2 $Header: /cvsroot/pgsql/doc/src/sgml/release.sgml,v 1.157 2002/09/02 20:04:39 tgl Exp $
5 <appendix id="release">
6 <title>Release Notes</title>
8 <sect1 id="release-devel">
9 <title>&version; Development Branch</title>
12 Below is a subset of the changes that have gone into the
13 development branch of PostgreSQL since version 7.2. 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 CREATE OR REPLACE VIEW, CREATE OR REPLACE RULE are available
28 No-autocommit mode is available (set autocommit to off)
29 Substantial improvements in functionality for functions returning sets
30 Client libraries older than 6.3 no longer supported (version 0 protocol removed)
31 PREPARE statement allows caching query plans for interactive statements
32 Type OPAQUE is now deprecated in favor of pseudo-types cstring, trigger, etc
33 Standalone composite types can now be created with CREATE TYPE
34 Files larger than 2 GB are now supported (if supported by the operating system)
35 SERIAL no longer implies UNIQUE; specify explicitly if index is wanted
36 pg_dump -n and -N options have been removed. The new behavior is like -n but knows about key words.
37 CLUSTER is no longer hazardous to your schema
38 COPY accepts a list of columns to copy
39 ALTER TABLE DROP COLUMN
40 CREATE OPERATOR CLASS/DROP OPERATOR CLASS
42 Sequences created by SERIAL column definitions now auto-drop with the column
43 Most forms of DROP now support RESTRICT and CASCADE options
44 Recursive SQL functions can be defined
45 User-defined procedural languages can register a validator function to check new functions as they are created
46 Functions can be executed with the privileges of the owner
47 Syntax of CREATE FUNCTION has been extended to resemble SQL99
48 Effects of SET within a transaction block now roll back if transaction aborts
49 New SET LOCAL syntax sets a parameter for the life of the current transaction
50 Datestyle, timezone, client_encoding can be set in postgresql.conf
51 The last vestiges of support for type names datetime and timespan are gone; use timestamp and interval instead
52 Rule names are now per-relation, not global; DROP RULE and COMMENT ON RULE syntax changes accordingly
53 Readline and Zlib are now required by default and must be turned off explicitly if their use is not desired
54 Define a third class of function volatility to allow indexscans in more cases
55 Locale support is now built by default; choice of locale is set by initdb and/or at run-time
56 ALTER TABLE ALTER COLUMN SET/DROP NOT NULL
57 EXPLAIN and SHOW output comes out as a query result, not a NOTICE message
58 DOMAINs (types that are constrained versions of base types)
59 Access privileges on functions
60 Access privileges on procedural languages
61 CREATE DATABASE has OWNER option so superuser can create DB for someone else
62 Kerberos 5 support now works with Heimdal
63 Database and user-specific session defaults for run-time configuration variables (ALTER DATABASE ... SET and ALTER USER ... SET)
64 String function OVERLAY() implemented per SQL99
65 Regular expression operator SIMILAR TO implemented per SQL99
66 Regular expression function SUBSTRING() implemented per SQL99
71 <sect1 id="release-7-2-2">
72 <title>Release 7.2.2</title>
75 <title>Release date</title>
76 <simpara>2002-08-23</simpara>
80 This has a variety of fixes from 7.2.1.
85 <title>Migration to version 7.2.2</title>
88 A dump/restore is <emphasis>not</emphasis> required for those running
94 <title>Changes</title>
98 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/PgSQL (Tom)
99 Fix for compressed transaction log id wraparound (Tom)
100 Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Tatsuo)
101 Fix for psql and pg_dump crashing when invoked with non-existand long
103 Fix crash when invoking geometric operators (Tom)
104 Allow OPEN cursor(args) (Tom)
105 Fix for rtree_gist index build (Teodor)
106 Fix for dumping user-defined aggregates (Tom)
107 Contrib/intarray fixes (Oleg)
108 Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)
109 Fix to pg_convert (Tatsuo)
110 Fix for crash with long DATA strings (Thomes, Neil)
111 Fix for repeat(), lpad(), rpad() and long strings (Neil)
118 <sect1 id="release-7-2-1">
119 <title>Release 7.2.1</title>
122 <title>Release date</title>
123 <simpara>2002-03-21</simpara>
127 This has a variety of fixes from 7.2.
132 <title>Migration to version 7.2.1</title>
135 A dump/restore is <emphasis>not</emphasis> required for those running
141 <title>Changes</title>
145 Ensure that sequence counters do not go backwards after a crash (Tom)
146 Fix pgaccess kanji-coversion key binding (Tatsuo)
147 Optimizer improvements (Tom)
148 cash I/O improvements (Tom)
150 Compile fix for missing AuthBlockSig (Heiko)
151 Additional time zones and time zone fixes (Thomas)
152 Allow psql \connect to handle mixed case database and user names (Tom)
153 Return proper OID on command completion even with ON INSERT rules (Tom)
154 Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)
155 Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)
156 Improve handling of multiple UNIONs with different lengths (Tom)
157 contrib/btree_gist improvements (Teodor Sigaev)
158 contrib/tsearch dictionary improvements, see README.tsearch for
159 an additional installation step (Thomas T. Thai, Teodor Sigaev)
160 Fix for array subscripts handling (Tom)
161 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/PgSQL (Tom)
168 <sect1 id="release-7-2">
169 <title>Release 7.2</title>
172 <title>Release date</title>
173 <simpara>2002-02-04</simpara>
177 <title>Overview</title>
180 This release improves <productname>PostgreSQL</> for use in
181 high-volume applications.
185 Major changes in this release:
192 Vacuuming no longer locks tables, thus allowing normal user
193 access during the vacuum. A new <command>VACUUM FULL</>
194 command does old-style vacuum by locking the table and
195 shrinking the on-disk copy of the table.
201 <term>Transactions</term>
204 There is no longer a problem with installations that exceed
205 four billion transactions.
214 OIDs are now optional. Users can now create tables without
215 OIDs for cases where OID usage is excessive.
221 <term>Optimizer</term>
224 The system now computes histogram column statistics during
225 <command>ANALYZE</>, allowing much better optimizer choices.
231 <term>Security</term>
234 A new MD5 encryption option allows more secure storage and
235 transfer of passwords. A new Unix-domain socket
236 authentication option is available on Linux and BSD systems.
242 <term>Statistics</term>
245 Administrators can use the new table access statistics module
246 to get fine-grained information about table and index usage.
252 <term>Internationalization</term>
255 Program and library messages can now be displayed in several
265 <title>Migration to version 7.2</title>
268 A dump/restore using <command>pg_dump</command> is required for
269 those wishing to migrate data from any previous release.
273 Observe the following incompatibilities:
278 The semantics of the <command>VACUUM</command> command have
279 changed in this release. You may wish to update your
280 maintenance procedures accordingly.
286 In this release, comparisons using <literal>= NULL</literal>
287 will always return false (or NULL, more precisely). Previous
288 releases automatically transformed this syntax to <literal>IS
289 NULL</literal>. The old behavior can be re-enabled using a
290 <filename>postgresql.conf</filename> parameter.
296 The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
297 configuration is now only reloaded after receiving a
298 <systemitem>SIGHUP</> signal, not with each connection.
304 The function <filename>octet_length()</> now returns the uncompressed data length.
310 The date/time value <literal>'current'</literal> is no longer
311 available. You will need to rewrite your applications.
317 The <literal>timestamp()</literal>, <literal>time()</literal>,
318 and <literal>interval()</literal> functions are no longer
319 available. Instead of <literal>timestamp()</literal>, use
320 <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
330 The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
331 in the next release. You should change your queries to use
332 separate LIMIT and OFFSET clauses, e.g. <literal>LIMIT 10 OFFSET
338 <title>Changes</title>
341 <title>Server Operation</title>
343 Create temporary files in a separate directory (Bruce)
344 Delete orphaned temporary files on postmaster startup (Bruce)
345 Added unique indexes to some system tables (Tom)
346 System table operator reorganization (Oleg Bartunov, Teodor Sigaev, Tom)
347 Renamed pg_log to pg_clog (Tom)
348 Enable SIGTERM, SIGQUIT to kill backends (Jan)
349 Removed compile-time limit on number of backends (Tom)
350 Better cleanup for semaphore resource failure (Tatsuo, Tom)
351 Allow safe transaction ID wraparound (Tom)
352 Removed OIDs from some system tables (Tom)
353 Removed "triggered data change violation" error check (Tom)
354 SPI portal creation of prepared/saved plans (Jan)
355 Allow SPI column functions to work for system columns (Tom)
356 Long value compression improvement (Tom)
357 Statistics collector for table, index access (Jan)
358 Truncate extra-long sequence names to a reasonable value (Tom)
359 Measure transaction times in milliseconds (Thomas)
360 Fix TID sequential scans (Hiroshi)
361 Superuser ID now fixed at 1 (Peter E)
362 New pg_ctl "reload" option (Tom)
367 <title>Performance</title>
369 Optimizer improvements (Tom)
370 New histogram column statistics for optimizer (Tom)
371 Reuse write-ahead log files rather than discarding them (Tom)
372 Cache improvements (Tom)
373 IS NULL, IS NOT NULL optimizer improvement (Tom)
374 Improve lock manager to reduce lock contention (Tom)
375 Keep relcache entries for index access support functions (Tom)
376 Allow better selectivity with NaN and infinities in NUMERIC (Tom)
377 R-tree performance improvements (Kenneth Been)
378 B-tree splits more efficient (Tom)
383 <title>Privileges</title>
385 Change UPDATE, DELETE permissions to be distinct (Peter E)
386 New REFERENCES, TRIGGER privileges (Peter E)
387 Allow GRANT/REVOKE to/from more than one user at a time (Peter E)
388 New has_table_privilege() function (Joe Conway)
389 Allow non-superuser to vacuum database (Tom)
390 New SET SESSION AUTHORIZATION command (Peter E)
391 Fix bug in privilege modifications on newly created tables (Tom)
392 Disallow access to pg_statistic for non-superuser, add user-accessible views (Tom)
397 <title>Client Authentication</title>
399 Fork postmaster before doing authentication to prevent hangs (Peter E)
400 Add ident authentication over Unix domain sockets on Linux, *BSD (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)
401 Add a password authentication method that uses MD5 encryption (Bruce)
402 Allow encryption of stored passwords using MD5 (Bruce)
403 PAM authentication (Dominic J. Eidson)
404 Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)
409 <title>Server Configuration</title>
411 Interpretation of some time zone abbreviations as Australian rather than North American now settable at run time (Bruce)
412 New parameter to set default transaction isolation level (Peter E)
413 New parameter to enable conversion of "expr = NULL" into "expr IS NULL", off by default (Peter E)
414 New parameter to control memory usage by VACUUM (Tom)
415 New parameter to set client authentication timeout (Tom)
416 New parameter to set maximum number of open files (Tom)
421 <title>Queries</title>
423 Statements added by INSERT rules now execute after the INSERT (Jan)
424 Prevent unadorned relation names in target list (Bruce)
425 NULLs now sort after all normal values in ORDER BY (Tom)
426 New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)
427 New SHARE UPDATE EXCLUSIVE lock mode (Tom)
428 New EXPLAIN ANALYZE command that shows run times and row counts (Martijn van Oosterhout)
429 Fix problem with LIMIT and subqueries (Tom)
430 Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)
431 Fix nested EXCEPT/INTERSECT (Tom)
436 <title>Schema Manipulation</title>
438 Fix SERIAL in temporary tables (Bruce)
439 Allow temporary sequences (Bruce)
440 Sequences now use int8 internally (Tom)
441 New SERIAL8 creates int8 columns with sequences, default still SERIAL4 (Tom)
442 Make OIDs optional using WITHOUT OIDS (Tom)
443 Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)
444 Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher Kings-Lynne)
445 New CREATE OR REPLACE FUNCTION to alter existing function (preserving the function OID) (Gavin Sherry)
446 Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)
447 Allow column renaming in views
448 Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent Verner)
449 Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (Stephan Szabo)
450 ALTER TABLE RENAME update foreign-key trigger arguments correctly (Brent Verner)
451 DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)
452 Add automatic return type data casting for SQL functions (Tom)
453 Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov, Teodor Sigaev, Tom)
454 Enable partial indexes (Martijn van Oosterhout)
459 <title>Utility Commands</title>
461 Add RESET ALL, SHOW ALL (Marko Kreen)
462 CREATE/ALTER USER/GROUP now allow options in any order (Vince)
463 Add LOCK A, B, C functionality (Neil Padgett)
464 New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)
465 New light-weight VACUUM does not lock table; old semantics are available as VACUUM FULL (Tom)
466 Disable COPY TO/FROM on views (Bruce)
467 COPY DELIMITERS string must be exactly one character (Tom)
468 VACUUM warning about index tuples fewer than heap now only appears when appropriate (Martijn van Oosterhout)
469 Fix permission checks for CREATE INDEX (Tom)
470 Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)
475 <title>Data Types and Functions</title>
477 SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)
478 Add convert(), convert2() (Tatsuo)
479 New function bit_length() (Peter E)
480 Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tatsuo)
481 CHAR(), VARCHAR() now reject strings that are too long (Peter E)
482 BIT VARYING now rejects bit strings that are too long (Peter E)
483 BIT now rejects bit strings that do not match declared size (Peter E)
484 INET, CIDR text conversion functions (Alex Pilosov)
485 INET, CIDR operators << and <<= indexable (Alex Pilosov)
486 Bytea \### now requires valid three digit octal number
487 Bytea comparison improvements, now supports =, <>, >, >=, <, and <=
488 Bytea now supports B-tree indexes
489 Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE
490 Bytea now supports concatenation
491 New bytea functions: position, substring, trim, btrim, and length
492 New encode() function mode, "escaped", converts minimally escaped bytea to/from text
493 Add pg_database_encoding_max_length() (Tatsuo)
494 Add pg_client_encoding() function (Tatsuo)
495 now() returns time with millisecond precision (Thomas)
496 New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)
497 Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)
498 New xid/int comparison functions (Hiroshi)
499 Add precision to TIME, TIMESTAMP, and INVERVAL data types (Thomas)
500 Modify type coercion logic to attempt binary-compatible functions first (Tom)
501 New encode() function installed by default (Marko Kreen)
502 Improved to_*() conversion functions (Karel Zak)
503 Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)
504 New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_salt() (Marko Kreen)
505 Correct description of translate() function (Bruce)
506 Add INTERVAL argument for SET TIME ZONE (Thomas)
507 Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)
508 Optimize length functions when using single-byte encodings (Tatsuo)
509 Fix path_inter, path_distance, path_length, dist_ppath to handle closed paths (Curtis Barrett, Tom)
510 octet_length(text) now returns non-compressed length (Tatsuo, Bruce)
511 Handle "July" full name in date/time literals (Greg Sabino Mullane)
512 Some datatype() function calls now evaluated differently
513 Add support for Julian and ISO time specifications (Thomas)
518 <title>Internationalization</title>
520 National language support in psql, pg_dump, libpq, and server (Peter E)
521 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)
522 Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware (Tatsuo)
523 Add LATIN5,6,7,8,9,10 support (Tatsuo)
524 Add ISO 8859-5,6,7,8 support (Tatsuo)
525 Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)
526 Make mic2ascii() non-ASCII aware (Tatsuo)
527 Reject invalid multibyte character sequences (Tatsuo)
532 <title>PL/pgSQL</title>
534 Now uses portals for SELECT loops, allowing huge result sets (Jan)
535 CURSOR and REFCURSOR support (Jan)
536 Can now return open cursors (Jan)
537 Add ELSEIF (Klaus Reger)
538 Improve PL/pgSQL error reporting, including location of error (Tom)
539 Allow IS or FOR key words in cursor declaration, for compatibility (Bruce)
540 Fix for SELECT ... FOR UPDATE (Tom)
541 Fix for PERFORM returning multiple rows (Tom)
542 Make PL/pgSQL use the server's type coercion code (Tom)
543 Memory leak fix (Jan, Tom)
544 Make trailing semicolon optional (Tom)
549 <title>PL/Perl</title>
551 New untrusted PL/Perl (Alex Pilosov)
552 PL/Perl is now built on some platforms even if libperl is not shared (Peter E)
557 <title>PL/Tcl</title>
559 Now reports errorInfo (Vsevolod Lobko)
560 Add spi_lastoid function (bob@redivi.com)
565 <title>PL/Python</title>
567 ...is new (Andrew Bosma)
574 \d displays indexes in unique, primary groupings (Christopher Kings-Lynne)
575 Allow trailing semicolons in backslash commands (Greg Sabino Mullane)
576 Read password from /dev/tty if possible
577 Force new password prompt when changing user and database (Tatsuo, Tom)
578 Format the correct number of columns for Unicode (Patrice)
585 New function PQescapeString() to escape quotes in command strings (Florian Weimer)
586 New function PQescapeBytea() escapes binary strings for use as SQL string literals
593 Return OID of INSERT (Ken K)
594 Handle more data types (Ken K)
595 Handle single quotes and newlines in strings (Ken K)
596 Handle NULL variables (Ken K)
597 Fix for time zone handling (Barry Lind)
598 Improved Druid support
599 Allow eight-bit characters with non-multibyte server (Barry Lind)
600 Support BIT, BINARY types (Ned Wolpert)
601 Reduce memory usage (Michael Stephens, Dave Cramer)
602 Update DatabaseMetaData (Peter E)
603 Add DatabaseMetaData.getCatalogs() (Peter E)
604 Encoding fixes (Anders Bengtsson)
605 Get/setCatalog methods (Jason Davies)
606 DatabaseMetaData.getColumns() now returns column defaults (Jason Davies)
607 DatabaseMetaData.getColumns() performance improvement (Jeroen van Vianen)
608 Some JDBC1 and JDBC2 merging (Anders Bengtsson)
609 Transaction performance improvements (Barry Lind)
610 Array fixes (Greg Zoller)
612 Fix batch processing (Rene Pijlman)
613 ExecSQL method reorganization (Anders Bengtsson)
614 GetColumn() fixes (Jeroen van Vianen)
615 Fix isWriteable() function (Rene Pijlman)
616 Improved passage of JDBC2 conformance tests (Rene Pijlman)
617 Add bytea type capability (Barry Lind)
618 Add isNullable() (Rene Pijlman)
619 JDBC date/time test suite fixes (Liam Stewart)
620 Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)
621 Fix DatabaseMetaData to show precision properly (Mark Lillywhite)
622 New getImported/getExported keys (Jason Davies)
623 MD5 password encryption support (Jeremy Wohl)
624 Fix to actually use type cache (Ned Wolpert)
631 Remove query size limit (Hiroshi)
632 Remove text field size limit (Hiroshi)
633 Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)
634 Allow ODBC procedure calls (Hiroshi)
635 Improve boolean handing (Aidan Mountford)
636 Most configuration options on setable via DSN (Hiroshi)
637 Multibyte, performance fixes (Hiroshi)
638 Allow driver to be used with iODBC or unixODBC (Peter E)
639 MD5 password encryption support (Bruce)
640 Add more compatibility functions to odbc.sql (Peter E)
647 EXECUTE ... INTO implemented (Christof Petig)
648 Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)
649 Fix for GRANT parameters (Lee Kindness)
650 Fix INITIALLY DEFERRED bug
651 Various bug fixes (Michael, Christof Petig)
652 Auto allocation for indicator variable arrays (int *ind_p=NULL)
653 Auto allocation for string arrays (char **foo_pp=NULL)
654 ECPGfree_auto_mem fixed
655 All function names with external linkage are now prefixed by ECPG
656 Fixes for arrays of structures (Michael)
661 <title>Misc. Interfaces</title>
663 Python fix fetchone() (Gerhard Haring)
664 Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Max)
665 Add Tcl COPY TO/FROM (ljb)
666 Prevent output of default index op class in pg_dump (Tom)
667 Fix libpgeasy memory leak (Bruce)
672 <title>Build and Install</title>
674 Configure, dynamic loader, and shared library fixes (Peter E)
675 Fixes in QNX 4 port (Bernd Tegge)
676 Fixes in Cygwin and Win32 ports (Jason Tishler, Gerhard Haring, Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)
677 Fix for Win32 socket communication failures (Magnus, Mikhail Terekhov)
678 Hurd compile fix (Oliver Elphick)
679 BeOS fixes (Cyril Velter)
680 Remove configure --enable-unicode-conversion, now enabled by multibyte (Tatsuo)
681 AIX fixes (Tatsuo, Andreas)
682 Fix parallel make (Peter E)
683 Install SQL language manual pages into OS-specific directories (Peter E)
684 Rename config.h to pg_config.h (Peter E)
685 Reorganize installation layout of header files (Peter E)
690 <title>Source Code</title>
692 Remove SEP_CHAR (Bruce)
694 Merge GUC and command line handling (Marko Kreen)
695 Remove EXTEND INDEX (Martijn van Oosterhout, Tom)
696 New pgjindent utility to indent java code (Bruce)
697 Remove define of true/false when compiling under C++ (Leandro Fanzone, Tom)
698 pgindent fixes (Bruce, Tom)
699 Replace strcasecmp() with strcmp() where appropriate (Peter E)
700 Dynahash portability improvements (Tom)
701 Add 'volatile' usage in spinlock structures
702 Improve signal handling logic (Tom)
707 <title>Contrib</title>
709 New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)
710 New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)
711 Add contrib/dblink for remote database access (Joe Conway)
712 contrib/ora2pg Oracle conversion utility (Gilles Darold)
713 contrib/xml XML conversion utility (John Gray)
714 contrib/fulltextindex fixes (Christopher Kings-Lynne)
715 New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merged (Joe Conway)
716 Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartunov)
717 New pg_upgrade utility (Bruce)
718 Add new pg_resetxlog options (Bruce, Tom)
725 <sect1 id="release-7-1-3">
726 <title>Release 7.1.3</title>
729 <title>Release date</title>
730 <simpara>2001-08-15</simpara>
734 <title>Migration to version 7.1.3</title>
737 A dump/restore is <emphasis>not</emphasis> required for those running
743 <title>Changes</title>
747 Remove unused WAL segements of large transactions (Tom)
748 Multiaction rule fix (Tom)
749 PL/pgSQL memory allocation fix (Jan)
750 VACUUM buffer fix (Tom)
751 Regression test fixes (Tom)
752 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
753 Fix subselects with DISTINCT ON or LIMIT (Tom)
755 Disable COPY TO/FROM a view (Tom)
756 Cygwin build (Jason Tishler)
763 <sect1 id="release-7-1-2">
764 <title>Release 7.1.2</title>
767 <title>Release date</title>
768 <simpara>2001-05-11</simpara>
772 This has one fix from 7.1.1.
777 <title>Migration to version 7.1.2</title>
780 A dump/restore is <emphasis>not</emphasis> required for those running
786 <title>Changes</title>
790 Fix PL/pgSQL SELECTs when returning no rows
791 Fix for psql backslash core dump
792 Referential integrity permission fix
801 <sect1 id="release-7-1-1">
802 <title>Release 7.1.1</title>
805 <title>Release date</title>
806 <simpara>2001-05-05</simpara>
810 This has a variety of fixes from 7.1.
815 <title>Migration to version 7.1.1</title>
818 A dump/restore is <emphasis>not</emphasis> required for those running
824 <title>Changes</title>
828 Fix for numeric MODULO operator (Tom)
829 pg_dump fixes (Philip)
830 pg_dump can dump 7.0 databases (Philip)
831 readline 4.2 fixes (Peter E)
833 AIX, MSWIN, VAX, N32K fixes (Tom)
834 Multibytes fixes (Tom)
835 Unicode fixes (Tatsuo)
836 Optimizer improvements (Tom)
837 Fix for whole tuples in functions (Tom)
838 Fix for pg_ctl and option strings with spaces (Peter E)
840 EXTRACT can now take string argument (Thomas)
848 <sect1 id="release-7-1">
849 <title>Release 7.1</title>
852 <title>Release date</title>
853 <simpara>2001-04-13</simpara>
857 This release focuses on removing limitations that have existed in the
858 PostgreSQL code for many years.
862 Major changes in this release:
867 Write-ahead Log (WAL)
871 To maintain database consistency in case
872 of an operating system crash, previous releases of PostgreSQL have
873 forced all data modifications to disk before each transaction commit.
874 With WAL, only one log file must be flushed to disk, greatly improving
875 performance. If you have been using -F in previous releases to disable
876 disk flushes, you may want to consider discontinuing its use.
887 TOAST - Previous releases had a compiled-in row length limit,
888 typically 8k - 32k. This limit made storage of long text fields
889 difficult. With TOAST, long rows of any length can be stored with good
901 We now support outer joins. The UNION/NOT IN
902 workaround for outer joins is no longer required. We use the SQL92
914 The previous C function manager did not
915 handle NULLs properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
916 function manager does. You can continue using your old custom
917 functions, but you may want to rewrite them in the future to use the new
918 function manager call interface.
929 A large number of complex queries that were
930 unsupported in previous releases now work. Many combinations of views,
931 aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
932 now work properly. Inherited tables are now accessed by default.
933 Subqueries in FROM are now supported.
942 <title>Migration to version 7.1</title>
945 A dump/restore using pg_dump is required for those wishing to migrate
946 data from any previous release.
951 <title>Changes</title>
957 Many multibyte/Unicode/locale fixes (Tatsuo and others)
958 More reliable ALTER TABLE RENAME (Tom)
959 Kerberos V fixes (David Wragg)
960 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
961 Prompt username/password on standard error (Bruce)
962 Large objects inv_read/inv_write fixes (Tom)
963 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
965 Prevent query expressions from leaking memory (Tom)
966 Allow UPDATE of arrays elements (Tom)
967 Wake up lock waiters during cancel (Hiroshi)
968 Fix rare cursor crash when using hash join (Tom)
969 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
970 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
971 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
972 Fix PL/perl (Alex Kapranoff)
973 Disallow LOCK on views (Mark Hollomon)
974 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
975 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
976 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
977 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
978 Fix rare failure with TRUNCATE command (Tom)
979 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
980 DISTINCT, ORDER BY, SELECT...INTO (Tom)
981 Fix parser failures during aborted transactions (Tom)
982 Allow temporary relations to properly clean up indexes (Bruce)
983 Fix VACUUM problem with moving rows in same page (Tom)
984 Modify pg_dump to better handle user-defined items in template1 (Philip)
985 Allow LIMIT in VIEW (Tom)
986 Require cursor FETCH to honor LIMIT (Tom)
987 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
988 Allow ORDER BY, LIMIT in sub-selects (Tom)
989 Allow UNION in CREATE RULE (Tom)
990 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
991 Store initdb collation in pg_control so collation cannot be changed (Tom)
992 Fix INSERT...SELECT with rules (Tom)
993 Fix FOR UPDATE inside views and subselects (Tom)
994 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
995 Fix lpad() and rpad() to handle length less than input string (Tom)
996 Fix use of NOTIFY in some rules (Tom)
997 Overhaul btree code (Tom)
998 Fix NOT NULL use in Pl/pgSQL variables (Tom)
999 Overhaul GIST code (Oleg)
1000 Fix CLUSTER to preserve constraints and column default (Tom)
1001 Improved deadlock detection handling (Tom)
1002 Allow multiple SERIAL columns in a table (Tom)
1003 Prevent occasional index corruption (Vadim)
1007 Add OUTER JOINs (Tom)
1008 Function manager overhaul (Tom)
1009 Allow ALTER TABLE RENAME on indexes (Tom)
1010 Improve CLUSTER (Tom)
1011 Improve ps status display for more platforms (Peter E, Marc)
1012 Improve CREATE FUNCTION failure message (Ross)
1013 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
1015 Grand Unified Configuration scheme/GUC. Many options can now be set in
1016 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
1017 Improved handling of file descriptor cache (Tom)
1018 New warning code about auto-created table alias entries (Bruce)
1019 Overhaul initdb process (Tom, Peter E)
1020 Overhaul of inherited tables; inherited tables now accessed by default;
1021 new ONLY keyword prevents it (Chris Bitmead, Tom)
1022 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
1024 Allow renaming of temp tables (Tom)
1025 Overhaul memory manager contexts (Tom)
1026 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
1027 Overhaul pg_dump (Philip Warner)
1028 Allow pg_hba.conf secondary password file to specify only username (Peter E)
1029 Allow TEMPORARY or TEMP keyword when creating temporary tables (Bruce)
1030 New memory leak checker (Karel)
1031 New SET SESSION CHARACTERISTICS (Thomas)
1032 Allow nested block comments (Thomas)
1033 Add WITHOUT TIME ZONE type qualifier (Thomas)
1034 New ALTER TABLE ADD CONSTRAINT (Stephan)
1035 Use NUMERIC accumulators for INTEGER aggregates (Tom)
1036 Overhaul aggregate code (Tom)
1037 New VARIANCE and STDDEV() aggregates
1038 Improve dependency ordering of pg_dump (Philip)
1039 New pg_restore command (Philip)
1040 New pg_dump tar output option (Philip)
1041 New pg_dump of large objects (Philip)
1042 New ESCAPE option to LIKE (Thomas)
1043 New case-insensitive LIKE - ILIKE (Thomas)
1044 Allow functional indexes to use binary-compatible type (Tom)
1045 Allow SQL functions to be used in more contexts (Tom)
1046 New pg_config utility (Peter E)
1047 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
1049 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
1050 New quote_identifiers() and quote_literal() functions (Jan)
1051 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
1052 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
1053 Update PyGreSQL to version 3.1 (D'Arcy)
1054 Store tables as files named by OID (Vadim)
1055 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
1056 Require DROP VIEW to remove views, no DROP TABLE (Mark)
1057 Allow DROP VIEW view1, view2 (Mark)
1058 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
1059 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
1060 New /contrib/pgcrypto hashing functions (Marko Kreen)
1061 New pg_dumpall --globals-only option (Peter E)
1062 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
1063 New AT TIME ZONE syntax (Thomas)
1064 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
1065 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
1066 Allow socket path name to be specified in hostname by using leading slash
1067 (David J. MacKenzie)
1068 Allow CREATE DATABASE to specify template database (Tom)
1069 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
1070 New /contrib/rserv replication toolkit (Vadim)
1071 New file format for COPY BINARY (Tom)
1072 New /contrib/oid2name to map numeric files to table names (B Palmer)
1073 New "idle in transaction" ps status message (Marc)
1074 Update to pgaccess 0.98.7 (Constantin Teodorescu)
1075 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
1076 Add rudimentary dependency checking to pg_dump (Philip)
1080 Fix INET/CIDR type ordering and add new functions (Tom)
1081 Make OID behave as an unsigned type (Tom)
1082 Allow BIGINT as synonym for INT8 (Peter E)
1083 New int2 and int8 comparison operators (Tom)
1084 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
1085 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
1086 New GIST seg/cube examples (Gene Selkov)
1087 Improved round(numeric) handling (Tom)
1088 Fix CIDR output formatting (Tom)
1089 New CIDR abbrev() function (Tom)
1093 Write-Ahead Log (WAL) to provide crash recovery with less performance
1095 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
1096 Reduced file seeks (Denis Perchine)
1097 Improve BTREE code for duplicate keys (Tom)
1098 Store all large objects in a single table (Denis Perchine, Tom)
1099 Improve memory allocation performance (Karel, Tom)
1103 New function manager call conventions (Tom)
1104 SGI portability fixes (David Kaelbling)
1105 New configure --enable-syslog option (Peter E)
1106 New BSDI README (Bruce)
1107 configure script moved to top level, not /src (Peter E)
1108 Makefile/configuration/compilation overhaul (Peter E)
1109 New configure --with-python option (Peter E)
1110 Solaris cleanups (Peter E)
1111 Overhaul /contrib Makefiles (Karel)
1112 New OpenSSL configuration option (Magnus, Peter E)
1114 QNX fixes (Maurizio)
1115 New heap_open(), heap_openr() API (Tom)
1116 Remove colon and semi-colon operators (Thomas)
1117 New pg_class.relkind value for views (Mark Hollomon)
1118 Rename ichar() to chr() (Karel)
1119 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
1120 Fixes for NT/Cygwin (Pete Forman)
1121 AIX port fixes (Andreas)
1122 New BeOS port (David Reid, Cyril Velter)
1123 Add proofreader's changes to docs (Addison-Wesley, Bruce)
1124 New Alpha spinlock code (Adriaan Joubert, Compaq)
1125 UnixWare port overhaul (Peter E)
1126 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
1127 New FreeBSD Alpha port (Alfred)
1128 Overhaul shared memory segments (Tom)
1129 Add IBM S/390 support (Neale Ferguson)
1130 Moved macmanuf to /contrib (Larry Rosenman)
1131 Syslog improvements (Larry Rosenman)
1132 New template0 database that contains no user additions (Tom)
1133 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
1134 Allow NetBSD's libedit instead of readline (Peter)
1135 Improved assembly language source code format (Bruce)
1136 New contrib/pg_logger
1137 New --template option to createdb
1138 New contrib/pg_control utility (Oliver)
1139 New FreeBSD tools ipc_check, start-scripts/freebsd
1146 <sect1 id="release-7-0-3">
1147 <title>Release 7.0.3</title>
1150 <title>Release date</title>
1151 <simpara>2000-11-11</simpara>
1155 This has a variety of fixes from 7.0.2.
1160 <title>Migration to version 7.0.3</title>
1163 A dump/restore is <emphasis>not</emphasis> required for those running
1169 <title>Changes</title>
1174 Large object fix (Tom)
1175 Fix lean in COPY WITH OIDS leak (Tom)
1176 Fix backwards-index-scan (Tom)
1177 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
1178 Add --enable-syslog to configure (Marc)
1179 Fix abort transaction at backend exit in rare cases (Tom)
1180 Fix for psql \l+ when multibyte enabled (Tatsuo)
1181 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
1182 Make vacuum always flush buffers (Tom)
1183 Fix to allow cancel while waiting for a lock (Hiroshi)
1184 Fix for memory aloocation problem in user authentication code (Tom)
1185 Remove bogus use of int4out() (Tom)
1186 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
1187 Fix for failure of triggers on heap open in certain cases (Jeroen van
1189 Fix for erroneous selectivity of not-equals (Tom)
1190 Fix for erroneous use of strcmp() (Tom)
1191 Fix for bug where storage manager accesses items beyond end of file
1193 Fix to include kernel errno message in all smgr elog messages (Tom)
1194 Fix for '.' not in PATH at build time (SL Baur)
1195 Fix for out-of-file-descriptors error (Tom)
1196 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
1197 Fix for subselect in targetlist of Append node (Tom)
1198 Fix for mergejoin plans (Tom)
1199 Fix TRUNCATE failure on relations with indexes (Tom)
1200 Avoid database-wide restart on write error (Hiroshi)
1201 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
1202 Fix VACUUM problem with moving chain of update tuples when source and
1203 destination of a tuple lie on the same page (Tom)
1204 Fix user.c CommandCounterIncrement (Tom)
1205 Fix for AM/PM boundary problem in to_char() (Karel Zak)
1206 Fix TIME aggregate handling (Tom)
1207 Fix to_char() to avoid coredump on NULL input (Tom)
1209 Fix for inserting/copying longer multibyte strings into char() data
1211 Fix for crash of backend, on abort (Tom)
1218 <sect1 id="release-7-0-2">
1219 <title>Release 7.0.2</title>
1222 <title>Release date</title>
1223 <simpara>2000-06-05</simpara>
1227 This is a repackaging of 7.0.1 with added documentation.
1232 <title>Migration to version 7.0.2</title>
1235 A dump/restore is <emphasis>not</emphasis> required for those running
1241 <title>Changes</title>
1245 Added documentation to tarball.
1252 <sect1 id="release-7-0-1">
1253 <title>Release 7.0.1</title>
1256 <title>Release date</title>
1257 <simpara>2000-06-01</simpara>
1261 This is a cleanup release for 7.0.
1265 <title>Migration to version 7.0.1</title>
1268 A dump/restore is <emphasis>not</emphasis> required for those running
1274 <title>Changes</title>
1278 Fix many CLUSTER failures (Tom)
1279 Allow ALTER TABLE RENAME works on indexes (Tom)
1280 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
1281 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
1282 Fix the off by one errors in ResultSet from 6.5.3, and more.
1283 jdbc ResultSet fixes (Joseph Shraibman)
1284 optimizer tunings (Tom)
1285 Fix create user for pgaccess
1286 Fix for UNLISTEN failure
1287 IRIX fixes (David Kaelbling)
1288 QNX fixes (Andreas Kardos)
1289 Reduce COPY IN lock level (Tom)
1290 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
1291 Fix pg_dump to handle OID indexes (Tom)
1292 Fix small memory leak (Tom)
1293 Solaris fix for createdb/dropdb (Tatsuo)
1294 Fix for non-blocking connections (Alfred Perlstein)
1295 Fix improper recovery after RENAME TABLE failures (Tom)
1296 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
1297 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
1298 Fix too long syslog message (Tatsuo)
1299 Fix problem with quoted indexes that are too long (Tom)
1300 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
1301 ecpg changes (Michael)
1307 <sect1 id="release-7-0">
1308 <title>Release 7.0</title>
1311 <title>Release date</title>
1312 <simpara>2000-05-08</simpara>
1316 This release contains improvements in many areas, demonstrating
1317 the continued growth of <productname>PostgreSQL</productname>.
1318 There are more improvements and fixes in 7.0 than in any previous
1319 release. The developers have confidence that this is the best
1320 release yet; we do our best to put out only solid releases, and
1321 this one is no exception.
1325 Major changes in this release:
1334 Foreign keys are now implemented, with the exception of PARTIAL MATCH
1335 foreign keys. Many users have been asking for this feature, and we are
1336 pleased to offer it.
1347 Continuing on work started a year ago, the optimizer has been
1348 improved, allowing better query plan selection and faster performance
1349 with less memory usage.
1356 Updated <application>psql</application>
1360 <application>psql</application>, our interactive terminal monitor, has been
1361 updated with a variety of new features. See the <application>psql</application> manual page for details.
1372 SQL92 join syntax is now supported, though only as INNER JOINs
1373 for this release. JOIN, NATURAL JOIN, JOIN/USING, JOIN/ON are
1374 available, as are column correlation names.
1386 In 7.1 or 7.2, we plan to have outer joins, storage for very long rows, and a
1387 write-ahead logging system.
1396 <title>Migration to version 7.0</title>
1399 A dump/restore using <application>pg_dump</application>
1400 is required for those wishing to migrate data from any
1401 previous release of <productname>PostgreSQL</productname>.
1402 For those upgrading from 6.5.*, you may instead use
1403 <application>pg_upgrade</application> to upgrade to this
1404 release; however, a full dump/reload installation is always the
1405 most robust method for upgrades.
1409 Interface and compatibility issues to consider for the new
1415 The date/time types <type>datetime</type> and
1416 <type>timespan</type> have been superseded by the
1417 SQL92-defined types <type>timestamp</type> and
1418 <type>interval</type>. Although there has been some effort to
1419 ease the transition by allowing
1420 <productname>PostgreSQL</productname> to recognize
1421 the deprecated type names and translate them to the new type
1422 names, this mechanism may not be completely transparent to
1423 your existing application.
1429 The optimizer has been substantially improved in the area of
1430 query cost estimation. In some cases, this will result in
1431 decreased query times as the optimizer makes a better choice
1432 for the preferred plan. However, in a small number of cases,
1433 usually involving pathological distributions of data, your
1434 query times may go up. If you are dealing with large amounts
1435 of data, you may want to check your queries to verify
1442 The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
1443 interfaces have been upgraded and extended.
1449 The string function <function>CHAR_LENGTH</function> is now a
1450 native function. Previous versions translated this into a call
1451 to <function>LENGTH</function>, which could result in
1452 ambiguity with other types implementing
1453 <function>LENGTH</function> such as the geometric types.
1462 <title>Changes</title>
1468 Prevent function calls exceeding maximum number of arguments (Tom)
1469 Improve CASE construct (Tom)
1470 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
1471 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
1472 Fix GROUP BY scan bug (Tom)
1473 Improvements in SQL grammar processing (Tom)
1474 Fix for views involved in INSERT ... SELECT ... (Tom)
1475 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
1476 Fix for subselects in INSERT ... SELECT (Tom)
1477 Prevent INSERT ... SELECT ... ORDER BY (Tom)
1478 Fixes for relations greater than 2GB, including vacuum
1479 Improve propagating system table changes to other backends (Tom)
1480 Improve propagating user table changes to other backends (Tom)
1481 Fix handling of temp tables in complex situations (Bruce, Tom)
1482 Allow table locking at table open, improving concurrent reliability (Tom)
1483 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
1484 Prevent DROP DATABASE while others accessing
1485 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
1486 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
1487 Fix pg_upgrade so it works for MVCC (Tom)
1488 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
1489 Fix for "f1 datetime DEFAULT 'now'" (Tom)
1490 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
1491 Allow comment-only lines, and ;;; lines too. (Tom)
1492 Improve recovery after failed disk writes, disk full (Hiroshi)
1493 Fix cases where table is mentioned in FROM but not joined (Tom)
1494 Allow HAVING clause without aggregate functions (Tom)
1495 Fix for "--" comment and no trailing newline, as seen in perl interface
1496 Improve pg_dump failure error reports (Bruce)
1497 Allow sorts and hashes to exceed 2GB file sizes (Tom)
1498 Fix for pg_dump dumping of inherited rules (Tom)
1499 Fix for NULL handling comparisons (Tom)
1500 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
1501 Fix for dbname with dash
1502 Prevent DROP INDEX from interfering with other backends (Tom)
1503 Fix file descriptor leak in verify_password()
1504 Fix for "Unable to identify an operator =$" problem
1505 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
1506 Fix for recursive exit call (Massimo)
1507 Fix for extra-long timezones (Jeroen van Vianen)
1508 Make pg_dump preserve primary key information (Peter E)
1509 Prevent databases with single quotes (Peter E)
1510 Prevent DROP DATABASE inside transaction (Peter E)
1511 ecpg memory leak fixes (Stephen Birch)
1512 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
1513 Y2K timestamp fix (Massimo)
1514 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
1515 Fix for views with tables/columns containing spaces (Tom)
1516 Prevent permissions on indexes (Peter E)
1517 Fix for spinlock stuck problem when error is generated (Hiroshi)
1518 Fix ipcclean on Linux
1519 Fix handling of NULL constraint conditions (Tom)
1520 Fix memory leak in odbc driver (Nick Gorham)
1521 Fix for permission check on UNION tables (Tom)
1522 Fix to allow SELECT 'a' LIKE 'a' (Tom)
1523 Fix for SELECT 1 + NULL (Tom)
1525 Fix log() on numeric type (Tom)
1526 Deprecate ':' and ';' operators
1527 Allow vacuum of temporary tables
1528 Disallow inherited columns with the same name as new columns
1529 Recover or force failure when disk space is exhausted (Hiroshi)
1530 Fix INSERT INTO ... SELECT with AS columns matching result columns
1531 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns (Tom)
1532 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
1533 Fix UNION with LIMIT
1534 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
1535 Fix CREATE TABLE test(col char(2) DEFAULT user)
1536 Fix mismatched types in CREATE TABLE ... DEFAULT
1537 Fix SELECT * FROM pg_class where oid in (0,-1)
1538 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
1539 Prevent user who can create databases can modifying pg_database table (Peter E)
1540 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
1541 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
1545 New CLI interface include file sqlcli.h, based on SQL3/SQL98
1546 Remove all limits on query length, row length limit still exists (Tom)
1547 Update jdbc protocol to 2.0 (Jens Glaser <email>jens@jens.de</email>)
1548 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
1549 Fix to give super user and createdb user proper update catalog rights (Peter E)
1550 Allow ecpg bool variables to have NULL values (Christof)
1551 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
1552 Allow ^C to cancel COPY command (Massimo)
1553 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
1554 Function name overloading for dynamically-loaded C functions (Frankpitt)
1555 Add CmdTuples() to libpq++(Vince)
1556 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
1557 Allow CREATE FUNCTION/WITH clause to be used for all language types
1558 configure --enable-debug adds -g (Peter E)
1559 configure --disable-debug removes -g (Peter E)
1560 Allow more complex default expressions (Tom)
1561 First real FOREIGN KEY constraint trigger functionality (Jan)
1562 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
1563 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
1564 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
1565 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
1566 Change pgeasy connectdb() parameter ordering (Bruce)
1567 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
1568 Add Oracle's COMMENT ON command (Mike Mascari <email>mascarim@yahoo.com</email>)
1569 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
1570 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
1571 Make USING in COPY optional (Bruce)
1572 Allow subselects in the target list (Tom)
1573 Allow subselects on the left side of comparison operators (Tom)
1574 New parallel regression test (Jan)
1575 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
1576 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
1577 Added psql LASTOID variable to return last inserted oid (Peter E)
1578 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
1579 Add permissions check for vacuum (Peter E)
1580 New libpq functions to allow asynchronous connections: PQconnectStart(),
1581 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
1582 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
1583 New libpq PQsetenv() function (Ewan Mellor)
1584 create/alter user extension (Peter E)
1585 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
1586 New scripts for create/drop user/db (Peter E)
1587 Major psql overhaul (Peter E)
1588 Add const to libpq interface (Peter E)
1589 New libpq function PQoidValue (Peter E)
1590 Show specific non-aggregate causing problem with GROUP BY (Tom)
1591 Make changes to pg_shadow recreate pg_pwd file (Peter E)
1592 Add aggregate(DISTINCT ...) (Tom)
1593 Allow flag to control COPY input/output of NULLs (Peter E)
1594 Make postgres user have a password by default (Peter E)
1595 Add CREATE/ALTER/DROP GROUP (Peter E)
1596 All administration scripts now support --long options (Peter E, Karel)
1597 Vacuumdb script now supports --all option (Peter E)
1598 ecpg new portable FETCH syntax
1599 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
1600 and EXEC SQL ENDIF directives
1601 Add pg_ctl script to control backend start-up (Tatsuo)
1602 Add postmaster.opts.default file to store start-up flags (Tatsuo)
1603 Allow --with-mb=SQL_ASCII
1604 Increase maximum number of index keys to 16 (Bruce)
1605 Increase maximum number of function arguments to 16 (Bruce)
1606 Allow configuration of maximum number of index keys and arguments (Bruce)
1607 Allow unprivileged users to change their passwords (Peter E)
1608 Password authentication enabled; required for new users (Peter E)
1609 Disallow dropping a user who owns a database (Peter E)
1610 Change initdb option --with-mb to --enable-multibyte
1611 Add option for initdb to prompts for superuser password (Peter E)
1612 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
1613 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
1614 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
1615 Libpq non-blocking mode (Alfred Perlstein)
1616 Improve conversion of types in casts that don't specify a length
1617 New plperl internal programming language (Mark Hollomon)
1618 Allow COPY IN to read file that do not end with a newline (Tom)
1619 Indicate when long identifiers are truncated (Tom)
1620 Allow aggregates to use type equivalency (Peter E)
1621 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
1622 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
1623 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
1624 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
1625 Add NUMERIC and int8 types to ODBC
1626 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
1627 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
1628 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
1629 Enable backward sequential scan even after reaching EOF (Hiroshi)
1630 Add btree indexing of boolean values, >= and <= (Don Baccus)
1631 Print current line number when COPY FROM fails (Massimo)
1632 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
1633 Add DEC as synonym for DECIMAL (Thomas)
1634 Add SESSION_USER as SQL92 keyword, same as CURRENT_USER (Thomas)
1635 Implement SQL92 column aliases (aka correlation names) (Thomas)
1636 Implement SQL92 join syntax (Thomas)
1637 Make INTERVAL reserved word allowed as a column identifier (Thomas)
1638 Implement REINDEX command (Hiroshi)
1639 Accept ALL in aggregate function SUM(ALL col) (Tom)
1640 Prevent GROUP BY from using column aliases (Tom)
1641 New psql \encoding option (Tatsuo)
1642 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
1643 Allow negation of a negative number in all cases
1644 Add ecpg descriptors (Christof, Michael)
1645 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
1646 Allow casts with length, like foo::char(8)
1647 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
1648 Add support for SJIS user defined characters (Tatsuo)
1649 Larger views/rules supported
1650 Make libpq's PQconndefaults() thread-safe (Tom)
1651 Disable // as comment to be ANSI conforming, should use -- (Tom)
1652 Allow column aliases on views CREATE VIEW name (collist)
1653 Fixes for views with subqueries (Tom)
1654 Allow UPDATE table SET fld = (SELECT ...) (Tom)
1655 SET command options no longer require quotes
1656 Update pgaccess to 0.98.6
1657 New SET SEED command
1658 New pg_options.sample file
1659 New SET FSYNC command (Massimo)
1660 Allow pg_descriptions when creating tables
1661 Allow pg_descriptions when creating types, columns, and functions
1662 Allow psql \copy to allow delimiters (Peter E)
1663 Allow psql to print nulls as distinct from "" [null] (Peter E)
1667 Many array fixes (Tom)
1668 Allow bare column names to be subscripted as arrays (Tom)
1669 Improve type casting of int and float constants (Tom)
1670 Cleanups for int8 inputs, range checking, and type conversion (Tom)
1671 Fix for SELECT timespan('21:11:26'::time) (Tom)
1672 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
1673 Add btree index on NUMERIC (Jan)
1674 Perl fix for large objects containing NUL characters (Douglas Thomson)
1675 ODBC fix for for large objects (free)
1676 Fix indexing of cidr data type
1677 Fix for Ethernet MAC addresses (macaddr type) comparisons
1678 Fix for date/time types when overflows happened in computations (Tom)
1679 Allow array on int8 (Peter E)
1680 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
1681 Allow NUMERIC arrays
1682 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
1683 Make char_length()/octet_length including trailing blanks (Tom)
1684 Made abstime/reltime use int4 instead of time_t (Peter E)
1685 New lztext data type for compressed text fields
1686 Revise code to handle coercion of int and float constants (Tom)
1687 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
1688 NUMERIC now accepts scientific notation (Tom)
1689 NUMERIC to int4 rounds (Tom)
1690 Convert float4/8 to NUMERIC properly (Tom)
1691 Allow type conversion with NUMERIC (Thomas)
1692 Make ISO date style (2000-02-16 09:33) the default (Thomas)
1693 Add NATIONAL CHAR [ VARYING ] (Thomas)
1694 Allow NUMERIC round and trunc to accept negative scales (Tom)
1695 New TIME WITH TIME ZONE type (Thomas)
1696 Add MAX()/MIN() on time type (Thomas)
1697 Add abs(), mod(), fac() for int8 (Thomas)
1698 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
1699 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
1700 Add exp() and ln() for NUMERIC type
1701 Rename NUMERIC power() to pow() (Thomas)
1702 Improved TRANSLATE() function (Edwin Ramirez, Tom)
1703 Allow X=-Y operators (Tom)
1704 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom)
1705 Allow LOCALE to use indexes in regular expression searches (Tom)
1706 Allow creation of functional indexes to use default types
1710 Prevent exponential space consumption with many AND's and OR's (Tom)
1711 Collect attribute selectivity values for system columns (Tom)
1712 Reduce memory usage of aggregates (Tom)
1713 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
1714 Fix r-tree index optimizer selectivity (Thomas)
1715 Improve optimizer selectivity computations and functions (Tom)
1716 Optimize btree searching for cases where many equal keys exist (Tom)
1717 Enable fast LIKE index processing only if index present (Tom)
1718 Re-use free space on index pages with duplicates (Tom)
1719 Improve hash join processing (Tom)
1720 Prevent descending sort if result is already sorted(Hiroshi)
1721 Allow commuting of index scan query qualifications (Tom)
1722 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
1723 Allocate large memory requests in fix-sized chunks for performance (Tom)
1724 Fix vacuum's performance by reducing memory allocation requests (Tom)
1725 Implement constant-expression simplification (Bernard Frankpitt, Tom)
1726 Use secondary columns to be used to determine start of index scan (Hiroshi)
1727 Prevent quadruple use of disk space when doing internal sorting (Tom)
1728 Faster sorting by calling fewer functions (Tom)
1729 Create system indexes to match all system caches (Bruce, Hiroshi)
1730 Make system caches use system indexes (Bruce)
1731 Make all system indexes unique (Bruce)
1732 Improve pg_statistics management for VACUUM speed improvement (Tom)
1733 Flush backend cache less frequently (Tom, Hiroshi)
1734 COPY now reuses previous memory allocation, improving performance (Tom)
1735 Improve optimization cost estimation (Tom)
1736 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom)
1737 Use DNF instead of CNF where appropriate (Tom, Taral)
1738 Further cleanup for OR-of-AND WHERE-clauses (Tom)
1739 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
1740 Smarter optimizer computations for random index page access (Tom)
1741 New SET variable to control optimizer costs (Tom)
1742 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
1743 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
1744 Major subquery speedup (Tom)
1745 Fewer fsync writes when fsync is not disabled (Tom)
1746 Improved LIKE optimizer estimates (Tom)
1747 Prevent fsync in SELECT-only queries (Vadim)
1748 Make index creation use psort code, because it is now faster (Tom)
1749 Allow creation of sort temp tables > 1 Gig
1753 Fix for linux PPC compile
1754 New generic expression-tree-walker subroutine (Tom)
1755 Change form() to varargform() to prevent portability problems
1756 Improved range checking for large integers on Alphas
1757 Clean up #include in /include directory (Bruce)
1758 Add scripts for checking includes (Bruce)
1759 Remove un-needed #include's from *.c files (Bruce)
1760 Change #include's to use <> and "" as appropriate (Bruce)
1761 Enable WIN32 compilation of libpq
1762 Alpha spinlock fix from Uncle George <email>gatgul@voicenet.com</email>
1763 Overhaul of optimizer data structures (Tom)
1764 Fix to cygipc library (Yutaka Tanida)
1765 Allow pgsql to work on newer Cygwin snapshots (Dan)
1766 New catalog version number (Tom)
1768 Rename heap_replace to heap_update
1769 Update for QNX (Dr. Andreas Kardos)
1770 New platform-specific regression handling (Tom)
1771 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
1772 Included all yacc and lex files into the distribution (Peter E.)
1773 Remove lextest, no longer needed (Peter E)
1774 Fix for libpq and psql on Win32 (Magnus)
1775 Internally change datetime and timespan into timestamp and interval (Thomas)
1776 Fix for plpgsql on BSD/OS
1777 Add SQL_ASCII test case to the regression test (Tatsuo)
1778 configure --with-mb now deprecated (Tatsuo)
1780 NetBSD fixes (Johnny C. Lam <email>lamj@stat.cmu.edu</email>)
1781 Fixes for Alpha compiles
1782 New multibyte encodings
1788 <sect1 id="release-6-5-3">
1789 <title>Release 6.5.3</title>
1792 <title>Release date</title>
1793 <simpara>1999-10-13</simpara>
1797 This is basically a cleanup release for 6.5.2. We have added a new
1798 <application>PgAccess</> that was missing in 6.5.2, and installed an NT-specific fix.
1803 <title>Migration to version 6.5.3</title>
1806 A dump/restore is <emphasis>not</emphasis> required for those running
1811 <title>Changes</title>
1815 Updated version of pgaccess 0.98
1817 Fix dumping rules on inherited tables
1824 <sect1 id="release-6-5-2">
1825 <title>Release 6.5.2</title>
1828 <title>Release date</title>
1829 <simpara>1999-09-15</simpara>
1833 This is basically a cleanup release for 6.5.1. We have fixed a variety of
1834 problems reported by 6.5.1 users.
1839 <title>Migration to version 6.5.2</title>
1842 A dump/restore is <emphasis>not</emphasis> required for those running
1848 <title>Changes</title>
1852 subselect+CASE fixes(Tom)
1853 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
1854 Fixes for CASE in WHERE join clauses(Tom)
1855 Fix BTScan abort(Tom)
1856 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
1857 Improve it so that it checks for multicolumn constraints(Thomas)
1858 Fix for Win32 making problem with MB enabled(Hiroki Kataoka)
1859 Allow BSD yacc and bison to compile pl code(Bruce)
1860 Fix SET NAMES working
1862 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
1863 Reduce the total memory consumption of vacuum(Tom)
1864 Fix for timestamp(datetime)
1865 Rule deparsing bugfixes(Tom)
1866 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(Tom)
1867 This is to re-use space on index pages freed by vacuum(Vadim)
1868 document -x for pg_dump(Bruce)
1869 Fix for unary operators in rule deparser(Tom)
1870 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
1871 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
1872 Repair logic error in LIKE: should not return LIKE_ABORT
1873 when reach end of pattern before end of text(Tom)
1874 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom)
1875 Updated version of pgaccess 0.98
1881 <sect1 id="release-6-5-1">
1882 <title>Release 6.5.1</title>
1885 <title>Release date</title>
1886 <simpara>1999-07-15</simpara>
1890 This is basically a cleanup release for 6.5. We have fixed a variety of
1891 problems reported by 6.5 users.
1895 <title>Migration to version 6.5.1</title>
1898 A dump/restore is <emphasis>not</emphasis> required for those running
1904 <title>Changes</title>
1909 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
1910 Remove QUERY_LIMIT, use SELECT...LIMIT
1911 Fix for EXPLAIN on inheritance(Tom)
1912 Patch to allow vacuum on multisegment tables(Hiroshi)
1913 R-Tree optimizer selectivity fix(Tom)
1914 ACL file descriptor leak fix(Atsushi Ogawa)
1915 New expresssion subtree code(Tom)
1916 Avoid disk writes for read-only transactions(Vadim)
1917 Fix for removal of temp tables if last transaction was aborted(Bruce)
1918 Fix to prevent too large tuple from being created(Bruce)
1920 Allow port numbers 32k - 64k(Bruce)
1921 Add ^ precidence(Bruce)
1922 Rename sort files called pg_temp to pg_sorttemp(Bruce)
1923 Fix for microseconds in time values(Tom)
1924 Tutorial source cleanup
1926 Fix for sorting of NULL's in some cases(Tom)
1927 Shared library dependencies fixed (Tom)
1928 Fixed glitches affecting GROUP BY in subselects(Tom)
1929 Fix some compiler warnings (Tomoaki Nishiyama)
1930 Add Win1250 (Czech) support (Pavel Behal)
1936 <sect1 id="release-6-5">
1937 <title>Release 6.5</title>
1940 <title>Release date</title>
1941 <simpara>1999-06-09</simpara>
1945 This release marks a major step in the development team's mastery of the source
1946 code we inherited from Berkeley. You will see we are now easily adding
1947 major features, thanks to the increasing size and experience of our
1948 world-wide development team.
1952 Here is a brief summary of the more notable changes:
1957 Multiversion concurrency control(MVCC)
1961 This removes our old
1962 table-level locking, and replaces it with a locking system that is
1963 superior to most commercial database systems. In a traditional system,
1964 each row that is modified is locked until committed, preventing reads by
1965 other users. MVCC uses the natural multiversion nature of PostgreSQL
1966 to allow readers to continue reading consistent data during writer
1967 activity. Writers continue to use the compact pg_log transaction
1968 system. This is all performed without having to allocate a lock for
1969 every row like traditional database systems. So, basically, we no
1970 longer are restricted by simple table-level locking;
1971 we have something better than row-level locking.
1978 Hot backups from <application>pg_dump</application>
1982 <application>pg_dump</application> takes advantage of the new
1983 MVCC features to give a consistent database dump/backup while
1984 the database stays online and available for queries.
1995 We now have a true numeric data type, with
1996 user-specified precision.
2007 Temporary tables are guaranteed to have unique names
2008 within a database session, and are destroyed on session exit.
2019 We now have CASE, INTERSECT, and EXCEPT statement
2020 support. We have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL,
2021 SELECT ... FOR UPDATE, and an improved LOCK TABLE command.
2032 We continue to speed up PostgreSQL, thanks to the variety of
2033 talents within our team. We have sped up memory allocation,
2034 optimization, table joins, and row transfer routines.
2045 We continue to expand our port list, this time including
2046 <systemitem class="osname">Windows NT</>/<systemitem>ix86</> and <systemitem class="osname">NetBSD</>/<systemitem>arm32</>.
2057 Most interfaces have new versions, and existing functionality
2069 New and updated material is present throughout the
2070 documentation. New <acronym>FAQ</acronym>s have been
2071 contributed for <systemitem class="osname">SGI</> and <systemitem class="osname">AIX</> platforms.
2072 The <citetitle>Tutorial</citetitle> has introductory information
2073 on <acronym>SQL</acronym> from Stefan Simkovics.
2074 For the <citetitle>User's Guide</citetitle>, there are
2075 reference pages covering the postmaster and more utility
2076 programs, and a new appendix
2077 contains details on date/time behavior.
2078 The <citetitle>Administrator's Guide</citetitle> has a new
2079 chapter on troubleshooting from Tom Lane.
2080 And the <citetitle>Programmer's Guide</citetitle> has a
2081 description of query processing, also from Stefan, and details
2082 on obtaining the <productname>PostgreSQL</productname> source
2083 tree via anonymous <productname>CVS</productname> and
2084 <productname>CVSup</productname>.
2092 <title>Migration to version 6.5</title>
2095 A dump/restore using <application>pg_dump</application>
2096 is required for those wishing to migrate data from any
2097 previous release of <productname>PostgreSQL</productname>.
2098 <application>pg_upgrade</application> can <emphasis>not</emphasis>
2099 be used to upgrade to this release because the on-disk structure
2100 of the tables has changed compared to previous releases.
2104 The new Multiversion Concurrency Control (MVCC) features can
2105 give somewhat different behaviors in multiuser
2106 environments. <emphasis>Read and understand the following section
2107 to ensure that your existing applications will give you the
2108 behavior you need.</emphasis>
2112 <title>Multiversion Concurrency Control</title>
2115 Because readers in 6.5 don't lock data, regardless of transaction
2116 isolation level, data read by one transaction can be overwritten by
2117 another. In other words, if a row is returned by
2118 <command>SELECT</command> it doesn't mean that this row really exists
2119 at the time it is returned (i.e. sometime after the statement or
2120 transaction began) nor that the row is protected from being deleted or
2121 updated by concurrent transactions before the current transaction does
2122 a commit or rollback.
2126 To ensure the actual existence of a row and protect it against
2127 concurrent updates one must use <command>SELECT FOR UPDATE</command> or
2128 an appropriate <command>LOCK TABLE</command> statement. This should be
2129 taken into account when porting applications from previous releases of
2130 <productname>PostgreSQL</productname> and other environments.
2134 Keep the above in mind if you are using
2135 <filename>contrib/refint.*</filename> triggers for
2136 referential integrity. Additional techniques are required now. One way is
2137 to use <command>LOCK parent_table IN SHARE ROW EXCLUSIVE MODE</command>
2138 command if a transaction is going to update/delete a primary key and
2139 use <command>LOCK parent_table IN SHARE MODE</command> command if a
2140 transaction is going to update/insert a foreign key.
2144 Note that if you run a transaction in SERIALIZABLE mode then you must
2145 execute the <command>LOCK</command> commands above before execution of any
2146 <acronym>DML</acronym> statement
2147 (<command>SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO</command>) in the
2154 These inconveniences will disappear in the future
2155 when the ability to read dirty
2156 (uncommitted) data (regardless of isolation level) and true referential
2157 integrity will be implemented.
2163 <title>Changes</title>
2169 Fix text<->float8 and text<->float4 conversion functions(Thomas)
2170 Fix for creating tables with mixed-case constraints(Billy)
2171 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
2172 Fix bug in pg_dump -z
2173 Memory overrun cleanups(Tatsuo)
2174 Fix for lo_import crash(Tatsuo)
2175 Adjust handling of data type names to suppress double quotes(Thomas)
2176 Use type coercion for matching columns and DEFAULT(Thomas)
2177 Fix deadlock so it only checks once after one second of sleep(Bruce)
2178 Fixes for aggregates and PL/pgsql(Hiroshi)
2179 Fix for subquery crash(Vadim)
2180 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
2181 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo)
2182 Fix for pg_dump -d or -D and quote special characters in INSERT
2183 Repair serious problems with dynahash(Tom)
2184 Fix INET/CIDR portability problems
2185 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
2186 Fix executor so mergejoin of different column types works(Tom)
2187 Fix for Alpha OR selectivity bug
2188 Fix OR index selectivity problem(Bruce)
2189 Fix so \d shows proper length for char()/varchar()(Ryan)
2190 Fix tutorial code(Clark)
2191 Improve destroyuser checking(Oliver)
2192 Fix for Kerberos(Rodney McDuff)
2193 Fix for dropping database while dirty buffers(Bruce)
2194 Fix so sequence nextval() can be case-sensitive(Bruce)
2196 Drop buffers before destroying database files(Bruce)
2197 Fix case where executor evaluates functions twice(Tatsuo)
2198 Allow sequence nextval actions to be case-sensitive(Bruce)
2199 Fix optimizer indexing not working for negative numbers(Bruce)
2200 Fix for memory leak in executor with fjIsNull
2201 Fix for aggregate memory leaks(Erik Riedel)
2202 Allow username containing a dash GRANT permissions
2203 Cleanup of NULL in inet types
2204 Clean up system table bugs(Tom)
2205 Fix problems of PAGER and \? command(Masaaki Sakaida)
2206 Reduce default multisegment file size limit to 1GB(Peter)
2207 Fix for dumping of CREATE OPERATOR(Tom)
2208 Fix for backward scanning of cursors(Hiroshi Inoue)
2209 Fix for COPY FROM STDIN when using \i(Tom)
2210 Fix for subselect is compared inside an expression(Jan)
2211 Fix handling of error reporting while returning rows(Tom)
2212 Fix problems with reference to array types(Tom,Jan)
2213 Prevent UPDATE SET oid(Jan)
2214 Fix pg_dump so -t option can handle case-sensitive tablenames
2215 Fixes for GROUP BY in special cases(Tom, Jan)
2216 Fix for memory leak in failed queries(Tom)
2217 DEFAULT now supports mixed-case identifiers(Tom)
2218 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
2219 Disable use of pg_dump with both -o and -d options(Bruce)
2220 Allow pg_dump to properly dump GROUP permissions(Bruce)
2221 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
2222 Fix for computations in views(Jan)
2223 Fix for aggregates on array indexes(Tom)
2224 Fix for DEFAULT handles single quotes in value requiring too many quotes
2225 Fix security problem with non-super users importing/exporting large objects(Tom)
2226 Rollback of transaction that creates table cleaned up properly(Tom)
2227 Fix to allow long table and column names to generate proper serial names(Tom)
2231 Add "vacuumdb" utility
2232 Speed up libpq by allocating memory better(Tom)
2233 EXPLAIN all indexes used(Tom)
2234 Implement CASE, COALESCE, NULLIF expression(Thomas)
2235 New pg_dump table output format(Constantin)
2236 Add string min()/max() functions(Thomas)
2237 Extend new type coercion techniques to aggregates(Thomas)
2238 New moddatetime contrib(Terry)
2239 Update to pgaccess 0.96(Constantin)
2240 Add routines for single-byte "char" type(Thomas)
2241 Improved substr() function(Thomas)
2242 Improved multibyte handling(Tatsuo)
2243 Multiversion concurrency control/MVCC(Vadim)
2244 New Serialized mode(Vadim)
2245 Fix for tables over 2gigs(Peter)
2246 New SET TRANSACTION ISOLATION LEVEL(Vadim)
2247 New LOCK TABLE IN ... MODE(Vadim)
2248 Update ODBC driver(Byron)
2249 New NUMERIC data type(Jan)
2250 New SELECT FOR UPDATE(Vadim)
2251 Handle "NaN" and "Infinity" for input values(Jan)
2252 Improved date/year handling(Thomas)
2253 Improved handling of backend connections(Magnus)
2254 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
2255 New TCL_ARRAYS option(Massimo)
2256 New INTERSECT and EXCEPT(Stefan)
2257 New pg_index.indisprimary for primary key tracking(D'Arcy)
2258 New pg_dump option to allow dropping of tables before creation(Brook)
2259 Speedup of row output routines(Tom)
2260 New READ COMMITTED isolation level(Vadim)
2261 New TEMP tables/indexes(Bruce)
2262 Prevent sorting if result is already sorted(Jan)
2263 New memory allocation optimization(Jan)
2264 Allow psql to do \p\g(Bruce)
2265 Allow multiple rule actions(Jan)
2266 Added LIMIT/OFFSET functionality(Jan)
2267 Improve optimizer when joining a large number of tables(Bruce)
2268 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
2269 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
2270 Improved int8 support(Ryan Bradetich, Thomas, Tom)
2271 New routines to convert between int8 and text/varchar types(Thomas)
2272 New bushy plans, where meta-tables are joined(Bruce)
2273 Enable right-hand queries by default(Bruce)
2274 Allow reliable maximum number of backends to be set at configure time
2275 (--with-maxbackends and postmaster switch (-N backends))(Tom)
2276 GEQO default now 10 tables because of optimizer speedups(Tom)
2277 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
2278 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
2279 Allow psql \d on a view show query(Ryan)
2280 Speedup for LIKE(Bruce)
2281 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
2282 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
2283 Make % operator have precedence like /(Bruce)
2284 Add new postgres -O option to allow system table structure changes(Bruce)
2285 Update contrib/pginterface/findoidjoins script(Tom)
2286 Major speedup in vacuum of deleted rows with indexes(Vadim)
2287 Allow non-SQL functions to run different versions based on arguments(Tom)
2288 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida)
2289 Add version number in start-up banners for psql(Masaaki Sakaida)
2290 New contrib/vacuumlo removes large objects not referenced(Peter)
2291 New initialization for table sizes so non-vacuumed tables perform better(Tom)
2292 Improve error messages when a connection is rejected(Tom)
2293 Support for arrays of char() and varchar() fields(Massimo)
2294 Overhaul of hash code to increase reliability and performance(Tom)
2295 Update to PyGreSQL 2.4(D'Arcy)
2296 Changed debug options so -d4 and -d5 produce different node displays(Jan)
2297 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
2298 Better optimization statistics for system table access(Tom)
2299 Better handling of non-default block sizes(Massimo)
2300 Improve GEQO optimizer memory consumption(Tom)
2301 UNION now suppports ORDER BY of columns not in target list(Jan)
2302 Major libpq++ improvements(Vince Vielhaber)
2303 pg_dump now uses -z(ACL's) as default(Bruce)
2304 backend cache, memory speedups(Tom)
2305 have pg_dump do everything in one snapshot transaction(Vadim)
2306 fix for large object memory leakage, fix for pg_dumping(Tom)
2307 INET type now respects netmask for comparisons
2308 Make VACUUM ANALYZE only use a readlock(Vadim)
2309 Allow VIEWs on UNIONS(Jan)
2310 pg_dump now can generate consistent snapshots on active databases(Vadim)
2314 Improve port matching(Tom)
2315 Portability fixes for SunOS
2316 Add NT/Win32 backend port and enable dynamic loading(Magnus and Daniel Horak)
2317 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
2318 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
2319 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
2320 Port to NetBSD/macppc(Toshimi Aoki)
2321 Fix for tcl/tk configuration(Vince)
2322 Removed CURRENT keyword for rule queries(Jan)
2323 NT dynamic loading now works(Daniel Horak)
2324 Add ARM32 support(Andrew McMurry)
2325 Better support for HP-UX 11 and UnixWare
2326 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
2327 New install commands for plpgsql(Jan)
2334 <sect1 id="release-6-4-2">
2335 <title>Release 6.4.2</title>
2338 <title>Release date</title>
2339 <simpara>1998-12-20</simpara>
2343 The 6.4.1 release was improperly packaged. This also has one additional
2349 <title>Migration to version 6.4.2</title>
2352 A dump/restore is <emphasis>not</emphasis> required for those running
2357 <title>Changes</title>
2361 Fix for datetime constant problem on some platforms(Thomas)
2369 <sect1 id="release-6-4-1">
2370 <title>Release 6.4.1</title>
2373 <title>Release date</title>
2374 <simpara>1998-12-18</simpara>
2378 This is basically a cleanup release for 6.4. We have fixed a variety of
2379 problems reported by 6.4 users.
2384 <title>Migration to version 6.4.1</title>
2387 A dump/restore is <emphasis>not</emphasis> required for those running
2392 <title>Changes</title>
2396 Add pg_dump -N flag to force double quotes around identifiers. This is
2398 Fix for NOT in where clause causing crash(Bruce)
2399 EXPLAIN VERBOSE coredump fix(Vadim)
2400 Fix shared-library problems on Linux
2401 Fix test for table existance to allow mixed-case and whitespace in
2402 the table name(Thomas)
2403 Fix a couple of pg_dump bugs
2404 Configure matches template/.similar entries better(Tom)
2405 Change builtin function names from SPI_* to spi_*
2406 OR WHERE clause fix(Vadim)
2407 Fixes for mixed-case table names(Billy)
2408 contrib/linux/postgres.init.csh/sh fix(Thomas)
2409 libpq memory overrun fix
2411 Change exp() behavior to generate error on underflow(Thomas)
2412 pg_dump fixes for memory leak, inheritance constraints, layout change
2413 update pgaccess to 0.93
2414 Fix prototype for 64-bit platforms
2415 Multibyte fixes(Tatsuo)
2417 Fix memory overruns(Tatsuo)
2418 Fix for lo_import() crash(Bruce)
2419 Better search for install program(Tom)
2422 Use implicit type coercion for matching DEFAULT values(Thomas)
2423 Add routines to help with single-byte (internal) character type(Thomas)
2424 Compilation of libpq for Win32 fixes(Magnus)
2425 Upgrade to PyGreSQL 2.2(D'Arcy)
2433 <sect1 id="release-6-4">
2434 <title>Release 6.4</title>
2437 <title>Release date</title>
2438 <simpara>1998-10-30</simpara>
2442 There are <emphasis>many</emphasis> new features and improvements in this release.
2443 Thanks to our developers and maintainers, nearly every aspect of the system
2444 has received some attention since the previous release.
2445 Here is a brief, incomplete summary:
2450 Views and rules are now functional thanks to extensive new code in the
2451 rewrite rules system from Jan Wieck. He also wrote a chapter on it
2452 for the <citetitle>Programmer's Guide</citetitle>.
2457 Jan also contributed a second procedural language, <application>PL/pgSQL</application>, to go with the
2458 original <application>PL/pgTCL</application> procedural language he contributed last release.
2464 We have optional multiple-byte character set support from Tatsuo Iishi
2465 to complement our existing locale support.
2471 Client/server communications has been cleaned up, with better support for
2472 asynchronous messages and interrupts thanks to Tom Lane.
2478 The parser will now perform automatic type coercion to match arguments
2479 to available operators and functions, and to match columns and expressions
2480 with target columns. This uses a generic mechanism which supports
2481 the type extensibility features of <productname>PostgreSQL</productname>.
2482 There is a new chapter in the <citetitle>User's Guide</citetitle>
2483 which covers this topic.
2489 Three new data types have been added.
2490 Two types, <type>inet</type> and <type>cidr</type>, support various forms
2491 of IP network, subnet, and machine addressing. There is now an 8-byte integer
2492 type available on some platforms. See the chapter on data types
2493 in the <citetitle>User's Guide</citetitle> for details.
2494 A fourth type, <type>serial</type>, is now supported by the parser as an
2495 amalgam of the <type>int4</type> type, a sequence, and a unique index.
2501 Several more <acronym>SQL92</acronym>-compatible syntax features have been
2502 added, including <command>INSERT DEFAULT VALUES</command>
2508 The automatic configuration and installation system has received some
2509 attention, and should be more robust for more platforms than it has ever
2518 <title>Migration to version 6.4</title>
2521 A dump/restore using <application>pg_dump</application>
2522 or <application>pg_dumpall</application>
2523 is required for those wishing to migrate data from any
2524 previous release of <productname>PostgreSQL</productname>.
2529 <title>Changes</title>
2535 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
2536 Remove char2-16 data types, use char/varchar(Darren)
2537 Pqfn not handles a NOTICE message(Anders)
2538 Reduced busywaiting overhead for spinlocks with many backends (dg)
2539 Stuck spinlock detection (dg)
2540 Fix up "ISO-style" timespan decoding and encoding(Thomas)
2541 Fix problem with table drop after rollback of transaction(Vadim)
2542 Change error message and remove non-functional update message(Vadim)
2543 Fix for COPY array checking
2544 Fix for SELECT 1 UNION SELECT NULL
2545 Fix for buffer leaks in large object calls(Pascal)
2546 Change owner from oid to int4 type(Bruce)
2547 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
2548 Fix for shared invalidation cache overflow(Massimo)
2549 Prevent file descriptor leaks in failed COPY's(Bruce)
2550 Fix memory leak in libpgtcl's pg_select(Constantin)
2551 Fix problems with username/passwords over 8 characters(Tom)
2552 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
2553 Fix of many bad system table entries(Tom)
2557 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
2558 Show the index used in an EXPLAIN(Zeugswetter)
2559 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
2560 Multibyte awareness of many data types and functions, via configure(Tatsuo)
2561 New configure --with-mb option(Tatsuo)
2562 New initdb --pgencoding option(Tatsuo)
2563 New createdb -E multibyte option(Tatsuo)
2564 Select version(); now returns PostgreSQL version(Jeroen)
2565 Libpq now allows asynchronous clients(Tom)
2566 Allow cancel from client of backend query(Tom)
2567 Psql now cancels query with Control-C(Tom)
2568 Libpq users need not issue dummy queries to get NOTIFY messages(Tom)
2569 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
2570 PGresult struct now includes associated error message, if any(Tom)
2571 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
2572 Add routines to convert between varchar and bpchar(Thomas)
2573 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
2574 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
2575 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas)
2576 Fixes for unary minus parsing with leading spaces(Thomas)
2577 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
2578 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
2579 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
2580 Enable HAVING clause but no fixes elsewhere yet.
2581 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thomas)
2582 Save string type if specified for DEFAULT clause handling(Thomas)
2583 Coerce operations involving different data types(Thomas)
2584 Allow some index use for columns of different types(Thomas)
2585 Add capabilities for automatic type conversion(Thomas)
2586 Cleanups for large objects, so file is truncated on open(Peter)
2587 Readline cleanups(Tom)
2588 Allow psql \f \ to make spaces as delimiter(Bruce)
2589 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
2590 Msql compatibility library in /contrib(Aldrin)
2591 Remove the requirement that ORDER/GROUP BY clause identifiers be
2592 included in the target list(David)
2593 Convert columns to match columns in UNION clauses(Thomas)
2594 Remove fork()/exec() and only do fork()(Bruce)
2595 Jdbc cleanups(Peter)
2596 Show backend status on ps command line(only works on some platforms)(Bruce)
2597 Pg_hba.conf now has a sameuser option in the database field
2598 Make lo_unlink take oid param, not int4
2599 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
2600 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
2601 libpgtcl cleanups(Tom)
2602 Add -error option to libpgtcl's pg_result command(Tom)
2603 New locale patch, see docs/README/locale(Oleg)
2604 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
2605 New contrib/lo code for large object orphan removal(Peter)
2606 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
2607 feature, see /doc/README.mb(Tatsuo)
2608 /contrib/noupdate code to revoke update permission on a column
2609 Libpq can now be compiled on win32(Magnus)
2610 Add PQsetdbLogin() in libpq
2611 New 8-byte integer type, checked by configure for OS support(Thomas)
2612 Better support for quoted table/column names(Thomas)
2613 Surround table and column names with double-quotes in pg_dump(Thomas)
2614 PQreset() now works with passwords(Tom)
2615 Handle case of GROUP BY target list column number out of range(David)
2616 Allow UNION in subselects
2617 Add auto-size to screen to \d? commands(Bruce)
2618 Use UNION to show all \d? results in one query(Bruce)
2619 Add \d? field search feature(Bruce)
2620 Pg_dump issues fewer \connect requests(Tom)
2621 Make pg_dump -z flag work better, document it in manual page(Tom)
2622 Add HAVING clause with full support for subselects and unions(Stephan)
2623 Full text indexing routines in contrib/fulltextindex(Maarten)
2624 Transaction ids now stored in shared memory(Vadim)
2625 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
2626 Support for SQL92 syntax "SET NAMES"(Tatsuo)
2627 Support for LATIN2-5(Tatsuo)
2628 Add UNICODE regression test case(Tatsuo)
2629 Lock manager cleanup, new locking modes for LLL(Vadim)
2630 Allow index use with OR clauses(Bruce)
2631 Allows "SELECT NULL ORDER BY 1;"
2632 Explain VERBOSE prints the plan, and now pretty-prints the plan to
2633 the postmaster log file(Bruce)
2634 Add indexes display to \d command(Bruce)
2635 Allow GROUP BY on functions(David)
2636 New pg_class.relkind for large objects(Bruce)
2637 New way to send libpq NOTICE messages to a different location(Tom)
2638 New \w write command to psql(Bruce)
2639 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
2640 Allow binary-compatible indexes to be considered when checking for valid
2641 Indexes for restriction clauses containing a constant(Thomas)
2642 New ISBN/ISSN code in /contrib/isbn_issn
2643 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
2644 New rewrite system fixes many problems with rules and views(Jan)
2645 * Rules on relations work
2646 * Event qualifications on insert/update/delete work
2647 * New OLD variable to reference CURRENT, CURRENT will be remove in future
2648 * Update rules can reference NEW and OLD in rule qualifications/actions
2649 * Insert/update/delete rules on views work
2650 * Multiple rule actions are now supported, surrounded by parentheses
2651 * Regular users can create views/rules on tables they have RULE permits
2652 * Rules and views inherit the permissions on the creator
2653 * No rules at the column level
2654 * No UPDATE NEW/OLD rules
2655 * New pg_tables, pg_indexes, pg_rules and pg_views system views
2656 * Only a single action on SELECT rules
2657 * Total rewrite overhaul, perhaps for 6.5
2659 * handle aggregates on views
2660 * handle insert into select from view works
2661 System indexes are now multikey(Bruce)
2662 Oidint2, oidint4, and oidname types are removed(Bruce)
2663 Use system cache for more system table lookups(Bruce)
2664 New backend programming language PL/pgSQL in backend/pl(Jan)
2665 New SERIAL data type, auto-creates sequence/index(Thomas)
2666 Enable assert checking without a recompile(Massimo)
2667 User lock enhancements(Massimo)
2668 New setval() command to set sequence value(Massimo)
2669 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
2670 Conditional trace package(Massimo)
2671 New UNLISTEN command(Massimo)
2672 Psql and libpq now compile under win32 using win32.mak(Magnus)
2673 Lo_read no longer stores trailing NULL(Bruce)
2674 Identifiers are now truncated to 31 characters internally(Bruce)
2675 Createuser options now availble on the command line
2676 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
2677 Prevent file descriptor leaf from failed COPY(Bruce)
2678 New pg_upgrade command(Bruce)
2679 Updated /contrib directories(Massimo)
2680 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
2681 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
2682 New DECLARE and FETCH feature(Thomas)
2683 libpq's internal structures now not exported(Tom)
2684 Allow up to 8 key indexes(Bruce)
2685 Remove ARCHIVE keyword, that is no longer used(Thomas)
2686 pg_dump -n flag to supress quotes around indentifiers
2687 disable system columns for views(Jan)
2688 new INET and CIDR types for network addresses(TomH, Paul)
2689 no more double quotes in psql output
2690 pg_dump now dumps views(Terry)
2691 new SET QUERY_LIMIT(Tatsuo,Jan)
2695 /contrib cleanup(Jun)
2696 Inline some small functions called for every row(Bruce)
2699 Multibyte regression tests(Soonmyung.)
2700 Remove --disabled options from configure
2701 Define PGDOC to use POSTGRESDIR by default
2702 Make regression optional
2703 Remove extra braces code to pgindent(Bruce)
2704 Add bsdi shared library support(Bruce)
2705 New --without-CXX support configure option(Brook)
2707 Update backend flowchart in tools/backend(Bruce)
2708 Change atttypmod from int16 to int32(Bruce, Tom)
2709 Getrusage() fix for platforms that do not have it(Tom)
2710 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
2711 NS32K platform fixes(Phil Nelson, John Buller)
2712 SCO 7/UnixWare 2.x fixes(Billy,others)
2713 Sparc/Solaris 2.5 fixes(Ryan)
2714 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
2715 Even more documention(Thomas)
2716 Nextstep support(Jacek)
2718 pginterface manual page(Bruce)
2719 shared libraries all have version numbers
2720 merged all OS-specific shared library defines into one file
2721 smarter TCL/TK configuration checking(Billy)
2722 smarter perl configuration(Brook)
2723 configure uses supplied install-sh if no install script found(Tom)
2724 new Makefile.shlib for shared library configuration(Tom)
2730 <sect1 id="release-6-3-2">
2731 <title>Release 6.3.2</title>
2734 <title>Release date</title>
2735 <simpara>1998-04-07</simpara>
2739 This is a bug-fix release for 6.3.x.
2740 Refer to the release notes for version 6.3 for a more complete summary of new features.
2748 Repairs automatic configuration support for some platforms, including Linux,
2749 from breakage inadvertently introduced in version 6.3.1.
2755 Correctly handles function calls on the left side of BETWEEN and LIKE clauses.
2762 A dump/restore is NOT required for those running 6.3 or 6.3.1. A
2763 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
2764 This last step should be performed while the postmaster is not running.
2765 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
2768 For upgrades from pre-6.3 installations,
2769 refer to the installation and migration instructions for version 6.3.
2773 <title>Changes</title>
2777 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
2778 Manual page improvements(Bruce)
2779 BETWEEN and LIKE fix(Thomas)
2780 fix for psql \connect used by pg_dump(Oliver Elphick)
2782 pgaccess, version 0.86
2783 qsort removed, now uses libc version, cleanups(Jeroen)
2784 fix for buffer over-runs detected(Maurice Gittens)
2785 fix for buffer overrun in libpgtcl(Randy Kunkee)
2786 fix for UNION with DISTINCT or ORDER BY(Bruce)
2787 gettimeofday configure check(Doug Winterburn)
2788 Fix "indexes not used" bug(Vadim)
2789 docs additions(Thomas)
2790 Fix for backend memory leak(Bruce)
2791 libreadline cleanup(Erwan MAS)
2792 Remove DISTDIR(Bruce)
2793 Makefile dependency cleanup(Jeroen van Vianen)
2800 <sect1 id="release-6-3-1">
2801 <title>Release 6.3.1</title>
2804 <title>Release date</title>
2805 <simpara>1998-03-23</simpara>
2814 Additional support for multibyte character sets.
2820 Repair byte ordering for mixed-endian clients and servers.
2826 Minor updates to allowed SQL syntax.
2832 Improvements to the configuration autodetection for installation.
2839 A dump/restore is NOT required for those running 6.3. A
2840 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
2841 This last step should be performed while the postmaster is not running.
2842 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
2845 For upgrades from pre-6.3 installations,
2846 refer to the installation and migration instructions for version 6.3.
2850 <title>Changes</title>
2854 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
2855 pg_user cleanup(Bruce)
2856 large object fix for pg_dump and tclsh (alvin)
2857 LIKE fix for multiple adjacent underscores
2858 fix for redefining builtin functions(Thomas)
2860 upgrade to pg_access 0.83
2861 updated CLUSTER manual page
2862 multibyte character set support, see doc/README.mb(Tatsuo)
2863 configure --with-pgport fix
2865 big-endian fix for backend communications(Kataoka)
2866 SUBSTR() and substring() fix(Jan)
2867 several jdbc fixes(Peter)
2868 libpgtcl improvements, see libptcl/README(Randy Kunkee)
2869 Fix for "Datasize = 0" error(Vadim)
2870 Prevent \do from wrapping(Bruce)
2871 Remove duplicate Russian character set entries
2873 Allow optional TABLE keyword in LOCK and SELECT INTO(Thomas)
2874 CREATE SEQUENCE options to allow a negative integer(Thomas)
2875 Add "PASSWORD" as an allowed column identifier(Thomas)
2876 Add checks for UNION target fields(Bruce)
2877 Fix Alpha port(Dwayne Bailey)
2878 Fix for text arrays containing quotes(Doug Gibson)
2879 Solaris compile fix(Albert Chin-A-Young)
2880 Better identify tcl and tk libs and includes(Bruce)
2886 <sect1 id="release-6-3">
2887 <title>Release 6.3</title>
2890 <title>Release date</title>
2891 <simpara>1998-03-01</simpara>
2895 There are <emphasis>many</emphasis> new features and improvements in this release.
2896 Here is a brief, incomplete summary:
2901 Many new SQL features, including
2902 full <acronym>SQL92</acronym> subselect capability
2903 (everything is here but target-list subselects).
2909 Support for client-side environment variables to specify time zone and date style.
2915 Socket interface for client/server connection. This is the default now
2916 so you may need to start <application>postmaster</application> with the
2917 <option>-i</option> flag.
2923 Better password authorization mechanisms. Default table permissions have changed.
2929 Old-style <firstterm>time travel</firstterm>
2930 has been removed. Performance has been improved.
2938 Bruce Momjian wrote the following notes to introduce the new release.
2943 There are some general 6.3 issues that I want to mention. These are
2944 only the big items that can not be described in one sentence. A review
2945 of the detailed changes list is still needed.
2948 First, we now have subselects. Now that we have them, I would like to
2949 mention that without subselects, SQL is a very limited language.
2950 Subselects are a major feature, and you should review your code for
2951 places where subselects provide a better solution for your queries. I
2952 think you will find that there are more uses for subselects than you may
2953 think. Vadim has put us on the big SQL map with subselects, and fully
2954 functional ones too. The only thing you can't do with subselects is to
2955 use them in the target list.
2958 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
2959 enable connections from other machines, you have to use the new
2960 postmaster -i option, and of course edit <filename>pg_hba.conf</filename>. Also, for this
2961 reason, the format of <filename>pg_hba.conf</filename> has changed.
2964 Third, <type>char()</type> fields will now allow faster access than <type>varchar()</type> or
2965 <type>text</type>. Specifically, the <type>text</> and <type>varchar()</type> have a penalty for access to
2966 any columns after the first column of this type. <type>char()</type> used to also
2967 have this access penalty, but it no longer does. This may suggest that
2968 you redesign some of your tables, especially if you have short character
2969 columns that you have defined as <type>varchar()</type> or <type>text</type>. This and other
2970 changes make 6.3 even faster than earlier releases.
2973 We now have passwords definable independent of any Unix file. There are
2974 new SQL USER commands.
2975 See the <citetitle>Administrator's Guide</citetitle> for more
2976 information. There is a new table, pg_shadow, which is used to store
2977 user information and user passwords, and it by default only SELECT-able
2978 by the <systemitem>postgres</systemitem> super-user. pg_user is now a view of pg_shadow, and is
2979 SELECT-able by PUBLIC. You should keep using pg_user in your
2980 application without changes.
2983 User-created tables now no longer have SELECT permission to PUBLIC by
2984 default. This was done because the ANSI standard requires it. You can
2985 of course GRANT any permissions you want after the table is created.
2986 System tables continue to be SELECT-able by PUBLIC.
2989 We also have real deadlock detection code. No more sixty-second
2990 timeouts. And the new locking code implements a <acronym>FIFO</acronym> better, so there
2991 should be less resource starvation during heavy use.
2994 Many complaints have been made about inadequate documentation in previous
2995 releases. Thomas has put much effort into many new manuals for this
2996 release. Check out the doc/ directory.
2999 For performance reasons, time travel is gone, but can be implemented
3000 using triggers (see <filename>pgsql/contrib/spi/README</filename>). Please check out the new
3001 \d command for types, operators, etc. Also, views have their own
3002 permissions now, not based on the underlying tables, so permissions on
3003 them have to be set separately. Check <filename>/pgsql/interfaces</filename> for some new
3004 ways to talk to <productname>PostgreSQL</productname>.
3007 This is the first release that really required an explanation for
3008 existing users. In many ways, this was necessary because the new
3009 release removes many limitations, and the work-arounds people were using
3010 are no longer needed.
3014 <title>Migration to version 6.3</title>
3017 A dump/restore using <application>pg_dump</application>
3018 or <application>pg_dumpall</application>
3019 is required for those wishing to migrate data from any
3020 previous release of <productname>PostgreSQL</productname>.
3025 <title>Changes</title>
3031 Fix binary cursors broken by MOVE implementation(Vadim)
3032 Fix for tcl library crash(Jan)
3033 Fix for array handling, from Gerhard Hintermayer
3034 Fix acl error, and remove duplicate pqtrace(Bruce)
3035 Fix psql \e for empty file(Bruce)
3036 Fix for textcat on varchar() fields(Bruce)
3037 Fix for DBT Sendproc (Zeugswetter Andres)
3038 Fix vacuum analyze syntax problem(Bruce)
3039 Fix for international identifiers(Tatsuo)
3040 Fix aggregates on inherited tables(Bruce)
3041 Fix substr() for out-of-bounds data
3042 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
3043 Fix notty output to show status result. -q option still turns it off(Bruce)
3044 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
3046 Fix for PQtrace start/stop several times(Bruce)
3047 Fix a variety of locking problems like newer lock waiters getting
3048 lock before older waiters, and having readlock people not share
3049 locks if a writer is waiting for a lock, and waiting writers not
3050 getting priority over waiting readers(Bruce)
3051 Fix crashes in psql when executing queries from external files(James)
3052 Fix problem with multiple order by columns, with the first one having
3054 Use correct hash table support functions for float8 and int4(Thomas)
3055 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
3056 Change precedence for boolean operators to match expected behavior(Thomas)
3057 Generate elog(ERROR) on over-large integer(Bruce)
3058 Allow multiple-argument functions in constraint clauses(Thomas)
3059 Check boolean input literals for 'true','false','yes','no','1','0'
3060 and throw elog(ERROR) if unrecognized(Thomas)
3061 Major large objects fix
3062 Fix for GROUP BY showing duplicates(Vadim)
3063 Fix for index scans in MergeJion(Vadim)
3067 Subselects with EXISTS, IN, ALL, ANY keywords (Vadim, Bruce, Thomas)
3068 New User Manual(Thomas, others)
3069 Speedup by inlining some frequently-called functions
3070 Real deadlock detection, no more timeouts(Bruce)
3071 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
3072 CURRENT_USER(Thomas)
3073 Modify constraint syntax to be SQL92-compliant(Thomas)
3074 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
3075 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
3076 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas)
3077 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
3078 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
3079 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
3080 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
3081 Allow SQL92 delimited identifiers(Thomas)
3082 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas)
3083 Support SQL92 syntax for type coercion of literal strings
3084 (e.g. "DATETIME 'now'")(Thomas)
3085 Add conversions for int2, int4, and OID types to and from text(Thomas)
3086 Use shared lock when building indexes(Vadim)
3087 Free memory allocated for an user query inside transaction block after
3088 this query is done, was turned off in <= 6.2.1(Vadim)
3089 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
3090 New <productname>PostgreSQL</productname> Procedural Language (PL) backend interface(Jan)
3091 Rename pg_dump -H option to -h(Bruce)
3092 Add Java support for passwords, European dates(Peter)
3093 Use indexes for LIKE and ~, !~ operations(Bruce)
3094 Add hash functions for datetime and timespan(Thomas)
3095 Time Travel removed(Vadim, Bruce)
3096 Add paging for \d and \z, and fix \i(Bruce)
3097 Add Unix domain socket support to backend and to frontend library(Goran)
3098 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
3099 Allow more SQL92 and/or <productname>PostgreSQL</productname> reserved words as column identifiers(Thomas)
3100 Augment support for SQL92 SET TIME ZONE...(Thomas)
3101 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
3102 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
3103 Enable SET TIME ZONE using TZ environment variable(Thomas)
3104 Add PGDATESTYLE environment variable to frontend and backend initialization(Thomas)
3105 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
3106 frontend library initialization environment variables(Thomas)
3107 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
3108 Add pg_description table for info on tables, columns, operators, types, and
3110 Increase 16 char limit on system table/index names to 32 characters(Bruce)
3111 Rename system indexes(Bruce)
3112 Add 'GERMAN' option to SET DATESTYLE(Thomas)
3113 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
3114 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
3115 Validate numeric input more carefully for delta times(Thomas)
3116 Implement day of year as possible input to date_part()(Thomas)
3117 Define timespan_finite() and text_timespan() functions(Thomas)
3118 Remove archive stuff(Bruce)
3119 Allow for a pg_password authentication database that is separate from
3120 the system password file(Todd)
3121 Dump ACLs, GRANT, REVOKE permissions(Matt)
3122 Define text, varchar, and bpchar string length functions(Thomas)
3123 Fix Query handling for inheritance, and cost computations(Bruce)
3124 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
3125 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
3126 Implement UNIONs for SELECT(Bruce)
3127 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
3128 varchar() stores only necessary bytes on disk(Bruce)
3129 Fix for BLOBs(Peter)
3130 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
3131 Remove unused "option" from PQconnectdb()
3132 New LOCK command and lock manual page describing deadlocks(Bruce)
3133 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
3134 Enhance psql \z to show sequences(Bruce)
3135 Show NOT NULL and DEFAULT in psql \d table(Bruce)
3136 New psql .psqlrc file start-up(Andrew)
3137 Modify sample start-up script in contrib/linux to show syslog(Thomas)
3138 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
3139 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
3140 Update of contrib stuff(Massimo)
3141 Add Unix socket support to DBD::Pg(Goran)
3142 New python interface (PyGreSQL 2.0)(D'Arcy)
3143 New frontend/backend protocol has a version number, network byte order(Phil)
3144 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
3145 CHAR() now faster access than VARCHAR() or TEXT
3146 ecpg embedded SQL preprocessor
3147 Reduce system column overhead(Vadmin)
3148 Remove pg_time table(Vadim)
3149 Add pg_type attribute to identify types that need length (bpchar, varchar)
3150 Add report of offending line when COPY command fails
3151 Allow VIEW permissions to be set separately from the underlying tables.
3152 For security, use GRANT/REVOKE on views as appropriate(Jan)
3153 Tables now have no default GRANT SELECT TO PUBLIC. You must
3154 explicitly grant such permissions.
3155 Clean up tutorial examples(Darren)
3159 Add new html development tools, and flow chart in /tools/backend
3160 Fix for SCO compiles
3161 Stratus computer port Robert Gillies
3162 Added support for shlib for BSD44_derived & i386_solaris
3163 Make configure more automated(Brook)
3164 Add script to check regression test results
3165 Break parser functions into smaller files, group together(Bruce)
3166 Rename heap_create to heap_create_and_catalog, rename heap_creatr
3167 to heap_create()(Bruce)
3168 Sparc/Linux patch for locking(TomS)
3169 Remove PORTNAME and reorganize port-specific stuff(Marc)
3170 Add optimizer README file(Bruce)
3171 Remove some recursion in optimizer and clean up some code there(Bruce)
3172 Fix for NetBSD locking(Henry)
3173 Fix for libptcl make(Tatsuo)
3175 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
3176 function calls to istrue() or isfalse() to allow optimization(Thomas)
3177 Various fixes NetBSD/Sparc related(TomH)
3178 Alpha linux locking(Travis,Ryan)
3179 Change elog(WARN) to elog(ERROR)(Bruce)
3180 FAQ for FreeBSD(Marc)
3181 Bring in the PostODBC source tree as part of our standard distribution(Marc)
3182 A minor patch for HP/UX 10 vs 9(Stan)
3183 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
3184 UnixWare patches(Billy)
3185 New i386 'lock' for spin lock asm(Billy)
3186 Support for multiplexed backends is removed
3187 Start an OpenBSD port
3190 Add string functions to regression suite(Thomas)
3191 Expand a few function names formerly truncated to 16 characters(Thomas)
3192 Remove un-needed malloc() calls and replace with palloc()(Bruce)
3198 <sect1 id="release-6-2-1">
3199 <title>Release 6.2.1</title>
3202 <title>Release date</title>
3203 <simpara>1997-10-17</simpara>
3207 6.2.1 is a bug-fix and usability release on 6.2.
3215 Allow strings to span lines, per <acronym>SQL92</acronym>.
3221 Include example trigger function for inserting user names on table updates.
3228 This is a minor bug-fix release on 6.2.
3229 For upgrades from pre-6.2 systems, a full dump/reload is required.
3230 Refer to the 6.2 release notes for instructions.
3234 <title>Migration from version 6.2 to version 6.2.1</title>
3237 This is a minor bug-fix release. A dump/reload is not required from version 6.2,
3238 but is required from any release prior to 6.2.
3241 In upgrading from version 6.2, if you choose to dump/reload you will find that
3242 avg(money) is now calculated correctly. All other bug fixes take effect
3243 upon updating the executables.
3246 Another way to avoid dump/reload is to use the following SQL command
3247 from <command>psql</command> to update the existing system table:
3250 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
3251 where aggname = 'avg' and aggbasetype = 790;
3255 This will need to be done to every existing database, including template1.
3260 <title>Changes</title>
3264 Allow TIME and TYPE column names(Thomas)
3265 Allow larger range of true/false as boolean values(Thomas)
3266 Support output of "now" and "current"(Thomas)
3267 Handle DEFAULT with INSERT of NULL properly(Vadim)
3268 Fix for relation reference counts problem in buffer manager(Vadim)
3269 Allow strings to span lines, like ANSI(Thomas)
3270 Fix for backward cursor with ORDER BY(Vadim)
3271 Fix avg(cash) computation(Thomas)
3272 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
3273 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
3274 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
3280 <sect1 id="release-6-2">
3281 <title>Release 6.2</title>
3284 <title>Release date</title>
3285 <simpara>1997-10-02</simpara>
3289 A dump/restore is required for those wishing to migrate data from
3290 previous releases of <productname>PostgreSQL</productname>.
3294 <title>Migration from version 6.1 to version 6.2</title>
3297 This migration requires a complete dump of the 6.1 database and a
3298 restore of the database in 6.2.
3301 Note that the <command>pg_dump</command> and <command>pg_dumpall</command> utility from 6.2 should be used
3302 to dump the 6.1 database.
3307 <title>Migration from version 1.<replaceable>x</> to version 6.2</title>
3310 Those migrating from earlier 1.* releases should first upgrade to 1.09
3311 because the COPY output format was improved from the 1.02 release.
3316 <title>Changes</title>
3322 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
3323 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
3324 from Solaris(Diab Jerius)
3325 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
3326 Check for geometric intersections at endpoints to avoid rounding ugliness(Thomas)
3327 Catch non-functional delete attempts(Vadim)
3328 Change time function names to be more consistent(Michael Reifenberg)
3329 Check for zero divides(Michael Reifenberg)
3330 Fix very old bug which made tuples changed/inserted by a commnd
3331 visible to the command itself (so we had multiple update of
3332 updated tuples, etc)(Vadim)
3333 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
3334 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
3335 Remove un-needed signal stuff from contrib/pginterface
3336 Fix OR (where x != 1 or x isnull didn't return tuples with x NULL) (Vadim)
3337 Fix time_cmp function (Vadim)
3338 Fix handling of functions with non-attribute first argument in
3339 WHERE clauses (Vadim)
3340 Fix GROUP BY when order of entries is different from order
3341 in target list (Vadim)
3342 Fix pg_dump for aggregates without sfunc1 (Vadim)
3346 Default genetic optimizer GEQO parameter is now 8(Bruce)
3347 Allow use parameters in target list having aggregates in functions(Vadim)
3348 Added JDBC driver as an interface(Adrian & Peter)
3350 Return number of tuples inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
3351 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
3352 SPI (Server Programming Interface) allows execution of queries inside
3354 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
3355 Include reserved words for string handling, outer joins, and unions(Thomas)
3356 Implement extended comments ("/* ... */") using exclusive states(Thomas)
3357 Add "//" single-line comments(Bruce)
3358 Remove some restrictions on characters in operator names(Thomas)
3359 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
3360 Add text concatenation operator and function (SQL92)(Thomas)
3361 Support WITH TIME ZONE syntax (SQL92)(Thomas)
3362 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
3363 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
3364 and CHARACTER VARYING (SQL92)(Thomas)
3365 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
3366 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
3367 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
3368 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
3369 Add more reserved words, mostly for SQL92 compliance(Thomas)
3370 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
3371 Add center() routines for lseg, path, polygon(Thomas)
3372 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
3373 Check explicitly for points and polygons contained within polygons
3374 using an axis-crossing algorithm(Thomas)
3375 Add routine to convert circle-box(Thomas)
3376 Merge conflicting operators for different geometric data types(Thomas)
3377 Replace distance operator "<===>" with "<->"(Thomas)
3378 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Thomas)
3379 Add routines for text trimming on both ends, substring, and string position(Thomas)
3380 Added conversion routines circle(box) and poly(circle)(Thomas)
3381 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
3382 Allow functions and operators on internally-identical types to succeed(Bruce)
3383 Speed up backend start-up after profiling analysis(Bruce)
3384 Inline frequently called functions for performance(Bruce)
3385 Reduce open() calls(Bruce)
3386 psql: Add PAGER for \h and \?,\C fix
3387 Fix for psql pager when no tty(Bruce)
3388 New entab utility(Bruce)
3389 General trigger functions for referential integrity (Vadim)
3390 General trigger functions for time travel (Vadim)
3391 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
3392 MOVE implementation (Vadim)
3396 HP-UX 10 patches (Vladimir Turin)
3397 Added SCO support, (Daniel Harris)
3398 MkLinux patches (Tatsuo Ishii)
3399 Change geometric box terminology from "length" to "width"(Thomas)
3400 Deprecate temporary unstored slope fields in geometric code(Thomas)
3401 Remove restart instructions from INSTALL(Bruce)
3402 Look in /usr/ucb first for install(Bruce)
3403 Fix c++ copy example code(Thomas)
3404 Add -o to psql manual page(Bruce)
3405 Prevent relname unallocated string length from being copied into database(Bruce)
3406 Cleanup for NAMEDATALEN use(Bruce)
3407 Fix pg_proc names over 15 chars in output(Bruce)
3408 Add strNcpy() function(Bruce)
3409 remove some (void) casts that are unnecessary(Bruce)
3410 new interfaces directory(Marc)
3411 Replace fopen() calls with calls to fd.c functions(Bruce)
3412 Make functions static where possible(Bruce)
3413 enclose unused functions in #ifdef NOT_USED(Bruce)
3414 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
3415 Changes for Digital Unix
3416 Portability fix for pg_dumpall(Bruce)
3417 Rename pg_attribute.attnvals to attdispersion(Bruce)
3418 "intro/unix" manual page now "pgintro"(Bruce)
3419 "built-in" manual page now "pgbuiltin"(Bruce)
3420 "drop" manual page now "drop_table"(Bruce)
3421 Add "create_trigger", "drop_trigger" manual pages(Thomas)
3422 Add constraints regression test(Vadim & Thomas)
3423 Add comments syntax regression test(Thomas)
3424 Add PGINDENT and support program(Bruce)
3425 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
3426 Files moved to /src/tools directory(Bruce)
3427 SPI and Trigger programming guides (Vadim & D'Arcy)
3433 <sect1 id="release-6-1-1">
3434 <title>Release 6.1.1</title>
3437 <title>Release date</title>
3438 <simpara>1997-07-22</simpara>
3442 <title>Migration from version 6.1 to version 6.1.1</title>
3445 This is a minor bug-fix release. A dump/reload is not required from version 6.1,
3446 but is required from any release prior to 6.1.
3447 Refer to the release notes for 6.1 for more details.
3452 <title>Changes</title>
3456 fix for SET with options (Thomas)
3457 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
3458 new psql \connect option allows changing usernames without changing databases
3459 fix for initdb --debug option(Yoshihiko Ichikawa))
3460 lextest cleanup(Bruce)
3462 fix date/time month boundary arithmetic(Thomas)
3463 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
3464 timestamp overhauled to use standard functions(Thomas)
3465 other code cleanup in date/time routines(Thomas)
3466 psql's \d now case-insensitive(Bruce)
3467 psql's backslash commands can now have trailing semicolon(Bruce)
3468 fix memory leak in psql when using \g(Bruce)
3469 major fix for endian handling of communication to server(Thomas, Tatsuo)
3470 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
3471 allow underscores in usernames(Bruce)
3472 pg_dumpall now returns proper status, portability fix(Bruce)
3478 <sect1 id="release-6-1">
3479 <title>Release 6.1</title>
3482 <title>Release date</title>
3483 <simpara>1997-06-08</simpara>
3487 The regression tests have been adapted and extensively modified for the
3488 6.1 release of <productname>PostgreSQL</productname>.
3492 Three new data types (<type>datetime</type>, <type>timespan</type>, and <type>circle</type>) have been added to
3493 the native set of <productname>PostgreSQL</productname> types. Points, boxes, paths, and polygons
3494 have had their output formats made consistent across the data types.
3495 The polygon output in misc.out has only been spot-checked for correctness
3496 relative to the original regression output.
3500 <productname>PostgreSQL</productname> 6.1 introduces a new, alternate
3501 optimizer which uses <firstterm>genetic</firstterm>
3502 algorithms. These algorithms introduce a random behavior in the ordering
3503 of query results when the query contains multiple qualifiers or multiple
3504 tables (giving the optimizer a choice on order of evaluation). Several
3505 regression tests have been modified to explicitly order the results, and
3506 hence are insensitive to optimizer choices. A few regression tests are
3507 for data types which are inherently unordered (e.g. points and time
3508 intervals) and tests involving those types are explicitly bracketed with
3509 <command>set geqo to 'off'</command> and <command>reset geqo</command>.
3513 The interpretation of array specifiers (the curly braces around atomic
3514 values) appears to have changed sometime after the original regression
3515 tests were generated. The current <filename>./expected/*.out</filename> files reflect this
3516 new interpretation, which may not be correct!
3520 The float8 regression test fails on at least some platforms. This is due
3521 to differences in implementations of <function>pow()</function> and <function>exp()</function> and the signaling
3522 mechanisms used for overflow and underflow conditions.
3526 The "random" results in the random test should cause the "random" test
3527 to be "failed", since the regression tests are evaluated using a simple
3528 diff. However, "random" does not seem to produce random results on my
3529 test machine (Linux/gcc/i686).
3533 <title>Migration to version 6.1</title>
3536 This migration requires a complete dump of the 6.0 database and a
3537 restore of the database in 6.1.
3540 Those migrating from earlier 1.* releases should first upgrade to 1.09
3541 because the COPY output format was improved from the 1.02 release.
3546 <title>Changes</title>
3552 packet length checking in library routines
3553 lock manager priority patch
3554 check for under/over flow of float8(Bruce)
3555 multitable join fix(Vadim)
3556 SIGPIPE crash fix(Darren)
3557 large object fixes(Sven)
3558 allow btree indexes to handle NULLs(Vadim)
3559 timezone fixes(D'Arcy)
3560 select SUM(x) can return NULL on no rows(Thomas)
3561 internal optimizer, executor bug fixes(Vadim)
3562 fix problem where inner loop in < or <= has no rows(Vadim)
3563 prevent re-commuting join index clauses(Vadim)
3564 fix join clauses for multiple tables(Vadim)
3565 fix hash, hashjoin for arrays(Vadim)
3566 fix btree for abstime type(Vadim)
3567 large object fixes(Raymond)
3568 fix buffer leak in hash indexes (Vadim)
3569 fix rtree for use in inner scan (Vadim)
3570 fix gist for use in inner scan, cleanups (Vadim, Andrea)
3571 avoid unnecessary local buffers allocation (Vadim, Massimo)
3572 fix local buffers leak in transaction aborts (Vadim)
3573 fix file manager memmory leaks, cleanups (Vadim, Massimo)
3574 fix storage manager memmory leaks (Vadim)
3575 fix btree duplicates handling (Vadim)
3576 fix deleted tuples re-incarnation caused by vacuum (Vadim)
3577 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
3578 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
3582 attribute optimization statistics(Bruce)
3583 much faster new btree bulk load code(Paul)
3584 BTREE UNIQUE added to bulk load code(Vadim)
3585 new lock debug code(Massimo)
3586 massive changes to libpg++(Leo)
3587 new GEQO optimizer speeds table multitable optimization(Martin)
3588 new WARN message for non-unique insert into unique key(Marc)
3589 update x=-3, no spaces, now valid(Bruce)
3590 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
3591 debug backend now pretty-prints tree(Darren)
3592 new Oracle character functions(Edmund)
3593 new plaintext password functions(Dan)
3594 no such class or insufficient privilege changed to distinct messages(Dan)
3595 new ANSI timestamp function(Dan)
3596 new ANSI Time and Date types (Thomas)
3597 move large chunks of data in backend(Martin)
3598 multicolumn btree indexes(Vadim)
3599 new SET var TO value command(Martin)
3600 update transaction status on reads(Dan)
3601 new locale settings for character types(Oleg)
3602 new SEQUENCE serial number generator(Vadim)
3603 GROUP BY function now possible(Vadim)
3604 re-organize regression test(Thomas,Marc)
3605 new optimizer operation weights(Vadim)
3606 new psql \z grant/permit option(Marc)
3607 new MONEY data type(D'Arcy,Thomas)
3608 tcp socket communication speed improved(Vadim)
3609 new VACUUM option for attribute statistics, and for certain columns (Vadim)
3610 many geometric type improvements(Thomas,Keith)
3611 additional regression tests(Thomas)
3612 new datestyle variable(Thomas,Vadim,Martin)
3613 more comparison operators for sorting types(Thomas)
3614 new conversion functions(Thomas)
3615 new more compact btree format(Vadim)
3616 allow pg_dumpall to preserve database ownership(Bruce)
3617 new SET GEQO=# and R_PLANS variable(Vadim)
3618 old (!GEQO) optimizer can use right-sided plans (Vadim)
3619 typechecking improvement in SQL parser(Bruce)
3620 new SET, SHOW, RESET commands(Thomas,Vadim)
3621 new \connect database USER option
3622 new destroydb -i option (Igor)
3623 new \dt and \di psql commands (Darren)
3624 SELECT "\n" now escapes newline (A. Duursma)
3625 new geometry conversion functions from old format (Thomas)
3629 new configuration script(Marc)
3630 readline configuration option added(Marc)
3631 OS-specific configuration options removed(Marc)
3632 new OS-specific template files(Marc)
3633 no more need to edit Makefile.global(Marc)
3634 re-arrange include files(Marc)
3635 nextstep patches (Gregor Hoffleit)
3636 removed WIN32-specific code(Bruce)
3637 removed postmaster -e option, now only postgres -e option (Bruce)
3638 merge duplicate library code in front/backends(Martin)
3639 now works with eBones, international Kerberos(Jun)
3640 more shared library support
3641 c++ include file cleanup(Bruce)
3642 warn about buggy flex(Bruce)
3643 DG/UX, Ultrix, IRIX, AIX portability fixes
3649 <sect1 id="release-6-0">
3650 <title>Release 6.0</title>
3653 <title>Release date</title>
3654 <simpara>1997-01-29</simpara>
3658 A dump/restore is required for those wishing to migrate data from
3659 previous releases of <productname>PostgreSQL</productname>.
3663 <title>Migration from version 1.09 to version 6.0</title>
3666 This migration requires a complete dump of the 1.09 database and a
3667 restore of the database in 6.0.
3672 <title>Migration from pre-1.09 to version 6.0</title>
3675 Those migrating from earlier 1.* releases should first upgrade to 1.09
3676 because the COPY output format was improved from the 1.02 release.
3681 <title>Changes</title>
3687 ALTER TABLE bug - running postgress process needs to re-read table definition
3688 Allow vacuum to be run on one table or entire database(Bruce)
3690 Fix array over-runs of memory writes(Kurt)
3691 Fix elusive btree range/non-range bug(Dan)
3692 Fix for hash indexes on some types like time and date
3693 Fix for pg_log size explosion
3694 Fix permissions on lo_export()(Bruce)
3695 Fix unitialized reads of memory(Kurt)
3696 Fixed ALTER TABLE ... char(3) bug(Bruce)
3697 Fixed a few small memory leaks
3698 Fixed EXPLAIN handling of options and changed full_path option name
3699 Fixed output of group acl permissions
3700 Memory leaks (hunt and destroy with tools like Purify(Kurt)
3701 Minor improvements to rules system
3703 New asserts for run-checking
3704 Overhauled parser/analyze code to properly report errors and increase speed
3705 Pg_dump -d now handles NULL's properly(Bruce)
3706 Prevent SELECT NULL from crashing server (Bruce)
3707 Properly report errors when INSERT ... SELECT columns did not match
3708 Properly report errors when insert column names were not correct
3709 Psql \g filename now works(Bruce)
3710 Psql fixed problem with multiple statements on one line with multiple outputs
3711 Removed duplicate system OIDs
3712 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
3713 Several fixes for queries that crashed the backend
3714 Starting quote in insert string errors(Bruce)
3715 Submitting an empty query now returns empty status, not just " " query(Bruce)
3719 Add EXPLAIN manual page(Bruce)
3720 Add UNIQUE index capability(Dan)
3721 Add hostname/user level access control rather than just hostname and user
3722 Add synonym of != for <>(Bruce)
3723 Allow "select oid,* from table"
3724 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Bruce)
3725 Allow COPY from the frontend(Bryan)
3726 Allow GROUP BY to use alias column name(Bruce)
3727 Allow actual compression, not just reuse on the same page(Vadim)
3728 Allow installation-configuration option to auto-add all local users(Bryan)
3729 Allow libpq to distinguish between text value '' and null(Bruce)
3730 Allow non-postgres users with createdb privs to destroydb's
3731 Allow restriction on who can create C functions(Bryan)
3732 Allow restriction on who can do backend COPY(Bryan)
3733 Can shrink tables, pg_time and pg_log(Vadim & Erich)
3734 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
3735 Change default decimal constant representation from float4 to float8(Bruce)
3736 European date format now set when postmaster is started
3737 Execute lowercase function names if not found with exact case
3738 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z'
3739 Gist now included in the distrubution(Marc)
3740 Idend authentication of local users(Bryan)
3741 Implement BETWEEN qualifier(Bruce)
3742 Implement IN qualifier(Bruce)
3743 Libpq has PQgetisnull()(Bruce)
3744 Libpq++ improvements
3745 New options to initdb(Bryan)
3746 Pg_dump allow dump of OIDs(Bruce)
3747 Pg_dump create indexes after tables are loaded for speed(Bruce)
3748 Pg_dumpall dumps all databases, and the user table
3749 Pginterface additions for NULL values(Bruce)
3750 Prevent postmaster from being run as root
3751 Psql \h and \? is now readable(Bruce)
3752 Psql allow backslashed, semicolons anywhere on the line(Bruce)
3753 Psql changed command prompt for lines in query or in quotes(Bruce)
3754 Psql char(3) now displays as (bp)char in \d output(Bruce)
3755 Psql return code now more accurate(Bryan?)
3756 Psql updated help syntax(Bruce)
3757 Re-visit and fix vacuum(Vadim)
3758 Reduce size of regression diffs, remove timezone name difference(Bruce)
3759 Remove compile-time parameters to enable binary distributions(Bryan)
3760 Reverse meaning of HBA masks(Bryan)
3761 Secure Authentication of local users(Bryan)
3762 Speed up vacuum(Vadim)
3763 Vacuum now had VERBOSE option(Bruce)
3767 All functions now have prototypes that are compared against the calls
3768 Allow asserts to be disabled easly from Makefile.global(Bruce)
3769 Change oid constants used in code to #define names
3770 Decoupled sparc and solaris defines(Kurt)
3771 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
3772 Major include file reorganization/reduction(Marc)
3773 Make now stops on compile failure(Bryan)
3774 Makefile restructuring(Bryan, Marc)
3775 Merge bsdi_2_1 to bsdi(Bruce)
3776 Monitor program removed
3777 Name change from Postgres95 to PostgreSQL
3778 New config.h file(Marc, Bryan)
3779 PG_VERSION now set to 6.0 and used by postmaster
3780 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
3781 Reduced the number of #define's, centeralized #define's
3782 Remove duplicate OIDS in system tables(Dan)
3783 Remove duplicate system catalog info or report mismatches(Dan)
3784 Removed many os-specific #define's
3785 Restructured object file generation/location(Bryan, Marc)
3786 Restructured port-specific file locations(Bryan, Marc)
3787 Unused/uninialized variables corrected
3793 <sect1 id="release-1-09">
3794 <title>Release 1.09</title>
3797 <title>Release date</title>
3798 <simpara>1996-11-04</simpara>
3802 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of
3803 the changes listed in 6.0 were actually included in the 1.02.1 to 1.09
3808 <sect1 id="release-1-02">
3809 <title>Release 1.02</title>
3812 <title>Release date</title>
3813 <simpara>1996-08-01</simpara>
3817 <title>Migration from version 1.02 to version 1.02.1</title>
3820 Here is a new migration file for 1.02.1. It includes the 'copy' change
3821 and a script to convert old <acronym>ASCII</acronym> files.
3825 The following notes are for the benefit of users who want to migrate
3826 databases from <productname>Postgres95</> 1.01 and 1.02 to <productname>Postgres95</> 1.02.1.
3829 If you are starting afresh with <productname>Postgres95</> 1.02.1 and do not need
3830 to migrate old databases, you do not need to read any further.
3835 In order to upgrade older <productname>Postgres95</> version 1.01 or 1.02 databases to
3836 version 1.02.1, the following steps are required:
3841 Start up a new 1.02.1 postmaster
3846 Add the new built-in functions and operators of 1.02.1 to 1.01 or 1.02
3847 databases. This is done by running the new 1.02.1 server against
3848 your own 1.01 or 1.02 database and applying the queries attached at
3849 the end of the file. This can be done easily through <command>psql</>. If your
3850 1.01 or 1.02 database is named <literal>testdb</literal> and you have cut the commands
3851 from the end of this file and saved them in <filename>addfunc.sql</filename>:
3853 % psql testdb -f addfunc.sql
3856 Those upgrading 1.02 databases will get a warning when executing the
3857 last two statements in the file because they are already present in 1.02. This is
3858 not a cause for concern.
3865 <title>Dump/Reload Procedure</title>
3868 If you are trying to reload a pg_dump or text-mode, <literal>copy tablename to
3869 stdout</literal> generated with a previous version, you will need to run the
3870 attached <command>sed</command> script on the ASCII file before loading it into the
3871 database. The old format used '.' as end-of-data, while '\.' is now the
3872 end-of-data marker. Also, empty strings are now loaded in as '' rather
3873 than NULL. See the copy manual page for full details.
3876 sed 's/^\.$/\\./g' <in_file >out_file
3880 If you are loading an older binary copy or non-stdout copy, there is no
3881 end-of-data character, and hence no conversion necessary.
3884 -- following lines added by agc to reflect the case-insensitive
3885 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
3886 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexeq);
3887 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregexne);
3888 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregexeq);
3889 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticregexne);
3895 <title>Changes</title>
3899 Source code maintenance and development
3900 * worldwide team of volunteers
3901 * the source tree now in CVS at ftp.ki.net
3904 * psql (and underlying libpq library) now has many more options for
3905 formatting output, including HTML
3906 * pg_dump now output the schema and/or the data, with many fixes to
3907 enhance completeness.
3908 * psql used in place of monitor in administration shell scripts.
3909 monitor to be depreciated in next release.
3910 * date/time functions enhanced
3911 * NULL insert/update/comparison fixed/enhanced
3912 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
3914 Bug Fixes (almost too numerous to mention)
3916 * storage management
3917 * check for NULL pointer before dereferencing
3921 * added SolarisX86 port
3922 * added BSD/OS 2.1 port
3927 Contributors (appologies to any missed)
3928 * Kurt J. Lidl <lidl@va.pubnix.com>
3929 (missed in first run, but no less important)
3930 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
3931 * Jason Wright <jason@shiloh.vnet.net>
3932 * Cees de Groot <C.deGroot@inter.NL.net>
3933 * ernst.molitor@uni-bonn.de
3934 * michael.siebenborn@ae3.Hypo.DE (Michael Siebenborn (6929))
3935 * Brian E. Gallew <geek+@cmu.edu>
3936 * Vadim B. Mikheev <vadim@sable.krasnoyarsk.su>
3937 * Adam Sussman <myddryn@vidya.com>
3938 * Chris Dunlop <chris@onthe.net.au>
3939 * Marc G. Fournier <scrappy@ki.net>
3940 * Dan McGuirk <mcguirk@indirect.com>
3941 * Dr_George_D_Detlefsen <drgeorge@ilt.com>
3942 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
3943 * Massimo Dal Zotto <dz@cs.unitn.it>
3944 * Randy Kunkee <kunkee@Starbase.NeoSoft.COM>
3945 * Rick Weldon <rick@wisetech.com>
3946 * Thomas van Reimersdahl <reimersd@dali.techinfo.rwth-aachen.de>
3947 * david bennett <dave@bensoft.com>
3948 * ernst.molitor@uni-bonn.de
3949 * Julian Assange <proff@suburbia.net>
3950 * Bruce Momjian <maillist@candle.pha.pa.us>
3951 * Paul "Shag" Walmsley <ccshag@cclabs.missouri.edu>
3952 * "Alistair G. Crooks" <azcb0@sde.uts.amdahl.com>
3957 <sect1 id="release-1-01">
3958 <title>Release 1.01</title>
3961 <title>Release date</title>
3962 <simpara>1996-02-23</simpara>
3967 <title>Migration from version 1.0 to version 1.01</title>
3970 The following notes are for the benefit of users who want to migrate
3971 databases from <productname>Postgres95</> 1.0 to <productname>Postgres95</> 1.01.
3974 If you are starting afresh with <productname>Postgres95</> 1.01 and do not need
3975 to migrate old databases, you do not need to read any further.
3978 In order to <productname>Postgres95</> version 1.01 with databases created with
3979 <productname>Postgres95</> version 1.0, the following steps are required:
3984 Set the definition of <symbol>NAMEDATALEN</symbol> in <filename>src/Makefile.global</filename> to 16
3985 and <symbol>OIDNAMELEN</symbol> to 20.
3990 Decide whether you want to use Host based authentication.
3995 If you do, you must create a file name <literal>pg_hba</literal> in your top-level data
3996 directory (typically the value of your <envar>$PGDATA</envar>). <filename>src/libpq/pg_hba</filename>
3997 shows an example syntax.
4002 If you do not want host-based authentication, you can comment out
4007 in <filename>src/Makefile.global</filename>
4010 Note that host-based authentication is turned on by default, and if
4011 you do not take steps A or B above, the out-of-the-box 1.01 will
4012 not allow you to connect to 1.0 databases.
4020 Compile and install 1.01, but DO NOT do the <command>initdb</command> step.
4025 Before doing anything else, terminate your 1.0 postmaster, and
4026 backup your existing <envar>$PGDATA</envar> directory.
4031 Set your <envar>PGDATA</envar> environment variable to your 1.0 databases, but set up
4032 path up so that 1.01 binaries are being used.
4037 Modify the file <filename><envar>$PGDATA</envar>/PG_VERSION</filename> from 5.0 to 5.1
4042 Start up a new 1.01 postmaster
4047 Add the new built-in functions and operators of 1.01 to 1.0
4048 databases. This is done by running the new 1.01 server against
4049 your own 1.0 database and applying the queries attached and saving
4050 in the file 1.0_to_1.01.sql. This can be done easily through <command>psql</command>.
4051 If your 1.0 database is name <literal>testdb</literal>:
4054 % psql testdb -f 1.0_to_1.01.sql
4057 and then execute the following commands (cut and paste from here):
4060 -- add builtin functions that are new to 1.01
4062 create function int4eqoid (int4, oid) returns bool as 'foo'
4063 language 'internal';
4064 create function oideqint4 (oid, int4) returns bool as 'foo'
4065 language 'internal';
4066 create function char2icregexeq (char2, text) returns bool as 'foo'
4067 language 'internal';
4068 create function char2icregexne (char2, text) returns bool as 'foo'
4069 language 'internal';
4070 create function char4icregexeq (char4, text) returns bool as 'foo'
4071 language 'internal';
4072 create function char4icregexne (char4, text) returns bool as 'foo'
4073 language 'internal';
4074 create function char8icregexeq (char8, text) returns bool as 'foo'
4075 language 'internal';
4076 create function char8icregexne (char8, text) returns bool as 'foo'
4077 language 'internal';
4078 create function char16icregexeq (char16, text) returns bool as 'foo'
4079 language 'internal';
4080 create function char16icregexne (char16, text) returns bool as 'foo'
4081 language 'internal';
4082 create function texticregexeq (text, text) returns bool as 'foo'
4083 language 'internal';
4084 create function texticregexne (text, text) returns bool as 'foo'
4085 language 'internal';
4087 -- add builtin functions that are new to 1.01
4089 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
4090 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
4091 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexeq);
4092 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregexne);
4093 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexeq);
4094 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregexne);
4095 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexeq);
4096 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregexne);
4097 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icregexeq);
4098 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icregexne);
4099 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq);
4100 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne);
4108 <title>Changes</title>
4113 * 1.01 is backwards compatible with 1.0 database provided the user
4114 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
4115 If those steps are not taken, 1.01 is not compatible with 1.0 database.
4118 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
4119 * added NeXT port (requires SysVIPC implementation)
4120 * added CAST .. AS ... syntax
4121 * added ASC and DESC keywords
4122 * added 'internal' as a possible language for CREATE FUNCTION
4123 internal functions are C functions which have been statically linked
4124 into the postgres backend.
4125 * a new type "name" has been added for system identifiers (table names,
4126 attribute names, etc.) This replaces the old char16 type. The
4127 of name is set by the NAMEDATALEN #define in src/Makefile.global
4128 * a readable reference manual that describes the query language.
4129 * added host-based access control. A configuration file ($PGDATA/pg_hba)
4130 is used to hold the configuration data. If host-based access control
4131 is not desired, comment out HBA=1 in src/Makefile.global.
4132 * changed regex handling to be uniform use of Henry Spencer's regex code
4133 regardless of platform. The regex code is included in the distribution
4134 * added functions and operators for case-insensitive regular expressions.
4135 The operators are ~* and !~*.
4136 * pg_dump uses COPY instead of SELECT loop for better performance
4139 * fixed an optimizer bug that was causing core dumps when
4140 functions calls were used in comparisons in the WHERE clause
4141 * changed all uses of getuid to geteuid so that effective uids are used
4142 * psql now returns non-zero status on errors when using -c
4143 * applied public patches 1-14
4149 <sect1 id="release-1-0">
4150 <title>Release 1.0</title>
4153 <title>Release date</title>
4154 <simpara>1995-09-05</simpara>
4158 <title>Changes</title>
4163 * The copyright of <productname>Postgres</productname> 1.0 has been loosened to be freely modifiable
4164 and modifiable for any purpose. Please read the COPYRIGHT file.
4165 Thanks to Professor Michael Stonebraker for making this possible.
4168 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
4169 EUROPEAN STYLE). This follows SQL-92 specs.
4170 * "delimiters" is now a keyword
4173 * sql LIKE syntax has been added
4174 * copy command now takes an optional USING DELIMITER specification.
4175 delimiters can be any single-character string.
4176 * IRIX 5.3 port has been added.
4177 Thanks to Paul Walmsley and others.
4178 * updated pg_dump to work with new libpq
4179 * \d has been added psql
4180 Thanks to Keith Parks
4181 * regexp performance for architectures that use POSIX regex has been
4182 improved due to caching of precompiled patterns.
4183 Thanks to Alistair Crooks
4184 * a new version of libpq++
4185 Thanks to William Wanders
4188 * arbitrary userids can be specified in the createuser script
4189 * \c to connect to other databases in psql now works.
4190 * bad pg_proc entry for float4inc() is fixed
4191 * users with usecreatedb field set can now create databases without
4192 having to be usesuper
4193 * remove access control entries when the entry no longer has any
4195 * fixed non-portable datetimes implementation
4196 * added kerberos flags to the src/backend/Makefile
4197 * libpq now works with kerberos
4198 * typographic errors in the user manual have been corrected.
4199 * btrees with multiple index never worked, now we tell you they don't
4200 work when you try to use them
4206 <sect1 id="release-0-03">
4207 <title><productname>Postgres95</productname> Release 0.03</title>
4210 <title>Release date</title>
4211 <simpara>1995-07-21</simpara>
4215 <title>Changes</title>
4218 Incompatible changes:
4219 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
4220 (due to system catalog changes and indexing structure changes).
4221 * double-quote (") is deprecated as a quoting character for string literals;
4222 you need to convert them to single quotes ('). <!-- " -->
4223 * name of aggregates (eg. int4sum) are renamed in accordance with the
4224 SQL standard (eg. sum).
4225 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
4226 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
4227 previous releases); you might have to do typecasting if you depend on it
4228 being of type float8. If you neglect to do the typecasting and you assign
4229 a float literal to a field of type float8, you may get incorrect values
4231 * LIBPQ has been totally revamped so that frontend applications
4232 can connect to multiple backends
4233 * the usesysid field in pg_user has been changed from int2 to int4 to
4234 allow wider range of Unix user ids.
4235 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
4236 single BSD44_derived port. (thanks to Alistair Crooks)
4238 SQL standard-compliance (the following details changes that makes postgres95
4239 more compliant to the SQL-92 standard):
4240 * the following SQL types are now built-in: smallint, int(eger), float, real,
4241 char(N), varchar(N), date and time.
4243 The following are aliases to existing postgres types:
4245 integer, int -> int4
4246 float, real -> float4
4247 char(N) and varchar(N) are implemented as truncated text types. In
4248 addition, char(N) does blank-padding.
4249 * single-quote (') is used for quoting string literals; '' (in addition to
4250 \') is supported as means of inserting a single quote in a string
4251 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
4252 (Also, aggregates can now be overloaded, i.e. you can define your
4253 own MAX aggregate to take in a user-defined type.)
4254 * CHANGE ACL removed. GRANT/REVOKE syntax added.
4255 - Privileges can be given to a group using the "GROUP" keyword.
4257 GRANT SELECT ON foobar TO GROUP my_group;
4258 The keyword 'PUBLIC' is also supported to mean all users.
4260 Privileges can only be granted or revoked to one user or group
4263 "WITH GRANT OPTION" is not supported. Only class owners can change
4265 - The default access control is to to grant users readonly access.
4266 You must explicitly grant insert/update access to users. To change
4267 this, modify the line in
4268 src/backend/utils/acl.h
4269 that defines ACL_WORLD_DEFAULT
4272 * the bug where aggregates of empty tables were not run has been fixed. Now,
4273 aggregates run on empty tables will return the initial conditions of the
4274 aggregates. Thus, COUNT of an empty table will now properly return 0.
4275 MAX/MIN of an empty table will return a tuple of value NULL.
4276 * allow the use of \; inside the monitor
4277 * the LISTEN/NOTIFY asynchronous notification mechanism now work
4278 * NOTIFY in rule action bodies now work
4279 * hash indexes work, and access methods in general should perform better.
4280 creation of large btree indexes should be much faster. (thanks to Paul
4283 Other changes and enhancements:
4284 * addition of an EXPLAIN statement used for explaining the query execution
4285 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
4287 * WARN and NOTICE messages no longer have timestamps on them. To turn on
4288 timestamps of error messages, uncomment the line in
4289 src/backend/utils/elog.h:
4290 /* define ELOG_TIMESTAMPS */
4291 * On an access control violation, the message
4292 "Either no such class or insufficient privilege"
4293 will be given. This is the same message that is returned when
4294 a class is not found. This dissuades non-privileged users from
4295 guessing the existence of privileged classes.
4296 * some additional system catalog changes have been made that are not
4297 visible to the user.
4300 * The -oid option has been added to the "pg_result" tcl command.
4301 pg_result -oid returns oid of the last tuple inserted. If the
4302 last command was not an INSERT, then pg_result -oid returns "".
4303 * the large object interface is available as pg_lo* tcl commands:
4304 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
4306 Portability enhancements and New Ports:
4307 * flex/lex problems have been cleared up. Now, you should be able to use
4308 flex instead of lex on any platforms. We no longer make assumptions of
4309 what lexer you use based on the platform you use.
4310 * The Linux-ELF port is now supported. Various configuration have been
4311 tested: The following configuration is known to work:
4312 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
4313 with everything in ELF format,
4316 * ipcclean added to the distribution
4317 ipcclean usually does not need to be run, but if your backend crashes
4318 and leaves shared memory segments hanging around, ipcclean will
4319 clean them up for you.
4322 * the user manual has been revised and libpq documentation added.
4328 <sect1 id="release-0-02">
4329 <title><productname>Postgres95</productname> Release 0.02</title>
4332 <title>Release date</title>
4333 <simpara>1995-05-25</simpara>
4337 <title>Changes</title>
4341 Incompatible changes:
4342 * The SQL statement for creating a database is 'CREATE DATABASE' instead
4343 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
4344 of 'DESTROYDB'. However, the names of the executables 'createdb' and
4345 'destroydb' remain the same.
4348 * pgperl - a Perl (4.036) interface to Postgres95
4349 * pg_dump - a utility for dumping out a postgres database into a
4350 script file containing query commands. The script files are in a ASCII
4351 format and can be used to reconstruct the database, even on other
4352 machines and other architectures. (Also good for converting
4353 a Postgres 4.2 database to Postgres95 database.)
4355 The following ports have been incorporated into postgres95-beta-0.02:
4356 * the NetBSD port by Alistair Crooks
4357 * the AIX port by Mike Tung
4358 * the Windows NT port by Jon Forrest (more stuff but not done yet)
4359 * the Linux ELF port by Brian Gallew
4361 The following bugs have been fixed in postgres95-beta-0.02:
4362 * new lines not escaped in COPY OUT and problem with COPY OUT when first
4364 * cannot type return to use the default user id in createuser
4365 * SELECT DISTINCT on big tables crashes
4366 * Linux installation problems
4367 * monitor doesn't allow use of 'localhost' as PGHOST
4368 * psql core dumps when doing \c or \l
4369 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
4370 * libpgtcl has a hard-wired default port number
4371 * SELECT DISTINCT INTO TABLE hangs
4372 * CREATE TYPE doesn't accept 'variable' as the internallength
4373 * wrong result using more than 1 aggregate in a SELECT
4379 <sect1 id="release-0-01">
4380 <title><productname>Postgres95</productname> Release 0.01</title>
4383 <title>Release date</title>
4384 <simpara>1995-05-01</simpara>
4393 <sect1 id="timing-results">
4394 <title>Timing Results</title>
4397 These timing results are from running the regression test with the commands
4400 % cd src/test/regress
4406 Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
4407 to run, presumably due to the scheduling vagaries of multitasking systems.
4411 <title>Version 6.5</title>
4414 As has been the case for previous releases, timing between
4415 releases is not directly comparable since new regression tests
4416 have been added. In general, 6.5 is faster than previous
4421 Timing with <function>fsync()</function> disabled:
4425 02:00 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
4426 04:38 Sparc Ultra 1 143MHz, 64MB, Solaris 2.6
4431 Timing with <function>fsync()</function> enabled:
4435 04:21 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
4438 For the <systemitem class="osname">Linux</systemitem> system above, using <acronym>UW-SCSI</acronym> disks rather than (older) <acronym>IDE</acronym>
4439 disks leads to a 50% improvement in speed on the regression test.
4444 <title>Version 6.4beta</title>
4447 The times for this release are not directly comparable to those for previous releases
4448 since some additional regression tests have been included.
4449 In general, however, 6.4 should be slightly faster than the previous release (thanks, Bruce!).
4454 02:26 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
4460 <title>Version 6.3</title>
4463 The times for this release are not directly comparable to those for previous releases
4464 since some additional regression tests have been included and some obsolete tests involving
4465 time travel have been removed.
4466 In general, however, 6.3 is substantially faster than previous releases (thanks, Bruce!).
4471 02:30 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
4472 04:12 Dual Pentium Pro 180, 96MB, EIDE, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
4478 <title>Version 6.1</title>
4483 06:12 Pentium Pro 180, 32MB, EIDE, Linux 2.0.30, gcc 2.7.2 -O2 -m486
4484 12:06 P-100, 48MB, Linux 2.0.29, gcc
4485 39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
4493 <!-- Keep this comment at the end of the file
4498 sgml-minimize-attributes:nil
4499 sgml-always-quote-attributes:t
4502 sgml-parent-document:nil
4503 sgml-default-dtd-file:"./reference.ced"
4504 sgml-exposed-tags:nil
4505 sgml-local-catalogs:("/usr/lib/sgml/catalog")
4506 sgml-local-ecat-files:nil