7 Valid as of 2003-08-01. Update release.sgml later.
12 Allow IPv6 server connections (Nigel Kukard, Johan Jordaan, Bruce, Tom, Kurt
13 Roeckx, Andrew Dunstan)
14 Fix SSL to handle errors cleanly (Nathan Mueller)
15 SSL protocol security and performance improvements (Sean Chittenden)
16 Print lock information when a deadlock is detected (Tom)
17 Update /tmp socket mod. times regularly to avoid their removal (Tom)
18 Enable PAM for MAC OS X (Aaron Hillegass)
19 Make btree indexes fully WAL-safe (Tom)
20 Allow btree index compaction and empty page reuse (Tom)
21 Fix inconsistent index lookups during split of first root page (Tom)
22 Improve free space map allocation logic (Tom)
23 Preserve free space information between postmaster restarts (Tom)
24 Set proper schema permissions in initdb (Peter)
25 Add start time to pg_stat_activity (Neil)
26 New code to detect corrupt disk pages; erase with zero_damaged_pages (Tom)
27 New client/server protocol: faster, no username length limit, allow
28 clean exit from COPY (Tom)
29 Add transaction status, tableid, columnid to backend protocol (Tom)
30 Add new binary I/O protocol (Tom)
31 Remove autocommit server setting; move to client applications (Tom)
32 New error message wording, error codes, and three levels of error detail (Tom)
34 _________________________________________________________________
38 Add hashing for GROUP BY aggregates (Tom)
39 Allow nested loops to be smarter about multicolumn indexes (Tom)
40 Allow multi-key hash joins (Tom)
41 Improve constant folding (Tom)
42 Add ability to inline simple SQL functions (Tom)
43 Reduce memory usage for queries using complex functions (Tom)
44 Improve GEQO optimizer performance (Tom)
45 Allow IN/NOT IN to be handled via hash tables (Tom)
46 Improve NOT IN (subquery) performance (Tom)
47 Allow most IN subqueries to be processed as joins (Tom)
48 Allow the postmaster to preload libraries using preload_libraries (Joe)
49 Improve optimizer cost computations, particularly for subqueries (Tom)
50 Avoid sort when subquery ORDER BY matches upper query (Tom)
51 Assume WHERE a.x = b.y and b.y = 42 also means a.x = 42 (Tom)
52 Allow hash/merge joins on complex joins (Tom)
53 Allow hash joins for more data types (Tom)
54 Allow join optimization of ANSI joins, disable with join_collapse_limit (Tom)
55 Add from_collapse_limit to control conversion of subqueries to joins (Tom)
56 Use faster regex code from TCL (Henry Spencer, Tom)
57 Use bit-mapped relation sets in the optimizer (Tom)
58 Improve backend startup time (Tom)
59 Improve trigger/constraint performance (Stephan)
61 _________________________________________________________________
65 Rename server parameter server_min_messages to log_min_messages (Bruce)
66 Rename show_*_stats to log_*_stats (Bruce)
67 Rename show_source_port to log_source_port (Bruce)
68 Rename hostname_lookup to log_hostname (Bruce)
69 Add checkpoint_warning to warn of excessive checkpointing (Bruce)
70 New read-only server parameters for localization (Tom)
71 Change debug server log messages to output as DEBUG rather than LOG (Bruce)
72 Prevent server log variables from being turned off by non-super users (Bruce)
73 log_min_messages/client_min_messages now controls debug_* output (Bruce)
74 Add Rendezvous server support (Chris Campbell)
75 Add ability to print only slow statements using log_min_duration_statement
77 Allow pg_hba.conf to accept netmasks in CIDR format (Andrew Dunstan)
78 New is_superuser read-only variable (Tom)
79 New server-side parameter log_error_verbosity to control error detail (Tom)
80 postgres --help-config now dumps server config variables (Aizaz Ahmed)
81 Make default shared_buffers 1000 and max_connections 100, if possible (Tom)
82 Add new columns in pg_settings: context, type, source , min_val, max_val (Joe)
83 New pg_hba.conf 'hostnossl' to prevent SSL connections (Jon Jensen)
85 _________________________________________________________________
89 New SQL-standard information schema (Peter)
90 Add read-only transactions (Peter)
91 Add server variable regex_flavor to control regular expression
93 Print key name and value in foreign-key violation messages (Dmitry Tkach)
94 Allow users to see their own queries in pg_stat_activity (Kevin Brown)
95 Fix subquery aggregates of upper query columns to match SQL spec. (Tom)
96 Add option to prevent auto-addition of tables referenced in query (Nigel J.
98 Allow UPDATE ... SET col = DEFAULT (Rod)
99 Allow expressions to be used in LIMIT/OFFSET (Tom)
100 Change EXECUTE INTO to CREATE TABLE AS EXECUTE (Peter)
102 _________________________________________________________________
106 Make CREATE SEQUENCE grammar more SQL1999 standards compliant (Neil)
107 Add FOR EACH STATEMENT statement-level triggers (Neil Conway)
108 Add DOMAIN CHECK constraints (Rod)
109 Add ALTER DOMAIN .. SET / DROP NOT NULL, SET / DROP DEFAULT, ADD / DROP
111 Fix several zero-column table bugs (Tom)
112 Have ALTER TABLE ... ADD PRIMARY KEY add NOT NULL constraint (Rod)
113 Add ALTER DOMAIN OWNER (Rod)
114 Add ALTER TABLE ... WITHOUT OIDS? (Rod)
115 Add ALTER SEQUENCE to modify min/max/increment/cache/cycle values (Rod)
116 Add ALTER TABLE ... CLUSTER ON (Alvaro Herrera)
117 Improve DOMAIN automatic type casting (Rod, Tom)
118 Allow dollar signs in identifiers, except as first character (Tom)
119 Disallow dollar signs in operator names, so x=$1 works (Tom)
120 Allow SQL200X inheritance syntax LIKE <subtable>, INCLUDING DEFAULTS (Rod)
121 Add WITH GRANT OPTION clause to GRANT, per SQL spec (Peter)
123 _________________________________________________________________
127 Add ON COMMIT PRESERVE ROWS for temp tables (Gavin)
128 Allow cursors outside transactions using WITH HOLD (Neil)
129 Make MOVE/FETCH 0 actually move/fetch 0 (Bruce)
130 Cause FETCH 1 to return the current cursor row, or zero if at
131 beginning/end of cursor, per SQL spec (Bruce)
132 Have MOVE return 0 or 1 depending on cursor position (Bruce)
133 Properly handle SCROLL with cursors, or report an error (Neil)
134 Implement SQL92-compatible FIRST, LAST, ABSOLUTE n, RELATIVE n options
135 for FETCH and MOVE (Tom)
136 Allow EXPLAIN on DECLARE CURSOR (Tom)
137 Allow CLUSTER to use index marked as pre-clustered by default (Alvaro Herrera)
138 Allow CLUSTER to cluster all tables (Alvaro Herrera)
139 Prevent CLUSTER on partial indexes (Tom)
140 Allow \r and \r\n termination for COPY files (Bruce)
141 Disallow literal carriage return as a data value, backslash-carriage-return
142 and \r still allowed (Bruce)
143 COPY changes (binary, \.)? (Tom)
144 Recover from COPY IN/OUT failure cleanly (Tom)
145 Prevent possible memory leaks in COPY (Tom)
146 Make TRUNCATE transaction-safe (Rod)
147 Multiple pg_dump fixes, including tar format and large objects
148 Allow pg_dump to dump specific schemas (Neil)
149 Allow pg_dump to preserve column storage characteristics (Christopher)
150 Allow pg_dump to preserve CLUSTER characteristics (Christopher)
151 Have pg_dumpall use GRANT/REVOKE to dump database-level permissions (Tom)
152 Allow pg_dumpall to support the -a, -s, -x options of pg_dump (Tom)
153 Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)
154 Allow PREPARE/bind of utility commands like FETCH and EXPLAIN (Tom)
155 Allow EXPLAIN EXECUTE (Neil)
156 Allow pg_get_constraintdef() to support UNIQUE, PRIMARY KEY and
157 CHECK constraints (Christopher)
158 Improve VACUUM performance on indexes by reducing WAL traffic (Tom)
159 Allow pg_ctl to better handle non-standard ports (Greg)
160 Functional indexes now support indexes on column expressions (Tom)
161 Syntax errors now reported as 'syntax error' rather than 'parse error' (Tom)
162 Have SHOW TRANSACTION_ISOLATION match input to SET TRANSACTION_ISOLATION (Tom)
163 Have COMMENT ON DATABASE on non-local database generate a warning (Rod)
165 _________________________________________________________________
167 Data Types and Functions
169 New extra_float_digits server parameter to control float precision display
170 (Pedro Ferreira, Tom)
171 Allow +1300 as a numeric timezone specifier, for FJST (Tom)
172 Remove rarely used oidrand(), oidsrand(), and userfntest() functions (Neil)
173 Add md5() function to main server, already in /contrib/pgcrypto (Joe)
174 Increase date range of timestamp (John Cochran)
175 Change EXTRACT(EPOCH FROM timestamp) so timestamp without time zone
176 is assumed to be in local time, not GMT (Tom)
177 Trap division by zero in case the operating system doesn't prevent it (Tom)
178 Change the NUMERIC data type internally to base 10000 (Tom)
179 New hostmask() function (Greg Wickham)
180 Fixes for to_char() (Karel)
181 Allow functions that can take any argument data type and return
182 any data type, using ANYELEMENT and ANYARRAY (Joe)
183 Arrays may now be specified as ARRAY[1,2,3], ARRAY[['a','b'],['c','d']],
184 or ARRAY[ARRAY[ARRAY[2]]] (Joe)
185 Allow proper comparisons for arrays (Joe)
186 Allow array concatenation with '||' (Joe)
187 Allow indexes on array columns, and used in ORDER BY and DISTINCT (Joe)
188 Allow WHERE qualification 'expr <oper> ANY/SOME/ALL (array-expr)' (Joe)
189 Allow polymorphic SQL functions (Joe)
190 New array functions array_append(), array_cat(), array_lower(),
191 array_prepend(), array_to_string(), array_upper(), string_to_array() (Joe)
192 Allow user defined aggregates to use polymorphic functions (Joe)
193 Allow polymorphic user defined aggregates (Joe)
194 Allow assignments to empty arrays (Joe)
195 Allow 60 in seconds fields of timestamp, time, interval input values (Tom)
196 Allow CIDR data type to be cast to text (Tom)
197 Allow the creation of special LIKE indexes for non-C locales (Peter)
198 Disallow invalid timezone names (Tom)
199 Trim trailing spaces when CHAR() data is cast to VARCHAR or TEXT (Tom)
200 Make FLOAT(p) measure the precision p in bits, not decimal digits (Tom)
201 Add IPv6 support to the inet and cidr data types (Michael Graff)
202 Add family() function to report whether address is IPv4 or IPv6 (Michael Graff)
203 Have SHOW DATESTYLE generate output similar to that used by SET DATESTYLE (Tom)
204 Change DATESTYLE to output its value in a more common format (Tom)
205 Make EXTRACT(TIMEZONE) and SET/SHOW TIMEZONE follow the SQL convention
206 for the sign of timezone offsets, ie, positive is east from UTC (Tom)
207 Fix date_trunc('quarter',...) (B?jthe Zolt?n)
208 Make initcap() more compatible with Oracle (Mike Nolan)
209 Allow only DateStyle field order for date values not in ISO format (Greg)
210 Add new DateStyle values MDY, DMY, and YMD, honor US and European for
211 backward compatibility (Tom)
212 'now' will no longer work as a column default, use now() (change required for
213 prepared statements) (Tom)
214 Assume NaN value to be larger than any other value in MIN()/MAX() (Tom)
215 Prevent interval from supressing ':00' seconds display
216 New pg_get_triggerdef(prettyprint) and pg_constraint_is_visible() functions
218 _________________________________________________________________
220 Server-side Languages
222 Prevent PL/pgSQL crash when RETURN NEXT is used on a zero-row record var. (Tom)
223 Make PL/python's spi_execute interface handle NULLs properly (Andrew Bosma)
224 Allow PL/pgSQL to declare variables of composite types without %ROWTYPE (Tom)
225 Fix PL/python _quote() function to handle big integers (?)
226 Make PL/python an untrusted language, now called plpythonu (Kevin Jacobs, Tom)
227 Allow polymorphic PL/pgSQL functions (Tom, Joe)
228 Improved compiled function caching mechanism in PL/pgSQL with full
229 support for polymorphism (Joe)
230 Add new $0 parameter in PL/pgSQL representing the function's actual
233 _________________________________________________________________
237 Add "\pset pager always" to always use pager (Greg)
238 Improve tab completion (Rod, Ross Reedstrom, Ian Barwick)
239 Reorder \? help into groupings (Harald Armin Massa, Bruce)
240 Add schema, cast, and conversion backslash commands (Christopher)
241 \encoding now tracks client_encoding server variable (Tom)
242 Save edit history into readline history (Ross)
243 Improve \d display (Christopher)
244 Enhance HTML mode to be more standards-compliant (Greg)
245 New '\set AUTOCOMMIT off' capability (Tom)
246 New '\set VERBOSITY' to control error detail (Tom)
247 New %T prompt string to show transaction status (Tom)
249 _________________________________________________________________
253 Allow PQcmdTuples() to return row counts for MOVE and FETCH (Neil)
254 Add PQfreemem() for freeing memory on Win32, suggest for NOTIFY (Bruce)
255 Document service capability, and add sample file (Bruce)
256 Make PQsetdbLogin() have the same defaults as PQconnectdb() (Tom)
257 Allow libpq to cleanly fail when result sets are too large (Tom)
258 Improve performance of PGunescapeBytea() (Ben Lamb)
259 Allow thread-safe libpq with --enable-thread-safety (Lee Kindness, Philip Yarra)
260 Allow pqInternalNotice() to accept a format string and args instead of
261 just a preformatted message (Tom, Sean Chittenden)
262 Allow control SSL negotiation with sslmode values "disable", "allow",
263 "Prefer", and "require" (Jon Jensen)
264 Allow new error codes and levels of text (Tom)
265 Allow access to the underlying table and column of a query result (Tom)
266 Allow access to the current transaction status (Tom)
267 Add ability to pass binary data directly to the backend (Tom)
269 _________________________________________________________________
273 Allow setNull on updateable resultsets
274 Allow executeBatch on a prepared statement (Barry)
275 Support SSL connections (Barry)
276 Handle schema names in result sets (Paul Sorenson)
277 Add refcursor support (Nic Ferrier)
279 _________________________________________________________________
281 Miscellaneous Interfaces
283 Prevent possible memory leak or core dump during libpgtcl shutdown (Tom)
284 Add ecpg Informix compatibility (Michael)
285 Add ecpg DECIMAL type that is fixed length, for Informix (Michael)
286 Allow thread-safe ecpg with --enable-thread-safety (Lee Kindness, Bruce)
287 Allow client interfaces to compile under MinGW/win32 (Bruce)
288 Move python client interface to http://www.pygresql.org (Marc)
290 _________________________________________________________________
294 Prevent need for separate platform geometry regression result files (Tom)
295 Improved PPC locking primitive (Reinhard Max)
296 Embed LD_LIBRARY_PATH used for build process into binaries (Billy)
297 New palloc0 to allocate and clear memory (Bruce)
298 Fix locking code for s390x CPU (64-bit) (Tom)
299 Allow OpenBSD to use local ident credentials (William Ahern)
300 Make query plan trees read-only to executor (Tom)
301 Add Darwin startup scripts (David Wheeler)
302 Allow libpq to compile with Borland C++ compiler (Lester Godwin, Karl Waclawek)
303 Use our own version of getopt_long() if needed (Peter)
304 Convert administration scripts to C (Peter)
305 Bison >=1.85 is now required for grammar changes
306 Merge documentation into one book (Peter)
307 Add Win32 compatibility functions (Bruce)
308 New ereport() function for error reporting (Tom)
309 Support Intel Linux compiler (Peter)
310 Improve Linux startup scripts (Slawomir Sudnik, Darko Prenosil)
311 Add support for AMD Opteron and Itanium (Jeffrey W. Baker, Bruce)
313 _________________________________________________________________
317 Change dbmirror license to BSD
318 Improve earthdistance (Bruno Wolff III)
319 Portability improvements to pgcrypto (Marko Kreen)
320 Prevent xml crash (John Gray, Michael Richards)
323 Update cube (Bruno Wolff III)
324 Update earthdistance to use cube (Bruno Wolff III)
325 Update btree_gist (Oleg)
326 New tsearch2 full-text search module (Oleg, Teodor)
327 Add hashed based crosstab function to tablefuncs (Joe)
328 Add serial column to order connectby() siblings in tablefuncs (Nabil Sayegh,Joe)
329 Add named persistent connections to dblink (Shridhar Daithanka)
330 New pg_autovacuum allows automatic VACUUM (Matthew T. O'Connor)
331 Allow pgbench to honor PGHOST, PGPORT, PGUSER env. variables (Tatsuo)
332 Improve intarray (Teodor Sigaev)
333 Improve pgstattuple (Rod)
334 Fix bug in metaphone() in fuzzystrmatch
335 Improve adddepend (Rod)
336 Update spi/timetravel (B?jthe Zolt?n)
337 Fix dbase -s option (Thomas Behr)
339 _________________________________________________________________
343 Release date: 2003-05-22
345 This has a variety of fixes from 7.3.2.
346 _________________________________________________________________
348 Migration to version 7.3.3
350 A dump/restore is *not* required for those running 7.3.*.
351 _________________________________________________________________
355 Repair sometimes-incorrect computation of StartUpID after a crash
356 Avoid slowness with lots of deferred triggers in one transaction (Stephan)
357 Don't lock referenced row when UPDATE doesn't change foreign key's value
359 Use -fPIC not -fpic on Sparc (Tom Callaway)
360 Repair lack of schema-awareness in contrib/reindexdb
361 Fix contrib/intarray error for zero-element result array (Teodor)
362 Ensure createuser script will exit on control-C (Oliver)
363 Fix errors when the type of a dropped column has itself been dropped
364 CHECKPOINT does not cause database panic on failure in noncritical steps
365 Accept 60 in seconds fields of timestamp, time, interval input values
366 Issue notice, not error, if TIMESTAMP, TIME, or INTERVAL precision too
368 Fix abstime-to-time cast function (fix is not applied unless you initdb)
369 Fix pg_proc entry for timestamptz_izone (fix is not applied unless you
371 Make EXTRACT(EPOCH FROM timestamp without time zone) treat input as
373 'now'::timestamptz gave wrong answer if timezone changed earlier in
375 HAVE_INT64_TIMESTAMP code for time with timezone overwrote its input
376 Accept GLOBAL TEMP/TEMPORARY as a synonym for TEMPORARY
377 Avoid improper schema-permissions-check failure in foreign-key triggers
378 Fix bugs in foreign-key triggers for SET DEFAULT action
379 Fix incorrect time-qual check in row fetch for UPDATE and DELETE triggers
380 Foreign-key clauses were parsed but ignored in ALTER TABLE ADD COLUMN
381 Fix createlang script breakage for case where handler function already
383 Fix misbehavior on zero-column tables in pg_dump, COPY, ANALYZE, other
385 Fix misbehavior of func_error() on type names containing '%'
386 Fix misbehavior of replace() on strings containing '%'
387 Regular-expression patterns containing certain multibyte characters failed
388 Account correctly for NULLs in more cases in join size estimation
389 Avoid conflict with system definition of isblank() function or macro
390 Fix failure to convert large code point values in EUC_TW conversions
392 Fix error recovery for SSL_read/SSL_write calls
393 Don't do early constant-folding of type coercion expressions
394 Validate page header fields immediately after reading in any page
395 Repair incorrect check for ungrouped variables in unnamed joins
396 Fix buffer overrun in to_ascii (Guido Notari)
397 contrib/ltree fixes (Teodor)
398 Fix core dump in deadlock detection on machines where char is unsigned
399 Avoid running out of buffers in many-way indexscan (bug introduced in 7.3)
400 Fix planner's selectivity estimation functions to handle domains properly
401 Fix dbmirror memory-allocation bug (Steven Singer)
402 Prevent infinite loop in ln(numeric) due to roundoff error.
403 GROUP BY got confused if there were multiple equal GROUP BY items
404 Fix bad plan when inherited UPDATE/DELETE references another inherited
406 Prevent clustering on incomplete (partial or non-NULL-storing) indexes
407 Service shutdown request at proper time if it arrives while still
409 Fix left-links in temporary indexes (could make backwards scans miss
411 Fix incorrect handling of client_encoding setting in postgresql.conf
413 Fix failure to respond to 'pg_ctl stop -m fast' after Async_NotifyHandler
415 Fix SPI for case where rule contains multiple statements of the same type
416 Fix problem with checking for wrong type of access permission in rule query
417 Fix problem with EXCEPT in CREATE RULE
418 Prevent problem with dropping temp tables having serial columns
419 Fix replace_vars_with_subplan_refs failure in complex views
420 Fix regexp slowness in single-byte encodings (Tatsuo)
421 Allow qualified type names in CREATE CAST and DROP CAST
422 Accept 'SETOF type[]', which formerly had to be written 'SETOF _type'
423 Fix pg_dump core dump in some cases with procedural languages
424 Force ISO datestyle in pg_dump output, for portability (Oliver)
425 pg_dump failed to handle error return from lo_read (Oleg Drokin)
426 pg_dumpall failed with groups having no members (Nick Eskelinen)
427 pg_dumpall failed to recognize --globals-only switch
428 pg_restore failed to restore blobs if -X disable-triggers is specified
429 Repair intrafunction memory leak in plpgsql
430 pltcl's elog command dumped core if given wrong parameters (Ian Harding)
431 plpython used wrong value of atttypmod (Brad McLean)
432 Fix improper quoting of boolean values in Python interface (D'Arcy)
433 Added addDataType() method to PGConnection interface for JDBC
434 Fixed various problems with updateable ResultSets for JDBC (Shawn Green)
435 Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka,
437 Fixed problem with parsing table ACLs in JDBC
438 Better error message for character set conversion problems in JDBC
439 _________________________________________________________________
443 Release date: 2003-02-04
445 This has a variety of fixes from 7.3.1.
446 _________________________________________________________________
448 Migration to version 7.3.2
450 A dump/restore is *not* required for those running 7.3.*.
451 _________________________________________________________________
455 Restore creation of OID column in CREATE TABLE AS / SELECT INTO
456 Fix pg_dump core dump when dumping views having comments
457 Dump DEFERRABLE/INITIALLY DEFERRED constraints properly
458 Fix UPDATE when child table's column numbering differs from parent
459 Increase default value of max_fsm_relations
460 Fix problem when fetching backwards in a cursor for a single-row query
461 Make backward fetch work properly with cursor on SELECT DISTINCT query
462 Fix problems with loading pg_dump files containing contrib/lo usage
463 Fix problem with all-numeric user names
464 Fix possible memory leak and core dump during disconnect in libpgtcl
465 Make plpython's spi_execute command handle nulls properly (Andrew Bosma)
466 Adjust plpython error reporting so that its regression test passes again
467 Work with bison 1.875
468 Handle mixed-case names properly in plpgsql's %type (Neil)
469 Fix core dump in pltcl when executing a query rewritten by a rule
470 Repair array subscript overruns (per report from Yichen Xie)
471 Reduce MAX_TIME_PRECISION from 13 to 10 in floating-point case
472 Correctly case-fold variable names in per-database and per-user settings
473 Fix coredump in plpgsql's RETURN NEXT when SELECT into record returns
475 Fix outdated use of pg_type.typprtlen in python client interface
476 Correctly handle fractional seconds in timestamps in JDBC driver
477 Improve performance of getImportedKeys() in JDBC
478 Make shared-library symlinks work standardly on HPUX (Giles)
479 Repair inconsistent rounding behavior for timestamp, time, interval
480 SSL negotiation fixes (Nathan Mueller)
481 Make libpq's ~/.pgpass feature work when connecting with PQconnectDB
484 Add casts between types lo and oid in contrib/lo
485 fastpath code now checks for privilege to call function
486 _________________________________________________________________
490 Release date: 2002-12-18
492 This has a variety of fixes from 7.3.
493 _________________________________________________________________
495 Migration to version 7.3.1
497 A dump/restore is *not* required for those running 7.3. However, it
498 should be noted that the main PostgreSQL interface library, libpq, has
499 a new major version number for this release, which may require
500 recompilation of client code in certain cases.
501 _________________________________________________________________
505 Fix a core dump of COPY TO when client/server encodings don't match (Tom)
506 Allow pg_dump to work with pre-7.2 servers (Philip)
507 /contrib/adddepend fixes (Tom)
508 Fix problem with deletion of per-user/per-database config settings (Tom)
509 /contrib/vacuumlo fix (Tom)
510 Allow 'password' encryption even when pg_shadow contains MD5 passwords
512 /contrib/dbmirror fix (Steven Singer)
513 Optimizer fixes (Tom)
514 /contrib/tsearch fixes (Teodor Sigaev, Magnus)
515 Allow locale names to be mixed case (Nicolai Tufar)
516 Increment libpq library's major version number (Bruce)
517 pg_hba.conf error reporting fixes (Bruce, Neil)
518 Add SCO Openserver 5.0.4 as a supported platform (Bruce)
519 Prevent EXPLAIN from crashing server (Tom)
520 SSL fixes (Nathan Mueller)
521 Prevent composite column creation via ALTER TABLE (Tom)
522 _________________________________________________________________
526 Release date: 2002-11-27
527 _________________________________________________________________
531 Major changes in this release:
534 Schemas allow users to create objects in separate namespaces,
535 so two people or applications can have tables with the same
536 name. There is also a public schema for shared tables.
537 Table/index creation can be restricted by removing permissions
538 on the public schema.
541 PostgreSQL now supports the ALTER TABLE ... DROP COLUMN
545 Functions returning multiple rows and/or multiple columns are
546 now much easier to use than before. You can call such a "table
547 function" in the SELECT FROM clause, treating its output like a
548 table. Also, PL/pgSQL functions can now return sets.
551 PostgreSQL now supports prepared queries, for improved
555 PostgreSQL now records object dependencies, which allows
556 improvements in many areas. "DROP" statements now take either
557 CASCADE or RESTRICT to control whether dependent objects are
561 Functions and procedural languages now have privileges, and
562 functions can be defined to run with the privileges of their
566 Both multibyte and locale support are now always enabled.
569 A variety of logging options have been enhanced.
572 A large number of interfaces have been moved to
573 http://gborg.postgresql.org where they can be developed and
574 released independently.
576 Functions/Identifiers
577 By default, functions can now take up to 32 parameters, and
578 identifiers can be up to 63 bytes long. Also, OPAQUE is now
579 deprecated: there are specific "pseudo-datatypes" to represent
580 each of the former meanings of OPAQUE in function argument and
582 _________________________________________________________________
584 Migration to version 7.3
586 A dump/restore using pg_dump is required for those wishing to migrate
587 data from any previous release. If your application examines the
588 system catalogs, additional changes will be required due to the
589 introduction of schemas in 7.3; for more information, see:
590 http://www.ca.postgresql.org/docs/momjian/upgrade_tips_7.3.
592 Observe the following incompatibilities:
594 * Pre-6.3 clients are no longer supported.
595 * "pg_hba.conf" now has a column for the user name and additional
596 features. Existing files need to be adjusted.
597 * Several "postgresql.conf" logging parameters have been renamed.
598 * LIMIT #,# has been disabled; use LIMIT # OFFSET #.
599 * "INSERT" statements with column lists must specify a value for
600 each specified column. For example, INSERT INTO tab (col1, col2)
601 VALUES ('val1') is now invalid. It's still allowed to supply fewer
602 columns than expected if the "INSERT" does not have a column list.
603 * serial columns are no longer automatically UNIQUE; thus, an index
604 will not automatically be created.
605 * A "SET" command inside an aborted transaction is now rolled back.
606 * "COPY" no longer considers missing trailing columns to be null.
607 All columns need to be specified. (However, one may achieve a
608 similar effect by specifying a column list in the "COPY" command.)
609 * The data type timestamp is now equivalent to timestamp without
610 time zone, instead of timestamp with time zone.
611 * Pre-7.3 databases loaded into 7.3 will not have the new object
612 dependencies for serial columns, unique constraints, and foreign
613 keys. See the directory "contrib/adddepend/" for a detailed
614 description and a script that will add such dependencies.
615 * An empty string ('') is no longer allowed as the input into an
616 integer field. Formerly, it was silently interpreted as 0.
617 _________________________________________________________________
623 Add pg_locks view to show locks (Neil)
624 Security fixes for password negotiation memory allocation (Neil)
625 Remove support for version 0 FE/BE protocol (PostgreSQL 6.2 and
627 Reserve the last few backend slots for superusers, add parameter
628 superuser_reserved_connections to control this (Nigel J. Andrews)
629 _________________________________________________________________
633 Improve startup by calling localtime() only once (Tom)
634 Cache system catalog information in flat files for faster startup
636 Improve caching of index information (Tom)
637 Optimizer improvements (Tom, Fernando Nasser)
638 Catalog caches now store failed lookups (Tom)
639 Hash function improvements (Neil)
640 Improve performance of query tokenization and network handling (Peter)
641 Speed improvement for large object restore (Mario Weilguni)
642 Mark expired index entries on first lookup, saving later heap fetches
644 Avoid excessive NULL bitmap padding (Manfred Koizar)
645 Add BSD-licensed qsort() for Solaris, for performance (Bruce)
646 Reduce per-row overhead by four bytes (Manfred Koizar)
647 Fix GEQO optimizer bug (Neil Conway)
648 Make WITHOUT OID actually save four bytes per row (Manfred Koizar)
649 Add default_statistics_target variable to specify ANALYZE buckets
651 Use local buffer cache for temporary tables so no WAL overhead (Tom)
652 Improve free space map performance on large tables (Stephen Marshall,
654 Improved WAL write concurrency (Tom)
655 _________________________________________________________________
659 Add privileges on functions and procedural languages (Peter)
660 Add OWNER to CREATE DATABASE so superusers can create databases
661 on behalf of unprivileged users. (Gavin Sherry, Tom)
662 Add new object permission bits EXECUTE and USAGE (Tom)
663 Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION
665 Allow functions to be executed with the privilege of the function
667 _________________________________________________________________
671 Server log messages now tagged with LOG, not DEBUG (Bruce)
672 Add user column to pg_hba.conf (Bruce)
673 Have log_connections output two lines in log file (Tom)
674 Remove debug_level from postgresql.conf, now server_min_messages
676 New ALTER DATABASE/USER ... SET command for per-user/database
677 initialization (Peter)
678 New parameters server_min_messages and client_min_messages to
679 control which messages are sent to the server logs or client
681 Allow pg_hba.conf to specify lists of users/databases separated by
682 commas, group names prepended with +, and file names prepended
684 Remove secondary password file capability and pg_password utility
686 Add variable db_user_namespace for database-local user names (Bruce)
687 SSL improvements (Bear Giles)
688 Make encryption of stored passwords the default (Bruce)
689 Allow pg_statistics to be reset by calling pg_stat_reset()
691 Add log_duration parameter (Bruce)
692 Rename debug_print_query to log_statement (Bruce)
693 Rename show_query_stats to show_statement_stats (Bruce)
694 Add param log_min_error_statement to print commands to logs on error
696 _________________________________________________________________
700 Make cursors insensitive, meaning their contents do not change (Tom)
701 Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)
702 Increase identifier length to 63 (Neil, Bruce)
703 UNION fixes for merging >= 3 columns of different lengths (Tom)
704 Add DEFAULT keyword to INSERT, e.g., INSERT ... (..., DEFAULT, ...)
706 Allow views to have default values using ALTER COLUMN ... SET DEFAULT
708 Fail on INSERTs with column lists that don't supply all column
709 values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)
710 Fix for join aliases (Tom)
711 Fix for FULL OUTER JOINs (Tom)
712 Improve reporting of invalid identifier and location (Tom, Gavin)
713 Fix OPEN cursor(args) (Tom)
714 Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)
715 Fix for CREATE TABLE AS with UNION (Tom)
716 SQL99 syntax improvements (Thomas)
717 Add statement_timeout variable to cancel queries (Bruce)
718 Allow prepared queries with PREPARE/EXECUTE (Neil)
719 Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)
720 Add variable autocommit (Tom, David Van Wie)
721 _________________________________________________________________
725 Make equals signs optional in CREATE DATABASE (Gavin Sherry)
726 Make ALTER TABLE OWNER change index ownership too (Neil)
727 New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls
728 TOAST storage, compression (John Gray)
729 Add schema support, CREATE/DROP SCHEMA (Tom)
730 Create schema for temporary tables (Tom)
731 Add variable search_path for schema search (Tom)
732 Add ALTER TABLE SET/DROP NOT NULL (Christopher)
733 New CREATE FUNCTION volatility levels (Tom)
734 Make rule names unique only per table (Tom)
735 Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)
736 Add ALTER TRIGGER RENAME (Joe)
737 New current_schema() and current_schemas() inquiry functions (Tom)
738 Allow functions to return multiple rows (table functions) (Joe)
739 Make WITH optional in CREATE DATABASE, for consistency (Bruce)
740 Add object dependency tracking (Rod, Tom)
741 Add RESTRICT/CASCADE to DROP commands (Rod)
742 Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)
743 Autodestroy sequence on DROP of table with SERIAL (Rod)
744 Prevent column dropping if column is used by foreign key (Rod)
745 Automatically drop constraints/functions when object is dropped (Rod)
746 Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)
747 Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)
748 Prevent inherited columns from being removed or renamed (Alvaro
750 Fix foreign key constraints to not error on intermediate database
752 Propagate column or table renaming to foreign key constraints
753 Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)
754 Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)
755 Have rules execute alphabetically, returning more predictable values
757 Triggers are now fired in alphabetical order (Tom)
758 Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)
759 Allow better casting when inserting/updating values (Tom)
760 _________________________________________________________________
764 Have COPY TO output embedded carriage returns and newlines as \r and
766 Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)
767 Make pg_dump use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)
768 Disable brackets in multistatement rules (Bruce)
769 Disable VACUUM from being called inside a function (Bruce)
770 Allow dropdb and other scripts to use identifiers with spaces (Bruce)
771 Restrict database comment changes to the current database
772 Allow comments on operators, independent of the underlying function
774 Rollback SET commands in aborted transactions (Tom)
775 EXPLAIN now outputs as a query (Tom)
776 Display condition expressions and sort keys in EXPLAIN (Tom)
777 Add 'SET LOCAL var = value' to set configuration variables for a
778 single transaction (Tom)
779 Allow ANALYZE to run in a transaction (Bruce)
780 Improve COPY syntax using new WITH clauses, keep backward
781 compatibility (Bruce)
782 Fix pg_dump to consistently output tags in non-ASCII dumps (Bruce)
783 Make foreign key constraints clearer in dump file (Rod)
784 Add COMMENT ON CONSTRAINT (Rod)
785 Allow COPY TO/FROM to specify column names (Brent Verner)
786 Dump UNIQUE and PRIMARY KEY constraints as ALTER TABLE (Rod)
787 Have SHOW output a query result (Joe)
788 Generate failure on short COPY lines rather than pad NULLs (Neil)
789 Fix CLUSTER to preserve all table attributes (Alvaro Herrera)
790 New pg_settings table to view/modify GUC settings (Joe)
791 Add smart quoting, portability improvements to pg_dump output (Peter)
792 Dump serial columns out as SERIAL (Tom)
793 Enable large file support, >2G for pg_dump (Peter, Philip Warner,
795 Disallow TRUNCATE on tables that are involved in referential
797 Have TRUNCATE also auto-truncate the toast table of the relation (Tom)
798 Add clusterdb utility that will auto-cluster an entire database
799 based on previous CLUSTER operations (Alvaro Herrera)
800 Overhaul pg_dumpall (Peter)
801 Allow REINDEX of TOAST tables (Tom)
802 Implemented START TRANSACTION, per SQL99 (Neil)
803 Fix rare index corruption when a page split affects bulk delete (Tom)
804 Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)
805 _________________________________________________________________
807 Data Types and Functions
809 Fix factorial(0) to return 1 (Bruce)
810 Date/time/timezone improvements (Thomas)
811 Fix for array slice extraction (Tom)
812 Fix extract/date_part to report proper microseconds for timestamp
814 Allow text_substr() and bytea_substr() to read TOAST values more
815 efficiently (John Gray)
816 Add domain support (Rod)
817 Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types
819 Allow alternate storage scheme of 64-bit integers for date/time types
820 using --enable-integer-datetimes in configure (Thomas)
821 Make timezone(timestamptz) return timestamp rather than a string
823 Allow fractional seconds in date/time types for dates prior to 1BC
825 Limit timestamp data types to 6 decimal places of precision (Thomas)
826 Change timezone conversion functions from timetz() to timezone()
828 Add configuration variables datestyle and timezone (Tom)
829 Add OVERLAY(), which allows substitution of a substring in a string
831 Add SIMILAR TO (Thomas, Tom)
832 Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)
833 Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)
834 Add named composite types using CREATE TYPE typename AS (column)
836 Allow composite type definition in the table alias clause (Joe)
837 Add new API to simplify creation of C language table functions (Joe)
838 Remove ODBC-compatible empty parentheses from calls to SQL99
839 functions for which these parentheses do not match the standard
841 Allow macaddr data type to accept 12 hex digits with no separators
843 Add CREATE/DROP CAST (Peter)
844 Add IS DISTINCT FROM operator (Thomas)
845 Add SQL99 TREAT() function, synonym for CAST() (Thomas)
846 Add pg_backend_pid() to output backend pid (Bruce)
847 Add IS OF / IS NOT OF type predicate (Thomas)
848 Allow bit string constants without fully-specified length (Thomas)
849 Allow conversion between 8-byte integers and bit strings (Thomas)
850 Implement hex literal conversion to bit string literal (Thomas)
851 Allow table functions to appear in the FROM clause (Joe)
852 Increase maximum number of function parameters to 32 (Bruce)
853 No longer automatically create index for SERIAL column (Tom)
854 Add current_database() (Rod)
855 Fix cash_words() to not overflow buffer (Tom)
856 Add functions replace(), split_part(), to_hex() (Joe)
857 Fix LIKE for bytea as a right-hand argument (Joe)
858 Prevent crashes caused by SELECT cash_out(2) (Tom)
859 Fix to_char(1,'FM999.99') to return a period (Karel)
860 Fix trigger/type/language functions returning OPAQUE to return
862 _________________________________________________________________
866 Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese
867 (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC)
869 Enable locale support by default (Peter)
870 Add locale variables (Peter)
871 Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea
873 Add locale awareness to regular expression character classes
874 Enable multibyte support by default (Tatso)
875 Add GB18030 multibyte support (Bill Huang)
876 Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo,
878 Add pg_conversion table (Tatsuo)
879 Add SQL99 CONVERT() function (Tatsuo)
880 pg_dumpall, pg_controldata, and pg_resetxlog now national-language
882 New and updated translations
883 _________________________________________________________________
885 Server-side Languages
887 Allow recursive SQL function (Peter)
888 Change PL/Tcl build to use configured compiler and Makefile.shlib
890 Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible
892 Allow PL/pgSQL to handle quoted identifiers (Tom)
893 Allow set-returning PL/pgSQL functions (Neil)
894 Make PL/pgSQL schema-aware (Joe)
895 Remove some memory leaks (Nigel J. Andrews, Tom)
896 _________________________________________________________________
900 Don't lowercase psql \connect database name for 7.2.0 compatibility
902 Add psql \timing to time user queries (Greg Sabino Mullane)
903 Have psql \d show index information (Greg Sabino Mullane)
904 New psql \dD shows domains (Jonathan Eisler)
905 Allow psql to show rules on views (Paul ?)
906 Fix for psql variable substitution (Tom)
907 Allow psql \d to show temporary table structure (Tom)
908 Allow psql \d to show foreign keys (Rod)
909 Fix \? to honor \pset pager (Bruce)
910 Have psql reports its version number on startup (Tom)
911 Allow \copy to specify column names (Tom)
912 _________________________________________________________________
916 Add $HOME/.pgpass to store host/user password combinations (Alvaro
918 Add PQunescapeBytea() function to libpq (Patrick Welche)
919 Fix for sending large queries over non-blocking connections
921 Fix for libpq using timers on Win9X (David Ford)
922 Allow libpq notify to handle servers with different-length
924 Add libpq PQescapeString() and PQescapeBytea() to Win32 (Bruce)
925 Fix for SSL with non-blocking connections (Jack Bates)
926 Add libpq connection timeout parameter (Denis A Ustimenko)
927 _________________________________________________________________
931 Allow JDBC to compile with JDK 1.4 (Dave)
932 Add JDBC 3 support (Barry)
933 Allows JDBC to set loglevel by adding ?loglevel=X to the connection
935 Add Driver.info() message that prints out the version number (Barry)
936 Add updateable result sets (Raghu Nidagal, Dave)
937 Add support for callable statements (Paul Bethe)
938 Add query cancel capability
939 Add refresh row (Dave)
940 Fix MD5 encryption handling for multibyte servers (Jun Kawai)
941 Add support for prepared statements (Barry)
942 _________________________________________________________________
944 Miscellaneous Interfaces
946 Fixed ECPG bug concerning octal numbers in single quotes (Michael)
947 Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc,
949 Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)
950 Add libpgtcl connection close event (Gerhard Hintermayer)
951 Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc,
953 Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)
954 Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc,
956 Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc,
958 Remove src/bin/pgaccess from main tree, now at
959 http://www.pgaccess.org (Bruce)
960 Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer,
962 _________________________________________________________________
966 Fix for parallel make (Peter)
967 AIX fixes for linking Tcl (Andreas Zeugswetter)
968 Allow PL/Perl to build under Cygwin (Jason Tishler)
969 Improve MIPS compiles (Peter, Oliver Elphick)
970 Require Autoconf version 2.53 (Peter)
971 Require readline and zlib by default in configure (Peter)
972 Allow Solaris to use Intimate Shared Memory (ISM), for performance
973 (Scott Brunza, P.J. Josh Rovero)
974 Always enable syslog in compile, remove --enable-syslog option
976 Always enable multibyte in compile, remove --enable-multibyte option
978 Always enable locale in compile, remove --enable-locale option
980 Fix for Win9x DLL creation (Magnus Naeslund)
981 Fix for link() usage by WAL code on Win32, BeOS (Jason Tishler)
982 Add sys/types.h to c.h, remove from main files (Peter, Bruce)
983 Fix AIX hang on SMP machines (Tomoyuki Niijima)
984 AIX SMP hang fix (Tomoyuki Niijima)
985 Fix pre-1970 date handling on newer glibc libraries (Tom)
986 Fix PowerPC SMP locking (Tom)
987 Prevent gcc -ffast-math from being used (Peter, Tom)
988 Bison >= 1.50 now required for developer builds
989 Kerberos 5 support now builds with Heimdal (Peter)
990 Add appendix in the User's Guide which lists SQL features (Thomas)
991 Improve loadable module linking to use RTLD_NOW (Tom)
992 New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)
993 New src/port directory holds replaced libc functions (Peter, Bruce)
994 New pg_namespace system catalog for schemas (Tom)
995 Add pg_class.relnamespace for schemas (Tom)
996 Add pg_type.typnamespace for schemas (Tom)
997 Add pg_proc.pronamespace for schemas (Tom)
998 Restructure aggregates to have pg_proc entries (Tom)
999 System relations now have their own namespace, pg_* test not required
1001 Rename TOAST index names to be *_index rather than *_idx (Neil)
1002 Add namespaces for operators, opclasses (Tom)
1003 Add additional checks to server control file (Thomas)
1004 New Polish FAQ (Marcin Mazurek)
1005 Add Posix semaphore support (Tom)
1006 Document need for reindex (Bruce)
1007 Rename some internal identifiers to simplify Win32 compile (Jan,
1009 Add documentation on computing disk space (Bruce)
1010 Remove KSQO from GUC (Bruce)
1011 Fix memory leak in rtree (Kenneth Been)
1012 Modify a few error messages for consistency (Bruce)
1013 Remove unused system table columns (Peter)
1014 Make system columns NOT NULL where appropriate (Tom)
1015 Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)
1016 Remove OPAQUE and create specific subtypes (Tom)
1017 Cleanups in array internal handling (Joe, Tom)
1018 Disallow pg_atoi('') (Bruce)
1019 Remove parameter wal_files because WAL files are now recycled (Bruce)
1020 Add version numbers to heap pages (Tom)
1021 _________________________________________________________________
1025 Allow inet arrays in /contrib/array (Neil)
1026 Gist fixes (Teodor Sigaev, Neil)
1027 Upgrade /contrib/mysql
1028 Add /contrib/dbsize which shows table sizes without vacuum (Peter)
1029 Add /contrib/intagg, integer aggregator routines (mlw)
1030 Improve /contrib/oid2name (Neil, Bruce)
1031 Improve /contrib/tsearch (Oleg, Teodor Sigaev)
1032 Cleanups of /contrib/rserver (Alexey V. Borzov)
1033 Update /contrib/oracle conversion utility (Gilles Darold)
1034 Update /contrib/dblink (Joe)
1035 Improve options supported by /contrib/vacuumlo (Mario Weilguni)
1036 Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey
1038 Add /contrib/reindexdb utility (Shaun Thomas)
1039 Add indexing to /contrib/isbn_issn (Dan Weston)
1040 Add /contrib/dbmirror (Steven Singer)
1041 Improve /contrib/pgbench (Neil)
1042 Add /contrib/tablefunc table function examples (Joe)
1043 Add /contrib/ltree data type for tree structures (Teodor Sigaev,
1045 Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)
1046 Fixes to /contrib/cube (Bruno Wolff)
1047 Improve /contrib/fulltextindex (Christopher)
1048 _________________________________________________________________
1052 Release date: 2003-01-30
1054 This has a variety of fixes from 7.2.3, including fixes to prevent
1056 _________________________________________________________________
1058 Migration to version 7.2.4
1060 A dump/restore is *not* required for those running 7.2.X.
1061 _________________________________________________________________
1065 Fix some additional cases of VACUUM "No one parent tuple was found"
1067 Prevent VACUUM from being called inside a function (Bruce)
1068 Ensure pg_clog updates are sync'd to disk before marking checkpoint
1070 Avoid integer overflow during large hash joins
1071 Make GROUP commands work when pg_group.grolist is large enough
1073 Fix errors in datetime tables; some timezone names weren't being
1075 Fix integer overflows in circle_poly(), path_encode(), path_add()
1077 Repair long-standing logic errors in lseg_eq(), lseg_ne(), lseg_center()
1078 _________________________________________________________________
1082 Release date: 2002-10-01
1084 This has a variety of fixes from 7.2.2, including fixes to prevent
1086 _________________________________________________________________
1088 Migration to version 7.2.3
1090 A dump/restore is *not* required for those running 7.2.X.
1091 _________________________________________________________________
1095 Prevent possible compressed transaction log loss (Tom)
1096 Prevent non-superuser from increasing most recent vacuum info (Tom)
1097 Handle pre-1970 date values in newer versions of glibc (Tom)
1098 Fix possible hang during server shutdown
1099 Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)
1100 Fix pg_dump to properly dump FULL JOIN USING (Tom)
1101 _________________________________________________________________
1105 Release date: 2002-08-23
1107 This has a variety of fixes from 7.2.1.
1108 _________________________________________________________________
1110 Migration to version 7.2.2
1112 A dump/restore is *not* required for those running 7.2.X.
1113 _________________________________________________________________
1117 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
1118 Fix for compressed transaction log id wraparound (Tom)
1119 Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Ta
1121 Fix for psql and pg_dump crashing when invoked with non-existand long
1123 Fix crash when invoking geometric operators (Tom)
1124 Allow OPEN cursor(args) (Tom)
1125 Fix for rtree_gist index build (Teodor)
1126 Fix for dumping user-defined aggregates (Tom)
1127 Contrib/intarray fixes (Oleg)
1128 Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)
1129 Fix to pg_convert (Tatsuo)
1130 Fix for crash with long DATA strings (Thomes, Neil)
1131 Fix for repeat(), lpad(), rpad() and long strings (Neil)
1132 _________________________________________________________________
1136 Release date: 2002-03-21
1138 This has a variety of fixes from 7.2.
1139 _________________________________________________________________
1141 Migration to version 7.2.1
1143 A dump/restore is *not* required for those running 7.2.
1144 _________________________________________________________________
1148 Ensure that sequence counters do not go backwards after a crash (Tom)
1149 Fix pgaccess kanji-coversion key binding (Tatsuo)
1150 Optimizer improvements (Tom)
1151 Cash I/O improvements (Tom)
1153 Compile fix for missing AuthBlockSig (Heiko)
1154 Additional time zones and time zone fixes (Thomas)
1155 Allow psql \connect to handle mixed case database and user names (Tom)
1156 Return proper OID on command completion even with ON INSERT rules (Tom
1158 Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)
1159 Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)
1160 Improve handling of multiple UNIONs with different lengths (Tom)
1161 contrib/btree_gist improvements (Teodor Sigaev)
1162 contrib/tsearch dictionary improvements, see README.tsearch for
1163 an additional installation step (Thomas T. Thai, Teodor Sigaev)
1164 Fix for array subscripts handling (Tom)
1165 Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
1166 _________________________________________________________________
1170 Release date: 2002-02-04
1171 _________________________________________________________________
1175 This release improves PostgreSQL for use in high-volume applications.
1177 Major changes in this release:
1180 Vacuuming no longer locks tables, thus allowing normal user
1181 access during the vacuum. A new "VACUUM FULL" command does
1182 old-style vacuum by locking the table and shrinking the on-disk
1186 There is no longer a problem with installations that exceed
1187 four billion transactions.
1190 OIDs are now optional. Users can now create tables without OIDs
1191 for cases where OID usage is excessive.
1194 The system now computes histogram column statistics during
1195 "ANALYZE", allowing much better optimizer choices.
1198 A new MD5 encryption option allows more secure storage and
1199 transfer of passwords. A new Unix-domain socket authentication
1200 option is available on Linux and BSD systems.
1203 Administrators can use the new table access statistics module
1204 to get fine-grained information about table and index usage.
1206 Internationalization
1207 Program and library messages can now be displayed in several
1209 _________________________________________________________________
1211 Migration to version 7.2
1213 A dump/restore using "pg_dump" is required for those wishing to
1214 migrate data from any previous release.
1216 Observe the following incompatibilities:
1218 * The semantics of the "VACUUM" command have changed in this
1219 release. You may wish to update your maintenance procedures
1221 * In this release, comparisons using = NULL will always return false
1222 (or NULL, more precisely). Previous releases automatically
1223 transformed this syntax to IS NULL. The old behavior can be
1224 re-enabled using a "postgresql.conf" parameter.
1225 * The "pg_hba.conf" and "pg_ident.conf" configuration is now only
1226 reloaded after receiving a SIGHUP signal, not with each
1228 * The function "octet_length()" now returns the uncompressed data
1230 * The date/time value 'current' is no longer available. You will
1231 need to rewrite your applications.
1232 * The timestamp(), time(), and interval() functions are no longer
1233 available. Instead of timestamp(), use timestamp 'string' or CAST.
1235 The SELECT ... LIMIT #,# syntax will be removed in the next release.
1236 You should change your queries to use separate LIMIT and OFFSET
1237 clauses, e.g. LIMIT 10 OFFSET 20.
1238 _________________________________________________________________
1244 Create temporary files in a separate directory (Bruce)
1245 Delete orphaned temporary files on postmaster startup (Bruce)
1246 Added unique indexes to some system tables (Tom)
1247 System table operator reorganization (Oleg Bartunov, Teodor Sigaev, To
1249 Renamed pg_log to pg_clog (Tom)
1250 Enable SIGTERM, SIGQUIT to kill backends (Jan)
1251 Removed compile-time limit on number of backends (Tom)
1252 Better cleanup for semaphore resource failure (Tatsuo, Tom)
1253 Allow safe transaction ID wraparound (Tom)
1254 Removed OIDs from some system tables (Tom)
1255 Removed "triggered data change violation" error check (Tom)
1256 SPI portal creation of prepared/saved plans (Jan)
1257 Allow SPI column functions to work for system columns (Tom)
1258 Long value compression improvement (Tom)
1259 Statistics collector for table, index access (Jan)
1260 Truncate extra-long sequence names to a reasonable value (Tom)
1261 Measure transaction times in milliseconds (Thomas)
1262 Fix TID sequential scans (Hiroshi)
1263 Superuser ID now fixed at 1 (Peter E)
1264 New pg_ctl "reload" option (Tom)
1265 _________________________________________________________________
1269 Optimizer improvements (Tom)
1270 New histogram column statistics for optimizer (Tom)
1271 Reuse write-ahead log files rather than discarding them (Tom)
1272 Cache improvements (Tom)
1273 IS NULL, IS NOT NULL optimizer improvement (Tom)
1274 Improve lock manager to reduce lock contention (Tom)
1275 Keep relcache entries for index access support functions (Tom)
1276 Allow better selectivity with NaN and infinities in NUMERIC (Tom)
1277 R-tree performance improvements (Kenneth Been)
1278 B-tree splits more efficient (Tom)
1279 _________________________________________________________________
1283 Change UPDATE, DELETE permissions to be distinct (Peter E)
1284 New REFERENCES, TRIGGER privileges (Peter E)
1285 Allow GRANT/REVOKE to/from more than one user at a time (Peter E)
1286 New has_table_privilege() function (Joe Conway)
1287 Allow non-superuser to vacuum database (Tom)
1288 New SET SESSION AUTHORIZATION command (Peter E)
1289 Fix bug in privilege modifications on newly created tables (Tom)
1290 Disallow access to pg_statistic for non-superuser, add user-accessible
1292 _________________________________________________________________
1294 Client Authentication
1296 Fork postmaster before doing authentication to prevent hangs (Peter E)
1297 Add ident authentication over Unix domain sockets on Linux, *BSD (Helg
1298 e Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)
1299 Add a password authentication method that uses MD5 encryption (Bruce)
1300 Allow encryption of stored passwords using MD5 (Bruce)
1301 PAM authentication (Dominic J. Eidson)
1302 Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)
1303 _________________________________________________________________
1305 Server Configuration
1307 Interpretation of some time zone abbreviations as Australian rather th
1308 an North American now settable at run time (Bruce)
1309 New parameter to set default transaction isolation level (Peter E)
1310 New parameter to enable conversion of "expr = NULL" into "expr IS NULL
1311 ", off by default (Peter E)
1312 New parameter to control memory usage by VACUUM (Tom)
1313 New parameter to set client authentication timeout (Tom)
1314 New parameter to set maximum number of open files (Tom)
1315 _________________________________________________________________
1319 Statements added by INSERT rules now execute after the INSERT (Jan)
1320 Prevent unadorned relation names in target list (Bruce)
1321 NULLs now sort after all normal values in ORDER BY (Tom)
1322 New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)
1323 New SHARE UPDATE EXCLUSIVE lock mode (Tom)
1324 New EXPLAIN ANALYZE command that shows run times and row counts (Marti
1326 Fix problem with LIMIT and subqueries (Tom)
1327 Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)
1328 Fix nested EXCEPT/INTERSECT (Tom)
1329 _________________________________________________________________
1333 Fix SERIAL in temporary tables (Bruce)
1334 Allow temporary sequences (Bruce)
1335 Sequences now use int8 internally (Tom)
1336 New SERIAL8 creates int8 columns with sequences, default still SERIAL4
1338 Make OIDs optional using WITHOUT OIDS (Tom)
1339 Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)
1340 Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher K
1342 New CREATE OR REPLACE FUNCTION to alter existing function (preserving
1343 the function OID) (Gavin Sherry)
1344 Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)
1345 Allow column renaming in views
1346 Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent
1348 Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (
1350 ALTER TABLE RENAME update foreign-key trigger arguments correctly (Bre
1352 DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)
1353 Add automatic return type data casting for SQL functions (Tom)
1354 Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov
1355 , Teodor Sigaev, Tom)
1356 Enable partial indexes (Martijn van Oosterhout)
1357 _________________________________________________________________
1361 Add RESET ALL, SHOW ALL (Marko Kreen)
1362 CREATE/ALTER USER/GROUP now allow options in any order (Vince)
1363 Add LOCK A, B, C functionality (Neil Padgett)
1364 New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)
1365 New light-weight VACUUM does not lock table; old semantics are availab
1366 le as VACUUM FULL (Tom)
1367 Disable COPY TO/FROM on views (Bruce)
1368 COPY DELIMITERS string must be exactly one character (Tom)
1369 VACUUM warning about index tuples fewer than heap now only appears whe
1370 n appropriate (Martijn van Oosterhout)
1371 Fix permission checks for CREATE INDEX (Tom)
1372 Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)
1373 _________________________________________________________________
1375 Data Types and Functions
1377 SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)
1378 Add convert(), convert2() (Tatsuo)
1379 New function bit_length() (Peter E)
1380 Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tats
1382 CHAR(), VARCHAR() now reject strings that are too long (Peter E)
1383 BIT VARYING now rejects bit strings that are too long (Peter E)
1384 BIT now rejects bit strings that do not match declared size (Peter E)
1385 INET, CIDR text conversion functions (Alex Pilosov)
1386 INET, CIDR operators << and <<= indexable (Alex Pilosov)
1387 Bytea \### now requires valid three digit octal number
1388 Bytea comparison improvements, now supports =, <>, >, >=, <, and <=
1389 Bytea now supports B-tree indexes
1390 Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE
1391 Bytea now supports concatenation
1392 New bytea functions: position, substring, trim, btrim, and length
1393 New encode() function mode, "escaped", converts minimally escaped byte
1395 Add pg_database_encoding_max_length() (Tatsuo)
1396 Add pg_client_encoding() function (Tatsuo)
1397 now() returns time with millisecond precision (Thomas)
1398 New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)
1399 Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)
1400 New xid/int comparison functions (Hiroshi)
1401 Add precision to TIME, TIMESTAMP, and INVERVAL data types (Thomas)
1402 Modify type coercion logic to attempt binary-compatible functions firs
1404 New encode() function installed by default (Marko Kreen)
1405 Improved to_*() conversion functions (Karel Zak)
1406 Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)
1407 New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_sal
1409 Correct description of translate() function (Bruce)
1410 Add INTERVAL argument for SET TIME ZONE (Thomas)
1411 Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)
1412 Optimize length functions when using single-byte encodings (Tatsuo)
1413 Fix path_inter, path_distance, path_length, dist_ppath to handle close
1414 d paths (Curtis Barrett, Tom)
1415 octet_length(text) now returns non-compressed length (Tatsuo, Bruce)
1416 Handle "July" full name in date/time literals (Greg Sabino Mullane)
1417 Some datatype() function calls now evaluated differently
1418 Add support for Julian and ISO time specifications (Thomas)
1419 _________________________________________________________________
1421 Internationalization
1423 National language support in psql, pg_dump, libpq, and server (Peter E
1425 Message translations in Chinese (simplified, traditional), Czech, Fren
1426 ch, German, Hungarian, Russian, Swedish (Peter E, Serguei A. Mokhov, K
1427 arel Zak, Weiping He, Zhenbang Wei, Kovacs Zoltan)
1428 Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware
1430 Add LATIN5,6,7,8,9,10 support (Tatsuo)
1431 Add ISO 8859-5,6,7,8 support (Tatsuo)
1432 Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)
1433 Make mic2ascii() non-ASCII aware (Tatsuo)
1434 Reject invalid multibyte character sequences (Tatsuo)
1435 _________________________________________________________________
1439 Now uses portals for SELECT loops, allowing huge result sets (Jan)
1440 CURSOR and REFCURSOR support (Jan)
1441 Can now return open cursors (Jan)
1442 Add ELSEIF (Klaus Reger)
1443 Improve PL/pgSQL error reporting, including location of error (Tom)
1444 Allow IS or FOR key words in cursor declaration, for compatibility (Br
1446 Fix for SELECT ... FOR UPDATE (Tom)
1447 Fix for PERFORM returning multiple rows (Tom)
1448 Make PL/pgSQL use the server's type coercion code (Tom)
1449 Memory leak fix (Jan, Tom)
1450 Make trailing semicolon optional (Tom)
1451 _________________________________________________________________
1455 New untrusted PL/Perl (Alex Pilosov)
1456 PL/Perl is now built on some platforms even if libperl is not shared (
1458 _________________________________________________________________
1462 Now reports errorInfo (Vsevolod Lobko)
1463 Add spi_lastoid function (bob@redivi.com)
1464 _________________________________________________________________
1468 ...is new (Andrew Bosma)
1469 _________________________________________________________________
1473 \d displays indexes in unique, primary groupings (Christopher Kings-Ly
1475 Allow trailing semicolons in backslash commands (Greg Sabino Mullane)
1476 Read password from /dev/tty if possible
1477 Force new password prompt when changing user and database (Tatsuo, Tom
1479 Format the correct number of columns for Unicode (Patrice)
1480 _________________________________________________________________
1484 New function PQescapeString() to escape quotes in command strings (Flo
1486 New function PQescapeBytea() escapes binary strings for use as SQL str
1488 _________________________________________________________________
1492 Return OID of INSERT (Ken K)
1493 Handle more data types (Ken K)
1494 Handle single quotes and newlines in strings (Ken K)
1495 Handle NULL variables (Ken K)
1496 Fix for time zone handling (Barry Lind)
1497 Improved Druid support
1498 Allow eight-bit characters with non-multibyte server (Barry Lind)
1499 Support BIT, BINARY types (Ned Wolpert)
1500 Reduce memory usage (Michael Stephens, Dave Cramer)
1501 Update DatabaseMetaData (Peter E)
1502 Add DatabaseMetaData.getCatalogs() (Peter E)
1503 Encoding fixes (Anders Bengtsson)
1504 Get/setCatalog methods (Jason Davies)
1505 DatabaseMetaData.getColumns() now returns column defaults (Jason Davie
1507 DatabaseMetaData.getColumns() performance improvement (Jeroen van Vian
1509 Some JDBC1 and JDBC2 merging (Anders Bengtsson)
1510 Transaction performance improvements (Barry Lind)
1511 Array fixes (Greg Zoller)
1513 Fix batch processing (Rene Pijlman)
1514 ExecSQL method reorganization (Anders Bengtsson)
1515 GetColumn() fixes (Jeroen van Vianen)
1516 Fix isWriteable() function (Rene Pijlman)
1517 Improved passage of JDBC2 conformance tests (Rene Pijlman)
1518 Add bytea type capability (Barry Lind)
1519 Add isNullable() (Rene Pijlman)
1520 JDBC date/time test suite fixes (Liam Stewart)
1521 Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)
1522 Fix DatabaseMetaData to show precision properly (Mark Lillywhite)
1523 New getImported/getExported keys (Jason Davies)
1524 MD5 password encryption support (Jeremy Wohl)
1525 Fix to actually use type cache (Ned Wolpert)
1526 _________________________________________________________________
1530 Remove query size limit (Hiroshi)
1531 Remove text field size limit (Hiroshi)
1532 Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)
1533 Allow ODBC procedure calls (Hiroshi)
1534 Improve boolean handing (Aidan Mountford)
1535 Most configuration options on setable via DSN (Hiroshi)
1536 Multibyte, performance fixes (Hiroshi)
1537 Allow driver to be used with iODBC or unixODBC (Peter E)
1538 MD5 password encryption support (Bruce)
1539 Add more compatibility functions to odbc.sql (Peter E)
1540 _________________________________________________________________
1544 EXECUTE ... INTO implemented (Christof Petig)
1545 Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)
1546 Fix for GRANT parameters (Lee Kindness)
1547 Fix INITIALLY DEFERRED bug
1548 Various bug fixes (Michael, Christof Petig)
1549 Auto allocation for indicator variable arrays (int *ind_p=NULL)
1550 Auto allocation for string arrays (char **foo_pp=NULL)
1551 ECPGfree_auto_mem fixed
1552 All function names with external linkage are now prefixed by ECPG
1553 Fixes for arrays of structures (Michael)
1554 _________________________________________________________________
1558 Python fix fetchone() (Gerhard Haring)
1559 Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Ma
1561 Add Tcl COPY TO/FROM (ljb)
1562 Prevent output of default index op class in pg_dump (Tom)
1563 Fix libpgeasy memory leak (Bruce)
1564 _________________________________________________________________
1568 Configure, dynamic loader, and shared library fixes (Peter E)
1569 Fixes in QNX 4 port (Bernd Tegge)
1570 Fixes in Cygwin and Win32 ports (Jason Tishler, Gerhard Haring, Dmitry
1571 Yurtaev, Darko Prenosil, Mikhail Terekhov)
1572 Fix for Win32 socket communication failures (Magnus, Mikhail Terekhov)
1573 Hurd compile fix (Oliver Elphick)
1574 BeOS fixes (Cyril Velter)
1575 Remove configure --enable-unicode-conversion, now enabled by multibyte
1577 AIX fixes (Tatsuo, Andreas)
1578 Fix parallel make (Peter E)
1579 Install SQL language manual pages into OS-specific directories (Peter
1581 Rename config.h to pg_config.h (Peter E)
1582 Reorganize installation layout of header files (Peter E)
1583 _________________________________________________________________
1587 Remove SEP_CHAR (Bruce)
1589 Merge GUC and command line handling (Marko Kreen)
1590 Remove EXTEND INDEX (Martijn van Oosterhout, Tom)
1591 New pgjindent utility to indent java code (Bruce)
1592 Remove define of true/false when compiling under C++ (Leandro Fanzone,
1594 pgindent fixes (Bruce, Tom)
1595 Replace strcasecmp() with strcmp() where appropriate (Peter E)
1596 Dynahash portability improvements (Tom)
1597 Add 'volatile' usage in spinlock structures
1598 Improve signal handling logic (Tom)
1599 _________________________________________________________________
1603 New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)
1604 New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)
1605 Add contrib/dblink for remote database access (Joe Conway)
1606 contrib/ora2pg Oracle conversion utility (Gilles Darold)
1607 contrib/xml XML conversion utility (John Gray)
1608 contrib/fulltextindex fixes (Christopher Kings-Lynne)
1609 New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merg
1611 Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartu
1613 New pg_upgrade utility (Bruce)
1614 Add new pg_resetxlog options (Bruce, Tom)
1615 _________________________________________________________________
1619 Release date: 2001-08-15
1620 _________________________________________________________________
1622 Migration to version 7.1.3
1624 A dump/restore is *not* required for those running 7.1.X.
1625 _________________________________________________________________
1629 Remove unused WAL segements of large transactions (Tom)
1630 Multiaction rule fix (Tom)
1631 PL/pgSQL memory allocation fix (Jan)
1632 VACUUM buffer fix (Tom)
1633 Regression test fixes (Tom)
1634 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
1635 Fix subselects with DISTINCT ON or LIMIT (Tom)
1637 Disable COPY TO/FROM a view (Tom)
1638 Cygwin build (Jason Tishler)
1640 _________________________________________________________________
1644 Release date: 2001-05-11
1646 This has one fix from 7.1.1.
1647 _________________________________________________________________
1649 Migration to version 7.1.2
1651 A dump/restore is *not* required for those running 7.1.X.
1652 _________________________________________________________________
1656 Fix PL/pgSQL SELECTs when returning no rows
1657 Fix for psql backslash core dump
1658 Referential integrity permission fix
1662 _________________________________________________________________
1666 Release date: 2001-05-05
1668 This has a variety of fixes from 7.1.
1669 _________________________________________________________________
1671 Migration to version 7.1.1
1673 A dump/restore is *not* required for those running 7.1.
1674 _________________________________________________________________
1678 Fix for numeric MODULO operator (Tom)
1679 pg_dump fixes (Philip)
1680 pg_dump can dump 7.0 databases (Philip)
1681 readline 4.2 fixes (Peter E)
1683 AIX, MSWIN, VAX, N32K fixes (Tom)
1684 Multibytes fixes (Tom)
1685 Unicode fixes (Tatsuo)
1686 Optimizer improvements (Tom)
1687 Fix for whole tuples in functions (Tom)
1688 Fix for pg_ctl and option strings with spaces (Peter E)
1689 ODBC fixes (Hiroshi)
1690 EXTRACT can now take string argument (Thomas)
1691 Python fixes (Darcy)
1693 _________________________________________________________________
1697 Release date: 2001-04-13
1699 This release focuses on removing limitations that have existed in the
1700 PostgreSQL code for many years.
1702 Major changes in this release:
1704 Write-ahead Log (WAL)
1705 To maintain database consistency in case of an operating system
1706 crash, previous releases of PostgreSQL have forced all data
1707 modifications to disk before each transaction commit. With WAL,
1708 only one log file must be flushed to disk, greatly improving
1709 performance. If you have been using -F in previous releases to
1710 disable disk flushes, you may want to consider discontinuing
1714 TOAST - Previous releases had a compiled-in row length limit,
1715 typically 8k - 32k. This limit made storage of long text fields
1716 difficult. With TOAST, long rows of any length can be stored
1717 with good performance.
1720 We now support outer joins. The UNION/NOT IN workaround for
1721 outer joins is no longer required. We use the SQL92 outer join
1725 The previous C function manager did not handle null values
1726 properly, nor did it support 64-bit CPU's (Alpha). The new
1727 function manager does. You can continue using your old custom
1728 functions, but you may want to rewrite them in the future to
1729 use the new function manager call interface.
1732 A large number of complex queries that were unsupported in
1733 previous releases now work. Many combinations of views,
1734 aggregates, UNION, LIMIT, cursors, subqueries, and inherited
1735 tables now work properly. Inherited tables are now accessed by
1736 default. Subqueries in FROM are now supported.
1737 _________________________________________________________________
1739 Migration to version 7.1
1741 A dump/restore using pg_dump is required for those wishing to migrate
1742 data from any previous release.
1743 _________________________________________________________________
1749 Many multibyte/Unicode/locale fixes (Tatsuo and others)
1750 More reliable ALTER TABLE RENAME (Tom)
1751 Kerberos V fixes (David Wragg)
1752 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
1753 Prompt username/password on standard error (Bruce)
1754 Large objects inv_read/inv_write fixes (Tom)
1755 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
1757 Prevent query expressions from leaking memory (Tom)
1758 Allow UPDATE of arrays elements (Tom)
1759 Wake up lock waiters during cancel (Hiroshi)
1760 Fix rare cursor crash when using hash join (Tom)
1761 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
1762 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
1763 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
1764 Fix PL/perl (Alex Kapranoff)
1765 Disallow LOCK on views (Mark Hollomon)
1766 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
1767 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
1768 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
1769 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
1770 Fix rare failure with TRUNCATE command (Tom)
1771 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
1772 DISTINCT, ORDER BY, SELECT...INTO (Tom)
1773 Fix parser failures during aborted transactions (Tom)
1774 Allow temporary relations to properly clean up indexes (Bruce)
1775 Fix VACUUM problem with moving rows in same page (Tom)
1776 Modify pg_dump to better handle user-defined items in template1 (Philip)
1777 Allow LIMIT in VIEW (Tom)
1778 Require cursor FETCH to honor LIMIT (Tom)
1779 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
1780 Allow ORDER BY, LIMIT in sub-selects (Tom)
1781 Allow UNION in CREATE RULE (Tom)
1782 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
1783 Store initdb collation in pg_control so collation cannot be changed (Tom)
1784 Fix INSERT...SELECT with rules (Tom)
1785 Fix FOR UPDATE inside views and subselects (Tom)
1786 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
1787 Fix lpad() and rpad() to handle length less than input string (Tom)
1788 Fix use of NOTIFY in some rules (Tom)
1789 Overhaul btree code (Tom)
1790 Fix NOT NULL use in Pl/pgSQL variables (Tom)
1791 Overhaul GIST code (Oleg)
1792 Fix CLUSTER to preserve constraints and column default (Tom)
1793 Improved deadlock detection handling (Tom)
1794 Allow multiple SERIAL columns in a table (Tom)
1795 Prevent occasional index corruption (Vadim)
1799 Add OUTER JOINs (Tom)
1800 Function manager overhaul (Tom)
1801 Allow ALTER TABLE RENAME on indexes (Tom)
1802 Improve CLUSTER (Tom)
1803 Improve ps status display for more platforms (Peter E, Marc)
1804 Improve CREATE FUNCTION failure message (Ross)
1805 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
1807 Grand Unified Configuration scheme/GUC. Many options can now be set in
1808 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
1809 Improved handling of file descriptor cache (Tom)
1810 New warning code about auto-created table alias entries (Bruce)
1811 Overhaul initdb process (Tom, Peter E)
1812 Overhaul of inherited tables; inherited tables now accessed by default;
1813 new ONLY keyword prevents it (Chris Bitmead, Tom)
1814 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
1816 Allow renaming of temp tables (Tom)
1817 Overhaul memory manager contexts (Tom)
1818 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
1819 Overhaul pg_dump (Philip Warner)
1820 Allow pg_hba.conf secondary password file to specify only username (Peter E)
1821 Allow TEMPORARY or TEMP keyword when creating temporary tables (Bruce)
1822 New memory leak checker (Karel)
1823 New SET SESSION CHARACTERISTICS (Thomas)
1824 Allow nested block comments (Thomas)
1825 Add WITHOUT TIME ZONE type qualifier (Thomas)
1826 New ALTER TABLE ADD CONSTRAINT (Stephan)
1827 Use NUMERIC accumulators for INTEGER aggregates (Tom)
1828 Overhaul aggregate code (Tom)
1829 New VARIANCE and STDDEV() aggregates
1830 Improve dependency ordering of pg_dump (Philip)
1831 New pg_restore command (Philip)
1832 New pg_dump tar output option (Philip)
1833 New pg_dump of large objects (Philip)
1834 New ESCAPE option to LIKE (Thomas)
1835 New case-insensitive LIKE - ILIKE (Thomas)
1836 Allow functional indexes to use binary-compatible type (Tom)
1837 Allow SQL functions to be used in more contexts (Tom)
1838 New pg_config utility (Peter E)
1839 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
1841 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
1842 New quote_identifiers() and quote_literal() functions (Jan)
1843 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
1844 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
1845 Update PyGreSQL to version 3.1 (D'Arcy)
1846 Store tables as files named by OID (Vadim)
1847 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
1848 Require DROP VIEW to remove views, no DROP TABLE (Mark)
1849 Allow DROP VIEW view1, view2 (Mark)
1850 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
1851 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
1852 New /contrib/pgcrypto hashing functions (Marko Kreen)
1853 New pg_dumpall --globals-only option (Peter E)
1854 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
1855 New AT TIME ZONE syntax (Thomas)
1856 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
1857 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
1858 Allow socket path name to be specified in hostname by using leading slash
1859 (David J. MacKenzie)
1860 Allow CREATE DATABASE to specify template database (Tom)
1861 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
1862 New /contrib/rserv replication toolkit (Vadim)
1863 New file format for COPY BINARY (Tom)
1864 New /contrib/oid2name to map numeric files to table names (B Palmer)
1865 New "idle in transaction" ps status message (Marc)
1866 Update to pgaccess 0.98.7 (Constantin Teodorescu)
1867 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
1868 Add rudimentary dependency checking to pg_dump (Philip)
1872 Fix INET/CIDR type ordering and add new functions (Tom)
1873 Make OID behave as an unsigned type (Tom)
1874 Allow BIGINT as synonym for INT8 (Peter E)
1875 New int2 and int8 comparison operators (Tom)
1876 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
1877 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
1878 New GIST seg/cube examples (Gene Selkov)
1879 Improved round(numeric) handling (Tom)
1880 Fix CIDR output formatting (Tom)
1881 New CIDR abbrev() function (Tom)
1885 Write-Ahead Log (WAL) to provide crash recovery with less performance
1887 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
1888 Reduced file seeks (Denis Perchine)
1889 Improve BTREE code for duplicate keys (Tom)
1890 Store all large objects in a single table (Denis Perchine, Tom)
1891 Improve memory allocation performance (Karel, Tom)
1895 New function manager call conventions (Tom)
1896 SGI portability fixes (David Kaelbling)
1897 New configure --enable-syslog option (Peter E)
1898 New BSDI README (Bruce)
1899 configure script moved to top level, not /src (Peter E)
1900 Makefile/configuration/compilation overhaul (Peter E)
1901 New configure --with-python option (Peter E)
1902 Solaris cleanups (Peter E)
1903 Overhaul /contrib Makefiles (Karel)
1904 New OpenSSL configuration option (Magnus, Peter E)
1906 QNX fixes (Maurizio)
1907 New heap_open(), heap_openr() API (Tom)
1908 Remove colon and semi-colon operators (Thomas)
1909 New pg_class.relkind value for views (Mark Hollomon)
1910 Rename ichar() to chr() (Karel)
1911 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
1912 Fixes for NT/Cygwin (Pete Forman)
1913 AIX port fixes (Andreas)
1914 New BeOS port (David Reid, Cyril Velter)
1915 Add proofreader's changes to docs (Addison-Wesley, Bruce)
1916 New Alpha spinlock code (Adriaan Joubert, Compaq)
1917 UnixWare port overhaul (Peter E)
1918 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
1919 New FreeBSD Alpha port (Alfred)
1920 Overhaul shared memory segments (Tom)
1921 Add IBM S/390 support (Neale Ferguson)
1922 Moved macmanuf to /contrib (Larry Rosenman)
1923 Syslog improvements (Larry Rosenman)
1924 New template0 database that contains no user additions (Tom)
1925 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
1926 Allow NetBSD's libedit instead of readline (Peter)
1927 Improved assembly language source code format (Bruce)
1928 New contrib/pg_logger
1929 New --template option to createdb
1930 New contrib/pg_control utility (Oliver)
1931 New FreeBSD tools ipc_check, start-scripts/freebsd
1933 _________________________________________________________________
1937 Release date: 2000-11-11
1939 This has a variety of fixes from 7.0.2.
1940 _________________________________________________________________
1942 Migration to version 7.0.3
1944 A dump/restore is *not* required for those running 7.0.*.
1945 _________________________________________________________________
1950 Large object fix (Tom)
1951 Fix lean in COPY WITH OIDS leak (Tom)
1952 Fix backwards-index-scan (Tom)
1953 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
1954 Add --enable-syslog to configure (Marc)
1955 Fix abort transaction at backend exit in rare cases (Tom)
1956 Fix for psql \l+ when multibyte enabled (Tatsuo)
1957 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
1958 Make vacuum always flush buffers (Tom)
1959 Fix to allow cancel while waiting for a lock (Hiroshi)
1960 Fix for memory aloocation problem in user authentication code (Tom)
1961 Remove bogus use of int4out() (Tom)
1962 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
1963 Fix for failure of triggers on heap open in certain cases (Jeroen van
1965 Fix for erroneous selectivity of not-equals (Tom)
1966 Fix for erroneous use of strcmp() (Tom)
1967 Fix for bug where storage manager accesses items beyond end of file
1969 Fix to include kernel errno message in all smgr elog messages (Tom)
1970 Fix for '.' not in PATH at build time (SL Baur)
1971 Fix for out-of-file-descriptors error (Tom)
1972 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
1973 Fix for subselect in targetlist of Append node (Tom)
1974 Fix for mergejoin plans (Tom)
1975 Fix TRUNCATE failure on relations with indexes (Tom)
1976 Avoid database-wide restart on write error (Hiroshi)
1977 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
1978 Fix VACUUM problem with moving chain of update tuples when source and
1979 destination of a tuple lie on the same page (Tom)
1980 Fix user.c CommandCounterIncrement (Tom)
1981 Fix for AM/PM boundary problem in to_char() (Karel Zak)
1982 Fix TIME aggregate handling (Tom)
1983 Fix to_char() to avoid coredump on NULL input (Tom)
1985 Fix for inserting/copying longer multibyte strings into char() data
1987 Fix for crash of backend, on abort (Tom)
1989 _________________________________________________________________
1993 Release date: 2000-06-05
1995 This is a repackaging of 7.0.1 with added documentation.
1996 _________________________________________________________________
1998 Migration to version 7.0.2
2000 A dump/restore is *not* required for those running 7.*.
2001 _________________________________________________________________
2005 Added documentation to tarball.
2007 _________________________________________________________________
2011 Release date: 2000-06-01
2013 This is a cleanup release for 7.0.
2014 _________________________________________________________________
2016 Migration to version 7.0.1
2018 A dump/restore is *not* required for those running 7.0.
2019 _________________________________________________________________
2023 Fix many CLUSTER failures (Tom)
2024 Allow ALTER TABLE RENAME works on indexes (Tom)
2025 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
2026 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
2027 Fix the off by one errors in ResultSet from 6.5.3, and more.
2028 jdbc ResultSet fixes (Joseph Shraibman)
2029 optimizer tunings (Tom)
2030 Fix create user for pgaccess
2031 Fix for UNLISTEN failure
2032 IRIX fixes (David Kaelbling)
2033 QNX fixes (Andreas Kardos)
2034 Reduce COPY IN lock level (Tom)
2035 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
2036 Fix pg_dump to handle OID indexes (Tom)
2037 Fix small memory leak (Tom)
2038 Solaris fix for createdb/dropdb (Tatsuo)
2039 Fix for non-blocking connections (Alfred Perlstein)
2040 Fix improper recovery after RENAME TABLE failures (Tom)
2041 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
2042 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
2043 Fix too long syslog message (Tatsuo)
2044 Fix problem with quoted indexes that are too long (Tom)
2045 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
2046 ecpg changes (Michael)
2048 _________________________________________________________________
2052 Release date: 2000-05-08
2054 This release contains improvements in many areas, demonstrating the
2055 continued growth of PostgreSQL. There are more improvements and fixes
2056 in 7.0 than in any previous release. The developers have confidence
2057 that this is the best release yet; we do our best to put out only
2058 solid releases, and this one is no exception.
2060 Major changes in this release:
2063 Foreign keys are now implemented, with the exception of PARTIAL
2064 MATCH foreign keys. Many users have been asking for this
2065 feature, and we are pleased to offer it.
2068 Continuing on work started a year ago, the optimizer has been
2069 improved, allowing better query plan selection and faster
2070 performance with less memory usage.
2073 psql, our interactive terminal monitor, has been updated with a
2074 variety of new features. See the psql manual page for details.
2077 SQL92 join syntax is now supported, though only as INNER JOIN
2078 for this release. JOIN, NATURAL JOIN, JOIN/USING, and JOIN/ON
2079 are available, as are column correlation names.
2080 _________________________________________________________________
2082 Migration to version 7.0
2084 A dump/restore using pg_dump is required for those wishing to migrate
2085 data from any previous release of PostgreSQL. For those upgrading from
2086 6.5.*, you may instead use pg_upgrade to upgrade to this release;
2087 however, a full dump/reload installation is always the most robust
2088 method for upgrades.
2090 Interface and compatibility issues to consider for the new release
2093 * The date/time types datetime and timespan have been superseded by
2094 the SQL92-defined types timestamp and interval. Although there has
2095 been some effort to ease the transition by allowing PostgreSQL to
2096 recognize the deprecated type names and translate them to the new
2097 type names, this mechanism may not be completely transparent to
2098 your existing application.
2099 * The optimizer has been substantially improved in the area of query
2100 cost estimation. In some cases, this will result in decreased
2101 query times as the optimizer makes a better choice for the
2102 preferred plan. However, in a small number of cases, usually
2103 involving pathological distributions of data, your query times may
2104 go up. If you are dealing with large amounts of data, you may want
2105 to check your queries to verify performance.
2106 * The JDBC and ODBC interfaces have been upgraded and extended.
2107 * The string function CHAR_LENGTH is now a native function. Previous
2108 versions translated this into a call to LENGTH, which could result
2109 in ambiguity with other types implementing LENGTH such as the
2111 _________________________________________________________________
2117 Prevent function calls exceeding maximum number of arguments (Tom)
2118 Improve CASE construct (Tom)
2119 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
2120 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
2121 Fix GROUP BY scan bug (Tom)
2122 Improvements in SQL grammar processing (Tom)
2123 Fix for views involved in INSERT ... SELECT ... (Tom)
2124 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
2125 Fix for subselects in INSERT ... SELECT (Tom)
2126 Prevent INSERT ... SELECT ... ORDER BY (Tom)
2127 Fixes for relations greater than 2GB, including vacuum
2128 Improve propagating system table changes to other backends (Tom)
2129 Improve propagating user table changes to other backends (Tom)
2130 Fix handling of temp tables in complex situations (Bruce, Tom)
2131 Allow table locking at table open, improving concurrent reliability (Tom)
2132 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
2133 Prevent DROP DATABASE while others accessing
2134 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
2135 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
2136 Fix pg_upgrade so it works for MVCC (Tom)
2137 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
2138 Fix for "f1 datetime DEFAULT 'now'" (Tom)
2139 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
2140 Allow comment-only lines, and ;;; lines too. (Tom)
2141 Improve recovery after failed disk writes, disk full (Hiroshi)
2142 Fix cases where table is mentioned in FROM but not joined (Tom)
2143 Allow HAVING clause without aggregate functions (Tom)
2144 Fix for "--" comment and no trailing newline, as seen in perl interface
2145 Improve pg_dump failure error reports (Bruce)
2146 Allow sorts and hashes to exceed 2GB file sizes (Tom)
2147 Fix for pg_dump dumping of inherited rules (Tom)
2148 Fix for NULL handling comparisons (Tom)
2149 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
2150 Fix for dbname with dash
2151 Prevent DROP INDEX from interfering with other backends (Tom)
2152 Fix file descriptor leak in verify_password()
2153 Fix for "Unable to identify an operator =$" problem
2154 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
2155 Fix for recursive exit call (Massimo)
2156 Fix for extra-long timezones (Jeroen van Vianen)
2157 Make pg_dump preserve primary key information (Peter E)
2158 Prevent databases with single quotes (Peter E)
2159 Prevent DROP DATABASE inside transaction (Peter E)
2160 ecpg memory leak fixes (Stephen Birch)
2161 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
2162 Y2K timestamp fix (Massimo)
2163 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
2164 Fix for views with tables/columns containing spaces (Tom)
2165 Prevent permissions on indexes (Peter E)
2166 Fix for spinlock stuck problem when error is generated (Hiroshi)
2167 Fix ipcclean on Linux
2168 Fix handling of NULL constraint conditions (Tom)
2169 Fix memory leak in odbc driver (Nick Gorham)
2170 Fix for permission check on UNION tables (Tom)
2171 Fix to allow SELECT 'a' LIKE 'a' (Tom)
2172 Fix for SELECT 1 + NULL (Tom)
2174 Fix log() on numeric type (Tom)
2175 Deprecate ':' and ';' operators
2176 Allow vacuum of temporary tables
2177 Disallow inherited columns with the same name as new columns
2178 Recover or force failure when disk space is exhausted (Hiroshi)
2179 Fix INSERT INTO ... SELECT with AS columns matching result columns
2180 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns
2182 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
2183 Fix UNION with LIMIT
2184 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
2185 Fix CREATE TABLE test(col char(2) DEFAULT user)
2186 Fix mismatched types in CREATE TABLE ... DEFAULT
2187 Fix SELECT * FROM pg_class where oid in (0,-1)
2188 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
2189 Prevent user who can create databases can modifying pg_database table (Peter E)
2190 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
2191 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
2195 New CLI interface include file sqlcli.h, based on SQL3/SQL98
2196 Remove all limits on query length, row length limit still exists (Tom)
2197 Update jdbc protocol to 2.0 (Jens Glaser <jens@jens.de>)
2198 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
2199 Fix to give super user and createdb user proper update catalog rights (Peter E)
2200 Allow ecpg bool variables to have NULL values (Christof)
2201 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
2202 Allow ^C to cancel COPY command (Massimo)
2203 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
2204 Function name overloading for dynamically-loaded C functions (Frankpitt)
2205 Add CmdTuples() to libpq++(Vince)
2206 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
2207 Allow CREATE FUNCTION/WITH clause to be used for all language types
2208 configure --enable-debug adds -g (Peter E)
2209 configure --disable-debug removes -g (Peter E)
2210 Allow more complex default expressions (Tom)
2211 First real FOREIGN KEY constraint trigger functionality (Jan)
2212 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
2213 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
2214 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
2215 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
2216 Change pgeasy connectdb() parameter ordering (Bruce)
2217 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
2218 Add Oracle's COMMENT ON command (Mike Mascari <mascarim@yahoo.com>)
2219 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
2220 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
2221 Make USING in COPY optional (Bruce)
2222 Allow subselects in the target list (Tom)
2223 Allow subselects on the left side of comparison operators (Tom)
2224 New parallel regression test (Jan)
2225 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
2226 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
2227 Added psql LASTOID variable to return last inserted oid (Peter E)
2228 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
2229 Add permissions check for vacuum (Peter E)
2230 New libpq functions to allow asynchronous connections: PQconnectStart(),
2231 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
2232 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
2233 New libpq PQsetenv() function (Ewan Mellor)
2234 create/alter user extension (Peter E)
2235 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
2236 New scripts for create/drop user/db (Peter E)
2237 Major psql overhaul (Peter E)
2238 Add const to libpq interface (Peter E)
2239 New libpq function PQoidValue (Peter E)
2240 Show specific non-aggregate causing problem with GROUP BY (Tom)
2241 Make changes to pg_shadow recreate pg_pwd file (Peter E)
2242 Add aggregate(DISTINCT ...) (Tom)
2243 Allow flag to control COPY input/output of NULLs (Peter E)
2244 Make postgres user have a password by default (Peter E)
2245 Add CREATE/ALTER/DROP GROUP (Peter E)
2246 All administration scripts now support --long options (Peter E, Karel)
2247 Vacuumdb script now supports --all option (Peter E)
2248 ecpg new portable FETCH syntax
2249 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
2250 and EXEC SQL ENDIF directives
2251 Add pg_ctl script to control backend start-up (Tatsuo)
2252 Add postmaster.opts.default file to store start-up flags (Tatsuo)
2253 Allow --with-mb=SQL_ASCII
2254 Increase maximum number of index keys to 16 (Bruce)
2255 Increase maximum number of function arguments to 16 (Bruce)
2256 Allow configuration of maximum number of index keys and arguments (Bruce)
2257 Allow unprivileged users to change their passwords (Peter E)
2258 Password authentication enabled; required for new users (Peter E)
2259 Disallow dropping a user who owns a database (Peter E)
2260 Change initdb option --with-mb to --enable-multibyte
2261 Add option for initdb to prompts for superuser password (Peter E)
2262 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
2263 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
2264 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
2265 Libpq non-blocking mode (Alfred Perlstein)
2266 Improve conversion of types in casts that don't specify a length
2267 New plperl internal programming language (Mark Hollomon)
2268 Allow COPY IN to read file that do not end with a newline (Tom)
2269 Indicate when long identifiers are truncated (Tom)
2270 Allow aggregates to use type equivalency (Peter E)
2271 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
2272 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
2273 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
2274 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
2275 Add NUMERIC and int8 types to ODBC
2276 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
2277 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
2278 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
2279 Enable backward sequential scan even after reaching EOF (Hiroshi)
2280 Add btree indexing of boolean values, >= and <= (Don Baccus)
2281 Print current line number when COPY FROM fails (Massimo)
2282 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
2283 Add DEC as synonym for DECIMAL (Thomas)
2284 Add SESSION_USER as SQL92 keyword, same as CURRENT_USER (Thomas)
2285 Implement SQL92 column aliases (aka correlation names) (Thomas)
2286 Implement SQL92 join syntax (Thomas)
2287 Make INTERVAL reserved word allowed as a column identifier (Thomas)
2288 Implement REINDEX command (Hiroshi)
2289 Accept ALL in aggregate function SUM(ALL col) (Tom)
2290 Prevent GROUP BY from using column aliases (Tom)
2291 New psql \encoding option (Tatsuo)
2292 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
2293 Allow negation of a negative number in all cases
2294 Add ecpg descriptors (Christof, Michael)
2295 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
2296 Allow casts with length, like foo::char(8)
2297 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
2298 Add support for SJIS user defined characters (Tatsuo)
2299 Larger views/rules supported
2300 Make libpq's PQconndefaults() thread-safe (Tom)
2301 Disable // as comment to be ANSI conforming, should use -- (Tom)
2302 Allow column aliases on views CREATE VIEW name (collist)
2303 Fixes for views with subqueries (Tom)
2304 Allow UPDATE table SET fld = (SELECT ...) (Tom)
2305 SET command options no longer require quotes
2306 Update pgaccess to 0.98.6
2307 New SET SEED command
2308 New pg_options.sample file
2309 New SET FSYNC command (Massimo)
2310 Allow pg_descriptions when creating tables
2311 Allow pg_descriptions when creating types, columns, and functions
2312 Allow psql \copy to allow delimiters (Peter E)
2313 Allow psql to print nulls as distinct from "" [null] (Peter E)
2317 Many array fixes (Tom)
2318 Allow bare column names to be subscripted as arrays (Tom)
2319 Improve type casting of int and float constants (Tom)
2320 Cleanups for int8 inputs, range checking, and type conversion (Tom)
2321 Fix for SELECT timespan('21:11:26'::time) (Tom)
2322 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
2323 Add btree index on NUMERIC (Jan)
2324 Perl fix for large objects containing NUL characters (Douglas Thomson)
2325 ODBC fix for for large objects (free)
2326 Fix indexing of cidr data type
2327 Fix for Ethernet MAC addresses (macaddr type) comparisons
2328 Fix for date/time types when overflows happened in computations (Tom)
2329 Allow array on int8 (Peter E)
2330 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
2331 Allow NUMERIC arrays
2332 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
2333 Make char_length()/octet_length including trailing blanks (Tom)
2334 Made abstime/reltime use int4 instead of time_t (Peter E)
2335 New lztext data type for compressed text fields
2336 Revise code to handle coercion of int and float constants (Tom)
2337 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
2338 NUMERIC now accepts scientific notation (Tom)
2339 NUMERIC to int4 rounds (Tom)
2340 Convert float4/8 to NUMERIC properly (Tom)
2341 Allow type conversion with NUMERIC (Thomas)
2342 Make ISO date style (2000-02-16 09:33) the default (Thomas)
2343 Add NATIONAL CHAR [ VARYING ] (Thomas)
2344 Allow NUMERIC round and trunc to accept negative scales (Tom)
2345 New TIME WITH TIME ZONE type (Thomas)
2346 Add MAX()/MIN() on time type (Thomas)
2347 Add abs(), mod(), fac() for int8 (Thomas)
2348 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
2349 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
2350 Add exp() and ln() for NUMERIC type
2351 Rename NUMERIC power() to pow() (Thomas)
2352 Improved TRANSLATE() function (Edwin Ramirez, Tom)
2353 Allow X=-Y operators (Tom)
2354 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom
2356 Allow LOCALE to use indexes in regular expression searches (Tom)
2357 Allow creation of functional indexes to use default types
2361 Prevent exponential space consumption with many AND's and OR's (Tom)
2362 Collect attribute selectivity values for system columns (Tom)
2363 Reduce memory usage of aggregates (Tom)
2364 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
2365 Fix r-tree index optimizer selectivity (Thomas)
2366 Improve optimizer selectivity computations and functions (Tom)
2367 Optimize btree searching for cases where many equal keys exist (Tom)
2368 Enable fast LIKE index processing only if index present (Tom)
2369 Re-use free space on index pages with duplicates (Tom)
2370 Improve hash join processing (Tom)
2371 Prevent descending sort if result is already sorted(Hiroshi)
2372 Allow commuting of index scan query qualifications (Tom)
2373 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
2374 Allocate large memory requests in fix-sized chunks for performance (Tom)
2375 Fix vacuum's performance by reducing memory allocation requests (Tom)
2376 Implement constant-expression simplification (Bernard Frankpitt, Tom)
2377 Use secondary columns to be used to determine start of index scan (Hiroshi)
2378 Prevent quadruple use of disk space when doing internal sorting (Tom)
2379 Faster sorting by calling fewer functions (Tom)
2380 Create system indexes to match all system caches (Bruce, Hiroshi)
2381 Make system caches use system indexes (Bruce)
2382 Make all system indexes unique (Bruce)
2383 Improve pg_statistics management for VACUUM speed improvement (Tom)
2384 Flush backend cache less frequently (Tom, Hiroshi)
2385 COPY now reuses previous memory allocation, improving performance (Tom)
2386 Improve optimization cost estimation (Tom)
2387 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom
2389 Use DNF instead of CNF where appropriate (Tom, Taral)
2390 Further cleanup for OR-of-AND WHERE-clauses (Tom)
2391 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
2392 Smarter optimizer computations for random index page access (Tom)
2393 New SET variable to control optimizer costs (Tom)
2394 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
2395 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
2396 Major subquery speedup (Tom)
2397 Fewer fsync writes when fsync is not disabled (Tom)
2398 Improved LIKE optimizer estimates (Tom)
2399 Prevent fsync in SELECT-only queries (Vadim)
2400 Make index creation use psort code, because it is now faster (Tom)
2401 Allow creation of sort temp tables > 1 Gig
2405 Fix for linux PPC compile
2406 New generic expression-tree-walker subroutine (Tom)
2407 Change form() to varargform() to prevent portability problems
2408 Improved range checking for large integers on Alphas
2409 Clean up #include in /include directory (Bruce)
2410 Add scripts for checking includes (Bruce)
2411 Remove un-needed #include's from *.c files (Bruce)
2412 Change #include's to use <> and "" as appropriate (Bruce)
2413 Enable WIN32 compilation of libpq
2414 Alpha spinlock fix from Uncle George <gatgul@voicenet.com>
2415 Overhaul of optimizer data structures (Tom)
2416 Fix to cygipc library (Yutaka Tanida)
2417 Allow pgsql to work on newer Cygwin snapshots (Dan)
2418 New catalog version number (Tom)
2420 Rename heap_replace to heap_update
2421 Update for QNX (Dr. Andreas Kardos)
2422 New platform-specific regression handling (Tom)
2423 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
2424 Included all yacc and lex files into the distribution (Peter E.)
2425 Remove lextest, no longer needed (Peter E)
2426 Fix for libpq and psql on Win32 (Magnus)
2427 Internally change datetime and timespan into timestamp and interval (Thomas)
2428 Fix for plpgsql on BSD/OS
2429 Add SQL_ASCII test case to the regression test (Tatsuo)
2430 configure --with-mb now deprecated (Tatsuo)
2432 NetBSD fixes (Johnny C. Lam <lamj@stat.cmu.edu>)
2433 Fixes for Alpha compiles
2434 New multibyte encodings
2436 _________________________________________________________________
2440 Release date: 1999-10-13
2442 This is basically a cleanup release for 6.5.2. We have added a new
2443 PgAccess that was missing in 6.5.2, and installed an NT-specific fix.
2444 _________________________________________________________________
2446 Migration to version 6.5.3
2448 A dump/restore is *not* required for those running 6.5.*.
2449 _________________________________________________________________
2453 Updated version of pgaccess 0.98
2455 Fix dumping rules on inherited tables
2457 _________________________________________________________________
2461 Release date: 1999-09-15
2463 This is basically a cleanup release for 6.5.1. We have fixed a variety
2464 of problems reported by 6.5.1 users.
2465 _________________________________________________________________
2467 Migration to version 6.5.2
2469 A dump/restore is *not* required for those running 6.5.*.
2470 _________________________________________________________________
2474 subselect+CASE fixes(Tom)
2475 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
2476 Fixes for CASE in WHERE join clauses(Tom)
2477 Fix BTScan abort(Tom)
2478 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
2479 Improve it so that it checks for multicolumn constraints(Thomas)
2480 Fix for Win32 making problem with MB enabled(Hiroki Kataoka)
2481 Allow BSD yacc and bison to compile pl code(Bruce)
2482 Fix SET NAMES working
2484 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
2485 Reduce the total memory consumption of vacuum(Tom)
2486 Fix for timestamp(datetime)
2487 Rule deparsing bugfixes(Tom)
2488 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(To
2490 This is to re-use space on index pages freed by vacuum(Vadim)
2491 document -x for pg_dump(Bruce)
2492 Fix for unary operators in rule deparser(Tom)
2493 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
2494 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
2495 Repair logic error in LIKE: should not return LIKE_ABORT
2496 when reach end of pattern before end of text(Tom)
2497 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom
2499 Updated version of pgaccess 0.98
2501 _________________________________________________________________
2505 Release date: 1999-07-15
2507 This is basically a cleanup release for 6.5. We have fixed a variety
2508 of problems reported by 6.5 users.
2509 _________________________________________________________________
2511 Migration to version 6.5.1
2513 A dump/restore is *not* required for those running 6.5.
2514 _________________________________________________________________
2519 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
2520 Remove QUERY_LIMIT, use SELECT...LIMIT
2521 Fix for EXPLAIN on inheritance(Tom)
2522 Patch to allow vacuum on multisegment tables(Hiroshi)
2523 R-Tree optimizer selectivity fix(Tom)
2524 ACL file descriptor leak fix(Atsushi Ogawa)
2525 New expresssion subtree code(Tom)
2526 Avoid disk writes for read-only transactions(Vadim)
2527 Fix for removal of temp tables if last transaction was aborted(Bruce)
2528 Fix to prevent too large tuple from being created(Bruce)
2530 Allow port numbers 32k - 64k(Bruce)
2531 Add ^ precidence(Bruce)
2532 Rename sort files called pg_temp to pg_sorttemp(Bruce)
2533 Fix for microseconds in time values(Tom)
2534 Tutorial source cleanup
2536 Fix for sorting of NULL's in some cases(Tom)
2537 Shared library dependencies fixed (Tom)
2538 Fixed glitches affecting GROUP BY in subselects(Tom)
2539 Fix some compiler warnings (Tomoaki Nishiyama)
2540 Add Win1250 (Czech) support (Pavel Behal)
2542 _________________________________________________________________
2546 Release date: 1999-06-09
2548 This release marks a major step in the development team's mastery of
2549 the source code we inherited from Berkeley. You will see we are now
2550 easily adding major features, thanks to the increasing size and
2551 experience of our world-wide development team.
2553 Here is a brief summary of the more notable changes:
2555 Multiversion concurrency control(MVCC)
2556 This removes our old table-level locking, and replaces it with
2557 a locking system that is superior to most commercial database
2558 systems. In a traditional system, each row that is modified is
2559 locked until committed, preventing reads by other users. MVCC
2560 uses the natural multiversion nature of PostgreSQL to allow
2561 readers to continue reading consistent data during writer
2562 activity. Writers continue to use the compact pg_log
2563 transaction system. This is all performed without having to
2564 allocate a lock for every row like traditional database
2565 systems. So, basically, we no longer are restricted by simple
2566 table-level locking; we have something better than row-level
2569 Hot backups from pg_dump
2570 pg_dump takes advantage of the new MVCC features to give a
2571 consistent database dump/backup while the database stays online
2572 and available for queries.
2575 We now have a true numeric data type, with user-specified
2579 Temporary tables are guaranteed to have unique names within a
2580 database session, and are destroyed on session exit.
2583 We now have CASE, INTERSECT, and EXCEPT statement support. We
2584 have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL, SELECT
2585 ... FOR UPDATE, and an improved LOCK TABLE command.
2588 We continue to speed up PostgreSQL, thanks to the variety of
2589 talents within our team. We have sped up memory allocation,
2590 optimization, table joins, and row transfer routines.
2593 We continue to expand our port list, this time including
2594 Windows NT/ix86 and NetBSD/arm32.
2597 Most interfaces have new versions, and existing functionality
2601 New and updated material is present throughout the
2602 documentation. New FAQs have been contributed for SGI and AIX
2603 platforms. The Tutorial has introductory information on SQL
2604 from Stefan Simkovics. For the User's Guide, there are
2605 reference pages covering the postmaster and more utility
2606 programs, and a new appendix contains details on date/time
2607 behavior. The Administrator's Guide has a new chapter on
2608 troubleshooting from Tom Lane. And the Programmer's Guide has a
2609 description of query processing, also from Stefan, and details
2610 on obtaining the PostgreSQL source tree via anonymous CVS and
2612 _________________________________________________________________
2614 Migration to version 6.5
2616 A dump/restore using pg_dump is required for those wishing to migrate
2617 data from any previous release of PostgreSQL. pg_upgrade can *not* be
2618 used to upgrade to this release because the on-disk structure of the
2619 tables has changed compared to previous releases.
2621 The new Multiversion Concurrency Control (MVCC) features can give
2622 somewhat different behaviors in multiuser environments. *Read and
2623 understand the following section to ensure that your existing
2624 applications will give you the behavior you need.*
2625 _________________________________________________________________
2627 Multiversion Concurrency Control
2629 Because readers in 6.5 don't lock data, regardless of transaction
2630 isolation level, data read by one transaction can be overwritten by
2631 another. In other words, if a row is returned by "SELECT" it doesn't
2632 mean that this row really exists at the time it is returned (i.e.
2633 sometime after the statement or transaction began) nor that the row is
2634 protected from being deleted or updated by concurrent transactions
2635 before the current transaction does a commit or rollback.
2637 To ensure the actual existence of a row and protect it against
2638 concurrent updates one must use "SELECT FOR UPDATE" or an appropriate
2639 "LOCK TABLE" statement. This should be taken into account when porting
2640 applications from previous releases of PostgreSQL and other
2643 Keep the above in mind if you are using "contrib/refint.*" triggers
2644 for referential integrity. Additional techniques are required now. One
2645 way is to use "LOCK parent_table IN SHARE ROW EXCLUSIVE MODE" command
2646 if a transaction is going to update/delete a primary key and use "LOCK
2647 parent_table IN SHARE MODE" command if a transaction is going to
2648 update/insert a foreign key.
2650 Note: Note that if you run a transaction in SERIALIZABLE mode then
2651 you must execute the "LOCK" commands above before execution of any
2652 DML statement ("SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO") in the
2655 These inconveniences will disappear in the future when the ability to
2656 read dirty (uncommitted) data (regardless of isolation level) and true
2657 referential integrity will be implemented.
2658 _________________________________________________________________
2664 Fix text<->float8 and text<->float4 conversion functions(Thomas)
2665 Fix for creating tables with mixed-case constraints(Billy)
2666 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
2667 Fix bug in pg_dump -z
2668 Memory overrun cleanups(Tatsuo)
2669 Fix for lo_import crash(Tatsuo)
2670 Adjust handling of data type names to suppress double quotes(Thomas)
2671 Use type coercion for matching columns and DEFAULT(Thomas)
2672 Fix deadlock so it only checks once after one second of sleep(Bruce)
2673 Fixes for aggregates and PL/pgsql(Hiroshi)
2674 Fix for subquery crash(Vadim)
2675 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
2676 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo
2678 Fix for pg_dump -d or -D and quote special characters in INSERT
2679 Repair serious problems with dynahash(Tom)
2680 Fix INET/CIDR portability problems
2681 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
2682 Fix executor so mergejoin of different column types works(Tom)
2683 Fix for Alpha OR selectivity bug
2684 Fix OR index selectivity problem(Bruce)
2685 Fix so \d shows proper length for char()/varchar()(Ryan)
2686 Fix tutorial code(Clark)
2687 Improve destroyuser checking(Oliver)
2688 Fix for Kerberos(Rodney McDuff)
2689 Fix for dropping database while dirty buffers(Bruce)
2690 Fix so sequence nextval() can be case-sensitive(Bruce)
2692 Drop buffers before destroying database files(Bruce)
2693 Fix case where executor evaluates functions twice(Tatsuo)
2694 Allow sequence nextval actions to be case-sensitive(Bruce)
2695 Fix optimizer indexing not working for negative numbers(Bruce)
2696 Fix for memory leak in executor with fjIsNull
2697 Fix for aggregate memory leaks(Erik Riedel)
2698 Allow username containing a dash GRANT permissions
2699 Cleanup of NULL in inet types
2700 Clean up system table bugs(Tom)
2701 Fix problems of PAGER and \? command(Masaaki Sakaida)
2702 Reduce default multisegment file size limit to 1GB(Peter)
2703 Fix for dumping of CREATE OPERATOR(Tom)
2704 Fix for backward scanning of cursors(Hiroshi Inoue)
2705 Fix for COPY FROM STDIN when using \i(Tom)
2706 Fix for subselect is compared inside an expression(Jan)
2707 Fix handling of error reporting while returning rows(Tom)
2708 Fix problems with reference to array types(Tom,Jan)
2709 Prevent UPDATE SET oid(Jan)
2710 Fix pg_dump so -t option can handle case-sensitive tablenames
2711 Fixes for GROUP BY in special cases(Tom, Jan)
2712 Fix for memory leak in failed queries(Tom)
2713 DEFAULT now supports mixed-case identifiers(Tom)
2714 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
2715 Disable use of pg_dump with both -o and -d options(Bruce)
2716 Allow pg_dump to properly dump GROUP permissions(Bruce)
2717 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
2718 Fix for computations in views(Jan)
2719 Fix for aggregates on array indexes(Tom)
2720 Fix for DEFAULT handles single quotes in value requiring too many quotes
2721 Fix security problem with non-super users importing/exporting large objects(Tom
2723 Rollback of transaction that creates table cleaned up properly(Tom)
2724 Fix to allow long table and column names to generate proper serial names(Tom)
2728 Add "vacuumdb" utility
2729 Speed up libpq by allocating memory better(Tom)
2730 EXPLAIN all indexes used(Tom)
2731 Implement CASE, COALESCE, NULLIF expression(Thomas)
2732 New pg_dump table output format(Constantin)
2733 Add string min()/max() functions(Thomas)
2734 Extend new type coercion techniques to aggregates(Thomas)
2735 New moddatetime contrib(Terry)
2736 Update to pgaccess 0.96(Constantin)
2737 Add routines for single-byte "char" type(Thomas)
2738 Improved substr() function(Thomas)
2739 Improved multibyte handling(Tatsuo)
2740 Multiversion concurrency control/MVCC(Vadim)
2741 New Serialized mode(Vadim)
2742 Fix for tables over 2gigs(Peter)
2743 New SET TRANSACTION ISOLATION LEVEL(Vadim)
2744 New LOCK TABLE IN ... MODE(Vadim)
2745 Update ODBC driver(Byron)
2746 New NUMERIC data type(Jan)
2747 New SELECT FOR UPDATE(Vadim)
2748 Handle "NaN" and "Infinity" for input values(Jan)
2749 Improved date/year handling(Thomas)
2750 Improved handling of backend connections(Magnus)
2751 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
2752 New TCL_ARRAYS option(Massimo)
2753 New INTERSECT and EXCEPT(Stefan)
2754 New pg_index.indisprimary for primary key tracking(D'Arcy)
2755 New pg_dump option to allow dropping of tables before creation(Brook)
2756 Speedup of row output routines(Tom)
2757 New READ COMMITTED isolation level(Vadim)
2758 New TEMP tables/indexes(Bruce)
2759 Prevent sorting if result is already sorted(Jan)
2760 New memory allocation optimization(Jan)
2761 Allow psql to do \p\g(Bruce)
2762 Allow multiple rule actions(Jan)
2763 Added LIMIT/OFFSET functionality(Jan)
2764 Improve optimizer when joining a large number of tables(Bruce)
2765 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
2766 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
2767 Improved int8 support(Ryan Bradetich, Thomas, Tom)
2768 New routines to convert between int8 and text/varchar types(Thomas)
2769 New bushy plans, where meta-tables are joined(Bruce)
2770 Enable right-hand queries by default(Bruce)
2771 Allow reliable maximum number of backends to be set at configure time
2772 (--with-maxbackends and postmaster switch (-N backends))(Tom)
2773 GEQO default now 10 tables because of optimizer speedups(Tom)
2774 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
2775 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
2776 Allow psql \d on a view show query(Ryan)
2777 Speedup for LIKE(Bruce)
2778 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
2779 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
2780 Make % operator have precedence like /(Bruce)
2781 Add new postgres -O option to allow system table structure changes(Bruce)
2782 Update contrib/pginterface/findoidjoins script(Tom)
2783 Major speedup in vacuum of deleted rows with indexes(Vadim)
2784 Allow non-SQL functions to run different versions based on arguments(Tom)
2785 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida
2787 Add version number in start-up banners for psql(Masaaki Sakaida)
2788 New contrib/vacuumlo removes large objects not referenced(Peter)
2789 New initialization for table sizes so non-vacuumed tables perform better(Tom)
2790 Improve error messages when a connection is rejected(Tom)
2791 Support for arrays of char() and varchar() fields(Massimo)
2792 Overhaul of hash code to increase reliability and performance(Tom)
2793 Update to PyGreSQL 2.4(D'Arcy)
2794 Changed debug options so -d4 and -d5 produce different node displays(Jan)
2795 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
2796 Better optimization statistics for system table access(Tom)
2797 Better handling of non-default block sizes(Massimo)
2798 Improve GEQO optimizer memory consumption(Tom)
2799 UNION now suppports ORDER BY of columns not in target list(Jan)
2800 Major libpq++ improvements(Vince Vielhaber)
2801 pg_dump now uses -z(ACL's) as default(Bruce)
2802 backend cache, memory speedups(Tom)
2803 have pg_dump do everything in one snapshot transaction(Vadim)
2804 fix for large object memory leakage, fix for pg_dumping(Tom)
2805 INET type now respects netmask for comparisons
2806 Make VACUUM ANALYZE only use a readlock(Vadim)
2807 Allow VIEWs on UNIONS(Jan)
2808 pg_dump now can generate consistent snapshots on active databases(Vadim)
2812 Improve port matching(Tom)
2813 Portability fixes for SunOS
2814 Add NT/Win32 backend port and enable dynamic loading(Magnus and Daniel Horak)
2815 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
2816 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
2817 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
2818 Port to NetBSD/macppc(Toshimi Aoki)
2819 Fix for tcl/tk configuration(Vince)
2820 Removed CURRENT keyword for rule queries(Jan)
2821 NT dynamic loading now works(Daniel Horak)
2822 Add ARM32 support(Andrew McMurry)
2823 Better support for HP-UX 11 and UnixWare
2824 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
2825 New install commands for plpgsql(Jan)
2827 _________________________________________________________________
2831 Release date: 1998-12-20
2833 The 6.4.1 release was improperly packaged. This also has one
2835 _________________________________________________________________
2837 Migration to version 6.4.2
2839 A dump/restore is *not* required for those running 6.4.*.
2840 _________________________________________________________________
2844 Fix for datetime constant problem on some platforms(Thomas)
2845 _________________________________________________________________
2849 Release date: 1998-12-18
2851 This is basically a cleanup release for 6.4. We have fixed a variety
2852 of problems reported by 6.4 users.
2853 _________________________________________________________________
2855 Migration to version 6.4.1
2857 A dump/restore is *not* required for those running 6.4.
2858 _________________________________________________________________
2862 Add pg_dump -N flag to force double quotes around identifiers. This is
2864 Fix for NOT in where clause causing crash(Bruce)
2865 EXPLAIN VERBOSE coredump fix(Vadim)
2866 Fix shared-library problems on Linux
2867 Fix test for table existance to allow mixed-case and whitespace in
2868 the table name(Thomas)
2869 Fix a couple of pg_dump bugs
2870 Configure matches template/.similar entries better(Tom)
2871 Change builtin function names from SPI_* to spi_*
2872 OR WHERE clause fix(Vadim)
2873 Fixes for mixed-case table names(Billy)
2874 contrib/linux/postgres.init.csh/sh fix(Thomas)
2875 libpq memory overrun fix
2877 Change exp() behavior to generate error on underflow(Thomas)
2878 pg_dump fixes for memory leak, inheritance constraints, layout change
2879 update pgaccess to 0.93
2880 Fix prototype for 64-bit platforms
2881 Multibyte fixes(Tatsuo)
2883 Fix memory overruns(Tatsuo)
2884 Fix for lo_import() crash(Bruce)
2885 Better search for install program(Tom)
2888 Use implicit type coercion for matching DEFAULT values(Thomas)
2889 Add routines to help with single-byte (internal) character type(Thomas)
2890 Compilation of libpq for Win32 fixes(Magnus)
2891 Upgrade to PyGreSQL 2.2(D'Arcy)
2892 _________________________________________________________________
2896 Release date: 1998-10-30
2898 There are *many* new features and improvements in this release. Thanks
2899 to our developers and maintainers, nearly every aspect of the system
2900 has received some attention since the previous release. Here is a
2901 brief, incomplete summary:
2903 * Views and rules are now functional thanks to extensive new code in
2904 the rewrite rules system from Jan Wieck. He also wrote a chapter
2905 on it for the Programmer's Guide.
2906 * Jan also contributed a second procedural language, PL/pgSQL, to go
2907 with the original PL/pgTCL procedural language he contributed last
2909 * We have optional multiple-byte character set support from Tatsuo
2910 Ishii to complement our existing locale support.
2911 * Client/server communications has been cleaned up, with better
2912 support for asynchronous messages and interrupts thanks to Tom
2914 * The parser will now perform automatic type coercion to match
2915 arguments to available operators and functions, and to match
2916 columns and expressions with target columns. This uses a generic
2917 mechanism which supports the type extensibility features of
2918 PostgreSQL. There is a new chapter in the User's Guide which
2920 * Three new data types have been added. Two types, inet and cidr,
2921 support various forms of IP network, subnet, and machine
2922 addressing. There is now an 8-byte integer type available on some
2923 platforms. See the chapter on data types in the User's Guide for
2924 details. A fourth type, serial, is now supported by the parser as
2925 an amalgam of the int4 type, a sequence, and a unique index.
2926 * Several more SQL92-compatible syntax features have been added,
2927 including "INSERT DEFAULT VALUES"
2928 * The automatic configuration and installation system has received
2929 some attention, and should be more robust for more platforms than
2931 _________________________________________________________________
2933 Migration to version 6.4
2935 A dump/restore using pg_dump or pg_dumpall is required for those
2936 wishing to migrate data from any previous release of PostgreSQL.
2937 _________________________________________________________________
2943 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
2944 Remove char2-16 data types, use char/varchar(Darren)
2945 Pqfn not handles a NOTICE message(Anders)
2946 Reduced busywaiting overhead for spinlocks with many backends (dg)
2947 Stuck spinlock detection (dg)
2948 Fix up "ISO-style" timespan decoding and encoding(Thomas)
2949 Fix problem with table drop after rollback of transaction(Vadim)
2950 Change error message and remove non-functional update message(Vadim)
2951 Fix for COPY array checking
2952 Fix for SELECT 1 UNION SELECT NULL
2953 Fix for buffer leaks in large object calls(Pascal)
2954 Change owner from oid to int4 type(Bruce)
2955 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
2956 Fix for shared invalidation cache overflow(Massimo)
2957 Prevent file descriptor leaks in failed COPY's(Bruce)
2958 Fix memory leak in libpgtcl's pg_select(Constantin)
2959 Fix problems with username/passwords over 8 characters(Tom)
2960 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
2961 Fix of many bad system table entries(Tom)
2965 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
2966 Show the index used in an EXPLAIN(Zeugswetter)
2967 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
2968 Multibyte awareness of many data types and functions, via configure(Tatsuo)
2969 New configure --with-mb option(Tatsuo)
2970 New initdb --pgencoding option(Tatsuo)
2971 New createdb -E multibyte option(Tatsuo)
2972 Select version(); now returns PostgreSQL version(Jeroen)
2973 Libpq now allows asynchronous clients(Tom)
2974 Allow cancel from client of backend query(Tom)
2975 Psql now cancels query with Control-C(Tom)
2976 Libpq users need not issue dummy queries to get NOTIFY messages(Tom)
2977 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
2978 PGresult struct now includes associated error message, if any(Tom)
2979 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
2980 Add routines to convert between varchar and bpchar(Thomas)
2981 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
2982 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
2983 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas
2985 Fixes for unary minus parsing with leading spaces(Thomas)
2986 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
2987 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
2988 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
2989 Enable HAVING clause but no fixes elsewhere yet.
2990 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thoma
2992 Save string type if specified for DEFAULT clause handling(Thomas)
2993 Coerce operations involving different data types(Thomas)
2994 Allow some index use for columns of different types(Thomas)
2995 Add capabilities for automatic type conversion(Thomas)
2996 Cleanups for large objects, so file is truncated on open(Peter)
2997 Readline cleanups(Tom)
2998 Allow psql \f \ to make spaces as delimiter(Bruce)
2999 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
3000 Msql compatibility library in /contrib(Aldrin)
3001 Remove the requirement that ORDER/GROUP BY clause identifiers be
3002 included in the target list(David)
3003 Convert columns to match columns in UNION clauses(Thomas)
3004 Remove fork()/exec() and only do fork()(Bruce)
3005 Jdbc cleanups(Peter)
3006 Show backend status on ps command line(only works on some platforms)(Bruce)
3007 Pg_hba.conf now has a sameuser option in the database field
3008 Make lo_unlink take oid param, not int4
3009 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
3010 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
3011 libpgtcl cleanups(Tom)
3012 Add -error option to libpgtcl's pg_result command(Tom)
3013 New locale patch, see docs/README/locale(Oleg)
3014 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
3015 New contrib/lo code for large object orphan removal(Peter)
3016 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
3017 feature, see /doc/README.mb(Tatsuo)
3018 /contrib/noupdate code to revoke update permission on a column
3019 Libpq can now be compiled on win32(Magnus)
3020 Add PQsetdbLogin() in libpq
3021 New 8-byte integer type, checked by configure for OS support(Thomas)
3022 Better support for quoted table/column names(Thomas)
3023 Surround table and column names with double-quotes in pg_dump(Thomas)
3024 PQreset() now works with passwords(Tom)
3025 Handle case of GROUP BY target list column number out of range(David)
3026 Allow UNION in subselects
3027 Add auto-size to screen to \d? commands(Bruce)
3028 Use UNION to show all \d? results in one query(Bruce)
3029 Add \d? field search feature(Bruce)
3030 Pg_dump issues fewer \connect requests(Tom)
3031 Make pg_dump -z flag work better, document it in manual page(Tom)
3032 Add HAVING clause with full support for subselects and unions(Stephan)
3033 Full text indexing routines in contrib/fulltextindex(Maarten)
3034 Transaction ids now stored in shared memory(Vadim)
3035 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
3036 Support for SQL92 syntax "SET NAMES"(Tatsuo)
3037 Support for LATIN2-5(Tatsuo)
3038 Add UNICODE regression test case(Tatsuo)
3039 Lock manager cleanup, new locking modes for LLL(Vadim)
3040 Allow index use with OR clauses(Bruce)
3041 Allows "SELECT NULL ORDER BY 1;"
3042 Explain VERBOSE prints the plan, and now pretty-prints the plan to
3043 the postmaster log file(Bruce)
3044 Add indexes display to \d command(Bruce)
3045 Allow GROUP BY on functions(David)
3046 New pg_class.relkind for large objects(Bruce)
3047 New way to send libpq NOTICE messages to a different location(Tom)
3048 New \w write command to psql(Bruce)
3049 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
3050 Allow binary-compatible indexes to be considered when checking for valid
3051 Indexes for restriction clauses containing a constant(Thomas)
3052 New ISBN/ISSN code in /contrib/isbn_issn
3053 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
3054 New rewrite system fixes many problems with rules and views(Jan)
3055 * Rules on relations work
3056 * Event qualifications on insert/update/delete work
3057 * New OLD variable to reference CURRENT, CURRENT will be remove in futu
3059 * Update rules can reference NEW and OLD in rule qualifications/actions
3060 * Insert/update/delete rules on views work
3061 * Multiple rule actions are now supported, surrounded by parentheses
3062 * Regular users can create views/rules on tables they have RULE permits
3063 * Rules and views inherit the permissions on the creator
3064 * No rules at the column level
3065 * No UPDATE NEW/OLD rules
3066 * New pg_tables, pg_indexes, pg_rules and pg_views system views
3067 * Only a single action on SELECT rules
3068 * Total rewrite overhaul, perhaps for 6.5
3070 * handle aggregates on views
3071 * handle insert into select from view works
3072 System indexes are now multikey(Bruce)
3073 Oidint2, oidint4, and oidname types are removed(Bruce)
3074 Use system cache for more system table lookups(Bruce)
3075 New backend programming language PL/pgSQL in backend/pl(Jan)
3076 New SERIAL data type, auto-creates sequence/index(Thomas)
3077 Enable assert checking without a recompile(Massimo)
3078 User lock enhancements(Massimo)
3079 New setval() command to set sequence value(Massimo)
3080 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
3081 Conditional trace package(Massimo)
3082 New UNLISTEN command(Massimo)
3083 Psql and libpq now compile under win32 using win32.mak(Magnus)
3084 Lo_read no longer stores trailing NULL(Bruce)
3085 Identifiers are now truncated to 31 characters internally(Bruce)
3086 Createuser options now availble on the command line
3087 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
3088 Prevent file descriptor leaf from failed COPY(Bruce)
3089 New pg_upgrade command(Bruce)
3090 Updated /contrib directories(Massimo)
3091 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
3092 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
3093 New DECLARE and FETCH feature(Thomas)
3094 libpq's internal structures now not exported(Tom)
3095 Allow up to 8 key indexes(Bruce)
3096 Remove ARCHIVE keyword, that is no longer used(Thomas)
3097 pg_dump -n flag to supress quotes around indentifiers
3098 disable system columns for views(Jan)
3099 new INET and CIDR types for network addresses(TomH, Paul)
3100 no more double quotes in psql output
3101 pg_dump now dumps views(Terry)
3102 new SET QUERY_LIMIT(Tatsuo,Jan)
3106 /contrib cleanup(Jun)
3107 Inline some small functions called for every row(Bruce)
3110 Multibyte regression tests(Soonmyung.)
3111 Remove --disabled options from configure
3112 Define PGDOC to use POSTGRESDIR by default
3113 Make regression optional
3114 Remove extra braces code to pgindent(Bruce)
3115 Add bsdi shared library support(Bruce)
3116 New --without-CXX support configure option(Brook)
3118 Update backend flowchart in tools/backend(Bruce)
3119 Change atttypmod from int16 to int32(Bruce, Tom)
3120 Getrusage() fix for platforms that do not have it(Tom)
3121 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
3122 NS32K platform fixes(Phil Nelson, John Buller)
3123 SCO 7/UnixWare 2.x fixes(Billy,others)
3124 Sparc/Solaris 2.5 fixes(Ryan)
3125 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
3126 Even more documention(Thomas)
3127 Nextstep support(Jacek)
3129 pginterface manual page(Bruce)
3130 shared libraries all have version numbers
3131 merged all OS-specific shared library defines into one file
3132 smarter TCL/TK configuration checking(Billy)
3133 smarter perl configuration(Brook)
3134 configure uses supplied install-sh if no install script found(Tom)
3135 new Makefile.shlib for shared library configuration(Tom)
3136 _________________________________________________________________
3140 Release date: 1998-04-07
3142 This is a bug-fix release for 6.3.x. Refer to the release notes for
3143 version 6.3 for a more complete summary of new features.
3147 * Repairs automatic configuration support for some platforms,
3148 including Linux, from breakage inadvertently introduced in version
3150 * Correctly handles function calls on the left side of BETWEEN and
3153 A dump/restore is NOT required for those running 6.3 or 6.3.1. A make
3154 distclean, make, and make install is all that is required. This last
3155 step should be performed while the postmaster is not running. You
3156 should re-link any custom applications that use PostgreSQL libraries.
3158 For upgrades from pre-6.3 installations, refer to the installation and
3159 migration instructions for version 6.3.
3160 _________________________________________________________________
3164 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
3165 Manual page improvements(Bruce)
3166 BETWEEN and LIKE fix(Thomas)
3167 fix for psql \connect used by pg_dump(Oliver Elphick)
3169 pgaccess, version 0.86
3170 qsort removed, now uses libc version, cleanups(Jeroen)
3171 fix for buffer over-runs detected(Maurice Gittens)
3172 fix for buffer overrun in libpgtcl(Randy Kunkee)
3173 fix for UNION with DISTINCT or ORDER BY(Bruce)
3174 gettimeofday configure check(Doug Winterburn)
3175 Fix "indexes not used" bug(Vadim)
3176 docs additions(Thomas)
3177 Fix for backend memory leak(Bruce)
3178 libreadline cleanup(Erwan MAS)
3179 Remove DISTDIR(Bruce)
3180 Makefile dependency cleanup(Jeroen van Vianen)
3183 _________________________________________________________________
3187 Release date: 1998-03-23
3191 * Additional support for multibyte character sets.
3192 * Repair byte ordering for mixed-endian clients and servers.
3193 * Minor updates to allowed SQL syntax.
3194 * Improvements to the configuration autodetection for installation.
3196 A dump/restore is NOT required for those running 6.3. A make
3197 distclean, make, and make install is all that is required. This last
3198 step should be performed while the postmaster is not running. You
3199 should re-link any custom applications that use PostgreSQL libraries.
3201 For upgrades from pre-6.3 installations, refer to the installation and
3202 migration instructions for version 6.3.
3203 _________________________________________________________________
3207 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
3208 pg_user cleanup(Bruce)
3209 large object fix for pg_dump and tclsh (alvin)
3210 LIKE fix for multiple adjacent underscores
3211 fix for redefining builtin functions(Thomas)
3213 upgrade to pg_access 0.83
3214 updated CLUSTER manual page
3215 multibyte character set support, see doc/README.mb(Tatsuo)
3216 configure --with-pgport fix
3218 big-endian fix for backend communications(Kataoka)
3219 SUBSTR() and substring() fix(Jan)
3220 several jdbc fixes(Peter)
3221 libpgtcl improvements, see libptcl/README(Randy Kunkee)
3222 Fix for "Datasize = 0" error(Vadim)
3223 Prevent \do from wrapping(Bruce)
3224 Remove duplicate Russian character set entries
3226 Allow optional TABLE keyword in LOCK and SELECT INTO(Thomas)
3227 CREATE SEQUENCE options to allow a negative integer(Thomas)
3228 Add "PASSWORD" as an allowed column identifier(Thomas)
3229 Add checks for UNION target fields(Bruce)
3230 Fix Alpha port(Dwayne Bailey)
3231 Fix for text arrays containing quotes(Doug Gibson)
3232 Solaris compile fix(Albert Chin-A-Young)
3233 Better identify tcl and tk libs and includes(Bruce)
3235 _________________________________________________________________
3239 Release date: 1998-03-01
3241 There are *many* new features and improvements in this release. Here
3242 is a brief, incomplete summary:
3244 * Many new SQL features, including full SQL92 subselect capability
3245 (everything is here but target-list subselects).
3246 * Support for client-side environment variables to specify time zone
3248 * Socket interface for client/server connection. This is the default
3249 now so you may need to start postmaster with the "-i" flag.
3250 * Better password authorization mechanisms. Default table
3251 permissions have changed.
3252 * Old-style time travel has been removed. Performance has been
3255 Note: Bruce Momjian wrote the following notes to introduce the new
3258 There are some general 6.3 issues that I want to mention. These are
3259 only the big items that can not be described in one sentence. A review
3260 of the detailed changes list is still needed.
3262 First, we now have subselects. Now that we have them, I would like to
3263 mention that without subselects, SQL is a very limited language.
3264 Subselects are a major feature, and you should review your code for
3265 places where subselects provide a better solution for your queries. I
3266 think you will find that there are more uses for subselects than you
3267 may think. Vadim has put us on the big SQL map with subselects, and
3268 fully functional ones too. The only thing you can't do with subselects
3269 is to use them in the target list.
3271 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
3272 enable connections from other machines, you have to use the new
3273 postmaster -i option, and of course edit "pg_hba.conf". Also, for this
3274 reason, the format of "pg_hba.conf" has changed.
3276 Third, char() fields will now allow faster access than varchar() or
3277 text. Specifically, the text and varchar() have a penalty for access
3278 to any columns after the first column of this type. char() used to
3279 also have this access penalty, but it no longer does. This may suggest
3280 that you redesign some of your tables, especially if you have short
3281 character columns that you have defined as varchar() or text. This and
3282 other changes make 6.3 even faster than earlier releases.
3284 We now have passwords definable independent of any Unix file. There
3285 are new SQL USER commands. See the Administrator's Guide for more
3286 information. There is a new table, pg_shadow, which is used to store
3287 user information and user passwords, and it by default only
3288 SELECT-able by the postgres super-user. pg_user is now a view of
3289 pg_shadow, and is SELECT-able by PUBLIC. You should keep using pg_user
3290 in your application without changes.
3292 User-created tables now no longer have SELECT permission to PUBLIC by
3293 default. This was done because the ANSI standard requires it. You can
3294 of course GRANT any permissions you want after the table is created.
3295 System tables continue to be SELECT-able by PUBLIC.
3297 We also have real deadlock detection code. No more sixty-second
3298 timeouts. And the new locking code implements a FIFO better, so there
3299 should be less resource starvation during heavy use.
3301 Many complaints have been made about inadequate documentation in
3302 previous releases. Thomas has put much effort into many new manuals
3303 for this release. Check out the doc/ directory.
3305 For performance reasons, time travel is gone, but can be implemented
3306 using triggers (see "pgsql/contrib/spi/README"). Please check out the
3307 new \d command for types, operators, etc. Also, views have their own
3308 permissions now, not based on the underlying tables, so permissions on
3309 them have to be set separately. Check "/pgsql/interfaces" for some new
3310 ways to talk to PostgreSQL.
3312 This is the first release that really required an explanation for
3313 existing users. In many ways, this was necessary because the new
3314 release removes many limitations, and the work-arounds people were
3315 using are no longer needed.
3316 _________________________________________________________________
3318 Migration to version 6.3
3320 A dump/restore using pg_dump or pg_dumpall is required for those
3321 wishing to migrate data from any previous release of PostgreSQL.
3322 _________________________________________________________________
3328 Fix binary cursors broken by MOVE implementation(Vadim)
3329 Fix for tcl library crash(Jan)
3330 Fix for array handling, from Gerhard Hintermayer
3331 Fix acl error, and remove duplicate pqtrace(Bruce)
3332 Fix psql \e for empty file(Bruce)
3333 Fix for textcat on varchar() fields(Bruce)
3334 Fix for DBT Sendproc (Zeugswetter Andres)
3335 Fix vacuum analyze syntax problem(Bruce)
3336 Fix for international identifiers(Tatsuo)
3337 Fix aggregates on inherited tables(Bruce)
3338 Fix substr() for out-of-bounds data
3339 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
3340 Fix notty output to show status result. -q option still turns it off(Bruce)
3341 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
3343 Fix for PQtrace start/stop several times(Bruce)
3344 Fix a variety of locking problems like newer lock waiters getting
3345 lock before older waiters, and having readlock people not share
3346 locks if a writer is waiting for a lock, and waiting writers not
3347 getting priority over waiting readers(Bruce)
3348 Fix crashes in psql when executing queries from external files(James)
3349 Fix problem with multiple order by columns, with the first one having
3351 Use correct hash table support functions for float8 and int4(Thomas)
3352 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
3353 Change precedence for boolean operators to match expected behavior(Thomas)
3354 Generate elog(ERROR) on over-large integer(Bruce)
3355 Allow multiple-argument functions in constraint clauses(Thomas)
3356 Check boolean input literals for 'true','false','yes','no','1','0'
3357 and throw elog(ERROR) if unrecognized(Thomas)
3358 Major large objects fix
3359 Fix for GROUP BY showing duplicates(Vadim)
3360 Fix for index scans in MergeJion(Vadim)
3364 Subselects with EXISTS, IN, ALL, ANY keywords (Vadim, Bruce, Thomas)
3365 New User Manual(Thomas, others)
3366 Speedup by inlining some frequently-called functions
3367 Real deadlock detection, no more timeouts(Bruce)
3368 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
3369 CURRENT_USER(Thomas)
3370 Modify constraint syntax to be SQL92-compliant(Thomas)
3371 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
3372 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
3373 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas
3375 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
3376 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
3377 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
3378 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
3379 Allow SQL92 delimited identifiers(Thomas)
3380 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas
3382 Support SQL92 syntax for type coercion of literal strings
3383 (e.g. "DATETIME 'now'")(Thomas)
3384 Add conversions for int2, int4, and OID types to and from text(Thomas)
3385 Use shared lock when building indexes(Vadim)
3386 Free memory allocated for an user query inside transaction block after
3387 this query is done, was turned off in <= 6.2.1(Vadim)
3388 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
3389 New PostgreSQL Procedural Language (PL) backend interface(Jan)
3390 Rename pg_dump -H option to -h(Bruce)
3391 Add Java support for passwords, European dates(Peter)
3392 Use indexes for LIKE and ~, !~ operations(Bruce)
3393 Add hash functions for datetime and timespan(Thomas)
3394 Time Travel removed(Vadim, Bruce)
3395 Add paging for \d and \z, and fix \i(Bruce)
3396 Add Unix domain socket support to backend and to frontend library(Goran)
3397 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
3398 Allow more SQL92 and/or PostgreSQL reserved words as column identifiers(Thomas)
3399 Augment support for SQL92 SET TIME ZONE...(Thomas)
3400 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
3401 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
3402 Enable SET TIME ZONE using TZ environment variable(Thomas)
3403 Add PGDATESTYLE environment variable to frontend and backend initialization(Tho
3405 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
3406 frontend library initialization environment variables(Thomas)
3407 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
3408 Add pg_description table for info on tables, columns, operators, types, and
3410 Increase 16 char limit on system table/index names to 32 characters(Bruce)
3411 Rename system indexes(Bruce)
3412 Add 'GERMAN' option to SET DATESTYLE(Thomas)
3413 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
3414 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
3415 Validate numeric input more carefully for delta times(Thomas)
3416 Implement day of year as possible input to date_part()(Thomas)
3417 Define timespan_finite() and text_timespan() functions(Thomas)
3418 Remove archive stuff(Bruce)
3419 Allow for a pg_password authentication database that is separate from
3420 the system password file(Todd)
3421 Dump ACLs, GRANT, REVOKE permissions(Matt)
3422 Define text, varchar, and bpchar string length functions(Thomas)
3423 Fix Query handling for inheritance, and cost computations(Bruce)
3424 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
3425 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
3426 Implement UNIONs for SELECT(Bruce)
3427 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
3428 varchar() stores only necessary bytes on disk(Bruce)
3429 Fix for BLOBs(Peter)
3430 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
3431 Remove unused "option" from PQconnectdb()
3432 New LOCK command and lock manual page describing deadlocks(Bruce)
3433 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
3434 Enhance psql \z to show sequences(Bruce)
3435 Show NOT NULL and DEFAULT in psql \d table(Bruce)
3436 New psql .psqlrc file start-up(Andrew)
3437 Modify sample start-up script in contrib/linux to show syslog(Thomas)
3438 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
3439 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
3440 Update of contrib stuff(Massimo)
3441 Add Unix socket support to DBD::Pg(Goran)
3442 New python interface (PyGreSQL 2.0)(D'Arcy)
3443 New frontend/backend protocol has a version number, network byte order(Phil)
3444 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
3445 CHAR() now faster access than VARCHAR() or TEXT
3446 ecpg embedded SQL preprocessor
3447 Reduce system column overhead(Vadmin)
3448 Remove pg_time table(Vadim)
3449 Add pg_type attribute to identify types that need length (bpchar, varchar)
3450 Add report of offending line when COPY command fails
3451 Allow VIEW permissions to be set separately from the underlying tables.
3452 For security, use GRANT/REVOKE on views as appropriate(Jan)
3453 Tables now have no default GRANT SELECT TO PUBLIC. You must
3454 explicitly grant such permissions.
3455 Clean up tutorial examples(Darren)
3459 Add new html development tools, and flow chart in /tools/backend
3460 Fix for SCO compiles
3461 Stratus computer port Robert Gillies
3462 Added support for shlib for BSD44_derived & i386_solaris
3463 Make configure more automated(Brook)
3464 Add script to check regression test results
3465 Break parser functions into smaller files, group together(Bruce)
3466 Rename heap_create to heap_create_and_catalog, rename heap_creatr
3467 to heap_create()(Bruce)
3468 Sparc/Linux patch for locking(TomS)
3469 Remove PORTNAME and reorganize port-specific stuff(Marc)
3470 Add optimizer README file(Bruce)
3471 Remove some recursion in optimizer and clean up some code there(Bruce)
3472 Fix for NetBSD locking(Henry)
3473 Fix for libptcl make(Tatsuo)
3475 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
3476 function calls to istrue() or isfalse() to allow optimization(Thomas)
3477 Various fixes NetBSD/Sparc related(TomH)
3478 Alpha linux locking(Travis,Ryan)
3479 Change elog(WARN) to elog(ERROR)(Bruce)
3480 FAQ for FreeBSD(Marc)
3481 Bring in the PostODBC source tree as part of our standard distribution(Marc)
3482 A minor patch for HP/UX 10 vs 9(Stan)
3483 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
3484 UnixWare patches(Billy)
3485 New i386 'lock' for spin lock asm(Billy)
3486 Support for multiplexed backends is removed
3487 Start an OpenBSD port
3490 Add string functions to regression suite(Thomas)
3491 Expand a few function names formerly truncated to 16 characters(Thomas)
3492 Remove un-needed malloc() calls and replace with palloc()(Bruce)
3493 _________________________________________________________________
3497 Release date: 1997-10-17
3499 6.2.1 is a bug-fix and usability release on 6.2.
3503 * Allow strings to span lines, per SQL92.
3504 * Include example trigger function for inserting user names on table
3507 This is a minor bug-fix release on 6.2. For upgrades from pre-6.2
3508 systems, a full dump/reload is required. Refer to the 6.2 release
3509 notes for instructions.
3510 _________________________________________________________________
3512 Migration from version 6.2 to version 6.2.1
3514 This is a minor bug-fix release. A dump/reload is not required from
3515 version 6.2, but is required from any release prior to 6.2.
3517 In upgrading from version 6.2, if you choose to dump/reload you will
3518 find that avg(money) is now calculated correctly. All other bug fixes
3519 take effect upon updating the executables.
3521 Another way to avoid dump/reload is to use the following SQL command
3522 from "psql" to update the existing system table:
3523 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
3524 where aggname = 'avg' and aggbasetype = 790;
3526 This will need to be done to every existing database, including
3528 _________________________________________________________________
3532 Allow TIME and TYPE column names(Thomas)
3533 Allow larger range of true/false as boolean values(Thomas)
3534 Support output of "now" and "current"(Thomas)
3535 Handle DEFAULT with INSERT of NULL properly(Vadim)
3536 Fix for relation reference counts problem in buffer manager(Vadim)
3537 Allow strings to span lines, like ANSI(Thomas)
3538 Fix for backward cursor with ORDER BY(Vadim)
3539 Fix avg(cash) computation(Thomas)
3540 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
3541 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
3542 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
3544 _________________________________________________________________
3548 Release date: 1997-10-02
3550 A dump/restore is required for those wishing to migrate data from
3551 previous releases of PostgreSQL.
3552 _________________________________________________________________
3554 Migration from version 6.1 to version 6.2
3556 This migration requires a complete dump of the 6.1 database and a
3557 restore of the database in 6.2.
3559 Note that the "pg_dump" and "pg_dumpall" utility from 6.2 should be
3560 used to dump the 6.1 database.
3561 _________________________________________________________________
3563 Migration from version 1.x to version 6.2
3565 Those migrating from earlier 1.* releases should first upgrade to 1.09
3566 because the COPY output format was improved from the 1.02 release.
3567 _________________________________________________________________
3573 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
3574 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
3575 from Solaris(Diab Jerius)
3576 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
3577 Check for geometric intersections at endpoints to avoid rounding ugliness(Thoma
3579 Catch non-functional delete attempts(Vadim)
3580 Change time function names to be more consistent(Michael Reifenberg)
3581 Check for zero divides(Michael Reifenberg)
3582 Fix very old bug which made tuples changed/inserted by a commnd
3583 visible to the command itself (so we had multiple update of
3584 updated tuples, etc)(Vadim)
3585 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
3586 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
3587 Remove un-needed signal stuff from contrib/pginterface
3588 Fix OR (where x != 1 or x isnull didn't return tuples with x NULL) (Vadim)
3589 Fix time_cmp function (Vadim)
3590 Fix handling of functions with non-attribute first argument in
3591 WHERE clauses (Vadim)
3592 Fix GROUP BY when order of entries is different from order
3593 in target list (Vadim)
3594 Fix pg_dump for aggregates without sfunc1 (Vadim)
3598 Default genetic optimizer GEQO parameter is now 8(Bruce)
3599 Allow use parameters in target list having aggregates in functions(Vadim)
3600 Added JDBC driver as an interface(Adrian & Peter)
3602 Return number of tuples inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
3603 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
3604 SPI (Server Programming Interface) allows execution of queries inside
3606 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
3607 Include reserved words for string handling, outer joins, and unions(Thomas)
3608 Implement extended comments ("/* ... */") using exclusive states(Thomas)
3609 Add "//" single-line comments(Bruce)
3610 Remove some restrictions on characters in operator names(Thomas)
3611 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
3612 Add text concatenation operator and function (SQL92)(Thomas)
3613 Support WITH TIME ZONE syntax (SQL92)(Thomas)
3614 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
3615 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
3616 and CHARACTER VARYING (SQL92)(Thomas)
3617 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
3618 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
3619 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
3620 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
3621 Add more reserved words, mostly for SQL92 compliance(Thomas)
3622 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
3623 Add center() routines for lseg, path, polygon(Thomas)
3624 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
3625 Check explicitly for points and polygons contained within polygons
3626 using an axis-crossing algorithm(Thomas)
3627 Add routine to convert circle-box(Thomas)
3628 Merge conflicting operators for different geometric data types(Thomas)
3629 Replace distance operator "<===>" with "<->"(Thomas)
3630 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Tho
3632 Add routines for text trimming on both ends, substring, and string position(Tho
3634 Added conversion routines circle(box) and poly(circle)(Thomas)
3635 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
3636 Allow functions and operators on internally-identical types to succeed(Bruce)
3637 Speed up backend start-up after profiling analysis(Bruce)
3638 Inline frequently called functions for performance(Bruce)
3639 Reduce open() calls(Bruce)
3640 psql: Add PAGER for \h and \?,\C fix
3641 Fix for psql pager when no tty(Bruce)
3642 New entab utility(Bruce)
3643 General trigger functions for referential integrity (Vadim)
3644 General trigger functions for time travel (Vadim)
3645 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
3646 MOVE implementation (Vadim)
3650 HP-UX 10 patches (Vladimir Turin)
3651 Added SCO support, (Daniel Harris)
3652 MkLinux patches (Tatsuo Ishii)
3653 Change geometric box terminology from "length" to "width"(Thomas)
3654 Deprecate temporary unstored slope fields in geometric code(Thomas)
3655 Remove restart instructions from INSTALL(Bruce)
3656 Look in /usr/ucb first for install(Bruce)
3657 Fix c++ copy example code(Thomas)
3658 Add -o to psql manual page(Bruce)
3659 Prevent relname unallocated string length from being copied into database(Bruce
3661 Cleanup for NAMEDATALEN use(Bruce)
3662 Fix pg_proc names over 15 chars in output(Bruce)
3663 Add strNcpy() function(Bruce)
3664 remove some (void) casts that are unnecessary(Bruce)
3665 new interfaces directory(Marc)
3666 Replace fopen() calls with calls to fd.c functions(Bruce)
3667 Make functions static where possible(Bruce)
3668 enclose unused functions in #ifdef NOT_USED(Bruce)
3669 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
3670 Changes for Digital Unix
3671 Portability fix for pg_dumpall(Bruce)
3672 Rename pg_attribute.attnvals to attdispersion(Bruce)
3673 "intro/unix" manual page now "pgintro"(Bruce)
3674 "built-in" manual page now "pgbuiltin"(Bruce)
3675 "drop" manual page now "drop_table"(Bruce)
3676 Add "create_trigger", "drop_trigger" manual pages(Thomas)
3677 Add constraints regression test(Vadim & Thomas)
3678 Add comments syntax regression test(Thomas)
3679 Add PGINDENT and support program(Bruce)
3680 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
3681 Files moved to /src/tools directory(Bruce)
3682 SPI and Trigger programming guides (Vadim & D'Arcy)
3683 _________________________________________________________________
3687 Release date: 1997-07-22
3688 _________________________________________________________________
3690 Migration from version 6.1 to version 6.1.1
3692 This is a minor bug-fix release. A dump/reload is not required from
3693 version 6.1, but is required from any release prior to 6.1. Refer to
3694 the release notes for 6.1 for more details.
3695 _________________________________________________________________
3699 fix for SET with options (Thomas)
3700 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
3701 new psql \connect option allows changing usernames without changing databases
3702 fix for initdb --debug option(Yoshihiko Ichikawa))
3703 lextest cleanup(Bruce)
3705 fix date/time month boundary arithmetic(Thomas)
3706 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
3707 timestamp overhauled to use standard functions(Thomas)
3708 other code cleanup in date/time routines(Thomas)
3709 psql's \d now case-insensitive(Bruce)
3710 psql's backslash commands can now have trailing semicolon(Bruce)
3711 fix memory leak in psql when using \g(Bruce)
3712 major fix for endian handling of communication to server(Thomas, Tatsuo)
3713 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
3714 allow underscores in usernames(Bruce)
3715 pg_dumpall now returns proper status, portability fix(Bruce)
3717 _________________________________________________________________
3721 Release date: 1997-06-08
3723 The regression tests have been adapted and extensively modified for
3724 the 6.1 release of PostgreSQL.
3726 Three new data types (datetime, timespan, and circle) have been added
3727 to the native set of PostgreSQL types. Points, boxes, paths, and
3728 polygons have had their output formats made consistent across the data
3729 types. The polygon output in misc.out has only been spot-checked for
3730 correctness relative to the original regression output.
3732 PostgreSQL 6.1 introduces a new, alternate optimizer which uses
3733 genetic algorithms. These algorithms introduce a random behavior in
3734 the ordering of query results when the query contains multiple
3735 qualifiers or multiple tables (giving the optimizer a choice on order
3736 of evaluation). Several regression tests have been modified to
3737 explicitly order the results, and hence are insensitive to optimizer
3738 choices. A few regression tests are for data types which are
3739 inherently unordered (e.g. points and time intervals) and tests
3740 involving those types are explicitly bracketed with "set geqo to
3741 'off'" and "reset geqo".
3743 The interpretation of array specifiers (the curly braces around atomic
3744 values) appears to have changed sometime after the original regression
3745 tests were generated. The current "./expected/*.out" files reflect
3746 this new interpretation, which may not be correct!
3748 The float8 regression test fails on at least some platforms. This is
3749 due to differences in implementations of pow() and exp() and the
3750 signaling mechanisms used for overflow and underflow conditions.
3752 The "random" results in the random test should cause the "random" test
3753 to be "failed", since the regression tests are evaluated using a
3754 simple diff. However, "random" does not seem to produce random results
3755 on my test machine (Linux/gcc/i686).
3756 _________________________________________________________________
3758 Migration to version 6.1
3760 This migration requires a complete dump of the 6.0 database and a
3761 restore of the database in 6.1.
3763 Those migrating from earlier 1.* releases should first upgrade to 1.09
3764 because the COPY output format was improved from the 1.02 release.
3765 _________________________________________________________________
3771 packet length checking in library routines
3772 lock manager priority patch
3773 check for under/over flow of float8(Bruce)
3774 multitable join fix(Vadim)
3775 SIGPIPE crash fix(Darren)
3776 large object fixes(Sven)
3777 allow btree indexes to handle NULLs(Vadim)
3778 timezone fixes(D'Arcy)
3779 select SUM(x) can return NULL on no rows(Thomas)
3780 internal optimizer, executor bug fixes(Vadim)
3781 fix problem where inner loop in < or <= has no rows(Vadim)
3782 prevent re-commuting join index clauses(Vadim)
3783 fix join clauses for multiple tables(Vadim)
3784 fix hash, hashjoin for arrays(Vadim)
3785 fix btree for abstime type(Vadim)
3786 large object fixes(Raymond)
3787 fix buffer leak in hash indexes (Vadim)
3788 fix rtree for use in inner scan (Vadim)
3789 fix gist for use in inner scan, cleanups (Vadim, Andrea)
3790 avoid unnecessary local buffers allocation (Vadim, Massimo)
3791 fix local buffers leak in transaction aborts (Vadim)
3792 fix file manager memmory leaks, cleanups (Vadim, Massimo)
3793 fix storage manager memmory leaks (Vadim)
3794 fix btree duplicates handling (Vadim)
3795 fix deleted tuples re-incarnation caused by vacuum (Vadim)
3796 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
3797 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
3801 attribute optimization statistics(Bruce)
3802 much faster new btree bulk load code(Paul)
3803 BTREE UNIQUE added to bulk load code(Vadim)
3804 new lock debug code(Massimo)
3805 massive changes to libpg++(Leo)
3806 new GEQO optimizer speeds table multitable optimization(Martin)
3807 new WARN message for non-unique insert into unique key(Marc)
3808 update x=-3, no spaces, now valid(Bruce)
3809 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
3810 debug backend now pretty-prints tree(Darren)
3811 new Oracle character functions(Edmund)
3812 new plaintext password functions(Dan)
3813 no such class or insufficient privilege changed to distinct messages(Dan)
3814 new ANSI timestamp function(Dan)
3815 new ANSI Time and Date types (Thomas)
3816 move large chunks of data in backend(Martin)
3817 multicolumn btree indexes(Vadim)
3818 new SET var TO value command(Martin)
3819 update transaction status on reads(Dan)
3820 new locale settings for character types(Oleg)
3821 new SEQUENCE serial number generator(Vadim)
3822 GROUP BY function now possible(Vadim)
3823 re-organize regression test(Thomas,Marc)
3824 new optimizer operation weights(Vadim)
3825 new psql \z grant/permit option(Marc)
3826 new MONEY data type(D'Arcy,Thomas)
3827 tcp socket communication speed improved(Vadim)
3828 new VACUUM option for attribute statistics, and for certain columns (Vadim)
3829 many geometric type improvements(Thomas,Keith)
3830 additional regression tests(Thomas)
3831 new datestyle variable(Thomas,Vadim,Martin)
3832 more comparison operators for sorting types(Thomas)
3833 new conversion functions(Thomas)
3834 new more compact btree format(Vadim)
3835 allow pg_dumpall to preserve database ownership(Bruce)
3836 new SET GEQO=# and R_PLANS variable(Vadim)
3837 old (!GEQO) optimizer can use right-sided plans (Vadim)
3838 typechecking improvement in SQL parser(Bruce)
3839 new SET, SHOW, RESET commands(Thomas,Vadim)
3840 new \connect database USER option
3841 new destroydb -i option (Igor)
3842 new \dt and \di psql commands (Darren)
3843 SELECT "\n" now escapes newline (A. Duursma)
3844 new geometry conversion functions from old format (Thomas)
3848 new configuration script(Marc)
3849 readline configuration option added(Marc)
3850 OS-specific configuration options removed(Marc)
3851 new OS-specific template files(Marc)
3852 no more need to edit Makefile.global(Marc)
3853 re-arrange include files(Marc)
3854 nextstep patches (Gregor Hoffleit)
3855 removed WIN32-specific code(Bruce)
3856 removed postmaster -e option, now only postgres -e option (Bruce)
3857 merge duplicate library code in front/backends(Martin)
3858 now works with eBones, international Kerberos(Jun)
3859 more shared library support
3860 c++ include file cleanup(Bruce)
3861 warn about buggy flex(Bruce)
3862 DG/UX, Ultrix, IRIX, AIX portability fixes
3863 _________________________________________________________________
3867 Release date: 1997-01-29
3869 A dump/restore is required for those wishing to migrate data from
3870 previous releases of PostgreSQL.
3871 _________________________________________________________________
3873 Migration from version 1.09 to version 6.0
3875 This migration requires a complete dump of the 1.09 database and a
3876 restore of the database in 6.0.
3877 _________________________________________________________________
3879 Migration from pre-1.09 to version 6.0
3881 Those migrating from earlier 1.* releases should first upgrade to 1.09
3882 because the COPY output format was improved from the 1.02 release.
3883 _________________________________________________________________
3889 ALTER TABLE bug - running postgress process needs to re-read table definition
3890 Allow vacuum to be run on one table or entire database(Bruce)
3892 Fix array over-runs of memory writes(Kurt)
3893 Fix elusive btree range/non-range bug(Dan)
3894 Fix for hash indexes on some types like time and date
3895 Fix for pg_log size explosion
3896 Fix permissions on lo_export()(Bruce)
3897 Fix unitialized reads of memory(Kurt)
3898 Fixed ALTER TABLE ... char(3) bug(Bruce)
3899 Fixed a few small memory leaks
3900 Fixed EXPLAIN handling of options and changed full_path option name
3901 Fixed output of group acl permissions
3902 Memory leaks (hunt and destroy with tools like Purify(Kurt)
3903 Minor improvements to rules system
3905 New asserts for run-checking
3906 Overhauled parser/analyze code to properly report errors and increase speed
3907 Pg_dump -d now handles NULL's properly(Bruce)
3908 Prevent SELECT NULL from crashing server (Bruce)
3909 Properly report errors when INSERT ... SELECT columns did not match
3910 Properly report errors when insert column names were not correct
3911 Psql \g filename now works(Bruce)
3912 Psql fixed problem with multiple statements on one line with multiple outputs
3913 Removed duplicate system OIDs
3914 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
3915 Several fixes for queries that crashed the backend
3916 Starting quote in insert string errors(Bruce)
3917 Submitting an empty query now returns empty status, not just " " query(Bruce)
3921 Add EXPLAIN manual page(Bruce)
3922 Add UNIQUE index capability(Dan)
3923 Add hostname/user level access control rather than just hostname and user
3924 Add synonym of != for <>(Bruce)
3925 Allow "select oid,* from table"
3926 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Br
3928 Allow COPY from the frontend(Bryan)
3929 Allow GROUP BY to use alias column name(Bruce)
3930 Allow actual compression, not just reuse on the same page(Vadim)
3931 Allow installation-configuration option to auto-add all local users(Bryan)
3932 Allow libpq to distinguish between text value '' and null(Bruce)
3933 Allow non-postgres users with createdb privs to destroydb's
3934 Allow restriction on who can create C functions(Bryan)
3935 Allow restriction on who can do backend COPY(Bryan)
3936 Can shrink tables, pg_time and pg_log(Vadim & Erich)
3937 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
3938 Change default decimal constant representation from float4 to float8(Bruce)
3939 European date format now set when postmaster is started
3940 Execute lowercase function names if not found with exact case
3941 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z
3943 Gist now included in the distrubution(Marc)
3944 Idend authentication of local users(Bryan)
3945 Implement BETWEEN qualifier(Bruce)
3946 Implement IN qualifier(Bruce)
3947 Libpq has PQgetisnull()(Bruce)
3948 Libpq++ improvements
3949 New options to initdb(Bryan)
3950 Pg_dump allow dump of OIDs(Bruce)
3951 Pg_dump create indexes after tables are loaded for speed(Bruce)
3952 Pg_dumpall dumps all databases, and the user table
3953 Pginterface additions for NULL values(Bruce)
3954 Prevent postmaster from being run as root
3955 Psql \h and \? is now readable(Bruce)
3956 Psql allow backslashed, semicolons anywhere on the line(Bruce)
3957 Psql changed command prompt for lines in query or in quotes(Bruce)
3958 Psql char(3) now displays as (bp)char in \d output(Bruce)
3959 Psql return code now more accurate(Bryan?)
3960 Psql updated help syntax(Bruce)
3961 Re-visit and fix vacuum(Vadim)
3962 Reduce size of regression diffs, remove timezone name difference(Bruce)
3963 Remove compile-time parameters to enable binary distributions(Bryan)
3964 Reverse meaning of HBA masks(Bryan)
3965 Secure Authentication of local users(Bryan)
3966 Speed up vacuum(Vadim)
3967 Vacuum now had VERBOSE option(Bruce)
3971 All functions now have prototypes that are compared against the calls
3972 Allow asserts to be disabled easly from Makefile.global(Bruce)
3973 Change oid constants used in code to #define names
3974 Decoupled sparc and solaris defines(Kurt)
3975 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
3976 Major include file reorganization/reduction(Marc)
3977 Make now stops on compile failure(Bryan)
3978 Makefile restructuring(Bryan, Marc)
3979 Merge bsdi_2_1 to bsdi(Bruce)
3980 Monitor program removed
3981 Name change from Postgres95 to PostgreSQL
3982 New config.h file(Marc, Bryan)
3983 PG_VERSION now set to 6.0 and used by postmaster
3984 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
3985 Reduced the number of #define's, centeralized #define's
3986 Remove duplicate OIDS in system tables(Dan)
3987 Remove duplicate system catalog info or report mismatches(Dan)
3988 Removed many os-specific #define's
3989 Restructured object file generation/location(Bryan, Marc)
3990 Restructured port-specific file locations(Bryan, Marc)
3991 Unused/uninialized variables corrected
3992 _________________________________________________________________
3996 Release date: 1996-11-04
3998 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of the
3999 changes listed in 6.0 were actually included in the 1.02.1 to 1.09
4001 _________________________________________________________________
4005 Release date: 1996-08-01
4006 _________________________________________________________________
4008 Migration from version 1.02 to version 1.02.1
4010 Here is a new migration file for 1.02.1. It includes the 'copy' change
4011 and a script to convert old ASCII files.
4013 Note: The following notes are for the benefit of users who want to
4014 migrate databases from Postgres95 1.01 and 1.02 to Postgres95
4017 If you are starting afresh with Postgres95 1.02.1 and do not need
4018 to migrate old databases, you do not need to read any further.
4020 In order to upgrade older Postgres95 version 1.01 or 1.02 databases to
4021 version 1.02.1, the following steps are required:
4022 1. Start up a new 1.02.1 postmaster
4023 2. Add the new built-in functions and operators of 1.02.1 to 1.01 or
4024 1.02 databases. This is done by running the new 1.02.1 server
4025 against your own 1.01 or 1.02 database and applying the queries
4026 attached at the end of the file. This can be done easily through
4027 "psql". If your 1.01 or 1.02 database is named testdb and you have
4028 cut the commands from the end of this file and saved them in
4030 % psql testdb -f addfunc.sql
4031 Those upgrading 1.02 databases will get a warning when executing
4032 the last two statements in the file because they are already
4033 present in 1.02. This is not a cause for concern.
4034 _________________________________________________________________
4036 Dump/Reload Procedure
4038 If you are trying to reload a pg_dump or text-mode, copy tablename to
4039 stdout generated with a previous version, you will need to run the
4040 attached "sed" script on the ASCII file before loading it into the
4041 database. The old format used '.' as end-of-data, while '\.' is now
4042 the end-of-data marker. Also, empty strings are now loaded in as ''
4043 rather than NULL. See the copy manual page for full details.
4044 sed 's/^\.$/\\./g' <in_file >out_file
4046 If you are loading an older binary copy or non-stdout copy, there is
4047 no end-of-data character, and hence no conversion necessary.
4048 -- following lines added by agc to reflect the case-insensitive
4049 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
4050 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexe
4052 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregex
4054 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregex
4056 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticrege
4058 _________________________________________________________________
4062 Source code maintenance and development
4063 * worldwide team of volunteers
4064 * the source tree now in CVS at ftp.ki.net
4067 * psql (and underlying libpq library) now has many more options for
4068 formatting output, including HTML
4069 * pg_dump now output the schema and/or the data, with many fixes to
4070 enhance completeness.
4071 * psql used in place of monitor in administration shell scripts.
4072 monitor to be deprecated in next release.
4073 * date/time functions enhanced
4074 * NULL insert/update/comparison fixed/enhanced
4075 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
4077 Bug Fixes (almost too numerous to mention)
4079 * storage management
4080 * check for NULL pointer before dereferencing
4084 * added SolarisX86 port
4085 * added BSD/OS 2.1 port
4087 _________________________________________________________________
4091 Release date: 1996-02-23
4092 _________________________________________________________________
4094 Migration from version 1.0 to version 1.01
4096 The following notes are for the benefit of users who want to migrate
4097 databases from Postgres95 1.0 to Postgres95 1.01.
4099 If you are starting afresh with Postgres95 1.01 and do not need to
4100 migrate old databases, you do not need to read any further.
4102 In order to Postgres95 version 1.01 with databases created with
4103 Postgres95 version 1.0, the following steps are required:
4104 1. Set the definition of NAMEDATALEN in "src/Makefile.global" to 16
4105 and OIDNAMELEN to 20.
4106 2. Decide whether you want to use Host based authentication.
4107 a. If you do, you must create a file name pg_hba in your
4108 top-level data directory (typically the value of your
4109 $PGDATA). "src/libpq/pg_hba" shows an example syntax.
4110 b. If you do not want host-based authentication, you can comment
4113 in "src/Makefile.global"
4114 Note that host-based authentication is turned on by default,
4115 and if you do not take steps A or B above, the out-of-the-box
4116 1.01 will not allow you to connect to 1.0 databases.
4117 3. Compile and install 1.01, but DO NOT do the "initdb" step.
4118 4. Before doing anything else, terminate your 1.0 postmaster, and
4119 backup your existing $PGDATA directory.
4120 5. Set your PGDATA environment variable to your 1.0 databases, but
4121 set up path up so that 1.01 binaries are being used.
4122 6. Modify the file "$PGDATA/PG_VERSION" from 5.0 to 5.1
4123 7. Start up a new 1.01 postmaster
4124 8. Add the new built-in functions and operators of 1.01 to 1.0
4125 databases. This is done by running the new 1.01 server against
4126 your own 1.0 database and applying the queries attached and saving
4127 in the file 1.0_to_1.01.sql. This can be done easily through
4128 "psql". If your 1.0 database is name testdb:
4129 % psql testdb -f 1.0_to_1.01.sql
4130 and then execute the following commands (cut and paste from here):
4131 -- add builtin functions that are new to 1.01
4133 create function int4eqoid (int4, oid) returns bool as 'foo'
4134 language 'internal';
4135 create function oideqint4 (oid, int4) returns bool as 'foo'
4136 language 'internal';
4137 create function char2icregexeq (char2, text) returns bool as 'foo'
4138 language 'internal';
4139 create function char2icregexne (char2, text) returns bool as 'foo'
4140 language 'internal';
4141 create function char4icregexeq (char4, text) returns bool as 'foo'
4142 language 'internal';
4143 create function char4icregexne (char4, text) returns bool as 'foo'
4144 language 'internal';
4145 create function char8icregexeq (char8, text) returns bool as 'foo'
4146 language 'internal';
4147 create function char8icregexne (char8, text) returns bool as 'foo'
4148 language 'internal';
4149 create function char16icregexeq (char16, text) returns bool as 'foo'
4150 language 'internal';
4151 create function char16icregexne (char16, text) returns bool as 'foo'
4152 language 'internal';
4153 create function texticregexeq (text, text) returns bool as 'foo'
4154 language 'internal';
4155 create function texticregexne (text, text) returns bool as 'foo'
4156 language 'internal';
4158 -- add builtin functions that are new to 1.01
4160 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
4161 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
4162 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexe
4164 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregex
4166 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexe
4168 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregex
4170 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexe
4172 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregex
4174 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icrege
4176 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icreg
4178 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq)
4180 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne
4182 _________________________________________________________________
4187 * 1.01 is backwards compatible with 1.0 database provided the user
4188 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
4189 If those steps are not taken, 1.01 is not compatible with 1.0 database.
4192 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
4193 * added NeXT port (requires SysVIPC implementation)
4194 * added CAST .. AS ... syntax
4195 * added ASC and DESC keywords
4196 * added 'internal' as a possible language for CREATE FUNCTION
4197 internal functions are C functions which have been statically linked
4198 into the postgres backend.
4199 * a new type "name" has been added for system identifiers (table names,
4200 attribute names, etc.) This replaces the old char16 type. The
4201 of name is set by the NAMEDATALEN #define in src/Makefile.global
4202 * a readable reference manual that describes the query language.
4203 * added host-based access control. A configuration file ($PGDATA/pg_hba)
4204 is used to hold the configuration data. If host-based access control
4205 is not desired, comment out HBA=1 in src/Makefile.global.
4206 * changed regex handling to be uniform use of Henry Spencer's regex code
4207 regardless of platform. The regex code is included in the distribution
4208 * added functions and operators for case-insensitive regular expressions.
4209 The operators are ~* and !~*.
4210 * pg_dump uses COPY instead of SELECT loop for better performance
4213 * fixed an optimizer bug that was causing core dumps when
4214 functions calls were used in comparisons in the WHERE clause
4215 * changed all uses of getuid to geteuid so that effective uids are used
4216 * psql now returns non-zero status on errors when using -c
4217 * applied public patches 1-14
4218 _________________________________________________________________
4222 Release date: 1995-09-05
4223 _________________________________________________________________
4228 * The copyright of Postgres 1.0 has been loosened to be freely modifiable
4229 and modifiable for any purpose. Please read the COPYRIGHT file.
4230 Thanks to Professor Michael Stonebraker for making this possible.
4233 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
4234 EUROPEAN STYLE). This follows SQL-92 specs.
4235 * "delimiters" is now a keyword
4238 * sql LIKE syntax has been added
4239 * copy command now takes an optional USING DELIMITER specification.
4240 delimiters can be any single-character string.
4241 * IRIX 5.3 port has been added.
4242 Thanks to Paul Walmsley and others.
4243 * updated pg_dump to work with new libpq
4244 * \d has been added psql
4245 Thanks to Keith Parks
4246 * regexp performance for architectures that use POSIX regex has been
4247 improved due to caching of precompiled patterns.
4248 Thanks to Alistair Crooks
4249 * a new version of libpq++
4250 Thanks to William Wanders
4253 * arbitrary userids can be specified in the createuser script
4254 * \c to connect to other databases in psql now works.
4255 * bad pg_proc entry for float4inc() is fixed
4256 * users with usecreatedb field set can now create databases without
4257 having to be usesuper
4258 * remove access control entries when the entry no longer has any
4260 * fixed non-portable datetimes implementation
4261 * added kerberos flags to the src/backend/Makefile
4262 * libpq now works with kerberos
4263 * typographic errors in the user manual have been corrected.
4264 * btrees with multiple index never worked, now we tell you they don't
4265 work when you try to use them
4266 _________________________________________________________________
4268 Postgres95 Release 0.03
4270 Release date: 1995-07-21
4271 _________________________________________________________________
4275 Incompatible changes:
4276 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
4277 (due to system catalog changes and indexing structure changes).
4278 * double-quote (") is deprecated as a quoting character for string literals;
4279 you need to convert them to single quotes (').
4280 * name of aggregates (eg. int4sum) are renamed in accordance with the
4281 SQL standard (eg. sum).
4282 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
4283 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
4284 previous releases); you might have to do typecasting if you depend on it
4285 being of type float8. If you neglect to do the typecasting and you assign
4286 a float literal to a field of type float8, you may get incorrect values
4288 * LIBPQ has been totally revamped so that frontend applications
4289 can connect to multiple backends
4290 * the usesysid field in pg_user has been changed from int2 to int4 to
4291 allow wider range of Unix user ids.
4292 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
4293 single BSD44_derived port. (thanks to Alistair Crooks)
4295 SQL standard-compliance (the following details changes that makes postgres95
4296 more compliant to the SQL-92 standard):
4297 * the following SQL types are now built-in: smallint, int(eger), float, real,
4298 char(N), varchar(N), date and time.
4300 The following are aliases to existing postgres types:
4302 integer, int -> int4
4303 float, real -> float4
4304 char(N) and varchar(N) are implemented as truncated text types. In
4305 addition, char(N) does blank-padding.
4306 * single-quote (') is used for quoting string literals; '' (in addition to
4307 \') is supported as means of inserting a single quote in a string
4308 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
4309 (Also, aggregates can now be overloaded, i.e. you can define your
4310 own MAX aggregate to take in a user-defined type.)
4311 * CHANGE ACL removed. GRANT/REVOKE syntax added.
4312 - Privileges can be given to a group using the "GROUP" keyword.
4314 GRANT SELECT ON foobar TO GROUP my_group;
4315 The keyword 'PUBLIC' is also supported to mean all users.
4317 Privileges can only be granted or revoked to one user or group
4320 "WITH GRANT OPTION" is not supported. Only class owners can change
4322 - The default access control is to to grant users readonly access.
4323 You must explicitly grant insert/update access to users. To change
4324 this, modify the line in
4325 src/backend/utils/acl.h
4326 that defines ACL_WORLD_DEFAULT
4329 * the bug where aggregates of empty tables were not run has been fixed. Now,
4330 aggregates run on empty tables will return the initial conditions of the
4331 aggregates. Thus, COUNT of an empty table will now properly return 0.
4332 MAX/MIN of an empty table will return a tuple of value NULL.
4333 * allow the use of \; inside the monitor
4334 * the LISTEN/NOTIFY asynchronous notification mechanism now work
4335 * NOTIFY in rule action bodies now work
4336 * hash indexes work, and access methods in general should perform better.
4337 creation of large btree indexes should be much faster. (thanks to Paul
4340 Other changes and enhancements:
4341 * addition of an EXPLAIN statement used for explaining the query execution
4342 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
4344 * WARN and NOTICE messages no longer have timestamps on them. To turn on
4345 timestamps of error messages, uncomment the line in
4346 src/backend/utils/elog.h:
4347 /* define ELOG_TIMESTAMPS */
4348 * On an access control violation, the message
4349 "Either no such class or insufficient privilege"
4350 will be given. This is the same message that is returned when
4351 a class is not found. This dissuades non-privileged users from
4352 guessing the existence of privileged classes.
4353 * some additional system catalog changes have been made that are not
4354 visible to the user.
4357 * The -oid option has been added to the "pg_result" tcl command.
4358 pg_result -oid returns oid of the last tuple inserted. If the
4359 last command was not an INSERT, then pg_result -oid returns "".
4360 * the large object interface is available as pg_lo* tcl commands:
4361 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
4363 Portability enhancements and New Ports:
4364 * flex/lex problems have been cleared up. Now, you should be able to use
4365 flex instead of lex on any platforms. We no longer make assumptions of
4366 what lexer you use based on the platform you use.
4367 * The Linux-ELF port is now supported. Various configuration have been
4368 tested: The following configuration is known to work:
4369 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
4370 with everything in ELF format,
4373 * ipcclean added to the distribution
4374 ipcclean usually does not need to be run, but if your backend crashes
4375 and leaves shared memory segments hanging around, ipcclean will
4376 clean them up for you.
4379 * the user manual has been revised and libpq documentation added.
4380 _________________________________________________________________
4382 Postgres95 Release 0.02
4384 Release date: 1995-05-25
4385 _________________________________________________________________
4389 Incompatible changes:
4390 * The SQL statement for creating a database is 'CREATE DATABASE' instead
4391 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
4392 of 'DESTROYDB'. However, the names of the executables 'createdb' and
4393 'destroydb' remain the same.
4396 * pgperl - a Perl (4.036) interface to Postgres95
4397 * pg_dump - a utility for dumping out a postgres database into a
4398 script file containing query commands. The script files are in a ASCII
4399 format and can be used to reconstruct the database, even on other
4400 machines and other architectures. (Also good for converting
4401 a Postgres 4.2 database to Postgres95 database.)
4403 The following ports have been incorporated into postgres95-beta-0.02:
4404 * the NetBSD port by Alistair Crooks
4405 * the AIX port by Mike Tung
4406 * the Windows NT port by Jon Forrest (more stuff but not done yet)
4407 * the Linux ELF port by Brian Gallew
4409 The following bugs have been fixed in postgres95-beta-0.02:
4410 * new lines not escaped in COPY OUT and problem with COPY OUT when first
4412 * cannot type return to use the default user id in createuser
4413 * SELECT DISTINCT on big tables crashes
4414 * Linux installation problems
4415 * monitor doesn't allow use of 'localhost' as PGHOST
4416 * psql core dumps when doing \c or \l
4417 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
4418 * libpgtcl has a hard-wired default port number
4419 * SELECT DISTINCT INTO TABLE hangs
4420 * CREATE TYPE doesn't accept 'variable' as the internallength
4421 * wrong result using more than 1 aggregate in a SELECT
4422 _________________________________________________________________
4424 Postgres95 Release 0.01
4426 Release date: 1995-05-01