2 $Header: /cvsroot/pgsql/doc/src/sgml/release.sgml,v 1.169 2002/12/12 20:35:07 tgl Exp $
5 <appendix id="release">
6 <title>Release Notes</title>
8 <sect1 id="release-devel">
9 <title>7.4 Development Branch</title>
12 Below is a subset of the changes that have gone into the
13 development branch of PostgreSQL since version 7.3. For a complete
14 list of changes, consult the CVS logs.
18 Developers: When you add a feature, mention it here. This avoids
19 lossiness when digging out the information from the CVS logs, and
20 furthermore it advertises your feature to external parties at the
21 earliest possible moment.
23 CDATA means the content is "SGML-free", so you can write without
24 worries about funny characters.
26 <literallayout><![CDATA[
27 Domains now support CHECK constraints
32 <sect1 id="release-7-3">
33 <title>Release 7.3</title>
36 <title>Release date</title>
37 <simpara>2002-11-27</simpara>
41 <title>Overview</title>
44 Major changes in this release:
51 Schemas allow users to create objects in separate namespaces,
52 so two people or applications can have tables with the same
53 name. There is also a public schema for shared tables.
54 Table/index creation can be restricted by removing permissions
61 <term>Drop Column</term>
64 PostgreSQL now supports the <literal>ALTER TABLE ... DROP
65 COLUMN</literal> functionality.
71 <term>Table Functions</term>
74 Functions returning multiple rows and/or multiple columns are
75 now much easier to use than before. You can call such a
76 <quote>table function</quote> in the <literal>SELECT</literal>
77 <literal>FROM</literal> clause, treating its output like a
78 table. Also, <application>PL/pgSQL</application> functions can
85 <term>Prepared Queries</term>
88 PostgreSQL now supports prepared queries, for improved
95 <term>Dependency Tracking</term>
98 PostgreSQL now records object dependencies, which allows
99 improvements in many areas. <command>DROP</command>
100 statements now take either <literal>CASCADE</> or
101 <literal>RESTRICT</> to control whether dependent objects are
108 <term>Privileges</term>
111 Functions and procedural languages now have privileges, and
112 functions can be defined to run with the privileges of their
119 <term>Internationalization</term>
122 Both multibyte and locale support are now always enabled.
131 A variety of logging options have been enhanced.
137 <term>Interfaces</term>
140 A large number of interfaces have been moved to <ulink
141 url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
142 where they can be developed and released independently.
148 <term>Functions/Identifiers</term>
151 By default, functions can now take up to 32 parameters, and
152 identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
153 is now deprecated: there are specific <quote>pseudo-datatypes</>
154 to represent each of the former meanings of <literal>OPAQUE</>
155 in function argument and result types.
165 <title>Migration to version 7.3</title>
168 A dump/restore using <application>pg_dump</> is required for those
169 wishing to migrate data from any previous release. If your
170 application examines the system catalogs, additional changes will
171 be required due to the introduction of schemas in 7.3; for more
172 information, see: <ulink
173 url="http://www.ca.postgresql.org/docs/momjian/upgrade_tips_7.3">
174 http://www.ca.postgresql.org/docs/momjian/upgrade_tips_7.3</>.
178 Observe the following incompatibilities:
183 Pre-6.3 clients are no longer supported.
189 <filename>pg_hba.conf</filename> now has a column for the user
190 name and additional features. Existing files need to be
197 Several <filename>postgresql.conf</filename> logging parameters
204 <literal>LIMIT #,#</literal> has been disabled; use
205 <literal>LIMIT # OFFSET #</literal>.
211 <command>INSERT</command> statements with column lists must
212 specify a value for each specified column. For example,
213 <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
214 is now invalid. It's still allowed to supply fewer columns than
215 expected if the <command>INSERT</command> does not have a column list.
221 <type>serial</type> columns are no longer automatically
222 <literal>UNIQUE</>; thus, an index will not automatically be
229 A <command>SET</command> command inside an aborted transaction
236 <command>COPY</command> no longer considers missing trailing
237 columns to be null. All columns need to be specified.
238 (However, one may achieve a similar effect by specifying a
239 column list in the <command>COPY</command> command.)
245 The data type <type>timestamp</type> is now equivalent to
246 <type>timestamp without time zone</type>, instead of
247 <type>timestamp with time zone</type>.
253 Pre-7.3 databases loaded into 7.3 will not have the new object
254 dependencies for <type>serial</type> columns, unique
255 constraints, and foreign keys. See the directory
256 <filename>contrib/adddepend/</filename> for a detailed
257 description and a script that will add such dependencies.
263 An empty string (<literal>''</literal>) is no longer allowed as
264 the input into an integer field. Formerly, it was silently
275 <title>Changes</title>
278 <title>Server Operation</title>
280 Add pg_locks view to show locks (Neil)
281 Security fixes for password negotiation memory allocation (Neil)
282 Remove support for version 0 FE/BE protocol (PostgreSQL 6.2 and
284 Reserve the last few backend slots for superusers, add parameter
285 superuser_reserved_connections to control this (Nigel J. Andrews)
290 <title>Performance</title>
292 Improve startup by calling localtime() only once (Tom)
293 Cache system catalog information in flat files for faster startup
295 Improve caching of index information (Tom)
296 Optimizer improvements (Tom, Fernando Nasser)
297 Catalog caches now store failed lookups (Tom)
298 Hash function improvements (Neil)
299 Improve performance of query tokenization and network handling (Peter)
300 Speed improvement for large object restore (Mario Weilguni)
301 Mark expired index entries on first lookup, saving later heap fetches
303 Avoid excessive NULL bitmap padding (Manfred Koizar)
304 Add BSD-licensed qsort() for Solaris, for performance (Bruce)
305 Reduce per-row overhead by four bytes (Manfred Koizar)
306 Fix GEQO optimizer bug (Neil Conway)
307 Make WITHOUT OID actually save four bytes per row (Manfred Koizar)
308 Add default_statistics_target variable to specify ANALYZE buckets
310 Use local buffer cache for temporary tables so no WAL overhead (Tom)
311 Improve free space map performance on large tables (Stephen Marshall,
313 Improved WAL write concurrency (Tom)
318 <title>Privileges</title>
320 Add privileges on functions and procedural languages (Peter)
321 Add OWNER to CREATE DATABASE so superusers can create databases
322 on behalf of unprivileged users. (Gavin Sherry, Tom)
323 Add new object permission bits EXECUTE and USAGE (Tom)
324 Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION
326 Allow functions to be executed with the privilege of the function
332 <title>Server Configuration</title>
334 Server log messages now tagged with LOG, not DEBUG (Bruce)
335 Add user column to pg_hba.conf (Bruce)
336 Have log_connections output two lines in log file (Tom)
337 Remove debug_level from postgresql.conf, now server_min_messages
339 New ALTER DATABASE/USER ... SET command for per-user/database
340 initialization (Peter)
341 New parameters server_min_messages and client_min_messages to
342 control which messages are sent to the server logs or client
344 Allow pg_hba.conf to specify lists of users/databases separated by
345 commas, group names prepended with +, and file names prepended
347 Remove secondary password file capability and pg_password utility
349 Add variable db_user_namespace for database-local user names (Bruce)
350 SSL improvements (Bear Giles)
351 Make encryption of stored passwords the default (Bruce)
352 Allow pg_statistics to be reset by calling pg_stat_reset()
354 Add log_duration parameter (Bruce)
355 Rename debug_print_query to log_statement (Bruce)
356 Rename show_query_stats to show_statement_stats (Bruce)
357 Add param log_min_error_statement to print commands to logs on error
363 <title>Queries</title>
365 Make cursors insensitive, meaning their contents do not change (Tom)
366 Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)
367 Increase identifier length to 63 (Neil, Bruce)
368 UNION fixes for merging >= 3 columns of different lengths (Tom)
369 Add DEFAULT keyword to INSERT, e.g., INSERT ... (..., DEFAULT, ...)
371 Allow views to have default values using ALTER COLUMN ... SET DEFAULT
373 Fail on INSERTs with column lists that don't supply all column
374 values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)
375 Fix for join aliases (Tom)
376 Fix for FULL OUTER JOINs (Tom)
377 Improve reporting of invalid identifier and location (Tom, Gavin)
378 Fix OPEN cursor(args) (Tom)
379 Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)
380 Fix for CREATE TABLE AS with UNION (Tom)
381 SQL99 syntax improvements (Thomas)
382 Add statement_timeout variable to cancel queries (Bruce)
383 Allow prepared queries with PREPARE/EXECUTE (Neil)
384 Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)
385 Add variable autocommit (Tom, David Van Wie)
390 <title>Object Manipulation</title>
392 Make equals signs optional in CREATE DATABASE (Gavin Sherry)
393 Make ALTER TABLE OWNER change index ownership too (Neil)
394 New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls
395 TOAST storage, compression (John Gray)
396 Add schema support, CREATE/DROP SCHEMA (Tom)
397 Create schema for temporary tables (Tom)
398 Add variable search_path for schema search (Tom)
399 Add ALTER TABLE SET/DROP NOT NULL (Christopher)
400 New CREATE FUNCTION volatility levels (Tom)
401 Make rule names unique only per table (Tom)
402 Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)
403 Add ALTER TRIGGER RENAME (Joe)
404 New current_schema() and current_schemas() inquiry functions (Tom)
405 Allow functions to return multiple rows (table functions) (Joe)
406 Make WITH optional in CREATE DATABASE, for consistency (Bruce)
407 Add object dependency tracking (Rod, Tom)
408 Add RESTRICT/CASCADE to DROP commands (Rod)
409 Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)
410 Autodestroy sequence on DROP of table with SERIAL (Rod)
411 Prevent column dropping if column is used by foreign key (Rod)
412 Automatically drop constraints/functions when object is dropped (Rod)
413 Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)
414 Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)
415 Prevent inherited columns from being removed or renamed (Alvaro
417 Fix foreign key constraints to not error on intermediate database
419 Propagate column or table renaming to foreign key constraints
420 Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)
421 Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)
422 Have rules execute alphabetically, returning more predictable values
424 Triggers are now fired in alphabetical order (Tom)
425 Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)
426 Allow better casting when inserting/updating values (Tom)
431 <title>Utility Commands</title>
433 Have COPY TO output embedded carriage returns and newlines as \r and
435 Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)
436 Make pg_dump use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)
437 Disable brackets in multistatement rules (Bruce)
438 Disable VACUUM from being called inside a function (Bruce)
439 Allow dropdb and other scripts to use identifiers with spaces (Bruce)
440 Restrict database comment changes to the current database
441 Allow comments on operators, independent of the underlying function
443 Rollback SET commands in aborted transactions (Tom)
444 EXPLAIN now outputs as a query (Tom)
445 Display condition expressions and sort keys in EXPLAIN (Tom)
446 Add 'SET LOCAL var = value' to set configuration variables for a
447 single transaction (Tom)
448 Allow ANALYZE to run in a transaction (Bruce)
449 Improve COPY syntax using new WITH clauses, keep backward
450 compatibility (Bruce)
451 Fix pg_dump to consistently output tags in non-ASCII dumps (Bruce)
452 Make foreign key constraints clearer in dump file (Rod)
453 Add COMMENT ON CONSTRAINT (Rod)
454 Allow COPY TO/FROM to specify column names (Brent Verner)
455 Dump UNIQUE and PRIMARY KEY contraints as ALTER TABLE (Rod)
456 Have SHOW output a query result (Joe)
457 Generate failure on short COPY lines rather than pad NULLs (Neil)
458 Fix CLUSTER to preserve all table attributes (Alvaro Herrera)
459 New pg_settings table to view/modify GUC settings (Joe)
460 Add smart quoting, portability improvements to pg_dump output (Peter)
461 Dump serial columns out as SERIAL (Tom)
462 Enable large file support, >2G for pg_dump (Peter, Philip Warner,
464 Disallow TRUNCATE on tables that are involved in referential
466 Have TRUNCATE also auto-truncate the toast table of the relation (Tom)
467 Add clusterdb utility that will auto-cluster an entire database
468 based on previous CLUSTER operations (Alvaro Herrera)
469 Overhaul pg_dumpall (Peter)
470 Allow REINDEX of TOAST tables (Tom)
471 Implemented START TRANSACTION, per SQL99 (Neil)
472 Fix rare index corruption when a page split affects bulk delete (Tom)
473 Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)
478 <title>Data Types and Functions</title>
480 Fix factorial(0) to return 1 (Bruce)
481 Date/time/timezone improvements (Thomas)
482 Fix for array slice extraction (Tom)
483 Fix extract/date_part to report proper microseconds for timestamp
485 Allow text_substr() and bytea_substr() to read TOAST values more
486 efficiently (John Gray)
487 Add domain support (Rod)
488 Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types
490 Allow alternate storage scheme of 64-bit integers for date/time types
491 using --enable-integer-datetimes in configure (Thomas)
492 Make timezone(timestamptz) return timestamp rather than a string
494 Allow fractional seconds in date/time types for dates prior to 1BC
496 Limit timestamp data types to 6 decimal places of precision (Thomas)
497 Change timezone conversion functions from timetz() to timezone()
499 Add configuration variables datestyle and timezone (Tom)
500 Add OVERLAY(), which allows substitution of a substring in a string
502 Add SIMILAR TO (Thomas, Tom)
503 Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)
504 Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)
505 Add named composite types using CREATE TYPE typename AS (column)
507 Allow composite type definition in the table alias clause (Joe)
508 Add new API to simplify creation of C language table functions (Joe)
509 Remove ODBC-compatible empty parentheses from calls to SQL99
510 functions for which these parentheses do not match the standard
512 Allow macaddr data type to accept 12 hex digits with no separators
514 Add CREATE/DROP CAST (Peter)
515 Add IS DISTINCT FROM operator (Thomas)
516 Add SQL99 TREAT() function, synonym for CAST() (Thomas)
517 Add pg_backend_pid() to output backend pid (Bruce)
518 Add IS OF / IS NOT OF type predicate (Thomas)
519 Allow bit string constants without fully-specified length (Thomas)
520 Allow conversion between 8-byte integers and bit strings (Thomas)
521 Implement hex literal conversion to bit string literal (Thomas)
522 Allow table functions to appear in the FROM clause (Joe)
523 Increase maximum number of function parameters to 32 (Bruce)
524 No longer automatically create index for SERIAL column (Tom)
525 Add current_database() (Rod)
526 Fix cash_words() to not overflow buffer (Tom)
527 Add functions replace(), split_part(), to_hex() (Joe)
528 Fix LIKE for bytea as a right-hand argument (Joe)
529 Prevent crashes caused by SELECT cash_out(2) (Tom)
530 Fix to_char(1,'FM999.99') to return a period (Karel)
531 Fix trigger/type/language functions returning OPAQUE to return
537 <title>Internationalization</title>
539 Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese
540 (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC)
542 Enable locale support by default (Peter)
543 Add locale variables (Peter)
544 Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea
546 Add locale awareness to regular expression character classes
547 Enable multibyte support by default (Tatso)
548 Add GB18030 multibyte support (Bill Huang)
549 Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo,
551 Add pg_conversion table (Tatsuo)
552 Add SQL99 CONVERT() function (Tatsuo)
553 pg_dumpall, pg_controldata, and pg_resetxlog now national-language
555 New and updated translations
560 <title>Server-side Languages</title>
562 Allow recursive SQL function (Peter)
563 Change PL/Tcl build to use configured compiler and Makefile.shlib
565 Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible
567 Allow PL/pgSQL to handle quoted identifiers (Tom)
568 Allow set-returning PL/pgSQL functions (Neil)
569 Make PL/pgSQL schema-aware (Joe)
570 Remove some memory leaks (Nigel J. Andrews, Tom)
577 Don't lowercase psql \connect database name for 7.2.0 compatibility
579 Add psql \timing to time user queries (Greg Sabino Mullane)
580 Have psql \d show index information (Greg Sabino Mullane)
581 New psql \dD shows domains (Jonathan Eisler)
582 Allow psql to show rules on views (Paul ?)
583 Fix for psql variable substitution (Tom)
584 Allow psql \d to show temporary table structure (Tom)
585 Allow psql \d to show foreign keys (Rod)
586 Fix \? to honor \pset pager (Bruce)
587 Have psql reports its version number on startup (Tom)
588 Allow \copy to specify column names (Tom)
595 Add $HOME/.pgpass to store host/user password combinations (Alvaro
597 Add PQunescapeBytea() function to libpq (Patrick Welche)
598 Fix for sending large queries over non-blocking connections
600 Fix for libpq using timers on Win9X (David Ford)
601 Allow libpq notify to handle servers with different-length
603 Add libpq PQescapeString() and PQescapeBytea() to Win32 (Bruce)
604 Fix for SSL with non-blocking connections (Jack Bates)
605 Add libpq connection timeout parameter (Denis A Ustimenko)
612 Allow JDBC to compile with JDK 1.4 (Dave)
613 Add JDBC 3 support (Barry)
614 Allows JDBC to set loglevel by adding ?loglevel=X to the connection
616 Add Driver.info() message that prints out the version number (Barry)
617 Add updateable result sets (Raghu Nidagal, Dave)
618 Add support for callable statements (Paul Bethe)
619 Add query cancel capability
620 Add refresh row (Dave)
621 Fix MD5 encryption handling for multibyte servers (Jun Kawai)
622 Add support for prepared statements (Barry)
627 <title>Miscellaneous Interfaces</title>
629 Fixed ECPG bug concerning octal numbers in single quotes (Michael)
630 Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc,
632 Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)
633 Add libpgtcl connection close event (Gerhard Hintermayer)
634 Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc,
636 Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)
637 Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc,
639 Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc,
641 Remove src/bin/pgaccess from main tree, now at
642 http://www.pgaccess.org (Bruce)
643 Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer,
649 <title>Source Code</title>
651 Fix for parallel make (Peter)
652 AIX fixes for linking Tcl (Andreas Zeugswetter)
653 Allow PL/Perl to build under Cygwin (Jason Tishler)
654 Improve MIPS compiles (Peter, Oliver Elphick)
655 Require Autoconf version 2.53 (Peter)
656 Require readline and zlib by default in configure (Peter)
657 Allow Solaris to use Intimate Shared Memory (ISM), for performance
658 (Scott Brunza, P.J. Josh Rovero)
659 Always enable syslog in compile, remove --enable-syslog option
661 Always enable multibyte in compile, remove --enable-multibyte option
663 Always enable locale in compile, remove --enable-locale option
665 Fix for Win9x DLL creation (Magnus Naeslund)
666 Fix for link() usage by WAL code on Win32, BeOS (Jason Tishler)
667 Add sys/types.h to c.h, remove from main files (Peter, Bruce)
668 Fix AIX hang on SMP machines (Tomoyuki Niijima)
669 AIX SMP hang fix (Tomoyuki Niijima)
670 Fix pre-1970 date handling on newer glibc libraries (Tom)
671 Fix PowerPC SMP locking (Tom)
672 Prevent gcc -ffast-math from being used (Peter, Tom)
673 Bison >= 1.50 now required for developer builds
674 Kerberos 5 support now builds with Heimdal (Peter)
675 Add appendix in the User's Guide which lists SQL features (Thomas)
676 Improve loadable module linking to use RTLD_NOW (Tom)
677 New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)
678 New src/port directory holds replaced libc functions (Peter, Bruce)
679 New pg_namespace system catalog for schemas (Tom)
680 Add pg_class.relnamespace for schemas (Tom)
681 Add pg_type.typnamespace for schemas (Tom)
682 Add pg_proc.pronamespace for schemas (Tom)
683 Restructure aggregates to have pg_proc entries (Tom)
684 System relations now have their own namespace, pg_* test not required
686 Rename TOAST index names to be *_index rather than *_idx (Neil)
687 Add namespaces for operators, opclasses (Tom)
688 Add additional checks to server control file (Thomas)
689 New Polish FAQ (Marcin Mazurek)
690 Add Posix semaphore support (Tom)
691 Document need for reindex (Bruce)
692 Rename some internal identifiers to simplify Win32 compile (Jan,
694 Add documentation on computing disk space (Bruce)
695 Remove KSQO from GUC (Bruce)
696 Fix memory leak in rtree (Kenneth Been)
697 Modify a few error messages for consistency (Bruce)
698 Remove unused system table columns (Peter)
699 Make system columns NOT NULL where appropriate (Tom)
700 Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)
701 Remove OPAQUE and create specific subtypes (Tom)
702 Cleanups in array internal handling (Joe, Tom)
703 Disallow pg_atoi('') (Bruce)
704 Remove parameter wal_files because WAL files are now recycled (Bruce)
705 Add version numbers to heap pages (Tom)
710 <title>Contrib</title>
712 Allow inet arrays in /contrib/array (Neil)
713 Gist fixes (Teodor Sigaev, Neil)
714 Upgrade /contrib/mysql
715 Add /contrib/dbsize which shows table sizes without vacuum (Peter)
716 Add /contrib/intagg, integer aggregator routines (mlw)
717 Improve /contrib/oid2name (Neil, Bruce)
718 Improve /contrib/tsearch (Oleg, Teodor Sigaev)
719 Cleanups of /contrib/rserver (Alexey V. Borzov)
720 Update /contrib/oracle conversion utility (Gilles Darold)
721 Update /contrib/dblink (Joe)
722 Improve options supported by /contrib/vacuumlo (Mario Weilguni)
723 Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey
725 Add /contrib/reindexdb utility (Shaun Thomas)
726 Add indexing to /contrib/isbn_issn (Dan Weston)
727 Add /contrib/dbmirror (Steven Singer)
728 Improve /contrib/pgbench (Neil)
729 Add /contrib/tablefunc table function examples (Joe)
730 Add /contrib/ltree data type for tree structures (Teodor Sigaev,
732 Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)
733 Fixes to /contrib/cube (Bruno Wolff)
734 Improve /contrib/fulltextindex (Christopher)
742 <sect1 id="release-7-2-3">
743 <title>Release 7.2.3</title>
746 <title>Release date</title>
747 <simpara>2002-10-01</simpara>
751 This has a variety of fixes from 7.2.2, including fixes to prevent
757 <title>Migration to version 7.2.3</title>
760 A dump/restore is <emphasis>not</emphasis> required for those running
766 <title>Changes</title>
770 Prevent possible compressed transaction log loss (Tom)
771 Prevent non-superuser from increasing most recent vacuum info (Tom)
772 Handle pre-1970 date values in newer versions of glibc (Tom)
773 Fix possible hang during server shutdown
774 Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)
775 Fix pg_dump to properly dump FULL JOIN USING (Tom)
782 <sect1 id="release-7-2-2">
783 <title>Release 7.2.2</title>
786 <title>Release date</title>
787 <simpara>2002-08-23</simpara>
791 This has a variety of fixes from 7.2.1.
796 <title>Migration to version 7.2.2</title>
799 A dump/restore is <emphasis>not</emphasis> required for those running
805 <title>Changes</title>
809 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
810 Fix for compressed transaction log id wraparound (Tom)
811 Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Tatsuo)
812 Fix for psql and pg_dump crashing when invoked with non-existand long
814 Fix crash when invoking geometric operators (Tom)
815 Allow OPEN cursor(args) (Tom)
816 Fix for rtree_gist index build (Teodor)
817 Fix for dumping user-defined aggregates (Tom)
818 Contrib/intarray fixes (Oleg)
819 Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)
820 Fix to pg_convert (Tatsuo)
821 Fix for crash with long DATA strings (Thomes, Neil)
822 Fix for repeat(), lpad(), rpad() and long strings (Neil)
829 <sect1 id="release-7-2-1">
830 <title>Release 7.2.1</title>
833 <title>Release date</title>
834 <simpara>2002-03-21</simpara>
838 This has a variety of fixes from 7.2.
843 <title>Migration to version 7.2.1</title>
846 A dump/restore is <emphasis>not</emphasis> required for those running
852 <title>Changes</title>
856 Ensure that sequence counters do not go backwards after a crash (Tom)
857 Fix pgaccess kanji-coversion key binding (Tatsuo)
858 Optimizer improvements (Tom)
859 cash I/O improvements (Tom)
861 Compile fix for missing AuthBlockSig (Heiko)
862 Additional time zones and time zone fixes (Thomas)
863 Allow psql \connect to handle mixed case database and user names (Tom)
864 Return proper OID on command completion even with ON INSERT rules (Tom)
865 Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)
866 Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)
867 Improve handling of multiple UNIONs with different lengths (Tom)
868 contrib/btree_gist improvements (Teodor Sigaev)
869 contrib/tsearch dictionary improvements, see README.tsearch for
870 an additional installation step (Thomas T. Thai, Teodor Sigaev)
871 Fix for array subscripts handling (Tom)
872 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
879 <sect1 id="release-7-2">
880 <title>Release 7.2</title>
883 <title>Release date</title>
884 <simpara>2002-02-04</simpara>
888 <title>Overview</title>
891 This release improves <productname>PostgreSQL</> for use in
892 high-volume applications.
896 Major changes in this release:
903 Vacuuming no longer locks tables, thus allowing normal user
904 access during the vacuum. A new <command>VACUUM FULL</>
905 command does old-style vacuum by locking the table and
906 shrinking the on-disk copy of the table.
912 <term>Transactions</term>
915 There is no longer a problem with installations that exceed
916 four billion transactions.
925 OIDs are now optional. Users can now create tables without
926 OIDs for cases where OID usage is excessive.
932 <term>Optimizer</term>
935 The system now computes histogram column statistics during
936 <command>ANALYZE</>, allowing much better optimizer choices.
942 <term>Security</term>
945 A new MD5 encryption option allows more secure storage and
946 transfer of passwords. A new Unix-domain socket
947 authentication option is available on Linux and BSD systems.
953 <term>Statistics</term>
956 Administrators can use the new table access statistics module
957 to get fine-grained information about table and index usage.
963 <term>Internationalization</term>
966 Program and library messages can now be displayed in several
976 <title>Migration to version 7.2</title>
979 A dump/restore using <command>pg_dump</command> is required for
980 those wishing to migrate data from any previous release.
984 Observe the following incompatibilities:
989 The semantics of the <command>VACUUM</command> command have
990 changed in this release. You may wish to update your
991 maintenance procedures accordingly.
997 In this release, comparisons using <literal>= NULL</literal>
998 will always return false (or NULL, more precisely). Previous
999 releases automatically transformed this syntax to <literal>IS
1000 NULL</literal>. The old behavior can be re-enabled using a
1001 <filename>postgresql.conf</filename> parameter.
1007 The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
1008 configuration is now only reloaded after receiving a
1009 <systemitem>SIGHUP</> signal, not with each connection.
1015 The function <filename>octet_length()</> now returns the uncompressed data length.
1021 The date/time value <literal>'current'</literal> is no longer
1022 available. You will need to rewrite your applications.
1028 The <literal>timestamp()</literal>, <literal>time()</literal>,
1029 and <literal>interval()</literal> functions are no longer
1030 available. Instead of <literal>timestamp()</literal>, use
1031 <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
1041 The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
1042 in the next release. You should change your queries to use
1043 separate LIMIT and OFFSET clauses, e.g. <literal>LIMIT 10 OFFSET
1049 <title>Changes</title>
1052 <title>Server Operation</title>
1054 Create temporary files in a separate directory (Bruce)
1055 Delete orphaned temporary files on postmaster startup (Bruce)
1056 Added unique indexes to some system tables (Tom)
1057 System table operator reorganization (Oleg Bartunov, Teodor Sigaev, Tom)
1058 Renamed pg_log to pg_clog (Tom)
1059 Enable SIGTERM, SIGQUIT to kill backends (Jan)
1060 Removed compile-time limit on number of backends (Tom)
1061 Better cleanup for semaphore resource failure (Tatsuo, Tom)
1062 Allow safe transaction ID wraparound (Tom)
1063 Removed OIDs from some system tables (Tom)
1064 Removed "triggered data change violation" error check (Tom)
1065 SPI portal creation of prepared/saved plans (Jan)
1066 Allow SPI column functions to work for system columns (Tom)
1067 Long value compression improvement (Tom)
1068 Statistics collector for table, index access (Jan)
1069 Truncate extra-long sequence names to a reasonable value (Tom)
1070 Measure transaction times in milliseconds (Thomas)
1071 Fix TID sequential scans (Hiroshi)
1072 Superuser ID now fixed at 1 (Peter E)
1073 New pg_ctl "reload" option (Tom)
1078 <title>Performance</title>
1080 Optimizer improvements (Tom)
1081 New histogram column statistics for optimizer (Tom)
1082 Reuse write-ahead log files rather than discarding them (Tom)
1083 Cache improvements (Tom)
1084 IS NULL, IS NOT NULL optimizer improvement (Tom)
1085 Improve lock manager to reduce lock contention (Tom)
1086 Keep relcache entries for index access support functions (Tom)
1087 Allow better selectivity with NaN and infinities in NUMERIC (Tom)
1088 R-tree performance improvements (Kenneth Been)
1089 B-tree splits more efficient (Tom)
1094 <title>Privileges</title>
1096 Change UPDATE, DELETE permissions to be distinct (Peter E)
1097 New REFERENCES, TRIGGER privileges (Peter E)
1098 Allow GRANT/REVOKE to/from more than one user at a time (Peter E)
1099 New has_table_privilege() function (Joe Conway)
1100 Allow non-superuser to vacuum database (Tom)
1101 New SET SESSION AUTHORIZATION command (Peter E)
1102 Fix bug in privilege modifications on newly created tables (Tom)
1103 Disallow access to pg_statistic for non-superuser, add user-accessible views (Tom)
1108 <title>Client Authentication</title>
1110 Fork postmaster before doing authentication to prevent hangs (Peter E)
1111 Add ident authentication over Unix domain sockets on Linux, *BSD (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)
1112 Add a password authentication method that uses MD5 encryption (Bruce)
1113 Allow encryption of stored passwords using MD5 (Bruce)
1114 PAM authentication (Dominic J. Eidson)
1115 Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)
1120 <title>Server Configuration</title>
1122 Interpretation of some time zone abbreviations as Australian rather than North American now settable at run time (Bruce)
1123 New parameter to set default transaction isolation level (Peter E)
1124 New parameter to enable conversion of "expr = NULL" into "expr IS NULL", off by default (Peter E)
1125 New parameter to control memory usage by VACUUM (Tom)
1126 New parameter to set client authentication timeout (Tom)
1127 New parameter to set maximum number of open files (Tom)
1132 <title>Queries</title>
1134 Statements added by INSERT rules now execute after the INSERT (Jan)
1135 Prevent unadorned relation names in target list (Bruce)
1136 NULLs now sort after all normal values in ORDER BY (Tom)
1137 New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)
1138 New SHARE UPDATE EXCLUSIVE lock mode (Tom)
1139 New EXPLAIN ANALYZE command that shows run times and row counts (Martijn van Oosterhout)
1140 Fix problem with LIMIT and subqueries (Tom)
1141 Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)
1142 Fix nested EXCEPT/INTERSECT (Tom)
1147 <title>Schema Manipulation</title>
1149 Fix SERIAL in temporary tables (Bruce)
1150 Allow temporary sequences (Bruce)
1151 Sequences now use int8 internally (Tom)
1152 New SERIAL8 creates int8 columns with sequences, default still SERIAL4 (Tom)
1153 Make OIDs optional using WITHOUT OIDS (Tom)
1154 Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)
1155 Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher Kings-Lynne)
1156 New CREATE OR REPLACE FUNCTION to alter existing function (preserving the function OID) (Gavin Sherry)
1157 Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)
1158 Allow column renaming in views
1159 Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent Verner)
1160 Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (Stephan Szabo)
1161 ALTER TABLE RENAME update foreign-key trigger arguments correctly (Brent Verner)
1162 DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)
1163 Add automatic return type data casting for SQL functions (Tom)
1164 Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov, Teodor Sigaev, Tom)
1165 Enable partial indexes (Martijn van Oosterhout)
1170 <title>Utility Commands</title>
1172 Add RESET ALL, SHOW ALL (Marko Kreen)
1173 CREATE/ALTER USER/GROUP now allow options in any order (Vince)
1174 Add LOCK A, B, C functionality (Neil Padgett)
1175 New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)
1176 New light-weight VACUUM does not lock table; old semantics are available as VACUUM FULL (Tom)
1177 Disable COPY TO/FROM on views (Bruce)
1178 COPY DELIMITERS string must be exactly one character (Tom)
1179 VACUUM warning about index tuples fewer than heap now only appears when appropriate (Martijn van Oosterhout)
1180 Fix permission checks for CREATE INDEX (Tom)
1181 Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)
1186 <title>Data Types and Functions</title>
1188 SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)
1189 Add convert(), convert2() (Tatsuo)
1190 New function bit_length() (Peter E)
1191 Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tatsuo)
1192 CHAR(), VARCHAR() now reject strings that are too long (Peter E)
1193 BIT VARYING now rejects bit strings that are too long (Peter E)
1194 BIT now rejects bit strings that do not match declared size (Peter E)
1195 INET, CIDR text conversion functions (Alex Pilosov)
1196 INET, CIDR operators << and <<= indexable (Alex Pilosov)
1197 Bytea \### now requires valid three digit octal number
1198 Bytea comparison improvements, now supports =, <>, >, >=, <, and <=
1199 Bytea now supports B-tree indexes
1200 Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE
1201 Bytea now supports concatenation
1202 New bytea functions: position, substring, trim, btrim, and length
1203 New encode() function mode, "escaped", converts minimally escaped bytea to/from text
1204 Add pg_database_encoding_max_length() (Tatsuo)
1205 Add pg_client_encoding() function (Tatsuo)
1206 now() returns time with millisecond precision (Thomas)
1207 New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)
1208 Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)
1209 New xid/int comparison functions (Hiroshi)
1210 Add precision to TIME, TIMESTAMP, and INVERVAL data types (Thomas)
1211 Modify type coercion logic to attempt binary-compatible functions first (Tom)
1212 New encode() function installed by default (Marko Kreen)
1213 Improved to_*() conversion functions (Karel Zak)
1214 Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)
1215 New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_salt() (Marko Kreen)
1216 Correct description of translate() function (Bruce)
1217 Add INTERVAL argument for SET TIME ZONE (Thomas)
1218 Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)
1219 Optimize length functions when using single-byte encodings (Tatsuo)
1220 Fix path_inter, path_distance, path_length, dist_ppath to handle closed paths (Curtis Barrett, Tom)
1221 octet_length(text) now returns non-compressed length (Tatsuo, Bruce)
1222 Handle "July" full name in date/time literals (Greg Sabino Mullane)
1223 Some datatype() function calls now evaluated differently
1224 Add support for Julian and ISO time specifications (Thomas)
1229 <title>Internationalization</title>
1231 National language support in psql, pg_dump, libpq, and server (Peter E)
1232 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)
1233 Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware (Tatsuo)
1234 Add LATIN5,6,7,8,9,10 support (Tatsuo)
1235 Add ISO 8859-5,6,7,8 support (Tatsuo)
1236 Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)
1237 Make mic2ascii() non-ASCII aware (Tatsuo)
1238 Reject invalid multibyte character sequences (Tatsuo)
1243 <title><application>PL/pgSQL</></title>
1245 Now uses portals for SELECT loops, allowing huge result sets (Jan)
1246 CURSOR and REFCURSOR support (Jan)
1247 Can now return open cursors (Jan)
1248 Add ELSEIF (Klaus Reger)
1249 Improve PL/pgSQL error reporting, including location of error (Tom)
1250 Allow IS or FOR key words in cursor declaration, for compatibility (Bruce)
1251 Fix for SELECT ... FOR UPDATE (Tom)
1252 Fix for PERFORM returning multiple rows (Tom)
1253 Make PL/pgSQL use the server's type coercion code (Tom)
1254 Memory leak fix (Jan, Tom)
1255 Make trailing semicolon optional (Tom)
1260 <title>PL/Perl</title>
1262 New untrusted PL/Perl (Alex Pilosov)
1263 PL/Perl is now built on some platforms even if libperl is not shared (Peter E)
1268 <title>PL/Tcl</title>
1270 Now reports errorInfo (Vsevolod Lobko)
1271 Add spi_lastoid function (bob@redivi.com)
1276 <title>PL/Python</title>
1278 ...is new (Andrew Bosma)
1283 <title><application>Psql</></title>
1285 \d displays indexes in unique, primary groupings (Christopher Kings-Lynne)
1286 Allow trailing semicolons in backslash commands (Greg Sabino Mullane)
1287 Read password from /dev/tty if possible
1288 Force new password prompt when changing user and database (Tatsuo, Tom)
1289 Format the correct number of columns for Unicode (Patrice)
1294 <title><application>Libpq</></title>
1296 New function PQescapeString() to escape quotes in command strings (Florian Weimer)
1297 New function PQescapeBytea() escapes binary strings for use as SQL string literals
1304 Return OID of INSERT (Ken K)
1305 Handle more data types (Ken K)
1306 Handle single quotes and newlines in strings (Ken K)
1307 Handle NULL variables (Ken K)
1308 Fix for time zone handling (Barry Lind)
1309 Improved Druid support
1310 Allow eight-bit characters with non-multibyte server (Barry Lind)
1311 Support BIT, BINARY types (Ned Wolpert)
1312 Reduce memory usage (Michael Stephens, Dave Cramer)
1313 Update DatabaseMetaData (Peter E)
1314 Add DatabaseMetaData.getCatalogs() (Peter E)
1315 Encoding fixes (Anders Bengtsson)
1316 Get/setCatalog methods (Jason Davies)
1317 DatabaseMetaData.getColumns() now returns column defaults (Jason Davies)
1318 DatabaseMetaData.getColumns() performance improvement (Jeroen van Vianen)
1319 Some JDBC1 and JDBC2 merging (Anders Bengtsson)
1320 Transaction performance improvements (Barry Lind)
1321 Array fixes (Greg Zoller)
1323 Fix batch processing (Rene Pijlman)
1324 ExecSQL method reorganization (Anders Bengtsson)
1325 GetColumn() fixes (Jeroen van Vianen)
1326 Fix isWriteable() function (Rene Pijlman)
1327 Improved passage of JDBC2 conformance tests (Rene Pijlman)
1328 Add bytea type capability (Barry Lind)
1329 Add isNullable() (Rene Pijlman)
1330 JDBC date/time test suite fixes (Liam Stewart)
1331 Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)
1332 Fix DatabaseMetaData to show precision properly (Mark Lillywhite)
1333 New getImported/getExported keys (Jason Davies)
1334 MD5 password encryption support (Jeremy Wohl)
1335 Fix to actually use type cache (Ned Wolpert)
1342 Remove query size limit (Hiroshi)
1343 Remove text field size limit (Hiroshi)
1344 Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)
1345 Allow ODBC procedure calls (Hiroshi)
1346 Improve boolean handing (Aidan Mountford)
1347 Most configuration options on setable via DSN (Hiroshi)
1348 Multibyte, performance fixes (Hiroshi)
1349 Allow driver to be used with iODBC or unixODBC (Peter E)
1350 MD5 password encryption support (Bruce)
1351 Add more compatibility functions to odbc.sql (Peter E)
1356 <title><application>ECPG</></title>
1358 EXECUTE ... INTO implemented (Christof Petig)
1359 Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)
1360 Fix for GRANT parameters (Lee Kindness)
1361 Fix INITIALLY DEFERRED bug
1362 Various bug fixes (Michael, Christof Petig)
1363 Auto allocation for indicator variable arrays (int *ind_p=NULL)
1364 Auto allocation for string arrays (char **foo_pp=NULL)
1365 ECPGfree_auto_mem fixed
1366 All function names with external linkage are now prefixed by ECPG
1367 Fixes for arrays of structures (Michael)
1372 <title>Misc. Interfaces</title>
1374 Python fix fetchone() (Gerhard Haring)
1375 Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Max)
1376 Add Tcl COPY TO/FROM (ljb)
1377 Prevent output of default index op class in pg_dump (Tom)
1378 Fix libpgeasy memory leak (Bruce)
1383 <title>Build and Install</title>
1385 Configure, dynamic loader, and shared library fixes (Peter E)
1386 Fixes in QNX 4 port (Bernd Tegge)
1387 Fixes in Cygwin and Win32 ports (Jason Tishler, Gerhard Haring, Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)
1388 Fix for Win32 socket communication failures (Magnus, Mikhail Terekhov)
1389 Hurd compile fix (Oliver Elphick)
1390 BeOS fixes (Cyril Velter)
1391 Remove configure --enable-unicode-conversion, now enabled by multibyte (Tatsuo)
1392 AIX fixes (Tatsuo, Andreas)
1393 Fix parallel make (Peter E)
1394 Install SQL language manual pages into OS-specific directories (Peter E)
1395 Rename config.h to pg_config.h (Peter E)
1396 Reorganize installation layout of header files (Peter E)
1401 <title>Source Code</title>
1403 Remove SEP_CHAR (Bruce)
1405 Merge GUC and command line handling (Marko Kreen)
1406 Remove EXTEND INDEX (Martijn van Oosterhout, Tom)
1407 New pgjindent utility to indent java code (Bruce)
1408 Remove define of true/false when compiling under C++ (Leandro Fanzone, Tom)
1409 pgindent fixes (Bruce, Tom)
1410 Replace strcasecmp() with strcmp() where appropriate (Peter E)
1411 Dynahash portability improvements (Tom)
1412 Add 'volatile' usage in spinlock structures
1413 Improve signal handling logic (Tom)
1418 <title>Contrib</title>
1420 New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)
1421 New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)
1422 Add contrib/dblink for remote database access (Joe Conway)
1423 contrib/ora2pg Oracle conversion utility (Gilles Darold)
1424 contrib/xml XML conversion utility (John Gray)
1425 contrib/fulltextindex fixes (Christopher Kings-Lynne)
1426 New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merged (Joe Conway)
1427 Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartunov)
1428 New pg_upgrade utility (Bruce)
1429 Add new pg_resetxlog options (Bruce, Tom)
1436 <sect1 id="release-7-1-3">
1437 <title>Release 7.1.3</title>
1440 <title>Release date</title>
1441 <simpara>2001-08-15</simpara>
1445 <title>Migration to version 7.1.3</title>
1448 A dump/restore is <emphasis>not</emphasis> required for those running
1454 <title>Changes</title>
1458 Remove unused WAL segements of large transactions (Tom)
1459 Multiaction rule fix (Tom)
1460 PL/pgSQL memory allocation fix (Jan)
1461 VACUUM buffer fix (Tom)
1462 Regression test fixes (Tom)
1463 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
1464 Fix subselects with DISTINCT ON or LIMIT (Tom)
1466 Disable COPY TO/FROM a view (Tom)
1467 Cygwin build (Jason Tishler)
1474 <sect1 id="release-7-1-2">
1475 <title>Release 7.1.2</title>
1478 <title>Release date</title>
1479 <simpara>2001-05-11</simpara>
1483 This has one fix from 7.1.1.
1488 <title>Migration to version 7.1.2</title>
1491 A dump/restore is <emphasis>not</emphasis> required for those running
1497 <title>Changes</title>
1501 Fix PL/pgSQL SELECTs when returning no rows
1502 Fix for psql backslash core dump
1503 Referential integrity permission fix
1512 <sect1 id="release-7-1-1">
1513 <title>Release 7.1.1</title>
1516 <title>Release date</title>
1517 <simpara>2001-05-05</simpara>
1521 This has a variety of fixes from 7.1.
1526 <title>Migration to version 7.1.1</title>
1529 A dump/restore is <emphasis>not</emphasis> required for those running
1535 <title>Changes</title>
1539 Fix for numeric MODULO operator (Tom)
1540 pg_dump fixes (Philip)
1541 pg_dump can dump 7.0 databases (Philip)
1542 readline 4.2 fixes (Peter E)
1544 AIX, MSWIN, VAX, N32K fixes (Tom)
1545 Multibytes fixes (Tom)
1546 Unicode fixes (Tatsuo)
1547 Optimizer improvements (Tom)
1548 Fix for whole tuples in functions (Tom)
1549 Fix for pg_ctl and option strings with spaces (Peter E)
1550 ODBC fixes (Hiroshi)
1551 EXTRACT can now take string argument (Thomas)
1552 Python fixes (Darcy)
1559 <sect1 id="release-7-1">
1560 <title>Release 7.1</title>
1563 <title>Release date</title>
1564 <simpara>2001-04-13</simpara>
1568 This release focuses on removing limitations that have existed in the
1569 PostgreSQL code for many years.
1573 Major changes in this release:
1578 Write-ahead Log (WAL)
1582 To maintain database consistency in case
1583 of an operating system crash, previous releases of PostgreSQL have
1584 forced all data modifications to disk before each transaction commit.
1585 With WAL, only one log file must be flushed to disk, greatly improving
1586 performance. If you have been using -F in previous releases to disable
1587 disk flushes, you may want to consider discontinuing its use.
1598 TOAST - Previous releases had a compiled-in row length limit,
1599 typically 8k - 32k. This limit made storage of long text fields
1600 difficult. With TOAST, long rows of any length can be stored with good
1612 We now support outer joins. The UNION/NOT IN
1613 workaround for outer joins is no longer required. We use the SQL92
1625 The previous C function manager did not
1626 handle null values properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
1627 function manager does. You can continue using your old custom
1628 functions, but you may want to rewrite them in the future to use the new
1629 function manager call interface.
1640 A large number of complex queries that were
1641 unsupported in previous releases now work. Many combinations of views,
1642 aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
1643 now work properly. Inherited tables are now accessed by default.
1644 Subqueries in FROM are now supported.
1653 <title>Migration to version 7.1</title>
1656 A dump/restore using pg_dump is required for those wishing to migrate
1657 data from any previous release.
1662 <title>Changes</title>
1668 Many multibyte/Unicode/locale fixes (Tatsuo and others)
1669 More reliable ALTER TABLE RENAME (Tom)
1670 Kerberos V fixes (David Wragg)
1671 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
1672 Prompt username/password on standard error (Bruce)
1673 Large objects inv_read/inv_write fixes (Tom)
1674 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
1676 Prevent query expressions from leaking memory (Tom)
1677 Allow UPDATE of arrays elements (Tom)
1678 Wake up lock waiters during cancel (Hiroshi)
1679 Fix rare cursor crash when using hash join (Tom)
1680 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
1681 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
1682 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
1683 Fix PL/perl (Alex Kapranoff)
1684 Disallow LOCK on views (Mark Hollomon)
1685 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
1686 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
1687 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
1688 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
1689 Fix rare failure with TRUNCATE command (Tom)
1690 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
1691 DISTINCT, ORDER BY, SELECT...INTO (Tom)
1692 Fix parser failures during aborted transactions (Tom)
1693 Allow temporary relations to properly clean up indexes (Bruce)
1694 Fix VACUUM problem with moving rows in same page (Tom)
1695 Modify pg_dump to better handle user-defined items in template1 (Philip)
1696 Allow LIMIT in VIEW (Tom)
1697 Require cursor FETCH to honor LIMIT (Tom)
1698 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
1699 Allow ORDER BY, LIMIT in sub-selects (Tom)
1700 Allow UNION in CREATE RULE (Tom)
1701 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
1702 Store initdb collation in pg_control so collation cannot be changed (Tom)
1703 Fix INSERT...SELECT with rules (Tom)
1704 Fix FOR UPDATE inside views and subselects (Tom)
1705 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
1706 Fix lpad() and rpad() to handle length less than input string (Tom)
1707 Fix use of NOTIFY in some rules (Tom)
1708 Overhaul btree code (Tom)
1709 Fix NOT NULL use in Pl/pgSQL variables (Tom)
1710 Overhaul GIST code (Oleg)
1711 Fix CLUSTER to preserve constraints and column default (Tom)
1712 Improved deadlock detection handling (Tom)
1713 Allow multiple SERIAL columns in a table (Tom)
1714 Prevent occasional index corruption (Vadim)
1718 Add OUTER JOINs (Tom)
1719 Function manager overhaul (Tom)
1720 Allow ALTER TABLE RENAME on indexes (Tom)
1721 Improve CLUSTER (Tom)
1722 Improve ps status display for more platforms (Peter E, Marc)
1723 Improve CREATE FUNCTION failure message (Ross)
1724 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
1726 Grand Unified Configuration scheme/GUC. Many options can now be set in
1727 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
1728 Improved handling of file descriptor cache (Tom)
1729 New warning code about auto-created table alias entries (Bruce)
1730 Overhaul initdb process (Tom, Peter E)
1731 Overhaul of inherited tables; inherited tables now accessed by default;
1732 new ONLY keyword prevents it (Chris Bitmead, Tom)
1733 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
1735 Allow renaming of temp tables (Tom)
1736 Overhaul memory manager contexts (Tom)
1737 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
1738 Overhaul pg_dump (Philip Warner)
1739 Allow pg_hba.conf secondary password file to specify only username (Peter E)
1740 Allow TEMPORARY or TEMP keyword when creating temporary tables (Bruce)
1741 New memory leak checker (Karel)
1742 New SET SESSION CHARACTERISTICS (Thomas)
1743 Allow nested block comments (Thomas)
1744 Add WITHOUT TIME ZONE type qualifier (Thomas)
1745 New ALTER TABLE ADD CONSTRAINT (Stephan)
1746 Use NUMERIC accumulators for INTEGER aggregates (Tom)
1747 Overhaul aggregate code (Tom)
1748 New VARIANCE and STDDEV() aggregates
1749 Improve dependency ordering of pg_dump (Philip)
1750 New pg_restore command (Philip)
1751 New pg_dump tar output option (Philip)
1752 New pg_dump of large objects (Philip)
1753 New ESCAPE option to LIKE (Thomas)
1754 New case-insensitive LIKE - ILIKE (Thomas)
1755 Allow functional indexes to use binary-compatible type (Tom)
1756 Allow SQL functions to be used in more contexts (Tom)
1757 New pg_config utility (Peter E)
1758 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
1760 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
1761 New quote_identifiers() and quote_literal() functions (Jan)
1762 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
1763 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
1764 Update PyGreSQL to version 3.1 (D'Arcy)
1765 Store tables as files named by OID (Vadim)
1766 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
1767 Require DROP VIEW to remove views, no DROP TABLE (Mark)
1768 Allow DROP VIEW view1, view2 (Mark)
1769 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
1770 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
1771 New /contrib/pgcrypto hashing functions (Marko Kreen)
1772 New pg_dumpall --globals-only option (Peter E)
1773 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
1774 New AT TIME ZONE syntax (Thomas)
1775 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
1776 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
1777 Allow socket path name to be specified in hostname by using leading slash
1778 (David J. MacKenzie)
1779 Allow CREATE DATABASE to specify template database (Tom)
1780 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
1781 New /contrib/rserv replication toolkit (Vadim)
1782 New file format for COPY BINARY (Tom)
1783 New /contrib/oid2name to map numeric files to table names (B Palmer)
1784 New "idle in transaction" ps status message (Marc)
1785 Update to pgaccess 0.98.7 (Constantin Teodorescu)
1786 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
1787 Add rudimentary dependency checking to pg_dump (Philip)
1791 Fix INET/CIDR type ordering and add new functions (Tom)
1792 Make OID behave as an unsigned type (Tom)
1793 Allow BIGINT as synonym for INT8 (Peter E)
1794 New int2 and int8 comparison operators (Tom)
1795 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
1796 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
1797 New GIST seg/cube examples (Gene Selkov)
1798 Improved round(numeric) handling (Tom)
1799 Fix CIDR output formatting (Tom)
1800 New CIDR abbrev() function (Tom)
1804 Write-Ahead Log (WAL) to provide crash recovery with less performance
1806 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
1807 Reduced file seeks (Denis Perchine)
1808 Improve BTREE code for duplicate keys (Tom)
1809 Store all large objects in a single table (Denis Perchine, Tom)
1810 Improve memory allocation performance (Karel, Tom)
1814 New function manager call conventions (Tom)
1815 SGI portability fixes (David Kaelbling)
1816 New configure --enable-syslog option (Peter E)
1817 New BSDI README (Bruce)
1818 configure script moved to top level, not /src (Peter E)
1819 Makefile/configuration/compilation overhaul (Peter E)
1820 New configure --with-python option (Peter E)
1821 Solaris cleanups (Peter E)
1822 Overhaul /contrib Makefiles (Karel)
1823 New OpenSSL configuration option (Magnus, Peter E)
1825 QNX fixes (Maurizio)
1826 New heap_open(), heap_openr() API (Tom)
1827 Remove colon and semi-colon operators (Thomas)
1828 New pg_class.relkind value for views (Mark Hollomon)
1829 Rename ichar() to chr() (Karel)
1830 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
1831 Fixes for NT/Cygwin (Pete Forman)
1832 AIX port fixes (Andreas)
1833 New BeOS port (David Reid, Cyril Velter)
1834 Add proofreader's changes to docs (Addison-Wesley, Bruce)
1835 New Alpha spinlock code (Adriaan Joubert, Compaq)
1836 UnixWare port overhaul (Peter E)
1837 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
1838 New FreeBSD Alpha port (Alfred)
1839 Overhaul shared memory segments (Tom)
1840 Add IBM S/390 support (Neale Ferguson)
1841 Moved macmanuf to /contrib (Larry Rosenman)
1842 Syslog improvements (Larry Rosenman)
1843 New template0 database that contains no user additions (Tom)
1844 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
1845 Allow NetBSD's libedit instead of readline (Peter)
1846 Improved assembly language source code format (Bruce)
1847 New contrib/pg_logger
1848 New --template option to createdb
1849 New contrib/pg_control utility (Oliver)
1850 New FreeBSD tools ipc_check, start-scripts/freebsd
1857 <sect1 id="release-7-0-3">
1858 <title>Release 7.0.3</title>
1861 <title>Release date</title>
1862 <simpara>2000-11-11</simpara>
1866 This has a variety of fixes from 7.0.2.
1871 <title>Migration to version 7.0.3</title>
1874 A dump/restore is <emphasis>not</emphasis> required for those running
1880 <title>Changes</title>
1885 Large object fix (Tom)
1886 Fix lean in COPY WITH OIDS leak (Tom)
1887 Fix backwards-index-scan (Tom)
1888 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
1889 Add --enable-syslog to configure (Marc)
1890 Fix abort transaction at backend exit in rare cases (Tom)
1891 Fix for psql \l+ when multibyte enabled (Tatsuo)
1892 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
1893 Make vacuum always flush buffers (Tom)
1894 Fix to allow cancel while waiting for a lock (Hiroshi)
1895 Fix for memory aloocation problem in user authentication code (Tom)
1896 Remove bogus use of int4out() (Tom)
1897 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
1898 Fix for failure of triggers on heap open in certain cases (Jeroen van
1900 Fix for erroneous selectivity of not-equals (Tom)
1901 Fix for erroneous use of strcmp() (Tom)
1902 Fix for bug where storage manager accesses items beyond end of file
1904 Fix to include kernel errno message in all smgr elog messages (Tom)
1905 Fix for '.' not in PATH at build time (SL Baur)
1906 Fix for out-of-file-descriptors error (Tom)
1907 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
1908 Fix for subselect in targetlist of Append node (Tom)
1909 Fix for mergejoin plans (Tom)
1910 Fix TRUNCATE failure on relations with indexes (Tom)
1911 Avoid database-wide restart on write error (Hiroshi)
1912 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
1913 Fix VACUUM problem with moving chain of update tuples when source and
1914 destination of a tuple lie on the same page (Tom)
1915 Fix user.c CommandCounterIncrement (Tom)
1916 Fix for AM/PM boundary problem in to_char() (Karel Zak)
1917 Fix TIME aggregate handling (Tom)
1918 Fix to_char() to avoid coredump on NULL input (Tom)
1920 Fix for inserting/copying longer multibyte strings into char() data
1922 Fix for crash of backend, on abort (Tom)
1929 <sect1 id="release-7-0-2">
1930 <title>Release 7.0.2</title>
1933 <title>Release date</title>
1934 <simpara>2000-06-05</simpara>
1938 This is a repackaging of 7.0.1 with added documentation.
1943 <title>Migration to version 7.0.2</title>
1946 A dump/restore is <emphasis>not</emphasis> required for those running
1952 <title>Changes</title>
1956 Added documentation to tarball.
1963 <sect1 id="release-7-0-1">
1964 <title>Release 7.0.1</title>
1967 <title>Release date</title>
1968 <simpara>2000-06-01</simpara>
1972 This is a cleanup release for 7.0.
1976 <title>Migration to version 7.0.1</title>
1979 A dump/restore is <emphasis>not</emphasis> required for those running
1985 <title>Changes</title>
1989 Fix many CLUSTER failures (Tom)
1990 Allow ALTER TABLE RENAME works on indexes (Tom)
1991 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
1992 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
1993 Fix the off by one errors in ResultSet from 6.5.3, and more.
1994 jdbc ResultSet fixes (Joseph Shraibman)
1995 optimizer tunings (Tom)
1996 Fix create user for pgaccess
1997 Fix for UNLISTEN failure
1998 IRIX fixes (David Kaelbling)
1999 QNX fixes (Andreas Kardos)
2000 Reduce COPY IN lock level (Tom)
2001 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
2002 Fix pg_dump to handle OID indexes (Tom)
2003 Fix small memory leak (Tom)
2004 Solaris fix for createdb/dropdb (Tatsuo)
2005 Fix for non-blocking connections (Alfred Perlstein)
2006 Fix improper recovery after RENAME TABLE failures (Tom)
2007 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
2008 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
2009 Fix too long syslog message (Tatsuo)
2010 Fix problem with quoted indexes that are too long (Tom)
2011 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
2012 ecpg changes (Michael)
2018 <sect1 id="release-7-0">
2019 <title>Release 7.0</title>
2022 <title>Release date</title>
2023 <simpara>2000-05-08</simpara>
2027 This release contains improvements in many areas, demonstrating
2028 the continued growth of <productname>PostgreSQL</productname>.
2029 There are more improvements and fixes in 7.0 than in any previous
2030 release. The developers have confidence that this is the best
2031 release yet; we do our best to put out only solid releases, and
2032 this one is no exception.
2036 Major changes in this release:
2045 Foreign keys are now implemented, with the exception of PARTIAL MATCH
2046 foreign keys. Many users have been asking for this feature, and we are
2047 pleased to offer it.
2058 Continuing on work started a year ago, the optimizer has been
2059 improved, allowing better query plan selection and faster performance
2060 with less memory usage.
2067 Updated <application>psql</application>
2071 <application>psql</application>, our interactive terminal monitor, has been
2072 updated with a variety of new features. See the <application>psql</application> manual page for details.
2083 SQL92 join syntax is now supported, though only as
2084 <literal>INNER JOIN</> for this release. <literal>JOIN</>,
2085 <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
2086 and <literal>JOIN</>/<literal>ON</> are available, as are
2087 column correlation names.
2099 In 7.1 or 7.2, we plan to have outer joins, storage for very long rows, and a
2100 write-ahead logging system.
2109 <title>Migration to version 7.0</title>
2112 A dump/restore using <application>pg_dump</application>
2113 is required for those wishing to migrate data from any
2114 previous release of <productname>PostgreSQL</productname>.
2115 For those upgrading from 6.5.*, you may instead use
2116 <application>pg_upgrade</application> to upgrade to this
2117 release; however, a full dump/reload installation is always the
2118 most robust method for upgrades.
2122 Interface and compatibility issues to consider for the new
2128 The date/time types <type>datetime</type> and
2129 <type>timespan</type> have been superseded by the
2130 SQL92-defined types <type>timestamp</type> and
2131 <type>interval</type>. Although there has been some effort to
2132 ease the transition by allowing
2133 <productname>PostgreSQL</productname> to recognize
2134 the deprecated type names and translate them to the new type
2135 names, this mechanism may not be completely transparent to
2136 your existing application.
2142 The optimizer has been substantially improved in the area of
2143 query cost estimation. In some cases, this will result in
2144 decreased query times as the optimizer makes a better choice
2145 for the preferred plan. However, in a small number of cases,
2146 usually involving pathological distributions of data, your
2147 query times may go up. If you are dealing with large amounts
2148 of data, you may want to check your queries to verify
2155 The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
2156 interfaces have been upgraded and extended.
2162 The string function <function>CHAR_LENGTH</function> is now a
2163 native function. Previous versions translated this into a call
2164 to <function>LENGTH</function>, which could result in
2165 ambiguity with other types implementing
2166 <function>LENGTH</function> such as the geometric types.
2175 <title>Changes</title>
2181 Prevent function calls exceeding maximum number of arguments (Tom)
2182 Improve CASE construct (Tom)
2183 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
2184 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
2185 Fix GROUP BY scan bug (Tom)
2186 Improvements in SQL grammar processing (Tom)
2187 Fix for views involved in INSERT ... SELECT ... (Tom)
2188 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
2189 Fix for subselects in INSERT ... SELECT (Tom)
2190 Prevent INSERT ... SELECT ... ORDER BY (Tom)
2191 Fixes for relations greater than 2GB, including vacuum
2192 Improve propagating system table changes to other backends (Tom)
2193 Improve propagating user table changes to other backends (Tom)
2194 Fix handling of temp tables in complex situations (Bruce, Tom)
2195 Allow table locking at table open, improving concurrent reliability (Tom)
2196 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
2197 Prevent DROP DATABASE while others accessing
2198 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
2199 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
2200 Fix pg_upgrade so it works for MVCC (Tom)
2201 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
2202 Fix for "f1 datetime DEFAULT 'now'" (Tom)
2203 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
2204 Allow comment-only lines, and ;;; lines too. (Tom)
2205 Improve recovery after failed disk writes, disk full (Hiroshi)
2206 Fix cases where table is mentioned in FROM but not joined (Tom)
2207 Allow HAVING clause without aggregate functions (Tom)
2208 Fix for "--" comment and no trailing newline, as seen in perl interface
2209 Improve pg_dump failure error reports (Bruce)
2210 Allow sorts and hashes to exceed 2GB file sizes (Tom)
2211 Fix for pg_dump dumping of inherited rules (Tom)
2212 Fix for NULL handling comparisons (Tom)
2213 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
2214 Fix for dbname with dash
2215 Prevent DROP INDEX from interfering with other backends (Tom)
2216 Fix file descriptor leak in verify_password()
2217 Fix for "Unable to identify an operator =$" problem
2218 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
2219 Fix for recursive exit call (Massimo)
2220 Fix for extra-long timezones (Jeroen van Vianen)
2221 Make pg_dump preserve primary key information (Peter E)
2222 Prevent databases with single quotes (Peter E)
2223 Prevent DROP DATABASE inside transaction (Peter E)
2224 ecpg memory leak fixes (Stephen Birch)
2225 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
2226 Y2K timestamp fix (Massimo)
2227 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
2228 Fix for views with tables/columns containing spaces (Tom)
2229 Prevent permissions on indexes (Peter E)
2230 Fix for spinlock stuck problem when error is generated (Hiroshi)
2231 Fix ipcclean on Linux
2232 Fix handling of NULL constraint conditions (Tom)
2233 Fix memory leak in odbc driver (Nick Gorham)
2234 Fix for permission check on UNION tables (Tom)
2235 Fix to allow SELECT 'a' LIKE 'a' (Tom)
2236 Fix for SELECT 1 + NULL (Tom)
2238 Fix log() on numeric type (Tom)
2239 Deprecate ':' and ';' operators
2240 Allow vacuum of temporary tables
2241 Disallow inherited columns with the same name as new columns
2242 Recover or force failure when disk space is exhausted (Hiroshi)
2243 Fix INSERT INTO ... SELECT with AS columns matching result columns
2244 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns (Tom)
2245 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
2246 Fix UNION with LIMIT
2247 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
2248 Fix CREATE TABLE test(col char(2) DEFAULT user)
2249 Fix mismatched types in CREATE TABLE ... DEFAULT
2250 Fix SELECT * FROM pg_class where oid in (0,-1)
2251 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
2252 Prevent user who can create databases can modifying pg_database table (Peter E)
2253 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
2254 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
2258 New CLI interface include file sqlcli.h, based on SQL3/SQL98
2259 Remove all limits on query length, row length limit still exists (Tom)
2260 Update jdbc protocol to 2.0 (Jens Glaser <email>jens@jens.de</email>)
2261 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
2262 Fix to give super user and createdb user proper update catalog rights (Peter E)
2263 Allow ecpg bool variables to have NULL values (Christof)
2264 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
2265 Allow ^C to cancel COPY command (Massimo)
2266 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
2267 Function name overloading for dynamically-loaded C functions (Frankpitt)
2268 Add CmdTuples() to libpq++(Vince)
2269 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
2270 Allow CREATE FUNCTION/WITH clause to be used for all language types
2271 configure --enable-debug adds -g (Peter E)
2272 configure --disable-debug removes -g (Peter E)
2273 Allow more complex default expressions (Tom)
2274 First real FOREIGN KEY constraint trigger functionality (Jan)
2275 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
2276 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
2277 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
2278 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
2279 Change pgeasy connectdb() parameter ordering (Bruce)
2280 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
2281 Add Oracle's COMMENT ON command (Mike Mascari <email>mascarim@yahoo.com</email>)
2282 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
2283 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
2284 Make USING in COPY optional (Bruce)
2285 Allow subselects in the target list (Tom)
2286 Allow subselects on the left side of comparison operators (Tom)
2287 New parallel regression test (Jan)
2288 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
2289 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
2290 Added psql LASTOID variable to return last inserted oid (Peter E)
2291 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
2292 Add permissions check for vacuum (Peter E)
2293 New libpq functions to allow asynchronous connections: PQconnectStart(),
2294 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
2295 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
2296 New libpq PQsetenv() function (Ewan Mellor)
2297 create/alter user extension (Peter E)
2298 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
2299 New scripts for create/drop user/db (Peter E)
2300 Major psql overhaul (Peter E)
2301 Add const to libpq interface (Peter E)
2302 New libpq function PQoidValue (Peter E)
2303 Show specific non-aggregate causing problem with GROUP BY (Tom)
2304 Make changes to pg_shadow recreate pg_pwd file (Peter E)
2305 Add aggregate(DISTINCT ...) (Tom)
2306 Allow flag to control COPY input/output of NULLs (Peter E)
2307 Make postgres user have a password by default (Peter E)
2308 Add CREATE/ALTER/DROP GROUP (Peter E)
2309 All administration scripts now support --long options (Peter E, Karel)
2310 Vacuumdb script now supports --all option (Peter E)
2311 ecpg new portable FETCH syntax
2312 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
2313 and EXEC SQL ENDIF directives
2314 Add pg_ctl script to control backend start-up (Tatsuo)
2315 Add postmaster.opts.default file to store start-up flags (Tatsuo)
2316 Allow --with-mb=SQL_ASCII
2317 Increase maximum number of index keys to 16 (Bruce)
2318 Increase maximum number of function arguments to 16 (Bruce)
2319 Allow configuration of maximum number of index keys and arguments (Bruce)
2320 Allow unprivileged users to change their passwords (Peter E)
2321 Password authentication enabled; required for new users (Peter E)
2322 Disallow dropping a user who owns a database (Peter E)
2323 Change initdb option --with-mb to --enable-multibyte
2324 Add option for initdb to prompts for superuser password (Peter E)
2325 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
2326 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
2327 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
2328 Libpq non-blocking mode (Alfred Perlstein)
2329 Improve conversion of types in casts that don't specify a length
2330 New plperl internal programming language (Mark Hollomon)
2331 Allow COPY IN to read file that do not end with a newline (Tom)
2332 Indicate when long identifiers are truncated (Tom)
2333 Allow aggregates to use type equivalency (Peter E)
2334 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
2335 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
2336 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
2337 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
2338 Add NUMERIC and int8 types to ODBC
2339 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
2340 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
2341 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
2342 Enable backward sequential scan even after reaching EOF (Hiroshi)
2343 Add btree indexing of boolean values, >= and <= (Don Baccus)
2344 Print current line number when COPY FROM fails (Massimo)
2345 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
2346 Add DEC as synonym for DECIMAL (Thomas)
2347 Add SESSION_USER as SQL92 keyword, same as CURRENT_USER (Thomas)
2348 Implement SQL92 column aliases (aka correlation names) (Thomas)
2349 Implement SQL92 join syntax (Thomas)
2350 Make INTERVAL reserved word allowed as a column identifier (Thomas)
2351 Implement REINDEX command (Hiroshi)
2352 Accept ALL in aggregate function SUM(ALL col) (Tom)
2353 Prevent GROUP BY from using column aliases (Tom)
2354 New psql \encoding option (Tatsuo)
2355 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
2356 Allow negation of a negative number in all cases
2357 Add ecpg descriptors (Christof, Michael)
2358 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
2359 Allow casts with length, like foo::char(8)
2360 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
2361 Add support for SJIS user defined characters (Tatsuo)
2362 Larger views/rules supported
2363 Make libpq's PQconndefaults() thread-safe (Tom)
2364 Disable // as comment to be ANSI conforming, should use -- (Tom)
2365 Allow column aliases on views CREATE VIEW name (collist)
2366 Fixes for views with subqueries (Tom)
2367 Allow UPDATE table SET fld = (SELECT ...) (Tom)
2368 SET command options no longer require quotes
2369 Update pgaccess to 0.98.6
2370 New SET SEED command
2371 New pg_options.sample file
2372 New SET FSYNC command (Massimo)
2373 Allow pg_descriptions when creating tables
2374 Allow pg_descriptions when creating types, columns, and functions
2375 Allow psql \copy to allow delimiters (Peter E)
2376 Allow psql to print nulls as distinct from "" [null] (Peter E)
2380 Many array fixes (Tom)
2381 Allow bare column names to be subscripted as arrays (Tom)
2382 Improve type casting of int and float constants (Tom)
2383 Cleanups for int8 inputs, range checking, and type conversion (Tom)
2384 Fix for SELECT timespan('21:11:26'::time) (Tom)
2385 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
2386 Add btree index on NUMERIC (Jan)
2387 Perl fix for large objects containing NUL characters (Douglas Thomson)
2388 ODBC fix for for large objects (free)
2389 Fix indexing of cidr data type
2390 Fix for Ethernet MAC addresses (macaddr type) comparisons
2391 Fix for date/time types when overflows happened in computations (Tom)
2392 Allow array on int8 (Peter E)
2393 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
2394 Allow NUMERIC arrays
2395 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
2396 Make char_length()/octet_length including trailing blanks (Tom)
2397 Made abstime/reltime use int4 instead of time_t (Peter E)
2398 New lztext data type for compressed text fields
2399 Revise code to handle coercion of int and float constants (Tom)
2400 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
2401 NUMERIC now accepts scientific notation (Tom)
2402 NUMERIC to int4 rounds (Tom)
2403 Convert float4/8 to NUMERIC properly (Tom)
2404 Allow type conversion with NUMERIC (Thomas)
2405 Make ISO date style (2000-02-16 09:33) the default (Thomas)
2406 Add NATIONAL CHAR [ VARYING ] (Thomas)
2407 Allow NUMERIC round and trunc to accept negative scales (Tom)
2408 New TIME WITH TIME ZONE type (Thomas)
2409 Add MAX()/MIN() on time type (Thomas)
2410 Add abs(), mod(), fac() for int8 (Thomas)
2411 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
2412 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
2413 Add exp() and ln() for NUMERIC type
2414 Rename NUMERIC power() to pow() (Thomas)
2415 Improved TRANSLATE() function (Edwin Ramirez, Tom)
2416 Allow X=-Y operators (Tom)
2417 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom)
2418 Allow LOCALE to use indexes in regular expression searches (Tom)
2419 Allow creation of functional indexes to use default types
2423 Prevent exponential space consumption with many AND's and OR's (Tom)
2424 Collect attribute selectivity values for system columns (Tom)
2425 Reduce memory usage of aggregates (Tom)
2426 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
2427 Fix r-tree index optimizer selectivity (Thomas)
2428 Improve optimizer selectivity computations and functions (Tom)
2429 Optimize btree searching for cases where many equal keys exist (Tom)
2430 Enable fast LIKE index processing only if index present (Tom)
2431 Re-use free space on index pages with duplicates (Tom)
2432 Improve hash join processing (Tom)
2433 Prevent descending sort if result is already sorted(Hiroshi)
2434 Allow commuting of index scan query qualifications (Tom)
2435 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
2436 Allocate large memory requests in fix-sized chunks for performance (Tom)
2437 Fix vacuum's performance by reducing memory allocation requests (Tom)
2438 Implement constant-expression simplification (Bernard Frankpitt, Tom)
2439 Use secondary columns to be used to determine start of index scan (Hiroshi)
2440 Prevent quadruple use of disk space when doing internal sorting (Tom)
2441 Faster sorting by calling fewer functions (Tom)
2442 Create system indexes to match all system caches (Bruce, Hiroshi)
2443 Make system caches use system indexes (Bruce)
2444 Make all system indexes unique (Bruce)
2445 Improve pg_statistics management for VACUUM speed improvement (Tom)
2446 Flush backend cache less frequently (Tom, Hiroshi)
2447 COPY now reuses previous memory allocation, improving performance (Tom)
2448 Improve optimization cost estimation (Tom)
2449 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom)
2450 Use DNF instead of CNF where appropriate (Tom, Taral)
2451 Further cleanup for OR-of-AND WHERE-clauses (Tom)
2452 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
2453 Smarter optimizer computations for random index page access (Tom)
2454 New SET variable to control optimizer costs (Tom)
2455 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
2456 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
2457 Major subquery speedup (Tom)
2458 Fewer fsync writes when fsync is not disabled (Tom)
2459 Improved LIKE optimizer estimates (Tom)
2460 Prevent fsync in SELECT-only queries (Vadim)
2461 Make index creation use psort code, because it is now faster (Tom)
2462 Allow creation of sort temp tables > 1 Gig
2466 Fix for linux PPC compile
2467 New generic expression-tree-walker subroutine (Tom)
2468 Change form() to varargform() to prevent portability problems
2469 Improved range checking for large integers on Alphas
2470 Clean up #include in /include directory (Bruce)
2471 Add scripts for checking includes (Bruce)
2472 Remove un-needed #include's from *.c files (Bruce)
2473 Change #include's to use <> and "" as appropriate (Bruce)
2474 Enable WIN32 compilation of libpq
2475 Alpha spinlock fix from Uncle George <email>gatgul@voicenet.com</email>
2476 Overhaul of optimizer data structures (Tom)
2477 Fix to cygipc library (Yutaka Tanida)
2478 Allow pgsql to work on newer Cygwin snapshots (Dan)
2479 New catalog version number (Tom)
2481 Rename heap_replace to heap_update
2482 Update for QNX (Dr. Andreas Kardos)
2483 New platform-specific regression handling (Tom)
2484 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
2485 Included all yacc and lex files into the distribution (Peter E.)
2486 Remove lextest, no longer needed (Peter E)
2487 Fix for libpq and psql on Win32 (Magnus)
2488 Internally change datetime and timespan into timestamp and interval (Thomas)
2489 Fix for plpgsql on BSD/OS
2490 Add SQL_ASCII test case to the regression test (Tatsuo)
2491 configure --with-mb now deprecated (Tatsuo)
2493 NetBSD fixes (Johnny C. Lam <email>lamj@stat.cmu.edu</email>)
2494 Fixes for Alpha compiles
2495 New multibyte encodings
2501 <sect1 id="release-6-5-3">
2502 <title>Release 6.5.3</title>
2505 <title>Release date</title>
2506 <simpara>1999-10-13</simpara>
2510 This is basically a cleanup release for 6.5.2. We have added a new
2511 <application>PgAccess</> that was missing in 6.5.2, and installed an NT-specific fix.
2516 <title>Migration to version 6.5.3</title>
2519 A dump/restore is <emphasis>not</emphasis> required for those running
2524 <title>Changes</title>
2528 Updated version of pgaccess 0.98
2530 Fix dumping rules on inherited tables
2537 <sect1 id="release-6-5-2">
2538 <title>Release 6.5.2</title>
2541 <title>Release date</title>
2542 <simpara>1999-09-15</simpara>
2546 This is basically a cleanup release for 6.5.1. We have fixed a variety of
2547 problems reported by 6.5.1 users.
2552 <title>Migration to version 6.5.2</title>
2555 A dump/restore is <emphasis>not</emphasis> required for those running
2561 <title>Changes</title>
2565 subselect+CASE fixes(Tom)
2566 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
2567 Fixes for CASE in WHERE join clauses(Tom)
2568 Fix BTScan abort(Tom)
2569 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
2570 Improve it so that it checks for multicolumn constraints(Thomas)
2571 Fix for Win32 making problem with MB enabled(Hiroki Kataoka)
2572 Allow BSD yacc and bison to compile pl code(Bruce)
2573 Fix SET NAMES working
2575 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
2576 Reduce the total memory consumption of vacuum(Tom)
2577 Fix for timestamp(datetime)
2578 Rule deparsing bugfixes(Tom)
2579 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(Tom)
2580 This is to re-use space on index pages freed by vacuum(Vadim)
2581 document -x for pg_dump(Bruce)
2582 Fix for unary operators in rule deparser(Tom)
2583 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
2584 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
2585 Repair logic error in LIKE: should not return LIKE_ABORT
2586 when reach end of pattern before end of text(Tom)
2587 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom)
2588 Updated version of pgaccess 0.98
2594 <sect1 id="release-6-5-1">
2595 <title>Release 6.5.1</title>
2598 <title>Release date</title>
2599 <simpara>1999-07-15</simpara>
2603 This is basically a cleanup release for 6.5. We have fixed a variety of
2604 problems reported by 6.5 users.
2608 <title>Migration to version 6.5.1</title>
2611 A dump/restore is <emphasis>not</emphasis> required for those running
2617 <title>Changes</title>
2622 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
2623 Remove QUERY_LIMIT, use SELECT...LIMIT
2624 Fix for EXPLAIN on inheritance(Tom)
2625 Patch to allow vacuum on multisegment tables(Hiroshi)
2626 R-Tree optimizer selectivity fix(Tom)
2627 ACL file descriptor leak fix(Atsushi Ogawa)
2628 New expresssion subtree code(Tom)
2629 Avoid disk writes for read-only transactions(Vadim)
2630 Fix for removal of temp tables if last transaction was aborted(Bruce)
2631 Fix to prevent too large tuple from being created(Bruce)
2633 Allow port numbers 32k - 64k(Bruce)
2634 Add ^ precidence(Bruce)
2635 Rename sort files called pg_temp to pg_sorttemp(Bruce)
2636 Fix for microseconds in time values(Tom)
2637 Tutorial source cleanup
2639 Fix for sorting of NULL's in some cases(Tom)
2640 Shared library dependencies fixed (Tom)
2641 Fixed glitches affecting GROUP BY in subselects(Tom)
2642 Fix some compiler warnings (Tomoaki Nishiyama)
2643 Add Win1250 (Czech) support (Pavel Behal)
2649 <sect1 id="release-6-5">
2650 <title>Release 6.5</title>
2653 <title>Release date</title>
2654 <simpara>1999-06-09</simpara>
2658 This release marks a major step in the development team's mastery of the source
2659 code we inherited from Berkeley. You will see we are now easily adding
2660 major features, thanks to the increasing size and experience of our
2661 world-wide development team.
2665 Here is a brief summary of the more notable changes:
2670 Multiversion concurrency control(MVCC)
2674 This removes our old
2675 table-level locking, and replaces it with a locking system that is
2676 superior to most commercial database systems. In a traditional system,
2677 each row that is modified is locked until committed, preventing reads by
2678 other users. MVCC uses the natural multiversion nature of PostgreSQL
2679 to allow readers to continue reading consistent data during writer
2680 activity. Writers continue to use the compact pg_log transaction
2681 system. This is all performed without having to allocate a lock for
2682 every row like traditional database systems. So, basically, we no
2683 longer are restricted by simple table-level locking;
2684 we have something better than row-level locking.
2691 Hot backups from <application>pg_dump</application>
2695 <application>pg_dump</application> takes advantage of the new
2696 MVCC features to give a consistent database dump/backup while
2697 the database stays online and available for queries.
2708 We now have a true numeric data type, with
2709 user-specified precision.
2720 Temporary tables are guaranteed to have unique names
2721 within a database session, and are destroyed on session exit.
2732 We now have CASE, INTERSECT, and EXCEPT statement
2733 support. We have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL,
2734 SELECT ... FOR UPDATE, and an improved LOCK TABLE command.
2745 We continue to speed up PostgreSQL, thanks to the variety of
2746 talents within our team. We have sped up memory allocation,
2747 optimization, table joins, and row transfer routines.
2758 We continue to expand our port list, this time including
2759 <systemitem class="osname">Windows NT</>/<systemitem>ix86</> and <systemitem class="osname">NetBSD</>/<systemitem>arm32</>.
2770 Most interfaces have new versions, and existing functionality
2782 New and updated material is present throughout the
2783 documentation. New <acronym>FAQ</acronym>s have been
2784 contributed for <systemitem class="osname">SGI</> and <systemitem class="osname">AIX</> platforms.
2785 The <citetitle>Tutorial</citetitle> has introductory information
2786 on <acronym>SQL</acronym> from Stefan Simkovics.
2787 For the <citetitle>User's Guide</citetitle>, there are
2788 reference pages covering the postmaster and more utility
2789 programs, and a new appendix
2790 contains details on date/time behavior.
2791 The <citetitle>Administrator's Guide</citetitle> has a new
2792 chapter on troubleshooting from Tom Lane.
2793 And the <citetitle>Programmer's Guide</citetitle> has a
2794 description of query processing, also from Stefan, and details
2795 on obtaining the <productname>PostgreSQL</productname> source
2796 tree via anonymous <productname>CVS</productname> and
2797 <productname>CVSup</productname>.
2805 <title>Migration to version 6.5</title>
2808 A dump/restore using <application>pg_dump</application>
2809 is required for those wishing to migrate data from any
2810 previous release of <productname>PostgreSQL</productname>.
2811 <application>pg_upgrade</application> can <emphasis>not</emphasis>
2812 be used to upgrade to this release because the on-disk structure
2813 of the tables has changed compared to previous releases.
2817 The new Multiversion Concurrency Control (MVCC) features can
2818 give somewhat different behaviors in multiuser
2819 environments. <emphasis>Read and understand the following section
2820 to ensure that your existing applications will give you the
2821 behavior you need.</emphasis>
2825 <title>Multiversion Concurrency Control</title>
2828 Because readers in 6.5 don't lock data, regardless of transaction
2829 isolation level, data read by one transaction can be overwritten by
2830 another. In other words, if a row is returned by
2831 <command>SELECT</command> it doesn't mean that this row really exists
2832 at the time it is returned (i.e. sometime after the statement or
2833 transaction began) nor that the row is protected from being deleted or
2834 updated by concurrent transactions before the current transaction does
2835 a commit or rollback.
2839 To ensure the actual existence of a row and protect it against
2840 concurrent updates one must use <command>SELECT FOR UPDATE</command> or
2841 an appropriate <command>LOCK TABLE</command> statement. This should be
2842 taken into account when porting applications from previous releases of
2843 <productname>PostgreSQL</productname> and other environments.
2847 Keep the above in mind if you are using
2848 <filename>contrib/refint.*</filename> triggers for
2849 referential integrity. Additional techniques are required now. One way is
2850 to use <command>LOCK parent_table IN SHARE ROW EXCLUSIVE MODE</command>
2851 command if a transaction is going to update/delete a primary key and
2852 use <command>LOCK parent_table IN SHARE MODE</command> command if a
2853 transaction is going to update/insert a foreign key.
2857 Note that if you run a transaction in SERIALIZABLE mode then you must
2858 execute the <command>LOCK</command> commands above before execution of any
2859 <acronym>DML</acronym> statement
2860 (<command>SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO</command>) in the
2867 These inconveniences will disappear in the future
2868 when the ability to read dirty
2869 (uncommitted) data (regardless of isolation level) and true referential
2870 integrity will be implemented.
2876 <title>Changes</title>
2882 Fix text<->float8 and text<->float4 conversion functions(Thomas)
2883 Fix for creating tables with mixed-case constraints(Billy)
2884 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
2885 Fix bug in pg_dump -z
2886 Memory overrun cleanups(Tatsuo)
2887 Fix for lo_import crash(Tatsuo)
2888 Adjust handling of data type names to suppress double quotes(Thomas)
2889 Use type coercion for matching columns and DEFAULT(Thomas)
2890 Fix deadlock so it only checks once after one second of sleep(Bruce)
2891 Fixes for aggregates and PL/pgsql(Hiroshi)
2892 Fix for subquery crash(Vadim)
2893 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
2894 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo)
2895 Fix for pg_dump -d or -D and quote special characters in INSERT
2896 Repair serious problems with dynahash(Tom)
2897 Fix INET/CIDR portability problems
2898 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
2899 Fix executor so mergejoin of different column types works(Tom)
2900 Fix for Alpha OR selectivity bug
2901 Fix OR index selectivity problem(Bruce)
2902 Fix so \d shows proper length for char()/varchar()(Ryan)
2903 Fix tutorial code(Clark)
2904 Improve destroyuser checking(Oliver)
2905 Fix for Kerberos(Rodney McDuff)
2906 Fix for dropping database while dirty buffers(Bruce)
2907 Fix so sequence nextval() can be case-sensitive(Bruce)
2909 Drop buffers before destroying database files(Bruce)
2910 Fix case where executor evaluates functions twice(Tatsuo)
2911 Allow sequence nextval actions to be case-sensitive(Bruce)
2912 Fix optimizer indexing not working for negative numbers(Bruce)
2913 Fix for memory leak in executor with fjIsNull
2914 Fix for aggregate memory leaks(Erik Riedel)
2915 Allow username containing a dash GRANT permissions
2916 Cleanup of NULL in inet types
2917 Clean up system table bugs(Tom)
2918 Fix problems of PAGER and \? command(Masaaki Sakaida)
2919 Reduce default multisegment file size limit to 1GB(Peter)
2920 Fix for dumping of CREATE OPERATOR(Tom)
2921 Fix for backward scanning of cursors(Hiroshi Inoue)
2922 Fix for COPY FROM STDIN when using \i(Tom)
2923 Fix for subselect is compared inside an expression(Jan)
2924 Fix handling of error reporting while returning rows(Tom)
2925 Fix problems with reference to array types(Tom,Jan)
2926 Prevent UPDATE SET oid(Jan)
2927 Fix pg_dump so -t option can handle case-sensitive tablenames
2928 Fixes for GROUP BY in special cases(Tom, Jan)
2929 Fix for memory leak in failed queries(Tom)
2930 DEFAULT now supports mixed-case identifiers(Tom)
2931 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
2932 Disable use of pg_dump with both -o and -d options(Bruce)
2933 Allow pg_dump to properly dump GROUP permissions(Bruce)
2934 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
2935 Fix for computations in views(Jan)
2936 Fix for aggregates on array indexes(Tom)
2937 Fix for DEFAULT handles single quotes in value requiring too many quotes
2938 Fix security problem with non-super users importing/exporting large objects(Tom)
2939 Rollback of transaction that creates table cleaned up properly(Tom)
2940 Fix to allow long table and column names to generate proper serial names(Tom)
2944 Add "vacuumdb" utility
2945 Speed up libpq by allocating memory better(Tom)
2946 EXPLAIN all indexes used(Tom)
2947 Implement CASE, COALESCE, NULLIF expression(Thomas)
2948 New pg_dump table output format(Constantin)
2949 Add string min()/max() functions(Thomas)
2950 Extend new type coercion techniques to aggregates(Thomas)
2951 New moddatetime contrib(Terry)
2952 Update to pgaccess 0.96(Constantin)
2953 Add routines for single-byte "char" type(Thomas)
2954 Improved substr() function(Thomas)
2955 Improved multibyte handling(Tatsuo)
2956 Multiversion concurrency control/MVCC(Vadim)
2957 New Serialized mode(Vadim)
2958 Fix for tables over 2gigs(Peter)
2959 New SET TRANSACTION ISOLATION LEVEL(Vadim)
2960 New LOCK TABLE IN ... MODE(Vadim)
2961 Update ODBC driver(Byron)
2962 New NUMERIC data type(Jan)
2963 New SELECT FOR UPDATE(Vadim)
2964 Handle "NaN" and "Infinity" for input values(Jan)
2965 Improved date/year handling(Thomas)
2966 Improved handling of backend connections(Magnus)
2967 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
2968 New TCL_ARRAYS option(Massimo)
2969 New INTERSECT and EXCEPT(Stefan)
2970 New pg_index.indisprimary for primary key tracking(D'Arcy)
2971 New pg_dump option to allow dropping of tables before creation(Brook)
2972 Speedup of row output routines(Tom)
2973 New READ COMMITTED isolation level(Vadim)
2974 New TEMP tables/indexes(Bruce)
2975 Prevent sorting if result is already sorted(Jan)
2976 New memory allocation optimization(Jan)
2977 Allow psql to do \p\g(Bruce)
2978 Allow multiple rule actions(Jan)
2979 Added LIMIT/OFFSET functionality(Jan)
2980 Improve optimizer when joining a large number of tables(Bruce)
2981 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
2982 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
2983 Improved int8 support(Ryan Bradetich, Thomas, Tom)
2984 New routines to convert between int8 and text/varchar types(Thomas)
2985 New bushy plans, where meta-tables are joined(Bruce)
2986 Enable right-hand queries by default(Bruce)
2987 Allow reliable maximum number of backends to be set at configure time
2988 (--with-maxbackends and postmaster switch (-N backends))(Tom)
2989 GEQO default now 10 tables because of optimizer speedups(Tom)
2990 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
2991 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
2992 Allow psql \d on a view show query(Ryan)
2993 Speedup for LIKE(Bruce)
2994 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
2995 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
2996 Make % operator have precedence like /(Bruce)
2997 Add new postgres -O option to allow system table structure changes(Bruce)
2998 Update contrib/pginterface/findoidjoins script(Tom)
2999 Major speedup in vacuum of deleted rows with indexes(Vadim)
3000 Allow non-SQL functions to run different versions based on arguments(Tom)
3001 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida)
3002 Add version number in start-up banners for psql(Masaaki Sakaida)
3003 New contrib/vacuumlo removes large objects not referenced(Peter)
3004 New initialization for table sizes so non-vacuumed tables perform better(Tom)
3005 Improve error messages when a connection is rejected(Tom)
3006 Support for arrays of char() and varchar() fields(Massimo)
3007 Overhaul of hash code to increase reliability and performance(Tom)
3008 Update to PyGreSQL 2.4(D'Arcy)
3009 Changed debug options so -d4 and -d5 produce different node displays(Jan)
3010 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
3011 Better optimization statistics for system table access(Tom)
3012 Better handling of non-default block sizes(Massimo)
3013 Improve GEQO optimizer memory consumption(Tom)
3014 UNION now suppports ORDER BY of columns not in target list(Jan)
3015 Major libpq++ improvements(Vince Vielhaber)
3016 pg_dump now uses -z(ACL's) as default(Bruce)
3017 backend cache, memory speedups(Tom)
3018 have pg_dump do everything in one snapshot transaction(Vadim)
3019 fix for large object memory leakage, fix for pg_dumping(Tom)
3020 INET type now respects netmask for comparisons
3021 Make VACUUM ANALYZE only use a readlock(Vadim)
3022 Allow VIEWs on UNIONS(Jan)
3023 pg_dump now can generate consistent snapshots on active databases(Vadim)
3027 Improve port matching(Tom)
3028 Portability fixes for SunOS
3029 Add NT/Win32 backend port and enable dynamic loading(Magnus and Daniel Horak)
3030 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
3031 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
3032 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
3033 Port to NetBSD/macppc(Toshimi Aoki)
3034 Fix for tcl/tk configuration(Vince)
3035 Removed CURRENT keyword for rule queries(Jan)
3036 NT dynamic loading now works(Daniel Horak)
3037 Add ARM32 support(Andrew McMurry)
3038 Better support for HP-UX 11 and UnixWare
3039 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
3040 New install commands for plpgsql(Jan)
3047 <sect1 id="release-6-4-2">
3048 <title>Release 6.4.2</title>
3051 <title>Release date</title>
3052 <simpara>1998-12-20</simpara>
3056 The 6.4.1 release was improperly packaged. This also has one additional
3062 <title>Migration to version 6.4.2</title>
3065 A dump/restore is <emphasis>not</emphasis> required for those running
3070 <title>Changes</title>
3074 Fix for datetime constant problem on some platforms(Thomas)
3082 <sect1 id="release-6-4-1">
3083 <title>Release 6.4.1</title>
3086 <title>Release date</title>
3087 <simpara>1998-12-18</simpara>
3091 This is basically a cleanup release for 6.4. We have fixed a variety of
3092 problems reported by 6.4 users.
3097 <title>Migration to version 6.4.1</title>
3100 A dump/restore is <emphasis>not</emphasis> required for those running
3105 <title>Changes</title>
3109 Add pg_dump -N flag to force double quotes around identifiers. This is
3111 Fix for NOT in where clause causing crash(Bruce)
3112 EXPLAIN VERBOSE coredump fix(Vadim)
3113 Fix shared-library problems on Linux
3114 Fix test for table existance to allow mixed-case and whitespace in
3115 the table name(Thomas)
3116 Fix a couple of pg_dump bugs
3117 Configure matches template/.similar entries better(Tom)
3118 Change builtin function names from SPI_* to spi_*
3119 OR WHERE clause fix(Vadim)
3120 Fixes for mixed-case table names(Billy)
3121 contrib/linux/postgres.init.csh/sh fix(Thomas)
3122 libpq memory overrun fix
3124 Change exp() behavior to generate error on underflow(Thomas)
3125 pg_dump fixes for memory leak, inheritance constraints, layout change
3126 update pgaccess to 0.93
3127 Fix prototype for 64-bit platforms
3128 Multibyte fixes(Tatsuo)
3130 Fix memory overruns(Tatsuo)
3131 Fix for lo_import() crash(Bruce)
3132 Better search for install program(Tom)
3135 Use implicit type coercion for matching DEFAULT values(Thomas)
3136 Add routines to help with single-byte (internal) character type(Thomas)
3137 Compilation of libpq for Win32 fixes(Magnus)
3138 Upgrade to PyGreSQL 2.2(D'Arcy)
3146 <sect1 id="release-6-4">
3147 <title>Release 6.4</title>
3150 <title>Release date</title>
3151 <simpara>1998-10-30</simpara>
3155 There are <emphasis>many</emphasis> new features and improvements in this release.
3156 Thanks to our developers and maintainers, nearly every aspect of the system
3157 has received some attention since the previous release.
3158 Here is a brief, incomplete summary:
3163 Views and rules are now functional thanks to extensive new code in the
3164 rewrite rules system from Jan Wieck. He also wrote a chapter on it
3165 for the <citetitle>Programmer's Guide</citetitle>.
3170 Jan also contributed a second procedural language, <application>PL/pgSQL</application>, to go with the
3171 original <application>PL/pgTCL</application> procedural language he contributed last release.
3177 We have optional multiple-byte character set support from Tatsuo Ishii
3178 to complement our existing locale support.
3184 Client/server communications has been cleaned up, with better support for
3185 asynchronous messages and interrupts thanks to Tom Lane.
3191 The parser will now perform automatic type coercion to match arguments
3192 to available operators and functions, and to match columns and expressions
3193 with target columns. This uses a generic mechanism which supports
3194 the type extensibility features of <productname>PostgreSQL</productname>.
3195 There is a new chapter in the <citetitle>User's Guide</citetitle>
3196 which covers this topic.
3202 Three new data types have been added.
3203 Two types, <type>inet</type> and <type>cidr</type>, support various forms
3204 of IP network, subnet, and machine addressing. There is now an 8-byte integer
3205 type available on some platforms. See the chapter on data types
3206 in the <citetitle>User's Guide</citetitle> for details.
3207 A fourth type, <type>serial</type>, is now supported by the parser as an
3208 amalgam of the <type>int4</type> type, a sequence, and a unique index.
3214 Several more <acronym>SQL92</acronym>-compatible syntax features have been
3215 added, including <command>INSERT DEFAULT VALUES</command>
3221 The automatic configuration and installation system has received some
3222 attention, and should be more robust for more platforms than it has ever
3231 <title>Migration to version 6.4</title>
3234 A dump/restore using <application>pg_dump</application>
3235 or <application>pg_dumpall</application>
3236 is required for those wishing to migrate data from any
3237 previous release of <productname>PostgreSQL</productname>.
3242 <title>Changes</title>
3248 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
3249 Remove char2-16 data types, use char/varchar(Darren)
3250 Pqfn not handles a NOTICE message(Anders)
3251 Reduced busywaiting overhead for spinlocks with many backends (dg)
3252 Stuck spinlock detection (dg)
3253 Fix up "ISO-style" timespan decoding and encoding(Thomas)
3254 Fix problem with table drop after rollback of transaction(Vadim)
3255 Change error message and remove non-functional update message(Vadim)
3256 Fix for COPY array checking
3257 Fix for SELECT 1 UNION SELECT NULL
3258 Fix for buffer leaks in large object calls(Pascal)
3259 Change owner from oid to int4 type(Bruce)
3260 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
3261 Fix for shared invalidation cache overflow(Massimo)
3262 Prevent file descriptor leaks in failed COPY's(Bruce)
3263 Fix memory leak in libpgtcl's pg_select(Constantin)
3264 Fix problems with username/passwords over 8 characters(Tom)
3265 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
3266 Fix of many bad system table entries(Tom)
3270 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
3271 Show the index used in an EXPLAIN(Zeugswetter)
3272 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
3273 Multibyte awareness of many data types and functions, via configure(Tatsuo)
3274 New configure --with-mb option(Tatsuo)
3275 New initdb --pgencoding option(Tatsuo)
3276 New createdb -E multibyte option(Tatsuo)
3277 Select version(); now returns PostgreSQL version(Jeroen)
3278 Libpq now allows asynchronous clients(Tom)
3279 Allow cancel from client of backend query(Tom)
3280 Psql now cancels query with Control-C(Tom)
3281 Libpq users need not issue dummy queries to get NOTIFY messages(Tom)
3282 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
3283 PGresult struct now includes associated error message, if any(Tom)
3284 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
3285 Add routines to convert between varchar and bpchar(Thomas)
3286 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
3287 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
3288 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas)
3289 Fixes for unary minus parsing with leading spaces(Thomas)
3290 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
3291 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
3292 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
3293 Enable HAVING clause but no fixes elsewhere yet.
3294 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thomas)
3295 Save string type if specified for DEFAULT clause handling(Thomas)
3296 Coerce operations involving different data types(Thomas)
3297 Allow some index use for columns of different types(Thomas)
3298 Add capabilities for automatic type conversion(Thomas)
3299 Cleanups for large objects, so file is truncated on open(Peter)
3300 Readline cleanups(Tom)
3301 Allow psql \f \ to make spaces as delimiter(Bruce)
3302 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
3303 Msql compatibility library in /contrib(Aldrin)
3304 Remove the requirement that ORDER/GROUP BY clause identifiers be
3305 included in the target list(David)
3306 Convert columns to match columns in UNION clauses(Thomas)
3307 Remove fork()/exec() and only do fork()(Bruce)
3308 Jdbc cleanups(Peter)
3309 Show backend status on ps command line(only works on some platforms)(Bruce)
3310 Pg_hba.conf now has a sameuser option in the database field
3311 Make lo_unlink take oid param, not int4
3312 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
3313 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
3314 libpgtcl cleanups(Tom)
3315 Add -error option to libpgtcl's pg_result command(Tom)
3316 New locale patch, see docs/README/locale(Oleg)
3317 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
3318 New contrib/lo code for large object orphan removal(Peter)
3319 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
3320 feature, see /doc/README.mb(Tatsuo)
3321 /contrib/noupdate code to revoke update permission on a column
3322 Libpq can now be compiled on win32(Magnus)
3323 Add PQsetdbLogin() in libpq
3324 New 8-byte integer type, checked by configure for OS support(Thomas)
3325 Better support for quoted table/column names(Thomas)
3326 Surround table and column names with double-quotes in pg_dump(Thomas)
3327 PQreset() now works with passwords(Tom)
3328 Handle case of GROUP BY target list column number out of range(David)
3329 Allow UNION in subselects
3330 Add auto-size to screen to \d? commands(Bruce)
3331 Use UNION to show all \d? results in one query(Bruce)
3332 Add \d? field search feature(Bruce)
3333 Pg_dump issues fewer \connect requests(Tom)
3334 Make pg_dump -z flag work better, document it in manual page(Tom)
3335 Add HAVING clause with full support for subselects and unions(Stephan)
3336 Full text indexing routines in contrib/fulltextindex(Maarten)
3337 Transaction ids now stored in shared memory(Vadim)
3338 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
3339 Support for SQL92 syntax "SET NAMES"(Tatsuo)
3340 Support for LATIN2-5(Tatsuo)
3341 Add UNICODE regression test case(Tatsuo)
3342 Lock manager cleanup, new locking modes for LLL(Vadim)
3343 Allow index use with OR clauses(Bruce)
3344 Allows "SELECT NULL ORDER BY 1;"
3345 Explain VERBOSE prints the plan, and now pretty-prints the plan to
3346 the postmaster log file(Bruce)
3347 Add indexes display to \d command(Bruce)
3348 Allow GROUP BY on functions(David)
3349 New pg_class.relkind for large objects(Bruce)
3350 New way to send libpq NOTICE messages to a different location(Tom)
3351 New \w write command to psql(Bruce)
3352 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
3353 Allow binary-compatible indexes to be considered when checking for valid
3354 Indexes for restriction clauses containing a constant(Thomas)
3355 New ISBN/ISSN code in /contrib/isbn_issn
3356 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
3357 New rewrite system fixes many problems with rules and views(Jan)
3358 * Rules on relations work
3359 * Event qualifications on insert/update/delete work
3360 * New OLD variable to reference CURRENT, CURRENT will be remove in future
3361 * Update rules can reference NEW and OLD in rule qualifications/actions
3362 * Insert/update/delete rules on views work
3363 * Multiple rule actions are now supported, surrounded by parentheses
3364 * Regular users can create views/rules on tables they have RULE permits
3365 * Rules and views inherit the permissions on the creator
3366 * No rules at the column level
3367 * No UPDATE NEW/OLD rules
3368 * New pg_tables, pg_indexes, pg_rules and pg_views system views
3369 * Only a single action on SELECT rules
3370 * Total rewrite overhaul, perhaps for 6.5
3372 * handle aggregates on views
3373 * handle insert into select from view works
3374 System indexes are now multikey(Bruce)
3375 Oidint2, oidint4, and oidname types are removed(Bruce)
3376 Use system cache for more system table lookups(Bruce)
3377 New backend programming language PL/pgSQL in backend/pl(Jan)
3378 New SERIAL data type, auto-creates sequence/index(Thomas)
3379 Enable assert checking without a recompile(Massimo)
3380 User lock enhancements(Massimo)
3381 New setval() command to set sequence value(Massimo)
3382 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
3383 Conditional trace package(Massimo)
3384 New UNLISTEN command(Massimo)
3385 Psql and libpq now compile under win32 using win32.mak(Magnus)
3386 Lo_read no longer stores trailing NULL(Bruce)
3387 Identifiers are now truncated to 31 characters internally(Bruce)
3388 Createuser options now availble on the command line
3389 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
3390 Prevent file descriptor leaf from failed COPY(Bruce)
3391 New pg_upgrade command(Bruce)
3392 Updated /contrib directories(Massimo)
3393 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
3394 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
3395 New DECLARE and FETCH feature(Thomas)
3396 libpq's internal structures now not exported(Tom)
3397 Allow up to 8 key indexes(Bruce)
3398 Remove ARCHIVE keyword, that is no longer used(Thomas)
3399 pg_dump -n flag to supress quotes around indentifiers
3400 disable system columns for views(Jan)
3401 new INET and CIDR types for network addresses(TomH, Paul)
3402 no more double quotes in psql output
3403 pg_dump now dumps views(Terry)
3404 new SET QUERY_LIMIT(Tatsuo,Jan)
3408 /contrib cleanup(Jun)
3409 Inline some small functions called for every row(Bruce)
3412 Multibyte regression tests(Soonmyung.)
3413 Remove --disabled options from configure
3414 Define PGDOC to use POSTGRESDIR by default
3415 Make regression optional
3416 Remove extra braces code to pgindent(Bruce)
3417 Add bsdi shared library support(Bruce)
3418 New --without-CXX support configure option(Brook)
3420 Update backend flowchart in tools/backend(Bruce)
3421 Change atttypmod from int16 to int32(Bruce, Tom)
3422 Getrusage() fix for platforms that do not have it(Tom)
3423 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
3424 NS32K platform fixes(Phil Nelson, John Buller)
3425 SCO 7/UnixWare 2.x fixes(Billy,others)
3426 Sparc/Solaris 2.5 fixes(Ryan)
3427 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
3428 Even more documention(Thomas)
3429 Nextstep support(Jacek)
3431 pginterface manual page(Bruce)
3432 shared libraries all have version numbers
3433 merged all OS-specific shared library defines into one file
3434 smarter TCL/TK configuration checking(Billy)
3435 smarter perl configuration(Brook)
3436 configure uses supplied install-sh if no install script found(Tom)
3437 new Makefile.shlib for shared library configuration(Tom)
3443 <sect1 id="release-6-3-2">
3444 <title>Release 6.3.2</title>
3447 <title>Release date</title>
3448 <simpara>1998-04-07</simpara>
3452 This is a bug-fix release for 6.3.x.
3453 Refer to the release notes for version 6.3 for a more complete summary of new features.
3461 Repairs automatic configuration support for some platforms, including Linux,
3462 from breakage inadvertently introduced in version 6.3.1.
3468 Correctly handles function calls on the left side of BETWEEN and LIKE clauses.
3475 A dump/restore is NOT required for those running 6.3 or 6.3.1. A
3476 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
3477 This last step should be performed while the postmaster is not running.
3478 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
3481 For upgrades from pre-6.3 installations,
3482 refer to the installation and migration instructions for version 6.3.
3486 <title>Changes</title>
3490 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
3491 Manual page improvements(Bruce)
3492 BETWEEN and LIKE fix(Thomas)
3493 fix for psql \connect used by pg_dump(Oliver Elphick)
3495 pgaccess, version 0.86
3496 qsort removed, now uses libc version, cleanups(Jeroen)
3497 fix for buffer over-runs detected(Maurice Gittens)
3498 fix for buffer overrun in libpgtcl(Randy Kunkee)
3499 fix for UNION with DISTINCT or ORDER BY(Bruce)
3500 gettimeofday configure check(Doug Winterburn)
3501 Fix "indexes not used" bug(Vadim)
3502 docs additions(Thomas)
3503 Fix for backend memory leak(Bruce)
3504 libreadline cleanup(Erwan MAS)
3505 Remove DISTDIR(Bruce)
3506 Makefile dependency cleanup(Jeroen van Vianen)
3513 <sect1 id="release-6-3-1">
3514 <title>Release 6.3.1</title>
3517 <title>Release date</title>
3518 <simpara>1998-03-23</simpara>
3527 Additional support for multibyte character sets.
3533 Repair byte ordering for mixed-endian clients and servers.
3539 Minor updates to allowed SQL syntax.
3545 Improvements to the configuration autodetection for installation.
3552 A dump/restore is NOT required for those running 6.3. A
3553 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
3554 This last step should be performed while the postmaster is not running.
3555 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
3558 For upgrades from pre-6.3 installations,
3559 refer to the installation and migration instructions for version 6.3.
3563 <title>Changes</title>
3567 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
3568 pg_user cleanup(Bruce)
3569 large object fix for pg_dump and tclsh (alvin)
3570 LIKE fix for multiple adjacent underscores
3571 fix for redefining builtin functions(Thomas)
3573 upgrade to pg_access 0.83
3574 updated CLUSTER manual page
3575 multibyte character set support, see doc/README.mb(Tatsuo)
3576 configure --with-pgport fix
3578 big-endian fix for backend communications(Kataoka)
3579 SUBSTR() and substring() fix(Jan)
3580 several jdbc fixes(Peter)
3581 libpgtcl improvements, see libptcl/README(Randy Kunkee)
3582 Fix for "Datasize = 0" error(Vadim)
3583 Prevent \do from wrapping(Bruce)
3584 Remove duplicate Russian character set entries
3586 Allow optional TABLE keyword in LOCK and SELECT INTO(Thomas)
3587 CREATE SEQUENCE options to allow a negative integer(Thomas)
3588 Add "PASSWORD" as an allowed column identifier(Thomas)
3589 Add checks for UNION target fields(Bruce)
3590 Fix Alpha port(Dwayne Bailey)
3591 Fix for text arrays containing quotes(Doug Gibson)
3592 Solaris compile fix(Albert Chin-A-Young)
3593 Better identify tcl and tk libs and includes(Bruce)
3599 <sect1 id="release-6-3">
3600 <title>Release 6.3</title>
3603 <title>Release date</title>
3604 <simpara>1998-03-01</simpara>
3608 There are <emphasis>many</emphasis> new features and improvements in this release.
3609 Here is a brief, incomplete summary:
3614 Many new SQL features, including
3615 full <acronym>SQL92</acronym> subselect capability
3616 (everything is here but target-list subselects).
3622 Support for client-side environment variables to specify time zone and date style.
3628 Socket interface for client/server connection. This is the default now
3629 so you may need to start <application>postmaster</application> with the
3630 <option>-i</option> flag.
3636 Better password authorization mechanisms. Default table permissions have changed.
3642 Old-style <firstterm>time travel</firstterm>
3643 has been removed. Performance has been improved.
3651 Bruce Momjian wrote the following notes to introduce the new release.
3656 There are some general 6.3 issues that I want to mention. These are
3657 only the big items that can not be described in one sentence. A review
3658 of the detailed changes list is still needed.
3661 First, we now have subselects. Now that we have them, I would like to
3662 mention that without subselects, SQL is a very limited language.
3663 Subselects are a major feature, and you should review your code for
3664 places where subselects provide a better solution for your queries. I
3665 think you will find that there are more uses for subselects than you may
3666 think. Vadim has put us on the big SQL map with subselects, and fully
3667 functional ones too. The only thing you can't do with subselects is to
3668 use them in the target list.
3671 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
3672 enable connections from other machines, you have to use the new
3673 postmaster -i option, and of course edit <filename>pg_hba.conf</filename>. Also, for this
3674 reason, the format of <filename>pg_hba.conf</filename> has changed.
3677 Third, <type>char()</type> fields will now allow faster access than <type>varchar()</type> or
3678 <type>text</type>. Specifically, the <type>text</> and <type>varchar()</type> have a penalty for access to
3679 any columns after the first column of this type. <type>char()</type> used to also
3680 have this access penalty, but it no longer does. This may suggest that
3681 you redesign some of your tables, especially if you have short character
3682 columns that you have defined as <type>varchar()</type> or <type>text</type>. This and other
3683 changes make 6.3 even faster than earlier releases.
3686 We now have passwords definable independent of any Unix file. There are
3687 new SQL USER commands.
3688 See the <citetitle>Administrator's Guide</citetitle> for more
3689 information. There is a new table, pg_shadow, which is used to store
3690 user information and user passwords, and it by default only SELECT-able
3691 by the <systemitem>postgres</systemitem> super-user. pg_user is now a view of pg_shadow, and is
3692 SELECT-able by PUBLIC. You should keep using pg_user in your
3693 application without changes.
3696 User-created tables now no longer have SELECT permission to PUBLIC by
3697 default. This was done because the ANSI standard requires it. You can
3698 of course GRANT any permissions you want after the table is created.
3699 System tables continue to be SELECT-able by PUBLIC.
3702 We also have real deadlock detection code. No more sixty-second
3703 timeouts. And the new locking code implements a <acronym>FIFO</acronym> better, so there
3704 should be less resource starvation during heavy use.
3707 Many complaints have been made about inadequate documentation in previous
3708 releases. Thomas has put much effort into many new manuals for this
3709 release. Check out the doc/ directory.
3712 For performance reasons, time travel is gone, but can be implemented
3713 using triggers (see <filename>pgsql/contrib/spi/README</filename>). Please check out the new
3714 \d command for types, operators, etc. Also, views have their own
3715 permissions now, not based on the underlying tables, so permissions on
3716 them have to be set separately. Check <filename>/pgsql/interfaces</filename> for some new
3717 ways to talk to <productname>PostgreSQL</productname>.
3720 This is the first release that really required an explanation for
3721 existing users. In many ways, this was necessary because the new
3722 release removes many limitations, and the work-arounds people were using
3723 are no longer needed.
3727 <title>Migration to version 6.3</title>
3730 A dump/restore using <application>pg_dump</application>
3731 or <application>pg_dumpall</application>
3732 is required for those wishing to migrate data from any
3733 previous release of <productname>PostgreSQL</productname>.
3738 <title>Changes</title>
3744 Fix binary cursors broken by MOVE implementation(Vadim)
3745 Fix for tcl library crash(Jan)
3746 Fix for array handling, from Gerhard Hintermayer
3747 Fix acl error, and remove duplicate pqtrace(Bruce)
3748 Fix psql \e for empty file(Bruce)
3749 Fix for textcat on varchar() fields(Bruce)
3750 Fix for DBT Sendproc (Zeugswetter Andres)
3751 Fix vacuum analyze syntax problem(Bruce)
3752 Fix for international identifiers(Tatsuo)
3753 Fix aggregates on inherited tables(Bruce)
3754 Fix substr() for out-of-bounds data
3755 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
3756 Fix notty output to show status result. -q option still turns it off(Bruce)
3757 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
3759 Fix for PQtrace start/stop several times(Bruce)
3760 Fix a variety of locking problems like newer lock waiters getting
3761 lock before older waiters, and having readlock people not share
3762 locks if a writer is waiting for a lock, and waiting writers not
3763 getting priority over waiting readers(Bruce)
3764 Fix crashes in psql when executing queries from external files(James)
3765 Fix problem with multiple order by columns, with the first one having
3767 Use correct hash table support functions for float8 and int4(Thomas)
3768 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
3769 Change precedence for boolean operators to match expected behavior(Thomas)
3770 Generate elog(ERROR) on over-large integer(Bruce)
3771 Allow multiple-argument functions in constraint clauses(Thomas)
3772 Check boolean input literals for 'true','false','yes','no','1','0'
3773 and throw elog(ERROR) if unrecognized(Thomas)
3774 Major large objects fix
3775 Fix for GROUP BY showing duplicates(Vadim)
3776 Fix for index scans in MergeJion(Vadim)
3780 Subselects with EXISTS, IN, ALL, ANY keywords (Vadim, Bruce, Thomas)
3781 New User Manual(Thomas, others)
3782 Speedup by inlining some frequently-called functions
3783 Real deadlock detection, no more timeouts(Bruce)
3784 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
3785 CURRENT_USER(Thomas)
3786 Modify constraint syntax to be SQL92-compliant(Thomas)
3787 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
3788 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
3789 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas)
3790 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
3791 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
3792 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
3793 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
3794 Allow SQL92 delimited identifiers(Thomas)
3795 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas)
3796 Support SQL92 syntax for type coercion of literal strings
3797 (e.g. "DATETIME 'now'")(Thomas)
3798 Add conversions for int2, int4, and OID types to and from text(Thomas)
3799 Use shared lock when building indexes(Vadim)
3800 Free memory allocated for an user query inside transaction block after
3801 this query is done, was turned off in <= 6.2.1(Vadim)
3802 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
3803 New <productname>PostgreSQL</productname> Procedural Language (PL) backend interface(Jan)
3804 Rename pg_dump -H option to -h(Bruce)
3805 Add Java support for passwords, European dates(Peter)
3806 Use indexes for LIKE and ~, !~ operations(Bruce)
3807 Add hash functions for datetime and timespan(Thomas)
3808 Time Travel removed(Vadim, Bruce)
3809 Add paging for \d and \z, and fix \i(Bruce)
3810 Add Unix domain socket support to backend and to frontend library(Goran)
3811 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
3812 Allow more SQL92 and/or <productname>PostgreSQL</productname> reserved words as column identifiers(Thomas)
3813 Augment support for SQL92 SET TIME ZONE...(Thomas)
3814 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
3815 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
3816 Enable SET TIME ZONE using TZ environment variable(Thomas)
3817 Add PGDATESTYLE environment variable to frontend and backend initialization(Thomas)
3818 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
3819 frontend library initialization environment variables(Thomas)
3820 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
3821 Add pg_description table for info on tables, columns, operators, types, and
3823 Increase 16 char limit on system table/index names to 32 characters(Bruce)
3824 Rename system indexes(Bruce)
3825 Add 'GERMAN' option to SET DATESTYLE(Thomas)
3826 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
3827 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
3828 Validate numeric input more carefully for delta times(Thomas)
3829 Implement day of year as possible input to date_part()(Thomas)
3830 Define timespan_finite() and text_timespan() functions(Thomas)
3831 Remove archive stuff(Bruce)
3832 Allow for a pg_password authentication database that is separate from
3833 the system password file(Todd)
3834 Dump ACLs, GRANT, REVOKE permissions(Matt)
3835 Define text, varchar, and bpchar string length functions(Thomas)
3836 Fix Query handling for inheritance, and cost computations(Bruce)
3837 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
3838 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
3839 Implement UNIONs for SELECT(Bruce)
3840 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
3841 varchar() stores only necessary bytes on disk(Bruce)
3842 Fix for BLOBs(Peter)
3843 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
3844 Remove unused "option" from PQconnectdb()
3845 New LOCK command and lock manual page describing deadlocks(Bruce)
3846 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
3847 Enhance psql \z to show sequences(Bruce)
3848 Show NOT NULL and DEFAULT in psql \d table(Bruce)
3849 New psql .psqlrc file start-up(Andrew)
3850 Modify sample start-up script in contrib/linux to show syslog(Thomas)
3851 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
3852 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
3853 Update of contrib stuff(Massimo)
3854 Add Unix socket support to DBD::Pg(Goran)
3855 New python interface (PyGreSQL 2.0)(D'Arcy)
3856 New frontend/backend protocol has a version number, network byte order(Phil)
3857 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
3858 CHAR() now faster access than VARCHAR() or TEXT
3859 ecpg embedded SQL preprocessor
3860 Reduce system column overhead(Vadmin)
3861 Remove pg_time table(Vadim)
3862 Add pg_type attribute to identify types that need length (bpchar, varchar)
3863 Add report of offending line when COPY command fails
3864 Allow VIEW permissions to be set separately from the underlying tables.
3865 For security, use GRANT/REVOKE on views as appropriate(Jan)
3866 Tables now have no default GRANT SELECT TO PUBLIC. You must
3867 explicitly grant such permissions.
3868 Clean up tutorial examples(Darren)
3872 Add new html development tools, and flow chart in /tools/backend
3873 Fix for SCO compiles
3874 Stratus computer port Robert Gillies
3875 Added support for shlib for BSD44_derived & i386_solaris
3876 Make configure more automated(Brook)
3877 Add script to check regression test results
3878 Break parser functions into smaller files, group together(Bruce)
3879 Rename heap_create to heap_create_and_catalog, rename heap_creatr
3880 to heap_create()(Bruce)
3881 Sparc/Linux patch for locking(TomS)
3882 Remove PORTNAME and reorganize port-specific stuff(Marc)
3883 Add optimizer README file(Bruce)
3884 Remove some recursion in optimizer and clean up some code there(Bruce)
3885 Fix for NetBSD locking(Henry)
3886 Fix for libptcl make(Tatsuo)
3888 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
3889 function calls to istrue() or isfalse() to allow optimization(Thomas)
3890 Various fixes NetBSD/Sparc related(TomH)
3891 Alpha linux locking(Travis,Ryan)
3892 Change elog(WARN) to elog(ERROR)(Bruce)
3893 FAQ for FreeBSD(Marc)
3894 Bring in the PostODBC source tree as part of our standard distribution(Marc)
3895 A minor patch for HP/UX 10 vs 9(Stan)
3896 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
3897 UnixWare patches(Billy)
3898 New i386 'lock' for spin lock asm(Billy)
3899 Support for multiplexed backends is removed
3900 Start an OpenBSD port
3903 Add string functions to regression suite(Thomas)
3904 Expand a few function names formerly truncated to 16 characters(Thomas)
3905 Remove un-needed malloc() calls and replace with palloc()(Bruce)
3911 <sect1 id="release-6-2-1">
3912 <title>Release 6.2.1</title>
3915 <title>Release date</title>
3916 <simpara>1997-10-17</simpara>
3920 6.2.1 is a bug-fix and usability release on 6.2.
3928 Allow strings to span lines, per <acronym>SQL92</acronym>.
3934 Include example trigger function for inserting user names on table updates.
3941 This is a minor bug-fix release on 6.2.
3942 For upgrades from pre-6.2 systems, a full dump/reload is required.
3943 Refer to the 6.2 release notes for instructions.
3947 <title>Migration from version 6.2 to version 6.2.1</title>
3950 This is a minor bug-fix release. A dump/reload is not required from version 6.2,
3951 but is required from any release prior to 6.2.
3954 In upgrading from version 6.2, if you choose to dump/reload you will find that
3955 avg(money) is now calculated correctly. All other bug fixes take effect
3956 upon updating the executables.
3959 Another way to avoid dump/reload is to use the following SQL command
3960 from <command>psql</command> to update the existing system table:
3963 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
3964 where aggname = 'avg' and aggbasetype = 790;
3968 This will need to be done to every existing database, including template1.
3973 <title>Changes</title>
3977 Allow TIME and TYPE column names(Thomas)
3978 Allow larger range of true/false as boolean values(Thomas)
3979 Support output of "now" and "current"(Thomas)
3980 Handle DEFAULT with INSERT of NULL properly(Vadim)
3981 Fix for relation reference counts problem in buffer manager(Vadim)
3982 Allow strings to span lines, like ANSI(Thomas)
3983 Fix for backward cursor with ORDER BY(Vadim)
3984 Fix avg(cash) computation(Thomas)
3985 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
3986 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
3987 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
3993 <sect1 id="release-6-2">
3994 <title>Release 6.2</title>
3997 <title>Release date</title>
3998 <simpara>1997-10-02</simpara>
4002 A dump/restore is required for those wishing to migrate data from
4003 previous releases of <productname>PostgreSQL</productname>.
4007 <title>Migration from version 6.1 to version 6.2</title>
4010 This migration requires a complete dump of the 6.1 database and a
4011 restore of the database in 6.2.
4014 Note that the <command>pg_dump</command> and <command>pg_dumpall</command> utility from 6.2 should be used
4015 to dump the 6.1 database.
4020 <title>Migration from version 1.<replaceable>x</> to version 6.2</title>
4023 Those migrating from earlier 1.* releases should first upgrade to 1.09
4024 because the COPY output format was improved from the 1.02 release.
4029 <title>Changes</title>
4035 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
4036 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
4037 from Solaris(Diab Jerius)
4038 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
4039 Check for geometric intersections at endpoints to avoid rounding ugliness(Thomas)
4040 Catch non-functional delete attempts(Vadim)
4041 Change time function names to be more consistent(Michael Reifenberg)
4042 Check for zero divides(Michael Reifenberg)
4043 Fix very old bug which made tuples changed/inserted by a commnd
4044 visible to the command itself (so we had multiple update of
4045 updated tuples, etc)(Vadim)
4046 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
4047 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
4048 Remove un-needed signal stuff from contrib/pginterface
4049 Fix OR (where x != 1 or x isnull didn't return tuples with x NULL) (Vadim)
4050 Fix time_cmp function (Vadim)
4051 Fix handling of functions with non-attribute first argument in
4052 WHERE clauses (Vadim)
4053 Fix GROUP BY when order of entries is different from order
4054 in target list (Vadim)
4055 Fix pg_dump for aggregates without sfunc1 (Vadim)
4059 Default genetic optimizer GEQO parameter is now 8(Bruce)
4060 Allow use parameters in target list having aggregates in functions(Vadim)
4061 Added JDBC driver as an interface(Adrian & Peter)
4063 Return number of tuples inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
4064 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
4065 SPI (Server Programming Interface) allows execution of queries inside
4067 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
4068 Include reserved words for string handling, outer joins, and unions(Thomas)
4069 Implement extended comments ("/* ... */") using exclusive states(Thomas)
4070 Add "//" single-line comments(Bruce)
4071 Remove some restrictions on characters in operator names(Thomas)
4072 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
4073 Add text concatenation operator and function (SQL92)(Thomas)
4074 Support WITH TIME ZONE syntax (SQL92)(Thomas)
4075 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
4076 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
4077 and CHARACTER VARYING (SQL92)(Thomas)
4078 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
4079 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
4080 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
4081 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
4082 Add more reserved words, mostly for SQL92 compliance(Thomas)
4083 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
4084 Add center() routines for lseg, path, polygon(Thomas)
4085 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
4086 Check explicitly for points and polygons contained within polygons
4087 using an axis-crossing algorithm(Thomas)
4088 Add routine to convert circle-box(Thomas)
4089 Merge conflicting operators for different geometric data types(Thomas)
4090 Replace distance operator "<===>" with "<->"(Thomas)
4091 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Thomas)
4092 Add routines for text trimming on both ends, substring, and string position(Thomas)
4093 Added conversion routines circle(box) and poly(circle)(Thomas)
4094 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
4095 Allow functions and operators on internally-identical types to succeed(Bruce)
4096 Speed up backend start-up after profiling analysis(Bruce)
4097 Inline frequently called functions for performance(Bruce)
4098 Reduce open() calls(Bruce)
4099 psql: Add PAGER for \h and \?,\C fix
4100 Fix for psql pager when no tty(Bruce)
4101 New entab utility(Bruce)
4102 General trigger functions for referential integrity (Vadim)
4103 General trigger functions for time travel (Vadim)
4104 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
4105 MOVE implementation (Vadim)
4109 HP-UX 10 patches (Vladimir Turin)
4110 Added SCO support, (Daniel Harris)
4111 MkLinux patches (Tatsuo Ishii)
4112 Change geometric box terminology from "length" to "width"(Thomas)
4113 Deprecate temporary unstored slope fields in geometric code(Thomas)
4114 Remove restart instructions from INSTALL(Bruce)
4115 Look in /usr/ucb first for install(Bruce)
4116 Fix c++ copy example code(Thomas)
4117 Add -o to psql manual page(Bruce)
4118 Prevent relname unallocated string length from being copied into database(Bruce)
4119 Cleanup for NAMEDATALEN use(Bruce)
4120 Fix pg_proc names over 15 chars in output(Bruce)
4121 Add strNcpy() function(Bruce)
4122 remove some (void) casts that are unnecessary(Bruce)
4123 new interfaces directory(Marc)
4124 Replace fopen() calls with calls to fd.c functions(Bruce)
4125 Make functions static where possible(Bruce)
4126 enclose unused functions in #ifdef NOT_USED(Bruce)
4127 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
4128 Changes for Digital Unix
4129 Portability fix for pg_dumpall(Bruce)
4130 Rename pg_attribute.attnvals to attdispersion(Bruce)
4131 "intro/unix" manual page now "pgintro"(Bruce)
4132 "built-in" manual page now "pgbuiltin"(Bruce)
4133 "drop" manual page now "drop_table"(Bruce)
4134 Add "create_trigger", "drop_trigger" manual pages(Thomas)
4135 Add constraints regression test(Vadim & Thomas)
4136 Add comments syntax regression test(Thomas)
4137 Add PGINDENT and support program(Bruce)
4138 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
4139 Files moved to /src/tools directory(Bruce)
4140 SPI and Trigger programming guides (Vadim & D'Arcy)
4146 <sect1 id="release-6-1-1">
4147 <title>Release 6.1.1</title>
4150 <title>Release date</title>
4151 <simpara>1997-07-22</simpara>
4155 <title>Migration from version 6.1 to version 6.1.1</title>
4158 This is a minor bug-fix release. A dump/reload is not required from version 6.1,
4159 but is required from any release prior to 6.1.
4160 Refer to the release notes for 6.1 for more details.
4165 <title>Changes</title>
4169 fix for SET with options (Thomas)
4170 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
4171 new psql \connect option allows changing usernames without changing databases
4172 fix for initdb --debug option(Yoshihiko Ichikawa))
4173 lextest cleanup(Bruce)
4175 fix date/time month boundary arithmetic(Thomas)
4176 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
4177 timestamp overhauled to use standard functions(Thomas)
4178 other code cleanup in date/time routines(Thomas)
4179 psql's \d now case-insensitive(Bruce)
4180 psql's backslash commands can now have trailing semicolon(Bruce)
4181 fix memory leak in psql when using \g(Bruce)
4182 major fix for endian handling of communication to server(Thomas, Tatsuo)
4183 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
4184 allow underscores in usernames(Bruce)
4185 pg_dumpall now returns proper status, portability fix(Bruce)
4191 <sect1 id="release-6-1">
4192 <title>Release 6.1</title>
4195 <title>Release date</title>
4196 <simpara>1997-06-08</simpara>
4200 The regression tests have been adapted and extensively modified for the
4201 6.1 release of <productname>PostgreSQL</productname>.
4205 Three new data types (<type>datetime</type>, <type>timespan</type>, and <type>circle</type>) have been added to
4206 the native set of <productname>PostgreSQL</productname> types. Points, boxes, paths, and polygons
4207 have had their output formats made consistent across the data types.
4208 The polygon output in misc.out has only been spot-checked for correctness
4209 relative to the original regression output.
4213 <productname>PostgreSQL</productname> 6.1 introduces a new, alternate
4214 optimizer which uses <firstterm>genetic</firstterm>
4215 algorithms. These algorithms introduce a random behavior in the ordering
4216 of query results when the query contains multiple qualifiers or multiple
4217 tables (giving the optimizer a choice on order of evaluation). Several
4218 regression tests have been modified to explicitly order the results, and
4219 hence are insensitive to optimizer choices. A few regression tests are
4220 for data types which are inherently unordered (e.g. points and time
4221 intervals) and tests involving those types are explicitly bracketed with
4222 <command>set geqo to 'off'</command> and <command>reset geqo</command>.
4226 The interpretation of array specifiers (the curly braces around atomic
4227 values) appears to have changed sometime after the original regression
4228 tests were generated. The current <filename>./expected/*.out</filename> files reflect this
4229 new interpretation, which may not be correct!
4233 The float8 regression test fails on at least some platforms. This is due
4234 to differences in implementations of <function>pow()</function> and <function>exp()</function> and the signaling
4235 mechanisms used for overflow and underflow conditions.
4239 The <quote>random</> results in the random test should cause the
4240 <quote>random</quote> test to be <quote>failed</quote>, since the
4241 regression tests are evaluated using a simple diff. However,
4242 <quote>random</> does not seem to produce random results on my test
4243 machine (Linux/<application>gcc</>/i686).
4247 <title>Migration to version 6.1</title>
4250 This migration requires a complete dump of the 6.0 database and a
4251 restore of the database in 6.1.
4254 Those migrating from earlier 1.* releases should first upgrade to 1.09
4255 because the COPY output format was improved from the 1.02 release.
4260 <title>Changes</title>
4266 packet length checking in library routines
4267 lock manager priority patch
4268 check for under/over flow of float8(Bruce)
4269 multitable join fix(Vadim)
4270 SIGPIPE crash fix(Darren)
4271 large object fixes(Sven)
4272 allow btree indexes to handle NULLs(Vadim)
4273 timezone fixes(D'Arcy)
4274 select SUM(x) can return NULL on no rows(Thomas)
4275 internal optimizer, executor bug fixes(Vadim)
4276 fix problem where inner loop in < or <= has no rows(Vadim)
4277 prevent re-commuting join index clauses(Vadim)
4278 fix join clauses for multiple tables(Vadim)
4279 fix hash, hashjoin for arrays(Vadim)
4280 fix btree for abstime type(Vadim)
4281 large object fixes(Raymond)
4282 fix buffer leak in hash indexes (Vadim)
4283 fix rtree for use in inner scan (Vadim)
4284 fix gist for use in inner scan, cleanups (Vadim, Andrea)
4285 avoid unnecessary local buffers allocation (Vadim, Massimo)
4286 fix local buffers leak in transaction aborts (Vadim)
4287 fix file manager memmory leaks, cleanups (Vadim, Massimo)
4288 fix storage manager memmory leaks (Vadim)
4289 fix btree duplicates handling (Vadim)
4290 fix deleted tuples re-incarnation caused by vacuum (Vadim)
4291 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
4292 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
4296 attribute optimization statistics(Bruce)
4297 much faster new btree bulk load code(Paul)
4298 BTREE UNIQUE added to bulk load code(Vadim)
4299 new lock debug code(Massimo)
4300 massive changes to libpg++(Leo)
4301 new GEQO optimizer speeds table multitable optimization(Martin)
4302 new WARN message for non-unique insert into unique key(Marc)
4303 update x=-3, no spaces, now valid(Bruce)
4304 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
4305 debug backend now pretty-prints tree(Darren)
4306 new Oracle character functions(Edmund)
4307 new plaintext password functions(Dan)
4308 no such class or insufficient privilege changed to distinct messages(Dan)
4309 new ANSI timestamp function(Dan)
4310 new ANSI Time and Date types (Thomas)
4311 move large chunks of data in backend(Martin)
4312 multicolumn btree indexes(Vadim)
4313 new SET var TO value command(Martin)
4314 update transaction status on reads(Dan)
4315 new locale settings for character types(Oleg)
4316 new SEQUENCE serial number generator(Vadim)
4317 GROUP BY function now possible(Vadim)
4318 re-organize regression test(Thomas,Marc)
4319 new optimizer operation weights(Vadim)
4320 new psql \z grant/permit option(Marc)
4321 new MONEY data type(D'Arcy,Thomas)
4322 tcp socket communication speed improved(Vadim)
4323 new VACUUM option for attribute statistics, and for certain columns (Vadim)
4324 many geometric type improvements(Thomas,Keith)
4325 additional regression tests(Thomas)
4326 new datestyle variable(Thomas,Vadim,Martin)
4327 more comparison operators for sorting types(Thomas)
4328 new conversion functions(Thomas)
4329 new more compact btree format(Vadim)
4330 allow pg_dumpall to preserve database ownership(Bruce)
4331 new SET GEQO=# and R_PLANS variable(Vadim)
4332 old (!GEQO) optimizer can use right-sided plans (Vadim)
4333 typechecking improvement in SQL parser(Bruce)
4334 new SET, SHOW, RESET commands(Thomas,Vadim)
4335 new \connect database USER option
4336 new destroydb -i option (Igor)
4337 new \dt and \di psql commands (Darren)
4338 SELECT "\n" now escapes newline (A. Duursma)
4339 new geometry conversion functions from old format (Thomas)
4343 new configuration script(Marc)
4344 readline configuration option added(Marc)
4345 OS-specific configuration options removed(Marc)
4346 new OS-specific template files(Marc)
4347 no more need to edit Makefile.global(Marc)
4348 re-arrange include files(Marc)
4349 nextstep patches (Gregor Hoffleit)
4350 removed WIN32-specific code(Bruce)
4351 removed postmaster -e option, now only postgres -e option (Bruce)
4352 merge duplicate library code in front/backends(Martin)
4353 now works with eBones, international Kerberos(Jun)
4354 more shared library support
4355 c++ include file cleanup(Bruce)
4356 warn about buggy flex(Bruce)
4357 DG/UX, Ultrix, IRIX, AIX portability fixes
4363 <sect1 id="release-6-0">
4364 <title>Release 6.0</title>
4367 <title>Release date</title>
4368 <simpara>1997-01-29</simpara>
4372 A dump/restore is required for those wishing to migrate data from
4373 previous releases of <productname>PostgreSQL</productname>.
4377 <title>Migration from version 1.09 to version 6.0</title>
4380 This migration requires a complete dump of the 1.09 database and a
4381 restore of the database in 6.0.
4386 <title>Migration from pre-1.09 to version 6.0</title>
4389 Those migrating from earlier 1.* releases should first upgrade to 1.09
4390 because the COPY output format was improved from the 1.02 release.
4395 <title>Changes</title>
4401 ALTER TABLE bug - running postgress process needs to re-read table definition
4402 Allow vacuum to be run on one table or entire database(Bruce)
4404 Fix array over-runs of memory writes(Kurt)
4405 Fix elusive btree range/non-range bug(Dan)
4406 Fix for hash indexes on some types like time and date
4407 Fix for pg_log size explosion
4408 Fix permissions on lo_export()(Bruce)
4409 Fix unitialized reads of memory(Kurt)
4410 Fixed ALTER TABLE ... char(3) bug(Bruce)
4411 Fixed a few small memory leaks
4412 Fixed EXPLAIN handling of options and changed full_path option name
4413 Fixed output of group acl permissions
4414 Memory leaks (hunt and destroy with tools like Purify(Kurt)
4415 Minor improvements to rules system
4417 New asserts for run-checking
4418 Overhauled parser/analyze code to properly report errors and increase speed
4419 Pg_dump -d now handles NULL's properly(Bruce)
4420 Prevent SELECT NULL from crashing server (Bruce)
4421 Properly report errors when INSERT ... SELECT columns did not match
4422 Properly report errors when insert column names were not correct
4423 Psql \g filename now works(Bruce)
4424 Psql fixed problem with multiple statements on one line with multiple outputs
4425 Removed duplicate system OIDs
4426 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
4427 Several fixes for queries that crashed the backend
4428 Starting quote in insert string errors(Bruce)
4429 Submitting an empty query now returns empty status, not just " " query(Bruce)
4433 Add EXPLAIN manual page(Bruce)
4434 Add UNIQUE index capability(Dan)
4435 Add hostname/user level access control rather than just hostname and user
4436 Add synonym of != for <>(Bruce)
4437 Allow "select oid,* from table"
4438 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Bruce)
4439 Allow COPY from the frontend(Bryan)
4440 Allow GROUP BY to use alias column name(Bruce)
4441 Allow actual compression, not just reuse on the same page(Vadim)
4442 Allow installation-configuration option to auto-add all local users(Bryan)
4443 Allow libpq to distinguish between text value '' and null(Bruce)
4444 Allow non-postgres users with createdb privs to destroydb's
4445 Allow restriction on who can create C functions(Bryan)
4446 Allow restriction on who can do backend COPY(Bryan)
4447 Can shrink tables, pg_time and pg_log(Vadim & Erich)
4448 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
4449 Change default decimal constant representation from float4 to float8(Bruce)
4450 European date format now set when postmaster is started
4451 Execute lowercase function names if not found with exact case
4452 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z'
4453 Gist now included in the distrubution(Marc)
4454 Idend authentication of local users(Bryan)
4455 Implement BETWEEN qualifier(Bruce)
4456 Implement IN qualifier(Bruce)
4457 Libpq has PQgetisnull()(Bruce)
4458 Libpq++ improvements
4459 New options to initdb(Bryan)
4460 Pg_dump allow dump of OIDs(Bruce)
4461 Pg_dump create indexes after tables are loaded for speed(Bruce)
4462 Pg_dumpall dumps all databases, and the user table
4463 Pginterface additions for NULL values(Bruce)
4464 Prevent postmaster from being run as root
4465 Psql \h and \? is now readable(Bruce)
4466 Psql allow backslashed, semicolons anywhere on the line(Bruce)
4467 Psql changed command prompt for lines in query or in quotes(Bruce)
4468 Psql char(3) now displays as (bp)char in \d output(Bruce)
4469 Psql return code now more accurate(Bryan?)
4470 Psql updated help syntax(Bruce)
4471 Re-visit and fix vacuum(Vadim)
4472 Reduce size of regression diffs, remove timezone name difference(Bruce)
4473 Remove compile-time parameters to enable binary distributions(Bryan)
4474 Reverse meaning of HBA masks(Bryan)
4475 Secure Authentication of local users(Bryan)
4476 Speed up vacuum(Vadim)
4477 Vacuum now had VERBOSE option(Bruce)
4481 All functions now have prototypes that are compared against the calls
4482 Allow asserts to be disabled easly from Makefile.global(Bruce)
4483 Change oid constants used in code to #define names
4484 Decoupled sparc and solaris defines(Kurt)
4485 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
4486 Major include file reorganization/reduction(Marc)
4487 Make now stops on compile failure(Bryan)
4488 Makefile restructuring(Bryan, Marc)
4489 Merge bsdi_2_1 to bsdi(Bruce)
4490 Monitor program removed
4491 Name change from Postgres95 to PostgreSQL
4492 New config.h file(Marc, Bryan)
4493 PG_VERSION now set to 6.0 and used by postmaster
4494 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
4495 Reduced the number of #define's, centeralized #define's
4496 Remove duplicate OIDS in system tables(Dan)
4497 Remove duplicate system catalog info or report mismatches(Dan)
4498 Removed many os-specific #define's
4499 Restructured object file generation/location(Bryan, Marc)
4500 Restructured port-specific file locations(Bryan, Marc)
4501 Unused/uninialized variables corrected
4507 <sect1 id="release-1-09">
4508 <title>Release 1.09</title>
4511 <title>Release date</title>
4512 <simpara>1996-11-04</simpara>
4516 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of
4517 the changes listed in 6.0 were actually included in the 1.02.1 to 1.09
4522 <sect1 id="release-1-02">
4523 <title>Release 1.02</title>
4526 <title>Release date</title>
4527 <simpara>1996-08-01</simpara>
4531 <title>Migration from version 1.02 to version 1.02.1</title>
4534 Here is a new migration file for 1.02.1. It includes the 'copy' change
4535 and a script to convert old <acronym>ASCII</acronym> files.
4539 The following notes are for the benefit of users who want to migrate
4540 databases from <productname>Postgres95</> 1.01 and 1.02 to <productname>Postgres95</> 1.02.1.
4543 If you are starting afresh with <productname>Postgres95</> 1.02.1 and do not need
4544 to migrate old databases, you do not need to read any further.
4549 In order to upgrade older <productname>Postgres95</> version 1.01 or 1.02 databases to
4550 version 1.02.1, the following steps are required:
4555 Start up a new 1.02.1 postmaster
4560 Add the new built-in functions and operators of 1.02.1 to 1.01 or 1.02
4561 databases. This is done by running the new 1.02.1 server against
4562 your own 1.01 or 1.02 database and applying the queries attached at
4563 the end of the file. This can be done easily through <command>psql</>. If your
4564 1.01 or 1.02 database is named <literal>testdb</literal> and you have cut the commands
4565 from the end of this file and saved them in <filename>addfunc.sql</filename>:
4567 % psql testdb -f addfunc.sql
4570 Those upgrading 1.02 databases will get a warning when executing the
4571 last two statements in the file because they are already present in 1.02. This is
4572 not a cause for concern.
4579 <title>Dump/Reload Procedure</title>
4582 If you are trying to reload a pg_dump or text-mode, <literal>copy tablename to
4583 stdout</literal> generated with a previous version, you will need to run the
4584 attached <command>sed</command> script on the ASCII file before loading it into the
4585 database. The old format used '.' as end-of-data, while '\.' is now the
4586 end-of-data marker. Also, empty strings are now loaded in as '' rather
4587 than NULL. See the copy manual page for full details.
4590 sed 's/^\.$/\\./g' <in_file >out_file
4594 If you are loading an older binary copy or non-<systemitem>stdout</> copy, there is no
4595 end-of-data character, and hence no conversion necessary.
4598 -- following lines added by agc to reflect the case-insensitive
4599 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
4600 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexeq);
4601 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregexne);
4602 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregexeq);
4603 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticregexne);
4609 <title>Changes</title>
4613 Source code maintenance and development
4614 * worldwide team of volunteers
4615 * the source tree now in CVS at ftp.ki.net
4618 * psql (and underlying libpq library) now has many more options for
4619 formatting output, including HTML
4620 * pg_dump now output the schema and/or the data, with many fixes to
4621 enhance completeness.
4622 * psql used in place of monitor in administration shell scripts.
4623 monitor to be deprecated in next release.
4624 * date/time functions enhanced
4625 * NULL insert/update/comparison fixed/enhanced
4626 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
4628 Bug Fixes (almost too numerous to mention)
4630 * storage management
4631 * check for NULL pointer before dereferencing
4635 * added SolarisX86 port
4636 * added BSD/OS 2.1 port
4641 Contributors (appologies to any missed)
4642 * Kurt J. Lidl <lidl@va.pubnix.com>
4643 (missed in first run, but no less important)
4644 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
4645 * Jason Wright <jason@shiloh.vnet.net>
4646 * Cees de Groot <C.deGroot@inter.NL.net>
4647 * ernst.molitor@uni-bonn.de
4648 * michael.siebenborn@ae3.Hypo.DE (Michael Siebenborn (6929))
4649 * Brian E. Gallew <geek+@cmu.edu>
4650 * Vadim B. Mikheev <vadim@sable.krasnoyarsk.su>
4651 * Adam Sussman <myddryn@vidya.com>
4652 * Chris Dunlop <chris@onthe.net.au>
4653 * Marc G. Fournier <scrappy@ki.net>
4654 * Dan McGuirk <mcguirk@indirect.com>
4655 * Dr_George_D_Detlefsen <drgeorge@ilt.com>
4656 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
4657 * Massimo Dal Zotto <dz@cs.unitn.it>
4658 * Randy Kunkee <kunkee@Starbase.NeoSoft.COM>
4659 * Rick Weldon <rick@wisetech.com>
4660 * Thomas van Reimersdahl <reimersd@dali.techinfo.rwth-aachen.de>
4661 * david bennett <dave@bensoft.com>
4662 * ernst.molitor@uni-bonn.de
4663 * Julian Assange <proff@suburbia.net>
4664 * Bruce Momjian <maillist@candle.pha.pa.us>
4665 * Paul "Shag" Walmsley <ccshag@cclabs.missouri.edu>
4666 * "Alistair G. Crooks" <azcb0@sde.uts.amdahl.com>
4671 <sect1 id="release-1-01">
4672 <title>Release 1.01</title>
4675 <title>Release date</title>
4676 <simpara>1996-02-23</simpara>
4681 <title>Migration from version 1.0 to version 1.01</title>
4684 The following notes are for the benefit of users who want to migrate
4685 databases from <productname>Postgres95</> 1.0 to <productname>Postgres95</> 1.01.
4688 If you are starting afresh with <productname>Postgres95</> 1.01 and do not need
4689 to migrate old databases, you do not need to read any further.
4692 In order to <productname>Postgres95</> version 1.01 with databases created with
4693 <productname>Postgres95</> version 1.0, the following steps are required:
4698 Set the definition of <symbol>NAMEDATALEN</symbol> in <filename>src/Makefile.global</filename> to 16
4699 and <symbol>OIDNAMELEN</symbol> to 20.
4704 Decide whether you want to use Host based authentication.
4709 If you do, you must create a file name <literal>pg_hba</literal> in your top-level data
4710 directory (typically the value of your <envar>$PGDATA</envar>). <filename>src/libpq/pg_hba</filename>
4711 shows an example syntax.
4716 If you do not want host-based authentication, you can comment out
4721 in <filename>src/Makefile.global</filename>
4724 Note that host-based authentication is turned on by default, and if
4725 you do not take steps A or B above, the out-of-the-box 1.01 will
4726 not allow you to connect to 1.0 databases.
4734 Compile and install 1.01, but DO NOT do the <command>initdb</command> step.
4739 Before doing anything else, terminate your 1.0 postmaster, and
4740 backup your existing <envar>$PGDATA</envar> directory.
4745 Set your <envar>PGDATA</envar> environment variable to your 1.0 databases, but set up
4746 path up so that 1.01 binaries are being used.
4751 Modify the file <filename><envar>$PGDATA</envar>/PG_VERSION</filename> from 5.0 to 5.1
4756 Start up a new 1.01 postmaster
4761 Add the new built-in functions and operators of 1.01 to 1.0
4762 databases. This is done by running the new 1.01 server against
4763 your own 1.0 database and applying the queries attached and saving
4764 in the file 1.0_to_1.01.sql. This can be done easily through <command>psql</command>.
4765 If your 1.0 database is name <literal>testdb</literal>:
4768 % psql testdb -f 1.0_to_1.01.sql
4771 and then execute the following commands (cut and paste from here):
4774 -- add builtin functions that are new to 1.01
4776 create function int4eqoid (int4, oid) returns bool as 'foo'
4777 language 'internal';
4778 create function oideqint4 (oid, int4) returns bool as 'foo'
4779 language 'internal';
4780 create function char2icregexeq (char2, text) returns bool as 'foo'
4781 language 'internal';
4782 create function char2icregexne (char2, text) returns bool as 'foo'
4783 language 'internal';
4784 create function char4icregexeq (char4, text) returns bool as 'foo'
4785 language 'internal';
4786 create function char4icregexne (char4, text) returns bool as 'foo'
4787 language 'internal';
4788 create function char8icregexeq (char8, text) returns bool as 'foo'
4789 language 'internal';
4790 create function char8icregexne (char8, text) returns bool as 'foo'
4791 language 'internal';
4792 create function char16icregexeq (char16, text) returns bool as 'foo'
4793 language 'internal';
4794 create function char16icregexne (char16, text) returns bool as 'foo'
4795 language 'internal';
4796 create function texticregexeq (text, text) returns bool as 'foo'
4797 language 'internal';
4798 create function texticregexne (text, text) returns bool as 'foo'
4799 language 'internal';
4801 -- add builtin functions that are new to 1.01
4803 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
4804 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
4805 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexeq);
4806 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregexne);
4807 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexeq);
4808 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregexne);
4809 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexeq);
4810 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregexne);
4811 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icregexeq);
4812 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icregexne);
4813 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq);
4814 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne);
4822 <title>Changes</title>
4827 * 1.01 is backwards compatible with 1.0 database provided the user
4828 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
4829 If those steps are not taken, 1.01 is not compatible with 1.0 database.
4832 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
4833 * added NeXT port (requires SysVIPC implementation)
4834 * added CAST .. AS ... syntax
4835 * added ASC and DESC keywords
4836 * added 'internal' as a possible language for CREATE FUNCTION
4837 internal functions are C functions which have been statically linked
4838 into the postgres backend.
4839 * a new type "name" has been added for system identifiers (table names,
4840 attribute names, etc.) This replaces the old char16 type. The
4841 of name is set by the NAMEDATALEN #define in src/Makefile.global
4842 * a readable reference manual that describes the query language.
4843 * added host-based access control. A configuration file ($PGDATA/pg_hba)
4844 is used to hold the configuration data. If host-based access control
4845 is not desired, comment out HBA=1 in src/Makefile.global.
4846 * changed regex handling to be uniform use of Henry Spencer's regex code
4847 regardless of platform. The regex code is included in the distribution
4848 * added functions and operators for case-insensitive regular expressions.
4849 The operators are ~* and !~*.
4850 * pg_dump uses COPY instead of SELECT loop for better performance
4853 * fixed an optimizer bug that was causing core dumps when
4854 functions calls were used in comparisons in the WHERE clause
4855 * changed all uses of getuid to geteuid so that effective uids are used
4856 * psql now returns non-zero status on errors when using -c
4857 * applied public patches 1-14
4863 <sect1 id="release-1-0">
4864 <title>Release 1.0</title>
4867 <title>Release date</title>
4868 <simpara>1995-09-05</simpara>
4872 <title>Changes</title>
4877 * The copyright of <productname>Postgres</productname> 1.0 has been loosened to be freely modifiable
4878 and modifiable for any purpose. Please read the COPYRIGHT file.
4879 Thanks to Professor Michael Stonebraker for making this possible.
4882 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
4883 EUROPEAN STYLE). This follows SQL-92 specs.
4884 * "delimiters" is now a keyword
4887 * sql LIKE syntax has been added
4888 * copy command now takes an optional USING DELIMITER specification.
4889 delimiters can be any single-character string.
4890 * IRIX 5.3 port has been added.
4891 Thanks to Paul Walmsley and others.
4892 * updated pg_dump to work with new libpq
4893 * \d has been added psql
4894 Thanks to Keith Parks
4895 * regexp performance for architectures that use POSIX regex has been
4896 improved due to caching of precompiled patterns.
4897 Thanks to Alistair Crooks
4898 * a new version of libpq++
4899 Thanks to William Wanders
4902 * arbitrary userids can be specified in the createuser script
4903 * \c to connect to other databases in psql now works.
4904 * bad pg_proc entry for float4inc() is fixed
4905 * users with usecreatedb field set can now create databases without
4906 having to be usesuper
4907 * remove access control entries when the entry no longer has any
4909 * fixed non-portable datetimes implementation
4910 * added kerberos flags to the src/backend/Makefile
4911 * libpq now works with kerberos
4912 * typographic errors in the user manual have been corrected.
4913 * btrees with multiple index never worked, now we tell you they don't
4914 work when you try to use them
4920 <sect1 id="release-0-03">
4921 <title><productname>Postgres95</productname> Release 0.03</title>
4924 <title>Release date</title>
4925 <simpara>1995-07-21</simpara>
4929 <title>Changes</title>
4932 Incompatible changes:
4933 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
4934 (due to system catalog changes and indexing structure changes).
4935 * double-quote (") is deprecated as a quoting character for string literals;
4936 you need to convert them to single quotes ('). <!-- " -->
4937 * name of aggregates (eg. int4sum) are renamed in accordance with the
4938 SQL standard (eg. sum).
4939 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
4940 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
4941 previous releases); you might have to do typecasting if you depend on it
4942 being of type float8. If you neglect to do the typecasting and you assign
4943 a float literal to a field of type float8, you may get incorrect values
4945 * LIBPQ has been totally revamped so that frontend applications
4946 can connect to multiple backends
4947 * the usesysid field in pg_user has been changed from int2 to int4 to
4948 allow wider range of Unix user ids.
4949 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
4950 single BSD44_derived port. (thanks to Alistair Crooks)
4952 SQL standard-compliance (the following details changes that makes postgres95
4953 more compliant to the SQL-92 standard):
4954 * the following SQL types are now built-in: smallint, int(eger), float, real,
4955 char(N), varchar(N), date and time.
4957 The following are aliases to existing postgres types:
4959 integer, int -> int4
4960 float, real -> float4
4961 char(N) and varchar(N) are implemented as truncated text types. In
4962 addition, char(N) does blank-padding.
4963 * single-quote (') is used for quoting string literals; '' (in addition to
4964 \') is supported as means of inserting a single quote in a string
4965 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
4966 (Also, aggregates can now be overloaded, i.e. you can define your
4967 own MAX aggregate to take in a user-defined type.)
4968 * CHANGE ACL removed. GRANT/REVOKE syntax added.
4969 - Privileges can be given to a group using the "GROUP" keyword.
4971 GRANT SELECT ON foobar TO GROUP my_group;
4972 The keyword 'PUBLIC' is also supported to mean all users.
4974 Privileges can only be granted or revoked to one user or group
4977 "WITH GRANT OPTION" is not supported. Only class owners can change
4979 - The default access control is to to grant users readonly access.
4980 You must explicitly grant insert/update access to users. To change
4981 this, modify the line in
4982 src/backend/utils/acl.h
4983 that defines ACL_WORLD_DEFAULT
4986 * the bug where aggregates of empty tables were not run has been fixed. Now,
4987 aggregates run on empty tables will return the initial conditions of the
4988 aggregates. Thus, COUNT of an empty table will now properly return 0.
4989 MAX/MIN of an empty table will return a tuple of value NULL.
4990 * allow the use of \; inside the monitor
4991 * the LISTEN/NOTIFY asynchronous notification mechanism now work
4992 * NOTIFY in rule action bodies now work
4993 * hash indexes work, and access methods in general should perform better.
4994 creation of large btree indexes should be much faster. (thanks to Paul
4997 Other changes and enhancements:
4998 * addition of an EXPLAIN statement used for explaining the query execution
4999 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
5001 * WARN and NOTICE messages no longer have timestamps on them. To turn on
5002 timestamps of error messages, uncomment the line in
5003 src/backend/utils/elog.h:
5004 /* define ELOG_TIMESTAMPS */
5005 * On an access control violation, the message
5006 "Either no such class or insufficient privilege"
5007 will be given. This is the same message that is returned when
5008 a class is not found. This dissuades non-privileged users from
5009 guessing the existence of privileged classes.
5010 * some additional system catalog changes have been made that are not
5011 visible to the user.
5014 * The -oid option has been added to the "pg_result" tcl command.
5015 pg_result -oid returns oid of the last tuple inserted. If the
5016 last command was not an INSERT, then pg_result -oid returns "".
5017 * the large object interface is available as pg_lo* tcl commands:
5018 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
5020 Portability enhancements and New Ports:
5021 * flex/lex problems have been cleared up. Now, you should be able to use
5022 flex instead of lex on any platforms. We no longer make assumptions of
5023 what lexer you use based on the platform you use.
5024 * The Linux-ELF port is now supported. Various configuration have been
5025 tested: The following configuration is known to work:
5026 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
5027 with everything in ELF format,
5030 * ipcclean added to the distribution
5031 ipcclean usually does not need to be run, but if your backend crashes
5032 and leaves shared memory segments hanging around, ipcclean will
5033 clean them up for you.
5036 * the user manual has been revised and libpq documentation added.
5042 <sect1 id="release-0-02">
5043 <title><productname>Postgres95</productname> Release 0.02</title>
5046 <title>Release date</title>
5047 <simpara>1995-05-25</simpara>
5051 <title>Changes</title>
5055 Incompatible changes:
5056 * The SQL statement for creating a database is 'CREATE DATABASE' instead
5057 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
5058 of 'DESTROYDB'. However, the names of the executables 'createdb' and
5059 'destroydb' remain the same.
5062 * pgperl - a Perl (4.036) interface to Postgres95
5063 * pg_dump - a utility for dumping out a postgres database into a
5064 script file containing query commands. The script files are in a ASCII
5065 format and can be used to reconstruct the database, even on other
5066 machines and other architectures. (Also good for converting
5067 a Postgres 4.2 database to Postgres95 database.)
5069 The following ports have been incorporated into postgres95-beta-0.02:
5070 * the NetBSD port by Alistair Crooks
5071 * the AIX port by Mike Tung
5072 * the Windows NT port by Jon Forrest (more stuff but not done yet)
5073 * the Linux ELF port by Brian Gallew
5075 The following bugs have been fixed in postgres95-beta-0.02:
5076 * new lines not escaped in COPY OUT and problem with COPY OUT when first
5078 * cannot type return to use the default user id in createuser
5079 * SELECT DISTINCT on big tables crashes
5080 * Linux installation problems
5081 * monitor doesn't allow use of 'localhost' as PGHOST
5082 * psql core dumps when doing \c or \l
5083 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
5084 * libpgtcl has a hard-wired default port number
5085 * SELECT DISTINCT INTO TABLE hangs
5086 * CREATE TYPE doesn't accept 'variable' as the internallength
5087 * wrong result using more than 1 aggregate in a SELECT
5093 <sect1 id="release-0-01">
5094 <title><productname>Postgres95</productname> Release 0.01</title>
5097 <title>Release date</title>
5098 <simpara>1995-05-01</simpara>
5107 <sect1 id="timing-results">
5108 <title>Timing Results</title>
5111 These timing results are from running the regression test with the commands
5114 % cd src/test/regress
5120 Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
5121 to run, presumably due to the scheduling vagaries of multitasking systems.
5125 <title>Version 6.5</title>
5128 As has been the case for previous releases, timing between
5129 releases is not directly comparable since new regression tests
5130 have been added. In general, 6.5 is faster than previous
5135 Timing with <function>fsync()</function> disabled:
5139 02:00 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
5140 04:38 Sparc Ultra 1 143MHz, 64MB, Solaris 2.6
5145 Timing with <function>fsync()</function> enabled:
5149 04:21 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
5152 For the <systemitem class="osname">Linux</systemitem> system above, using <acronym>UW-SCSI</acronym> disks rather than (older) <acronym>IDE</acronym>
5153 disks leads to a 50% improvement in speed on the regression test.
5158 <title>Version 6.4beta</title>
5161 The times for this release are not directly comparable to those for previous releases
5162 since some additional regression tests have been included.
5163 In general, however, 6.4 should be slightly faster than the previous release (thanks, Bruce!).
5168 02:26 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
5174 <title>Version 6.3</title>
5177 The times for this release are not directly comparable to those for previous releases
5178 since some additional regression tests have been included and some obsolete tests involving
5179 time travel have been removed.
5180 In general, however, 6.3 is substantially faster than previous releases (thanks, Bruce!).
5185 02:30 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
5186 04:12 Dual Pentium Pro 180, 96MB, EIDE, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
5192 <title>Version 6.1</title>
5197 06:12 Pentium Pro 180, 32MB, EIDE, Linux 2.0.30, gcc 2.7.2 -O2 -m486
5198 12:06 P-100, 48MB, Linux 2.0.29, gcc
5199 39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
5207 <!-- Keep this comment at the end of the file
5212 sgml-minimize-attributes:nil
5213 sgml-always-quote-attributes:t
5216 sgml-parent-document:nil
5217 sgml-default-dtd-file:"./reference.ced"
5218 sgml-exposed-tags:nil
5219 sgml-local-catalogs:("/usr/lib/sgml/catalog")
5220 sgml-local-ecat-files:nil
projects / postgresql / blob