6 Below is a subset of the changes that have gone into the development
7 branch of PostgreSQL since version 7.3. For a complete list of
8 changes, consult the CVS logs.
11 Domains now support CHECK constraints
12 _________________________________________________________________
16 Release date: 2002-12-18
18 This has a variety of fixes from 7.3.
19 _________________________________________________________________
21 Migration to version 7.3.1
23 A dump/restore is *not* required for those running 7.3. However, it
24 should be noted that the main PostgreSQL interface library, libpq, has
25 a new major version number for this release, which may require
26 recompilation of client code in certain cases.
27 _________________________________________________________________
31 Fix a core dump of COPY TO when client/server encodings don't match (Tom)
32 Allow pg_dump to work with pre-7.2 servers (Philip)
33 /contrib/adddepend fixes (Tom)
34 Fix problem with deletion of per-user/per-database config settings (Tom)
35 /contrib/vacuumlo fix (Tom)
36 Allow 'password' encryption even when pg_shadow contains MD5 passwords
38 /contrib/dbmirror fix (Steven Singer)
40 /contrib/tsearch fixes (Teodor Sigaev, Magnus)
41 Allow locale names to be mixed case (Nicolai Tufar)
42 Increment libpq library's major version number (Bruce)
43 pg_hba.conf error reporting fixes (Bruce, Neil)
44 Add SCO Openserver 5.0.4 as a supported platform (Bruce)
45 Prevent EXPLAIN from crashing server (Tom)
46 SSL fixes (Nathan Mueller)
47 Prevent composite column creation via ALTER TABLE (Tom)
48 _________________________________________________________________
52 Release date: 2002-11-27
53 _________________________________________________________________
57 Major changes in this release:
60 Schemas allow users to create objects in separate namespaces,
61 so two people or applications can have tables with the same
62 name. There is also a public schema for shared tables.
63 Table/index creation can be restricted by removing permissions
67 PostgreSQL now supports the ALTER TABLE ... DROP COLUMN
71 Functions returning multiple rows and/or multiple columns are
72 now much easier to use than before. You can call such a "table
73 function" in the SELECT FROM clause, treating its output like a
74 table. Also, PL/pgSQL functions can now return sets.
77 PostgreSQL now supports prepared queries, for improved
81 PostgreSQL now records object dependencies, which allows
82 improvements in many areas. "DROP" statements now take either
83 CASCADE or RESTRICT to control whether dependent objects are
87 Functions and procedural languages now have privileges, and
88 functions can be defined to run with the privileges of their
92 Both multibyte and locale support are now always enabled.
95 A variety of logging options have been enhanced.
98 A large number of interfaces have been moved to
99 http://gborg.postgresql.org where they can be developed and
100 released independently.
102 Functions/Identifiers
103 By default, functions can now take up to 32 parameters, and
104 identifiers can be up to 63 bytes long. Also, OPAQUE is now
105 deprecated: there are specific "pseudo-datatypes" to represent
106 each of the former meanings of OPAQUE in function argument and
108 _________________________________________________________________
110 Migration to version 7.3
112 A dump/restore using pg_dump is required for those wishing to migrate
113 data from any previous release. If your application examines the
114 system catalogs, additional changes will be required due to the
115 introduction of schemas in 7.3; for more information, see:
116 http://www.ca.postgresql.org/docs/momjian/upgrade_tips_7.3.
118 Observe the following incompatibilities:
120 * Pre-6.3 clients are no longer supported.
121 * "pg_hba.conf" now has a column for the user name and additional
122 features. Existing files need to be adjusted.
123 * Several "postgresql.conf" logging parameters have been renamed.
124 * LIMIT #,# has been disabled; use LIMIT # OFFSET #.
125 * "INSERT" statements with column lists must specify a value for
126 each specified column. For example, INSERT INTO tab (col1, col2)
127 VALUES ('val1') is now invalid. It's still allowed to supply fewer
128 columns than expected if the "INSERT" does not have a column list.
129 * serial columns are no longer automatically UNIQUE; thus, an index
130 will not automatically be created.
131 * A "SET" command inside an aborted transaction is now rolled back.
132 * "COPY" no longer considers missing trailing columns to be null.
133 All columns need to be specified. (However, one may achieve a
134 similar effect by specifying a column list in the "COPY" command.)
135 * The data type timestamp is now equivalent to timestamp without
136 time zone, instead of timestamp with time zone.
137 * Pre-7.3 databases loaded into 7.3 will not have the new object
138 dependencies for serial columns, unique constraints, and foreign
139 keys. See the directory "contrib/adddepend/" for a detailed
140 description and a script that will add such dependencies.
141 * An empty string ('') is no longer allowed as the input into an
142 integer field. Formerly, it was silently interpreted as 0.
143 _________________________________________________________________
149 Add pg_locks view to show locks (Neil)
150 Security fixes for password negotiation memory allocation (Neil)
151 Remove support for version 0 FE/BE protocol (PostgreSQL 6.2 and
153 Reserve the last few backend slots for superusers, add parameter
154 superuser_reserved_connections to control this (Nigel J. Andrews)
155 _________________________________________________________________
159 Improve startup by calling localtime() only once (Tom)
160 Cache system catalog information in flat files for faster startup
162 Improve caching of index information (Tom)
163 Optimizer improvements (Tom, Fernando Nasser)
164 Catalog caches now store failed lookups (Tom)
165 Hash function improvements (Neil)
166 Improve performance of query tokenization and network handling (Peter)
167 Speed improvement for large object restore (Mario Weilguni)
168 Mark expired index entries on first lookup, saving later heap fetches
170 Avoid excessive NULL bitmap padding (Manfred Koizar)
171 Add BSD-licensed qsort() for Solaris, for performance (Bruce)
172 Reduce per-row overhead by four bytes (Manfred Koizar)
173 Fix GEQO optimizer bug (Neil Conway)
174 Make WITHOUT OID actually save four bytes per row (Manfred Koizar)
175 Add default_statistics_target variable to specify ANALYZE buckets
177 Use local buffer cache for temporary tables so no WAL overhead (Tom)
178 Improve free space map performance on large tables (Stephen Marshall,
180 Improved WAL write concurrency (Tom)
181 _________________________________________________________________
185 Add privileges on functions and procedural languages (Peter)
186 Add OWNER to CREATE DATABASE so superusers can create databases
187 on behalf of unprivileged users. (Gavin Sherry, Tom)
188 Add new object permission bits EXECUTE and USAGE (Tom)
189 Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION
191 Allow functions to be executed with the privilege of the function
193 _________________________________________________________________
197 Server log messages now tagged with LOG, not DEBUG (Bruce)
198 Add user column to pg_hba.conf (Bruce)
199 Have log_connections output two lines in log file (Tom)
200 Remove debug_level from postgresql.conf, now server_min_messages
202 New ALTER DATABASE/USER ... SET command for per-user/database
203 initialization (Peter)
204 New parameters server_min_messages and client_min_messages to
205 control which messages are sent to the server logs or client
207 Allow pg_hba.conf to specify lists of users/databases separated by
208 commas, group names prepended with +, and file names prepended
210 Remove secondary password file capability and pg_password utility
212 Add variable db_user_namespace for database-local user names (Bruce)
213 SSL improvements (Bear Giles)
214 Make encryption of stored passwords the default (Bruce)
215 Allow pg_statistics to be reset by calling pg_stat_reset()
217 Add log_duration parameter (Bruce)
218 Rename debug_print_query to log_statement (Bruce)
219 Rename show_query_stats to show_statement_stats (Bruce)
220 Add param log_min_error_statement to print commands to logs on error
222 _________________________________________________________________
226 Make cursors insensitive, meaning their contents do not change (Tom)
227 Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)
228 Increase identifier length to 63 (Neil, Bruce)
229 UNION fixes for merging >= 3 columns of different lengths (Tom)
230 Add DEFAULT keyword to INSERT, e.g., INSERT ... (..., DEFAULT, ...)
232 Allow views to have default values using ALTER COLUMN ... SET DEFAULT
234 Fail on INSERTs with column lists that don't supply all column
235 values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)
236 Fix for join aliases (Tom)
237 Fix for FULL OUTER JOINs (Tom)
238 Improve reporting of invalid identifier and location (Tom, Gavin)
239 Fix OPEN cursor(args) (Tom)
240 Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)
241 Fix for CREATE TABLE AS with UNION (Tom)
242 SQL99 syntax improvements (Thomas)
243 Add statement_timeout variable to cancel queries (Bruce)
244 Allow prepared queries with PREPARE/EXECUTE (Neil)
245 Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)
246 Add variable autocommit (Tom, David Van Wie)
247 _________________________________________________________________
251 Make equals signs optional in CREATE DATABASE (Gavin Sherry)
252 Make ALTER TABLE OWNER change index ownership too (Neil)
253 New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls
254 TOAST storage, compression (John Gray)
255 Add schema support, CREATE/DROP SCHEMA (Tom)
256 Create schema for temporary tables (Tom)
257 Add variable search_path for schema search (Tom)
258 Add ALTER TABLE SET/DROP NOT NULL (Christopher)
259 New CREATE FUNCTION volatility levels (Tom)
260 Make rule names unique only per table (Tom)
261 Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)
262 Add ALTER TRIGGER RENAME (Joe)
263 New current_schema() and current_schemas() inquiry functions (Tom)
264 Allow functions to return multiple rows (table functions) (Joe)
265 Make WITH optional in CREATE DATABASE, for consistency (Bruce)
266 Add object dependency tracking (Rod, Tom)
267 Add RESTRICT/CASCADE to DROP commands (Rod)
268 Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)
269 Autodestroy sequence on DROP of table with SERIAL (Rod)
270 Prevent column dropping if column is used by foreign key (Rod)
271 Automatically drop constraints/functions when object is dropped (Rod)
272 Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)
273 Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)
274 Prevent inherited columns from being removed or renamed (Alvaro
276 Fix foreign key constraints to not error on intermediate database
278 Propagate column or table renaming to foreign key constraints
279 Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)
280 Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)
281 Have rules execute alphabetically, returning more predictable values
283 Triggers are now fired in alphabetical order (Tom)
284 Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)
285 Allow better casting when inserting/updating values (Tom)
286 _________________________________________________________________
290 Have COPY TO output embedded carriage returns and newlines as \r and
292 Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)
293 Make pg_dump use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)
294 Disable brackets in multistatement rules (Bruce)
295 Disable VACUUM from being called inside a function (Bruce)
296 Allow dropdb and other scripts to use identifiers with spaces (Bruce)
297 Restrict database comment changes to the current database
298 Allow comments on operators, independent of the underlying function
300 Rollback SET commands in aborted transactions (Tom)
301 EXPLAIN now outputs as a query (Tom)
302 Display condition expressions and sort keys in EXPLAIN (Tom)
303 Add 'SET LOCAL var = value' to set configuration variables for a
304 single transaction (Tom)
305 Allow ANALYZE to run in a transaction (Bruce)
306 Improve COPY syntax using new WITH clauses, keep backward
307 compatibility (Bruce)
308 Fix pg_dump to consistently output tags in non-ASCII dumps (Bruce)
309 Make foreign key constraints clearer in dump file (Rod)
310 Add COMMENT ON CONSTRAINT (Rod)
311 Allow COPY TO/FROM to specify column names (Brent Verner)
312 Dump UNIQUE and PRIMARY KEY contraints as ALTER TABLE (Rod)
313 Have SHOW output a query result (Joe)
314 Generate failure on short COPY lines rather than pad NULLs (Neil)
315 Fix CLUSTER to preserve all table attributes (Alvaro Herrera)
316 New pg_settings table to view/modify GUC settings (Joe)
317 Add smart quoting, portability improvements to pg_dump output (Peter)
318 Dump serial columns out as SERIAL (Tom)
319 Enable large file support, >2G for pg_dump (Peter, Philip Warner,
321 Disallow TRUNCATE on tables that are involved in referential
323 Have TRUNCATE also auto-truncate the toast table of the relation (Tom)
324 Add clusterdb utility that will auto-cluster an entire database
325 based on previous CLUSTER operations (Alvaro Herrera)
326 Overhaul pg_dumpall (Peter)
327 Allow REINDEX of TOAST tables (Tom)
328 Implemented START TRANSACTION, per SQL99 (Neil)
329 Fix rare index corruption when a page split affects bulk delete (Tom)
330 Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)
331 _________________________________________________________________
333 Data Types and Functions
335 Fix factorial(0) to return 1 (Bruce)
336 Date/time/timezone improvements (Thomas)
337 Fix for array slice extraction (Tom)
338 Fix extract/date_part to report proper microseconds for timestamp
340 Allow text_substr() and bytea_substr() to read TOAST values more
341 efficiently (John Gray)
342 Add domain support (Rod)
343 Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types
345 Allow alternate storage scheme of 64-bit integers for date/time types
346 using --enable-integer-datetimes in configure (Thomas)
347 Make timezone(timestamptz) return timestamp rather than a string
349 Allow fractional seconds in date/time types for dates prior to 1BC
351 Limit timestamp data types to 6 decimal places of precision (Thomas)
352 Change timezone conversion functions from timetz() to timezone()
354 Add configuration variables datestyle and timezone (Tom)
355 Add OVERLAY(), which allows substitution of a substring in a string
357 Add SIMILAR TO (Thomas, Tom)
358 Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)
359 Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)
360 Add named composite types using CREATE TYPE typename AS (column)
362 Allow composite type definition in the table alias clause (Joe)
363 Add new API to simplify creation of C language table functions (Joe)
364 Remove ODBC-compatible empty parentheses from calls to SQL99
365 functions for which these parentheses do not match the standard
367 Allow macaddr data type to accept 12 hex digits with no separators
369 Add CREATE/DROP CAST (Peter)
370 Add IS DISTINCT FROM operator (Thomas)
371 Add SQL99 TREAT() function, synonym for CAST() (Thomas)
372 Add pg_backend_pid() to output backend pid (Bruce)
373 Add IS OF / IS NOT OF type predicate (Thomas)
374 Allow bit string constants without fully-specified length (Thomas)
375 Allow conversion between 8-byte integers and bit strings (Thomas)
376 Implement hex literal conversion to bit string literal (Thomas)
377 Allow table functions to appear in the FROM clause (Joe)
378 Increase maximum number of function parameters to 32 (Bruce)
379 No longer automatically create index for SERIAL column (Tom)
380 Add current_database() (Rod)
381 Fix cash_words() to not overflow buffer (Tom)
382 Add functions replace(), split_part(), to_hex() (Joe)
383 Fix LIKE for bytea as a right-hand argument (Joe)
384 Prevent crashes caused by SELECT cash_out(2) (Tom)
385 Fix to_char(1,'FM999.99') to return a period (Karel)
386 Fix trigger/type/language functions returning OPAQUE to return
388 _________________________________________________________________
392 Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese
393 (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC)
395 Enable locale support by default (Peter)
396 Add locale variables (Peter)
397 Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea
399 Add locale awareness to regular expression character classes
400 Enable multibyte support by default (Tatso)
401 Add GB18030 multibyte support (Bill Huang)
402 Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo,
404 Add pg_conversion table (Tatsuo)
405 Add SQL99 CONVERT() function (Tatsuo)
406 pg_dumpall, pg_controldata, and pg_resetxlog now national-language
408 New and updated translations
409 _________________________________________________________________
411 Server-side Languages
413 Allow recursive SQL function (Peter)
414 Change PL/Tcl build to use configured compiler and Makefile.shlib
416 Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible
418 Allow PL/pgSQL to handle quoted identifiers (Tom)
419 Allow set-returning PL/pgSQL functions (Neil)
420 Make PL/pgSQL schema-aware (Joe)
421 Remove some memory leaks (Nigel J. Andrews, Tom)
422 _________________________________________________________________
426 Don't lowercase psql \connect database name for 7.2.0 compatibility
428 Add psql \timing to time user queries (Greg Sabino Mullane)
429 Have psql \d show index information (Greg Sabino Mullane)
430 New psql \dD shows domains (Jonathan Eisler)
431 Allow psql to show rules on views (Paul ?)
432 Fix for psql variable substitution (Tom)
433 Allow psql \d to show temporary table structure (Tom)
434 Allow psql \d to show foreign keys (Rod)
435 Fix \? to honor \pset pager (Bruce)
436 Have psql reports its version number on startup (Tom)
437 Allow \copy to specify column names (Tom)
438 _________________________________________________________________
442 Add $HOME/.pgpass to store host/user password combinations (Alvaro
444 Add PQunescapeBytea() function to libpq (Patrick Welche)
445 Fix for sending large queries over non-blocking connections
447 Fix for libpq using timers on Win9X (David Ford)
448 Allow libpq notify to handle servers with different-length
450 Add libpq PQescapeString() and PQescapeBytea() to Win32 (Bruce)
451 Fix for SSL with non-blocking connections (Jack Bates)
452 Add libpq connection timeout parameter (Denis A Ustimenko)
453 _________________________________________________________________
457 Allow JDBC to compile with JDK 1.4 (Dave)
458 Add JDBC 3 support (Barry)
459 Allows JDBC to set loglevel by adding ?loglevel=X to the connection
461 Add Driver.info() message that prints out the version number (Barry)
462 Add updateable result sets (Raghu Nidagal, Dave)
463 Add support for callable statements (Paul Bethe)
464 Add query cancel capability
465 Add refresh row (Dave)
466 Fix MD5 encryption handling for multibyte servers (Jun Kawai)
467 Add support for prepared statements (Barry)
468 _________________________________________________________________
470 Miscellaneous Interfaces
472 Fixed ECPG bug concerning octal numbers in single quotes (Michael)
473 Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc,
475 Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)
476 Add libpgtcl connection close event (Gerhard Hintermayer)
477 Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc,
479 Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)
480 Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc,
482 Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc,
484 Remove src/bin/pgaccess from main tree, now at
485 http://www.pgaccess.org (Bruce)
486 Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer,
488 _________________________________________________________________
492 Fix for parallel make (Peter)
493 AIX fixes for linking Tcl (Andreas Zeugswetter)
494 Allow PL/Perl to build under Cygwin (Jason Tishler)
495 Improve MIPS compiles (Peter, Oliver Elphick)
496 Require Autoconf version 2.53 (Peter)
497 Require readline and zlib by default in configure (Peter)
498 Allow Solaris to use Intimate Shared Memory (ISM), for performance
499 (Scott Brunza, P.J. Josh Rovero)
500 Always enable syslog in compile, remove --enable-syslog option
502 Always enable multibyte in compile, remove --enable-multibyte option
504 Always enable locale in compile, remove --enable-locale option
506 Fix for Win9x DLL creation (Magnus Naeslund)
507 Fix for link() usage by WAL code on Win32, BeOS (Jason Tishler)
508 Add sys/types.h to c.h, remove from main files (Peter, Bruce)
509 Fix AIX hang on SMP machines (Tomoyuki Niijima)
510 AIX SMP hang fix (Tomoyuki Niijima)
511 Fix pre-1970 date handling on newer glibc libraries (Tom)
512 Fix PowerPC SMP locking (Tom)
513 Prevent gcc -ffast-math from being used (Peter, Tom)
514 Bison >= 1.50 now required for developer builds
515 Kerberos 5 support now builds with Heimdal (Peter)
516 Add appendix in the User's Guide which lists SQL features (Thomas)
517 Improve loadable module linking to use RTLD_NOW (Tom)
518 New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)
519 New src/port directory holds replaced libc functions (Peter, Bruce)
520 New pg_namespace system catalog for schemas (Tom)
521 Add pg_class.relnamespace for schemas (Tom)
522 Add pg_type.typnamespace for schemas (Tom)
523 Add pg_proc.pronamespace for schemas (Tom)
524 Restructure aggregates to have pg_proc entries (Tom)
525 System relations now have their own namespace, pg_* test not required
527 Rename TOAST index names to be *_index rather than *_idx (Neil)
528 Add namespaces for operators, opclasses (Tom)
529 Add additional checks to server control file (Thomas)
530 New Polish FAQ (Marcin Mazurek)
531 Add Posix semaphore support (Tom)
532 Document need for reindex (Bruce)
533 Rename some internal identifiers to simplify Win32 compile (Jan,
535 Add documentation on computing disk space (Bruce)
536 Remove KSQO from GUC (Bruce)
537 Fix memory leak in rtree (Kenneth Been)
538 Modify a few error messages for consistency (Bruce)
539 Remove unused system table columns (Peter)
540 Make system columns NOT NULL where appropriate (Tom)
541 Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)
542 Remove OPAQUE and create specific subtypes (Tom)
543 Cleanups in array internal handling (Joe, Tom)
544 Disallow pg_atoi('') (Bruce)
545 Remove parameter wal_files because WAL files are now recycled (Bruce)
546 Add version numbers to heap pages (Tom)
547 _________________________________________________________________
551 Allow inet arrays in /contrib/array (Neil)
552 Gist fixes (Teodor Sigaev, Neil)
553 Upgrade /contrib/mysql
554 Add /contrib/dbsize which shows table sizes without vacuum (Peter)
555 Add /contrib/intagg, integer aggregator routines (mlw)
556 Improve /contrib/oid2name (Neil, Bruce)
557 Improve /contrib/tsearch (Oleg, Teodor Sigaev)
558 Cleanups of /contrib/rserver (Alexey V. Borzov)
559 Update /contrib/oracle conversion utility (Gilles Darold)
560 Update /contrib/dblink (Joe)
561 Improve options supported by /contrib/vacuumlo (Mario Weilguni)
562 Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey
564 Add /contrib/reindexdb utility (Shaun Thomas)
565 Add indexing to /contrib/isbn_issn (Dan Weston)
566 Add /contrib/dbmirror (Steven Singer)
567 Improve /contrib/pgbench (Neil)
568 Add /contrib/tablefunc table function examples (Joe)
569 Add /contrib/ltree data type for tree structures (Teodor Sigaev,
571 Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)
572 Fixes to /contrib/cube (Bruno Wolff)
573 Improve /contrib/fulltextindex (Christopher)
574 _________________________________________________________________
578 Release date: 2002-10-01
580 This has a variety of fixes from 7.2.2, including fixes to prevent
582 _________________________________________________________________
584 Migration to version 7.2.3
586 A dump/restore is *not* required for those running 7.2.X.
587 _________________________________________________________________
591 Prevent possible compressed transaction log loss (Tom)
592 Prevent non-superuser from increasing most recent vacuum info (Tom)
593 Handle pre-1970 date values in newer versions of glibc (Tom)
594 Fix possible hang during server shutdown
595 Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)
596 Fix pg_dump to properly dump FULL JOIN USING (Tom)
597 _________________________________________________________________
601 Release date: 2002-08-23
603 This has a variety of fixes from 7.2.1.
604 _________________________________________________________________
606 Migration to version 7.2.2
608 A dump/restore is *not* required for those running 7.2.X.
609 _________________________________________________________________
613 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
614 Fix for compressed transaction log id wraparound (Tom)
615 Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Ta
617 Fix for psql and pg_dump crashing when invoked with non-existand long
619 Fix crash when invoking geometric operators (Tom)
620 Allow OPEN cursor(args) (Tom)
621 Fix for rtree_gist index build (Teodor)
622 Fix for dumping user-defined aggregates (Tom)
623 Contrib/intarray fixes (Oleg)
624 Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)
625 Fix to pg_convert (Tatsuo)
626 Fix for crash with long DATA strings (Thomes, Neil)
627 Fix for repeat(), lpad(), rpad() and long strings (Neil)
628 _________________________________________________________________
632 Release date: 2002-03-21
634 This has a variety of fixes from 7.2.
635 _________________________________________________________________
637 Migration to version 7.2.1
639 A dump/restore is *not* required for those running 7.2.
640 _________________________________________________________________
644 Ensure that sequence counters do not go backwards after a crash (Tom)
645 Fix pgaccess kanji-coversion key binding (Tatsuo)
646 Optimizer improvements (Tom)
647 Cash I/O improvements (Tom)
649 Compile fix for missing AuthBlockSig (Heiko)
650 Additional time zones and time zone fixes (Thomas)
651 Allow psql \connect to handle mixed case database and user names (Tom)
652 Return proper OID on command completion even with ON INSERT rules (Tom
654 Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)
655 Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)
656 Improve handling of multiple UNIONs with different lengths (Tom)
657 contrib/btree_gist improvements (Teodor Sigaev)
658 contrib/tsearch dictionary improvements, see README.tsearch for
659 an additional installation step (Thomas T. Thai, Teodor Sigaev)
660 Fix for array subscripts handling (Tom)
661 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
662 _________________________________________________________________
666 Release date: 2002-02-04
667 _________________________________________________________________
671 This release improves PostgreSQL for use in high-volume applications.
673 Major changes in this release:
676 Vacuuming no longer locks tables, thus allowing normal user
677 access during the vacuum. A new "VACUUM FULL" command does
678 old-style vacuum by locking the table and shrinking the on-disk
682 There is no longer a problem with installations that exceed
683 four billion transactions.
686 OIDs are now optional. Users can now create tables without OIDs
687 for cases where OID usage is excessive.
690 The system now computes histogram column statistics during
691 "ANALYZE", allowing much better optimizer choices.
694 A new MD5 encryption option allows more secure storage and
695 transfer of passwords. A new Unix-domain socket authentication
696 option is available on Linux and BSD systems.
699 Administrators can use the new table access statistics module
700 to get fine-grained information about table and index usage.
703 Program and library messages can now be displayed in several
705 _________________________________________________________________
707 Migration to version 7.2
709 A dump/restore using "pg_dump" is required for those wishing to
710 migrate data from any previous release.
712 Observe the following incompatibilities:
714 * The semantics of the "VACUUM" command have changed in this
715 release. You may wish to update your maintenance procedures
717 * In this release, comparisons using = NULL will always return false
718 (or NULL, more precisely). Previous releases automatically
719 transformed this syntax to IS NULL. The old behavior can be
720 re-enabled using a "postgresql.conf" parameter.
721 * The "pg_hba.conf" and "pg_ident.conf" configuration is now only
722 reloaded after receiving a SIGHUP signal, not with each
724 * The function "octet_length()" now returns the uncompressed data
726 * The date/time value 'current' is no longer available. You will
727 need to rewrite your applications.
728 * The timestamp(), time(), and interval() functions are no longer
729 available. Instead of timestamp(), use timestamp 'string' or CAST.
731 The SELECT ... LIMIT #,# syntax will be removed in the next release.
732 You should change your queries to use separate LIMIT and OFFSET
733 clauses, e.g. LIMIT 10 OFFSET 20.
734 _________________________________________________________________
740 Create temporary files in a separate directory (Bruce)
741 Delete orphaned temporary files on postmaster startup (Bruce)
742 Added unique indexes to some system tables (Tom)
743 System table operator reorganization (Oleg Bartunov, Teodor Sigaev, To
745 Renamed pg_log to pg_clog (Tom)
746 Enable SIGTERM, SIGQUIT to kill backends (Jan)
747 Removed compile-time limit on number of backends (Tom)
748 Better cleanup for semaphore resource failure (Tatsuo, Tom)
749 Allow safe transaction ID wraparound (Tom)
750 Removed OIDs from some system tables (Tom)
751 Removed "triggered data change violation" error check (Tom)
752 SPI portal creation of prepared/saved plans (Jan)
753 Allow SPI column functions to work for system columns (Tom)
754 Long value compression improvement (Tom)
755 Statistics collector for table, index access (Jan)
756 Truncate extra-long sequence names to a reasonable value (Tom)
757 Measure transaction times in milliseconds (Thomas)
758 Fix TID sequential scans (Hiroshi)
759 Superuser ID now fixed at 1 (Peter E)
760 New pg_ctl "reload" option (Tom)
761 _________________________________________________________________
765 Optimizer improvements (Tom)
766 New histogram column statistics for optimizer (Tom)
767 Reuse write-ahead log files rather than discarding them (Tom)
768 Cache improvements (Tom)
769 IS NULL, IS NOT NULL optimizer improvement (Tom)
770 Improve lock manager to reduce lock contention (Tom)
771 Keep relcache entries for index access support functions (Tom)
772 Allow better selectivity with NaN and infinities in NUMERIC (Tom)
773 R-tree performance improvements (Kenneth Been)
774 B-tree splits more efficient (Tom)
775 _________________________________________________________________
779 Change UPDATE, DELETE permissions to be distinct (Peter E)
780 New REFERENCES, TRIGGER privileges (Peter E)
781 Allow GRANT/REVOKE to/from more than one user at a time (Peter E)
782 New has_table_privilege() function (Joe Conway)
783 Allow non-superuser to vacuum database (Tom)
784 New SET SESSION AUTHORIZATION command (Peter E)
785 Fix bug in privilege modifications on newly created tables (Tom)
786 Disallow access to pg_statistic for non-superuser, add user-accessible
788 _________________________________________________________________
790 Client Authentication
792 Fork postmaster before doing authentication to prevent hangs (Peter E)
793 Add ident authentication over Unix domain sockets on Linux, *BSD (Helg
794 e Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)
795 Add a password authentication method that uses MD5 encryption (Bruce)
796 Allow encryption of stored passwords using MD5 (Bruce)
797 PAM authentication (Dominic J. Eidson)
798 Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)
799 _________________________________________________________________
803 Interpretation of some time zone abbreviations as Australian rather th
804 an North American now settable at run time (Bruce)
805 New parameter to set default transaction isolation level (Peter E)
806 New parameter to enable conversion of "expr = NULL" into "expr IS NULL
807 ", off by default (Peter E)
808 New parameter to control memory usage by VACUUM (Tom)
809 New parameter to set client authentication timeout (Tom)
810 New parameter to set maximum number of open files (Tom)
811 _________________________________________________________________
815 Statements added by INSERT rules now execute after the INSERT (Jan)
816 Prevent unadorned relation names in target list (Bruce)
817 NULLs now sort after all normal values in ORDER BY (Tom)
818 New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)
819 New SHARE UPDATE EXCLUSIVE lock mode (Tom)
820 New EXPLAIN ANALYZE command that shows run times and row counts (Marti
822 Fix problem with LIMIT and subqueries (Tom)
823 Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)
824 Fix nested EXCEPT/INTERSECT (Tom)
825 _________________________________________________________________
829 Fix SERIAL in temporary tables (Bruce)
830 Allow temporary sequences (Bruce)
831 Sequences now use int8 internally (Tom)
832 New SERIAL8 creates int8 columns with sequences, default still SERIAL4
834 Make OIDs optional using WITHOUT OIDS (Tom)
835 Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)
836 Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher K
838 New CREATE OR REPLACE FUNCTION to alter existing function (preserving
839 the function OID) (Gavin Sherry)
840 Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)
841 Allow column renaming in views
842 Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent
844 Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (
846 ALTER TABLE RENAME update foreign-key trigger arguments correctly (Bre
848 DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)
849 Add automatic return type data casting for SQL functions (Tom)
850 Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov
851 , Teodor Sigaev, Tom)
852 Enable partial indexes (Martijn van Oosterhout)
853 _________________________________________________________________
857 Add RESET ALL, SHOW ALL (Marko Kreen)
858 CREATE/ALTER USER/GROUP now allow options in any order (Vince)
859 Add LOCK A, B, C functionality (Neil Padgett)
860 New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)
861 New light-weight VACUUM does not lock table; old semantics are availab
862 le as VACUUM FULL (Tom)
863 Disable COPY TO/FROM on views (Bruce)
864 COPY DELIMITERS string must be exactly one character (Tom)
865 VACUUM warning about index tuples fewer than heap now only appears whe
866 n appropriate (Martijn van Oosterhout)
867 Fix permission checks for CREATE INDEX (Tom)
868 Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)
869 _________________________________________________________________
871 Data Types and Functions
873 SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)
874 Add convert(), convert2() (Tatsuo)
875 New function bit_length() (Peter E)
876 Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tats
878 CHAR(), VARCHAR() now reject strings that are too long (Peter E)
879 BIT VARYING now rejects bit strings that are too long (Peter E)
880 BIT now rejects bit strings that do not match declared size (Peter E)
881 INET, CIDR text conversion functions (Alex Pilosov)
882 INET, CIDR operators << and <<= indexable (Alex Pilosov)
883 Bytea \### now requires valid three digit octal number
884 Bytea comparison improvements, now supports =, <>, >, >=, <, and <=
885 Bytea now supports B-tree indexes
886 Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE
887 Bytea now supports concatenation
888 New bytea functions: position, substring, trim, btrim, and length
889 New encode() function mode, "escaped", converts minimally escaped byte
891 Add pg_database_encoding_max_length() (Tatsuo)
892 Add pg_client_encoding() function (Tatsuo)
893 now() returns time with millisecond precision (Thomas)
894 New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)
895 Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)
896 New xid/int comparison functions (Hiroshi)
897 Add precision to TIME, TIMESTAMP, and INVERVAL data types (Thomas)
898 Modify type coercion logic to attempt binary-compatible functions firs
900 New encode() function installed by default (Marko Kreen)
901 Improved to_*() conversion functions (Karel Zak)
902 Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)
903 New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_sal
905 Correct description of translate() function (Bruce)
906 Add INTERVAL argument for SET TIME ZONE (Thomas)
907 Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)
908 Optimize length functions when using single-byte encodings (Tatsuo)
909 Fix path_inter, path_distance, path_length, dist_ppath to handle close
910 d paths (Curtis Barrett, Tom)
911 octet_length(text) now returns non-compressed length (Tatsuo, Bruce)
912 Handle "July" full name in date/time literals (Greg Sabino Mullane)
913 Some datatype() function calls now evaluated differently
914 Add support for Julian and ISO time specifications (Thomas)
915 _________________________________________________________________
919 National language support in psql, pg_dump, libpq, and server (Peter E
921 Message translations in Chinese (simplified, traditional), Czech, Fren
922 ch, German, Hungarian, Russian, Swedish (Peter E, Serguei A. Mokhov, K
923 arel Zak, Weiping He, Zhenbang Wei, Kovacs Zoltan)
924 Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware
926 Add LATIN5,6,7,8,9,10 support (Tatsuo)
927 Add ISO 8859-5,6,7,8 support (Tatsuo)
928 Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)
929 Make mic2ascii() non-ASCII aware (Tatsuo)
930 Reject invalid multibyte character sequences (Tatsuo)
931 _________________________________________________________________
935 Now uses portals for SELECT loops, allowing huge result sets (Jan)
936 CURSOR and REFCURSOR support (Jan)
937 Can now return open cursors (Jan)
938 Add ELSEIF (Klaus Reger)
939 Improve PL/pgSQL error reporting, including location of error (Tom)
940 Allow IS or FOR key words in cursor declaration, for compatibility (Br
942 Fix for SELECT ... FOR UPDATE (Tom)
943 Fix for PERFORM returning multiple rows (Tom)
944 Make PL/pgSQL use the server's type coercion code (Tom)
945 Memory leak fix (Jan, Tom)
946 Make trailing semicolon optional (Tom)
947 _________________________________________________________________
951 New untrusted PL/Perl (Alex Pilosov)
952 PL/Perl is now built on some platforms even if libperl is not shared (
954 _________________________________________________________________
958 Now reports errorInfo (Vsevolod Lobko)
959 Add spi_lastoid function (bob@redivi.com)
960 _________________________________________________________________
964 ...is new (Andrew Bosma)
965 _________________________________________________________________
969 \d displays indexes in unique, primary groupings (Christopher Kings-Ly
971 Allow trailing semicolons in backslash commands (Greg Sabino Mullane)
972 Read password from /dev/tty if possible
973 Force new password prompt when changing user and database (Tatsuo, Tom
975 Format the correct number of columns for Unicode (Patrice)
976 _________________________________________________________________
980 New function PQescapeString() to escape quotes in command strings (Flo
982 New function PQescapeBytea() escapes binary strings for use as SQL str
984 _________________________________________________________________
988 Return OID of INSERT (Ken K)
989 Handle more data types (Ken K)
990 Handle single quotes and newlines in strings (Ken K)
991 Handle NULL variables (Ken K)
992 Fix for time zone handling (Barry Lind)
993 Improved Druid support
994 Allow eight-bit characters with non-multibyte server (Barry Lind)
995 Support BIT, BINARY types (Ned Wolpert)
996 Reduce memory usage (Michael Stephens, Dave Cramer)
997 Update DatabaseMetaData (Peter E)
998 Add DatabaseMetaData.getCatalogs() (Peter E)
999 Encoding fixes (Anders Bengtsson)
1000 Get/setCatalog methods (Jason Davies)
1001 DatabaseMetaData.getColumns() now returns column defaults (Jason Davie
1003 DatabaseMetaData.getColumns() performance improvement (Jeroen van Vian
1005 Some JDBC1 and JDBC2 merging (Anders Bengtsson)
1006 Transaction performance improvements (Barry Lind)
1007 Array fixes (Greg Zoller)
1009 Fix batch processing (Rene Pijlman)
1010 ExecSQL method reorganization (Anders Bengtsson)
1011 GetColumn() fixes (Jeroen van Vianen)
1012 Fix isWriteable() function (Rene Pijlman)
1013 Improved passage of JDBC2 conformance tests (Rene Pijlman)
1014 Add bytea type capability (Barry Lind)
1015 Add isNullable() (Rene Pijlman)
1016 JDBC date/time test suite fixes (Liam Stewart)
1017 Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)
1018 Fix DatabaseMetaData to show precision properly (Mark Lillywhite)
1019 New getImported/getExported keys (Jason Davies)
1020 MD5 password encryption support (Jeremy Wohl)
1021 Fix to actually use type cache (Ned Wolpert)
1022 _________________________________________________________________
1026 Remove query size limit (Hiroshi)
1027 Remove text field size limit (Hiroshi)
1028 Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)
1029 Allow ODBC procedure calls (Hiroshi)
1030 Improve boolean handing (Aidan Mountford)
1031 Most configuration options on setable via DSN (Hiroshi)
1032 Multibyte, performance fixes (Hiroshi)
1033 Allow driver to be used with iODBC or unixODBC (Peter E)
1034 MD5 password encryption support (Bruce)
1035 Add more compatibility functions to odbc.sql (Peter E)
1036 _________________________________________________________________
1040 EXECUTE ... INTO implemented (Christof Petig)
1041 Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)
1042 Fix for GRANT parameters (Lee Kindness)
1043 Fix INITIALLY DEFERRED bug
1044 Various bug fixes (Michael, Christof Petig)
1045 Auto allocation for indicator variable arrays (int *ind_p=NULL)
1046 Auto allocation for string arrays (char **foo_pp=NULL)
1047 ECPGfree_auto_mem fixed
1048 All function names with external linkage are now prefixed by ECPG
1049 Fixes for arrays of structures (Michael)
1050 _________________________________________________________________
1054 Python fix fetchone() (Gerhard Haring)
1055 Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Ma
1057 Add Tcl COPY TO/FROM (ljb)
1058 Prevent output of default index op class in pg_dump (Tom)
1059 Fix libpgeasy memory leak (Bruce)
1060 _________________________________________________________________
1064 Configure, dynamic loader, and shared library fixes (Peter E)
1065 Fixes in QNX 4 port (Bernd Tegge)
1066 Fixes in Cygwin and Win32 ports (Jason Tishler, Gerhard Haring, Dmitry
1067 Yurtaev, Darko Prenosil, Mikhail Terekhov)
1068 Fix for Win32 socket communication failures (Magnus, Mikhail Terekhov)
1069 Hurd compile fix (Oliver Elphick)
1070 BeOS fixes (Cyril Velter)
1071 Remove configure --enable-unicode-conversion, now enabled by multibyte
1073 AIX fixes (Tatsuo, Andreas)
1074 Fix parallel make (Peter E)
1075 Install SQL language manual pages into OS-specific directories (Peter
1077 Rename config.h to pg_config.h (Peter E)
1078 Reorganize installation layout of header files (Peter E)
1079 _________________________________________________________________
1083 Remove SEP_CHAR (Bruce)
1085 Merge GUC and command line handling (Marko Kreen)
1086 Remove EXTEND INDEX (Martijn van Oosterhout, Tom)
1087 New pgjindent utility to indent java code (Bruce)
1088 Remove define of true/false when compiling under C++ (Leandro Fanzone,
1090 pgindent fixes (Bruce, Tom)
1091 Replace strcasecmp() with strcmp() where appropriate (Peter E)
1092 Dynahash portability improvements (Tom)
1093 Add 'volatile' usage in spinlock structures
1094 Improve signal handling logic (Tom)
1095 _________________________________________________________________
1099 New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)
1100 New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)
1101 Add contrib/dblink for remote database access (Joe Conway)
1102 contrib/ora2pg Oracle conversion utility (Gilles Darold)
1103 contrib/xml XML conversion utility (John Gray)
1104 contrib/fulltextindex fixes (Christopher Kings-Lynne)
1105 New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merg
1107 Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartu
1109 New pg_upgrade utility (Bruce)
1110 Add new pg_resetxlog options (Bruce, Tom)
1111 _________________________________________________________________
1115 Release date: 2001-08-15
1116 _________________________________________________________________
1118 Migration to version 7.1.3
1120 A dump/restore is *not* required for those running 7.1.X.
1121 _________________________________________________________________
1125 Remove unused WAL segements of large transactions (Tom)
1126 Multiaction rule fix (Tom)
1127 PL/pgSQL memory allocation fix (Jan)
1128 VACUUM buffer fix (Tom)
1129 Regression test fixes (Tom)
1130 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
1131 Fix subselects with DISTINCT ON or LIMIT (Tom)
1133 Disable COPY TO/FROM a view (Tom)
1134 Cygwin build (Jason Tishler)
1136 _________________________________________________________________
1140 Release date: 2001-05-11
1142 This has one fix from 7.1.1.
1143 _________________________________________________________________
1145 Migration to version 7.1.2
1147 A dump/restore is *not* required for those running 7.1.X.
1148 _________________________________________________________________
1152 Fix PL/pgSQL SELECTs when returning no rows
1153 Fix for psql backslash core dump
1154 Referential integrity permission fix
1158 _________________________________________________________________
1162 Release date: 2001-05-05
1164 This has a variety of fixes from 7.1.
1165 _________________________________________________________________
1167 Migration to version 7.1.1
1169 A dump/restore is *not* required for those running 7.1.
1170 _________________________________________________________________
1174 Fix for numeric MODULO operator (Tom)
1175 pg_dump fixes (Philip)
1176 pg_dump can dump 7.0 databases (Philip)
1177 readline 4.2 fixes (Peter E)
1179 AIX, MSWIN, VAX, N32K fixes (Tom)
1180 Multibytes fixes (Tom)
1181 Unicode fixes (Tatsuo)
1182 Optimizer improvements (Tom)
1183 Fix for whole tuples in functions (Tom)
1184 Fix for pg_ctl and option strings with spaces (Peter E)
1185 ODBC fixes (Hiroshi)
1186 EXTRACT can now take string argument (Thomas)
1187 Python fixes (Darcy)
1189 _________________________________________________________________
1193 Release date: 2001-04-13
1195 This release focuses on removing limitations that have existed in the
1196 PostgreSQL code for many years.
1198 Major changes in this release:
1200 Write-ahead Log (WAL)
1201 To maintain database consistency in case of an operating system
1202 crash, previous releases of PostgreSQL have forced all data
1203 modifications to disk before each transaction commit. With WAL,
1204 only one log file must be flushed to disk, greatly improving
1205 performance. If you have been using -F in previous releases to
1206 disable disk flushes, you may want to consider discontinuing
1210 TOAST - Previous releases had a compiled-in row length limit,
1211 typically 8k - 32k. This limit made storage of long text fields
1212 difficult. With TOAST, long rows of any length can be stored
1213 with good performance.
1216 We now support outer joins. The UNION/NOT IN workaround for
1217 outer joins is no longer required. We use the SQL92 outer join
1221 The previous C function manager did not handle null values
1222 properly, nor did it support 64-bit CPU's (Alpha). The new
1223 function manager does. You can continue using your old custom
1224 functions, but you may want to rewrite them in the future to
1225 use the new function manager call interface.
1228 A large number of complex queries that were unsupported in
1229 previous releases now work. Many combinations of views,
1230 aggregates, UNION, LIMIT, cursors, subqueries, and inherited
1231 tables now work properly. Inherited tables are now accessed by
1232 default. Subqueries in FROM are now supported.
1233 _________________________________________________________________
1235 Migration to version 7.1
1237 A dump/restore using pg_dump is required for those wishing to migrate
1238 data from any previous release.
1239 _________________________________________________________________
1245 Many multibyte/Unicode/locale fixes (Tatsuo and others)
1246 More reliable ALTER TABLE RENAME (Tom)
1247 Kerberos V fixes (David Wragg)
1248 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
1249 Prompt username/password on standard error (Bruce)
1250 Large objects inv_read/inv_write fixes (Tom)
1251 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
1253 Prevent query expressions from leaking memory (Tom)
1254 Allow UPDATE of arrays elements (Tom)
1255 Wake up lock waiters during cancel (Hiroshi)
1256 Fix rare cursor crash when using hash join (Tom)
1257 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
1258 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
1259 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
1260 Fix PL/perl (Alex Kapranoff)
1261 Disallow LOCK on views (Mark Hollomon)
1262 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
1263 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
1264 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
1265 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
1266 Fix rare failure with TRUNCATE command (Tom)
1267 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
1268 DISTINCT, ORDER BY, SELECT...INTO (Tom)
1269 Fix parser failures during aborted transactions (Tom)
1270 Allow temporary relations to properly clean up indexes (Bruce)
1271 Fix VACUUM problem with moving rows in same page (Tom)
1272 Modify pg_dump to better handle user-defined items in template1 (Philip)
1273 Allow LIMIT in VIEW (Tom)
1274 Require cursor FETCH to honor LIMIT (Tom)
1275 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
1276 Allow ORDER BY, LIMIT in sub-selects (Tom)
1277 Allow UNION in CREATE RULE (Tom)
1278 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
1279 Store initdb collation in pg_control so collation cannot be changed (Tom)
1280 Fix INSERT...SELECT with rules (Tom)
1281 Fix FOR UPDATE inside views and subselects (Tom)
1282 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
1283 Fix lpad() and rpad() to handle length less than input string (Tom)
1284 Fix use of NOTIFY in some rules (Tom)
1285 Overhaul btree code (Tom)
1286 Fix NOT NULL use in Pl/pgSQL variables (Tom)
1287 Overhaul GIST code (Oleg)
1288 Fix CLUSTER to preserve constraints and column default (Tom)
1289 Improved deadlock detection handling (Tom)
1290 Allow multiple SERIAL columns in a table (Tom)
1291 Prevent occasional index corruption (Vadim)
1295 Add OUTER JOINs (Tom)
1296 Function manager overhaul (Tom)
1297 Allow ALTER TABLE RENAME on indexes (Tom)
1298 Improve CLUSTER (Tom)
1299 Improve ps status display for more platforms (Peter E, Marc)
1300 Improve CREATE FUNCTION failure message (Ross)
1301 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
1303 Grand Unified Configuration scheme/GUC. Many options can now be set in
1304 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
1305 Improved handling of file descriptor cache (Tom)
1306 New warning code about auto-created table alias entries (Bruce)
1307 Overhaul initdb process (Tom, Peter E)
1308 Overhaul of inherited tables; inherited tables now accessed by default;
1309 new ONLY keyword prevents it (Chris Bitmead, Tom)
1310 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
1312 Allow renaming of temp tables (Tom)
1313 Overhaul memory manager contexts (Tom)
1314 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
1315 Overhaul pg_dump (Philip Warner)
1316 Allow pg_hba.conf secondary password file to specify only username (Peter E)
1317 Allow TEMPORARY or TEMP keyword when creating temporary tables (Bruce)
1318 New memory leak checker (Karel)
1319 New SET SESSION CHARACTERISTICS (Thomas)
1320 Allow nested block comments (Thomas)
1321 Add WITHOUT TIME ZONE type qualifier (Thomas)
1322 New ALTER TABLE ADD CONSTRAINT (Stephan)
1323 Use NUMERIC accumulators for INTEGER aggregates (Tom)
1324 Overhaul aggregate code (Tom)
1325 New VARIANCE and STDDEV() aggregates
1326 Improve dependency ordering of pg_dump (Philip)
1327 New pg_restore command (Philip)
1328 New pg_dump tar output option (Philip)
1329 New pg_dump of large objects (Philip)
1330 New ESCAPE option to LIKE (Thomas)
1331 New case-insensitive LIKE - ILIKE (Thomas)
1332 Allow functional indexes to use binary-compatible type (Tom)
1333 Allow SQL functions to be used in more contexts (Tom)
1334 New pg_config utility (Peter E)
1335 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
1337 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
1338 New quote_identifiers() and quote_literal() functions (Jan)
1339 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
1340 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
1341 Update PyGreSQL to version 3.1 (D'Arcy)
1342 Store tables as files named by OID (Vadim)
1343 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
1344 Require DROP VIEW to remove views, no DROP TABLE (Mark)
1345 Allow DROP VIEW view1, view2 (Mark)
1346 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
1347 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
1348 New /contrib/pgcrypto hashing functions (Marko Kreen)
1349 New pg_dumpall --globals-only option (Peter E)
1350 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
1351 New AT TIME ZONE syntax (Thomas)
1352 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
1353 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
1354 Allow socket path name to be specified in hostname by using leading slash
1355 (David J. MacKenzie)
1356 Allow CREATE DATABASE to specify template database (Tom)
1357 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
1358 New /contrib/rserv replication toolkit (Vadim)
1359 New file format for COPY BINARY (Tom)
1360 New /contrib/oid2name to map numeric files to table names (B Palmer)
1361 New "idle in transaction" ps status message (Marc)
1362 Update to pgaccess 0.98.7 (Constantin Teodorescu)
1363 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
1364 Add rudimentary dependency checking to pg_dump (Philip)
1368 Fix INET/CIDR type ordering and add new functions (Tom)
1369 Make OID behave as an unsigned type (Tom)
1370 Allow BIGINT as synonym for INT8 (Peter E)
1371 New int2 and int8 comparison operators (Tom)
1372 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
1373 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
1374 New GIST seg/cube examples (Gene Selkov)
1375 Improved round(numeric) handling (Tom)
1376 Fix CIDR output formatting (Tom)
1377 New CIDR abbrev() function (Tom)
1381 Write-Ahead Log (WAL) to provide crash recovery with less performance
1383 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
1384 Reduced file seeks (Denis Perchine)
1385 Improve BTREE code for duplicate keys (Tom)
1386 Store all large objects in a single table (Denis Perchine, Tom)
1387 Improve memory allocation performance (Karel, Tom)
1391 New function manager call conventions (Tom)
1392 SGI portability fixes (David Kaelbling)
1393 New configure --enable-syslog option (Peter E)
1394 New BSDI README (Bruce)
1395 configure script moved to top level, not /src (Peter E)
1396 Makefile/configuration/compilation overhaul (Peter E)
1397 New configure --with-python option (Peter E)
1398 Solaris cleanups (Peter E)
1399 Overhaul /contrib Makefiles (Karel)
1400 New OpenSSL configuration option (Magnus, Peter E)
1402 QNX fixes (Maurizio)
1403 New heap_open(), heap_openr() API (Tom)
1404 Remove colon and semi-colon operators (Thomas)
1405 New pg_class.relkind value for views (Mark Hollomon)
1406 Rename ichar() to chr() (Karel)
1407 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
1408 Fixes for NT/Cygwin (Pete Forman)
1409 AIX port fixes (Andreas)
1410 New BeOS port (David Reid, Cyril Velter)
1411 Add proofreader's changes to docs (Addison-Wesley, Bruce)
1412 New Alpha spinlock code (Adriaan Joubert, Compaq)
1413 UnixWare port overhaul (Peter E)
1414 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
1415 New FreeBSD Alpha port (Alfred)
1416 Overhaul shared memory segments (Tom)
1417 Add IBM S/390 support (Neale Ferguson)
1418 Moved macmanuf to /contrib (Larry Rosenman)
1419 Syslog improvements (Larry Rosenman)
1420 New template0 database that contains no user additions (Tom)
1421 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
1422 Allow NetBSD's libedit instead of readline (Peter)
1423 Improved assembly language source code format (Bruce)
1424 New contrib/pg_logger
1425 New --template option to createdb
1426 New contrib/pg_control utility (Oliver)
1427 New FreeBSD tools ipc_check, start-scripts/freebsd
1429 _________________________________________________________________
1433 Release date: 2000-11-11
1435 This has a variety of fixes from 7.0.2.
1436 _________________________________________________________________
1438 Migration to version 7.0.3
1440 A dump/restore is *not* required for those running 7.0.*.
1441 _________________________________________________________________
1446 Large object fix (Tom)
1447 Fix lean in COPY WITH OIDS leak (Tom)
1448 Fix backwards-index-scan (Tom)
1449 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
1450 Add --enable-syslog to configure (Marc)
1451 Fix abort transaction at backend exit in rare cases (Tom)
1452 Fix for psql \l+ when multibyte enabled (Tatsuo)
1453 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
1454 Make vacuum always flush buffers (Tom)
1455 Fix to allow cancel while waiting for a lock (Hiroshi)
1456 Fix for memory aloocation problem in user authentication code (Tom)
1457 Remove bogus use of int4out() (Tom)
1458 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
1459 Fix for failure of triggers on heap open in certain cases (Jeroen van
1461 Fix for erroneous selectivity of not-equals (Tom)
1462 Fix for erroneous use of strcmp() (Tom)
1463 Fix for bug where storage manager accesses items beyond end of file
1465 Fix to include kernel errno message in all smgr elog messages (Tom)
1466 Fix for '.' not in PATH at build time (SL Baur)
1467 Fix for out-of-file-descriptors error (Tom)
1468 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
1469 Fix for subselect in targetlist of Append node (Tom)
1470 Fix for mergejoin plans (Tom)
1471 Fix TRUNCATE failure on relations with indexes (Tom)
1472 Avoid database-wide restart on write error (Hiroshi)
1473 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
1474 Fix VACUUM problem with moving chain of update tuples when source and
1475 destination of a tuple lie on the same page (Tom)
1476 Fix user.c CommandCounterIncrement (Tom)
1477 Fix for AM/PM boundary problem in to_char() (Karel Zak)
1478 Fix TIME aggregate handling (Tom)
1479 Fix to_char() to avoid coredump on NULL input (Tom)
1481 Fix for inserting/copying longer multibyte strings into char() data
1483 Fix for crash of backend, on abort (Tom)
1485 _________________________________________________________________
1489 Release date: 2000-06-05
1491 This is a repackaging of 7.0.1 with added documentation.
1492 _________________________________________________________________
1494 Migration to version 7.0.2
1496 A dump/restore is *not* required for those running 7.*.
1497 _________________________________________________________________
1501 Added documentation to tarball.
1503 _________________________________________________________________
1507 Release date: 2000-06-01
1509 This is a cleanup release for 7.0.
1510 _________________________________________________________________
1512 Migration to version 7.0.1
1514 A dump/restore is *not* required for those running 7.0.
1515 _________________________________________________________________
1519 Fix many CLUSTER failures (Tom)
1520 Allow ALTER TABLE RENAME works on indexes (Tom)
1521 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
1522 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
1523 Fix the off by one errors in ResultSet from 6.5.3, and more.
1524 jdbc ResultSet fixes (Joseph Shraibman)
1525 optimizer tunings (Tom)
1526 Fix create user for pgaccess
1527 Fix for UNLISTEN failure
1528 IRIX fixes (David Kaelbling)
1529 QNX fixes (Andreas Kardos)
1530 Reduce COPY IN lock level (Tom)
1531 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
1532 Fix pg_dump to handle OID indexes (Tom)
1533 Fix small memory leak (Tom)
1534 Solaris fix for createdb/dropdb (Tatsuo)
1535 Fix for non-blocking connections (Alfred Perlstein)
1536 Fix improper recovery after RENAME TABLE failures (Tom)
1537 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
1538 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
1539 Fix too long syslog message (Tatsuo)
1540 Fix problem with quoted indexes that are too long (Tom)
1541 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
1542 ecpg changes (Michael)
1544 _________________________________________________________________
1548 Release date: 2000-05-08
1550 This release contains improvements in many areas, demonstrating the
1551 continued growth of PostgreSQL. There are more improvements and fixes
1552 in 7.0 than in any previous release. The developers have confidence
1553 that this is the best release yet; we do our best to put out only
1554 solid releases, and this one is no exception.
1556 Major changes in this release:
1559 Foreign keys are now implemented, with the exception of PARTIAL
1560 MATCH foreign keys. Many users have been asking for this
1561 feature, and we are pleased to offer it.
1564 Continuing on work started a year ago, the optimizer has been
1565 improved, allowing better query plan selection and faster
1566 performance with less memory usage.
1569 psql, our interactive terminal monitor, has been updated with a
1570 variety of new features. See the psql manual page for details.
1573 SQL92 join syntax is now supported, though only as INNER JOIN
1574 for this release. JOIN, NATURAL JOIN, JOIN/USING, and JOIN/ON
1575 are available, as are column correlation names.
1576 _________________________________________________________________
1578 Migration to version 7.0
1580 A dump/restore using pg_dump is required for those wishing to migrate
1581 data from any previous release of PostgreSQL. For those upgrading from
1582 6.5.*, you may instead use pg_upgrade to upgrade to this release;
1583 however, a full dump/reload installation is always the most robust
1584 method for upgrades.
1586 Interface and compatibility issues to consider for the new release
1589 * The date/time types datetime and timespan have been superseded by
1590 the SQL92-defined types timestamp and interval. Although there has
1591 been some effort to ease the transition by allowing PostgreSQL to
1592 recognize the deprecated type names and translate them to the new
1593 type names, this mechanism may not be completely transparent to
1594 your existing application.
1595 * The optimizer has been substantially improved in the area of query
1596 cost estimation. In some cases, this will result in decreased
1597 query times as the optimizer makes a better choice for the
1598 preferred plan. However, in a small number of cases, usually
1599 involving pathological distributions of data, your query times may
1600 go up. If you are dealing with large amounts of data, you may want
1601 to check your queries to verify performance.
1602 * The JDBC and ODBC interfaces have been upgraded and extended.
1603 * The string function CHAR_LENGTH is now a native function. Previous
1604 versions translated this into a call to LENGTH, which could result
1605 in ambiguity with other types implementing LENGTH such as the
1607 _________________________________________________________________
1613 Prevent function calls exceeding maximum number of arguments (Tom)
1614 Improve CASE construct (Tom)
1615 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
1616 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
1617 Fix GROUP BY scan bug (Tom)
1618 Improvements in SQL grammar processing (Tom)
1619 Fix for views involved in INSERT ... SELECT ... (Tom)
1620 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
1621 Fix for subselects in INSERT ... SELECT (Tom)
1622 Prevent INSERT ... SELECT ... ORDER BY (Tom)
1623 Fixes for relations greater than 2GB, including vacuum
1624 Improve propagating system table changes to other backends (Tom)
1625 Improve propagating user table changes to other backends (Tom)
1626 Fix handling of temp tables in complex situations (Bruce, Tom)
1627 Allow table locking at table open, improving concurrent reliability (Tom)
1628 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
1629 Prevent DROP DATABASE while others accessing
1630 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
1631 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
1632 Fix pg_upgrade so it works for MVCC (Tom)
1633 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
1634 Fix for "f1 datetime DEFAULT 'now'" (Tom)
1635 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
1636 Allow comment-only lines, and ;;; lines too. (Tom)
1637 Improve recovery after failed disk writes, disk full (Hiroshi)
1638 Fix cases where table is mentioned in FROM but not joined (Tom)
1639 Allow HAVING clause without aggregate functions (Tom)
1640 Fix for "--" comment and no trailing newline, as seen in perl interface
1641 Improve pg_dump failure error reports (Bruce)
1642 Allow sorts and hashes to exceed 2GB file sizes (Tom)
1643 Fix for pg_dump dumping of inherited rules (Tom)
1644 Fix for NULL handling comparisons (Tom)
1645 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
1646 Fix for dbname with dash
1647 Prevent DROP INDEX from interfering with other backends (Tom)
1648 Fix file descriptor leak in verify_password()
1649 Fix for "Unable to identify an operator =$" problem
1650 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
1651 Fix for recursive exit call (Massimo)
1652 Fix for extra-long timezones (Jeroen van Vianen)
1653 Make pg_dump preserve primary key information (Peter E)
1654 Prevent databases with single quotes (Peter E)
1655 Prevent DROP DATABASE inside transaction (Peter E)
1656 ecpg memory leak fixes (Stephen Birch)
1657 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
1658 Y2K timestamp fix (Massimo)
1659 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
1660 Fix for views with tables/columns containing spaces (Tom)
1661 Prevent permissions on indexes (Peter E)
1662 Fix for spinlock stuck problem when error is generated (Hiroshi)
1663 Fix ipcclean on Linux
1664 Fix handling of NULL constraint conditions (Tom)
1665 Fix memory leak in odbc driver (Nick Gorham)
1666 Fix for permission check on UNION tables (Tom)
1667 Fix to allow SELECT 'a' LIKE 'a' (Tom)
1668 Fix for SELECT 1 + NULL (Tom)
1670 Fix log() on numeric type (Tom)
1671 Deprecate ':' and ';' operators
1672 Allow vacuum of temporary tables
1673 Disallow inherited columns with the same name as new columns
1674 Recover or force failure when disk space is exhausted (Hiroshi)
1675 Fix INSERT INTO ... SELECT with AS columns matching result columns
1676 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns
1678 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
1679 Fix UNION with LIMIT
1680 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
1681 Fix CREATE TABLE test(col char(2) DEFAULT user)
1682 Fix mismatched types in CREATE TABLE ... DEFAULT
1683 Fix SELECT * FROM pg_class where oid in (0,-1)
1684 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
1685 Prevent user who can create databases can modifying pg_database table (Peter E)
1686 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
1687 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
1691 New CLI interface include file sqlcli.h, based on SQL3/SQL98
1692 Remove all limits on query length, row length limit still exists (Tom)
1693 Update jdbc protocol to 2.0 (Jens Glaser <jens@jens.de>)
1694 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
1695 Fix to give super user and createdb user proper update catalog rights (Peter E)
1696 Allow ecpg bool variables to have NULL values (Christof)
1697 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
1698 Allow ^C to cancel COPY command (Massimo)
1699 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
1700 Function name overloading for dynamically-loaded C functions (Frankpitt)
1701 Add CmdTuples() to libpq++(Vince)
1702 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
1703 Allow CREATE FUNCTION/WITH clause to be used for all language types
1704 configure --enable-debug adds -g (Peter E)
1705 configure --disable-debug removes -g (Peter E)
1706 Allow more complex default expressions (Tom)
1707 First real FOREIGN KEY constraint trigger functionality (Jan)
1708 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
1709 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
1710 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
1711 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
1712 Change pgeasy connectdb() parameter ordering (Bruce)
1713 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
1714 Add Oracle's COMMENT ON command (Mike Mascari <mascarim@yahoo.com>)
1715 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
1716 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
1717 Make USING in COPY optional (Bruce)
1718 Allow subselects in the target list (Tom)
1719 Allow subselects on the left side of comparison operators (Tom)
1720 New parallel regression test (Jan)
1721 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
1722 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
1723 Added psql LASTOID variable to return last inserted oid (Peter E)
1724 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
1725 Add permissions check for vacuum (Peter E)
1726 New libpq functions to allow asynchronous connections: PQconnectStart(),
1727 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
1728 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
1729 New libpq PQsetenv() function (Ewan Mellor)
1730 create/alter user extension (Peter E)
1731 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
1732 New scripts for create/drop user/db (Peter E)
1733 Major psql overhaul (Peter E)
1734 Add const to libpq interface (Peter E)
1735 New libpq function PQoidValue (Peter E)
1736 Show specific non-aggregate causing problem with GROUP BY (Tom)
1737 Make changes to pg_shadow recreate pg_pwd file (Peter E)
1738 Add aggregate(DISTINCT ...) (Tom)
1739 Allow flag to control COPY input/output of NULLs (Peter E)
1740 Make postgres user have a password by default (Peter E)
1741 Add CREATE/ALTER/DROP GROUP (Peter E)
1742 All administration scripts now support --long options (Peter E, Karel)
1743 Vacuumdb script now supports --all option (Peter E)
1744 ecpg new portable FETCH syntax
1745 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
1746 and EXEC SQL ENDIF directives
1747 Add pg_ctl script to control backend start-up (Tatsuo)
1748 Add postmaster.opts.default file to store start-up flags (Tatsuo)
1749 Allow --with-mb=SQL_ASCII
1750 Increase maximum number of index keys to 16 (Bruce)
1751 Increase maximum number of function arguments to 16 (Bruce)
1752 Allow configuration of maximum number of index keys and arguments (Bruce)
1753 Allow unprivileged users to change their passwords (Peter E)
1754 Password authentication enabled; required for new users (Peter E)
1755 Disallow dropping a user who owns a database (Peter E)
1756 Change initdb option --with-mb to --enable-multibyte
1757 Add option for initdb to prompts for superuser password (Peter E)
1758 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
1759 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
1760 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
1761 Libpq non-blocking mode (Alfred Perlstein)
1762 Improve conversion of types in casts that don't specify a length
1763 New plperl internal programming language (Mark Hollomon)
1764 Allow COPY IN to read file that do not end with a newline (Tom)
1765 Indicate when long identifiers are truncated (Tom)
1766 Allow aggregates to use type equivalency (Peter E)
1767 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
1768 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
1769 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
1770 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
1771 Add NUMERIC and int8 types to ODBC
1772 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
1773 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
1774 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
1775 Enable backward sequential scan even after reaching EOF (Hiroshi)
1776 Add btree indexing of boolean values, >= and <= (Don Baccus)
1777 Print current line number when COPY FROM fails (Massimo)
1778 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
1779 Add DEC as synonym for DECIMAL (Thomas)
1780 Add SESSION_USER as SQL92 keyword, same as CURRENT_USER (Thomas)
1781 Implement SQL92 column aliases (aka correlation names) (Thomas)
1782 Implement SQL92 join syntax (Thomas)
1783 Make INTERVAL reserved word allowed as a column identifier (Thomas)
1784 Implement REINDEX command (Hiroshi)
1785 Accept ALL in aggregate function SUM(ALL col) (Tom)
1786 Prevent GROUP BY from using column aliases (Tom)
1787 New psql \encoding option (Tatsuo)
1788 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
1789 Allow negation of a negative number in all cases
1790 Add ecpg descriptors (Christof, Michael)
1791 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
1792 Allow casts with length, like foo::char(8)
1793 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
1794 Add support for SJIS user defined characters (Tatsuo)
1795 Larger views/rules supported
1796 Make libpq's PQconndefaults() thread-safe (Tom)
1797 Disable // as comment to be ANSI conforming, should use -- (Tom)
1798 Allow column aliases on views CREATE VIEW name (collist)
1799 Fixes for views with subqueries (Tom)
1800 Allow UPDATE table SET fld = (SELECT ...) (Tom)
1801 SET command options no longer require quotes
1802 Update pgaccess to 0.98.6
1803 New SET SEED command
1804 New pg_options.sample file
1805 New SET FSYNC command (Massimo)
1806 Allow pg_descriptions when creating tables
1807 Allow pg_descriptions when creating types, columns, and functions
1808 Allow psql \copy to allow delimiters (Peter E)
1809 Allow psql to print nulls as distinct from "" [null] (Peter E)
1813 Many array fixes (Tom)
1814 Allow bare column names to be subscripted as arrays (Tom)
1815 Improve type casting of int and float constants (Tom)
1816 Cleanups for int8 inputs, range checking, and type conversion (Tom)
1817 Fix for SELECT timespan('21:11:26'::time) (Tom)
1818 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
1819 Add btree index on NUMERIC (Jan)
1820 Perl fix for large objects containing NUL characters (Douglas Thomson)
1821 ODBC fix for for large objects (free)
1822 Fix indexing of cidr data type
1823 Fix for Ethernet MAC addresses (macaddr type) comparisons
1824 Fix for date/time types when overflows happened in computations (Tom)
1825 Allow array on int8 (Peter E)
1826 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
1827 Allow NUMERIC arrays
1828 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
1829 Make char_length()/octet_length including trailing blanks (Tom)
1830 Made abstime/reltime use int4 instead of time_t (Peter E)
1831 New lztext data type for compressed text fields
1832 Revise code to handle coercion of int and float constants (Tom)
1833 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
1834 NUMERIC now accepts scientific notation (Tom)
1835 NUMERIC to int4 rounds (Tom)
1836 Convert float4/8 to NUMERIC properly (Tom)
1837 Allow type conversion with NUMERIC (Thomas)
1838 Make ISO date style (2000-02-16 09:33) the default (Thomas)
1839 Add NATIONAL CHAR [ VARYING ] (Thomas)
1840 Allow NUMERIC round and trunc to accept negative scales (Tom)
1841 New TIME WITH TIME ZONE type (Thomas)
1842 Add MAX()/MIN() on time type (Thomas)
1843 Add abs(), mod(), fac() for int8 (Thomas)
1844 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
1845 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
1846 Add exp() and ln() for NUMERIC type
1847 Rename NUMERIC power() to pow() (Thomas)
1848 Improved TRANSLATE() function (Edwin Ramirez, Tom)
1849 Allow X=-Y operators (Tom)
1850 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom
1852 Allow LOCALE to use indexes in regular expression searches (Tom)
1853 Allow creation of functional indexes to use default types
1857 Prevent exponential space consumption with many AND's and OR's (Tom)
1858 Collect attribute selectivity values for system columns (Tom)
1859 Reduce memory usage of aggregates (Tom)
1860 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
1861 Fix r-tree index optimizer selectivity (Thomas)
1862 Improve optimizer selectivity computations and functions (Tom)
1863 Optimize btree searching for cases where many equal keys exist (Tom)
1864 Enable fast LIKE index processing only if index present (Tom)
1865 Re-use free space on index pages with duplicates (Tom)
1866 Improve hash join processing (Tom)
1867 Prevent descending sort if result is already sorted(Hiroshi)
1868 Allow commuting of index scan query qualifications (Tom)
1869 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
1870 Allocate large memory requests in fix-sized chunks for performance (Tom)
1871 Fix vacuum's performance by reducing memory allocation requests (Tom)
1872 Implement constant-expression simplification (Bernard Frankpitt, Tom)
1873 Use secondary columns to be used to determine start of index scan (Hiroshi)
1874 Prevent quadruple use of disk space when doing internal sorting (Tom)
1875 Faster sorting by calling fewer functions (Tom)
1876 Create system indexes to match all system caches (Bruce, Hiroshi)
1877 Make system caches use system indexes (Bruce)
1878 Make all system indexes unique (Bruce)
1879 Improve pg_statistics management for VACUUM speed improvement (Tom)
1880 Flush backend cache less frequently (Tom, Hiroshi)
1881 COPY now reuses previous memory allocation, improving performance (Tom)
1882 Improve optimization cost estimation (Tom)
1883 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom
1885 Use DNF instead of CNF where appropriate (Tom, Taral)
1886 Further cleanup for OR-of-AND WHERE-clauses (Tom)
1887 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
1888 Smarter optimizer computations for random index page access (Tom)
1889 New SET variable to control optimizer costs (Tom)
1890 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
1891 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
1892 Major subquery speedup (Tom)
1893 Fewer fsync writes when fsync is not disabled (Tom)
1894 Improved LIKE optimizer estimates (Tom)
1895 Prevent fsync in SELECT-only queries (Vadim)
1896 Make index creation use psort code, because it is now faster (Tom)
1897 Allow creation of sort temp tables > 1 Gig
1901 Fix for linux PPC compile
1902 New generic expression-tree-walker subroutine (Tom)
1903 Change form() to varargform() to prevent portability problems
1904 Improved range checking for large integers on Alphas
1905 Clean up #include in /include directory (Bruce)
1906 Add scripts for checking includes (Bruce)
1907 Remove un-needed #include's from *.c files (Bruce)
1908 Change #include's to use <> and "" as appropriate (Bruce)
1909 Enable WIN32 compilation of libpq
1910 Alpha spinlock fix from Uncle George <gatgul@voicenet.com>
1911 Overhaul of optimizer data structures (Tom)
1912 Fix to cygipc library (Yutaka Tanida)
1913 Allow pgsql to work on newer Cygwin snapshots (Dan)
1914 New catalog version number (Tom)
1916 Rename heap_replace to heap_update
1917 Update for QNX (Dr. Andreas Kardos)
1918 New platform-specific regression handling (Tom)
1919 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
1920 Included all yacc and lex files into the distribution (Peter E.)
1921 Remove lextest, no longer needed (Peter E)
1922 Fix for libpq and psql on Win32 (Magnus)
1923 Internally change datetime and timespan into timestamp and interval (Thomas)
1924 Fix for plpgsql on BSD/OS
1925 Add SQL_ASCII test case to the regression test (Tatsuo)
1926 configure --with-mb now deprecated (Tatsuo)
1928 NetBSD fixes (Johnny C. Lam <lamj@stat.cmu.edu>)
1929 Fixes for Alpha compiles
1930 New multibyte encodings
1932 _________________________________________________________________
1936 Release date: 1999-10-13
1938 This is basically a cleanup release for 6.5.2. We have added a new
1939 PgAccess that was missing in 6.5.2, and installed an NT-specific fix.
1940 _________________________________________________________________
1942 Migration to version 6.5.3
1944 A dump/restore is *not* required for those running 6.5.*.
1945 _________________________________________________________________
1949 Updated version of pgaccess 0.98
1951 Fix dumping rules on inherited tables
1953 _________________________________________________________________
1957 Release date: 1999-09-15
1959 This is basically a cleanup release for 6.5.1. We have fixed a variety
1960 of problems reported by 6.5.1 users.
1961 _________________________________________________________________
1963 Migration to version 6.5.2
1965 A dump/restore is *not* required for those running 6.5.*.
1966 _________________________________________________________________
1970 subselect+CASE fixes(Tom)
1971 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
1972 Fixes for CASE in WHERE join clauses(Tom)
1973 Fix BTScan abort(Tom)
1974 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
1975 Improve it so that it checks for multicolumn constraints(Thomas)
1976 Fix for Win32 making problem with MB enabled(Hiroki Kataoka)
1977 Allow BSD yacc and bison to compile pl code(Bruce)
1978 Fix SET NAMES working
1980 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
1981 Reduce the total memory consumption of vacuum(Tom)
1982 Fix for timestamp(datetime)
1983 Rule deparsing bugfixes(Tom)
1984 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(To
1986 This is to re-use space on index pages freed by vacuum(Vadim)
1987 document -x for pg_dump(Bruce)
1988 Fix for unary operators in rule deparser(Tom)
1989 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
1990 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
1991 Repair logic error in LIKE: should not return LIKE_ABORT
1992 when reach end of pattern before end of text(Tom)
1993 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom
1995 Updated version of pgaccess 0.98
1997 _________________________________________________________________
2001 Release date: 1999-07-15
2003 This is basically a cleanup release for 6.5. We have fixed a variety
2004 of problems reported by 6.5 users.
2005 _________________________________________________________________
2007 Migration to version 6.5.1
2009 A dump/restore is *not* required for those running 6.5.
2010 _________________________________________________________________
2015 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
2016 Remove QUERY_LIMIT, use SELECT...LIMIT
2017 Fix for EXPLAIN on inheritance(Tom)
2018 Patch to allow vacuum on multisegment tables(Hiroshi)
2019 R-Tree optimizer selectivity fix(Tom)
2020 ACL file descriptor leak fix(Atsushi Ogawa)
2021 New expresssion subtree code(Tom)
2022 Avoid disk writes for read-only transactions(Vadim)
2023 Fix for removal of temp tables if last transaction was aborted(Bruce)
2024 Fix to prevent too large tuple from being created(Bruce)
2026 Allow port numbers 32k - 64k(Bruce)
2027 Add ^ precidence(Bruce)
2028 Rename sort files called pg_temp to pg_sorttemp(Bruce)
2029 Fix for microseconds in time values(Tom)
2030 Tutorial source cleanup
2032 Fix for sorting of NULL's in some cases(Tom)
2033 Shared library dependencies fixed (Tom)
2034 Fixed glitches affecting GROUP BY in subselects(Tom)
2035 Fix some compiler warnings (Tomoaki Nishiyama)
2036 Add Win1250 (Czech) support (Pavel Behal)
2038 _________________________________________________________________
2042 Release date: 1999-06-09
2044 This release marks a major step in the development team's mastery of
2045 the source code we inherited from Berkeley. You will see we are now
2046 easily adding major features, thanks to the increasing size and
2047 experience of our world-wide development team.
2049 Here is a brief summary of the more notable changes:
2051 Multiversion concurrency control(MVCC)
2052 This removes our old table-level locking, and replaces it with
2053 a locking system that is superior to most commercial database
2054 systems. In a traditional system, each row that is modified is
2055 locked until committed, preventing reads by other users. MVCC
2056 uses the natural multiversion nature of PostgreSQL to allow
2057 readers to continue reading consistent data during writer
2058 activity. Writers continue to use the compact pg_log
2059 transaction system. This is all performed without having to
2060 allocate a lock for every row like traditional database
2061 systems. So, basically, we no longer are restricted by simple
2062 table-level locking; we have something better than row-level
2065 Hot backups from pg_dump
2066 pg_dump takes advantage of the new MVCC features to give a
2067 consistent database dump/backup while the database stays online
2068 and available for queries.
2071 We now have a true numeric data type, with user-specified
2075 Temporary tables are guaranteed to have unique names within a
2076 database session, and are destroyed on session exit.
2079 We now have CASE, INTERSECT, and EXCEPT statement support. We
2080 have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL, SELECT
2081 ... FOR UPDATE, and an improved LOCK TABLE command.
2084 We continue to speed up PostgreSQL, thanks to the variety of
2085 talents within our team. We have sped up memory allocation,
2086 optimization, table joins, and row transfer routines.
2089 We continue to expand our port list, this time including
2090 Windows NT/ix86 and NetBSD/arm32.
2093 Most interfaces have new versions, and existing functionality
2097 New and updated material is present throughout the
2098 documentation. New FAQs have been contributed for SGI and AIX
2099 platforms. The Tutorial has introductory information on SQL
2100 from Stefan Simkovics. For the User's Guide, there are
2101 reference pages covering the postmaster and more utility
2102 programs, and a new appendix contains details on date/time
2103 behavior. The Administrator's Guide has a new chapter on
2104 troubleshooting from Tom Lane. And the Programmer's Guide has a
2105 description of query processing, also from Stefan, and details
2106 on obtaining the PostgreSQL source tree via anonymous CVS and
2108 _________________________________________________________________
2110 Migration to version 6.5
2112 A dump/restore using pg_dump is required for those wishing to migrate
2113 data from any previous release of PostgreSQL. pg_upgrade can *not* be
2114 used to upgrade to this release because the on-disk structure of the
2115 tables has changed compared to previous releases.
2117 The new Multiversion Concurrency Control (MVCC) features can give
2118 somewhat different behaviors in multiuser environments. *Read and
2119 understand the following section to ensure that your existing
2120 applications will give you the behavior you need.*
2121 _________________________________________________________________
2123 Multiversion Concurrency Control
2125 Because readers in 6.5 don't lock data, regardless of transaction
2126 isolation level, data read by one transaction can be overwritten by
2127 another. In other words, if a row is returned by "SELECT" it doesn't
2128 mean that this row really exists at the time it is returned (i.e.
2129 sometime after the statement or transaction began) nor that the row is
2130 protected from being deleted or updated by concurrent transactions
2131 before the current transaction does a commit or rollback.
2133 To ensure the actual existence of a row and protect it against
2134 concurrent updates one must use "SELECT FOR UPDATE" or an appropriate
2135 "LOCK TABLE" statement. This should be taken into account when porting
2136 applications from previous releases of PostgreSQL and other
2139 Keep the above in mind if you are using "contrib/refint.*" triggers
2140 for referential integrity. Additional techniques are required now. One
2141 way is to use "LOCK parent_table IN SHARE ROW EXCLUSIVE MODE" command
2142 if a transaction is going to update/delete a primary key and use "LOCK
2143 parent_table IN SHARE MODE" command if a transaction is going to
2144 update/insert a foreign key.
2146 Note: Note that if you run a transaction in SERIALIZABLE mode then
2147 you must execute the "LOCK" commands above before execution of any
2148 DML statement ("SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO") in the
2151 These inconveniences will disappear in the future when the ability to
2152 read dirty (uncommitted) data (regardless of isolation level) and true
2153 referential integrity will be implemented.
2154 _________________________________________________________________
2160 Fix text<->float8 and text<->float4 conversion functions(Thomas)
2161 Fix for creating tables with mixed-case constraints(Billy)
2162 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
2163 Fix bug in pg_dump -z
2164 Memory overrun cleanups(Tatsuo)
2165 Fix for lo_import crash(Tatsuo)
2166 Adjust handling of data type names to suppress double quotes(Thomas)
2167 Use type coercion for matching columns and DEFAULT(Thomas)
2168 Fix deadlock so it only checks once after one second of sleep(Bruce)
2169 Fixes for aggregates and PL/pgsql(Hiroshi)
2170 Fix for subquery crash(Vadim)
2171 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
2172 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo
2174 Fix for pg_dump -d or -D and quote special characters in INSERT
2175 Repair serious problems with dynahash(Tom)
2176 Fix INET/CIDR portability problems
2177 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
2178 Fix executor so mergejoin of different column types works(Tom)
2179 Fix for Alpha OR selectivity bug
2180 Fix OR index selectivity problem(Bruce)
2181 Fix so \d shows proper length for char()/varchar()(Ryan)
2182 Fix tutorial code(Clark)
2183 Improve destroyuser checking(Oliver)
2184 Fix for Kerberos(Rodney McDuff)
2185 Fix for dropping database while dirty buffers(Bruce)
2186 Fix so sequence nextval() can be case-sensitive(Bruce)
2188 Drop buffers before destroying database files(Bruce)
2189 Fix case where executor evaluates functions twice(Tatsuo)
2190 Allow sequence nextval actions to be case-sensitive(Bruce)
2191 Fix optimizer indexing not working for negative numbers(Bruce)
2192 Fix for memory leak in executor with fjIsNull
2193 Fix for aggregate memory leaks(Erik Riedel)
2194 Allow username containing a dash GRANT permissions
2195 Cleanup of NULL in inet types
2196 Clean up system table bugs(Tom)
2197 Fix problems of PAGER and \? command(Masaaki Sakaida)
2198 Reduce default multisegment file size limit to 1GB(Peter)
2199 Fix for dumping of CREATE OPERATOR(Tom)
2200 Fix for backward scanning of cursors(Hiroshi Inoue)
2201 Fix for COPY FROM STDIN when using \i(Tom)
2202 Fix for subselect is compared inside an expression(Jan)
2203 Fix handling of error reporting while returning rows(Tom)
2204 Fix problems with reference to array types(Tom,Jan)
2205 Prevent UPDATE SET oid(Jan)
2206 Fix pg_dump so -t option can handle case-sensitive tablenames
2207 Fixes for GROUP BY in special cases(Tom, Jan)
2208 Fix for memory leak in failed queries(Tom)
2209 DEFAULT now supports mixed-case identifiers(Tom)
2210 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
2211 Disable use of pg_dump with both -o and -d options(Bruce)
2212 Allow pg_dump to properly dump GROUP permissions(Bruce)
2213 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
2214 Fix for computations in views(Jan)
2215 Fix for aggregates on array indexes(Tom)
2216 Fix for DEFAULT handles single quotes in value requiring too many quotes
2217 Fix security problem with non-super users importing/exporting large objects(Tom
2219 Rollback of transaction that creates table cleaned up properly(Tom)
2220 Fix to allow long table and column names to generate proper serial names(Tom)
2224 Add "vacuumdb" utility
2225 Speed up libpq by allocating memory better(Tom)
2226 EXPLAIN all indexes used(Tom)
2227 Implement CASE, COALESCE, NULLIF expression(Thomas)
2228 New pg_dump table output format(Constantin)
2229 Add string min()/max() functions(Thomas)
2230 Extend new type coercion techniques to aggregates(Thomas)
2231 New moddatetime contrib(Terry)
2232 Update to pgaccess 0.96(Constantin)
2233 Add routines for single-byte "char" type(Thomas)
2234 Improved substr() function(Thomas)
2235 Improved multibyte handling(Tatsuo)
2236 Multiversion concurrency control/MVCC(Vadim)
2237 New Serialized mode(Vadim)
2238 Fix for tables over 2gigs(Peter)
2239 New SET TRANSACTION ISOLATION LEVEL(Vadim)
2240 New LOCK TABLE IN ... MODE(Vadim)
2241 Update ODBC driver(Byron)
2242 New NUMERIC data type(Jan)
2243 New SELECT FOR UPDATE(Vadim)
2244 Handle "NaN" and "Infinity" for input values(Jan)
2245 Improved date/year handling(Thomas)
2246 Improved handling of backend connections(Magnus)
2247 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
2248 New TCL_ARRAYS option(Massimo)
2249 New INTERSECT and EXCEPT(Stefan)
2250 New pg_index.indisprimary for primary key tracking(D'Arcy)
2251 New pg_dump option to allow dropping of tables before creation(Brook)
2252 Speedup of row output routines(Tom)
2253 New READ COMMITTED isolation level(Vadim)
2254 New TEMP tables/indexes(Bruce)
2255 Prevent sorting if result is already sorted(Jan)
2256 New memory allocation optimization(Jan)
2257 Allow psql to do \p\g(Bruce)
2258 Allow multiple rule actions(Jan)
2259 Added LIMIT/OFFSET functionality(Jan)
2260 Improve optimizer when joining a large number of tables(Bruce)
2261 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
2262 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
2263 Improved int8 support(Ryan Bradetich, Thomas, Tom)
2264 New routines to convert between int8 and text/varchar types(Thomas)
2265 New bushy plans, where meta-tables are joined(Bruce)
2266 Enable right-hand queries by default(Bruce)
2267 Allow reliable maximum number of backends to be set at configure time
2268 (--with-maxbackends and postmaster switch (-N backends))(Tom)
2269 GEQO default now 10 tables because of optimizer speedups(Tom)
2270 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
2271 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
2272 Allow psql \d on a view show query(Ryan)
2273 Speedup for LIKE(Bruce)
2274 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
2275 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
2276 Make % operator have precedence like /(Bruce)
2277 Add new postgres -O option to allow system table structure changes(Bruce)
2278 Update contrib/pginterface/findoidjoins script(Tom)
2279 Major speedup in vacuum of deleted rows with indexes(Vadim)
2280 Allow non-SQL functions to run different versions based on arguments(Tom)
2281 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida
2283 Add version number in start-up banners for psql(Masaaki Sakaida)
2284 New contrib/vacuumlo removes large objects not referenced(Peter)
2285 New initialization for table sizes so non-vacuumed tables perform better(Tom)
2286 Improve error messages when a connection is rejected(Tom)
2287 Support for arrays of char() and varchar() fields(Massimo)
2288 Overhaul of hash code to increase reliability and performance(Tom)
2289 Update to PyGreSQL 2.4(D'Arcy)
2290 Changed debug options so -d4 and -d5 produce different node displays(Jan)
2291 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
2292 Better optimization statistics for system table access(Tom)
2293 Better handling of non-default block sizes(Massimo)
2294 Improve GEQO optimizer memory consumption(Tom)
2295 UNION now suppports ORDER BY of columns not in target list(Jan)
2296 Major libpq++ improvements(Vince Vielhaber)
2297 pg_dump now uses -z(ACL's) as default(Bruce)
2298 backend cache, memory speedups(Tom)
2299 have pg_dump do everything in one snapshot transaction(Vadim)
2300 fix for large object memory leakage, fix for pg_dumping(Tom)
2301 INET type now respects netmask for comparisons
2302 Make VACUUM ANALYZE only use a readlock(Vadim)
2303 Allow VIEWs on UNIONS(Jan)
2304 pg_dump now can generate consistent snapshots on active databases(Vadim)
2308 Improve port matching(Tom)
2309 Portability fixes for SunOS
2310 Add NT/Win32 backend port and enable dynamic loading(Magnus and Daniel Horak)
2311 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
2312 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
2313 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
2314 Port to NetBSD/macppc(Toshimi Aoki)
2315 Fix for tcl/tk configuration(Vince)
2316 Removed CURRENT keyword for rule queries(Jan)
2317 NT dynamic loading now works(Daniel Horak)
2318 Add ARM32 support(Andrew McMurry)
2319 Better support for HP-UX 11 and UnixWare
2320 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
2321 New install commands for plpgsql(Jan)
2323 _________________________________________________________________
2327 Release date: 1998-12-20
2329 The 6.4.1 release was improperly packaged. This also has one
2331 _________________________________________________________________
2333 Migration to version 6.4.2
2335 A dump/restore is *not* required for those running 6.4.*.
2336 _________________________________________________________________
2340 Fix for datetime constant problem on some platforms(Thomas)
2341 _________________________________________________________________
2345 Release date: 1998-12-18
2347 This is basically a cleanup release for 6.4. We have fixed a variety
2348 of problems reported by 6.4 users.
2349 _________________________________________________________________
2351 Migration to version 6.4.1
2353 A dump/restore is *not* required for those running 6.4.
2354 _________________________________________________________________
2358 Add pg_dump -N flag to force double quotes around identifiers. This is
2360 Fix for NOT in where clause causing crash(Bruce)
2361 EXPLAIN VERBOSE coredump fix(Vadim)
2362 Fix shared-library problems on Linux
2363 Fix test for table existance to allow mixed-case and whitespace in
2364 the table name(Thomas)
2365 Fix a couple of pg_dump bugs
2366 Configure matches template/.similar entries better(Tom)
2367 Change builtin function names from SPI_* to spi_*
2368 OR WHERE clause fix(Vadim)
2369 Fixes for mixed-case table names(Billy)
2370 contrib/linux/postgres.init.csh/sh fix(Thomas)
2371 libpq memory overrun fix
2373 Change exp() behavior to generate error on underflow(Thomas)
2374 pg_dump fixes for memory leak, inheritance constraints, layout change
2375 update pgaccess to 0.93
2376 Fix prototype for 64-bit platforms
2377 Multibyte fixes(Tatsuo)
2379 Fix memory overruns(Tatsuo)
2380 Fix for lo_import() crash(Bruce)
2381 Better search for install program(Tom)
2384 Use implicit type coercion for matching DEFAULT values(Thomas)
2385 Add routines to help with single-byte (internal) character type(Thomas)
2386 Compilation of libpq for Win32 fixes(Magnus)
2387 Upgrade to PyGreSQL 2.2(D'Arcy)
2388 _________________________________________________________________
2392 Release date: 1998-10-30
2394 There are *many* new features and improvements in this release. Thanks
2395 to our developers and maintainers, nearly every aspect of the system
2396 has received some attention since the previous release. Here is a
2397 brief, incomplete summary:
2399 * Views and rules are now functional thanks to extensive new code in
2400 the rewrite rules system from Jan Wieck. He also wrote a chapter
2401 on it for the Programmer's Guide.
2402 * Jan also contributed a second procedural language, PL/pgSQL, to go
2403 with the original PL/pgTCL procedural language he contributed last
2405 * We have optional multiple-byte character set support from Tatsuo
2406 Ishii to complement our existing locale support.
2407 * Client/server communications has been cleaned up, with better
2408 support for asynchronous messages and interrupts thanks to Tom
2410 * The parser will now perform automatic type coercion to match
2411 arguments to available operators and functions, and to match
2412 columns and expressions with target columns. This uses a generic
2413 mechanism which supports the type extensibility features of
2414 PostgreSQL. There is a new chapter in the User's Guide which
2416 * Three new data types have been added. Two types, inet and cidr,
2417 support various forms of IP network, subnet, and machine
2418 addressing. There is now an 8-byte integer type available on some
2419 platforms. See the chapter on data types in the User's Guide for
2420 details. A fourth type, serial, is now supported by the parser as
2421 an amalgam of the int4 type, a sequence, and a unique index.
2422 * Several more SQL92-compatible syntax features have been added,
2423 including "INSERT DEFAULT VALUES"
2424 * The automatic configuration and installation system has received
2425 some attention, and should be more robust for more platforms than
2427 _________________________________________________________________
2429 Migration to version 6.4
2431 A dump/restore using pg_dump or pg_dumpall is required for those
2432 wishing to migrate data from any previous release of PostgreSQL.
2433 _________________________________________________________________
2439 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
2440 Remove char2-16 data types, use char/varchar(Darren)
2441 Pqfn not handles a NOTICE message(Anders)
2442 Reduced busywaiting overhead for spinlocks with many backends (dg)
2443 Stuck spinlock detection (dg)
2444 Fix up "ISO-style" timespan decoding and encoding(Thomas)
2445 Fix problem with table drop after rollback of transaction(Vadim)
2446 Change error message and remove non-functional update message(Vadim)
2447 Fix for COPY array checking
2448 Fix for SELECT 1 UNION SELECT NULL
2449 Fix for buffer leaks in large object calls(Pascal)
2450 Change owner from oid to int4 type(Bruce)
2451 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
2452 Fix for shared invalidation cache overflow(Massimo)
2453 Prevent file descriptor leaks in failed COPY's(Bruce)
2454 Fix memory leak in libpgtcl's pg_select(Constantin)
2455 Fix problems with username/passwords over 8 characters(Tom)
2456 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
2457 Fix of many bad system table entries(Tom)
2461 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
2462 Show the index used in an EXPLAIN(Zeugswetter)
2463 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
2464 Multibyte awareness of many data types and functions, via configure(Tatsuo)
2465 New configure --with-mb option(Tatsuo)
2466 New initdb --pgencoding option(Tatsuo)
2467 New createdb -E multibyte option(Tatsuo)
2468 Select version(); now returns PostgreSQL version(Jeroen)
2469 Libpq now allows asynchronous clients(Tom)
2470 Allow cancel from client of backend query(Tom)
2471 Psql now cancels query with Control-C(Tom)
2472 Libpq users need not issue dummy queries to get NOTIFY messages(Tom)
2473 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
2474 PGresult struct now includes associated error message, if any(Tom)
2475 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
2476 Add routines to convert between varchar and bpchar(Thomas)
2477 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
2478 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
2479 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas
2481 Fixes for unary minus parsing with leading spaces(Thomas)
2482 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
2483 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
2484 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
2485 Enable HAVING clause but no fixes elsewhere yet.
2486 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thoma
2488 Save string type if specified for DEFAULT clause handling(Thomas)
2489 Coerce operations involving different data types(Thomas)
2490 Allow some index use for columns of different types(Thomas)
2491 Add capabilities for automatic type conversion(Thomas)
2492 Cleanups for large objects, so file is truncated on open(Peter)
2493 Readline cleanups(Tom)
2494 Allow psql \f \ to make spaces as delimiter(Bruce)
2495 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
2496 Msql compatibility library in /contrib(Aldrin)
2497 Remove the requirement that ORDER/GROUP BY clause identifiers be
2498 included in the target list(David)
2499 Convert columns to match columns in UNION clauses(Thomas)
2500 Remove fork()/exec() and only do fork()(Bruce)
2501 Jdbc cleanups(Peter)
2502 Show backend status on ps command line(only works on some platforms)(Bruce)
2503 Pg_hba.conf now has a sameuser option in the database field
2504 Make lo_unlink take oid param, not int4
2505 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
2506 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
2507 libpgtcl cleanups(Tom)
2508 Add -error option to libpgtcl's pg_result command(Tom)
2509 New locale patch, see docs/README/locale(Oleg)
2510 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
2511 New contrib/lo code for large object orphan removal(Peter)
2512 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
2513 feature, see /doc/README.mb(Tatsuo)
2514 /contrib/noupdate code to revoke update permission on a column
2515 Libpq can now be compiled on win32(Magnus)
2516 Add PQsetdbLogin() in libpq
2517 New 8-byte integer type, checked by configure for OS support(Thomas)
2518 Better support for quoted table/column names(Thomas)
2519 Surround table and column names with double-quotes in pg_dump(Thomas)
2520 PQreset() now works with passwords(Tom)
2521 Handle case of GROUP BY target list column number out of range(David)
2522 Allow UNION in subselects
2523 Add auto-size to screen to \d? commands(Bruce)
2524 Use UNION to show all \d? results in one query(Bruce)
2525 Add \d? field search feature(Bruce)
2526 Pg_dump issues fewer \connect requests(Tom)
2527 Make pg_dump -z flag work better, document it in manual page(Tom)
2528 Add HAVING clause with full support for subselects and unions(Stephan)
2529 Full text indexing routines in contrib/fulltextindex(Maarten)
2530 Transaction ids now stored in shared memory(Vadim)
2531 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
2532 Support for SQL92 syntax "SET NAMES"(Tatsuo)
2533 Support for LATIN2-5(Tatsuo)
2534 Add UNICODE regression test case(Tatsuo)
2535 Lock manager cleanup, new locking modes for LLL(Vadim)
2536 Allow index use with OR clauses(Bruce)
2537 Allows "SELECT NULL ORDER BY 1;"
2538 Explain VERBOSE prints the plan, and now pretty-prints the plan to
2539 the postmaster log file(Bruce)
2540 Add indexes display to \d command(Bruce)
2541 Allow GROUP BY on functions(David)
2542 New pg_class.relkind for large objects(Bruce)
2543 New way to send libpq NOTICE messages to a different location(Tom)
2544 New \w write command to psql(Bruce)
2545 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
2546 Allow binary-compatible indexes to be considered when checking for valid
2547 Indexes for restriction clauses containing a constant(Thomas)
2548 New ISBN/ISSN code in /contrib/isbn_issn
2549 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
2550 New rewrite system fixes many problems with rules and views(Jan)
2551 * Rules on relations work
2552 * Event qualifications on insert/update/delete work
2553 * New OLD variable to reference CURRENT, CURRENT will be remove in futu
2555 * Update rules can reference NEW and OLD in rule qualifications/actions
2556 * Insert/update/delete rules on views work
2557 * Multiple rule actions are now supported, surrounded by parentheses
2558 * Regular users can create views/rules on tables they have RULE permits
2559 * Rules and views inherit the permissions on the creator
2560 * No rules at the column level
2561 * No UPDATE NEW/OLD rules
2562 * New pg_tables, pg_indexes, pg_rules and pg_views system views
2563 * Only a single action on SELECT rules
2564 * Total rewrite overhaul, perhaps for 6.5
2566 * handle aggregates on views
2567 * handle insert into select from view works
2568 System indexes are now multikey(Bruce)
2569 Oidint2, oidint4, and oidname types are removed(Bruce)
2570 Use system cache for more system table lookups(Bruce)
2571 New backend programming language PL/pgSQL in backend/pl(Jan)
2572 New SERIAL data type, auto-creates sequence/index(Thomas)
2573 Enable assert checking without a recompile(Massimo)
2574 User lock enhancements(Massimo)
2575 New setval() command to set sequence value(Massimo)
2576 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
2577 Conditional trace package(Massimo)
2578 New UNLISTEN command(Massimo)
2579 Psql and libpq now compile under win32 using win32.mak(Magnus)
2580 Lo_read no longer stores trailing NULL(Bruce)
2581 Identifiers are now truncated to 31 characters internally(Bruce)
2582 Createuser options now availble on the command line
2583 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
2584 Prevent file descriptor leaf from failed COPY(Bruce)
2585 New pg_upgrade command(Bruce)
2586 Updated /contrib directories(Massimo)
2587 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
2588 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
2589 New DECLARE and FETCH feature(Thomas)
2590 libpq's internal structures now not exported(Tom)
2591 Allow up to 8 key indexes(Bruce)
2592 Remove ARCHIVE keyword, that is no longer used(Thomas)
2593 pg_dump -n flag to supress quotes around indentifiers
2594 disable system columns for views(Jan)
2595 new INET and CIDR types for network addresses(TomH, Paul)
2596 no more double quotes in psql output
2597 pg_dump now dumps views(Terry)
2598 new SET QUERY_LIMIT(Tatsuo,Jan)
2602 /contrib cleanup(Jun)
2603 Inline some small functions called for every row(Bruce)
2606 Multibyte regression tests(Soonmyung.)
2607 Remove --disabled options from configure
2608 Define PGDOC to use POSTGRESDIR by default
2609 Make regression optional
2610 Remove extra braces code to pgindent(Bruce)
2611 Add bsdi shared library support(Bruce)
2612 New --without-CXX support configure option(Brook)
2614 Update backend flowchart in tools/backend(Bruce)
2615 Change atttypmod from int16 to int32(Bruce, Tom)
2616 Getrusage() fix for platforms that do not have it(Tom)
2617 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
2618 NS32K platform fixes(Phil Nelson, John Buller)
2619 SCO 7/UnixWare 2.x fixes(Billy,others)
2620 Sparc/Solaris 2.5 fixes(Ryan)
2621 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
2622 Even more documention(Thomas)
2623 Nextstep support(Jacek)
2625 pginterface manual page(Bruce)
2626 shared libraries all have version numbers
2627 merged all OS-specific shared library defines into one file
2628 smarter TCL/TK configuration checking(Billy)
2629 smarter perl configuration(Brook)
2630 configure uses supplied install-sh if no install script found(Tom)
2631 new Makefile.shlib for shared library configuration(Tom)
2632 _________________________________________________________________
2636 Release date: 1998-04-07
2638 This is a bug-fix release for 6.3.x. Refer to the release notes for
2639 version 6.3 for a more complete summary of new features.
2643 * Repairs automatic configuration support for some platforms,
2644 including Linux, from breakage inadvertently introduced in version
2646 * Correctly handles function calls on the left side of BETWEEN and
2649 A dump/restore is NOT required for those running 6.3 or 6.3.1. A make
2650 distclean, make, and make install is all that is required. This last
2651 step should be performed while the postmaster is not running. You
2652 should re-link any custom applications that use PostgreSQL libraries.
2654 For upgrades from pre-6.3 installations, refer to the installation and
2655 migration instructions for version 6.3.
2656 _________________________________________________________________
2660 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
2661 Manual page improvements(Bruce)
2662 BETWEEN and LIKE fix(Thomas)
2663 fix for psql \connect used by pg_dump(Oliver Elphick)
2665 pgaccess, version 0.86
2666 qsort removed, now uses libc version, cleanups(Jeroen)
2667 fix for buffer over-runs detected(Maurice Gittens)
2668 fix for buffer overrun in libpgtcl(Randy Kunkee)
2669 fix for UNION with DISTINCT or ORDER BY(Bruce)
2670 gettimeofday configure check(Doug Winterburn)
2671 Fix "indexes not used" bug(Vadim)
2672 docs additions(Thomas)
2673 Fix for backend memory leak(Bruce)
2674 libreadline cleanup(Erwan MAS)
2675 Remove DISTDIR(Bruce)
2676 Makefile dependency cleanup(Jeroen van Vianen)
2679 _________________________________________________________________
2683 Release date: 1998-03-23
2687 * Additional support for multibyte character sets.
2688 * Repair byte ordering for mixed-endian clients and servers.
2689 * Minor updates to allowed SQL syntax.
2690 * Improvements to the configuration autodetection for installation.
2692 A dump/restore is NOT required for those running 6.3. A make
2693 distclean, make, and make install is all that is required. This last
2694 step should be performed while the postmaster is not running. You
2695 should re-link any custom applications that use PostgreSQL libraries.
2697 For upgrades from pre-6.3 installations, refer to the installation and
2698 migration instructions for version 6.3.
2699 _________________________________________________________________
2703 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
2704 pg_user cleanup(Bruce)
2705 large object fix for pg_dump and tclsh (alvin)
2706 LIKE fix for multiple adjacent underscores
2707 fix for redefining builtin functions(Thomas)
2709 upgrade to pg_access 0.83
2710 updated CLUSTER manual page
2711 multibyte character set support, see doc/README.mb(Tatsuo)
2712 configure --with-pgport fix
2714 big-endian fix for backend communications(Kataoka)
2715 SUBSTR() and substring() fix(Jan)
2716 several jdbc fixes(Peter)
2717 libpgtcl improvements, see libptcl/README(Randy Kunkee)
2718 Fix for "Datasize = 0" error(Vadim)
2719 Prevent \do from wrapping(Bruce)
2720 Remove duplicate Russian character set entries
2722 Allow optional TABLE keyword in LOCK and SELECT INTO(Thomas)
2723 CREATE SEQUENCE options to allow a negative integer(Thomas)
2724 Add "PASSWORD" as an allowed column identifier(Thomas)
2725 Add checks for UNION target fields(Bruce)
2726 Fix Alpha port(Dwayne Bailey)
2727 Fix for text arrays containing quotes(Doug Gibson)
2728 Solaris compile fix(Albert Chin-A-Young)
2729 Better identify tcl and tk libs and includes(Bruce)
2731 _________________________________________________________________
2735 Release date: 1998-03-01
2737 There are *many* new features and improvements in this release. Here
2738 is a brief, incomplete summary:
2740 * Many new SQL features, including full SQL92 subselect capability
2741 (everything is here but target-list subselects).
2742 * Support for client-side environment variables to specify time zone
2744 * Socket interface for client/server connection. This is the default
2745 now so you may need to start postmaster with the "-i" flag.
2746 * Better password authorization mechanisms. Default table
2747 permissions have changed.
2748 * Old-style time travel has been removed. Performance has been
2751 Note: Bruce Momjian wrote the following notes to introduce the new
2754 There are some general 6.3 issues that I want to mention. These are
2755 only the big items that can not be described in one sentence. A review
2756 of the detailed changes list is still needed.
2758 First, we now have subselects. Now that we have them, I would like to
2759 mention that without subselects, SQL is a very limited language.
2760 Subselects are a major feature, and you should review your code for
2761 places where subselects provide a better solution for your queries. I
2762 think you will find that there are more uses for subselects than you
2763 may think. Vadim has put us on the big SQL map with subselects, and
2764 fully functional ones too. The only thing you can't do with subselects
2765 is to use them in the target list.
2767 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
2768 enable connections from other machines, you have to use the new
2769 postmaster -i option, and of course edit "pg_hba.conf". Also, for this
2770 reason, the format of "pg_hba.conf" has changed.
2772 Third, char() fields will now allow faster access than varchar() or
2773 text. Specifically, the text and varchar() have a penalty for access
2774 to any columns after the first column of this type. char() used to
2775 also have this access penalty, but it no longer does. This may suggest
2776 that you redesign some of your tables, especially if you have short
2777 character columns that you have defined as varchar() or text. This and
2778 other changes make 6.3 even faster than earlier releases.
2780 We now have passwords definable independent of any Unix file. There
2781 are new SQL USER commands. See the Administrator's Guide for more
2782 information. There is a new table, pg_shadow, which is used to store
2783 user information and user passwords, and it by default only
2784 SELECT-able by the postgres super-user. pg_user is now a view of
2785 pg_shadow, and is SELECT-able by PUBLIC. You should keep using pg_user
2786 in your application without changes.
2788 User-created tables now no longer have SELECT permission to PUBLIC by
2789 default. This was done because the ANSI standard requires it. You can
2790 of course GRANT any permissions you want after the table is created.
2791 System tables continue to be SELECT-able by PUBLIC.
2793 We also have real deadlock detection code. No more sixty-second
2794 timeouts. And the new locking code implements a FIFO better, so there
2795 should be less resource starvation during heavy use.
2797 Many complaints have been made about inadequate documentation in
2798 previous releases. Thomas has put much effort into many new manuals
2799 for this release. Check out the doc/ directory.
2801 For performance reasons, time travel is gone, but can be implemented
2802 using triggers (see "pgsql/contrib/spi/README"). Please check out the
2803 new \d command for types, operators, etc. Also, views have their own
2804 permissions now, not based on the underlying tables, so permissions on
2805 them have to be set separately. Check "/pgsql/interfaces" for some new
2806 ways to talk to PostgreSQL.
2808 This is the first release that really required an explanation for
2809 existing users. In many ways, this was necessary because the new
2810 release removes many limitations, and the work-arounds people were
2811 using are no longer needed.
2812 _________________________________________________________________
2814 Migration to version 6.3
2816 A dump/restore using pg_dump or pg_dumpall is required for those
2817 wishing to migrate data from any previous release of PostgreSQL.
2818 _________________________________________________________________
2824 Fix binary cursors broken by MOVE implementation(Vadim)
2825 Fix for tcl library crash(Jan)
2826 Fix for array handling, from Gerhard Hintermayer
2827 Fix acl error, and remove duplicate pqtrace(Bruce)
2828 Fix psql \e for empty file(Bruce)
2829 Fix for textcat on varchar() fields(Bruce)
2830 Fix for DBT Sendproc (Zeugswetter Andres)
2831 Fix vacuum analyze syntax problem(Bruce)
2832 Fix for international identifiers(Tatsuo)
2833 Fix aggregates on inherited tables(Bruce)
2834 Fix substr() for out-of-bounds data
2835 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
2836 Fix notty output to show status result. -q option still turns it off(Bruce)
2837 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
2839 Fix for PQtrace start/stop several times(Bruce)
2840 Fix a variety of locking problems like newer lock waiters getting
2841 lock before older waiters, and having readlock people not share
2842 locks if a writer is waiting for a lock, and waiting writers not
2843 getting priority over waiting readers(Bruce)
2844 Fix crashes in psql when executing queries from external files(James)
2845 Fix problem with multiple order by columns, with the first one having
2847 Use correct hash table support functions for float8 and int4(Thomas)
2848 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
2849 Change precedence for boolean operators to match expected behavior(Thomas)
2850 Generate elog(ERROR) on over-large integer(Bruce)
2851 Allow multiple-argument functions in constraint clauses(Thomas)
2852 Check boolean input literals for 'true','false','yes','no','1','0'
2853 and throw elog(ERROR) if unrecognized(Thomas)
2854 Major large objects fix
2855 Fix for GROUP BY showing duplicates(Vadim)
2856 Fix for index scans in MergeJion(Vadim)
2860 Subselects with EXISTS, IN, ALL, ANY keywords (Vadim, Bruce, Thomas)
2861 New User Manual(Thomas, others)
2862 Speedup by inlining some frequently-called functions
2863 Real deadlock detection, no more timeouts(Bruce)
2864 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
2865 CURRENT_USER(Thomas)
2866 Modify constraint syntax to be SQL92-compliant(Thomas)
2867 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
2868 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
2869 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas
2871 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
2872 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
2873 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
2874 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
2875 Allow SQL92 delimited identifiers(Thomas)
2876 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas
2878 Support SQL92 syntax for type coercion of literal strings
2879 (e.g. "DATETIME 'now'")(Thomas)
2880 Add conversions for int2, int4, and OID types to and from text(Thomas)
2881 Use shared lock when building indexes(Vadim)
2882 Free memory allocated for an user query inside transaction block after
2883 this query is done, was turned off in <= 6.2.1(Vadim)
2884 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
2885 New PostgreSQL Procedural Language (PL) backend interface(Jan)
2886 Rename pg_dump -H option to -h(Bruce)
2887 Add Java support for passwords, European dates(Peter)
2888 Use indexes for LIKE and ~, !~ operations(Bruce)
2889 Add hash functions for datetime and timespan(Thomas)
2890 Time Travel removed(Vadim, Bruce)
2891 Add paging for \d and \z, and fix \i(Bruce)
2892 Add Unix domain socket support to backend and to frontend library(Goran)
2893 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
2894 Allow more SQL92 and/or PostgreSQL reserved words as column identifiers(Thomas)
2895 Augment support for SQL92 SET TIME ZONE...(Thomas)
2896 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
2897 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
2898 Enable SET TIME ZONE using TZ environment variable(Thomas)
2899 Add PGDATESTYLE environment variable to frontend and backend initialization(Tho
2901 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
2902 frontend library initialization environment variables(Thomas)
2903 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
2904 Add pg_description table for info on tables, columns, operators, types, and
2906 Increase 16 char limit on system table/index names to 32 characters(Bruce)
2907 Rename system indexes(Bruce)
2908 Add 'GERMAN' option to SET DATESTYLE(Thomas)
2909 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
2910 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
2911 Validate numeric input more carefully for delta times(Thomas)
2912 Implement day of year as possible input to date_part()(Thomas)
2913 Define timespan_finite() and text_timespan() functions(Thomas)
2914 Remove archive stuff(Bruce)
2915 Allow for a pg_password authentication database that is separate from
2916 the system password file(Todd)
2917 Dump ACLs, GRANT, REVOKE permissions(Matt)
2918 Define text, varchar, and bpchar string length functions(Thomas)
2919 Fix Query handling for inheritance, and cost computations(Bruce)
2920 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
2921 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
2922 Implement UNIONs for SELECT(Bruce)
2923 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
2924 varchar() stores only necessary bytes on disk(Bruce)
2925 Fix for BLOBs(Peter)
2926 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
2927 Remove unused "option" from PQconnectdb()
2928 New LOCK command and lock manual page describing deadlocks(Bruce)
2929 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
2930 Enhance psql \z to show sequences(Bruce)
2931 Show NOT NULL and DEFAULT in psql \d table(Bruce)
2932 New psql .psqlrc file start-up(Andrew)
2933 Modify sample start-up script in contrib/linux to show syslog(Thomas)
2934 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
2935 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
2936 Update of contrib stuff(Massimo)
2937 Add Unix socket support to DBD::Pg(Goran)
2938 New python interface (PyGreSQL 2.0)(D'Arcy)
2939 New frontend/backend protocol has a version number, network byte order(Phil)
2940 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
2941 CHAR() now faster access than VARCHAR() or TEXT
2942 ecpg embedded SQL preprocessor
2943 Reduce system column overhead(Vadmin)
2944 Remove pg_time table(Vadim)
2945 Add pg_type attribute to identify types that need length (bpchar, varchar)
2946 Add report of offending line when COPY command fails
2947 Allow VIEW permissions to be set separately from the underlying tables.
2948 For security, use GRANT/REVOKE on views as appropriate(Jan)
2949 Tables now have no default GRANT SELECT TO PUBLIC. You must
2950 explicitly grant such permissions.
2951 Clean up tutorial examples(Darren)
2955 Add new html development tools, and flow chart in /tools/backend
2956 Fix for SCO compiles
2957 Stratus computer port Robert Gillies
2958 Added support for shlib for BSD44_derived & i386_solaris
2959 Make configure more automated(Brook)
2960 Add script to check regression test results
2961 Break parser functions into smaller files, group together(Bruce)
2962 Rename heap_create to heap_create_and_catalog, rename heap_creatr
2963 to heap_create()(Bruce)
2964 Sparc/Linux patch for locking(TomS)
2965 Remove PORTNAME and reorganize port-specific stuff(Marc)
2966 Add optimizer README file(Bruce)
2967 Remove some recursion in optimizer and clean up some code there(Bruce)
2968 Fix for NetBSD locking(Henry)
2969 Fix for libptcl make(Tatsuo)
2971 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
2972 function calls to istrue() or isfalse() to allow optimization(Thomas)
2973 Various fixes NetBSD/Sparc related(TomH)
2974 Alpha linux locking(Travis,Ryan)
2975 Change elog(WARN) to elog(ERROR)(Bruce)
2976 FAQ for FreeBSD(Marc)
2977 Bring in the PostODBC source tree as part of our standard distribution(Marc)
2978 A minor patch for HP/UX 10 vs 9(Stan)
2979 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
2980 UnixWare patches(Billy)
2981 New i386 'lock' for spin lock asm(Billy)
2982 Support for multiplexed backends is removed
2983 Start an OpenBSD port
2986 Add string functions to regression suite(Thomas)
2987 Expand a few function names formerly truncated to 16 characters(Thomas)
2988 Remove un-needed malloc() calls and replace with palloc()(Bruce)
2989 _________________________________________________________________
2993 Release date: 1997-10-17
2995 6.2.1 is a bug-fix and usability release on 6.2.
2999 * Allow strings to span lines, per SQL92.
3000 * Include example trigger function for inserting user names on table
3003 This is a minor bug-fix release on 6.2. For upgrades from pre-6.2
3004 systems, a full dump/reload is required. Refer to the 6.2 release
3005 notes for instructions.
3006 _________________________________________________________________
3008 Migration from version 6.2 to version 6.2.1
3010 This is a minor bug-fix release. A dump/reload is not required from
3011 version 6.2, but is required from any release prior to 6.2.
3013 In upgrading from version 6.2, if you choose to dump/reload you will
3014 find that avg(money) is now calculated correctly. All other bug fixes
3015 take effect upon updating the executables.
3017 Another way to avoid dump/reload is to use the following SQL command
3018 from "psql" to update the existing system table:
3019 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
3020 where aggname = 'avg' and aggbasetype = 790;
3022 This will need to be done to every existing database, including
3024 _________________________________________________________________
3028 Allow TIME and TYPE column names(Thomas)
3029 Allow larger range of true/false as boolean values(Thomas)
3030 Support output of "now" and "current"(Thomas)
3031 Handle DEFAULT with INSERT of NULL properly(Vadim)
3032 Fix for relation reference counts problem in buffer manager(Vadim)
3033 Allow strings to span lines, like ANSI(Thomas)
3034 Fix for backward cursor with ORDER BY(Vadim)
3035 Fix avg(cash) computation(Thomas)
3036 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
3037 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
3038 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
3040 _________________________________________________________________
3044 Release date: 1997-10-02
3046 A dump/restore is required for those wishing to migrate data from
3047 previous releases of PostgreSQL.
3048 _________________________________________________________________
3050 Migration from version 6.1 to version 6.2
3052 This migration requires a complete dump of the 6.1 database and a
3053 restore of the database in 6.2.
3055 Note that the "pg_dump" and "pg_dumpall" utility from 6.2 should be
3056 used to dump the 6.1 database.
3057 _________________________________________________________________
3059 Migration from version 1.x to version 6.2
3061 Those migrating from earlier 1.* releases should first upgrade to 1.09
3062 because the COPY output format was improved from the 1.02 release.
3063 _________________________________________________________________
3069 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
3070 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
3071 from Solaris(Diab Jerius)
3072 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
3073 Check for geometric intersections at endpoints to avoid rounding ugliness(Thoma
3075 Catch non-functional delete attempts(Vadim)
3076 Change time function names to be more consistent(Michael Reifenberg)
3077 Check for zero divides(Michael Reifenberg)
3078 Fix very old bug which made tuples changed/inserted by a commnd
3079 visible to the command itself (so we had multiple update of
3080 updated tuples, etc)(Vadim)
3081 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
3082 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
3083 Remove un-needed signal stuff from contrib/pginterface
3084 Fix OR (where x != 1 or x isnull didn't return tuples with x NULL) (Vadim)
3085 Fix time_cmp function (Vadim)
3086 Fix handling of functions with non-attribute first argument in
3087 WHERE clauses (Vadim)
3088 Fix GROUP BY when order of entries is different from order
3089 in target list (Vadim)
3090 Fix pg_dump for aggregates without sfunc1 (Vadim)
3094 Default genetic optimizer GEQO parameter is now 8(Bruce)
3095 Allow use parameters in target list having aggregates in functions(Vadim)
3096 Added JDBC driver as an interface(Adrian & Peter)
3098 Return number of tuples inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
3099 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
3100 SPI (Server Programming Interface) allows execution of queries inside
3102 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
3103 Include reserved words for string handling, outer joins, and unions(Thomas)
3104 Implement extended comments ("/* ... */") using exclusive states(Thomas)
3105 Add "//" single-line comments(Bruce)
3106 Remove some restrictions on characters in operator names(Thomas)
3107 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
3108 Add text concatenation operator and function (SQL92)(Thomas)
3109 Support WITH TIME ZONE syntax (SQL92)(Thomas)
3110 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
3111 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
3112 and CHARACTER VARYING (SQL92)(Thomas)
3113 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
3114 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
3115 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
3116 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
3117 Add more reserved words, mostly for SQL92 compliance(Thomas)
3118 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
3119 Add center() routines for lseg, path, polygon(Thomas)
3120 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
3121 Check explicitly for points and polygons contained within polygons
3122 using an axis-crossing algorithm(Thomas)
3123 Add routine to convert circle-box(Thomas)
3124 Merge conflicting operators for different geometric data types(Thomas)
3125 Replace distance operator "<===>" with "<->"(Thomas)
3126 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Tho
3128 Add routines for text trimming on both ends, substring, and string position(Tho
3130 Added conversion routines circle(box) and poly(circle)(Thomas)
3131 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
3132 Allow functions and operators on internally-identical types to succeed(Bruce)
3133 Speed up backend start-up after profiling analysis(Bruce)
3134 Inline frequently called functions for performance(Bruce)
3135 Reduce open() calls(Bruce)
3136 psql: Add PAGER for \h and \?,\C fix
3137 Fix for psql pager when no tty(Bruce)
3138 New entab utility(Bruce)
3139 General trigger functions for referential integrity (Vadim)
3140 General trigger functions for time travel (Vadim)
3141 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
3142 MOVE implementation (Vadim)
3146 HP-UX 10 patches (Vladimir Turin)
3147 Added SCO support, (Daniel Harris)
3148 MkLinux patches (Tatsuo Ishii)
3149 Change geometric box terminology from "length" to "width"(Thomas)
3150 Deprecate temporary unstored slope fields in geometric code(Thomas)
3151 Remove restart instructions from INSTALL(Bruce)
3152 Look in /usr/ucb first for install(Bruce)
3153 Fix c++ copy example code(Thomas)
3154 Add -o to psql manual page(Bruce)
3155 Prevent relname unallocated string length from being copied into database(Bruce
3157 Cleanup for NAMEDATALEN use(Bruce)
3158 Fix pg_proc names over 15 chars in output(Bruce)
3159 Add strNcpy() function(Bruce)
3160 remove some (void) casts that are unnecessary(Bruce)
3161 new interfaces directory(Marc)
3162 Replace fopen() calls with calls to fd.c functions(Bruce)
3163 Make functions static where possible(Bruce)
3164 enclose unused functions in #ifdef NOT_USED(Bruce)
3165 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
3166 Changes for Digital Unix
3167 Portability fix for pg_dumpall(Bruce)
3168 Rename pg_attribute.attnvals to attdispersion(Bruce)
3169 "intro/unix" manual page now "pgintro"(Bruce)
3170 "built-in" manual page now "pgbuiltin"(Bruce)
3171 "drop" manual page now "drop_table"(Bruce)
3172 Add "create_trigger", "drop_trigger" manual pages(Thomas)
3173 Add constraints regression test(Vadim & Thomas)
3174 Add comments syntax regression test(Thomas)
3175 Add PGINDENT and support program(Bruce)
3176 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
3177 Files moved to /src/tools directory(Bruce)
3178 SPI and Trigger programming guides (Vadim & D'Arcy)
3179 _________________________________________________________________
3183 Release date: 1997-07-22
3184 _________________________________________________________________
3186 Migration from version 6.1 to version 6.1.1
3188 This is a minor bug-fix release. A dump/reload is not required from
3189 version 6.1, but is required from any release prior to 6.1. Refer to
3190 the release notes for 6.1 for more details.
3191 _________________________________________________________________
3195 fix for SET with options (Thomas)
3196 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
3197 new psql \connect option allows changing usernames without changing databases
3198 fix for initdb --debug option(Yoshihiko Ichikawa))
3199 lextest cleanup(Bruce)
3201 fix date/time month boundary arithmetic(Thomas)
3202 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
3203 timestamp overhauled to use standard functions(Thomas)
3204 other code cleanup in date/time routines(Thomas)
3205 psql's \d now case-insensitive(Bruce)
3206 psql's backslash commands can now have trailing semicolon(Bruce)
3207 fix memory leak in psql when using \g(Bruce)
3208 major fix for endian handling of communication to server(Thomas, Tatsuo)
3209 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
3210 allow underscores in usernames(Bruce)
3211 pg_dumpall now returns proper status, portability fix(Bruce)
3213 _________________________________________________________________
3217 Release date: 1997-06-08
3219 The regression tests have been adapted and extensively modified for
3220 the 6.1 release of PostgreSQL.
3222 Three new data types (datetime, timespan, and circle) have been added
3223 to the native set of PostgreSQL types. Points, boxes, paths, and
3224 polygons have had their output formats made consistent across the data
3225 types. The polygon output in misc.out has only been spot-checked for
3226 correctness relative to the original regression output.
3228 PostgreSQL 6.1 introduces a new, alternate optimizer which uses
3229 genetic algorithms. These algorithms introduce a random behavior in
3230 the ordering of query results when the query contains multiple
3231 qualifiers or multiple tables (giving the optimizer a choice on order
3232 of evaluation). Several regression tests have been modified to
3233 explicitly order the results, and hence are insensitive to optimizer
3234 choices. A few regression tests are for data types which are
3235 inherently unordered (e.g. points and time intervals) and tests
3236 involving those types are explicitly bracketed with "set geqo to
3237 'off'" and "reset geqo".
3239 The interpretation of array specifiers (the curly braces around atomic
3240 values) appears to have changed sometime after the original regression
3241 tests were generated. The current "./expected/*.out" files reflect
3242 this new interpretation, which may not be correct!
3244 The float8 regression test fails on at least some platforms. This is
3245 due to differences in implementations of pow() and exp() and the
3246 signaling mechanisms used for overflow and underflow conditions.
3248 The "random" results in the random test should cause the "random" test
3249 to be "failed", since the regression tests are evaluated using a
3250 simple diff. However, "random" does not seem to produce random results
3251 on my test machine (Linux/gcc/i686).
3252 _________________________________________________________________
3254 Migration to version 6.1
3256 This migration requires a complete dump of the 6.0 database and a
3257 restore of the database in 6.1.
3259 Those migrating from earlier 1.* releases should first upgrade to 1.09
3260 because the COPY output format was improved from the 1.02 release.
3261 _________________________________________________________________
3267 packet length checking in library routines
3268 lock manager priority patch
3269 check for under/over flow of float8(Bruce)
3270 multitable join fix(Vadim)
3271 SIGPIPE crash fix(Darren)
3272 large object fixes(Sven)
3273 allow btree indexes to handle NULLs(Vadim)
3274 timezone fixes(D'Arcy)
3275 select SUM(x) can return NULL on no rows(Thomas)
3276 internal optimizer, executor bug fixes(Vadim)
3277 fix problem where inner loop in < or <= has no rows(Vadim)
3278 prevent re-commuting join index clauses(Vadim)
3279 fix join clauses for multiple tables(Vadim)
3280 fix hash, hashjoin for arrays(Vadim)
3281 fix btree for abstime type(Vadim)
3282 large object fixes(Raymond)
3283 fix buffer leak in hash indexes (Vadim)
3284 fix rtree for use in inner scan (Vadim)
3285 fix gist for use in inner scan, cleanups (Vadim, Andrea)
3286 avoid unnecessary local buffers allocation (Vadim, Massimo)
3287 fix local buffers leak in transaction aborts (Vadim)
3288 fix file manager memmory leaks, cleanups (Vadim, Massimo)
3289 fix storage manager memmory leaks (Vadim)
3290 fix btree duplicates handling (Vadim)
3291 fix deleted tuples re-incarnation caused by vacuum (Vadim)
3292 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
3293 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
3297 attribute optimization statistics(Bruce)
3298 much faster new btree bulk load code(Paul)
3299 BTREE UNIQUE added to bulk load code(Vadim)
3300 new lock debug code(Massimo)
3301 massive changes to libpg++(Leo)
3302 new GEQO optimizer speeds table multitable optimization(Martin)
3303 new WARN message for non-unique insert into unique key(Marc)
3304 update x=-3, no spaces, now valid(Bruce)
3305 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
3306 debug backend now pretty-prints tree(Darren)
3307 new Oracle character functions(Edmund)
3308 new plaintext password functions(Dan)
3309 no such class or insufficient privilege changed to distinct messages(Dan)
3310 new ANSI timestamp function(Dan)
3311 new ANSI Time and Date types (Thomas)
3312 move large chunks of data in backend(Martin)
3313 multicolumn btree indexes(Vadim)
3314 new SET var TO value command(Martin)
3315 update transaction status on reads(Dan)
3316 new locale settings for character types(Oleg)
3317 new SEQUENCE serial number generator(Vadim)
3318 GROUP BY function now possible(Vadim)
3319 re-organize regression test(Thomas,Marc)
3320 new optimizer operation weights(Vadim)
3321 new psql \z grant/permit option(Marc)
3322 new MONEY data type(D'Arcy,Thomas)
3323 tcp socket communication speed improved(Vadim)
3324 new VACUUM option for attribute statistics, and for certain columns (Vadim)
3325 many geometric type improvements(Thomas,Keith)
3326 additional regression tests(Thomas)
3327 new datestyle variable(Thomas,Vadim,Martin)
3328 more comparison operators for sorting types(Thomas)
3329 new conversion functions(Thomas)
3330 new more compact btree format(Vadim)
3331 allow pg_dumpall to preserve database ownership(Bruce)
3332 new SET GEQO=# and R_PLANS variable(Vadim)
3333 old (!GEQO) optimizer can use right-sided plans (Vadim)
3334 typechecking improvement in SQL parser(Bruce)
3335 new SET, SHOW, RESET commands(Thomas,Vadim)
3336 new \connect database USER option
3337 new destroydb -i option (Igor)
3338 new \dt and \di psql commands (Darren)
3339 SELECT "\n" now escapes newline (A. Duursma)
3340 new geometry conversion functions from old format (Thomas)
3344 new configuration script(Marc)
3345 readline configuration option added(Marc)
3346 OS-specific configuration options removed(Marc)
3347 new OS-specific template files(Marc)
3348 no more need to edit Makefile.global(Marc)
3349 re-arrange include files(Marc)
3350 nextstep patches (Gregor Hoffleit)
3351 removed WIN32-specific code(Bruce)
3352 removed postmaster -e option, now only postgres -e option (Bruce)
3353 merge duplicate library code in front/backends(Martin)
3354 now works with eBones, international Kerberos(Jun)
3355 more shared library support
3356 c++ include file cleanup(Bruce)
3357 warn about buggy flex(Bruce)
3358 DG/UX, Ultrix, IRIX, AIX portability fixes
3359 _________________________________________________________________
3363 Release date: 1997-01-29
3365 A dump/restore is required for those wishing to migrate data from
3366 previous releases of PostgreSQL.
3367 _________________________________________________________________
3369 Migration from version 1.09 to version 6.0
3371 This migration requires a complete dump of the 1.09 database and a
3372 restore of the database in 6.0.
3373 _________________________________________________________________
3375 Migration from pre-1.09 to version 6.0
3377 Those migrating from earlier 1.* releases should first upgrade to 1.09
3378 because the COPY output format was improved from the 1.02 release.
3379 _________________________________________________________________
3385 ALTER TABLE bug - running postgress process needs to re-read table definition
3386 Allow vacuum to be run on one table or entire database(Bruce)
3388 Fix array over-runs of memory writes(Kurt)
3389 Fix elusive btree range/non-range bug(Dan)
3390 Fix for hash indexes on some types like time and date
3391 Fix for pg_log size explosion
3392 Fix permissions on lo_export()(Bruce)
3393 Fix unitialized reads of memory(Kurt)
3394 Fixed ALTER TABLE ... char(3) bug(Bruce)
3395 Fixed a few small memory leaks
3396 Fixed EXPLAIN handling of options and changed full_path option name
3397 Fixed output of group acl permissions
3398 Memory leaks (hunt and destroy with tools like Purify(Kurt)
3399 Minor improvements to rules system
3401 New asserts for run-checking
3402 Overhauled parser/analyze code to properly report errors and increase speed
3403 Pg_dump -d now handles NULL's properly(Bruce)
3404 Prevent SELECT NULL from crashing server (Bruce)
3405 Properly report errors when INSERT ... SELECT columns did not match
3406 Properly report errors when insert column names were not correct
3407 Psql \g filename now works(Bruce)
3408 Psql fixed problem with multiple statements on one line with multiple outputs
3409 Removed duplicate system OIDs
3410 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
3411 Several fixes for queries that crashed the backend
3412 Starting quote in insert string errors(Bruce)
3413 Submitting an empty query now returns empty status, not just " " query(Bruce)
3417 Add EXPLAIN manual page(Bruce)
3418 Add UNIQUE index capability(Dan)
3419 Add hostname/user level access control rather than just hostname and user
3420 Add synonym of != for <>(Bruce)
3421 Allow "select oid,* from table"
3422 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Br
3424 Allow COPY from the frontend(Bryan)
3425 Allow GROUP BY to use alias column name(Bruce)
3426 Allow actual compression, not just reuse on the same page(Vadim)
3427 Allow installation-configuration option to auto-add all local users(Bryan)
3428 Allow libpq to distinguish between text value '' and null(Bruce)
3429 Allow non-postgres users with createdb privs to destroydb's
3430 Allow restriction on who can create C functions(Bryan)
3431 Allow restriction on who can do backend COPY(Bryan)
3432 Can shrink tables, pg_time and pg_log(Vadim & Erich)
3433 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
3434 Change default decimal constant representation from float4 to float8(Bruce)
3435 European date format now set when postmaster is started
3436 Execute lowercase function names if not found with exact case
3437 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z
3439 Gist now included in the distrubution(Marc)
3440 Idend authentication of local users(Bryan)
3441 Implement BETWEEN qualifier(Bruce)
3442 Implement IN qualifier(Bruce)
3443 Libpq has PQgetisnull()(Bruce)
3444 Libpq++ improvements
3445 New options to initdb(Bryan)
3446 Pg_dump allow dump of OIDs(Bruce)
3447 Pg_dump create indexes after tables are loaded for speed(Bruce)
3448 Pg_dumpall dumps all databases, and the user table
3449 Pginterface additions for NULL values(Bruce)
3450 Prevent postmaster from being run as root
3451 Psql \h and \? is now readable(Bruce)
3452 Psql allow backslashed, semicolons anywhere on the line(Bruce)
3453 Psql changed command prompt for lines in query or in quotes(Bruce)
3454 Psql char(3) now displays as (bp)char in \d output(Bruce)
3455 Psql return code now more accurate(Bryan?)
3456 Psql updated help syntax(Bruce)
3457 Re-visit and fix vacuum(Vadim)
3458 Reduce size of regression diffs, remove timezone name difference(Bruce)
3459 Remove compile-time parameters to enable binary distributions(Bryan)
3460 Reverse meaning of HBA masks(Bryan)
3461 Secure Authentication of local users(Bryan)
3462 Speed up vacuum(Vadim)
3463 Vacuum now had VERBOSE option(Bruce)
3467 All functions now have prototypes that are compared against the calls
3468 Allow asserts to be disabled easly from Makefile.global(Bruce)
3469 Change oid constants used in code to #define names
3470 Decoupled sparc and solaris defines(Kurt)
3471 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
3472 Major include file reorganization/reduction(Marc)
3473 Make now stops on compile failure(Bryan)
3474 Makefile restructuring(Bryan, Marc)
3475 Merge bsdi_2_1 to bsdi(Bruce)
3476 Monitor program removed
3477 Name change from Postgres95 to PostgreSQL
3478 New config.h file(Marc, Bryan)
3479 PG_VERSION now set to 6.0 and used by postmaster
3480 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
3481 Reduced the number of #define's, centeralized #define's
3482 Remove duplicate OIDS in system tables(Dan)
3483 Remove duplicate system catalog info or report mismatches(Dan)
3484 Removed many os-specific #define's
3485 Restructured object file generation/location(Bryan, Marc)
3486 Restructured port-specific file locations(Bryan, Marc)
3487 Unused/uninialized variables corrected
3488 _________________________________________________________________
3492 Release date: 1996-11-04
3494 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of the
3495 changes listed in 6.0 were actually included in the 1.02.1 to 1.09
3497 _________________________________________________________________
3501 Release date: 1996-08-01
3502 _________________________________________________________________
3504 Migration from version 1.02 to version 1.02.1
3506 Here is a new migration file for 1.02.1. It includes the 'copy' change
3507 and a script to convert old ASCII files.
3509 Note: The following notes are for the benefit of users who want to
3510 migrate databases from Postgres95 1.01 and 1.02 to Postgres95
3513 If you are starting afresh with Postgres95 1.02.1 and do not need
3514 to migrate old databases, you do not need to read any further.
3516 In order to upgrade older Postgres95 version 1.01 or 1.02 databases to
3517 version 1.02.1, the following steps are required:
3518 1. Start up a new 1.02.1 postmaster
3519 2. Add the new built-in functions and operators of 1.02.1 to 1.01 or
3520 1.02 databases. This is done by running the new 1.02.1 server
3521 against your own 1.01 or 1.02 database and applying the queries
3522 attached at the end of the file. This can be done easily through
3523 "psql". If your 1.01 or 1.02 database is named testdb and you have
3524 cut the commands from the end of this file and saved them in
3526 % psql testdb -f addfunc.sql
3527 Those upgrading 1.02 databases will get a warning when executing
3528 the last two statements in the file because they are already
3529 present in 1.02. This is not a cause for concern.
3530 _________________________________________________________________
3532 Dump/Reload Procedure
3534 If you are trying to reload a pg_dump or text-mode, copy tablename to
3535 stdout generated with a previous version, you will need to run the
3536 attached "sed" script on the ASCII file before loading it into the
3537 database. The old format used '.' as end-of-data, while '\.' is now
3538 the end-of-data marker. Also, empty strings are now loaded in as ''
3539 rather than NULL. See the copy manual page for full details.
3540 sed 's/^\.$/\\./g' <in_file >out_file
3542 If you are loading an older binary copy or non-stdout copy, there is
3543 no end-of-data character, and hence no conversion necessary.
3544 -- following lines added by agc to reflect the case-insensitive
3545 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
3546 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexe
3548 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregex
3550 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregex
3552 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticrege
3554 _________________________________________________________________
3558 Source code maintenance and development
3559 * worldwide team of volunteers
3560 * the source tree now in CVS at ftp.ki.net
3563 * psql (and underlying libpq library) now has many more options for
3564 formatting output, including HTML
3565 * pg_dump now output the schema and/or the data, with many fixes to
3566 enhance completeness.
3567 * psql used in place of monitor in administration shell scripts.
3568 monitor to be deprecated in next release.
3569 * date/time functions enhanced
3570 * NULL insert/update/comparison fixed/enhanced
3571 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
3573 Bug Fixes (almost too numerous to mention)
3575 * storage management
3576 * check for NULL pointer before dereferencing
3580 * added SolarisX86 port
3581 * added BSD/OS 2.1 port
3583 _________________________________________________________________
3587 Release date: 1996-02-23
3588 _________________________________________________________________
3590 Migration from version 1.0 to version 1.01
3592 The following notes are for the benefit of users who want to migrate
3593 databases from Postgres95 1.0 to Postgres95 1.01.
3595 If you are starting afresh with Postgres95 1.01 and do not need to
3596 migrate old databases, you do not need to read any further.
3598 In order to Postgres95 version 1.01 with databases created with
3599 Postgres95 version 1.0, the following steps are required:
3600 1. Set the definition of NAMEDATALEN in "src/Makefile.global" to 16
3601 and OIDNAMELEN to 20.
3602 2. Decide whether you want to use Host based authentication.
3603 a. If you do, you must create a file name pg_hba in your
3604 top-level data directory (typically the value of your
3605 $PGDATA). "src/libpq/pg_hba" shows an example syntax.
3606 b. If you do not want host-based authentication, you can comment
3609 in "src/Makefile.global"
3610 Note that host-based authentication is turned on by default,
3611 and if you do not take steps A or B above, the out-of-the-box
3612 1.01 will not allow you to connect to 1.0 databases.
3613 3. Compile and install 1.01, but DO NOT do the "initdb" step.
3614 4. Before doing anything else, terminate your 1.0 postmaster, and
3615 backup your existing $PGDATA directory.
3616 5. Set your PGDATA environment variable to your 1.0 databases, but
3617 set up path up so that 1.01 binaries are being used.
3618 6. Modify the file "$PGDATA/PG_VERSION" from 5.0 to 5.1
3619 7. Start up a new 1.01 postmaster
3620 8. Add the new built-in functions and operators of 1.01 to 1.0
3621 databases. This is done by running the new 1.01 server against
3622 your own 1.0 database and applying the queries attached and saving
3623 in the file 1.0_to_1.01.sql. This can be done easily through
3624 "psql". If your 1.0 database is name testdb:
3625 % psql testdb -f 1.0_to_1.01.sql
3626 and then execute the following commands (cut and paste from here):
3627 -- add builtin functions that are new to 1.01
3629 create function int4eqoid (int4, oid) returns bool as 'foo'
3630 language 'internal';
3631 create function oideqint4 (oid, int4) returns bool as 'foo'
3632 language 'internal';
3633 create function char2icregexeq (char2, text) returns bool as 'foo'
3634 language 'internal';
3635 create function char2icregexne (char2, text) returns bool as 'foo'
3636 language 'internal';
3637 create function char4icregexeq (char4, text) returns bool as 'foo'
3638 language 'internal';
3639 create function char4icregexne (char4, text) returns bool as 'foo'
3640 language 'internal';
3641 create function char8icregexeq (char8, text) returns bool as 'foo'
3642 language 'internal';
3643 create function char8icregexne (char8, text) returns bool as 'foo'
3644 language 'internal';
3645 create function char16icregexeq (char16, text) returns bool as 'foo'
3646 language 'internal';
3647 create function char16icregexne (char16, text) returns bool as 'foo'
3648 language 'internal';
3649 create function texticregexeq (text, text) returns bool as 'foo'
3650 language 'internal';
3651 create function texticregexne (text, text) returns bool as 'foo'
3652 language 'internal';
3654 -- add builtin functions that are new to 1.01
3656 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
3657 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
3658 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexe
3660 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregex
3662 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexe
3664 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregex
3666 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexe
3668 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregex
3670 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icrege
3672 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icreg
3674 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq)
3676 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne
3678 _________________________________________________________________
3683 * 1.01 is backwards compatible with 1.0 database provided the user
3684 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
3685 If those steps are not taken, 1.01 is not compatible with 1.0 database.
3688 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
3689 * added NeXT port (requires SysVIPC implementation)
3690 * added CAST .. AS ... syntax
3691 * added ASC and DESC keywords
3692 * added 'internal' as a possible language for CREATE FUNCTION
3693 internal functions are C functions which have been statically linked
3694 into the postgres backend.
3695 * a new type "name" has been added for system identifiers (table names,
3696 attribute names, etc.) This replaces the old char16 type. The
3697 of name is set by the NAMEDATALEN #define in src/Makefile.global
3698 * a readable reference manual that describes the query language.
3699 * added host-based access control. A configuration file ($PGDATA/pg_hba)
3700 is used to hold the configuration data. If host-based access control
3701 is not desired, comment out HBA=1 in src/Makefile.global.
3702 * changed regex handling to be uniform use of Henry Spencer's regex code
3703 regardless of platform. The regex code is included in the distribution
3704 * added functions and operators for case-insensitive regular expressions.
3705 The operators are ~* and !~*.
3706 * pg_dump uses COPY instead of SELECT loop for better performance
3709 * fixed an optimizer bug that was causing core dumps when
3710 functions calls were used in comparisons in the WHERE clause
3711 * changed all uses of getuid to geteuid so that effective uids are used
3712 * psql now returns non-zero status on errors when using -c
3713 * applied public patches 1-14
3714 _________________________________________________________________
3718 Release date: 1995-09-05
3719 _________________________________________________________________
3724 * The copyright of Postgres 1.0 has been loosened to be freely modifiable
3725 and modifiable for any purpose. Please read the COPYRIGHT file.
3726 Thanks to Professor Michael Stonebraker for making this possible.
3729 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
3730 EUROPEAN STYLE). This follows SQL-92 specs.
3731 * "delimiters" is now a keyword
3734 * sql LIKE syntax has been added
3735 * copy command now takes an optional USING DELIMITER specification.
3736 delimiters can be any single-character string.
3737 * IRIX 5.3 port has been added.
3738 Thanks to Paul Walmsley and others.
3739 * updated pg_dump to work with new libpq
3740 * \d has been added psql
3741 Thanks to Keith Parks
3742 * regexp performance for architectures that use POSIX regex has been
3743 improved due to caching of precompiled patterns.
3744 Thanks to Alistair Crooks
3745 * a new version of libpq++
3746 Thanks to William Wanders
3749 * arbitrary userids can be specified in the createuser script
3750 * \c to connect to other databases in psql now works.
3751 * bad pg_proc entry for float4inc() is fixed
3752 * users with usecreatedb field set can now create databases without
3753 having to be usesuper
3754 * remove access control entries when the entry no longer has any
3756 * fixed non-portable datetimes implementation
3757 * added kerberos flags to the src/backend/Makefile
3758 * libpq now works with kerberos
3759 * typographic errors in the user manual have been corrected.
3760 * btrees with multiple index never worked, now we tell you they don't
3761 work when you try to use them
3762 _________________________________________________________________
3764 Postgres95 Release 0.03
3766 Release date: 1995-07-21
3767 _________________________________________________________________
3771 Incompatible changes:
3772 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
3773 (due to system catalog changes and indexing structure changes).
3774 * double-quote (") is deprecated as a quoting character for string literals;
3775 you need to convert them to single quotes (').
3776 * name of aggregates (eg. int4sum) are renamed in accordance with the
3777 SQL standard (eg. sum).
3778 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
3779 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
3780 previous releases); you might have to do typecasting if you depend on it
3781 being of type float8. If you neglect to do the typecasting and you assign
3782 a float literal to a field of type float8, you may get incorrect values
3784 * LIBPQ has been totally revamped so that frontend applications
3785 can connect to multiple backends
3786 * the usesysid field in pg_user has been changed from int2 to int4 to
3787 allow wider range of Unix user ids.
3788 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
3789 single BSD44_derived port. (thanks to Alistair Crooks)
3791 SQL standard-compliance (the following details changes that makes postgres95
3792 more compliant to the SQL-92 standard):
3793 * the following SQL types are now built-in: smallint, int(eger), float, real,
3794 char(N), varchar(N), date and time.
3796 The following are aliases to existing postgres types:
3798 integer, int -> int4
3799 float, real -> float4
3800 char(N) and varchar(N) are implemented as truncated text types. In
3801 addition, char(N) does blank-padding.
3802 * single-quote (') is used for quoting string literals; '' (in addition to
3803 \') is supported as means of inserting a single quote in a string
3804 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
3805 (Also, aggregates can now be overloaded, i.e. you can define your
3806 own MAX aggregate to take in a user-defined type.)
3807 * CHANGE ACL removed. GRANT/REVOKE syntax added.
3808 - Privileges can be given to a group using the "GROUP" keyword.
3810 GRANT SELECT ON foobar TO GROUP my_group;
3811 The keyword 'PUBLIC' is also supported to mean all users.
3813 Privileges can only be granted or revoked to one user or group
3816 "WITH GRANT OPTION" is not supported. Only class owners can change
3818 - The default access control is to to grant users readonly access.
3819 You must explicitly grant insert/update access to users. To change
3820 this, modify the line in
3821 src/backend/utils/acl.h
3822 that defines ACL_WORLD_DEFAULT
3825 * the bug where aggregates of empty tables were not run has been fixed. Now,
3826 aggregates run on empty tables will return the initial conditions of the
3827 aggregates. Thus, COUNT of an empty table will now properly return 0.
3828 MAX/MIN of an empty table will return a tuple of value NULL.
3829 * allow the use of \; inside the monitor
3830 * the LISTEN/NOTIFY asynchronous notification mechanism now work
3831 * NOTIFY in rule action bodies now work
3832 * hash indexes work, and access methods in general should perform better.
3833 creation of large btree indexes should be much faster. (thanks to Paul
3836 Other changes and enhancements:
3837 * addition of an EXPLAIN statement used for explaining the query execution
3838 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
3840 * WARN and NOTICE messages no longer have timestamps on them. To turn on
3841 timestamps of error messages, uncomment the line in
3842 src/backend/utils/elog.h:
3843 /* define ELOG_TIMESTAMPS */
3844 * On an access control violation, the message
3845 "Either no such class or insufficient privilege"
3846 will be given. This is the same message that is returned when
3847 a class is not found. This dissuades non-privileged users from
3848 guessing the existence of privileged classes.
3849 * some additional system catalog changes have been made that are not
3850 visible to the user.
3853 * The -oid option has been added to the "pg_result" tcl command.
3854 pg_result -oid returns oid of the last tuple inserted. If the
3855 last command was not an INSERT, then pg_result -oid returns "".
3856 * the large object interface is available as pg_lo* tcl commands:
3857 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
3859 Portability enhancements and New Ports:
3860 * flex/lex problems have been cleared up. Now, you should be able to use
3861 flex instead of lex on any platforms. We no longer make assumptions of
3862 what lexer you use based on the platform you use.
3863 * The Linux-ELF port is now supported. Various configuration have been
3864 tested: The following configuration is known to work:
3865 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
3866 with everything in ELF format,
3869 * ipcclean added to the distribution
3870 ipcclean usually does not need to be run, but if your backend crashes
3871 and leaves shared memory segments hanging around, ipcclean will
3872 clean them up for you.
3875 * the user manual has been revised and libpq documentation added.
3876 _________________________________________________________________
3878 Postgres95 Release 0.02
3880 Release date: 1995-05-25
3881 _________________________________________________________________
3885 Incompatible changes:
3886 * The SQL statement for creating a database is 'CREATE DATABASE' instead
3887 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
3888 of 'DESTROYDB'. However, the names of the executables 'createdb' and
3889 'destroydb' remain the same.
3892 * pgperl - a Perl (4.036) interface to Postgres95
3893 * pg_dump - a utility for dumping out a postgres database into a
3894 script file containing query commands. The script files are in a ASCII
3895 format and can be used to reconstruct the database, even on other
3896 machines and other architectures. (Also good for converting
3897 a Postgres 4.2 database to Postgres95 database.)
3899 The following ports have been incorporated into postgres95-beta-0.02:
3900 * the NetBSD port by Alistair Crooks
3901 * the AIX port by Mike Tung
3902 * the Windows NT port by Jon Forrest (more stuff but not done yet)
3903 * the Linux ELF port by Brian Gallew
3905 The following bugs have been fixed in postgres95-beta-0.02:
3906 * new lines not escaped in COPY OUT and problem with COPY OUT when first
3908 * cannot type return to use the default user id in createuser
3909 * SELECT DISTINCT on big tables crashes
3910 * Linux installation problems
3911 * monitor doesn't allow use of 'localhost' as PGHOST
3912 * psql core dumps when doing \c or \l
3913 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
3914 * libpgtcl has a hard-wired default port number
3915 * SELECT DISTINCT INTO TABLE hangs
3916 * CREATE TYPE doesn't accept 'variable' as the internallength
3917 * wrong result using more than 1 aggregate in a SELECT
3918 _________________________________________________________________
3920 Postgres95 Release 0.01
3922 Release date: 1995-05-01