (Tom)
6081 <para>Add <command>EXPLAIN EXECUTE</command> (Neil)</para>
6085 <para>Improve <command>VACUUM</command> performance on indexes by reducing WAL traffic (Tom)</para>
6089 <para>Functional indexes have been generalized into indexes on expressions (Tom)</para>
6091 In prior releases, functional indexes only supported a simple
6092 function applied to one or more column names. This release
6093 allows any type of scalar expression.
6099 Have <command>SHOW TRANSACTION ISOLATION</command> match input
6100 to <command>SET TRANSACTION ISOLATION</command>
6107 Have <command>COMMENT ON DATABASE</command> on nonlocal
6108 database generate a warning, rather than an error (Rod)
6112 Database comments are stored in database-local tables so
6113 comments on a database have to be stored in each database.
6119 Improve reliability of <command>LISTEN</>/<command>NOTIFY</> (Tom)
6124 <para>Allow <command>REINDEX</command> to reliably reindex nonshared system catalog indexes (Tom)</para>
6126 This allows system tables to be reindexed without the
6127 requirement of a standalone session, which was necessary in
6128 previous releases. The only tables that now require a standalone
6129 session for reindexing are the global system tables
6130 <literal>pg_database</>, <literal>pg_shadow</>, and
6131 <literal>pg_group</>.
6138 <title>Data Type and Function Changes</title>
6143 New server parameter <varname>extra_float_digits</varname> to
6144 control precision display of floating-point numbers (Pedro
6148 This controls output precision which was causing regression
6153 <listitem><para>Allow <literal>+1300</literal> as a numeric time-zone specifier, for FJST (Tom)</para></listitem>
6157 Remove rarely used functions <function>oidrand</>,
6158 <function>oidsrand</>, and <function>userfntest</> functions
6164 <para>Add <function>md5()</> function to main server, already in <filename>contrib/pgcrypto</filename> (Joe)</para>
6166 An MD5 function was frequently requested. For more complex
6167 encryption capabilities, use
6168 <filename>contrib/pgcrypto</filename>.
6172 <listitem><para>Increase date range of <type>timestamp</type> (John Cochran)</para></listitem>
6176 Change <literal>EXTRACT(EPOCH FROM timestamp)</literal> so
6177 <type>timestamp without time zone</type> is assumed to be in
6178 local time, not GMT (Tom)
6182 <listitem><para>Trap division by zero in case the operating system doesn't prevent it (Tom)</para></listitem>
6183 <listitem><para>Change the <type>numeric</type> data type internally to base 10000 (Tom)</para></listitem>
6184 <listitem><para>New <function>hostmask()</function> function (Greg Wickham)</para></listitem>
6185 <listitem><para>Fixes for <function>to_char()</function> and <function>to_timestamp()</function> (Karel)</para></listitem>
6189 Allow functions that can take any argument data type and return
6190 any data type, using <type>anyelement</type> and
6191 <type>anyarray</type> (Joe)
6194 This allows the creation of functions that can work with any
6201 Arrays may now be specified as <literal>ARRAY[1,2,3]</literal>,
6202 <literal>ARRAY[['a','b'],['c','d']]</literal>, or
6203 <literal>ARRAY[ARRAY[ARRAY[2]]]</literal> (Joe)
6209 Allow proper comparisons for arrays, including <literal>ORDER
6210 BY</literal> and <literal>DISTINCT</literal> support
6215 <listitem><para>Allow indexes on array columns (Joe)</para></listitem>
6216 <listitem><para>Allow array concatenation with <literal>||</literal> (Joe)</para></listitem>
6220 Allow <literal>WHERE</literal> qualification
6221 <literal><replaceable>expr</> <replaceable>op</> ANY/SOME/ALL
6222 (<replaceable>array_expr</>)</literal> (Joe)
6225 This allows arrays to behave like a list of values, for purposes
6226 like <literal>SELECT * FROM tab WHERE col IN
6227 (array_val)</literal>.
6233 New array functions <function>array_append</>,
6234 <function>array_cat</>, <function>array_lower</>,
6235 <function>array_prepend</>, <function>array_to_string</>,
6236 <function>array_upper</>, <function>string_to_array</> (Joe)
6240 <listitem><para>Allow user defined aggregates to use polymorphic functions (Joe)</para></listitem>
6241 <listitem><para>Allow assignments to empty arrays (Joe)</para></listitem>
6245 Allow 60 in seconds fields of <type>time</type>,
6246 <type>timestamp</type>, and <type>interval</type> input values
6250 Sixty-second values are needed for leap seconds.
6254 <listitem><para>Allow <type>cidr</type> data type to be cast to <type>text</type> (Tom)</para></listitem>
6256 <listitem><para>Disallow invalid time zone names in SET TIMEZONE</para></listitem>
6260 Trim trailing spaces when <type>char</type> is cast to
6261 <type>varchar</> or <type>text</> (Tom)
6267 Make <type>float(<replaceable>p</>)</> measure the precision
6268 <replaceable>p</> in binary digits, not decimal digits
6274 <para>Add IPv6 support to the <type>inet</type> and <type>cidr</type> data types (Michael Graff)</para>
6278 <para>Add <function>family()</function> function to report whether address is IPv4 or IPv6 (Michael Graff)</para>
6283 Have <literal>SHOW datestyle</literal> generate output similar
6284 to that used by <literal>SET datestyle</literal> (Tom)
6290 Make <literal>EXTRACT(TIMEZONE)</literal> and <literal>SET/SHOW
6291 TIME ZONE</literal> follow the SQL convention for the sign of
6292 time zone offsets, i.e., positive is east from UTC (Tom)
6297 <para>Fix <literal>date_trunc('quarter', ...)</literal> (Böjthe Zoltán)</para>
6299 Prior releases returned an incorrect value for this function call.
6304 <para>Make <function>initcap()</function> more compatible with Oracle (Mike Nolan)</para>
6306 <function>initcap()</function> now uppercases a letter appearing
6307 after any non-alphanumeric character, rather than only after
6313 <para>Allow only <varname>datestyle</varname> field order for date values not in ISO-8601 format (Greg)</para>
6318 Add new <varname>datestyle</varname> values <literal>MDY</>,
6319 <literal>DMY</>, and <literal>YMD</> to set input field order;
6320 honor <literal>US</> and <literal>European</> for backward
6327 String literals like <literal>'now'</literal> or
6328 <literal>'today'</literal> will no longer work as a column
6329 default. Use functions such as <function>now()</function>,
6330 <function>current_timestamp</function> instead. (change
6331 required for prepared statements) (Tom)
6336 <para>Treat NaN as larger than any other value in <function>min()</>/<function>max()</> (Tom)</para>
6338 NaN was already sorted after ordinary numeric values for most
6339 purposes, but <function>min()</> and <function>max()</> didn't
6345 <para>Prevent interval from suppressing <literal>:00</literal>
6346 seconds display</para>
6351 New functions <function>pg_get_triggerdef(prettyprint)</function>
6352 and <function>pg_conversion_is_visible()</function> (Christopher)
6357 <para>Allow time to be specified as <literal>040506</> or <literal>0405</> (Tom)</para>
6362 Input date order must now be <literal>YYYY-MM-DD</literal> (with 4-digit year) or
6363 match <varname>datestyle</varname>
6369 Make <function>pg_get_constraintdef</function> support
6370 unique, primary-key, and check constraints (Christopher)
6377 <title>Server-Side Language Changes</title>
6382 Prevent PL/pgSQL crash when <literal>RETURN NEXT</literal> is
6383 used on a zero-row record variable (Tom)
6389 Make PL/Python's <function>spi_execute</function> interface
6390 handle null values properly (Andrew Bosma)
6395 <para>Allow PL/pgSQL to declare variables of composite types without <literal>%ROWTYPE</literal> (Tom)</para>
6399 <para>Fix PL/Python's <function>_quote()</function> function to handle big integers</para>
6403 <para>Make PL/Python an untrusted language, now called <literal>plpythonu</literal> (Kevin Jacobs, Tom)</para>
6405 The Python language no longer supports a restricted execution
6406 environment, so the trusted version of PL/Python was removed. If
6407 this situation changes, a version of PL/Python that can be used
6408 by non-superusers will be readded.
6413 <para>Allow polymorphic PL/pgSQL functions (Joe, Tom)</para>
6417 <para>Allow polymorphic SQL functions (Joe)</para>
6422 Improved compiled function caching mechanism in PL/pgSQL with
6423 full support for polymorphism (Joe)
6429 Add new parameter <literal>$0</> in PL/pgSQL representing the
6430 function's actual return type (Joe)
6436 Allow PL/Tcl and PL/Python to use the same trigger on multiple tables (Tom)
6442 Fixed PL/Tcl's <function>spi_prepare</function> to accept fully
6443 qualified type names in the parameter type list
6451 <title>psql Changes</title>
6455 <para>Add <literal>\pset pager always</literal> to always use pager (Greg)</para>
6457 This forces the pager to be used even if the number of rows is
6458 less than the screen height. This is valuable for rows that
6459 wrap across several screen rows.
6463 <listitem><para>Improve tab completion (Rod, Ross Reedstrom, Ian Barwick)</para></listitem>
6464 <listitem><para>Reorder <literal>\?</> help into groupings (Harald Armin Massa, Bruce)</para></listitem>
6465 <listitem><para>Add backslash commands for listing schemas, casts, and conversions (Christopher)</para></listitem>
6469 <command>\encoding</> now changes based on the server parameter
6470 <varname>client_encoding</varname> (Tom)
6473 In previous versions, <command>\encoding</command> was not aware
6474 of encoding changes made using <literal>SET
6475 client_encoding</literal>.
6480 <para>Save editor buffer into readline history (Ross)</para>
6482 When <command>\e</> is used to edit a query, the result is saved
6483 in the readline history for retrieval using the up arrow.
6487 <listitem><para>Improve <command>\d</command> display (Christopher)</para></listitem>
6488 <listitem><para>Enhance HTML mode to be more standards-conforming (Greg)</para></listitem>
6491 <para>New <command>\set AUTOCOMMIT off</command> capability (Tom)</para>
6493 This takes the place of the removed server parameter <varname>autocommit</varname>.
6498 <para>New <command>\set VERBOSITY</command> to control error detail (Tom)</para>
6500 This controls the new error reporting details.
6504 <listitem><para>New prompt escape sequence <literal>%x</literal> to show transaction status (Tom)</para></listitem>
6505 <listitem><para>Long options for <application>psql</application> are now available on all platforms</para></listitem>
6510 <title>pg_dump Changes</title>
6513 <listitem><para>Multiple pg_dump fixes, including tar format and large objects</para></listitem>
6514 <listitem><para>Allow pg_dump to dump specific schemas (Neil)</para></listitem>
6517 <para>Make pg_dump preserve column storage characteristics (Christopher)</para>
6519 This preserves <literal>ALTER TABLE ... SET STORAGE</literal> information.
6523 <listitem><para>Make pg_dump preserve <command>CLUSTER</command> characteristics (Christopher)</para></listitem>
6527 Have pg_dumpall use <command>GRANT</>/<command>REVOKE</> to dump database-level privleges (Tom)
6533 Allow pg_dumpall to support the options <option>-a</>,
6534 <option>-s</>, <option>-x</> of pg_dump (Tom)
6538 <listitem><para>Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)</para></listitem>
6542 pg_dump options <option>--use-set-session-authorization</option>
6543 and <option>--no-reconnect</option> now do nothing, all dumps
6544 use <command>SET SESSION AUTHORIZATION</command>
6547 pg_dump no longer reconnects to switch users, but instead always
6548 uses <command>SET SESSION AUTHORIZATION</command>. This will
6549 reduce password prompting during restores.
6554 <para>Long options for <application>pg_dump</application> are now available on all platforms</para>
6556 <productname>PostgreSQL</productname> now includes its own
6557 long-option processing routines.
6564 <title>libpq Changes</title>
6569 Add function <function>PQfreemem</function> for freeing memory on
6570 Windows, suggested for <command>NOTIFY</command> (Bruce)
6573 Windows requires that memory allocated in a library be freed by
6574 a function in the same library, hence
6575 <function>free()</function> doesn't work for freeing memory
6576 allocated by libpq. <function>PQfreemem</function> is the proper
6577 way to free libpq memory, especially on Windows, and is
6578 recommended for other platforms as well.
6583 <para>Document service capability, and add sample file (Bruce)</para>
6585 This allows clients to look up connection information in a
6586 central file on the client machine.
6592 Make <function>PQsetdbLogin</function> have the same defaults as
6593 <function>PQconnectdb</function> (Tom)
6597 <listitem><para>Allow libpq to cleanly fail when result sets are too large (Tom)</para></listitem>
6601 Improve performance of function <function>PQunescapeBytea</function> (Ben Lamb)
6607 Allow thread-safe libpq with <filename>configure</filename>
6608 option <option>--enable-thread-safety</option> (Lee Kindness,
6615 Allow function <function>pqInternalNotice</function> to accept a
6616 format string and arguments instead of just a preformatted
6617 message (Tom, Sean Chittenden)
6623 Control SSL negotiation with <literal>sslmode</literal> values
6624 <literal>disable</literal>, <literal>allow</literal>,
6625 <literal>prefer</literal>, and <literal>require</literal> (Jon
6631 <para>Allow new error codes and levels of text (Tom)</para>
6635 <para>Allow access to the underlying table and column of a query result (Tom)</para>
6637 This is helpful for query-builder applications that want to know
6638 the underlying table and column names associated with a specific
6643 <listitem><para>Allow access to the current transaction status (Tom)</para></listitem>
6644 <listitem><para>Add ability to pass binary data directly to the server (Tom)</para></listitem>
6648 Add function <function>PQexecPrepared</function> and
6649 <function>PQsendQueryPrepared</function> functions which perform
6650 bind/execute of previously prepared statements (Tom)
6657 <title>JDBC Changes</title>
6660 <listitem><para>Allow <function>setNull</function> on updateable result sets</para></listitem>
6661 <listitem><para>Allow <function>executeBatch</function> on a prepared statement (Barry)</para></listitem>
6662 <listitem><para>Support SSL connections (Barry)</para></listitem>
6663 <listitem><para>Handle schema names in result sets (Paul Sorenson)</para></listitem>
6664 <listitem><para>Add refcursor support (Nic Ferrier)</para></listitem>
6669 <title>Miscellaneous Interface Changes</title>
6673 <para>Prevent possible memory leak or core dump during libpgtcl shutdown (Tom)</para>
6676 <para>Add Informix compatibility to ECPG (Michael)</para>
6678 This allows ECPG to process embedded C programs that were
6679 written using certain Informix extensions.
6684 <para>Add type <type>decimal</type> to ECPG that is fixed length, for Informix (Michael)</para>
6689 Allow thread-safe embedded SQL programs with
6690 <filename>configure</filename> option
6691 <option>--enable-thread-safety</option> (Lee Kindness, Bruce)
6694 This allows multiple threads to access the database at the same
6700 <para>Moved Python client PyGreSQL to <ulink url="http://www.pygresql.org"></ulink> (Marc)</para>
6706 <title>Source Code Changes</title>
6709 <listitem><para>Prevent need for separate platform geometry regression result files (Tom)</para></listitem>
6710 <listitem><para>Improved PPC locking primitive (Reinhard Max)</para></listitem>
6711 <listitem><para>New function <function>palloc0</function> to allocate and clear memory (Bruce)</para></listitem>
6712 <listitem><para>Fix locking code for s390x CPU (64-bit) (Tom)</para></listitem>
6713 <listitem><para>Allow OpenBSD to use local ident credentials (William Ahern)</para></listitem>
6714 <listitem><para>Make query plan trees read-only to executor (Tom)</para></listitem>
6715 <listitem><para>Add Darwin startup scripts (David Wheeler)</para></listitem>
6716 <listitem><para>Allow libpq to compile with Borland C++ compiler (Lester Godwin, Karl Waclawek)</para></listitem>
6717 <listitem><para>Use our own version of <function>getopt_long()</function> if needed (Peter)</para></listitem>
6718 <listitem><para>Convert administration scripts to C (Peter)</para></listitem>
6719 <listitem><para> Bison >= 1.85 is now required to build the <productname>PostgreSQL</> grammar, if building from CVS</para></listitem>
6720 <listitem><para>Merge documentation into one book (Peter)</para></listitem>
6721 <listitem><para>Add Windows compatibility functions (Bruce)</para></listitem>
6722 <listitem><para>Allow client interfaces to compile under MinGW (Bruce)</para></listitem>
6723 <listitem><para>New <function>ereport()</function> function for error reporting (Tom)</para></listitem>
6724 <listitem><para>Support Intel compiler on Linux (Peter)</para></listitem>
6725 <listitem><para>Improve Linux startup scripts (Slawomir Sudnik, Darko Prenosil)</para></listitem>
6726 <listitem><para>Add support for AMD Opteron and Itanium (Jeffrey W. Baker, Bruce)</para></listitem>
6728 <para>Remove <option>--enable-recode</option> option from <command>configure</command></para>
6730 This was no longer needed now that we have <command>CREATE CONVERSION</command>.
6734 <para>Generate a compile error if spinlock code is not found (Bruce)</para>
6736 Platforms without spinlock code will now fail to compile, rather
6737 than silently using semaphores. This failure can be disabled
6738 with a new <command>configure</command> option.
6745 <title>Contrib Changes</title>
6748 <listitem><para>Change dbmirror license to BSD</para></listitem>
6749 <listitem><para>Improve earthdistance (Bruno Wolff III)</para></listitem>
6750 <listitem><para>Portability improvements to pgcrypto (Marko Kreen)</para></listitem>
6751 <listitem><para>Prevent crash in xml (John Gray, Michael Richards)</para></listitem>
6752 <listitem><para>Update oracle</para></listitem>
6753 <listitem><para>Update mysql</para></listitem>
6754 <listitem><para>Update cube (Bruno Wolff III)</para></listitem>
6755 <listitem><para>Update earthdistance to use cube (Bruno Wolff III)</para></listitem>
6756 <listitem><para>Update btree_gist (Oleg)</para></listitem>
6757 <listitem><para>New tsearch2 full-text search module (Oleg, Teodor)</para></listitem>
6758 <listitem><para>Add hash-based crosstab function to tablefuncs (Joe)</para></listitem>
6759 <listitem><para>Add serial column to order <function>connectby()</> siblings in tablefuncs (Nabil Sayegh,Joe)</para></listitem>
6760 <listitem><para>Add named persistent connections to dblink (Shridhar Daithanka)</para></listitem>
6761 <listitem><para>New pg_autovacuum allows automatic <command>VACUUM</command> (Matthew T. O'Connor)</para></listitem>
6762 <listitem><para>Make pgbench honor environment variables <envar>PGHOST</>, <envar>PGPORT</>, <envar>PGUSER</> (Tatsuo)</para></listitem>
6763 <listitem><para>Improve intarray (Teodor Sigaev)</para></listitem>
6764 <listitem><para>Improve pgstattuple (Rod)</para></listitem>
6765 <listitem><para>Fix bug in <function>metaphone()</function> in fuzzystrmatch</para></listitem>
6766 <listitem><para>Improve adddepend (Rod)</para></listitem>
6767 <listitem><para>Update spi/timetravel (Böjthe Zoltán)</para></listitem>
6768 <listitem><para>Fix dbase <option>-s</> option and improve non-ASCII handling (Thomas Behr, Márcio Smiderle)</para></listitem>
6769 <listitem><para>Remove array module because features now included by default (Joe)</para></listitem>
6775 <sect1 id="release-7-3-10">
6776 <title>Release 7.3.10</title>
6779 <title>Release date</title>
6780 <simpara>2005-05-09</simpara>
6784 This release contains a variety of fixes from 7.3.9, including several
6785 security-related issues.
6789 <title>Migration to version 7.3.10</title>
6792 A dump/restore is not required for those running 7.3.X. However,
6793 it is one possible way of handling a significant security problem
6794 that has been found in the initial contents of 7.3.X system
6795 catalogs. A dump/initdb/reload sequence using 7.3.10's initdb will
6796 automatically correct this problem.
6800 The security problem is that the built-in character set encoding
6801 conversion functions can be invoked from SQL commands by unprivileged
6802 users, but the functions were not designed for such use and are not
6803 secure against malicious choices of arguments. The fix involves changing
6804 the declared parameter list of these functions so that they can no longer
6805 be invoked from SQL commands. (This does not affect their normal use
6806 by the encoding conversion machinery.)
6807 It is strongly recommended that all installations repair this error,
6808 either by initdb or by following the manual repair procedure given
6809 below. The error at least allows unprivileged database users to crash
6810 their server process, and may allow unprivileged users to gain the
6811 privileges of a database superuser.
6815 If you wish not to do an initdb, perform the following procedure instead.
6816 As the database superuser, do:
6820 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
6821 WHERE pronamespace = 11 AND pronargs = 5
6822 AND proargtypes[2] = 'cstring'::regtype;
6823 -- The command should report having updated 90 rows;
6824 -- if not, rollback and investigate instead of committing!
6830 The above procedure must be carried out in <emphasis>each</> database
6831 of an installation, including <literal>template1</>, and ideally
6832 including <literal>template0</> as well. If you do not fix the
6833 template databases then any subsequently created databases will contain
6834 the same error. <literal>template1</> can be fixed in the same way
6835 as any other database, but fixing <literal>template0</> requires
6836 additional steps. First, from any database issue
6838 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
6840 Next connect to <literal>template0</> and perform the above repair
6841 procedure. Finally, do
6843 -- re-freeze template0:
6845 -- and protect it against future alterations:
6846 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
6852 <title>Changes</title>
6855 <listitem><para>Change encoding function signature to prevent
6856 misuse</para></listitem>
6857 <listitem><para>Repair ancient race condition that allowed a transaction to be
6858 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
6859 than for other purposes</para>
6860 <para>This is an extremely serious bug since it could lead to apparent
6861 data inconsistencies being briefly visible to applications.</para></listitem>
6862 <listitem><para>Repair race condition between relation extension and
6864 <para>This could theoretically have caused loss of a page's worth of
6865 freshly-inserted data, although the scenario seems of very low probability.
6866 There are no known cases of it having caused more than an Assert failure.
6868 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
6870 The comparison code was wrong in the case where the
6871 <literal>--enable-integer-datetimes</> configuration switch had been used.
6872 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
6873 it will need to be <command>REINDEX</>ed after installing this update, because
6874 the fix corrects the sort order of column values.
6876 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
6877 <type>TIME WITH TIME ZONE</> values</para></listitem>
6878 <listitem><para>Fix mis-display of negative fractional seconds in
6879 <type>INTERVAL</> values</para>
6881 This error only occurred when the
6882 <literal>--enable-integer-datetimes</> configuration switch had been used.
6884 <listitem><para>Additional buffer overrun checks in plpgsql
6885 (Neil)</para></listitem>
6886 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
6887 correctly (Neil)</para></listitem>
6888 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
6889 month-related formats</para></listitem>
6890 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
6891 (Marko Kreen)</para></listitem>
6892 <listitem><para>Still more 64-bit fixes for
6893 <filename>contrib/intagg</></para></listitem>
6894 <listitem><para>Prevent incorrect optimization of functions returning
6895 <type>RECORD</></para></listitem>
6901 <sect1 id="release-7-3-9">
6902 <title>Release 7.3.9</title>
6905 <title>Release date</title>
6906 <simpara>2005-01-31</simpara>
6910 This release contains a variety of fixes from 7.3.8, including several
6911 security-related issues.
6915 <title>Migration to version 7.3.9</title>
6918 A dump/restore is not required for those running 7.3.X.
6923 <title>Changes</title>
6926 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
6928 On platforms that will automatically execute initialization functions of a
6929 shared library (this includes at least Windows and ELF-based Unixen),
6930 <command>LOAD</> can be used to make the server execute arbitrary code.
6931 Thanks to NGS Software for reporting this.</para></listitem>
6932 <listitem><para>Check that creator of an aggregate function has the right to
6933 execute the specified transition functions</para>
6935 This oversight made it possible to bypass denial of EXECUTE
6936 permission on a function.</para></listitem>
6937 <listitem><para>Fix security and 64-bit issues in
6938 contrib/intagg</para></listitem>
6939 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
6940 Jurka)</para></listitem>
6941 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
6942 many parameters (Neil)</para></listitem>
6943 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
6945 The result of the join was mistakenly supposed to be sorted the same as the
6946 left input. This could not only deliver mis-sorted output to the user, but
6947 in case of nested merge joins could give outright wrong answers.
6949 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
6950 <listitem><para>Fix display of negative intervals in SQL and GERMAN
6951 datestyles</para></listitem>
6957 <sect1 id="release-7-3-8">
6958 <title>Release 7.3.8</title>
6961 <title>Release date</title>
6962 <simpara>2004-10-22</simpara>
6966 This release contains a variety of fixes from 7.3.7.
6971 <title>Migration to version 7.3.8</title>
6974 A dump/restore is not required for those running 7.3.X.
6979 <title>Changes</title>
6982 <listitem><para>Repair possible failure to update hint bits on disk</para>
6984 Under rare circumstances this oversight could lead to
6985 <quote>could not access transaction status</> failures, which qualifies
6986 it as a potential-data-loss bug.
6988 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
6990 Very large left joins using a hash join plan could fail to output unmatched
6991 left-side rows given just the right data distribution.
6993 <listitem><para>Disallow running pg_ctl as root</para>
6995 This is to guard against any possible security issues.
6997 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
6999 This has been reported as a security issue, though it's hardly worthy of
7000 concern since there is no reason for non-developers to use this script anyway.
7007 <sect1 id="release-7-3-7">
7008 <title>Release 7.3.7</title>
7011 <title>Release date</title>
7012 <simpara>2004-08-16</simpara>
7016 This release contains one critical fix over 7.3.6, and some minor items.
7021 <title>Migration to version 7.3.7</title>
7024 A dump/restore is not required for those running 7.3.X.
7029 <title>Changes</title>
7032 <listitem><para>Prevent possible loss of committed transactions during crash</para>
7034 Due to insufficient interlocking between transaction commit and checkpointing,
7035 it was possible for transactions committed just before the most recent
7036 checkpoint to be lost, in whole or in part, following a database crash and
7037 restart. This is a serious bug that has existed
7038 since <productname>PostgreSQL</productname> 7.1.
7040 <listitem><para>Remove asymmetrical word processing in tsearch (Teodor)</para></listitem>
7041 <listitem><para>Properly schema-qualify function names when pg_dump'ing a CAST</para></listitem>
7047 <sect1 id="release-7-3-6">
7048 <title>Release 7.3.6</title>
7051 <title>Release date</title>
7052 <simpara>2004-03-02</simpara>
7056 This release contains a variety of fixes from 7.3.5.
7061 <title>Migration to version 7.3.6</title>
7064 A dump/restore is <emphasis>not</emphasis> required for those
7071 <title>Changes</title>
7074 <listitem><para>Revert erroneous changes in rule permissions checking</para>
7075 <para>A patch applied in 7.3.3 to fix a corner case in rule permissions checks
7076 turns out to have disabled rule-related permissions checks in many
7077 not-so-corner cases. This would for example allow users to insert into views
7078 they weren't supposed to have permission to insert into. We have therefore
7079 reverted the 7.3.3 patch. The original bug will be fixed in 8.0.
7081 <listitem><para>Repair incorrect order of operations in
7082 GetNewTransactionId()</para>
7084 This bug could result in failure under out-of-disk-space conditions, including
7085 inability to restart even after disk space is freed.
7087 <listitem><para>Ensure configure selects -fno-strict-aliasing even when
7088 an external value for CFLAGS is supplied</para>
7090 On some platforms, building with -fstrict-aliasing causes bugs.
7092 <listitem><para>Make pg_restore handle 64-bit off_t correctly</para>
7094 This bug prevented proper restoration from archive files exceeding 4Gb.
7096 <listitem><para>Make contrib/dblink not assume that local and remote type OIDs
7097 match (Joe)</para></listitem>
7098 <listitem><para>Quote connectby()'s start_with argument properly (Joe)</para></listitem>
7099 <listitem><para>Don't crash when a rowtype argument to a plpgsql function is
7100 NULL</para></listitem>
7101 <listitem><para>Avoid generating invalid character encoding sequences in
7102 corner cases when planning LIKE operations</para></listitem>
7103 <listitem><para>Ensure text_position() cannot scan past end of source string
7104 in multibyte cases (Korea PostgreSQL Users' Group)</para></listitem>
7105 <listitem><para>Fix index optimization and selectivity estimates for LIKE
7106 operations on bytea columns (Joe)</para></listitem>
7112 <sect1 id="release-7-3-5">
7113 <title>Release 7.3.5</title>
7116 <title>Release date</title>
7117 <simpara>2003-12-03</simpara>
7121 This has a variety of fixes from 7.3.4.
7126 <title>Migration to version 7.3.5</title>
7129 A dump/restore is <emphasis>not</emphasis> required for those
7135 <title>Changes</title>
7138 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
7139 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
7140 <listitem><para>Force stats processes to detach from shared memory, ensuring cleaner shutdown</para></listitem>
7141 <listitem><para>Make PQescapeBytea and byteaout consistent with each other (Joe)</para></listitem>
7142 <listitem><para>Added missing SPI_finish() calls to dblink's get_tuple_of_interest() (Joe)</para></listitem>
7143 <listitem><para>Fix for possible foreign key violation when rule rewrites INSERT (Jan)</para></listitem>
7144 <listitem><para>Support qualified type names in PL/Tcl's spi_prepare command (Jan)</para></listitem>
7145 <listitem><para>Make pg_dump handle a procedural language handler located in pg_catalog</para></listitem>
7146 <listitem><para>Make pg_dump handle cases where a custom opclass is in another schema</para></listitem>
7147 <listitem><para>Make pg_dump dump binary-compatible casts correctly (Jan)</para></listitem>
7148 <listitem><para>Fix insertion of expressions containing subqueries into rule bodies</para></listitem>
7149 <listitem><para>Fix incorrect argument processing in clusterdb script (Anand Ranganathan)</para></listitem>
7150 <listitem><para>Fix problems with dropped columns in plpython triggers</para></listitem>
7151 <listitem><para>Repair problems with to_char() reading past end of its input string (Karel)</para></listitem>
7152 <listitem><para>Fix GB18030 mapping errors (Tatsuo)</para></listitem>
7153 <listitem><para>Fix several problems with SSL error handling and asynchronous SSL I/O</para></listitem>
7154 <listitem><para>Remove ability to bind a list of values to a single parameter in JDBC
7155 (prevents possible SQL-injection attacks)</para></listitem>
7156 <listitem><para>Fix some errors in HAVE_INT64_TIMESTAMP code paths</para></listitem>
7157 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
7163 <sect1 id="release-7-3-4">
7164 <title>Release 7.3.4</title>
7167 <title>Release date</title>
7168 <simpara>2003-07-24</simpara>
7172 This has a variety of fixes from 7.3.3.
7177 <title>Migration to version 7.3.4</title>
7180 A dump/restore is <emphasis>not</emphasis> required for those
7186 <title>Changes</title>
7189 <listitem><para>Repair breakage in timestamp-to-date conversion for dates before 2000</para></listitem>
7190 <listitem><para>Prevent rare possibility of server startup failure (Tom)</para></listitem>
7191 <listitem><para>Fix bugs in interval-to-time conversion (Tom)</para></listitem>
7192 <listitem><para>Add constraint names in a few places in pg_dump (Rod)</para></listitem>
7193 <listitem><para>Improve performance of functions with many parameters (Tom)</para></listitem>
7194 <listitem><para>Fix to_ascii() buffer overruns (Tom)</para></listitem>
7195 <listitem><para>Prevent restore of database comments from throwing an error (Tom)</para></listitem>
7196 <listitem><para>Work around buggy strxfrm() present in some Solaris releases (Tom)</para></listitem>
7197 <listitem><para>Properly escape jdbc setObject() strings to improve security (Barry)</para></listitem>
7203 <sect1 id="release-7-3-3">
7204 <title>Release 7.3.3</title>
7207 <title>Release date</title>
7208 <simpara>2003-05-22</simpara>
7212 This release contains a variety of fixes for version 7.3.2.
7216 <title>Migration to version 7.3.3</title>
7219 A dump/restore is <emphasis>not</emphasis> required for those
7220 running version 7.3.*.
7225 <title>Changes</title>
7228 <listitem><para>Repair sometimes-incorrect computation of StartUpID after a crash</para></listitem>
7229 <listitem><para>Avoid slowness with lots of deferred triggers in one transaction (Stephan)</para></listitem>
7230 <listitem><para>Don't lock referenced row when <command>UPDATE</command> doesn't change foreign key's value (Jan)</para></listitem>
7231 <listitem><para>Use <command>-fPIC</command> not <command>-fpic</command> on Sparc (Tom Callaway)</para></listitem>
7232 <listitem><para>Repair lack of schema-awareness in contrib/reindexdb</para></listitem>
7233 <listitem><para>Fix contrib/intarray error for zero-element result array (Teodor)</para></listitem>
7234 <listitem><para>Ensure createuser script will exit on control-C (Oliver)</para></listitem>
7235 <listitem><para>Fix errors when the type of a dropped column has itself been dropped</para></listitem>
7236 <listitem><para><command>CHECKPOINT</command> does not cause database panic on failure in noncritical steps</para></listitem>
7237 <listitem><para>Accept 60 in seconds fields of timestamp, time, interval input values</para></listitem>
7238 <listitem><para>Issue notice, not error, if <type>TIMESTAMP</type>,
7239 <type> TIME</type>, or <type>INTERVAL</type> precision too large</para></listitem>
7240 <listitem><para>Fix <function>abstime-to-time</function> cast function (fix is
7241 not applied unless you <application>initdb</application>)</para></listitem>
7242 <listitem><para>Fix <application>pg_proc</application> entry for
7243 <type>timestampt_izone</type> (fix is not applied unless you
7244 <application>initdb</application>)</para></listitem>
7245 <listitem><para>Make <function>EXTRACT(EPOCH FROM timestamp without time zone)</function> treat input as local time</para></listitem>
7246 <listitem><para><command>'now'::timestamptz</command> gave wrong answer if timezone changed earlier in transaction</para></listitem>
7247 <listitem><para><envar>HAVE_INT64_TIMESTAMP</envar> code for time with timezone overwrote its input</para></listitem>
7248 <listitem><para>Accept <command>GLOBAL TEMP/TEMPORARY</command> as a
7249 synonym for <command>TEMPORARY</command></para></listitem>
7250 <listitem><para>Avoid improper schema-privilege-check failure in foreign-key triggers</para></listitem>
7251 <listitem><para>Fix bugs in foreign-key triggers for <command>SET DEFAULT</command> action</para></listitem>
7252 <listitem><para>Fix incorrect time-qual check in row fetch for
7253 <command>UPDATE</command> and <command>DELETE</command> triggers</para></listitem>
7254 <listitem><para>Foreign-key clauses were parsed but ignored in
7255 <command>ALTER TABLE ADD COLUMN</command></para></listitem>
7256 <listitem><para>Fix createlang script breakage for case where handler function already exists</para></listitem>
7257 <listitem><para>Fix misbehavior on zero-column tables in <application>pg_dump</application>, COPY, ANALYZE, other places</para></listitem>
7258 <listitem><para>Fix misbehavior of <function>func_error()</function> on type names containing '%'</para></listitem>
7259 <listitem><para>Fix misbehavior of <function>replace()</function> on strings containing '%'</para></listitem>
7260 <listitem><para>Regular-expression patterns containing certain multibyte characters failed</para></listitem>
7261 <listitem><para>Account correctly for <command>NULL</command>s in more cases in join size estimation</para></listitem>
7262 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
7263 <listitem><para>Fix failure to convert large code point values in EUC_TW conversions (Tatsuo)</para></listitem>
7264 <listitem><para>Fix error recovery for <function>SSL_read</function>/<function>SSL_write</function> calls</para></listitem>
7265 <listitem><para>Don't do early constant-folding of type coercion expressions</para></listitem>
7266 <listitem><para>Validate page header fields immediately after reading in any page</para></listitem>
7267 <listitem><para>Repair incorrect check for ungrouped variables in unnamed joins</para></listitem>
7268 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
7269 <listitem><para>contrib/ltree fixes (Teodor)</para></listitem>
7270 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
7271 <listitem><para>Avoid running out of buffers in many-way indexscan (bug introduced in 7.3)</para></listitem>
7272 <listitem><para>Fix planner's selectivity estimation functions to handle domains properly</para></listitem>
7273 <listitem><para>Fix <application>dbmirror</application> memory-allocation bug (Steven Singer)</para></listitem>
7274 <listitem><para>Prevent infinite loop in <function>ln(numeric)</function> due to roundoff error</para></listitem>
7275 <listitem><para><command>GROUP BY</command> got confused if there were multiple equal GROUP BY items</para></listitem>
7276 <listitem><para>Fix bad plan when inherited <command>UPDATE</command>/<command>DELETE</command> references another inherited table</para></listitem>
7277 <listitem><para>Prevent clustering on incomplete (partial or non-NULL-storing) indexes</para></listitem>
7278 <listitem><para>Service shutdown request at proper time if it arrives while still starting up</para></listitem>
7279 <listitem><para>Fix left-links in temporary indexes (could make backwards scans miss entries)</para></listitem>
7280 <listitem><para>Fix incorrect handling of client_encoding setting in postgresql.conf (Tatsuo)</para></listitem>
7281 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
7282 <listitem><para>Fix SPI for case where rule contains multiple statements of the same type</para></listitem>
7283 <listitem><para>Fix problem with checking for wrong type of access privilege in rule query</para></listitem>
7284 <listitem><para>Fix problem with <command>EXCEPT</command> in <command>CREATE RULE</command></para></listitem>
7285 <listitem><para>Prevent problem with dropping temp tables having serial columns</para></listitem>
7286 <listitem><para>Fix replace_vars_with_subplan_refs failure in complex views</para></listitem>
7287 <listitem><para>Fix regexp slowness in single-byte encodings (Tatsuo)</para></listitem>
7288 <listitem><para>Allow qualified type names in <command>CREATE CAST</command>
7289 and <command> DROP CAST</command></para></listitem>
7290 <listitem><para>Accept <function>SETOF type[]</function>, which formerly had to
7291 be written <function>SETOF _type</function></para></listitem>
7292 <listitem><para>Fix <application>pg_dump</application> core dump in some cases with procedural languages</para></listitem>
7293 <listitem><para>Force ISO datestyle in <application>pg_dump</application> output, for portability (Oliver)</para></listitem>
7294 <listitem><para><application>pg_dump</application> failed to handle error return
7295 from <function>lo_read</function> (Oleg Drokin)</para></listitem>
7296 <listitem><para><application>pg_dumpall</application> failed with groups having no members (Nick Eskelinen)</para></listitem>
7297 <listitem><para><application>pg_dumpall</application> failed to recognize --globals-only switch</para></listitem>
7298 <listitem><para>pg_restore failed to restore blobs if -X disable-triggers is specified</para></listitem>
7299 <listitem><para>Repair intrafunction memory leak in plpgsql</para></listitem>
7300 <listitem><para>pltcl's <command>elog</command> command dumped core if given wrong parameters (Ian Harding)</para></listitem>
7301 <listitem><para>plpython used wrong value of <envar>atttypmod</envar> (Brad McLean)</para></listitem>
7302 <listitem><para>Fix improper quoting of boolean values in Python interface (D'Arcy)</para></listitem>
7303 <listitem><para>Added <function>addDataType()</function> method to PGConnection interface for JDBC</para></listitem>
7304 <listitem><para>Fixed various problems with updateable ResultSets for JDBC (Shawn Green)</para></listitem>
7305 <listitem><para>Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka, Peter Royal)</para></listitem>
7306 <listitem><para>Fixed problem with parsing table ACLs in JDBC</para></listitem>
7307 <listitem><para>Better error message for character set conversion problems in JDBC</para></listitem>
7313 <sect1 id="release-7-3-2">
7314 <title>Release 7.3.2</title>
7317 <title>Release date</title>
7318 <simpara>2003-02-04</simpara>
7322 This release contains a variety of fixes for version 7.3.1.
7327 <title>Migration to version 7.3.2</title>
7330 A dump/restore is <emphasis>not</emphasis> required for those
7331 running version 7.3.*.
7336 <title>Changes</title>
7339 <listitem><para>Restore creation of OID column in CREATE TABLE AS / SELECT INTO</para></listitem>
7340 <listitem><para>Fix <application>pg_dump</> core dump when dumping views having comments</para></listitem>
7341 <listitem><para>Dump DEFERRABLE/INITIALLY DEFERRED constraints properly</para></listitem>
7342 <listitem><para>Fix UPDATE when child table's column numbering differs from parent</para></listitem>
7343 <listitem><para>Increase default value of max_fsm_relations</para></listitem>
7344 <listitem><para>Fix problem when fetching backwards in a cursor for a single-row query</para></listitem>
7345 <listitem><para>Make backward fetch work properly with cursor on SELECT DISTINCT query</para></listitem>
7346 <listitem><para>Fix problems with loading <application>pg_dump</> files containing contrib/lo usage</para></listitem>
7347 <listitem><para>Fix problem with all-numeric user names</para></listitem>
7348 <listitem><para>Fix possible memory leak and core dump during disconnect in libpgtcl</para></listitem>
7349 <listitem><para>Make plpython's spi_execute command handle nulls properly (Andrew Bosma)</para></listitem>
7350 <listitem><para>Adjust plpython error reporting so that its regression test passes again</para></listitem>
7351 <listitem><para>Work with bison 1.875</para></listitem>
7352 <listitem><para>Handle mixed-case names properly in plpgsql's %type (Neil)</para></listitem>
7353 <listitem><para>Fix core dump in pltcl when executing a query rewritten by a rule</para></listitem>
7354 <listitem><para>Repair array subscript overruns (per report from Yichen Xie)</para></listitem>
7355 <listitem><para>Reduce MAX_TIME_PRECISION from 13 to 10 in floating-point case</para></listitem>
7356 <listitem><para>Correctly case-fold variable names in per-database and per-user settings</para></listitem>
7357 <listitem><para>Fix coredump in plpgsql's RETURN NEXT when SELECT into record returns no rows</para></listitem>
7358 <listitem><para>Fix outdated use of pg_type.typprtlen in python client interface</para></listitem>
7359 <listitem><para>Correctly handle fractional seconds in timestamps in JDBC driver</para></listitem>
7360 <listitem><para>Improve performance of getImportedKeys() in JDBC</para></listitem>
7361 <listitem><para>Make shared-library symlinks work standardly on HPUX (Giles)</para></listitem>
7362 <listitem><para>Repair inconsistent rounding behavior for timestamp, time, interval</para></listitem>
7363 <listitem><para>SSL negotiation fixes (Nathan Mueller)</para></listitem>
7364 <listitem><para>Make libpq's ~/.pgpass feature work when connecting with PQconnectDB</para></listitem>
7365 <listitem><para>Update my2pg, ora2pg</para></listitem>
7366 <listitem><para>Translation updates</para></listitem>
7367 <listitem><para>Add casts between types lo and oid in contrib/lo</para></listitem>
7368 <listitem><para>fastpath code now checks for privilege to call function</para></listitem>
7374 <sect1 id="release-7-3-1">
7375 <title>Release 7.3.1</title>
7378 <title>Release date</title>
7379 <simpara>2002-12-18</simpara>
7383 This release contains a variety of fixes for version 7.3.
7388 <title>Migration to version 7.3.1</title>
7391 A dump/restore is <emphasis>not</emphasis> required for those
7392 running version 7.3. However, it should be noted that the main
7393 <productname>PostgreSQL</productname> interface library, libpq,
7394 has a new major version number for this release, which may require
7395 recompilation of client code in certain cases.
7400 <title>Changes</title>
7403 <listitem><para>Fix a core dump of COPY TO when client/server encodings don't match (Tom)</para></listitem>
7404 <listitem><para>Allow <application>pg_dump</> to work with pre-7.2 servers (Philip)</para></listitem>
7405 <listitem><para>contrib/adddepend fixes (Tom)</para></listitem>
7406 <listitem><para>Fix problem with deletion of per-user/per-database config settings (Tom)</para></listitem>
7407 <listitem><para>contrib/vacuumlo fix (Tom)</para></listitem>
7408 <listitem><para>Allow 'password' encryption even when pg_shadow contains MD5 passwords (Bruce)</para></listitem>
7409 <listitem><para>contrib/dbmirror fix (Steven Singer)</para></listitem>
7410 <listitem><para>Optimizer fixes (Tom)</para></listitem>
7411 <listitem><para>contrib/tsearch fixes (Teodor Sigaev, Magnus)</para></listitem>
7412 <listitem><para>Allow locale names to be mixed case (Nicolai Tufar)</para></listitem>
7413 <listitem><para>Increment libpq library's major version number (Bruce)</para></listitem>
7414 <listitem><para>pg_hba.conf error reporting fixes (Bruce, Neil)</para></listitem>
7415 <listitem><para>Add SCO Openserver 5.0.4 as a supported platform (Bruce)</para></listitem>
7416 <listitem><para>Prevent EXPLAIN from crashing server (Tom)</para></listitem>
7417 <listitem><para>SSL fixes (Nathan Mueller)</para></listitem>
7418 <listitem><para>Prevent composite column creation via ALTER TABLE (Tom)</para></listitem>
7424 <sect1 id="release-7-3">
7425 <title>Release 7.3</title>
7428 <title>Release date</title>
7429 <simpara>2002-11-27</simpara>
7433 <title>Overview</title>
7436 Major changes in this release:
7441 <term>Schemas</term>
7444 Schemas allow users to create objects in separate namespaces,
7445 so two people or applications can have tables with the same
7446 name. There is also a public schema for shared tables.
7447 Table/index creation can be restricted by removing privileges
7448 on the public schema.
7454 <term>Drop Column</term>
7457 <productname>PostgreSQL</productname> now supports the
7458 <literal>ALTER TABLE ... DROP COLUMN</literal> functionality.
7464 <term>Table Functions</term>
7467 Functions returning multiple rows and/or multiple columns are
7468 now much easier to use than before. You can call such a
7469 <quote>table function</quote> in the <literal>SELECT</literal>
7470 <literal>FROM</literal> clause, treating its output like a
7471 table. Also, <application>PL/pgSQL</application> functions can
7478 <term>Prepared Queries</term>
7481 <productname>PostgreSQL</productname> now supports prepared
7482 queries, for improved performance.
7488 <term>Dependency Tracking</term>
7491 <productname>PostgreSQL</productname> now records object
7492 dependencies, which allows improvements in many areas.
7493 <command>DROP</command> statements now take either
7494 <literal>CASCADE</> or <literal>RESTRICT</> to control whether
7495 dependent objects are also dropped.
7501 <term>Privileges</term>
7504 Functions and procedural languages now have privileges, and
7505 functions can be defined to run with the privileges of their
7512 <term>Internationalization</term>
7515 Both multibyte and locale support are now always enabled.
7521 <term>Logging</term>
7524 A variety of logging options have been enhanced.
7530 <term>Interfaces</term>
7533 A large number of interfaces have been moved to <ulink
7534 url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
7535 where they can be developed and released independently.
7541 <term>Functions/Identifiers</term>
7544 By default, functions can now take up to 32 parameters, and
7545 identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
7546 is now deprecated: there are specific <quote>pseudo-datatypes</>
7547 to represent each of the former meanings of <literal>OPAQUE</>
7548 in function argument and result types.
7557 <title>Migration to version 7.3</title>
7560 A dump/restore using <application>pg_dump</> is required for those
7561 wishing to migrate data from any previous release. If your
7562 application examines the system catalogs, additional changes will
7563 be required due to the introduction of schemas in 7.3; for more
7564 information, see: <ulink
7565 url="http://developer.postgresql.org/~momjian/upgrade_tips_7.3"></>.
7569 Observe the following incompatibilities:
7574 Pre-6.3 clients are no longer supported.
7580 <filename>pg_hba.conf</filename> now has a column for the user
7581 name and additional features. Existing files need to be
7588 Several <filename>postgresql.conf</filename> logging parameters
7595 <literal>LIMIT #,#</literal> has been disabled; use
7596 <literal>LIMIT # OFFSET #</literal>.
7602 <command>INSERT</command> statements with column lists must
7603 specify a value for each specified column. For example,
7604 <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
7605 is now invalid. It's still allowed to supply fewer columns than
7606 expected if the <command>INSERT</command> does not have a column list.
7612 <type>serial</type> columns are no longer automatically
7613 <literal>UNIQUE</>; thus, an index will not automatically be
7620 A <command>SET</command> command inside an aborted transaction
7627 <command>COPY</command> no longer considers missing trailing
7628 columns to be null. All columns need to be specified.
7629 (However, one may achieve a similar effect by specifying a
7630 column list in the <command>COPY</command> command.)
7636 The data type <type>timestamp</type> is now equivalent to
7637 <type>timestamp without time zone</type>, instead of
7638 <type>timestamp with time zone</type>.
7644 Pre-7.3 databases loaded into 7.3 will not have the new object
7645 dependencies for <type>serial</type> columns, unique
7646 constraints, and foreign keys. See the directory
7647 <filename>contrib/adddepend/</filename> for a detailed
7648 description and a script that will add such dependencies.
7654 An empty string (<literal>''</literal>) is no longer allowed as
7655 the input into an integer field. Formerly, it was silently
7666 <title>Changes</title>
7669 <title>Server Operation</title>
7671 <listitem><para>Add pg_locks view to show locks (Neil)</para></listitem>
7672 <listitem><para>Security fixes for password negotiation memory allocation (Neil)</para></listitem>
7673 <listitem><para>Remove support for version 0 FE/BE protocol (<productname>PostgreSQL</productname> 6.2 and earlier) (Tom)</para></listitem>
7674 <listitem><para>Reserve the last few backend slots for superusers, add parameter superuser_reserved_connections to control this (Nigel J. Andrews)</para></listitem>
7679 <title>Performance</title>
7681 <listitem><para>Improve startup by calling localtime() only once (Tom)</para></listitem>
7682 <listitem><para>Cache system catalog information in flat files for faster startup (Tom)</para></listitem>
7683 <listitem><para>Improve caching of index information (Tom)</para></listitem>
7684 <listitem><para>Optimizer improvements (Tom, Fernando Nasser)</para></listitem>
7685 <listitem><para>Catalog caches now store failed lookups (Tom)</para></listitem>
7686 <listitem><para>Hash function improvements (Neil)</para></listitem>
7687 <listitem><para>Improve performance of query tokenization and network handling (Peter)</para></listitem>
7688 <listitem><para>Speed improvement for large object restore (Mario Weilguni)</para></listitem>
7689 <listitem><para>Mark expired index entries on first lookup, saving later heap fetches (Tom)</para></listitem>
7690 <listitem><para>Avoid excessive NULL bitmap padding (Manfred Koizar)</para></listitem>
7691 <listitem><para>Add BSD-licensed qsort() for Solaris, for performance (Bruce)</para></listitem>
7692 <listitem><para>Reduce per-row overhead by four bytes (Manfred Koizar)</para></listitem>
7693 <listitem><para>Fix GEQO optimizer bug (Neil Conway)</para></listitem>
7694 <listitem><para>Make WITHOUT OID actually save four bytes per row (Manfred Koizar)</para></listitem>
7695 <listitem><para>Add default_statistics_target variable to specify ANALYZE buckets (Neil)</para></listitem>
7696 <listitem><para>Use local buffer cache for temporary tables so no WAL overhead (Tom)</para></listitem>
7697 <listitem><para>Improve free space map performance on large tables (Stephen Marshall, Tom)</para></listitem>
7698 <listitem><para>Improved WAL write concurrency (Tom)</para></listitem>
7703 <title>Privileges</title>
7705 <listitem><para>Add privileges on functions and procedural languages (Peter)</para></listitem>
7706 <listitem><para>Add OWNER to CREATE DATABASE so superusers can create databases on behalf of unprivileged users (Gavin Sherry, Tom)</para></listitem>
7707 <listitem><para>Add new object privilege bits EXECUTE and USAGE (Tom)</para></listitem>
7708 <listitem><para>Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION (Tom)</para></listitem>
7709 <listitem><para>Allow functions to be executed with the privilege of the function owner (Peter)</para></listitem>
7714 <title>Server Configuration</title>
7716 <listitem><para>Server log messages now tagged with LOG, not DEBUG (Bruce)</para></listitem>
7717 <listitem><para>Add user column to pg_hba.conf (Bruce)</para></listitem>
7718 <listitem><para>Have log_connections output two lines in log file (Tom)</para></listitem>
7719 <listitem><para>Remove debug_level from postgresql.conf, now server_min_messages (Bruce)</para></listitem>
7720 <listitem><para>New ALTER DATABASE/USER ... SET command for per-user/database initialization (Peter)</para></listitem>
7721 <listitem><para>New parameters server_min_messages and client_min_messages to control which messages are sent to the server logs or client applications (Bruce)</para></listitem>
7722 <listitem><para>Allow pg_hba.conf to specify lists of users/databases separated by commas, group names prepended with +, and file names prepended with @ (Bruce)</para></listitem>
7723 <listitem><para>Remove secondary password file capability and pg_password utility (Bruce)</para></listitem>
7724 <listitem><para>Add variable db_user_namespace for database-local user names (Bruce)</para></listitem>
7725 <listitem><para>SSL improvements (Bear Giles)</para></listitem>
7726 <listitem><para>Make encryption of stored passwords the default (Bruce)</para></listitem>
7727 <listitem><para>Allow pg_statistics to be reset by calling pg_stat_reset() (Christopher)</para></listitem>
7728 <listitem><para>Add log_duration parameter (Bruce)</para></listitem>
7729 <listitem><para>Rename debug_print_query to log_statement (Bruce)</para></listitem>
7730 <listitem><para>Rename show_query_stats to show_statement_stats (Bruce)</para></listitem>
7731 <listitem><para>Add param log_min_error_statement to print commands to logs on error (Gavin)</para></listitem>
7736 <title>Queries</title>
7738 <listitem><para>Make cursors insensitive, meaning their contents do not change (Tom)</para></listitem>
7739 <listitem><para>Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)</para></listitem>
7740 <listitem><para>Increase identifier length to 63 (Neil, Bruce)</para></listitem>
7741 <listitem><para>UNION fixes for merging >= 3 columns of different lengths (Tom)</para></listitem>
7742 <listitem><para>Add DEFAULT key word to INSERT, e.g., INSERT ... (..., DEFAULT, ...) (Rod)</para></listitem>
7743 <listitem><para>Allow views to have default values using ALTER COLUMN ... SET DEFAULT (Neil)</para></listitem>
7744 <listitem><para>Fail on INSERTs with column lists that don't supply all column values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)</para></listitem>
7745 <listitem><para>Fix for join aliases (Tom)</para></listitem>
7746 <listitem><para>Fix for FULL OUTER JOINs (Tom)</para></listitem>
7747 <listitem><para>Improve reporting of invalid identifier and location (Tom, Gavin)</para></listitem>
7748 <listitem><para>Fix OPEN cursor(args) (Tom)</para></listitem>
7749 <listitem><para>Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)</para></listitem>
7750 <listitem><para>Fix for CREATE TABLE AS with UNION (Tom)</para></listitem>
7751 <listitem><para>SQL99 syntax improvements (Thomas)</para></listitem>
7752 <listitem><para>Add statement_timeout variable to cancel queries (Bruce)</para></listitem>
7753 <listitem><para>Allow prepared queries with PREPARE/EXECUTE (Neil)</para></listitem>
7754 <listitem><para>Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)</para></listitem>
7755 <listitem><para>Add variable autocommit (Tom, David Van Wie)</para></listitem>
7760 <title>Object Manipulation</title>
7762 <listitem><para>Make equals signs optional in CREATE DATABASE (Gavin Sherry)</para></listitem>
7763 <listitem><para>Make ALTER TABLE OWNER change index ownership too (Neil)</para></listitem>
7764 <listitem><para>New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls TOAST storage, compression (John Gray)</para></listitem>
7765 <listitem><para>Add schema support, CREATE/DROP SCHEMA (Tom)</para></listitem>
7766 <listitem><para>Create schema for temporary tables (Tom)</para></listitem>
7767 <listitem><para>Add variable search_path for schema search (Tom)</para></listitem>
7768 <listitem><para>Add ALTER TABLE SET/DROP NOT NULL (Christopher)</para></listitem>
7769 <listitem><para>New CREATE FUNCTION volatility levels (Tom)</para></listitem>
7770 <listitem><para>Make rule names unique only per table (Tom)</para></listitem>
7771 <listitem><para>Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)</para></listitem>
7772 <listitem><para>Add ALTER TRIGGER RENAME (Joe)</para></listitem>
7773 <listitem><para>New current_schema() and current_schemas() inquiry functions (Tom)</para></listitem>
7774 <listitem><para>Allow functions to return multiple rows (table functions) (Joe)</para></listitem>
7775 <listitem><para>Make WITH optional in CREATE DATABASE, for consistency (Bruce)</para></listitem>
7776 <listitem><para>Add object dependency tracking (Rod, Tom)</para></listitem>
7777 <listitem><para>Add RESTRICT/CASCADE to DROP commands (Rod)</para></listitem>
7778 <listitem><para>Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)</para></listitem>
7779 <listitem><para>Autodestroy sequence on DROP of table with SERIAL (Rod)</para></listitem>
7780 <listitem><para>Prevent column dropping if column is used by foreign key (Rod)</para></listitem>
7781 <listitem><para>Automatically drop constraints/functions when object is dropped (Rod)</para></listitem>
7782 <listitem><para>Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)</para></listitem>
7783 <listitem><para>Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)</para></listitem>
7784 <listitem><para>Prevent inherited columns from being removed or renamed (Alvaro Herrera)</para></listitem>
7785 <listitem><para>Fix foreign key constraints to not error on intermediate database states (Stephan)</para></listitem>
7786 <listitem><para>Propagate column or table renaming to foreign key constraints</para></listitem>
7787 <listitem><para>Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)</para></listitem>
7788 <listitem><para>Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)</para></listitem>
7789 <listitem><para>Have rules execute alphabetically, returning more predictable values (Tom)</para></listitem>
7790 <listitem><para>Triggers are now fired in alphabetical order (Tom)</para></listitem>
7791 <listitem><para>Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)</para></listitem>
7792 <listitem><para>Allow better casting when inserting/updating values (Tom)</para></listitem>
7797 <title>Utility Commands</title>
7799 <listitem><para>Have COPY TO output embedded carriage returns and newlines as \r and \n (Tom)</para></listitem>
7800 <listitem><para>Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)</para></listitem>
7801 <listitem><para>Make <application>pg_dump</> use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)</para></listitem>
7802 <listitem><para>Disable brackets in multistatement rules (Bruce)</para></listitem>
7803 <listitem><para>Disable VACUUM from being called inside a function (Bruce)</para></listitem>
7804 <listitem><para>Allow dropdb and other scripts to use identifiers with spaces (Bruce)</para></listitem>
7805 <listitem><para>Restrict database comment changes to the current database</para></listitem>
7806 <listitem><para>Allow comments on operators, independent of the underlying function (Rod)</para></listitem>
7807 <listitem><para>Rollback SET commands in aborted transactions (Tom)</para></listitem>
7808 <listitem><para>EXPLAIN now outputs as a query (Tom)</para></listitem>
7809 <listitem><para>Display condition expressions and sort keys in EXPLAIN (Tom)</para></listitem>
7810 <listitem><para>Add 'SET LOCAL var = value' to set configuration variables for a single transaction (Tom)</para></listitem>
7811 <listitem><para>Allow ANALYZE to run in a transaction (Bruce)</para></listitem>
7812 <listitem><para>Improve COPY syntax using new WITH clauses, keep backward compatibility (Bruce)</para></listitem>
7813 <listitem><para>Fix <application>pg_dump</> to consistently output tags in non-ASCII dumps (Bruce)</para></listitem>
7814 <listitem><para>Make foreign key constraints clearer in dump file (Rod)</para></listitem>
7815 <listitem><para>Add COMMENT ON CONSTRAINT (Rod)</para></listitem>
7816 <listitem><para>Allow COPY TO/FROM to specify column names (Brent Verner)</para></listitem>
7817 <listitem><para>Dump UNIQUE and PRIMARY KEY constraints as ALTER TABLE (Rod)</para></listitem>
7818 <listitem><para>Have SHOW output a query result (Joe)</para></listitem>
7819 <listitem><para>Generate failure on short COPY lines rather than pad NULLs (Neil)</para></listitem>
7820 <listitem><para>Fix CLUSTER to preserve all table attributes (Alvaro Herrera)</para></listitem>
7821 <listitem><para>New pg_settings table to view/modify GUC settings (Joe)</para></listitem>
7822 <listitem><para>Add smart quoting, portability improvements to <application>pg_dump</> output (Peter)</para></listitem>
7823 <listitem><para>Dump serial columns out as SERIAL (Tom)</para></listitem>
7824 <listitem><para>Enable large file support, >2G for <application>pg_dump</> (Peter, Philip Warner, Bruce)</para></listitem>
7825 <listitem><para>Disallow TRUNCATE on tables that are involved in referential constraints (Rod)</para></listitem>
7826 <listitem><para>Have TRUNCATE also auto-truncate the toast table of the relation (Tom)</para></listitem>
7827 <listitem><para>Add clusterdb utility that will auto-cluster an entire database based on previous CLUSTER operations (Alvaro Herrera)</para></listitem>
7828 <listitem><para>Overhaul pg_dumpall (Peter)</para></listitem>
7829 <listitem><para>Allow REINDEX of TOAST tables (Tom)</para></listitem>
7830 <listitem><para>Implemented START TRANSACTION, per SQL99 (Neil)</para></listitem>
7831 <listitem><para>Fix rare index corruption when a page split affects bulk delete (Tom)</para></listitem>
7832 <listitem><para>Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)</para></listitem>
7837 <title>Data Types and Functions</title>
7839 <listitem><para>Fix factorial(0) to return 1 (Bruce)</para></listitem>
7840 <listitem><para>Date/time/timezone improvements (Thomas)</para></listitem>
7841 <listitem><para>Fix for array slice extraction (Tom)</para></listitem>
7842 <listitem><para>Fix extract/date_part to report proper microseconds for timestamp (Tatsuo)</para></listitem>
7843 <listitem><para>Allow text_substr() and bytea_substr() to read TOAST values more efficiently (John Gray)</para></listitem>
7844 <listitem><para>Add domain support (Rod)</para></listitem>
7845 <listitem><para>Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types (Thomas)</para></listitem>
7846 <listitem><para>Allow alternate storage scheme of 64-bit integers for date/time types using --enable-integer-datetimes in configure (Thomas)</para></listitem>
7847 <listitem><para>Make timezone(timestamptz) return timestamp rather than a string (Thomas)</para></listitem>
7848 <listitem><para>Allow fractional seconds in date/time types for dates prior to 1BC (Thomas)</para></listitem>
7849 <listitem><para>Limit timestamp data types to 6 decimal places of precision (Thomas)</para></listitem>
7850 <listitem><para>Change timezone conversion functions from timetz() to timezone() (Thomas)</para></listitem>
7851 <listitem><para>Add configuration variables datestyle and timezone (Tom)</para></listitem>
7852 <listitem><para>Add OVERLAY(), which allows substitution of a substring in a string (Thomas)</para></listitem>
7853 <listitem><para>Add SIMILAR TO (Thomas, Tom)</para></listitem>
7854 <listitem><para>Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)</para></listitem>
7855 <listitem><para>Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)</para></listitem>
7856 <listitem><para>Add named composite types using CREATE TYPE typename AS (column) (Joe)</para></listitem>
7857 <listitem><para>Allow composite type definition in the table alias clause (Joe)</para></listitem>
7858 <listitem><para>Add new API to simplify creation of C language table functions (Joe)</para></listitem>
7859 <listitem><para>Remove ODBC-compatible empty parentheses from calls to SQL99 functions for which these parentheses do not match the standard (Thomas)</para></listitem>
7860 <listitem><para>Allow macaddr data type to accept 12 hex digits with no separators (Mike Wyer)</para></listitem>
7861 <listitem><para>Add CREATE/DROP CAST (Peter)</para></listitem>
7862 <listitem><para>Add IS DISTINCT FROM operator (Thomas)</para></listitem>
7863 <listitem><para>Add SQL99 TREAT() function, synonym for CAST() (Thomas)</para></listitem>
7864 <listitem><para>Add pg_backend_pid() to output backend pid (Bruce)</para></listitem>
7865 <listitem><para>Add IS OF / IS NOT OF type predicate (Thomas)</para></listitem>
7866 <listitem><para>Allow bit string constants without fully-specified length (Thomas)</para></listitem>
7867 <listitem><para>Allow conversion between 8-byte integers and bit strings (Thomas)</para></listitem>
7868 <listitem><para>Implement hex literal conversion to bit string literal (Thomas)</para></listitem>
7869 <listitem><para>Allow table functions to appear in the FROM clause (Joe)</para></listitem>
7870 <listitem><para>Increase maximum number of function parameters to 32 (Bruce)</para></listitem>
7871 <listitem><para>No longer automatically create index for SERIAL column (Tom)</para></listitem>
7872 <listitem><para>Add current_database() (Rod)</para></listitem>
7873 <listitem><para>Fix cash_words() to not overflow buffer (Tom)</para></listitem>
7874 <listitem><para>Add functions replace(), split_part(), to_hex() (Joe)</para></listitem>
7875 <listitem><para>Fix LIKE for bytea as a right-hand argument (Joe)</para></listitem>
7876 <listitem><para>Prevent crashes caused by SELECT cash_out(2) (Tom)</para></listitem>
7877 <listitem><para>Fix to_char(1,'FM999.99') to return a period (Karel)</para></listitem>
7878 <listitem><para>Fix trigger/type/language functions returning OPAQUE to return proper type (Tom)</para></listitem>
7883 <title>Internationalization</title>
7885 <listitem><para>Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC) (Eiji Tokuya)</para></listitem>
7886 <listitem><para>Enable locale support by default (Peter)</para></listitem>
7887 <listitem><para>Add locale variables (Peter)</para></listitem>
7888 <listitem><para>Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea (Tatsuo)</para></listitem>
7889 <listitem><para>Add locale awareness to regular expression character classes</para></listitem>
7890 <listitem><para>Enable multibyte support by default (Tatsuo)</para></listitem>
7891 <listitem><para>Add GB18030 multibyte support (Bill Huang)</para></listitem>
7892 <listitem><para>Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo, Kaori)</para></listitem>
7893 <listitem><para>Add pg_conversion table (Tatsuo)</para></listitem>
7894 <listitem><para>Add SQL99 CONVERT() function (Tatsuo)</para></listitem>
7895 <listitem><para>pg_dumpall, pg_controldata, and pg_resetxlog now national-language aware (Peter)</para></listitem>
7896 <listitem><para>New and updated translations</para></listitem>
7901 <title>Server-side Languages</title>
7903 <listitem><para>Allow recursive SQL function (Peter)</para></listitem>
7904 <listitem><para>Change PL/Tcl build to use configured compiler and Makefile.shlib (Peter)</para></listitem>
7905 <listitem><para>Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible (Neil, Tom)</para></listitem>
7906 <listitem><para>Allow PL/pgSQL to handle quoted identifiers (Tom)</para></listitem>
7907 <listitem><para>Allow set-returning PL/pgSQL functions (Neil)</para></listitem>
7908 <listitem><para>Make PL/pgSQL schema-aware (Joe)</para></listitem>
7909 <listitem><para>Remove some memory leaks (Nigel J. Andrews, Tom)</para></listitem>
7916 <listitem><para>Don't lowercase psql \connect database name for 7.2.0 compatibility (Tom)</para></listitem>
7917 <listitem><para>Add psql \timing to time user queries (Greg Sabino Mullane)</para></listitem>
7918 <listitem><para>Have psql \d show index information (Greg Sabino Mullane)</para></listitem>
7919 <listitem><para>New psql \dD shows domains (Jonathan Eisler)</para></listitem>
7920 <listitem><para>Allow psql to show rules on views (Paul ?)</para></listitem>
7921 <listitem><para>Fix for psql variable substitution (Tom)</para></listitem>
7922 <listitem><para>Allow psql \d to show temporary table structure (Tom)</para></listitem>
7923 <listitem><para>Allow psql \d to show foreign keys (Rod)</para></listitem>
7924 <listitem><para>Fix \? to honor \pset pager (Bruce)</para></listitem>
7925 <listitem><para>Have psql reports its version number on startup (Tom)</para></listitem>
7926 <listitem><para>Allow \copy to specify column names (Tom)</para></listitem>
7931 <title>libpq</title>
7933 <listitem><para>Add ~/.pgpass to store host/user password combinations (Alvaro Herrera)</para></listitem>
7934 <listitem><para>Add PQunescapeBytea() function to libpq (Patrick Welche)</para></listitem>
7935 <listitem><para>Fix for sending large queries over non-blocking connections (Bernhard Herzog)</para></listitem>
7936 <listitem><para>Fix for libpq using timers on Win9X (David Ford)</para></listitem>
7937 <listitem><para>Allow libpq notify to handle servers with different-length identifiers (Tom)</para></listitem>
7938 <listitem><para>Add libpq PQescapeString() and PQescapeBytea() to Windows (Bruce)</para></listitem>
7939 <listitem><para>Fix for SSL with non-blocking connections (Jack Bates)</para></listitem>
7940 <listitem><para>Add libpq connection timeout parameter (Denis A Ustimenko)</para></listitem>
7947 <listitem><para>Allow JDBC to compile with JDK 1.4 (Dave)</para></listitem>
7948 <listitem><para>Add JDBC 3 support (Barry)</para></listitem>
7949 <listitem><para>Allows JDBC to set loglevel by adding ?loglevel=X to the connection URL (Barry)</para></listitem>
7950 <listitem><para>Add Driver.info() message that prints out the version number (Barry)</para></listitem>
7951 <listitem><para>Add updateable result sets (Raghu Nidagal, Dave)</para></listitem>
7952 <listitem><para>Add support for callable statements (Paul Bethe)</para></listitem>
7953 <listitem><para>Add query cancel capability</para></listitem>
7954 <listitem><para>Add refresh row (Dave)</para></listitem>
7955 <listitem><para>Fix MD5 encryption handling for multibyte servers (Jun Kawai)</para></listitem>
7956 <listitem><para>Add support for prepared statements (Barry)</para></listitem>
7961 <title>Miscellaneous Interfaces</title>
7963 <listitem><para>Fixed ECPG bug concerning octal numbers in single quotes (Michael)</para></listitem>
7964 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
7965 <listitem><para>Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)</para></listitem>
7966 <listitem><para>Add libpgtcl connection close event (Gerhard Hintermayer)</para></listitem>
7967 <listitem><para>Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
7968 <listitem><para>Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)</para></listitem>
7969 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
7970 <listitem><para>Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
7971 <listitem><para>Remove src/bin/pgaccess from main tree, now at http://www.pgaccess.org (Bruce)</para></listitem>
7972 <listitem><para>Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer, Tom)</para></listitem>
7977 <title>Source Code</title>
7979 <listitem><para>Fix for parallel make (Peter)</para></listitem>
7980 <listitem><para>AIX fixes for linking Tcl (Andreas Zeugswetter)</para></listitem>
7981 <listitem><para>Allow PL/Perl to build under Cygwin (Jason Tishler)</para></listitem>
7982 <listitem><para>Improve MIPS compiles (Peter, Oliver Elphick)</para></listitem>
7983 <listitem><para>Require Autoconf version 2.53 (Peter)</para></listitem>
7984 <listitem><para>Require readline and zlib by default in configure (Peter)</para></listitem>
7985 <listitem><para>Allow Solaris to use Intimate Shared Memory (ISM), for performance (Scott Brunza, P.J. Josh Rovero)</para></listitem>
7986 <listitem><para>Always enable syslog in compile, remove --enable-syslog option (Tatsuo)</para></listitem>
7987 <listitem><para>Always enable multibyte in compile, remove --enable-multibyte option (Tatsuo)</para></listitem>
7988 <listitem><para>Always enable locale in compile, remove --enable-locale option (Peter)</para></listitem>
7989 <listitem><para>Fix for Win9x DLL creation (Magnus Naeslund)</para></listitem>
7990 <listitem><para>Fix for link() usage by WAL code on Windows, BeOS (Jason Tishler)</para></listitem>
7991 <listitem><para>Add sys/types.h to c.h, remove from main files (Peter, Bruce)</para></listitem>
7992 <listitem><para>Fix AIX hang on SMP machines (Tomoyuki Niijima)</para></listitem>
7993 <listitem><para>AIX SMP hang fix (Tomoyuki Niijima)</para></listitem>
7994 <listitem><para>Fix pre-1970 date handling on newer glibc libraries (Tom)</para></listitem>
7995 <listitem><para>Fix PowerPC SMP locking (Tom)</para></listitem>
7996 <listitem><para>Prevent gcc -ffast-math from being used (Peter, Tom)</para></listitem>
7997 <listitem><para>Bison >= 1.50 now required for developer builds</para></listitem>
7998 <listitem><para>Kerberos 5 support now builds with Heimdal (Peter)</para></listitem>
7999 <listitem><para>Add appendix in the User's Guide which lists SQL features (Thomas)</para></listitem>
8000 <listitem><para>Improve loadable module linking to use RTLD_NOW (Tom)</para></listitem>
8001 <listitem><para>New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)</para></listitem>
8002 <listitem><para>New src/port directory holds replaced libc functions (Peter, Bruce)</para></listitem>
8003 <listitem><para>New pg_namespace system catalog for schemas (Tom)</para></listitem>
8004 <listitem><para>Add pg_class.relnamespace for schemas (Tom)</para></listitem>
8005 <listitem><para>Add pg_type.typnamespace for schemas (Tom)</para></listitem>
8006 <listitem><para>Add pg_proc.pronamespace for schemas (Tom)</para></listitem>
8007 <listitem><para>Restructure aggregates to have pg_proc entries (Tom)</para></listitem>
8008 <listitem><para>System relations now have their own namespace, pg_* test not required (Fernando Nasser)</para></listitem>
8009 <listitem><para>Rename TOAST index names to be *_index rather than *_idx (Neil)</para></listitem>
8010 <listitem><para>Add namespaces for operators, opclasses (Tom)</para></listitem>
8011 <listitem><para>Add additional checks to server control file (Thomas)</para></listitem>
8012 <listitem><para>New Polish FAQ (Marcin Mazurek)</para></listitem>
8013 <listitem><para>Add Posix semaphore support (Tom)</para></listitem>
8014 <listitem><para>Document need for reindex (Bruce)</para></listitem>
8015 <listitem><para>Rename some internal identifiers to simplify Windows compile (Jan, Katherine Ward)</para></listitem>
8016 <listitem><para>Add documentation on computing disk space (Bruce)</para></listitem>
8017 <listitem><para>Remove KSQO from GUC (Bruce)</para></listitem>
8018 <listitem><para>Fix memory leak in rtree (Kenneth Been)</para></listitem>
8019 <listitem><para>Modify a few error messages for consistency (Bruce)</para></listitem>
8020 <listitem><para>Remove unused system table columns (Peter)</para></listitem>
8021 <listitem><para>Make system columns NOT NULL where appropriate (Tom)</para></listitem>
8022 <listitem><para>Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)</para></listitem>
8023 <listitem><para>Remove OPAQUE and create specific subtypes (Tom)</para></listitem>
8024 <listitem><para>Cleanups in array internal handling (Joe, Tom)</para></listitem>
8025 <listitem><para>Disallow pg_atoi('') (Bruce)</para></listitem>
8026 <listitem><para>Remove parameter wal_files because WAL files are now recycled (Bruce)</para></listitem>
8027 <listitem><para>Add version numbers to heap pages (Tom)</para></listitem>
8032 <title>Contrib</title>
8034 <listitem><para>Allow inet arrays in /contrib/array (Neil)</para></listitem>
8035 <listitem><para>GiST fixes (Teodor Sigaev, Neil)</para></listitem>
8036 <listitem><para>Upgrade /contrib/mysql</para></listitem>
8037 <listitem><para>Add /contrib/dbsize which shows table sizes without vacuum (Peter)</para></listitem>
8038 <listitem><para>Add /contrib/intagg, integer aggregator routines (mlw)</para></listitem>
8039 <listitem><para>Improve /contrib/oid2name (Neil, Bruce)</para></listitem>
8040 <listitem><para>Improve /contrib/tsearch (Oleg, Teodor Sigaev)</para></listitem>
8041 <listitem><para>Cleanups of /contrib/rserver (Alexey V. Borzov)</para></listitem>
8042 <listitem><para>Update /contrib/oracle conversion utility (Gilles Darold)</para></listitem>
8043 <listitem><para>Update /contrib/dblink (Joe)</para></listitem>
8044 <listitem><para>Improve options supported by /contrib/vacuumlo (Mario Weilguni)</para></listitem>
8045 <listitem><para>Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey Oktyabrski)</para></listitem>
8046 <listitem><para>Add /contrib/reindexdb utility (Shaun Thomas)</para></listitem>
8047 <listitem><para>Add indexing to /contrib/isbn_issn (Dan Weston)</para></listitem>
8048 <listitem><para>Add /contrib/dbmirror (Steven Singer)</para></listitem>
8049 <listitem><para>Improve /contrib/pgbench (Neil)</para></listitem>
8050 <listitem><para>Add /contrib/tablefunc table function examples (Joe)</para></listitem>
8051 <listitem><para>Add /contrib/ltree data type for tree structures (Teodor Sigaev, Oleg Bartunov)</para></listitem>
8052 <listitem><para>Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)</para></listitem>
8053 <listitem><para>Fixes to /contrib/cube (Bruno Wolff)</para></listitem>
8054 <listitem><para>Improve /contrib/fulltextindex (Christopher)</para></listitem>
8061 <sect1 id="release-7-2-8">
8062 <title>Release 7.2.8</title>
8065 <title>Release date</title>
8066 <simpara>2005-05-09</simpara>
8070 This release contains a variety of fixes from 7.2.7, including one
8071 security-related issue.
8075 <title>Migration to version 7.2.8</title>
8078 A dump/restore is not required for those running 7.2.X.
8083 <title>Changes</title>
8086 <listitem><para>Repair ancient race condition that allowed a transaction to be
8087 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
8088 than for other purposes</para>
8089 <para>This is an extremely serious bug since it could lead to apparent
8090 data inconsistencies being briefly visible to applications.</para></listitem>
8091 <listitem><para>Repair race condition between relation extension and
8093 <para>This could theoretically have caused loss of a page's worth of
8094 freshly-inserted data, although the scenario seems of very low probability.
8095 There are no known cases of it having caused more than an Assert failure.
8097 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
8098 <type>TIME WITH TIME ZONE</> values</para></listitem>
8099 <listitem><para>Additional buffer overrun checks in plpgsql
8100 (Neil)</para></listitem>
8101 <listitem><para>Fix pg_dump to dump index names and trigger names containing
8102 <literal>%</> correctly (Neil)</para></listitem>
8103 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
8104 month-related formats</para></listitem>
8105 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
8106 (Marko Kreen)</para></listitem>
8112 <sect1 id="release-7-2-7">
8113 <title>Release 7.2.7</title>
8116 <title>Release date</title>
8117 <simpara>2005-01-31</simpara>
8121 This release contains a variety of fixes from 7.2.6, including several
8122 security-related issues.
8126 <title>Migration to version 7.2.7</title>
8129 A dump/restore is not required for those running 7.2.X.
8134 <title>Changes</title>
8137 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
8139 On platforms that will automatically execute initialization functions of a
8140 shared library (this includes at least Windows and ELF-based Unixen),
8141 <command>LOAD</> can be used to make the server execute arbitrary code.
8142 Thanks to NGS Software for reporting this.</para></listitem>
8143 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
8144 Jurka)</para></listitem>
8145 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
8146 many parameters (Neil)</para></listitem>
8147 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
8149 The result of the join was mistakenly supposed to be sorted the same as the
8150 left input. This could not only deliver mis-sorted output to the user, but
8151 in case of nested merge joins could give outright wrong answers.
8153 <listitem><para>Fix display of negative intervals in SQL and GERMAN
8154 datestyles</para></listitem>
8160 <sect1 id="release-7-2-6">
8161 <title>Release 7.2.6</title>
8164 <title>Release date</title>
8165 <simpara>2004-10-22</simpara>
8169 This release contains a variety of fixes from 7.2.5.
8174 <title>Migration to version 7.2.6</title>
8177 A dump/restore is not required for those running 7.2.X.
8182 <title>Changes</title>
8185 <listitem><para>Repair possible failure to update hint bits on disk</para>
8187 Under rare circumstances this oversight could lead to
8188 <quote>could not access transaction status</> failures, which qualifies
8189 it as a potential-data-loss bug.
8191 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
8193 Very large left joins using a hash join plan could fail to output unmatched
8194 left-side rows given just the right data distribution.
8196 <listitem><para>Disallow running pg_ctl as root</para>
8198 This is to guard against any possible security issues.
8200 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
8202 This has been reported as a security issue, though it's hardly worthy of
8203 concern since there is no reason for non-developers to use this script anyway.
8205 <listitem><para>Update to newer versions of Bison</para></listitem>
8211 <sect1 id="release-7-2-5">
8212 <title>Release 7.2.5</title>
8215 <title>Release date</title>
8216 <simpara>2004-08-16</simpara>
8220 This release contains a variety of fixes from 7.2.4.
8225 <title>Migration to version 7.2.5</title>
8228 A dump/restore is not required for those running 7.2.X.
8233 <title>Changes</title>
8236 <listitem><para>Prevent possible loss of committed transactions during crash</para>
8238 Due to insufficient interlocking between transaction commit and checkpointing,
8239 it was possible for transactions committed just before the most recent
8240 checkpoint to be lost, in whole or in part, following a database crash and
8241 restart. This is a serious bug that has existed
8242 since <productname>PostgreSQL</productname> 7.1.
8244 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
8245 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
8246 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
8247 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
8248 <listitem><para>Repair memory leaks in pg_dump</para></listitem>
8249 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
8254 <sect1 id="release-7-2-4">
8255 <title>Release 7.2.4</title>
8258 <title>Release date</title>
8259 <simpara>2003-01-30</simpara>
8263 This release contains a variety of fixes for version 7.2.3,
8264 including fixes to prevent possible data loss.
8268 <title>Migration to version 7.2.4</title>
8271 A dump/restore is <emphasis>not</emphasis> required for those
8272 running version 7.2.*.
8277 <title>Changes</title>
8280 <listitem><para>Fix some additional cases of VACUUM "No one parent tuple was found" error</para></listitem>
8281 <listitem><para>Prevent VACUUM from being called inside a function (Bruce)</para></listitem>
8282 <listitem><para>Ensure pg_clog updates are sync'd to disk before marking checkpoint complete</para></listitem>
8283 <listitem><para>Avoid integer overflow during large hash joins</para></listitem>
8284 <listitem><para>Make GROUP commands work when pg_group.grolist is large enough to be toasted</para></listitem>
8285 <listitem><para>Fix errors in datetime tables; some timezone names weren't being recognized</para></listitem>
8286 <listitem><para>Fix integer overflows in circle_poly(), path_encode(), path_add() (Neil)</para></listitem>
8287 <listitem><para>Repair long-standing logic errors in lseg_eq(), lseg_ne(), lseg_center()</para></listitem>
8293 <sect1 id="release-7-2-3">
8294 <title>Release 7.2.3</title>
8297 <title>Release date</title>
8298 <simpara>2002-10-01</simpara>
8302 This release contains a variety of fixes for version 7.2.2,
8303 including fixes to prevent possible data loss.
8307 <title>Migration to version 7.2.3</title>
8310 A dump/restore is <emphasis>not</emphasis> required for those
8311 running version 7.2.*.
8316 <title>Changes</title>
8319 <listitem><para>Prevent possible compressed transaction log loss (Tom)</para></listitem>
8320 <listitem><para>Prevent non-superuser from increasing most recent vacuum info (Tom)</para></listitem>
8321 <listitem><para>Handle pre-1970 date values in newer versions of glibc (Tom)</para></listitem>
8322 <listitem><para>Fix possible hang during server shutdown</para></listitem>
8323 <listitem><para>Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)</para></listitem>
8324 <listitem><para>Fix <application>pg_dump</> to properly dump FULL JOIN USING (Tom)</para></listitem>
8330 <sect1 id="release-7-2-2">
8331 <title>Release 7.2.2</title>
8334 <title>Release date</title>
8335 <simpara>2002-08-23</simpara>
8339 This release contains a variety of fixes for version 7.2.1.
8343 <title>Migration to version 7.2.2</title>
8346 A dump/restore is <emphasis>not</emphasis> required for those
8347 running version 7.2.*.
8352 <title>Changes</title>
8355 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8356 <listitem><para>Fix for compressed transaction log id wraparound (Tom)</para></listitem>
8357 <listitem><para>Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Tatsuo)</para></listitem>
8358 <listitem><para>Fix for psql and <application>pg_dump</> crashing when invoked with non-existent long options (Tatsuo)</para></listitem>
8359 <listitem><para>Fix crash when invoking geometric operators (Tom)</para></listitem>
8360 <listitem><para>Allow OPEN cursor(args) (Tom)</para></listitem>
8361 <listitem><para>Fix for rtree_gist index build (Teodor)</para></listitem>
8362 <listitem><para>Fix for dumping user-defined aggregates (Tom)</para></listitem>
8363 <listitem><para>contrib/intarray fixes (Oleg)</para></listitem>
8364 <listitem><para>Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)</para></listitem>
8365 <listitem><para>Fix to pg_convert (Tatsuo)</para></listitem>
8366 <listitem><para>Fix for crash with long DATA strings (Thomas, Neil)</para></listitem>
8367 <listitem><para>Fix for repeat(), lpad(), rpad() and long strings (Neil)</para></listitem>
8373 <sect1 id="release-7-2-1">
8374 <title>Release 7.2.1</title>
8377 <title>Release date</title>
8378 <simpara>2002-03-21</simpara>
8382 This release contains a variety of fixes for version 7.2.
8386 <title>Migration to version 7.2.1</title>
8389 A dump/restore is <emphasis>not</emphasis> required for those
8390 running version 7.2.
8395 <title>Changes</title>
8398 <listitem><para>Ensure that sequence counters do not go backwards after a crash (Tom)</para></listitem>
8399 <listitem><para>Fix pgaccess kanji-conversion key binding (Tatsuo)</para></listitem>
8400 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8401 <listitem><para>Cash I/O improvements (Tom)</para></listitem>
8402 <listitem><para>New Russian FAQ</para></listitem>
8403 <listitem><para>Compile fix for missing AuthBlockSig (Heiko)</para></listitem>
8404 <listitem><para>Additional time zones and time zone fixes (Thomas)</para></listitem>
8405 <listitem><para>Allow psql \connect to handle mixed case database and user names (Tom)</para></listitem>
8406 <listitem><para>Return proper OID on command completion even with ON INSERT rules (Tom)</para></listitem>
8407 <listitem><para>Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)</para></listitem>
8408 <listitem><para>Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)</para></listitem>
8409 <listitem><para>Improve handling of multiple UNIONs with different lengths (Tom)</para></listitem>
8410 <listitem><para>contrib/btree_gist improvements (Teodor Sigaev)</para></listitem>
8411 <listitem><para>contrib/tsearch dictionary improvements, see README.tsearch for an additional installation step (Thomas T. Thai, Teodor Sigaev)</para></listitem>
8412 <listitem><para>Fix for array subscripts handling (Tom)</para></listitem>
8413 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8419 <sect1 id="release-7-2">
8420 <title>Release 7.2</title>
8423 <title>Release date</title>
8424 <simpara>2002-02-04</simpara>
8428 <title>Overview</title>
8431 This release improves <productname>PostgreSQL</> for use in
8432 high-volume applications.
8436 Major changes in this release:
8444 Vacuuming no longer locks tables, thus allowing normal user
8445 access during the vacuum. A new <command>VACUUM FULL</>
8446 command does old-style vacuum by locking the table and
8447 shrinking the on-disk copy of the table.
8453 <term>Transactions</term>
8456 There is no longer a problem with installations that exceed
8457 four billion transactions.
8466 OIDs are now optional. Users can now create tables without
8467 OIDs for cases where OID usage is excessive.
8473 <term>Optimizer</term>
8476 The system now computes histogram column statistics during
8477 <command>ANALYZE</>, allowing much better optimizer choices.
8483 <term>Security</term>
8486 A new MD5 encryption option allows more secure storage and
8487 transfer of passwords. A new Unix-domain socket
8488 authentication option is available on Linux and BSD systems.
8494 <term>Statistics</term>
8497 Administrators can use the new table access statistics module
8498 to get fine-grained information about table and index usage.
8504 <term>Internationalization</term>
8507 Program and library messages can now be displayed in several
8517 <title>Migration to version 7.2</title>
8520 A dump/restore using <command>pg_dump</command> is required for
8521 those wishing to migrate data from any previous release.
8525 Observe the following incompatibilities:
8530 The semantics of the <command>VACUUM</command> command have
8531 changed in this release. You may wish to update your
8532 maintenance procedures accordingly.
8538 In this release, comparisons using <literal>= NULL</literal>
8539 will always return false (or NULL, more precisely). Previous
8540 releases automatically transformed this syntax to <literal>IS
8541 NULL</literal>. The old behavior can be re-enabled using a
8542 <filename>postgresql.conf</filename> parameter.
8548 The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
8549 configuration is now only reloaded after receiving a
8550 <systemitem>SIGHUP</> signal, not with each connection.
8556 The function <filename>octet_length()</> now returns the uncompressed data length.
8562 The date/time value <literal>'current'</literal> is no longer
8563 available. You will need to rewrite your applications.
8569 The <literal>timestamp()</literal>, <literal>time()</literal>,
8570 and <literal>interval()</literal> functions are no longer
8571 available. Instead of <literal>timestamp()</literal>, use
8572 <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
8579 The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
8580 in the next release. You should change your queries to use
8581 separate LIMIT and OFFSET clauses, e.g. <literal>LIMIT 10 OFFSET
8587 <title>Changes</title>
8590 <title>Server Operation</title>
8592 <listitem><para>Create temporary files in a separate directory (Bruce)</para></listitem>
8593 <listitem><para>Delete orphaned temporary files on postmaster startup (Bruce)</para></listitem>
8594 <listitem><para>Added unique indexes to some system tables (Tom)</para></listitem>
8595 <listitem><para>System table operator reorganization (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
8596 <listitem><para>Renamed pg_log to pg_clog (Tom)</para></listitem>
8597 <listitem><para>Enable SIGTERM, SIGQUIT to kill backends (Jan)</para></listitem>
8598 <listitem><para>Removed compile-time limit on number of backends (Tom)</para></listitem>
8599 <listitem><para>Better cleanup for semaphore resource failure (Tatsuo, Tom)</para></listitem>
8600 <listitem><para>Allow safe transaction ID wraparound (Tom)</para></listitem>
8601 <listitem><para>Removed OIDs from some system tables (Tom)</para></listitem>
8602 <listitem><para>Removed "triggered data change violation" error check (Tom)</para></listitem>
8603 <listitem><para>SPI portal creation of prepared/saved plans (Jan)</para></listitem>
8604 <listitem><para>Allow SPI column functions to work for system columns (Tom)</para></listitem>
8605 <listitem><para>Long value compression improvement (Tom)</para></listitem>
8606 <listitem><para>Statistics collector for table, index access (Jan)</para></listitem>
8607 <listitem><para>Truncate extra-long sequence names to a reasonable value (Tom)</para></listitem>
8608 <listitem><para>Measure transaction times in milliseconds (Thomas)</para></listitem>
8609 <listitem><para>Fix TID sequential scans (Hiroshi)</para></listitem>
8610 <listitem><para>Superuser ID now fixed at 1 (Peter E)</para></listitem>
8611 <listitem><para>New pg_ctl "reload" option (Tom)</para></listitem>
8616 <title>Performance</title>
8618 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8619 <listitem><para>New histogram column statistics for optimizer (Tom)</para></listitem>
8620 <listitem><para>Reuse write-ahead log files rather than discarding them (Tom)</para></listitem>
8621 <listitem><para>Cache improvements (Tom)</para></listitem>
8622 <listitem><para>IS NULL, IS NOT NULL optimizer improvement (Tom)</para></listitem>
8623 <listitem><para>Improve lock manager to reduce lock contention (Tom)</para></listitem>
8624 <listitem><para>Keep relcache entries for index access support functions (Tom)</para></listitem>
8625 <listitem><para>Allow better selectivity with NaN and infinities in NUMERIC (Tom)</para></listitem>
8626 <listitem><para>R-tree performance improvements (Kenneth Been)</para></listitem>
8627 <listitem><para>B-tree splits more efficient (Tom)</para></listitem>
8632 <title>Privileges</title>
8634 <listitem><para>Change UPDATE, DELETE privileges to be distinct (Peter E)</para></listitem>
8635 <listitem><para>New REFERENCES, TRIGGER privileges (Peter E)</para></listitem>
8636 <listitem><para>Allow GRANT/REVOKE to/from more than one user at a time (Peter E)</para></listitem>
8637 <listitem><para>New has_table_privilege() function (Joe Conway)</para></listitem>
8638 <listitem><para>Allow non-superuser to vacuum database (Tom)</para></listitem>
8639 <listitem><para>New SET SESSION AUTHORIZATION command (Peter E)</para></listitem>
8640 <listitem><para>Fix bug in privilege modifications on newly created tables (Tom)</para></listitem>
8641 <listitem><para>Disallow access to pg_statistic for non-superuser, add user-accessible views (Tom)</para></listitem>
8646 <title>Client Authentication</title>
8648 <listitem><para>Fork postmaster before doing authentication to prevent hangs (Peter E)</para></listitem>
8649 <listitem><para>Add ident authentication over Unix domain sockets on Linux, *BSD (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)</para></listitem>
8650 <listitem><para>Add a password authentication method that uses MD5 encryption (Bruce)</para></listitem>
8651 <listitem><para>Allow encryption of stored passwords using MD5 (Bruce)</para></listitem>
8652 <listitem><para>PAM authentication (Dominic J. Eidson)</para></listitem>
8653 <listitem><para>Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)</para></listitem>
8658 <title>Server Configuration</title>
8660 <listitem><para>Interpretation of some time zone abbreviations as Australian rather than North American now settable at run time (Bruce)</para></listitem>
8661 <listitem><para>New parameter to set default transaction isolation level (Peter E)</para></listitem>
8662 <listitem><para>New parameter to enable conversion of "expr = NULL" into "expr IS NULL", off by default (Peter E)</para></listitem>
8663 <listitem><para>New parameter to control memory usage by VACUUM (Tom)</para></listitem>
8664 <listitem><para>New parameter to set client authentication timeout (Tom)</para></listitem>
8665 <listitem><para>New parameter to set maximum number of open files (Tom)</para></listitem>
8670 <title>Queries</title>
8672 <listitem><para>Statements added by INSERT rules now execute after the INSERT (Jan)</para></listitem>
8673 <listitem><para>Prevent unadorned relation names in target list (Bruce)</para></listitem>
8674 <listitem><para>NULLs now sort after all normal values in ORDER BY (Tom)</para></listitem>
8675 <listitem><para>New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)</para></listitem>
8676 <listitem><para>New SHARE UPDATE EXCLUSIVE lock mode (Tom)</para></listitem>
8677 <listitem><para>New EXPLAIN ANALYZE command that shows run times and row counts (Martijn van Oosterhout)</para></listitem>
8678 <listitem><para>Fix problem with LIMIT and subqueries (Tom)</para></listitem>
8679 <listitem><para>Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)</para></listitem>
8680 <listitem><para>Fix nested EXCEPT/INTERSECT (Tom)</para></listitem>
8685 <title>Schema Manipulation</title>
8687 <listitem><para>Fix SERIAL in temporary tables (Bruce)</para></listitem>
8688 <listitem><para>Allow temporary sequences (Bruce)</para></listitem>
8689 <listitem><para>Sequences now use int8 internally (Tom)</para></listitem>
8690 <listitem><para>New SERIAL8 creates int8 columns with sequences, default still SERIAL4 (Tom)</para></listitem>
8691 <listitem><para>Make OIDs optional using WITHOUT OIDS (Tom)</para></listitem>
8692 <listitem><para>Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)</para></listitem>
8693 <listitem><para>Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher Kings-Lynne)</para></listitem>
8694 <listitem><para>New CREATE OR REPLACE FUNCTION to alter existing function (preserving the function OID) (Gavin Sherry)</para></listitem>
8695 <listitem><para>Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)</para></listitem>
8696 <listitem><para>Allow column renaming in views</para></listitem>
8697 <listitem><para>Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent Verner)</para></listitem>
8698 <listitem><para>Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (Stephan Szabo)</para></listitem>
8699 <listitem><para>ALTER TABLE RENAME update foreign-key trigger arguments correctly (Brent Verner)</para></listitem>
8700 <listitem><para>DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)</para></listitem>
8701 <listitem><para>Add automatic return type data casting for SQL functions (Tom)</para></listitem>
8702 <listitem><para>Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
8703 <listitem><para>Enable partial indexes (Martijn van Oosterhout)</para></listitem>
8708 <title>Utility Commands</title>
8710 <listitem><para>Add RESET ALL, SHOW ALL (Marko Kreen)</para></listitem>
8711 <listitem><para>CREATE/ALTER USER/GROUP now allow options in any order (Vince)</para></listitem>
8712 <listitem><para>Add LOCK A, B, C functionality (Neil Padgett)</para></listitem>
8713 <listitem><para>New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)</para></listitem>
8714 <listitem><para>New light-weight VACUUM does not lock table; old semantics are available as VACUUM FULL (Tom)</para></listitem>
8715 <listitem><para>Disable COPY TO/FROM on views (Bruce)</para></listitem>
8716 <listitem><para>COPY DELIMITERS string must be exactly one character (Tom)</para></listitem>
8717 <listitem><para>VACUUM warning about index tuples fewer than heap now only appears when appropriate (Martijn van Oosterhout)</para></listitem>
8718 <listitem><para>Fix privilege checks for CREATE INDEX (Tom)</para></listitem>
8719 <listitem><para>Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)</para></listitem>
8724 <title>Data Types and Functions</title>
8726 <listitem><para>SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)</para></listitem>
8727 <listitem><para>Add convert(), convert2() (Tatsuo)</para></listitem>
8728 <listitem><para>New function bit_length() (Peter E)</para></listitem>
8729 <listitem><para>Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tatsuo)</para></listitem>
8730 <listitem><para>CHAR(), VARCHAR() now reject strings that are too long (Peter E)</para></listitem>
8731 <listitem><para>BIT VARYING now rejects bit strings that are too long (Peter E)</para></listitem>
8732 <listitem><para>BIT now rejects bit strings that do not match declared size (Peter E)</para></listitem>
8733 <listitem><para>INET, CIDR text conversion functions (Alex Pilosov)</para></listitem>
8734 <listitem><para>INET, CIDR operators << and <<= indexable (Alex Pilosov)</para></listitem>
8735 <listitem><para>Bytea \### now requires valid three digit octal number</para></listitem>
8736 <listitem><para>Bytea comparison improvements, now supports =, <>, >, >=, <, and <=</para></listitem>
8737 <listitem><para>Bytea now supports B-tree indexes</para></listitem>
8738 <listitem><para>Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE</para></listitem>
8739 <listitem><para>Bytea now supports concatenation</para></listitem>
8740 <listitem><para>New bytea functions: position, substring, trim, btrim, and length</para></listitem>
8741 <listitem><para>New encode() function mode, "escaped", converts minimally escaped bytea to/from text</para></listitem>
8742 <listitem><para>Add pg_database_encoding_max_length() (Tatsuo)</para></listitem>
8743 <listitem><para>Add pg_client_encoding() function (Tatsuo)</para></listitem>
8744 <listitem><para>now() returns time with millisecond precision (Thomas)</para></listitem>
8745 <listitem><para>New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)</para></listitem>
8746 <listitem><para>Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)</para></listitem>
8747 <listitem><para>New xid/int comparison functions (Hiroshi)</para></listitem>
8748 <listitem><para>Add precision to TIME, TIMESTAMP, and INTERVAL data types (Thomas)</para></listitem>
8749 <listitem><para>Modify type coercion logic to attempt binary-compatible functions first (Tom)</para></listitem>
8750 <listitem><para>New encode() function installed by default (Marko Kreen)</para></listitem>
8751 <listitem><para>Improved to_*() conversion functions (Karel Zak)</para></listitem>
8752 <listitem><para>Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)</para></listitem>
8753 <listitem><para>New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_salt() (Marko Kreen)</para></listitem>
8754 <listitem><para>Correct description of translate() function (Bruce)</para></listitem>
8755 <listitem><para>Add INTERVAL argument for SET TIME ZONE (Thomas)</para></listitem>
8756 <listitem><para>Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)</para></listitem>
8757 <listitem><para>Optimize length functions when using single-byte encodings (Tatsuo)</para></listitem>
8758 <listitem><para>Fix path_inter, path_distance, path_length, dist_ppath to handle closed paths (Curtis Barrett, Tom)</para></listitem>
8759 <listitem><para>octet_length(text) now returns non-compressed length (Tatsuo, Bruce)</para></listitem>
8760 <listitem><para>Handle "July" full name in date/time literals (Greg Sabino Mullane)</para></listitem>
8761 <listitem><para>Some datatype() function calls now evaluated differently</para></listitem>
8762 <listitem><para>Add support for Julian and ISO time specifications (Thomas)</para></listitem>
8767 <title>Internationalization</title>
8769 <listitem><para>National language support in psql, <application>pg_dump</>, libpq, and server (Peter E)</para></listitem>
8770 <listitem><para>Message translations in Chinese (simplified, traditional), Czech, French, German, Hungarian, Russian, Swedish (Peter E, Serguei A. Mokhov, Karel Zak, Weiping He, Zhenbang Wei, Kovacs Zoltan)</para></listitem>
8771 <listitem><para>Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware (Tatsuo)</para></listitem>
8772 <listitem><para>Add LATIN5,6,7,8,9,10 support (Tatsuo)</para></listitem>
8773 <listitem><para>Add ISO 8859-5,6,7,8 support (Tatsuo)</para></listitem>
8774 <listitem><para>Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)</para></listitem>
8775 <listitem><para>Make mic2ascii() non-ASCII aware (Tatsuo)</para></listitem>
8776 <listitem><para>Reject invalid multibyte character sequences (Tatsuo)</para></listitem>
8781 <title><application>PL/pgSQL</></title>
8783 <listitem><para>Now uses portals for SELECT loops, allowing huge result sets (Jan)</para></listitem>
8784 <listitem><para>CURSOR and REFCURSOR support (Jan)</para></listitem>
8785 <listitem><para>Can now return open cursors (Jan)</para></listitem>
8786 <listitem><para>Add ELSEIF (Klaus Reger)</para></listitem>
8787 <listitem><para>Improve PL/pgSQL error reporting, including location of error (Tom)</para></listitem>
8788 <listitem><para>Allow IS or FOR key words in cursor declaration, for compatibility (Bruce)</para></listitem>
8789 <listitem><para>Fix for SELECT ... FOR UPDATE (Tom)</para></listitem>
8790 <listitem><para>Fix for PERFORM returning multiple rows (Tom)</para></listitem>
8791 <listitem><para>Make PL/pgSQL use the server's type coercion code (Tom)</para></listitem>
8792 <listitem><para>Memory leak fix (Jan, Tom)</para></listitem>
8793 <listitem><para>Make trailing semicolon optional (Tom)</para></listitem>
8798 <title>PL/Perl</title>
8800 <listitem><para>New untrusted PL/Perl (Alex Pilosov)</para></listitem>
8801 <listitem><para>PL/Perl is now built on some platforms even if libperl is not shared (Peter E)</para></listitem>
8806 <title>PL/Tcl</title>
8808 <listitem><para>Now reports errorInfo (Vsevolod Lobko)</para></listitem>
8809 <listitem><para>Add spi_lastoid function (bob@redivi.com)</para></listitem>
8814 <title>PL/Python</title>
8816 <listitem><para>...is new (Andrew Bosma)</para></listitem>
8821 <title><application>psql</></title>
8823 <listitem><para>\d displays indexes in unique, primary groupings (Christopher Kings-Lynne)</para></listitem>
8824 <listitem><para>Allow trailing semicolons in backslash commands (Greg Sabino Mullane)</para></listitem>
8825 <listitem><para>Read password from /dev/tty if possible</para></listitem>
8826 <listitem><para>Force new password prompt when changing user and database (Tatsuo, Tom)</para></listitem>
8827 <listitem><para>Format the correct number of columns for Unicode (Patrice)</para></listitem>
8832 <title><application>libpq</></title>
8834 <listitem><para>New function PQescapeString() to escape quotes in command strings (Florian Weimer)</para></listitem>
8835 <listitem><para>New function PQescapeBytea() escapes binary strings for use as SQL string literals</para></listitem>
8842 <listitem><para>Return OID of INSERT (Ken K)</para></listitem>
8843 <listitem><para>Handle more data types (Ken K)</para></listitem>
8844 <listitem><para>Handle single quotes and newlines in strings (Ken K)</para></listitem>
8845 <listitem><para>Handle NULL variables (Ken K)</para></listitem>
8846 <listitem><para>Fix for time zone handling (Barry Lind)</para></listitem>
8847 <listitem><para>Improved Druid support</para></listitem>
8848 <listitem><para>Allow eight-bit characters with non-multibyte server (Barry Lind)</para></listitem>
8849 <listitem><para>Support BIT, BINARY types (Ned Wolpert)</para></listitem>
8850 <listitem><para>Reduce memory usage (Michael Stephens, Dave Cramer)</para></listitem>
8851 <listitem><para>Update DatabaseMetaData (Peter E)</para></listitem>
8852 <listitem><para>Add DatabaseMetaData.getCatalogs() (Peter E)</para></listitem>
8853 <listitem><para>Encoding fixes (Anders Bengtsson)</para></listitem>
8854 <listitem><para>Get/setCatalog methods (Jason Davies)</para></listitem>
8855 <listitem><para>DatabaseMetaData.getColumns() now returns column defaults (Jason Davies)</para></listitem>
8856 <listitem><para>DatabaseMetaData.getColumns() performance improvement (Jeroen van Vianen)</para></listitem>
8857 <listitem><para>Some JDBC1 and JDBC2 merging (Anders Bengtsson)</para></listitem>
8858 <listitem><para>Transaction performance improvements (Barry Lind)</para></listitem>
8859 <listitem><para>Array fixes (Greg Zoller)</para></listitem>
8860 <listitem><para>Serialize addition </para></listitem>
8861 <listitem><para>Fix batch processing (Rene Pijlman)</para></listitem>
8862 <listitem><para>ExecSQL method reorganization (Anders Bengtsson)</para></listitem>
8863 <listitem><para>GetColumn() fixes (Jeroen van Vianen)</para></listitem>
8864 <listitem><para>Fix isWriteable() function (Rene Pijlman)</para></listitem>
8865 <listitem><para>Improved passage of JDBC2 conformance tests (Rene Pijlman)</para></listitem>
8866 <listitem><para>Add bytea type capability (Barry Lind)</para></listitem>
8867 <listitem><para>Add isNullable() (Rene Pijlman)</para></listitem>
8868 <listitem><para>JDBC date/time test suite fixes (Liam Stewart)</para></listitem>
8869 <listitem><para>Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)</para></listitem>
8870 <listitem><para>Fix DatabaseMetaData to show precision properly (Mark Lillywhite)</para></listitem>
8871 <listitem><para>New getImported/getExported keys (Jason Davies)</para></listitem>
8872 <listitem><para>MD5 password encryption support (Jeremy Wohl)</para></listitem>
8873 <listitem><para>Fix to actually use type cache (Ned Wolpert)</para></listitem>
8880 <listitem><para>Remove query size limit (Hiroshi)</para></listitem>
8881 <listitem><para>Remove text field size limit (Hiroshi)</para></listitem>
8882 <listitem><para>Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)</para></listitem>
8883 <listitem><para>Allow ODBC procedure calls (Hiroshi)</para></listitem>
8884 <listitem><para>Improve boolean handing (Aidan Mountford)</para></listitem>
8885 <listitem><para>Most configuration options now settable via DSN (Hiroshi)</para></listitem>
8886 <listitem><para>Multibyte, performance fixes (Hiroshi)</para></listitem>
8887 <listitem><para>Allow driver to be used with iODBC or unixODBC (Peter E)</para></listitem>
8888 <listitem><para>MD5 password encryption support (Bruce)</para></listitem>
8889 <listitem><para>Add more compatibility functions to odbc.sql (Peter E)</para></listitem>
8894 <title><application>ECPG</></title>
8896 <listitem><para>EXECUTE ... INTO implemented (Christof Petig)</para></listitem>
8897 <listitem><para>Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)</para></listitem>
8898 <listitem><para>Fix for GRANT parameters (Lee Kindness)</para></listitem>
8899 <listitem><para>Fix INITIALLY DEFERRED bug</para></listitem>
8900 <listitem><para>Various bug fixes (Michael, Christof Petig)</para></listitem>
8901 <listitem><para>Auto allocation for indicator variable arrays (int *ind_p=NULL)</para></listitem>
8902 <listitem><para>Auto allocation for string arrays (char **foo_pp=NULL)</para></listitem>
8903 <listitem><para>ECPGfree_auto_mem fixed</para></listitem>
8904 <listitem><para>All function names with external linkage are now prefixed by ECPG</para></listitem>
8905 <listitem><para>Fixes for arrays of structures (Michael)</para></listitem>
8910 <title>Misc. Interfaces</title>
8912 <listitem><para>Python fix fetchone() (Gerhard Haring)</para></listitem>
8913 <listitem><para>Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Max)</para></listitem>
8914 <listitem><para>Add Tcl COPY TO/FROM (ljb)</para></listitem>
8915 <listitem><para>Prevent output of default index op class in <application>pg_dump</> (Tom)</para></listitem>
8916 <listitem><para>Fix libpgeasy memory leak (Bruce)</para></listitem>
8921 <title>Build and Install</title>
8923 <listitem><para>Configure, dynamic loader, and shared library fixes (Peter E)</para></listitem>
8924 <listitem><para>Fixes in QNX 4 port (Bernd Tegge)</para></listitem>
8925 <listitem><para>Fixes in Cygwin and Windows ports (Jason Tishler, Gerhard Haring, Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)</para></listitem>
8926 <listitem><para>Fix for Windows socket communication failures (Magnus, Mikhail Terekhov)</para></listitem>
8927 <listitem><para>Hurd compile fix (Oliver Elphick)</para></listitem>
8928 <listitem><para>BeOS fixes (Cyril Velter)</para></listitem>
8929 <listitem><para>Remove configure --enable-unicode-conversion, now enabled by multibyte (Tatsuo)</para></listitem>
8930 <listitem><para>AIX fixes (Tatsuo, Andreas)</para></listitem>
8931 <listitem><para>Fix parallel make (Peter E)</para></listitem>
8932 <listitem><para>Install SQL language manual pages into OS-specific directories (Peter E)</para></listitem>
8933 <listitem><para>Rename config.h to pg_config.h (Peter E)</para></listitem>
8934 <listitem><para>Reorganize installation layout of header files (Peter E)</para></listitem>
8939 <title>Source Code</title>
8941 <listitem><para>Remove SEP_CHAR (Bruce)</para></listitem>
8942 <listitem><para>New GUC hooks (Tom)</para></listitem>
8943 <listitem><para>Merge GUC and command line handling (Marko Kreen)</para></listitem>
8944 <listitem><para>Remove EXTEND INDEX (Martijn van Oosterhout, Tom)</para></listitem>
8945 <listitem><para>New pgjindent utility to indent java code (Bruce)</para></listitem>
8946 <listitem><para>Remove define of true/false when compiling under C++ (Leandro Fanzone, Tom)</para></listitem>
8947 <listitem><para>pgindent fixes (Bruce, Tom)</para></listitem>
8948 <listitem><para>Replace strcasecmp() with strcmp() where appropriate (Peter E)</para></listitem>
8949 <listitem><para>Dynahash portability improvements (Tom)</para></listitem>
8950 <listitem><para>Add 'volatile' usage in spinlock structures</para></listitem>
8951 <listitem><para>Improve signal handling logic (Tom)</para></listitem>
8956 <title>Contrib</title>
8958 <listitem><para>New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)</para></listitem>
8959 <listitem><para>New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)</para></listitem>
8960 <listitem><para>Add contrib/dblink for remote database access (Joe Conway)</para></listitem>
8961 <listitem><para>contrib/ora2pg Oracle conversion utility (Gilles Darold)</para></listitem>
8962 <listitem><para>contrib/xml XML conversion utility (John Gray)</para></listitem>
8963 <listitem><para>contrib/fulltextindex fixes (Christopher Kings-Lynne)</para></listitem>
8964 <listitem><para>New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merged (Joe Conway)</para></listitem>
8965 <listitem><para>Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartunov)</para></listitem>
8966 <listitem><para>New pg_upgrade utility (Bruce)</para></listitem>
8967 <listitem><para>Add new pg_resetxlog options (Bruce, Tom)</para></listitem>
8974 <sect1 id="release-7-1-3">
8975 <title>Release 7.1.3</title>
8978 <title>Release date</title>
8979 <simpara>2001-08-15</simpara>
8983 <title>Migration to version 7.1.3</title>
8986 A dump/restore is <emphasis>not</emphasis> required for those running
8992 <title>Changes</title>
8996 Remove unused WAL segements of large transactions (Tom)
8997 Multiaction rule fix (Tom)
8998 PL/pgSQL memory allocation fix (Jan)
8999 VACUUM buffer fix (Tom)
9000 Regression test fixes (Tom)
9001 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
9002 Fix subselects with DISTINCT ON or LIMIT (Tom)
9004 Disable COPY TO/FROM a view (Tom)
9005 Cygwin build (Jason Tishler)
9012 <sect1 id="release-7-1-2">
9013 <title>Release 7.1.2</title>
9016 <title>Release date</title>
9017 <simpara>2001-05-11</simpara>
9021 This has one fix from 7.1.1.
9026 <title>Migration to version 7.1.2</title>
9029 A dump/restore is <emphasis>not</emphasis> required for those running
9035 <title>Changes</title>
9039 Fix PL/pgSQL SELECTs when returning no rows
9040 Fix for psql backslash core dump
9041 Referential integrity privilege fix
9050 <sect1 id="release-7-1-1">
9051 <title>Release 7.1.1</title>
9054 <title>Release date</title>
9055 <simpara>2001-05-05</simpara>
9059 This has a variety of fixes from 7.1.
9064 <title>Migration to version 7.1.1</title>
9067 A dump/restore is <emphasis>not</emphasis> required for those running
9073 <title>Changes</title>
9077 Fix for numeric MODULO operator (Tom)
9078 pg_dump fixes (Philip)
9079 pg_dump can dump 7.0 databases (Philip)
9080 readline 4.2 fixes (Peter E)
9082 AIX, MSWIN, VAX, N32K fixes (Tom)
9083 Multibytes fixes (Tom)
9084 Unicode fixes (Tatsuo)
9085 Optimizer improvements (Tom)
9086 Fix for whole rows in functions (Tom)
9087 Fix for pg_ctl and option strings with spaces (Peter E)
9088 ODBC fixes (Hiroshi)
9089 EXTRACT can now take string argument (Thomas)
9090 Python fixes (Darcy)
9097 <sect1 id="release-7-1">
9098 <title>Release 7.1</title>
9101 <title>Release date</title>
9102 <simpara>2001-04-13</simpara>
9106 This release focuses on removing limitations that have existed in the
9107 <productname>PostgreSQL</productname> code for many years.
9111 Major changes in this release:
9117 Write-ahead Log (WAL)
9121 To maintain database consistency in case of an operating system crash,
9122 previous releases of <productname>PostgreSQL</productname> have forced
9123 all data modifications to disk before each transaction commit. With
9124 WAL, only one log file must be flushed to disk, greatly improving
9125 performance. If you have been using -F in previous releases to
9126 disable disk flushes, you may want to consider discontinuing its use.
9137 TOAST - Previous releases had a compiled-in row length limit,
9138 typically 8k - 32k. This limit made storage of long text fields
9139 difficult. With TOAST, long rows of any length can be stored with good
9151 We now support outer joins. The UNION/NOT IN
9152 workaround for outer joins is no longer required. We use the SQL92
9164 The previous C function manager did not
9165 handle null values properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
9166 function manager does. You can continue using your old custom
9167 functions, but you may want to rewrite them in the future to use the new
9168 function manager call interface.
9179 A large number of complex queries that were
9180 unsupported in previous releases now work. Many combinations of views,
9181 aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
9182 now work properly. Inherited tables are now accessed by default.
9183 Subqueries in FROM are now supported.
9192 <title>Migration to version 7.1</title>
9195 A dump/restore using pg_dump is required for those wishing to migrate
9196 data from any previous release.
9201 <title>Changes</title>
9207 Many multibyte/Unicode/locale fixes (Tatsuo and others)
9208 More reliable ALTER TABLE RENAME (Tom)
9209 Kerberos V fixes (David Wragg)
9210 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
9211 Prompt username/password on standard error (Bruce)
9212 Large objects inv_read/inv_write fixes (Tom)
9213 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
9215 Prevent query expressions from leaking memory (Tom)
9216 Allow UPDATE of arrays elements (Tom)
9217 Wake up lock waiters during cancel (Hiroshi)
9218 Fix rare cursor crash when using hash join (Tom)
9219 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
9220 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
9221 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
9222 Fix PL/perl (Alex Kapranoff)
9223 Disallow LOCK on views (Mark Hollomon)
9224 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
9225 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
9226 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
9227 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
9228 Fix rare failure with TRUNCATE command (Tom)
9229 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
9230 DISTINCT, ORDER BY, SELECT...INTO (Tom)
9231 Fix parser failures during aborted transactions (Tom)
9232 Allow temporary relations to properly clean up indexes (Bruce)
9233 Fix VACUUM problem with moving rows in same page (Tom)
9234 Modify pg_dump to better handle user-defined items in template1 (Philip)
9235 Allow LIMIT in VIEW (Tom)
9236 Require cursor FETCH to honor LIMIT (Tom)
9237 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
9238 Allow ORDER BY, LIMIT in subqueries (Tom)
9239 Allow UNION in CREATE RULE (Tom)
9240 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
9241 Store initdb collation in pg_control so collation cannot be changed (Tom)
9242 Fix INSERT...SELECT with rules (Tom)
9243 Fix FOR UPDATE inside views and subselects (Tom)
9244 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
9245 Fix lpad() and rpad() to handle length less than input string (Tom)
9246 Fix use of NOTIFY in some rules (Tom)
9247 Overhaul btree code (Tom)
9248 Fix NOT NULL use in Pl/pgSQL variables (Tom)
9249 Overhaul GIST code (Oleg)
9250 Fix CLUSTER to preserve constraints and column default (Tom)
9251 Improved deadlock detection handling (Tom)
9252 Allow multiple SERIAL columns in a table (Tom)
9253 Prevent occasional index corruption (Vadim)
9257 Add OUTER JOINs (Tom)
9258 Function manager overhaul (Tom)
9259 Allow ALTER TABLE RENAME on indexes (Tom)
9260 Improve CLUSTER (Tom)
9261 Improve ps status display for more platforms (Peter E, Marc)
9262 Improve CREATE FUNCTION failure message (Ross)
9263 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
9265 Grand Unified Configuration scheme/GUC. Many options can now be set in
9266 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
9267 Improved handling of file descriptor cache (Tom)
9268 New warning code about auto-created table alias entries (Bruce)
9269 Overhaul initdb process (Tom, Peter E)
9270 Overhaul of inherited tables; inherited tables now accessed by default;
9271 new ONLY key word prevents it (Chris Bitmead, Tom)
9272 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
9274 Allow renaming of temp tables (Tom)
9275 Overhaul memory manager contexts (Tom)
9276 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
9277 Overhaul pg_dump (Philip Warner)
9278 Allow pg_hba.conf secondary password file to specify only username (Peter E)
9279 Allow TEMPORARY or TEMP key word when creating temporary tables (Bruce)
9280 New memory leak checker (Karel)
9281 New SET SESSION CHARACTERISTICS (Thomas)
9282 Allow nested block comments (Thomas)
9283 Add WITHOUT TIME ZONE type qualifier (Thomas)
9284 New ALTER TABLE ADD CONSTRAINT (Stephan)
9285 Use NUMERIC accumulators for INTEGER aggregates (Tom)
9286 Overhaul aggregate code (Tom)
9287 New VARIANCE and STDDEV() aggregates
9288 Improve dependency ordering of pg_dump (Philip)
9289 New pg_restore command (Philip)
9290 New pg_dump tar output option (Philip)
9291 New pg_dump of large objects (Philip)
9292 New ESCAPE option to LIKE (Thomas)
9293 New case-insensitive LIKE - ILIKE (Thomas)
9294 Allow functional indexes to use binary-compatible type (Tom)
9295 Allow SQL functions to be used in more contexts (Tom)
9296 New pg_config utility (Peter E)
9297 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
9299 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
9300 New quote_identifiers() and quote_literal() functions (Jan)
9301 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
9302 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
9303 Update PyGreSQL to version 3.1 (D'Arcy)
9304 Store tables as files named by OID (Vadim)
9305 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
9306 Require DROP VIEW to remove views, no DROP TABLE (Mark)
9307 Allow DROP VIEW view1, view2 (Mark)
9308 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
9309 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
9310 New /contrib/pgcrypto hashing functions (Marko Kreen)
9311 New pg_dumpall --globals-only option (Peter E)
9312 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
9313 New AT TIME ZONE syntax (Thomas)
9314 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
9315 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
9316 Allow socket path name to be specified in hostname by using leading slash
9317 (David J. MacKenzie)
9318 Allow CREATE DATABASE to specify template database (Tom)
9319 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
9320 New /contrib/rserv replication toolkit (Vadim)
9321 New file format for COPY BINARY (Tom)
9322 New /contrib/oid2name to map numeric files to table names (B Palmer)
9323 New "idle in transaction" ps status message (Marc)
9324 Update to pgaccess 0.98.7 (Constantin Teodorescu)
9325 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
9326 Add rudimentary dependency checking to pg_dump (Philip)
9330 Fix INET/CIDR type ordering and add new functions (Tom)
9331 Make OID behave as an unsigned type (Tom)
9332 Allow BIGINT as synonym for INT8 (Peter E)
9333 New int2 and int8 comparison operators (Tom)
9334 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
9335 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
9336 New GIST seg/cube examples (Gene Selkov)
9337 Improved round(numeric) handling (Tom)
9338 Fix CIDR output formatting (Tom)
9339 New CIDR abbrev() function (Tom)
9343 Write-Ahead Log (WAL) to provide crash recovery with less performance
9345 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
9346 Reduced file seeks (Denis Perchine)
9347 Improve BTREE code for duplicate keys (Tom)
9348 Store all large objects in a single table (Denis Perchine, Tom)
9349 Improve memory allocation performance (Karel, Tom)
9353 New function manager call conventions (Tom)
9354 SGI portability fixes (David Kaelbling)
9355 New configure --enable-syslog option (Peter E)
9356 New BSDI README (Bruce)
9357 configure script moved to top level, not /src (Peter E)
9358 Makefile/configuration/compilation overhaul (Peter E)
9359 New configure --with-python option (Peter E)
9360 Solaris cleanups (Peter E)
9361 Overhaul /contrib Makefiles (Karel)
9362 New OpenSSL configuration option (Magnus, Peter E)
9364 QNX fixes (Maurizio)
9365 New heap_open(), heap_openr() API (Tom)
9366 Remove colon and semi-colon operators (Thomas)
9367 New pg_class.relkind value for views (Mark Hollomon)
9368 Rename ichar() to chr() (Karel)
9369 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
9370 Fixes for NT/Cygwin (Pete Forman)
9371 AIX port fixes (Andreas)
9372 New BeOS port (David Reid, Cyril Velter)
9373 Add proofreader's changes to docs (Addison-Wesley, Bruce)
9374 New Alpha spinlock code (Adriaan Joubert, Compaq)
9375 UnixWare port overhaul (Peter E)
9376 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
9377 New FreeBSD Alpha port (Alfred)
9378 Overhaul shared memory segments (Tom)
9379 Add IBM S/390 support (Neale Ferguson)
9380 Moved macmanuf to /contrib (Larry Rosenman)
9381 Syslog improvements (Larry Rosenman)
9382 New template0 database that contains no user additions (Tom)
9383 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
9384 Allow NetBSD's libedit instead of readline (Peter)
9385 Improved assembly language source code format (Bruce)
9386 New contrib/pg_logger
9387 New --template option to createdb
9388 New contrib/pg_control utility (Oliver)
9389 New FreeBSD tools ipc_check, start-scripts/freebsd
9396 <sect1 id="release-7-0-3">
9397 <title>Release 7.0.3</title>
9400 <title>Release date</title>
9401 <simpara>2000-11-11</simpara>
9405 This has a variety of fixes from 7.0.2.
9410 <title>Migration to version 7.0.3</title>
9413 A dump/restore is <emphasis>not</emphasis> required for those running
9419 <title>Changes</title>
9424 Large object fix (Tom)
9425 Fix lean in COPY WITH OIDS leak (Tom)
9426 Fix backwards-index-scan (Tom)
9427 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
9428 Add --enable-syslog to configure (Marc)
9429 Fix abort transaction at backend exit in rare cases (Tom)
9430 Fix for psql \l+ when multibyte enabled (Tatsuo)
9431 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
9432 Make vacuum always flush buffers (Tom)
9433 Fix to allow cancel while waiting for a lock (Hiroshi)
9434 Fix for memory aloocation problem in user authentication code (Tom)
9435 Remove bogus use of int4out() (Tom)
9436 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
9437 Fix for failure of triggers on heap open in certain cases (Jeroen van
9439 Fix for erroneous selectivity of not-equals (Tom)
9440 Fix for erroneous use of strcmp() (Tom)
9441 Fix for bug where storage manager accesses items beyond end of file
9443 Fix to include kernel errno message in all smgr elog messages (Tom)
9444 Fix for '.' not in PATH at build time (SL Baur)
9445 Fix for out-of-file-descriptors error (Tom)
9446 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
9447 Fix for subselect in targetlist of Append node (Tom)
9448 Fix for mergejoin plans (Tom)
9449 Fix TRUNCATE failure on relations with indexes (Tom)
9450 Avoid database-wide restart on write error (Hiroshi)
9451 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
9452 Fix VACUUM problem with moving chain of update row versions when source
9453 and destination of a row version lie on the same page (Tom)
9454 Fix user.c CommandCounterIncrement (Tom)
9455 Fix for AM/PM boundary problem in to_char() (Karel Zak)
9456 Fix TIME aggregate handling (Tom)
9457 Fix to_char() to avoid coredump on NULL input (Tom)
9459 Fix for inserting/copying longer multibyte strings into char() data
9461 Fix for crash of backend, on abort (Tom)
9468 <sect1 id="release-7-0-2">
9469 <title>Release 7.0.2</title>
9472 <title>Release date</title>
9473 <simpara>2000-06-05</simpara>
9477 This is a repackaging of 7.0.1 with added documentation.
9482 <title>Migration to version 7.0.2</title>
9485 A dump/restore is <emphasis>not</emphasis> required for those running
9491 <title>Changes</title>
9495 Added documentation to tarball.
9502 <sect1 id="release-7-0-1">
9503 <title>Release 7.0.1</title>
9506 <title>Release date</title>
9507 <simpara>2000-06-01</simpara>
9511 This is a cleanup release for 7.0.
9515 <title>Migration to version 7.0.1</title>
9518 A dump/restore is <emphasis>not</emphasis> required for those running
9524 <title>Changes</title>
9528 Fix many CLUSTER failures (Tom)
9529 Allow ALTER TABLE RENAME works on indexes (Tom)
9530 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
9531 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
9532 Fix the off by one errors in ResultSet from 6.5.3, and more.
9533 jdbc ResultSet fixes (Joseph Shraibman)
9534 optimizer tunings (Tom)
9535 Fix create user for pgaccess
9536 Fix for UNLISTEN failure
9537 IRIX fixes (David Kaelbling)
9538 QNX fixes (Andreas Kardos)
9539 Reduce COPY IN lock level (Tom)
9540 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
9541 Fix pg_dump to handle OID indexes (Tom)
9542 Fix small memory leak (Tom)
9543 Solaris fix for createdb/dropdb (Tatsuo)
9544 Fix for non-blocking connections (Alfred Perlstein)
9545 Fix improper recovery after RENAME TABLE failures (Tom)
9546 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
9547 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
9548 Fix too long syslog message (Tatsuo)
9549 Fix problem with quoted indexes that are too long (Tom)
9550 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
9551 ecpg changes (Michael)
9557 <sect1 id="release-7-0">
9558 <title>Release 7.0</title>
9561 <title>Release date</title>
9562 <simpara>2000-05-08</simpara>
9566 This release contains improvements in many areas, demonstrating
9567 the continued growth of <productname>PostgreSQL</productname>.
9568 There are more improvements and fixes in 7.0 than in any previous
9569 release. The developers have confidence that this is the best
9570 release yet; we do our best to put out only solid releases, and
9571 this one is no exception.
9575 Major changes in this release:
9585 Foreign keys are now implemented, with the exception of PARTIAL MATCH
9586 foreign keys. Many users have been asking for this feature, and we are
9587 pleased to offer it.
9598 Continuing on work started a year ago, the optimizer has been
9599 improved, allowing better query plan selection and faster performance
9600 with less memory usage.
9607 Updated <application>psql</application>
9611 <application>psql</application>, our interactive terminal monitor, has been
9612 updated with a variety of new features. See the <application>psql</application> manual page for details.
9623 SQL92 join syntax is now supported, though only as
9624 <literal>INNER JOIN</> for this release. <literal>JOIN</>,
9625 <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
9626 and <literal>JOIN</>/<literal>ON</> are available, as are
9627 column correlation names.
9635 <title>Migration to version 7.0</title>
9638 A dump/restore using <application>pg_dump</application>
9639 is required for those wishing to migrate data from any
9640 previous release of <productname>PostgreSQL</productname>.
9641 For those upgrading from 6.5.*, you may instead use
9642 <application>pg_upgrade</application> to upgrade to this
9643 release; however, a full dump/reload installation is always the
9644 most robust method for upgrades.
9648 Interface and compatibility issues to consider for the new
9655 The date/time types <type>datetime</type> and
9656 <type>timespan</type> have been superseded by the
9657 SQL92-defined types <type>timestamp</type> and
9658 <type>interval</type>. Although there has been some effort to
9659 ease the transition by allowing
9660 <productname>PostgreSQL</productname> to recognize
9661 the deprecated type names and translate them to the new type
9662 names, this mechanism may not be completely transparent to
9663 your existing application.
9669 The optimizer has been substantially improved in the area of
9670 query cost estimation. In some cases, this will result in
9671 decreased query times as the optimizer makes a better choice
9672 for the preferred plan. However, in a small number of cases,
9673 usually involving pathological distributions of data, your
9674 query times may go up. If you are dealing with large amounts
9675 of data, you may want to check your queries to verify
9682 The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
9683 interfaces have been upgraded and extended.
9689 The string function <function>CHAR_LENGTH</function> is now a
9690 native function. Previous versions translated this into a call
9691 to <function>LENGTH</function>, which could result in
9692 ambiguity with other types implementing
9693 <function>LENGTH</function> such as the geometric types.
9700 <title>Changes</title>
9706 Prevent function calls exceeding maximum number of arguments (Tom)
9707 Improve CASE construct (Tom)
9708 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
9709 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
9710 Fix GROUP BY scan bug (Tom)
9711 Improvements in SQL grammar processing (Tom)
9712 Fix for views involved in INSERT ... SELECT ... (Tom)
9713 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
9714 Fix for subselects in INSERT ... SELECT (Tom)
9715 Prevent INSERT ... SELECT ... ORDER BY (Tom)
9716 Fixes for relations greater than 2GB, including vacuum
9717 Improve propagating system table changes to other backends (Tom)
9718 Improve propagating user table changes to other backends (Tom)
9719 Fix handling of temp tables in complex situations (Bruce, Tom)
9720 Allow table locking at table open, improving concurrent reliability (Tom)
9721 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
9722 Prevent DROP DATABASE while others accessing
9723 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
9724 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
9725 Fix pg_upgrade so it works for MVCC (Tom)
9726 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
9727 Fix for "f1 datetime DEFAULT 'now'" (Tom)
9728 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
9729 Allow comment-only lines, and ;;; lines too. (Tom)
9730 Improve recovery after failed disk writes, disk full (Hiroshi)
9731 Fix cases where table is mentioned in FROM but not joined (Tom)
9732 Allow HAVING clause without aggregate functions (Tom)
9733 Fix for "--" comment and no trailing newline, as seen in perl interface
9734 Improve pg_dump failure error reports (Bruce)
9735 Allow sorts and hashes to exceed 2GB file sizes (Tom)
9736 Fix for pg_dump dumping of inherited rules (Tom)
9737 Fix for NULL handling comparisons (Tom)
9738 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
9739 Fix for dbname with dash
9740 Prevent DROP INDEX from interfering with other backends (Tom)
9741 Fix file descriptor leak in verify_password()
9742 Fix for "Unable to identify an operator =$" problem
9743 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
9744 Fix for recursive exit call (Massimo)
9745 Fix for extra-long timezones (Jeroen van Vianen)
9746 Make pg_dump preserve primary key information (Peter E)
9747 Prevent databases with single quotes (Peter E)
9748 Prevent DROP DATABASE inside transaction (Peter E)
9749 ecpg memory leak fixes (Stephen Birch)
9750 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
9751 Y2K timestamp fix (Massimo)
9752 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
9753 Fix for views with tables/columns containing spaces (Tom)
9754 Prevent privileges on indexes (Peter E)
9755 Fix for spinlock stuck problem when error is generated (Hiroshi)
9756 Fix ipcclean on Linux
9757 Fix handling of NULL constraint conditions (Tom)
9758 Fix memory leak in odbc driver (Nick Gorham)
9759 Fix for privilege check on UNION tables (Tom)
9760 Fix to allow SELECT 'a' LIKE 'a' (Tom)
9761 Fix for SELECT 1 + NULL (Tom)
9763 Fix log() on numeric type (Tom)
9764 Deprecate ':' and ';' operators
9765 Allow vacuum of temporary tables
9766 Disallow inherited columns with the same name as new columns
9767 Recover or force failure when disk space is exhausted (Hiroshi)
9768 Fix INSERT INTO ... SELECT with AS columns matching result columns
9769 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns (Tom)
9770 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
9771 Fix UNION with LIMIT
9772 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
9773 Fix CREATE TABLE test(col char(2) DEFAULT user)
9774 Fix mismatched types in CREATE TABLE ... DEFAULT
9775 Fix SELECT * FROM pg_class where oid in (0,-1)
9776 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
9777 Prevent user who can create databases can modifying pg_database table (Peter E)
9778 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
9779 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
9783 New CLI interface include file sqlcli.h, based on SQL3/SQL98
9784 Remove all limits on query length, row length limit still exists (Tom)
9785 Update jdbc protocol to 2.0 (Jens Glaser <email>jens@jens.de</email>)
9786 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
9787 Fix to give super user and createdb user proper update catalog rights (Peter E)
9788 Allow ecpg bool variables to have NULL values (Christof)
9789 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
9790 Allow ^C to cancel COPY command (Massimo)
9791 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
9792 Function name overloading for dynamically-loaded C functions (Frankpitt)
9793 Add CmdTuples() to libpq++(Vince)
9794 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
9795 Allow CREATE FUNCTION/WITH clause to be used for all language types
9796 configure --enable-debug adds -g (Peter E)
9797 configure --disable-debug removes -g (Peter E)
9798 Allow more complex default expressions (Tom)
9799 First real FOREIGN KEY constraint trigger functionality (Jan)
9800 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
9801 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
9802 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
9803 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
9804 Change pgeasy connectdb() parameter ordering (Bruce)
9805 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
9806 Add Oracle's COMMENT ON command (Mike Mascari <email>mascarim@yahoo.com</email>)
9807 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
9808 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
9809 Make USING in COPY optional (Bruce)
9810 Allow subselects in the target list (Tom)
9811 Allow subselects on the left side of comparison operators (Tom)
9812 New parallel regression test (Jan)
9813 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
9814 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
9815 Added psql LASTOID variable to return last inserted oid (Peter E)
9816 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
9817 Add privilege check for vacuum (Peter E)
9818 New libpq functions to allow asynchronous connections: PQconnectStart(),
9819 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
9820 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
9821 New libpq PQsetenv() function (Ewan Mellor)
9822 create/alter user extension (Peter E)
9823 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
9824 New scripts for create/drop user/db (Peter E)
9825 Major psql overhaul (Peter E)
9826 Add const to libpq interface (Peter E)
9827 New libpq function PQoidValue (Peter E)
9828 Show specific non-aggregate causing problem with GROUP BY (Tom)
9829 Make changes to pg_shadow recreate pg_pwd file (Peter E)
9830 Add aggregate(DISTINCT ...) (Tom)
9831 Allow flag to control COPY input/output of NULLs (Peter E)
9832 Make postgres user have a password by default (Peter E)
9833 Add CREATE/ALTER/DROP GROUP (Peter E)
9834 All administration scripts now support --long options (Peter E, Karel)
9835 Vacuumdb script now supports --all option (Peter E)
9836 ecpg new portable FETCH syntax
9837 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
9838 and EXEC SQL ENDIF directives
9839 Add pg_ctl script to control backend start-up (Tatsuo)
9840 Add postmaster.opts.default file to store start-up flags (Tatsuo)
9841 Allow --with-mb=SQL_ASCII
9842 Increase maximum number of index keys to 16 (Bruce)
9843 Increase maximum number of function arguments to 16 (Bruce)
9844 Allow configuration of maximum number of index keys and arguments (Bruce)
9845 Allow unprivileged users to change their passwords (Peter E)
9846 Password authentication enabled; required for new users (Peter E)
9847 Disallow dropping a user who owns a database (Peter E)
9848 Change initdb option --with-mb to --enable-multibyte
9849 Add option for initdb to prompts for superuser password (Peter E)
9850 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
9851 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
9852 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
9853 libpq non-blocking mode (Alfred Perlstein)
9854 Improve conversion of types in casts that don't specify a length
9855 New plperl internal programming language (Mark Hollomon)
9856 Allow COPY IN to read file that do not end with a newline (Tom)
9857 Indicate when long identifiers are truncated (Tom)
9858 Allow aggregates to use type equivalency (Peter E)
9859 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
9860 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
9861 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
9862 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
9863 Add NUMERIC and int8 types to ODBC
9864 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
9865 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
9866 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
9867 Enable backward sequential scan even after reaching EOF (Hiroshi)
9868 Add btree indexing of boolean values, >= and <= (Don Baccus)
9869 Print current line number when COPY FROM fails (Massimo)
9870 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
9871 Add DEC as synonym for DECIMAL (Thomas)
9872 Add SESSION_USER as SQL92 key word, same as CURRENT_USER (Thomas)
9873 Implement SQL92 column aliases (aka correlation names) (Thomas)
9874 Implement SQL92 join syntax (Thomas)
9875 Make INTERVAL reserved word allowed as a column identifier (Thomas)
9876 Implement REINDEX command (Hiroshi)
9877 Accept ALL in aggregate function SUM(ALL col) (Tom)
9878 Prevent GROUP BY from using column aliases (Tom)
9879 New psql \encoding option (Tatsuo)
9880 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
9881 Allow negation of a negative number in all cases
9882 Add ecpg descriptors (Christof, Michael)
9883 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
9884 Allow casts with length, like foo::char(8)
9885 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
9886 Add support for SJIS user defined characters (Tatsuo)
9887 Larger views/rules supported
9888 Make libpq's PQconndefaults() thread-safe (Tom)
9889 Disable // as comment to be ANSI conforming, should use -- (Tom)
9890 Allow column aliases on views CREATE VIEW name (collist)
9891 Fixes for views with subqueries (Tom)
9892 Allow UPDATE table SET fld = (SELECT ...) (Tom)
9893 SET command options no longer require quotes
9894 Update pgaccess to 0.98.6
9895 New SET SEED command
9896 New pg_options.sample file
9897 New SET FSYNC command (Massimo)
9898 Allow pg_descriptions when creating tables
9899 Allow pg_descriptions when creating types, columns, and functions
9900 Allow psql \copy to allow delimiters (Peter E)
9901 Allow psql to print nulls as distinct from "" [null] (Peter E)
9905 Many array fixes (Tom)
9906 Allow bare column names to be subscripted as arrays (Tom)
9907 Improve type casting of int and float constants (Tom)
9908 Cleanups for int8 inputs, range checking, and type conversion (Tom)
9909 Fix for SELECT timespan('21:11:26'::time) (Tom)
9910 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
9911 Add btree index on NUMERIC (Jan)
9912 Perl fix for large objects containing NUL characters (Douglas Thomson)
9913 ODBC fix for for large objects (free)
9914 Fix indexing of cidr data type
9915 Fix for Ethernet MAC addresses (macaddr type) comparisons
9916 Fix for date/time types when overflows happened in computations (Tom)
9917 Allow array on int8 (Peter E)
9918 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
9919 Allow NUMERIC arrays
9920 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
9921 Make char_length()/octet_length including trailing blanks (Tom)
9922 Made abstime/reltime use int4 instead of time_t (Peter E)
9923 New lztext data type for compressed text fields
9924 Revise code to handle coercion of int and float constants (Tom)
9925 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
9926 NUMERIC now accepts scientific notation (Tom)
9927 NUMERIC to int4 rounds (Tom)
9928 Convert float4/8 to NUMERIC properly (Tom)
9929 Allow type conversion with NUMERIC (Thomas)
9930 Make ISO date style (2000-02-16 09:33) the default (Thomas)
9931 Add NATIONAL CHAR [ VARYING ] (Thomas)
9932 Allow NUMERIC round and trunc to accept negative scales (Tom)
9933 New TIME WITH TIME ZONE type (Thomas)
9934 Add MAX()/MIN() on time type (Thomas)
9935 Add abs(), mod(), fac() for int8 (Thomas)
9936 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
9937 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
9938 Add exp() and ln() for NUMERIC type
9939 Rename NUMERIC power() to pow() (Thomas)
9940 Improved TRANSLATE() function (Edwin Ramirez, Tom)
9941 Allow X=-Y operators (Tom)
9942 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom)
9943 Allow LOCALE to use indexes in regular expression searches (Tom)
9944 Allow creation of functional indexes to use default types
9948 Prevent exponential space consumption with many AND's and OR's (Tom)
9949 Collect attribute selectivity values for system columns (Tom)
9950 Reduce memory usage of aggregates (Tom)
9951 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
9952 Fix r-tree index optimizer selectivity (Thomas)
9953 Improve optimizer selectivity computations and functions (Tom)
9954 Optimize btree searching for cases where many equal keys exist (Tom)
9955 Enable fast LIKE index processing only if index present (Tom)
9956 Re-use free space on index pages with duplicates (Tom)
9957 Improve hash join processing (Tom)
9958 Prevent descending sort if result is already sorted(Hiroshi)
9959 Allow commuting of index scan query qualifications (Tom)
9960 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
9961 Allocate large memory requests in fix-sized chunks for performance (Tom)
9962 Fix vacuum's performance by reducing memory allocation requests (Tom)
9963 Implement constant-expression simplification (Bernard Frankpitt, Tom)
9964 Use secondary columns to be used to determine start of index scan (Hiroshi)
9965 Prevent quadruple use of disk space when doing internal sorting (Tom)
9966 Faster sorting by calling fewer functions (Tom)
9967 Create system indexes to match all system caches (Bruce, Hiroshi)
9968 Make system caches use system indexes (Bruce)
9969 Make all system indexes unique (Bruce)
9970 Improve pg_statistics management for VACUUM speed improvement (Tom)
9971 Flush backend cache less frequently (Tom, Hiroshi)
9972 COPY now reuses previous memory allocation, improving performance (Tom)
9973 Improve optimization cost estimation (Tom)
9974 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom)
9975 Use DNF instead of CNF where appropriate (Tom, Taral)
9976 Further cleanup for OR-of-AND WHERE-clauses (Tom)
9977 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
9978 Smarter optimizer computations for random index page access (Tom)
9979 New SET variable to control optimizer costs (Tom)
9980 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
9981 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
9982 Major subquery speedup (Tom)
9983 Fewer fsync writes when fsync is not disabled (Tom)
9984 Improved LIKE optimizer estimates (Tom)
9985 Prevent fsync in SELECT-only queries (Vadim)
9986 Make index creation use psort code, because it is now faster (Tom)
9987 Allow creation of sort temp tables > 1 Gig
9991 Fix for linux PPC compile
9992 New generic expression-tree-walker subroutine (Tom)
9993 Change form() to varargform() to prevent portability problems
9994 Improved range checking for large integers on Alphas
9995 Clean up #include in /include directory (Bruce)
9996 Add scripts for checking includes (Bruce)
9997 Remove un-needed #include's from *.c files (Bruce)
9998 Change #include's to use <> and "" as appropriate (Bruce)
9999 Enable Windows compilation of libpq
10000 Alpha spinlock fix from Uncle George <email>gatgul@voicenet.com</email>
10001 Overhaul of optimizer data structures (Tom)
10002 Fix to cygipc library (Yutaka Tanida)
10003 Allow pgsql to work on newer Cygwin snapshots (Dan)
10004 New catalog version number (Tom)
10006 Rename heap_replace to heap_update
10007 Update for QNX (Dr. Andreas Kardos)
10008 New platform-specific regression handling (Tom)
10009 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
10010 Included all yacc and lex files into the distribution (Peter E.)
10011 Remove lextest, no longer needed (Peter E)
10012 Fix for libpq and psql on Windows (Magnus)
10013 Internally change datetime and timespan into timestamp and interval (Thomas)
10014 Fix for plpgsql on BSD/OS
10015 Add SQL_ASCII test case to the regression test (Tatsuo)
10016 configure --with-mb now deprecated (Tatsuo)
10018 NetBSD fixes (Johnny C. Lam <email>lamj@stat.cmu.edu</email>)
10019 Fixes for Alpha compiles
10020 New multibyte encodings
10026 <sect1 id="release-6-5-3">
10027 <title>Release 6.5.3</title>
10030 <title>Release date</title>
10031 <simpara>1999-10-13</simpara>
10035 This is basically a cleanup release for 6.5.2. We have added a new
10036 <application>PgAccess</> that was missing in 6.5.2, and installed an NT-specific fix.
10041 <title>Migration to version 6.5.3</title>
10044 A dump/restore is <emphasis>not</emphasis> required for those running
10049 <title>Changes</title>
10053 Updated version of pgaccess 0.98
10055 Fix dumping rules on inherited tables
10062 <sect1 id="release-6-5-2">
10063 <title>Release 6.5.2</title>
10066 <title>Release date</title>
10067 <simpara>1999-09-15</simpara>
10071 This is basically a cleanup release for 6.5.1. We have fixed a variety of
10072 problems reported by 6.5.1 users.
10077 <title>Migration to version 6.5.2</title>
10080 A dump/restore is <emphasis>not</emphasis> required for those running
10086 <title>Changes</title>
10090 subselect+CASE fixes(Tom)
10091 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
10092 Fixes for CASE in WHERE join clauses(Tom)
10093 Fix BTScan abort(Tom)
10094 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
10095 Improve it so that it checks for multicolumn constraints(Thomas)
10096 Fix for Windows making problem with MB enabled(Hiroki Kataoka)
10097 Allow BSD yacc and bison to compile pl code(Bruce)
10098 Fix SET NAMES working
10100 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
10101 Reduce the total memory consumption of vacuum(Tom)
10102 Fix for timestamp(datetime)
10103 Rule deparsing bugfixes(Tom)
10104 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(Tom)
10105 This is to re-use space on index pages freed by vacuum(Vadim)
10106 document -x for pg_dump(Bruce)
10107 Fix for unary operators in rule deparser(Tom)
10108 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
10109 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
10110 Repair logic error in LIKE: should not return LIKE_ABORT
10111 when reach end of pattern before end of text(Tom)
10112 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom)
10113 Updated version of pgaccess 0.98
10119 <sect1 id="release-6-5-1">
10120 <title>Release 6.5.1</title>
10123 <title>Release date</title>
10124 <simpara>1999-07-15</simpara>
10128 This is basically a cleanup release for 6.5. We have fixed a variety of
10129 problems reported by 6.5 users.
10133 <title>Migration to version 6.5.1</title>
10136 A dump/restore is <emphasis>not</emphasis> required for those running
10142 <title>Changes</title>
10147 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
10148 Remove QUERY_LIMIT, use SELECT...LIMIT
10149 Fix for EXPLAIN on inheritance(Tom)
10150 Patch to allow vacuum on multisegment tables(Hiroshi)
10151 R-Tree optimizer selectivity fix(Tom)
10152 ACL file descriptor leak fix(Atsushi Ogawa)
10153 New expresssion subtree code(Tom)
10154 Avoid disk writes for read-only transactions(Vadim)
10155 Fix for removal of temp tables if last transaction was aborted(Bruce)
10156 Fix to prevent too large row from being created(Bruce)
10158 Allow port numbers 32k - 64k(Bruce)
10159 Add ^ precidence(Bruce)
10160 Rename sort files called pg_temp to pg_sorttemp(Bruce)
10161 Fix for microseconds in time values(Tom)
10162 Tutorial source cleanup
10163 New linux_m68k port
10164 Fix for sorting of NULL's in some cases(Tom)
10165 Shared library dependencies fixed (Tom)
10166 Fixed glitches affecting GROUP BY in subselects(Tom)
10167 Fix some compiler warnings (Tomoaki Nishiyama)
10168 Add Win1250 (Czech) support (Pavel Behal)
10174 <sect1 id="release-6-5">
10175 <title>Release 6.5</title>
10178 <title>Release date</title>
10179 <simpara>1999-06-09</simpara>
10183 This release marks a major step in the development team's mastery of the source
10184 code we inherited from Berkeley. You will see we are now easily adding
10185 major features, thanks to the increasing size and experience of our
10186 world-wide development team.
10190 Here is a brief summary of the more notable changes:
10195 Multiversion concurrency control(MVCC)
10199 This removes our old table-level locking, and replaces it with
10200 a locking system that is superior to most commercial database
10201 systems. In a traditional system, each row that is modified
10202 is locked until committed, preventing reads by other users.
10203 MVCC uses the natural multiversion nature of
10204 <productname>PostgreSQL</productname> to allow readers to
10205 continue reading consistent data during writer activity.
10206 Writers continue to use the compact pg_log transaction system.
10207 This is all performed without having to allocate a lock for
10208 every row like traditional database systems. So, basically,
10209 we no longer are restricted by simple table-level locking; we
10210 have something better than row-level locking.
10217 Hot backups from <application>pg_dump</application>
10221 <application>pg_dump</application> takes advantage of the new
10222 MVCC features to give a consistent database dump/backup while
10223 the database stays online and available for queries.
10234 We now have a true numeric data type, with
10235 user-specified precision.
10246 Temporary tables are guaranteed to have unique names
10247 within a database session, and are destroyed on session exit.
10258 We now have CASE, INTERSECT, and EXCEPT statement
10259 support. We have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL,
10260 SELECT ... FOR UPDATE, and an improved LOCK TABLE command.
10271 We continue to speed up <productname>PostgreSQL</productname>,
10272 thanks to the variety of talents within our team. We have
10273 sped up memory allocation, optimization, table joins, and row
10285 We continue to expand our port list, this time including
10286 <systemitem class="osname">Windows NT</>/<systemitem>ix86</> and <systemitem class="osname">NetBSD</>/<systemitem>arm32</>.
10297 Most interfaces have new versions, and existing functionality
10309 New and updated material is present throughout the
10310 documentation. New <acronym>FAQ</acronym>s have been
10311 contributed for <systemitem class="osname">SGI</> and <systemitem class="osname">AIX</> platforms.
10312 The <citetitle>Tutorial</citetitle> has introductory information
10313 on <acronym>SQL</acronym> from Stefan Simkovics.
10314 For the <citetitle>User's Guide</citetitle>, there are
10315 reference pages covering the postmaster and more utility
10316 programs, and a new appendix
10317 contains details on date/time behavior.
10318 The <citetitle>Administrator's Guide</citetitle> has a new
10319 chapter on troubleshooting from Tom Lane.
10320 And the <citetitle>Programmer's Guide</citetitle> has a
10321 description of query processing, also from Stefan, and details
10322 on obtaining the <productname>PostgreSQL</productname> source
10323 tree via anonymous <productname>CVS</productname> and
10324 <productname>CVSup</productname>.
10332 <title>Migration to version 6.5</title>
10335 A dump/restore using <application>pg_dump</application>
10336 is required for those wishing to migrate data from any
10337 previous release of <productname>PostgreSQL</productname>.
10338 <application>pg_upgrade</application> can <emphasis>not</emphasis>
10339 be used to upgrade to this release because the on-disk structure
10340 of the tables has changed compared to previous releases.
10344 The new Multiversion Concurrency Control (MVCC) features can
10345 give somewhat different behaviors in multiuser
10346 environments. <emphasis>Read and understand the following section
10347 to ensure that your existing applications will give you the
10348 behavior you need.</emphasis>
10352 <title>Multiversion Concurrency Control</title>
10355 Because readers in 6.5 don't lock data, regardless of transaction
10356 isolation level, data read by one transaction can be overwritten by
10357 another. In other words, if a row is returned by
10358 <command>SELECT</command> it doesn't mean that this row really exists
10359 at the time it is returned (i.e. sometime after the statement or
10360 transaction began) nor that the row is protected from being deleted or
10361 updated by concurrent transactions before the current transaction does
10362 a commit or rollback.
10366 To ensure the actual existence of a row and protect it against
10367 concurrent updates one must use <command>SELECT FOR UPDATE</command> or
10368 an appropriate <command>LOCK TABLE</command> statement. This should be
10369 taken into account when porting applications from previous releases of
10370 <productname>PostgreSQL</productname> and other environments.
10374 Keep the above in mind if you are using
10375 <filename>contrib/refint.*</filename> triggers for
10376 referential integrity. Additional techniques are required now. One way is
10377 to use <command>LOCK parent_table IN SHARE ROW EXCLUSIVE MODE</command>
10378 command if a transaction is going to update/delete a primary key and
10379 use <command>LOCK parent_table IN SHARE MODE</command> command if a
10380 transaction is going to update/insert a foreign key.
10384 Note that if you run a transaction in SERIALIZABLE mode then you must
10385 execute the <command>LOCK</command> commands above before execution of any
10386 <acronym>DML</acronym> statement
10387 (<command>SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO</command>) in the
10394 These inconveniences will disappear in the future
10395 when the ability to read dirty
10396 (uncommitted) data (regardless of isolation level) and true referential
10397 integrity will be implemented.
10403 <title>Changes</title>
10409 Fix text<->float8 and text<->float4 conversion functions(Thomas)
10410 Fix for creating tables with mixed-case constraints(Billy)
10411 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
10412 Fix bug in pg_dump -z
10413 Memory overrun cleanups(Tatsuo)
10414 Fix for lo_import crash(Tatsuo)
10415 Adjust handling of data type names to suppress double quotes(Thomas)
10416 Use type coercion for matching columns and DEFAULT(Thomas)
10417 Fix deadlock so it only checks once after one second of sleep(Bruce)
10418 Fixes for aggregates and PL/pgsql(Hiroshi)
10419 Fix for subquery crash(Vadim)
10420 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
10421 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo)
10422 Fix for pg_dump -d or -D and quote special characters in INSERT
10423 Repair serious problems with dynahash(Tom)
10424 Fix INET/CIDR portability problems
10425 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
10426 Fix executor so mergejoin of different column types works(Tom)
10427 Fix for Alpha OR selectivity bug
10428 Fix OR index selectivity problem(Bruce)
10429 Fix so \d shows proper length for char()/varchar()(Ryan)
10430 Fix tutorial code(Clark)
10431 Improve destroyuser checking(Oliver)
10432 Fix for Kerberos(Rodney McDuff)
10433 Fix for dropping database while dirty buffers(Bruce)
10434 Fix so sequence nextval() can be case-sensitive(Bruce)
10436 Drop buffers before destroying database files(Bruce)
10437 Fix case where executor evaluates functions twice(Tatsuo)
10438 Allow sequence nextval actions to be case-sensitive(Bruce)
10439 Fix optimizer indexing not working for negative numbers(Bruce)
10440 Fix for memory leak in executor with fjIsNull
10441 Fix for aggregate memory leaks(Erik Riedel)
10442 Allow user name containing a dash to grant privileges
10443 Cleanup of NULL in inet types
10444 Clean up system table bugs(Tom)
10445 Fix problems of PAGER and \? command(Masaaki Sakaida)
10446 Reduce default multisegment file size limit to 1GB(Peter)
10447 Fix for dumping of CREATE OPERATOR(Tom)
10448 Fix for backward scanning of cursors(Hiroshi Inoue)
10449 Fix for COPY FROM STDIN when using \i(Tom)
10450 Fix for subselect is compared inside an expression(Jan)
10451 Fix handling of error reporting while returning rows(Tom)
10452 Fix problems with reference to array types(Tom,Jan)
10453 Prevent UPDATE SET oid(Jan)
10454 Fix pg_dump so -t option can handle case-sensitive tablenames
10455 Fixes for GROUP BY in special cases(Tom, Jan)
10456 Fix for memory leak in failed queries(Tom)
10457 DEFAULT now supports mixed-case identifiers(Tom)
10458 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
10459 Disable use of pg_dump with both -o and -d options(Bruce)
10460 Allow pg_dump to properly dump group privileges(Bruce)
10461 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
10462 Fix for computations in views(Jan)
10463 Fix for aggregates on array indexes(Tom)
10464 Fix for DEFAULT handles single quotes in value requiring too many quotes
10465 Fix security problem with non-super users importing/exporting large objects(Tom)
10466 Rollback of transaction that creates table cleaned up properly(Tom)
10467 Fix to allow long table and column names to generate proper serial names(Tom)
10471 Add "vacuumdb" utility
10472 Speed up libpq by allocating memory better(Tom)
10473 EXPLAIN all indexes used(Tom)
10474 Implement CASE, COALESCE, NULLIF expression(Thomas)
10475 New pg_dump table output format(Constantin)
10476 Add string min()/max() functions(Thomas)
10477 Extend new type coercion techniques to aggregates(Thomas)
10478 New moddatetime contrib(Terry)
10479 Update to pgaccess 0.96(Constantin)
10480 Add routines for single-byte "char" type(Thomas)
10481 Improved substr() function(Thomas)
10482 Improved multibyte handling(Tatsuo)
10483 Multiversion concurrency control/MVCC(Vadim)
10484 New Serialized mode(Vadim)
10485 Fix for tables over 2gigs(Peter)
10486 New SET TRANSACTION ISOLATION LEVEL(Vadim)
10487 New LOCK TABLE IN ... MODE(Vadim)
10488 Update ODBC driver(Byron)
10489 New NUMERIC data type(Jan)
10490 New SELECT FOR UPDATE(Vadim)
10491 Handle "NaN" and "Infinity" for input values(Jan)
10492 Improved date/year handling(Thomas)
10493 Improved handling of backend connections(Magnus)
10494 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
10495 New TCL_ARRAYS option(Massimo)
10496 New INTERSECT and EXCEPT(Stefan)
10497 New pg_index.indisprimary for primary key tracking(D'Arcy)
10498 New pg_dump option to allow dropping of tables before creation(Brook)
10499 Speedup of row output routines(Tom)
10500 New READ COMMITTED isolation level(Vadim)
10501 New TEMP tables/indexes(Bruce)
10502 Prevent sorting if result is already sorted(Jan)
10503 New memory allocation optimization(Jan)
10504 Allow psql to do \p\g(Bruce)
10505 Allow multiple rule actions(Jan)
10506 Added LIMIT/OFFSET functionality(Jan)
10507 Improve optimizer when joining a large number of tables(Bruce)
10508 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
10509 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
10510 Improved int8 support(Ryan Bradetich, Thomas, Tom)
10511 New routines to convert between int8 and text/varchar types(Thomas)
10512 New bushy plans, where meta-tables are joined(Bruce)
10513 Enable right-hand queries by default(Bruce)
10514 Allow reliable maximum number of backends to be set at configure time
10515 (--with-maxbackends and postmaster switch (-N backends))(Tom)
10516 GEQO default now 10 tables because of optimizer speedups(Tom)
10517 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
10518 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
10519 Allow psql \d on a view show query(Ryan)
10520 Speedup for LIKE(Bruce)
10521 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
10522 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
10523 Make % operator have precedence like /(Bruce)
10524 Add new postgres -O option to allow system table structure changes(Bruce)
10525 Update contrib/pginterface/findoidjoins script(Tom)
10526 Major speedup in vacuum of deleted rows with indexes(Vadim)
10527 Allow non-SQL functions to run different versions based on arguments(Tom)
10528 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida)
10529 Add version number in start-up banners for psql(Masaaki Sakaida)
10530 New contrib/vacuumlo removes large objects not referenced(Peter)
10531 New initialization for table sizes so non-vacuumed tables perform better(Tom)
10532 Improve error messages when a connection is rejected(Tom)
10533 Support for arrays of char() and varchar() fields(Massimo)
10534 Overhaul of hash code to increase reliability and performance(Tom)
10535 Update to PyGreSQL 2.4(D'Arcy)
10536 Changed debug options so -d4 and -d5 produce different node displays(Jan)
10537 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
10538 Better optimization statistics for system table access(Tom)
10539 Better handling of non-default block sizes(Massimo)
10540 Improve GEQO optimizer memory consumption(Tom)
10541 UNION now suppports ORDER BY of columns not in target list(Jan)
10542 Major libpq++ improvements(Vince Vielhaber)
10543 pg_dump now uses -z(ACL's) as default(Bruce)
10544 backend cache, memory speedups(Tom)
10545 have pg_dump do everything in one snapshot transaction(Vadim)
10546 fix for large object memory leakage, fix for pg_dumping(Tom)
10547 INET type now respects netmask for comparisons
10548 Make VACUUM ANALYZE only use a readlock(Vadim)
10549 Allow VIEWs on UNIONS(Jan)
10550 pg_dump now can generate consistent snapshots on active databases(Vadim)
10552 Source Tree Changes
10553 -------------------
10554 Improve port matching(Tom)
10555 Portability fixes for SunOS
10556 Add Windows NT backend port and enable dynamic loading(Magnus and Daniel Horak)
10557 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
10558 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
10559 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
10560 Port to NetBSD/macppc(Toshimi Aoki)
10561 Fix for tcl/tk configuration(Vince)
10562 Removed CURRENT key word for rule queries(Jan)
10563 NT dynamic loading now works(Daniel Horak)
10564 Add ARM32 support(Andrew McMurry)
10565 Better support for HP-UX 11 and UnixWare
10566 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
10567 New install commands for plpgsql(Jan)
10574 <sect1 id="release-6-4-2">
10575 <title>Release 6.4.2</title>
10578 <title>Release date</title>
10579 <simpara>1998-12-20</simpara>
10583 The 6.4.1 release was improperly packaged. This also has one additional
10589 <title>Migration to version 6.4.2</title>
10592 A dump/restore is <emphasis>not</emphasis> required for those running
10597 <title>Changes</title>
10601 Fix for datetime constant problem on some platforms(Thomas)
10609 <sect1 id="release-6-4-1">
10610 <title>Release 6.4.1</title>
10613 <title>Release date</title>
10614 <simpara>1998-12-18</simpara>
10618 This is basically a cleanup release for 6.4. We have fixed a variety of
10619 problems reported by 6.4 users.
10624 <title>Migration to version 6.4.1</title>
10627 A dump/restore is <emphasis>not</emphasis> required for those running
10632 <title>Changes</title>
10636 Add pg_dump -N flag to force double quotes around identifiers. This is
10637 the default(Thomas)
10638 Fix for NOT in where clause causing crash(Bruce)
10639 EXPLAIN VERBOSE coredump fix(Vadim)
10640 Fix shared-library problems on Linux
10641 Fix test for table existence to allow mixed-case and whitespace in
10642 the table name(Thomas)
10643 Fix a couple of pg_dump bugs
10644 Configure matches template/.similar entries better(Tom)
10645 Change builtin function names from SPI_* to spi_*
10646 OR WHERE clause fix(Vadim)
10647 Fixes for mixed-case table names(Billy)
10648 contrib/linux/postgres.init.csh/sh fix(Thomas)
10649 libpq memory overrun fix
10651 Change exp() behavior to generate error on underflow(Thomas)
10652 pg_dump fixes for memory leak, inheritance constraints, layout change
10653 update pgaccess to 0.93
10654 Fix prototype for 64-bit platforms
10655 Multibyte fixes(Tatsuo)
10657 Fix memory overruns(Tatsuo)
10658 Fix for lo_import() crash(Bruce)
10659 Better search for install program(Tom)
10660 Timezone fixes(Tom)
10662 Use implicit type coercion for matching DEFAULT values(Thomas)
10663 Add routines to help with single-byte (internal) character type(Thomas)
10664 Compilation of libpq for Windows fixes(Magnus)
10665 Upgrade to PyGreSQL 2.2(D'Arcy)
10673 <sect1 id="release-6-4">
10674 <title>Release 6.4</title>
10677 <title>Release date</title>
10678 <simpara>1998-10-30</simpara>
10682 There are <emphasis>many</emphasis> new features and improvements in this release.
10683 Thanks to our developers and maintainers, nearly every aspect of the system
10684 has received some attention since the previous release.
10685 Here is a brief, incomplete summary:
10690 Views and rules are now functional thanks to extensive new code in the
10691 rewrite rules system from Jan Wieck. He also wrote a chapter on it
10692 for the <citetitle>Programmer's Guide</citetitle>.
10697 Jan also contributed a second procedural language, <application>PL/pgSQL</application>, to go with the
10698 original <application>PL/pgTCL</application> procedural language he contributed last release.
10704 We have optional multiple-byte character set support from Tatsuo Ishii
10705 to complement our existing locale support.
10711 Client/server communications has been cleaned up, with better support for
10712 asynchronous messages and interrupts thanks to Tom Lane.
10718 The parser will now perform automatic type coercion to match arguments
10719 to available operators and functions, and to match columns and expressions
10720 with target columns. This uses a generic mechanism which supports
10721 the type extensibility features of <productname>PostgreSQL</productname>.
10722 There is a new chapter in the <citetitle>User's Guide</citetitle>
10723 which covers this topic.
10729 Three new data types have been added.
10730 Two types, <type>inet</type> and <type>cidr</type>, support various forms
10731 of IP network, subnet, and machine addressing. There is now an 8-byte integer
10732 type available on some platforms. See the chapter on data types
10733 in the <citetitle>User's Guide</citetitle> for details.
10734 A fourth type, <type>serial</type>, is now supported by the parser as an
10735 amalgam of the <type>int4</type> type, a sequence, and a unique index.
10741 Several more <acronym>SQL92</acronym>-compatible syntax features have been
10742 added, including <command>INSERT DEFAULT VALUES</command>
10748 The automatic configuration and installation system has received some
10749 attention, and should be more robust for more platforms than it has ever
10758 <title>Migration to version 6.4</title>
10761 A dump/restore using <application>pg_dump</application>
10762 or <application>pg_dumpall</application>
10763 is required for those wishing to migrate data from any
10764 previous release of <productname>PostgreSQL</productname>.
10769 <title>Changes</title>
10775 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
10776 Remove char2-16 data types, use char/varchar(Darren)
10777 Pqfn not handles a NOTICE message(Anders)
10778 Reduced busywaiting overhead for spinlocks with many backends (dg)
10779 Stuck spinlock detection (dg)
10780 Fix up "ISO-style" timespan decoding and encoding(Thomas)
10781 Fix problem with table drop after rollback of transaction(Vadim)
10782 Change error message and remove non-functional update message(Vadim)
10783 Fix for COPY array checking
10784 Fix for SELECT 1 UNION SELECT NULL
10785 Fix for buffer leaks in large object calls(Pascal)
10786 Change owner from oid to int4 type(Bruce)
10787 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
10788 Fix for shared invalidation cache overflow(Massimo)
10789 Prevent file descriptor leaks in failed COPY's(Bruce)
10790 Fix memory leak in libpgtcl's pg_select(Constantin)
10791 Fix problems with username/passwords over 8 characters(Tom)
10792 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
10793 Fix of many bad system table entries(Tom)
10797 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
10798 Show the index used in an EXPLAIN(Zeugswetter)
10799 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
10800 Multibyte awareness of many data types and functions, via configure(Tatsuo)
10801 New configure --with-mb option(Tatsuo)
10802 New initdb --pgencoding option(Tatsuo)
10803 New createdb -E multibyte option(Tatsuo)
10804 Select version(); now returns PostgreSQL version(Jeroen)
10805 libpq now allows asynchronous clients(Tom)
10806 Allow cancel from client of backend query(Tom)
10807 psql now cancels query with Control-C(Tom)
10808 libpq users need not issue dummy queries to get NOTIFY messages(Tom)
10809 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
10810 PGresult struct now includes associated error message, if any(Tom)
10811 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
10812 Add routines to convert between varchar and bpchar(Thomas)
10813 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
10814 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
10815 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas)
10816 Fixes for unary minus parsing with leading spaces(Thomas)
10817 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
10818 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
10819 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
10820 Enable HAVING clause but no fixes elsewhere yet.
10821 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thomas)
10822 Save string type if specified for DEFAULT clause handling(Thomas)
10823 Coerce operations involving different data types(Thomas)
10824 Allow some index use for columns of different types(Thomas)
10825 Add capabilities for automatic type conversion(Thomas)
10826 Cleanups for large objects, so file is truncated on open(Peter)
10827 Readline cleanups(Tom)
10828 Allow psql \f \ to make spaces as delimiter(Bruce)
10829 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
10830 Msql compatibility library in /contrib(Aldrin)
10831 Remove the requirement that ORDER/GROUP BY clause identifiers be
10832 included in the target list(David)
10833 Convert columns to match columns in UNION clauses(Thomas)
10834 Remove fork()/exec() and only do fork()(Bruce)
10835 Jdbc cleanups(Peter)
10836 Show backend status on ps command line(only works on some platforms)(Bruce)
10837 Pg_hba.conf now has a sameuser option in the database field
10838 Make lo_unlink take oid param, not int4
10839 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
10840 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
10841 libpgtcl cleanups(Tom)
10842 Add -error option to libpgtcl's pg_result command(Tom)
10843 New locale patch, see docs/README/locale(Oleg)
10844 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
10845 New contrib/lo code for large object orphan removal(Peter)
10846 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
10847 feature, see /doc/README.mb(Tatsuo)
10848 contrib/noupdate code to revoke update permission on a column
10849 libpq can now be compiled on Windows(Magnus)
10850 Add PQsetdbLogin() in libpq
10851 New 8-byte integer type, checked by configure for OS support(Thomas)
10852 Better support for quoted table/column names(Thomas)
10853 Surround table and column names with double-quotes in pg_dump(Thomas)
10854 PQreset() now works with passwords(Tom)
10855 Handle case of GROUP BY target list column number out of range(David)
10856 Allow UNION in subselects
10857 Add auto-size to screen to \d? commands(Bruce)
10858 Use UNION to show all \d? results in one query(Bruce)
10859 Add \d? field search feature(Bruce)
10860 Pg_dump issues fewer \connect requests(Tom)
10861 Make pg_dump -z flag work better, document it in manual page(Tom)
10862 Add HAVING clause with full support for subselects and unions(Stephan)
10863 Full text indexing routines in contrib/fulltextindex(Maarten)
10864 Transaction ids now stored in shared memory(Vadim)
10865 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
10866 Support for SQL92 syntax "SET NAMES"(Tatsuo)
10867 Support for LATIN2-5(Tatsuo)
10868 Add UNICODE regression test case(Tatsuo)
10869 Lock manager cleanup, new locking modes for LLL(Vadim)
10870 Allow index use with OR clauses(Bruce)
10871 Allows "SELECT NULL ORDER BY 1;"
10872 Explain VERBOSE prints the plan, and now pretty-prints the plan to
10873 the postmaster log file(Bruce)
10874 Add indexes display to \d command(Bruce)
10875 Allow GROUP BY on functions(David)
10876 New pg_class.relkind for large objects(Bruce)
10877 New way to send libpq NOTICE messages to a different location(Tom)
10878 New \w write command to psql(Bruce)
10879 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
10880 Allow binary-compatible indexes to be considered when checking for valid
10881 Indexes for restriction clauses containing a constant(Thomas)
10882 New ISBN/ISSN code in /contrib/isbn_issn
10883 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
10884 New rewrite system fixes many problems with rules and views(Jan)
10885 * Rules on relations work
10886 * Event qualifications on insert/update/delete work
10887 * New OLD variable to reference CURRENT, CURRENT will be remove in future
10888 * Update rules can reference NEW and OLD in rule qualifications/actions
10889 * Insert/update/delete rules on views work
10890 * Multiple rule actions are now supported, surrounded by parentheses
10891 * Regular users can create views/rules on tables they have RULE permits
10892 * Rules and views inherit the privileges of the creator
10893 * No rules at the column level
10894 * No UPDATE NEW/OLD rules
10895 * New pg_tables, pg_indexes, pg_rules and pg_views system views
10896 * Only a single action on SELECT rules
10897 * Total rewrite overhaul, perhaps for 6.5
10898 * handle subselects
10899 * handle aggregates on views
10900 * handle insert into select from view works
10901 System indexes are now multikey(Bruce)
10902 Oidint2, oidint4, and oidname types are removed(Bruce)
10903 Use system cache for more system table lookups(Bruce)
10904 New backend programming language PL/pgSQL in backend/pl(Jan)
10905 New SERIAL data type, auto-creates sequence/index(Thomas)
10906 Enable assert checking without a recompile(Massimo)
10907 User lock enhancements(Massimo)
10908 New setval() command to set sequence value(Massimo)
10909 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
10910 Conditional trace package(Massimo)
10911 New UNLISTEN command(Massimo)
10912 psql and libpq now compile under Windows using win32.mak(Magnus)
10913 Lo_read no longer stores trailing NULL(Bruce)
10914 Identifiers are now truncated to 31 characters internally(Bruce)
10915 Createuser options now availble on the command line
10916 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
10917 Prevent file descriptor leaf from failed COPY(Bruce)
10918 New pg_upgrade command(Bruce)
10919 Updated /contrib directories(Massimo)
10920 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
10921 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
10922 New DECLARE and FETCH feature(Thomas)
10923 libpq's internal structures now not exported(Tom)
10924 Allow up to 8 key indexes(Bruce)
10925 Remove ARCHIVE key word, that is no longer used(Thomas)
10926 pg_dump -n flag to supress quotes around indentifiers
10927 disable system columns for views(Jan)
10928 new INET and CIDR types for network addresses(TomH, Paul)
10929 no more double quotes in psql output
10930 pg_dump now dumps views(Terry)
10931 new SET QUERY_LIMIT(Tatsuo,Jan)
10933 Source Tree Changes
10934 -------------------
10935 /contrib cleanup(Jun)
10936 Inline some small functions called for every row(Bruce)
10938 HP-UX cleanups(Tom)
10939 Multibyte regression tests(Soonmyung.)
10940 Remove --disabled options from configure
10941 Define PGDOC to use POSTGRESDIR by default
10942 Make regression optional
10943 Remove extra braces code to pgindent(Bruce)
10944 Add bsdi shared library support(Bruce)
10945 New --without-CXX support configure option(Brook)
10947 Update backend flowchart in tools/backend(Bruce)
10948 Change atttypmod from int16 to int32(Bruce, Tom)
10949 Getrusage() fix for platforms that do not have it(Tom)
10950 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
10951 NS32K platform fixes(Phil Nelson, John Buller)
10952 SCO 7/UnixWare 2.x fixes(Billy,others)
10953 Sparc/Solaris 2.5 fixes(Ryan)
10954 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
10955 Even more documention(Thomas)
10956 Nextstep support(Jacek)
10958 pginterface manual page(Bruce)
10959 shared libraries all have version numbers
10960 merged all OS-specific shared library defines into one file
10961 smarter TCL/TK configuration checking(Billy)
10962 smarter perl configuration(Brook)
10963 configure uses supplied install-sh if no install script found(Tom)
10964 new Makefile.shlib for shared library configuration(Tom)
10970 <sect1 id="release-6-3-2">
10971 <title>Release 6.3.2</title>
10974 <title>Release date</title>
10975 <simpara>1998-04-07</simpara>
10979 This is a bug-fix release for 6.3.x.
10980 Refer to the release notes for version 6.3 for a more complete summary of new features.
10988 Repairs automatic configuration support for some platforms, including Linux,
10989 from breakage inadvertently introduced in version 6.3.1.
10995 Correctly handles function calls on the left side of BETWEEN and LIKE clauses.
11002 A dump/restore is NOT required for those running 6.3 or 6.3.1. A
11003 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11004 This last step should be performed while the postmaster is not running.
11005 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11008 For upgrades from pre-6.3 installations,
11009 refer to the installation and migration instructions for version 6.3.
11013 <title>Changes</title>
11017 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
11018 Manual page improvements(Bruce)
11019 BETWEEN and LIKE fix(Thomas)
11020 fix for psql \connect used by pg_dump(Oliver Elphick)
11022 pgaccess, version 0.86
11023 qsort removed, now uses libc version, cleanups(Jeroen)
11024 fix for buffer over-runs detected(Maurice Gittens)
11025 fix for buffer overrun in libpgtcl(Randy Kunkee)
11026 fix for UNION with DISTINCT or ORDER BY(Bruce)
11027 gettimeofday configure check(Doug Winterburn)
11028 Fix "indexes not used" bug(Vadim)
11029 docs additions(Thomas)
11030 Fix for backend memory leak(Bruce)
11031 libreadline cleanup(Erwan MAS)
11032 Remove DISTDIR(Bruce)
11033 Makefile dependency cleanup(Jeroen van Vianen)
11034 ASSERT fixes(Bruce)
11040 <sect1 id="release-6-3-1">
11041 <title>Release 6.3.1</title>
11044 <title>Release date</title>
11045 <simpara>1998-03-23</simpara>
11054 Additional support for multibyte character sets.
11060 Repair byte ordering for mixed-endian clients and servers.
11066 Minor updates to allowed SQL syntax.
11072 Improvements to the configuration autodetection for installation.
11079 A dump/restore is NOT required for those running 6.3. A
11080 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11081 This last step should be performed while the postmaster is not running.
11082 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11085 For upgrades from pre-6.3 installations,
11086 refer to the installation and migration instructions for version 6.3.
11090 <title>Changes</title>
11094 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
11095 pg_user cleanup(Bruce)
11096 large object fix for pg_dump and tclsh (alvin)
11097 LIKE fix for multiple adjacent underscores
11098 fix for redefining builtin functions(Thomas)
11100 upgrade to pg_access 0.83
11101 updated CLUSTER manual page
11102 multibyte character set support, see doc/README.mb(Tatsuo)
11103 configure --with-pgport fix
11105 big-endian fix for backend communications(Kataoka)
11106 SUBSTR() and substring() fix(Jan)
11107 several jdbc fixes(Peter)
11108 libpgtcl improvements, see libptcl/README(Randy Kunkee)
11109 Fix for "Datasize = 0" error(Vadim)
11110 Prevent \do from wrapping(Bruce)
11111 Remove duplicate Russian character set entries
11113 Allow optional TABLE key word in LOCK and SELECT INTO(Thomas)
11114 CREATE SEQUENCE options to allow a negative integer(Thomas)
11115 Add "PASSWORD" as an allowed column identifier(Thomas)
11116 Add checks for UNION target fields(Bruce)
11117 Fix Alpha port(Dwayne Bailey)
11118 Fix for text arrays containing quotes(Doug Gibson)
11119 Solaris compile fix(Albert Chin-A-Young)
11120 Better identify tcl and tk libs and includes(Bruce)
11126 <sect1 id="release-6-3">
11127 <title>Release 6.3</title>
11130 <title>Release date</title>
11131 <simpara>1998-03-01</simpara>
11135 There are <emphasis>many</emphasis> new features and improvements in this release.
11136 Here is a brief, incomplete summary:
11141 Many new SQL features, including
11142 full <acronym>SQL92</acronym> subselect capability
11143 (everything is here but target-list subselects).
11149 Support for client-side environment variables to specify time zone and date style.
11155 Socket interface for client/server connection. This is the default now
11156 so you may need to start <application>postmaster</application> with the
11157 <option>-i</option> flag.
11163 Better password authorization mechanisms. Default table privileges have changed.
11169 Old-style <firstterm>time travel</firstterm>
11170 has been removed. Performance has been improved.
11178 Bruce Momjian wrote the following notes to introduce the new release.
11183 There are some general 6.3 issues that I want to mention. These are
11184 only the big items that cannot be described in one sentence. A review
11185 of the detailed changes list is still needed.
11188 First, we now have subselects. Now that we have them, I would like to
11189 mention that without subselects, SQL is a very limited language.
11190 Subselects are a major feature, and you should review your code for
11191 places where subselects provide a better solution for your queries. I
11192 think you will find that there are more uses for subselects than you may
11193 think. Vadim has put us on the big SQL map with subselects, and fully
11194 functional ones too. The only thing you can't do with subselects is to
11195 use them in the target list.
11198 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
11199 enable connections from other machines, you have to use the new
11200 postmaster -i option, and of course edit <filename>pg_hba.conf</filename>. Also, for this
11201 reason, the format of <filename>pg_hba.conf</filename> has changed.
11204 Third, <type>char()</type> fields will now allow faster access than <type>varchar()</type> or
11205 <type>text</type>. Specifically, the <type>text</> and <type>varchar()</type> have a penalty for access to
11206 any columns after the first column of this type. <type>char()</type> used to also
11207 have this access penalty, but it no longer does. This may suggest that
11208 you redesign some of your tables, especially if you have short character
11209 columns that you have defined as <type>varchar()</type> or <type>text</type>. This and other
11210 changes make 6.3 even faster than earlier releases.
11213 We now have passwords definable independent of any Unix file. There are
11214 new SQL USER commands.
11215 See the <citetitle>Administrator's Guide</citetitle> for more
11216 information. There is a new table, pg_shadow, which is used to store
11217 user information and user passwords, and it by default only SELECT-able
11218 by the <systemitem>postgres</systemitem> super-user. pg_user is now a view of pg_shadow, and is
11219 SELECT-able by PUBLIC. You should keep using pg_user in your
11220 application without changes.
11223 User-created tables now no longer have SELECT privilege to PUBLIC by
11224 default. This was done because the ANSI standard requires it. You can
11225 of course GRANT any privileges you want after the table is created.
11226 System tables continue to be SELECT-able by PUBLIC.
11229 We also have real deadlock detection code. No more sixty-second
11230 timeouts. And the new locking code implements a <acronym>FIFO</acronym> better, so there
11231 should be less resource starvation during heavy use.
11234 Many complaints have been made about inadequate documentation in previous
11235 releases. Thomas has put much effort into many new manuals for this
11236 release. Check out the doc/ directory.
11239 For performance reasons, time travel is gone, but can be implemented
11240 using triggers (see <filename>pgsql/contrib/spi/README</filename>). Please check out the new
11241 \d command for types, operators, etc. Also, views have their own
11242 privileges now, not based on the underlying tables, so privileges on
11243 them have to be set separately. Check <filename>/pgsql/interfaces</filename> for some new
11244 ways to talk to <productname>PostgreSQL</productname>.
11247 This is the first release that really required an explanation for
11248 existing users. In many ways, this was necessary because the new
11249 release removes many limitations, and the work-arounds people were using
11250 are no longer needed.
11254 <title>Migration to version 6.3</title>
11257 A dump/restore using <application>pg_dump</application>
11258 or <application>pg_dumpall</application>
11259 is required for those wishing to migrate data from any
11260 previous release of <productname>PostgreSQL</productname>.
11265 <title>Changes</title>
11271 Fix binary cursors broken by MOVE implementation(Vadim)
11272 Fix for tcl library crash(Jan)
11273 Fix for array handling, from Gerhard Hintermayer
11274 Fix acl error, and remove duplicate pqtrace(Bruce)
11275 Fix psql \e for empty file(Bruce)
11276 Fix for textcat on varchar() fields(Bruce)
11277 Fix for DBT Sendproc (Zeugswetter Andres)
11278 Fix vacuum analyze syntax problem(Bruce)
11279 Fix for international identifiers(Tatsuo)
11280 Fix aggregates on inherited tables(Bruce)
11281 Fix substr() for out-of-bounds data
11282 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
11283 Fix notty output to show status result. -q option still turns it off(Bruce)
11284 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
11286 Fix for PQtrace start/stop several times(Bruce)
11287 Fix a variety of locking problems like newer lock waiters getting
11288 lock before older waiters, and having readlock people not share
11289 locks if a writer is waiting for a lock, and waiting writers not
11290 getting priority over waiting readers(Bruce)
11291 Fix crashes in psql when executing queries from external files(James)
11292 Fix problem with multiple order by columns, with the first one having
11293 NULL values(Jeroen)
11294 Use correct hash table support functions for float8 and int4(Thomas)
11295 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
11296 Change precedence for boolean operators to match expected behavior(Thomas)
11297 Generate elog(ERROR) on over-large integer(Bruce)
11298 Allow multiple-argument functions in constraint clauses(Thomas)
11299 Check boolean input literals for 'true','false','yes','no','1','0'
11300 and throw elog(ERROR) if unrecognized(Thomas)
11301 Major large objects fix
11302 Fix for GROUP BY showing duplicates(Vadim)
11303 Fix for index scans in MergeJion(Vadim)
11307 Subselects with EXISTS, IN, ALL, ANY key words (Vadim, Bruce, Thomas)
11308 New User Manual(Thomas, others)
11309 Speedup by inlining some frequently-called functions
11310 Real deadlock detection, no more timeouts(Bruce)
11311 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
11312 CURRENT_USER(Thomas)
11313 Modify constraint syntax to be SQL92-compliant(Thomas)
11314 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
11315 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
11316 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas)
11317 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
11318 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
11319 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
11320 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
11321 Allow SQL92 delimited identifiers(Thomas)
11322 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas)
11323 Support SQL92 syntax for type coercion of literal strings
11324 (e.g. "DATETIME 'now'")(Thomas)
11325 Add conversions for int2, int4, and OID types to and from text(Thomas)
11326 Use shared lock when building indexes(Vadim)
11327 Free memory allocated for an user query inside transaction block after
11328 this query is done, was turned off in <= 6.2.1(Vadim)
11329 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
11330 New <productname>PostgreSQL</productname> Procedural Language (PL) backend interface(Jan)
11331 Rename pg_dump -H option to -h(Bruce)
11332 Add Java support for passwords, European dates(Peter)
11333 Use indexes for LIKE and ~, !~ operations(Bruce)
11334 Add hash functions for datetime and timespan(Thomas)
11335 Time Travel removed(Vadim, Bruce)
11336 Add paging for \d and \z, and fix \i(Bruce)
11337 Add Unix domain socket support to backend and to frontend library(Goran)
11338 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
11339 Allow more SQL92 and/or <productname>PostgreSQL</productname> reserved words as column identifiers(Thomas)
11340 Augment support for SQL92 SET TIME ZONE...(Thomas)
11341 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
11342 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
11343 Enable SET TIME ZONE using TZ environment variable(Thomas)
11344 Add PGDATESTYLE environment variable to frontend and backend initialization(Thomas)
11345 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
11346 frontend library initialization environment variables(Thomas)
11347 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
11348 Add pg_description table for info on tables, columns, operators, types, and
11350 Increase 16 char limit on system table/index names to 32 characters(Bruce)
11351 Rename system indexes(Bruce)
11352 Add 'GERMAN' option to SET DATESTYLE(Thomas)
11353 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
11354 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
11355 Validate numeric input more carefully for delta times(Thomas)
11356 Implement day of year as possible input to date_part()(Thomas)
11357 Define timespan_finite() and text_timespan() functions(Thomas)
11358 Remove archive stuff(Bruce)
11359 Allow for a pg_password authentication database that is separate from
11360 the system password file(Todd)
11361 Dump ACLs, GRANT, REVOKE privileges(Matt)
11362 Define text, varchar, and bpchar string length functions(Thomas)
11363 Fix Query handling for inheritance, and cost computations(Bruce)
11364 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
11365 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
11366 Implement UNIONs for SELECT(Bruce)
11367 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
11368 varchar() stores only necessary bytes on disk(Bruce)
11369 Fix for BLOBs(Peter)
11370 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
11371 Remove unused "option" from PQconnectdb()
11372 New LOCK command and lock manual page describing deadlocks(Bruce)
11373 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
11374 Enhance psql \z to show sequences(Bruce)
11375 Show NOT NULL and DEFAULT in psql \d table(Bruce)
11376 New psql .psqlrc file start-up(Andrew)
11377 Modify sample start-up script in contrib/linux to show syslog(Thomas)
11378 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
11379 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
11380 Update of contrib stuff(Massimo)
11381 Add Unix socket support to DBD::Pg(Goran)
11382 New python interface (PyGreSQL 2.0)(D'Arcy)
11383 New frontend/backend protocol has a version number, network byte order(Phil)
11384 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
11385 CHAR() now faster access than VARCHAR() or TEXT
11386 ecpg embedded SQL preprocessor
11387 Reduce system column overhead(Vadmin)
11388 Remove pg_time table(Vadim)
11389 Add pg_type attribute to identify types that need length (bpchar, varchar)
11390 Add report of offending line when COPY command fails
11391 Allow VIEW privileges to be set separately from the underlying tables.
11392 For security, use GRANT/REVOKE on views as appropriate(Jan)
11393 Tables now have no default GRANT SELECT TO PUBLIC. You must
11394 explicitly grant such privileges.
11395 Clean up tutorial examples(Darren)
11397 Source Tree Changes
11398 -------------------
11399 Add new html development tools, and flow chart in /tools/backend
11400 Fix for SCO compiles
11401 Stratus computer port Robert Gillies
11402 Added support for shlib for BSD44_derived & i386_solaris
11403 Make configure more automated(Brook)
11404 Add script to check regression test results
11405 Break parser functions into smaller files, group together(Bruce)
11406 Rename heap_create to heap_create_and_catalog, rename heap_creatr
11407 to heap_create()(Bruce)
11408 Sparc/Linux patch for locking(TomS)
11409 Remove PORTNAME and reorganize port-specific stuff(Marc)
11410 Add optimizer README file(Bruce)
11411 Remove some recursion in optimizer and clean up some code there(Bruce)
11412 Fix for NetBSD locking(Henry)
11413 Fix for libptcl make(Tatsuo)
11415 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
11416 function calls to istrue() or isfalse() to allow optimization(Thomas)
11417 Various fixes NetBSD/Sparc related(TomH)
11418 Alpha linux locking(Travis,Ryan)
11419 Change elog(WARN) to elog(ERROR)(Bruce)
11420 FAQ for FreeBSD(Marc)
11421 Bring in the PostODBC source tree as part of our standard distribution(Marc)
11422 A minor patch for HP/UX 10 vs 9(Stan)
11423 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
11424 UnixWare patches(Billy)
11425 New i386 'lock' for spinlock asm(Billy)
11426 Support for multiplexed backends is removed
11427 Start an OpenBSD port
11429 Start a Cygnus port
11430 Add string functions to regression suite(Thomas)
11431 Expand a few function names formerly truncated to 16 characters(Thomas)
11432 Remove un-needed malloc() calls and replace with palloc()(Bruce)
11438 <sect1 id="release-6-2-1">
11439 <title>Release 6.2.1</title>
11442 <title>Release date</title>
11443 <simpara>1997-10-17</simpara>
11447 6.2.1 is a bug-fix and usability release on 6.2.
11455 Allow strings to span lines, per <acronym>SQL92</acronym>.
11461 Include example trigger function for inserting user names on table updates.
11468 This is a minor bug-fix release on 6.2.
11469 For upgrades from pre-6.2 systems, a full dump/reload is required.
11470 Refer to the 6.2 release notes for instructions.
11474 <title>Migration from version 6.2 to version 6.2.1</title>
11477 This is a minor bug-fix release. A dump/reload is not required from version 6.2,
11478 but is required from any release prior to 6.2.
11481 In upgrading from version 6.2, if you choose to dump/reload you will find that
11482 avg(money) is now calculated correctly. All other bug fixes take effect
11483 upon updating the executables.
11486 Another way to avoid dump/reload is to use the following SQL command
11487 from <command>psql</command> to update the existing system table:
11490 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
11491 where aggname = 'avg' and aggbasetype = 790;
11495 This will need to be done to every existing database, including template1.
11500 <title>Changes</title>
11504 Allow TIME and TYPE column names(Thomas)
11505 Allow larger range of true/false as boolean values(Thomas)
11506 Support output of "now" and "current"(Thomas)
11507 Handle DEFAULT with INSERT of NULL properly(Vadim)
11508 Fix for relation reference counts problem in buffer manager(Vadim)
11509 Allow strings to span lines, like ANSI(Thomas)
11510 Fix for backward cursor with ORDER BY(Vadim)
11511 Fix avg(cash) computation(Thomas)
11512 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
11513 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
11514 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
11520 <sect1 id="release-6-2">
11521 <title>Release 6.2</title>
11524 <title>Release date</title>
11525 <simpara>1997-10-02</simpara>
11529 A dump/restore is required for those wishing to migrate data from
11530 previous releases of <productname>PostgreSQL</productname>.
11534 <title>Migration from version 6.1 to version 6.2</title>
11537 This migration requires a complete dump of the 6.1 database and a
11538 restore of the database in 6.2.
11541 Note that the <command>pg_dump</command> and <command>pg_dumpall</command> utility from 6.2 should be used
11542 to dump the 6.1 database.
11547 <title>Migration from version 1.<replaceable>x</> to version 6.2</title>
11550 Those migrating from earlier 1.* releases should first upgrade to 1.09
11551 because the COPY output format was improved from the 1.02 release.
11556 <title>Changes</title>
11562 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
11563 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
11564 from Solaris(Diab Jerius)
11565 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
11566 Check for geometric intersections at endpoints to avoid rounding ugliness(Thomas)
11567 Catch non-functional delete attempts(Vadim)
11568 Change time function names to be more consistent(Michael Reifenberg)
11569 Check for zero divides(Michael Reifenberg)
11570 Fix very old bug which made rows changed/inserted by a command
11571 visible to the command itself (so we had multiple update of
11572 updated rows, etc.)(Vadim)
11573 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
11574 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
11575 Remove un-needed signal stuff from contrib/pginterface
11576 Fix OR (where x != 1 or x isnull didn't return rows with x NULL) (Vadim)
11577 Fix time_cmp function (Vadim)
11578 Fix handling of functions with non-attribute first argument in
11579 WHERE clauses (Vadim)
11580 Fix GROUP BY when order of entries is different from order
11581 in target list (Vadim)
11582 Fix pg_dump for aggregates without sfunc1 (Vadim)
11586 Default genetic optimizer GEQO parameter is now 8(Bruce)
11587 Allow use parameters in target list having aggregates in functions(Vadim)
11588 Added JDBC driver as an interface(Adrian & Peter)
11589 pg_password utility
11590 Return number of rows inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
11591 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
11592 SPI (Server Programming Interface) allows execution of queries inside
11593 C-functions (Vadim)
11594 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
11595 Include reserved words for string handling, outer joins, and unions(Thomas)
11596 Implement extended comments ("/* ... */") using exclusive states(Thomas)
11597 Add "//" single-line comments(Bruce)
11598 Remove some restrictions on characters in operator names(Thomas)
11599 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
11600 Add text concatenation operator and function (SQL92)(Thomas)
11601 Support WITH TIME ZONE syntax (SQL92)(Thomas)
11602 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
11603 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
11604 and CHARACTER VARYING (SQL92)(Thomas)
11605 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
11606 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
11607 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
11608 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
11609 Add more reserved words, mostly for SQL92 compliance(Thomas)
11610 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
11611 Add center() routines for lseg, path, polygon(Thomas)
11612 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
11613 Check explicitly for points and polygons contained within polygons
11614 using an axis-crossing algorithm(Thomas)
11615 Add routine to convert circle-box(Thomas)
11616 Merge conflicting operators for different geometric data types(Thomas)
11617 Replace distance operator "<===>" with "<->"(Thomas)
11618 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Thomas)
11619 Add routines for text trimming on both ends, substring, and string position(Thomas)
11620 Added conversion routines circle(box) and poly(circle)(Thomas)
11621 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
11622 Allow functions and operators on internally-identical types to succeed(Bruce)
11623 Speed up backend start-up after profiling analysis(Bruce)
11624 Inline frequently called functions for performance(Bruce)
11625 Reduce open() calls(Bruce)
11626 psql: Add PAGER for \h and \?,\C fix
11627 Fix for psql pager when no tty(Bruce)
11628 New entab utility(Bruce)
11629 General trigger functions for referential integrity (Vadim)
11630 General trigger functions for time travel (Vadim)
11631 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
11632 MOVE implementation (Vadim)
11634 Source Tree Changes
11635 -------------------
11636 HP-UX 10 patches (Vladimir Turin)
11637 Added SCO support, (Daniel Harris)
11638 MkLinux patches (Tatsuo Ishii)
11639 Change geometric box terminology from "length" to "width"(Thomas)
11640 Deprecate temporary unstored slope fields in geometric code(Thomas)
11641 Remove restart instructions from INSTALL(Bruce)
11642 Look in /usr/ucb first for install(Bruce)
11643 Fix c++ copy example code(Thomas)
11644 Add -o to psql manual page(Bruce)
11645 Prevent relname unallocated string length from being copied into database(Bruce)
11646 Cleanup for NAMEDATALEN use(Bruce)
11647 Fix pg_proc names over 15 chars in output(Bruce)
11648 Add strNcpy() function(Bruce)
11649 remove some (void) casts that are unnecessary(Bruce)
11650 new interfaces directory(Marc)
11651 Replace fopen() calls with calls to fd.c functions(Bruce)
11652 Make functions static where possible(Bruce)
11653 enclose unused functions in #ifdef NOT_USED(Bruce)
11654 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
11655 Changes for Digital Unix
11656 Portability fix for pg_dumpall(Bruce)
11657 Rename pg_attribute.attnvals to attdispersion(Bruce)
11658 "intro/unix" manual page now "pgintro"(Bruce)
11659 "built-in" manual page now "pgbuiltin"(Bruce)
11660 "drop" manual page now "drop_table"(Bruce)
11661 Add "create_trigger", "drop_trigger" manual pages(Thomas)
11662 Add constraints regression test(Vadim & Thomas)
11663 Add comments syntax regression test(Thomas)
11664 Add PGINDENT and support program(Bruce)
11665 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
11666 Files moved to /src/tools directory(Bruce)
11667 SPI and Trigger programming guides (Vadim & D'Arcy)
11673 <sect1 id="release-6-1-1">
11674 <title>Release 6.1.1</title>
11677 <title>Release date</title>
11678 <simpara>1997-07-22</simpara>
11682 <title>Migration from version 6.1 to version 6.1.1</title>
11685 This is a minor bug-fix release. A dump/reload is not required from version 6.1,
11686 but is required from any release prior to 6.1.
11687 Refer to the release notes for 6.1 for more details.
11692 <title>Changes</title>
11696 fix for SET with options (Thomas)
11697 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
11698 new psql \connect option allows changing usernames without changing databases
11699 fix for initdb --debug option(Yoshihiko Ichikawa))
11700 lextest cleanup(Bruce)
11702 fix date/time month boundary arithmetic(Thomas)
11703 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
11704 timestamp overhauled to use standard functions(Thomas)
11705 other code cleanup in date/time routines(Thomas)
11706 psql's \d now case-insensitive(Bruce)
11707 psql's backslash commands can now have trailing semicolon(Bruce)
11708 fix memory leak in psql when using \g(Bruce)
11709 major fix for endian handling of communication to server(Thomas, Tatsuo)
11710 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
11711 allow underscores in usernames(Bruce)
11712 pg_dumpall now returns proper status, portability fix(Bruce)
11718 <sect1 id="release-6-1">
11719 <title>Release 6.1</title>
11722 <title>Release date</title>
11723 <simpara>1997-06-08</simpara>
11727 The regression tests have been adapted and extensively modified for the
11728 6.1 release of <productname>PostgreSQL</productname>.
11732 Three new data types (<type>datetime</type>, <type>timespan</type>, and <type>circle</type>) have been added to
11733 the native set of <productname>PostgreSQL</productname> types. Points, boxes, paths, and polygons
11734 have had their output formats made consistent across the data types.
11735 The polygon output in misc.out has only been spot-checked for correctness
11736 relative to the original regression output.
11740 <productname>PostgreSQL</productname> 6.1 introduces a new, alternate
11741 optimizer which uses <firstterm>genetic</firstterm>
11742 algorithms. These algorithms introduce a random behavior in the ordering
11743 of query results when the query contains multiple qualifiers or multiple
11744 tables (giving the optimizer a choice on order of evaluation). Several
11745 regression tests have been modified to explicitly order the results, and
11746 hence are insensitive to optimizer choices. A few regression tests are
11747 for data types which are inherently unordered (e.g. points and time
11748 intervals) and tests involving those types are explicitly bracketed with
11749 <command>set geqo to 'off'</command> and <command>reset geqo</command>.
11753 The interpretation of array specifiers (the curly braces around atomic
11754 values) appears to have changed sometime after the original regression
11755 tests were generated. The current <filename>./expected/*.out</filename> files reflect this
11756 new interpretation, which may not be correct!
11760 The float8 regression test fails on at least some platforms. This is due
11761 to differences in implementations of <function>pow()</function> and <function>exp()</function> and the signaling
11762 mechanisms used for overflow and underflow conditions.
11766 The <quote>random</> results in the random test should cause the
11767 <quote>random</quote> test to be <quote>failed</quote>, since the
11768 regression tests are evaluated using a simple diff. However,
11769 <quote>random</> does not seem to produce random results on my test
11770 machine (Linux/<application>gcc</>/i686).
11774 <title>Migration to version 6.1</title>
11777 This migration requires a complete dump of the 6.0 database and a
11778 restore of the database in 6.1.
11781 Those migrating from earlier 1.* releases should first upgrade to 1.09
11782 because the COPY output format was improved from the 1.02 release.
11787 <title>Changes</title>
11793 packet length checking in library routines
11794 lock manager priority patch
11795 check for under/over flow of float8(Bruce)
11796 multitable join fix(Vadim)
11797 SIGPIPE crash fix(Darren)
11798 large object fixes(Sven)
11799 allow btree indexes to handle NULLs(Vadim)
11800 timezone fixes(D'Arcy)
11801 select SUM(x) can return NULL on no rows(Thomas)
11802 internal optimizer, executor bug fixes(Vadim)
11803 fix problem where inner loop in < or <= has no rows(Vadim)
11804 prevent re-commuting join index clauses(Vadim)
11805 fix join clauses for multiple tables(Vadim)
11806 fix hash, hashjoin for arrays(Vadim)
11807 fix btree for abstime type(Vadim)
11808 large object fixes(Raymond)
11809 fix buffer leak in hash indexes (Vadim)
11810 fix rtree for use in inner scan (Vadim)
11811 fix gist for use in inner scan, cleanups (Vadim, Andrea)
11812 avoid unnecessary local buffers allocation (Vadim, Massimo)
11813 fix local buffers leak in transaction aborts (Vadim)
11814 fix file manager memmory leaks, cleanups (Vadim, Massimo)
11815 fix storage manager memmory leaks (Vadim)
11816 fix btree duplicates handling (Vadim)
11817 fix deleted rows reincarnation caused by vacuum (Vadim)
11818 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
11819 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
11823 attribute optimization statistics(Bruce)
11824 much faster new btree bulk load code(Paul)
11825 BTREE UNIQUE added to bulk load code(Vadim)
11826 new lock debug code(Massimo)
11827 massive changes to libpg++(Leo)
11828 new GEQO optimizer speeds table multitable optimization(Martin)
11829 new WARN message for non-unique insert into unique key(Marc)
11830 update x=-3, no spaces, now valid(Bruce)
11831 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
11832 debug backend now pretty-prints tree(Darren)
11833 new Oracle character functions(Edmund)
11834 new plaintext password functions(Dan)
11835 no such class or insufficient privilege changed to distinct messages(Dan)
11836 new ANSI timestamp function(Dan)
11837 new ANSI Time and Date types (Thomas)
11838 move large chunks of data in backend(Martin)
11839 multicolumn btree indexes(Vadim)
11840 new SET var TO value command(Martin)
11841 update transaction status on reads(Dan)
11842 new locale settings for character types(Oleg)
11843 new SEQUENCE serial number generator(Vadim)
11844 GROUP BY function now possible(Vadim)
11845 re-organize regression test(Thomas,Marc)
11846 new optimizer operation weights(Vadim)
11847 new psql \z grant/permit option(Marc)
11848 new MONEY data type(D'Arcy,Thomas)
11849 tcp socket communication speed improved(Vadim)
11850 new VACUUM option for attribute statistics, and for certain columns (Vadim)
11851 many geometric type improvements(Thomas,Keith)
11852 additional regression tests(Thomas)
11853 new datestyle variable(Thomas,Vadim,Martin)
11854 more comparison operators for sorting types(Thomas)
11855 new conversion functions(Thomas)
11856 new more compact btree format(Vadim)
11857 allow pg_dumpall to preserve database ownership(Bruce)
11858 new SET GEQO=# and R_PLANS variable(Vadim)
11859 old (!GEQO) optimizer can use right-sided plans (Vadim)
11860 typechecking improvement in SQL parser(Bruce)
11861 new SET, SHOW, RESET commands(Thomas,Vadim)
11862 new \connect database USER option
11863 new destroydb -i option (Igor)
11864 new \dt and \di psql commands (Darren)
11865 SELECT "\n" now escapes newline (A. Duursma)
11866 new geometry conversion functions from old format (Thomas)
11868 Source tree changes
11869 -------------------
11870 new configuration script(Marc)
11871 readline configuration option added(Marc)
11872 OS-specific configuration options removed(Marc)
11873 new OS-specific template files(Marc)
11874 no more need to edit Makefile.global(Marc)
11875 re-arrange include files(Marc)
11876 nextstep patches (Gregor Hoffleit)
11877 removed Windows-specific code(Bruce)
11878 removed postmaster -e option, now only postgres -e option (Bruce)
11879 merge duplicate library code in front/backends(Martin)
11880 now works with eBones, international Kerberos(Jun)
11881 more shared library support
11882 c++ include file cleanup(Bruce)
11883 warn about buggy flex(Bruce)
11884 DG/UX, Ultrix, IRIX, AIX portability fixes
11890 <sect1 id="release-6-0">
11891 <title>Release 6.0</title>
11894 <title>Release date</title>
11895 <simpara>1997-01-29</simpara>
11899 A dump/restore is required for those wishing to migrate data from
11900 previous releases of <productname>PostgreSQL</productname>.
11904 <title>Migration from version 1.09 to version 6.0</title>
11907 This migration requires a complete dump of the 1.09 database and a
11908 restore of the database in 6.0.
11913 <title>Migration from pre-1.09 to version 6.0</title>
11916 Those migrating from earlier 1.* releases should first upgrade to 1.09
11917 because the COPY output format was improved from the 1.02 release.
11922 <title>Changes</title>
11928 ALTER TABLE bug - running postgress process needs to re-read table definition
11929 Allow vacuum to be run on one table or entire database(Bruce)
11931 Fix array over-runs of memory writes(Kurt)
11932 Fix elusive btree range/non-range bug(Dan)
11933 Fix for hash indexes on some types like time and date
11934 Fix for pg_log size explosion
11935 Fix permissions on lo_export()(Bruce)
11936 Fix unitialized reads of memory(Kurt)
11937 Fixed ALTER TABLE ... char(3) bug(Bruce)
11938 Fixed a few small memory leaks
11939 Fixed EXPLAIN handling of options and changed full_path option name
11940 Fixed output of group acl privileges
11941 Memory leaks (hunt and destroy with tools like Purify(Kurt)
11942 Minor improvements to rules system
11944 New asserts for run-checking
11945 Overhauled parser/analyze code to properly report errors and increase speed
11946 Pg_dump -d now handles NULL's properly(Bruce)
11947 Prevent SELECT NULL from crashing server (Bruce)
11948 Properly report errors when INSERT ... SELECT columns did not match
11949 Properly report errors when insert column names were not correct
11950 psql \g filename now works(Bruce)
11951 psql fixed problem with multiple statements on one line with multiple outputs
11952 Removed duplicate system OIDs
11953 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
11954 Several fixes for queries that crashed the backend
11955 Starting quote in insert string errors(Bruce)
11956 Submitting an empty query now returns empty status, not just " " query(Bruce)
11960 Add EXPLAIN manual page(Bruce)
11961 Add UNIQUE index capability(Dan)
11962 Add hostname/user level access control rather than just hostname and user
11963 Add synonym of != for <>(Bruce)
11964 Allow "select oid,* from table"
11965 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Bruce)
11966 Allow COPY from the frontend(Bryan)
11967 Allow GROUP BY to use alias column name(Bruce)
11968 Allow actual compression, not just reuse on the same page(Vadim)
11969 Allow installation-configuration option to auto-add all local users(Bryan)
11970 Allow libpq to distinguish between text value '' and null(Bruce)
11971 Allow non-postgres users with createdb privs to destroydb's
11972 Allow restriction on who can create C functions(Bryan)
11973 Allow restriction on who can do backend COPY(Bryan)
11974 Can shrink tables, pg_time and pg_log(Vadim & Erich)
11975 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
11976 Change default decimal constant representation from float4 to float8(Bruce)
11977 European date format now set when postmaster is started
11978 Execute lowercase function names if not found with exact case
11979 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z'
11980 Gist now included in the distrubution(Marc)
11981 Idend authentication of local users(Bryan)
11982 Implement BETWEEN qualifier(Bruce)
11983 Implement IN qualifier(Bruce)
11984 libpq has PQgetisnull()(Bruce)
11985 libpq++ improvements
11986 New options to initdb(Bryan)
11987 Pg_dump allow dump of OIDs(Bruce)
11988 Pg_dump create indexes after tables are loaded for speed(Bruce)
11989 Pg_dumpall dumps all databases, and the user table
11990 Pginterface additions for NULL values(Bruce)
11991 Prevent postmaster from being run as root
11992 psql \h and \? is now readable(Bruce)
11993 psql allow backslashed, semicolons anywhere on the line(Bruce)
11994 psql changed command prompt for lines in query or in quotes(Bruce)
11995 psql char(3) now displays as (bp)char in \d output(Bruce)
11996 psql return code now more accurate(Bryan?)
11997 psql updated help syntax(Bruce)
11998 Re-visit and fix vacuum(Vadim)
11999 Reduce size of regression diffs, remove timezone name difference(Bruce)
12000 Remove compile-time parameters to enable binary distributions(Bryan)
12001 Reverse meaning of HBA masks(Bryan)
12002 Secure Authentication of local users(Bryan)
12003 Speed up vacuum(Vadim)
12004 Vacuum now had VERBOSE option(Bruce)
12006 Source tree changes
12007 -------------------
12008 All functions now have prototypes that are compared against the calls
12009 Allow asserts to be disabled easly from Makefile.global(Bruce)
12010 Change oid constants used in code to #define names
12011 Decoupled sparc and solaris defines(Kurt)
12012 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
12013 Major include file reorganization/reduction(Marc)
12014 Make now stops on compile failure(Bryan)
12015 Makefile restructuring(Bryan, Marc)
12016 Merge bsdi_2_1 to bsdi(Bruce)
12017 Monitor program removed
12018 Name change from Postgres95 to PostgreSQL
12019 New config.h file(Marc, Bryan)
12020 PG_VERSION now set to 6.0 and used by postmaster
12021 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
12022 Reduced the number of #define's, centeralized #define's
12023 Remove duplicate OIDS in system tables(Dan)
12024 Remove duplicate system catalog info or report mismatches(Dan)
12025 Removed many os-specific #define's
12026 Restructured object file generation/location(Bryan, Marc)
12027 Restructured port-specific file locations(Bryan, Marc)
12028 Unused/uninialized variables corrected
12034 <sect1 id="release-1-09">
12035 <title>Release 1.09</title>
12038 <title>Release date</title>
12039 <simpara>1996-11-04</simpara>
12043 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of
12044 the changes listed in 6.0 were actually included in the 1.02.1 to 1.09
12049 <sect1 id="release-1-02">
12050 <title>Release 1.02</title>
12053 <title>Release date</title>
12054 <simpara>1996-08-01</simpara>
12058 <title>Migration from version 1.02 to version 1.02.1</title>
12061 Here is a new migration file for 1.02.1. It includes the 'copy' change
12062 and a script to convert old <acronym>ASCII</acronym> files.
12066 The following notes are for the benefit of users who want to migrate
12067 databases from <productname>Postgres95</> 1.01 and 1.02 to <productname>Postgres95</> 1.02.1.
12070 If you are starting afresh with <productname>Postgres95</> 1.02.1 and do not need
12071 to migrate old databases, you do not need to read any further.
12076 In order to upgrade older <productname>Postgres95</> version 1.01 or 1.02 databases to
12077 version 1.02.1, the following steps are required:
12082 Start up a new 1.02.1 postmaster
12087 Add the new built-in functions and operators of 1.02.1 to 1.01 or 1.02
12088 databases. This is done by running the new 1.02.1 server against
12089 your own 1.01 or 1.02 database and applying the queries attached at
12090 the end of the file. This can be done easily through <command>psql</>. If your
12091 1.01 or 1.02 database is named <literal>testdb</literal> and you have cut the commands
12092 from the end of this file and saved them in <filename>addfunc.sql</filename>:
12094 % psql testdb -f addfunc.sql
12097 Those upgrading 1.02 databases will get a warning when executing the
12098 last two statements in the file because they are already present in 1.02. This is
12099 not a cause for concern.
12106 <title>Dump/Reload Procedure</title>
12109 If you are trying to reload a pg_dump or text-mode, <literal>copy tablename to
12110 stdout</literal> generated with a previous version, you will need to run the
12111 attached <command>sed</command> script on the ASCII file before loading it into the
12112 database. The old format used '.' as end-of-data, while '\.' is now the
12113 end-of-data marker. Also, empty strings are now loaded in as '' rather
12114 than NULL. See the copy manual page for full details.
12117 sed 's/^\.$/\\./g' <in_file >out_file
12121 If you are loading an older binary copy or non-<systemitem>stdout</> copy, there is no
12122 end-of-data character, and hence no conversion necessary.
12125 -- following lines added by agc to reflect the case-insensitive
12126 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
12127 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexeq);
12128 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregexne);
12129 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregexeq);
12130 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticregexne);
12136 <title>Changes</title>
12140 Source code maintenance and development
12141 * worldwide team of volunteers
12142 * the source tree now in CVS at ftp.ki.net
12145 * psql (and underlying libpq library) now has many more options for
12146 formatting output, including HTML
12147 * pg_dump now output the schema and/or the data, with many fixes to
12148 enhance completeness.
12149 * psql used in place of monitor in administration shell scripts.
12150 monitor to be deprecated in next release.
12151 * date/time functions enhanced
12152 * NULL insert/update/comparison fixed/enhanced
12153 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
12155 Bug Fixes (almost too numerous to mention)
12157 * storage management
12158 * check for NULL pointer before dereferencing
12162 * added SolarisX86 port
12163 * added BSD/OS 2.1 port
12168 Contributors (appologies to any missed)
12169 * Kurt J. Lidl <lidl@va.pubnix.com>
12170 (missed in first run, but no less important)
12171 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12172 * Jason Wright <jason@shiloh.vnet.net>
12173 * Cees de Groot <C.deGroot@inter.NL.net>
12174 * ernst.molitor@uni-bonn.de
12175 * michael.siebenborn@ae3.Hypo.DE (Michael Siebenborn (6929))
12176 * Brian E. Gallew <geek+@cmu.edu>
12177 * Vadim B. Mikheev <vadim@sable.krasnoyarsk.su>
12178 * Adam Sussman <myddryn@vidya.com>
12179 * Chris Dunlop <chris@onthe.net.au>
12180 * Marc G. Fournier <scrappy@ki.net>
12181 * Dan McGuirk <mcguirk@indirect.com>
12182 * Dr_George_D_Detlefsen <drgeorge@ilt.com>
12183 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12184 * Massimo Dal Zotto <dz@cs.unitn.it>
12185 * Randy Kunkee <kunkee@Starbase.NeoSoft.COM>
12186 * Rick Weldon <rick@wisetech.com>
12187 * Thomas van Reimersdahl <reimersd@dali.techinfo.rwth-aachen.de>
12188 * david bennett <dave@bensoft.com>
12189 * ernst.molitor@uni-bonn.de
12190 * Julian Assange <proff@suburbia.net>
12191 * Bruce Momjian <pgman@candle.pha.pa.us>
12192 * Paul "Shag" Walmsley <ccshag@cclabs.missouri.edu>
12193 * "Alistair G. Crooks" <azcb0@sde.uts.amdahl.com>
12198 <sect1 id="release-1-01">
12199 <title>Release 1.01</title>
12202 <title>Release date</title>
12203 <simpara>1996-02-23</simpara>
12208 <title>Migration from version 1.0 to version 1.01</title>
12211 The following notes are for the benefit of users who want to migrate
12212 databases from <productname>Postgres95</> 1.0 to <productname>Postgres95</> 1.01.
12215 If you are starting afresh with <productname>Postgres95</> 1.01 and do not need
12216 to migrate old databases, you do not need to read any further.
12219 In order to <productname>Postgres95</> version 1.01 with databases created with
12220 <productname>Postgres95</> version 1.0, the following steps are required:
12225 Set the definition of <symbol>NAMEDATALEN</symbol> in <filename>src/Makefile.global</filename> to 16
12226 and <symbol>OIDNAMELEN</symbol> to 20.
12231 Decide whether you want to use Host based authentication.
12236 If you do, you must create a file name <literal>pg_hba</literal> in your top-level data
12237 directory (typically the value of your <envar>$PGDATA</envar>). <filename>src/libpq/pg_hba</filename>
12238 shows an example syntax.
12243 If you do not want host-based authentication, you can comment out
12248 in <filename>src/Makefile.global</filename>
12251 Note that host-based authentication is turned on by default, and if
12252 you do not take steps A or B above, the out-of-the-box 1.01 will
12253 not allow you to connect to 1.0 databases.
12261 Compile and install 1.01, but DO NOT do the <command>initdb</command> step.
12266 Before doing anything else, terminate your 1.0 postmaster, and
12267 backup your existing <envar>$PGDATA</envar> directory.
12272 Set your <envar>PGDATA</envar> environment variable to your 1.0 databases, but set up
12273 path up so that 1.01 binaries are being used.
12278 Modify the file <filename><envar>$PGDATA</envar>/PG_VERSION</filename> from 5.0 to 5.1
12283 Start up a new 1.01 postmaster
12288 Add the new built-in functions and operators of 1.01 to 1.0
12289 databases. This is done by running the new 1.01 server against
12290 your own 1.0 database and applying the queries attached and saving
12291 in the file 1.0_to_1.01.sql. This can be done easily through <command>psql</command>.
12292 If your 1.0 database is name <literal>testdb</literal>:
12295 % psql testdb -f 1.0_to_1.01.sql
12298 and then execute the following commands (cut and paste from here):
12301 -- add builtin functions that are new to 1.01
12303 create function int4eqoid (int4, oid) returns bool as 'foo'
12304 language 'internal';
12305 create function oideqint4 (oid, int4) returns bool as 'foo'
12306 language 'internal';
12307 create function char2icregexeq (char2, text) returns bool as 'foo'
12308 language 'internal';
12309 create function char2icregexne (char2, text) returns bool as 'foo'
12310 language 'internal';
12311 create function char4icregexeq (char4, text) returns bool as 'foo'
12312 language 'internal';
12313 create function char4icregexne (char4, text) returns bool as 'foo'
12314 language 'internal';
12315 create function char8icregexeq (char8, text) returns bool as 'foo'
12316 language 'internal';
12317 create function char8icregexne (char8, text) returns bool as 'foo'
12318 language 'internal';
12319 create function char16icregexeq (char16, text) returns bool as 'foo'
12320 language 'internal';
12321 create function char16icregexne (char16, text) returns bool as 'foo'
12322 language 'internal';
12323 create function texticregexeq (text, text) returns bool as 'foo'
12324 language 'internal';
12325 create function texticregexne (text, text) returns bool as 'foo'
12326 language 'internal';
12328 -- add builtin functions that are new to 1.01
12330 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
12331 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
12332 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexeq);
12333 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregexne);
12334 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexeq);
12335 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregexne);
12336 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexeq);
12337 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregexne);
12338 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icregexeq);
12339 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icregexne);
12340 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq);
12341 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne);
12349 <title>Changes</title>
12354 * 1.01 is backwards compatible with 1.0 database provided the user
12355 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
12356 If those steps are not taken, 1.01 is not compatible with 1.0 database.
12359 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
12360 * added NeXT port (requires SysVIPC implementation)
12361 * added CAST .. AS ... syntax
12362 * added ASC and DESC key words
12363 * added 'internal' as a possible language for CREATE FUNCTION
12364 internal functions are C functions which have been statically linked
12365 into the postgres backend.
12366 * a new type "name" has been added for system identifiers (table names,
12367 attribute names, etc.) This replaces the old char16 type. The
12368 of name is set by the NAMEDATALEN #define in src/Makefile.global
12369 * a readable reference manual that describes the query language.
12370 * added host-based access control. A configuration file ($PGDATA/pg_hba)
12371 is used to hold the configuration data. If host-based access control
12372 is not desired, comment out HBA=1 in src/Makefile.global.
12373 * changed regex handling to be uniform use of Henry Spencer's regex code
12374 regardless of platform. The regex code is included in the distribution
12375 * added functions and operators for case-insensitive regular expressions.
12376 The operators are ~* and !~*.
12377 * pg_dump uses COPY instead of SELECT loop for better performance
12380 * fixed an optimizer bug that was causing core dumps when
12381 functions calls were used in comparisons in the WHERE clause
12382 * changed all uses of getuid to geteuid so that effective uids are used
12383 * psql now returns non-zero status on errors when using -c
12384 * applied public patches 1-14
12390 <sect1 id="release-1-0">
12391 <title>Release 1.0</title>
12394 <title>Release date</title>
12395 <simpara>1995-09-05</simpara>
12399 <title>Changes</title>
12404 * The copyright of <productname>Postgres</productname> 1.0 has been loosened to be freely modifiable
12405 and modifiable for any purpose. Please read the COPYRIGHT file.
12406 Thanks to Professor Michael Stonebraker for making this possible.
12409 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
12410 EUROPEAN STYLE). This follows SQL-92 specs.
12411 * "delimiters" is now a key word
12414 * sql LIKE syntax has been added
12415 * copy command now takes an optional USING DELIMITER specification.
12416 delimiters can be any single-character string.
12417 * IRIX 5.3 port has been added.
12418 Thanks to Paul Walmsley and others.
12419 * updated pg_dump to work with new libpq
12420 * \d has been added psql
12421 Thanks to Keith Parks
12422 * regexp performance for architectures that use POSIX regex has been
12423 improved due to caching of precompiled patterns.
12424 Thanks to Alistair Crooks
12425 * a new version of libpq++
12426 Thanks to William Wanders
12429 * arbitrary userids can be specified in the createuser script
12430 * \c to connect to other databases in psql now works.
12431 * bad pg_proc entry for float4inc() is fixed
12432 * users with usecreatedb field set can now create databases without
12433 having to be usesuper
12434 * remove access control entries when the entry no longer has any
12436 * fixed non-portable datetimes implementation
12437 * added kerberos flags to the src/backend/Makefile
12438 * libpq now works with kerberos
12439 * typographic errors in the user manual have been corrected.
12440 * btrees with multiple index never worked, now we tell you they don't
12441 work when you try to use them
12447 <sect1 id="release-0-03">
12448 <title><productname>Postgres95</productname> Release 0.03</title>
12451 <title>Release date</title>
12452 <simpara>1995-07-21</simpara>
12456 <title>Changes</title>
12459 Incompatible changes:
12460 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
12461 (due to system catalog changes and indexing structure changes).
12462 * double-quote (") is deprecated as a quoting character for string literals;
12463 you need to convert them to single quotes ('). <!-- " -->
12464 * name of aggregates (eg. int4sum) are renamed in accordance with the
12465 SQL standard (eg. sum).
12466 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
12467 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
12468 previous releases); you might have to do typecasting if you depend on it
12469 being of type float8. If you neglect to do the typecasting and you assign
12470 a float literal to a field of type float8, you may get incorrect values
12472 * LIBPQ has been totally revamped so that frontend applications
12473 can connect to multiple backends
12474 * the usesysid field in pg_user has been changed from int2 to int4 to
12475 allow wider range of Unix user ids.
12476 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
12477 single BSD44_derived port. (thanks to Alistair Crooks)
12479 SQL standard-compliance (the following details changes that makes postgres95
12480 more compliant to the SQL-92 standard):
12481 * the following SQL types are now built-in: smallint, int(eger), float, real,
12482 char(N), varchar(N), date and time.
12484 The following are aliases to existing postgres types:
12485 smallint -> int2
12486 integer, int -> int4
12487 float, real -> float4
12488 char(N) and varchar(N) are implemented as truncated text types. In
12489 addition, char(N) does blank-padding.
12490 * single-quote (') is used for quoting string literals; '' (in addition to
12491 \') is supported as means of inserting a single quote in a string
12492 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
12493 (Also, aggregates can now be overloaded, i.e. you can define your
12494 own MAX aggregate to take in a user-defined type.)
12495 * CHANGE ACL removed. GRANT/REVOKE syntax added.
12496 - Privileges can be given to a group using the "GROUP" key word.
12498 GRANT SELECT ON foobar TO GROUP my_group;
12499 The key word 'PUBLIC' is also supported to mean all users.
12501 Privileges can only be granted or revoked to one user or group
12504 "WITH GRANT OPTION" is not supported. Only class owners can change
12506 - The default access control is to to grant users readonly access.
12507 You must explicitly grant insert/update access to users. To change
12508 this, modify the line in
12509 src/backend/utils/acl.h
12510 that defines ACL_WORLD_DEFAULT
12513 * the bug where aggregates of empty tables were not run has been fixed. Now,
12514 aggregates run on empty tables will return the initial conditions of the
12515 aggregates. Thus, COUNT of an empty table will now properly return 0.
12516 MAX/MIN of an empty table will return a row of value NULL.
12517 * allow the use of \; inside the monitor
12518 * the LISTEN/NOTIFY asynchronous notification mechanism now work
12519 * NOTIFY in rule action bodies now work
12520 * hash indexes work, and access methods in general should perform better.
12521 creation of large btree indexes should be much faster. (thanks to Paul
12524 Other changes and enhancements:
12525 * addition of an EXPLAIN statement used for explaining the query execution
12526 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
12528 * WARN and NOTICE messages no longer have timestamps on them. To turn on
12529 timestamps of error messages, uncomment the line in
12530 src/backend/utils/elog.h:
12531 /* define ELOG_TIMESTAMPS */
12532 * On an access control violation, the message
12533 "Either no such class or insufficient privilege"
12534 will be given. This is the same message that is returned when
12535 a class is not found. This dissuades non-privileged users from
12536 guessing the existence of privileged classes.
12537 * some additional system catalog changes have been made that are not
12538 visible to the user.
12541 * The -oid option has been added to the "pg_result" tcl command.
12542 pg_result -oid returns oid of the last row inserted. If the
12543 last command was not an INSERT, then pg_result -oid returns "".
12544 * the large object interface is available as pg_lo* tcl commands:
12545 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
12547 Portability enhancements and New Ports:
12548 * flex/lex problems have been cleared up. Now, you should be able to use
12549 flex instead of lex on any platforms. We no longer make assumptions of
12550 what lexer you use based on the platform you use.
12551 * The Linux-ELF port is now supported. Various configuration have been
12552 tested: The following configuration is known to work:
12553 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
12554 with everything in ELF format,
12557 * ipcclean added to the distribution
12558 ipcclean usually does not need to be run, but if your backend crashes
12559 and leaves shared memory segments hanging around, ipcclean will
12560 clean them up for you.
12563 * the user manual has been revised and libpq documentation added.
12569 <sect1 id="release-0-02">
12570 <title><productname>Postgres95</productname> Release 0.02</title>
12573 <title>Release date</title>
12574 <simpara>1995-05-25</simpara>
12578 <title>Changes</title>
12582 Incompatible changes:
12583 * The SQL statement for creating a database is 'CREATE DATABASE' instead
12584 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
12585 of 'DESTROYDB'. However, the names of the executables 'createdb' and
12586 'destroydb' remain the same.
12589 * pgperl - a Perl (4.036) interface to Postgres95
12590 * pg_dump - a utility for dumping out a postgres database into a
12591 script file containing query commands. The script files are in a ASCII
12592 format and can be used to reconstruct the database, even on other
12593 machines and other architectures. (Also good for converting
12594 a Postgres 4.2 database to Postgres95 database.)
12596 The following ports have been incorporated into postgres95-beta-0.02:
12597 * the NetBSD port by Alistair Crooks
12598 * the AIX port by Mike Tung
12599 * the Windows NT port by Jon Forrest (more stuff but not done yet)
12600 * the Linux ELF port by Brian Gallew
12602 The following bugs have been fixed in postgres95-beta-0.02:
12603 * new lines not escaped in COPY OUT and problem with COPY OUT when first
12605 * cannot type return to use the default user id in createuser
12606 * SELECT DISTINCT on big tables crashes
12607 * Linux installation problems
12608 * monitor doesn't allow use of 'localhost' as PGHOST
12609 * psql core dumps when doing \c or \l
12610 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
12611 * libpgtcl has a hard-wired default port number
12612 * SELECT DISTINCT INTO TABLE hangs
12613 * CREATE TYPE doesn't accept 'variable' as the internallength
12614 * wrong result using more than 1 aggregate in a SELECT
12620 <sect1 id="release-0-01">
12621 <title><productname>Postgres95</productname> Release 0.01</title>
12624 <title>Release date</title>
12625 <simpara>1995-05-01</simpara>
12634 <sect1 id="timing-results">
12635 <title>Timing Results</title>
12638 These timing results are from running the regression test with the commands
12641 % cd src/test/regress
12643 % time make runtest
12647 Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
12648 to run, presumably due to the scheduling vagaries of multitasking systems.
12652 <title>Version 6.5</title>
12655 As has been the case for previous releases, timing between
12656 releases is not directly comparable since new regression tests
12657 have been added. In general, 6.5 is faster than previous
12662 Timing with <function>fsync()</function> disabled:
12666 02:00 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
12667 04:38 Sparc Ultra 1 143MHz, 64MB, Solaris 2.6
12672 Timing with <function>fsync()</function> enabled:
12676 04:21 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
12679 For the <systemitem class="osname">Linux</systemitem> system above, using <acronym>UW-SCSI</acronym> disks rather than (older) <acronym>IDE</acronym>
12680 disks leads to a 50% improvement in speed on the regression test.
12685 <title>Version 6.4beta</title>
12688 The times for this release are not directly comparable to those for previous releases
12689 since some additional regression tests have been included.
12690 In general, however, 6.4 should be slightly faster than the previous release (thanks, Bruce!).
12695 02:26 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
12701 <title>Version 6.3</title>
12704 The times for this release are not directly comparable to those for previous releases
12705 since some additional regression tests have been included and some obsolete tests involving
12706 time travel have been removed.
12707 In general, however, 6.3 is substantially faster than previous releases (thanks, Bruce!).
12712 02:30 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
12713 04:12 Dual Pentium Pro 180, 96MB, EIDE, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
12719 <title>Version 6.1</title>
12724 06:12 Pentium Pro 180, 32MB, EIDE, Linux 2.0.30, gcc 2.7.2 -O2 -m486
12725 12:06 P-100, 48MB, Linux 2.0.29, gcc
12726 39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
12734 <!-- Keep this comment at the end of the file
12739 sgml-minimize-attributes:nil
12740 sgml-always-quote-attributes:t
12743 sgml-parent-document:nil
12744 sgml-default-dtd-file:"./reference.ced"
12745 sgml-exposed-tags:nil
12746 sgml-local-catalogs:("/usr/lib/sgml/catalog")
12747 sgml-local-ecat-files:nil