6 Release date: 2003-12-22
8 This release contains a variety of fixes from 7.4.
9 _________________________________________________________________
11 Migration to version 7.4.1
13 A dump/restore is *not* required for those running 7.4.
15 If you want to install the fixes in the information schema you need to
16 reload it into the database. This is either accomplished by
17 initializing a new cluster by running "initdb", or by running the
18 following sequence of SQL commands in each database (ideally including
19 template1) as a superuser in psql, after installing the new release:
20 DROP SCHEMA information_schema CASCADE;
21 \i /usr/local/pgsql/share/information_schema.sql
23 Substitute your installation path in the second command.
24 _________________________________________________________________
28 * Fixed bug in "CREATE SCHEMA" parsing in ECPG (Michael)
29 * Fix compile error when "--enable-thread-safety" and "--with-perl"
30 are used together (Peter)
31 * Fix for subqueries that used hash joins (Tom)
32 Certain subqueries that used hash joins would crash because of
33 improperly shared structures.
34 * Fix free space map compaction bug (Tom)
35 This fixes a bug where compaction of the free space map could lead
36 to a database server shutdown.
37 * Fix for Borland compiler build of libpq (Bruce)
38 * Fix netmask() and hostmask() to return the maximum-length masklen
40 Fix these functions to return values consistent with pre-7.4
42 * Several "contrib/pg_autovacuum" fixes
43 Fixes include improper variable initialization, missing vacuum
44 after "TRUNCATE", and duration computation overflow for long
46 * Allow compile of "contrib/cube" under Cygwin (Jason Tishler)
47 * Fix Solaris use of password file when no passwords are defined
49 Fix crash on Solaris caused by use of any type of password
50 authentication when no passwords were defined.
51 * JDBC fix for thread problems, other fixes
52 * Fix for bytea index lookups (Joe)
53 * Fix information schema for bit data types (Peter)
54 * Force zero_damaged_pages to be on during recovery from WAL
55 * Prevent some obscure cases of "variable not in subplan target
57 * Make PQescapeBytea and byteaout consistent with each other (Joe)
58 * Escape bytea output for bytes > 0x7e(Joe)
59 If different client encodings are used for bytea output and input,
60 it is possible for bytea values to be corrupted by the differing
61 encodings. This fix escapes all bytes that might be affected.
62 * Added missing SPI_finish() calls to dblink's
63 get_tuple_of_interest() (Joe)
65 * Fix information schema view constraint_column_usage for foreign
67 * ECPG fixes (Michael)
68 * Fix bug with multiple IN subqueries and joins in the subqueries
70 * Alllow COUNT('x') to work (Tom)
71 * Install ECPG include files for Informix compatibility into
72 separate directory (Peter)
73 Some names of ECPG include files for Informix compatibility
74 conflicted with operating system include files. By installing them
75 in their own directory, name conflicts have been reduced.
76 * Fix SSL memory leak (Neil)
77 This release fixes a bug in 7.4 where SSL didn't free all memory
79 * Prevent "pg_service.conf" from using service name as default
81 * Fix local ident authentication on FreeBSD (Tom)
82 _________________________________________________________________
86 Release date: 2003-11-17
87 _________________________________________________________________
91 Major changes in this release:
93 IN / NOT IN subqueries are now much more efficient
94 In previous releases, IN/NOT IN subqueries were joined to the
95 upper query by sequentially scanning the subquery looking for a
96 match. The 7.4 code uses the same sophisticated techniques used
97 by ordinary joins and so is much faster. An IN will now usually
98 be as fast as or faster than an equivalent EXISTS subquery;
99 this reverses the conventional wisdom that applied to previous
102 Improved GROUP BY processing by using hash buckets
103 In previous releases, rows to be grouped had to be sorted
104 first. The 7.4 code can do GROUP BY without sorting, by
105 accumulating results into a hash table with one entry per
106 group. It will still use the sort technique, however, if the
107 hash table is estimated to be too large to fit in sort_mem.
109 New multikey hash join capability
110 In previous releases, hash joins could only occur on single
111 keys. This release allows multicolumn hash joins.
113 Queries using the explicit JOIN syntax are now better optimized
114 Prior releases evaluated queries using the explicit JOIN syntax
115 only in the order implied by the syntax. 7.4 allows full
116 optimization of these queries, meaning the optimizer considers
117 all possible join orderings and chooses the most efficient.
118 Outer joins, however, must still follow the declared ordering.
120 Faster and more powerful regular expression code
121 The entire regular expression module has been replaced with a
122 new version by Henry Spencer, originally written for Tcl. The
123 code greatly improves performance and supports several flavors
124 of regular expressions.
126 Function-inlining for simple SQL functions
127 Simple SQL functions can now be inlined by including their SQL
128 in the main query. This improves performance by eliminating
129 per-call overhead. That means simple SQL functions now behave
132 Full support for IPv6 connections and IPv6 address data types
133 Previous releases allowed only IPv4 connections, and the IP
134 data types only supported IPv4 addresses. This release adds
135 full IPv6 support in both of these areas.
137 Major improvements in SSL performance and reliability
138 Several people very familiar with the SSL API have overhauled
139 our SSL code to improve SSL key negotiation and error recovery.
141 Make free space map efficiently reuse empty index pages, and other
142 free space management improvements
143 In previous releases, B-tree index pages that were left empty
144 because of deleted rows could only be reused by rows with index
145 values similar to the rows originally indexed on that page. In
146 7.4, "VACUUM" records empty index pages and allows them to be
147 reused for any future index rows.
149 SQL-standard information schema
150 The information schema provides a standardized and stable way
151 to access information about the schema objects defined in a
154 Cursors conform more closely to the SQL standard
155 The commands "FETCH" and "MOVE" have been overhauled to conform
156 more closely to the SQL standard.
158 Cursors can exist outside transactions
159 These cursors are also called holdable cursors.
161 New client-to-server protocol
162 The new protocol adds error codes, more status information,
163 faster startup, better support for binary data transmission,
164 parameter values separated from SQL commands, prepared
165 statements available at the protocol level, and cleaner
166 recovery from "COPY" failures. The older protocol is still
167 supported by both server and clients.
169 libpq and ECPG applications are now fully thread-safe
170 While previous libpq releases already supported threads, this
171 release improves thread safety by fixing some non-thread-safe
172 code that was used during database connection startup. The
173 "configure" option "--enable-thread-safety" must be used to
176 New version of full-text indexing
177 A new full-text indexing suite is available in
181 The new autovacuum tool in "contrib/autovacuum" monitors the
182 database statistics tables for "INSERT"/"UPDATE"/"DELETE"
183 activity and automatically vacuums tables when needed.
185 Array handling has been improved and moved into the server core
186 Many array limitations have been removed, and arrays behave
187 more like fully-supported data types.
188 _________________________________________________________________
190 Migration to version 7.4
192 A dump/restore using pg_dump is required for those wishing to migrate
193 data from any previous release.
195 Observe the following incompatibilities:
197 * The server-side autocommit setting was removed and reimplemented
198 in client applications and languages. Server-side autocommit was
199 causing too many problems with languages and applications that
200 wanted to control their own autocommit behavior, so autocommit was
201 removed from the server and added to individual client APIs as
203 * Error message wording has changed substantially in this release.
204 Significant effort was invested to make the messages more
205 consistent and user-oriented. If your applications try to detect
206 different error conditions by parsing the error message, you are
207 strongly encouraged to use the new error code facility instead.
208 * Inner joins using the explicit JOIN syntax may behave differently
209 because they are now better optimized.
210 * A number of server configuration parameters have been renamed for
211 clarity, primarily those related to logging.
212 * FETCH 0 or MOVE 0 now does nothing. In prior releases, FETCH 0
213 would fetch all remaining rows, and MOVE 0 would move to the end
215 * "FETCH" and "MOVE" now return the actual number of rows
216 fetched/moved, or zero if at the beginning/end of the cursor.
217 Prior releases would return the row count passed to the command,
218 not the number of rows actually fetched or moved.
219 * "COPY" now can process files that use carriage-return or
220 carriage-return/line-feed end-of-line sequences. Literal
221 carriage-returns and line-feeds are no longer accepted in data
222 values; use \r and \n instead.
223 * Trailing spaces are now trimmed when converting from type char(n)
224 to varchar(n) or text. This is what most people always expected to
226 * The data type float(p) now measures "p" in binary digits, not
227 decimal digits. The new behavior follows the SQL standard.
228 * Ambiguous date values now must match the ordering specified by the
229 datestyle setting. In prior releases, a date specification of
230 10/20/03 was interpreted as a date in October even if datestyle
231 specified that the day should be first. 7.4 will throw an error if
232 a date specification is invalid for the current setting of
234 * The functions oidrand, oidsrand, and userfntest have been removed.
235 These functions were determined to be no longer useful.
236 * String literals specifying time-varying date/time values, such as
237 'now' or 'today' will no longer work as expected in column default
238 expressions; they now cause the time of the table creation to be
239 the default, not the time of the insertion. Functions such as
240 now(), current_timestamp, or current_date should be used instead.
241 In previous releases, there was special code so that strings such
242 as 'now' were interpreted at "INSERT" time and not at table
243 creation time, but this work around didn't cover all cases.
244 Release 7.4 now requires that defaults be defined properly using
245 functions such as now() or current_timestamp. These will work in
247 * The dollar sign ($) is no longer allowed in operator names. It can
248 instead be a non-first character in identifiers. This was done to
249 improve compatibility with other database systems, and to avoid
250 syntax problems when parameter placeholders ($n) are written
251 adjacent to operators.
252 _________________________________________________________________
256 Below you will find a detailed account of the changes between release
257 7.4 and the previous major release.
258 _________________________________________________________________
260 Server Operation Changes
262 * Allow IPv6 server connections (Nigel Kukard, Johan Jordaan, Bruce,
263 Tom, Kurt Roeckx, Andrew Dunstan)
264 * Fix SSL to handle errors cleanly (Nathan Mueller)
265 In prior releases, certain SSL API error reports were not handled
266 correctly. This release fixes those problems.
267 * SSL protocol security and performance improvements (Sean
269 SSL key renegotiation was happening too frequently, causing poor
270 SSL performance. Also, initial key handling was improved.
271 * Print lock information when a deadlock is detected (Tom)
272 This allows easier debugging of deadlock situations.
273 * Update "/tmp" socket modification times regularly to avoid their
275 This should help prevent "/tmp" directory cleaner administration
276 scripts from removing server socket files.
277 * Enable PAM for Mac OS X (Aaron Hillegass)
278 * Make B-tree indexes fully WAL-safe (Tom)
279 In prior releases, under certain rare cases, a server crash could
280 cause B-tree indexes to become corrupt. This release removes those
282 * Allow B-tree index compaction and empty page reuse (Tom)
283 * Fix inconsistent index lookups during split of first root page
285 In prior releases, when a single-page index split into two pages,
286 there was a brief period when another database session could miss
287 seeing an index entry. This release fixes that rare failure case.
288 * Improve free space map allocation logic (Tom)
289 * Preserve free space information between server restarts (Tom)
290 In prior releases, the free space map was not saved when the
291 postmaster was stopped, so newly started servers had no free space
292 information. This release saves the free space map, and reloads it
293 when the server is restarted.
294 * Add start time to pg_stat_activity (Neil)
295 * New code to detect corrupt disk pages; erase with
296 zero_damaged_pages (Tom)
297 * New client/server protocol: faster, no username length limit,
298 allow clean exit from "COPY" (Tom)
299 * Add transaction status, table ID, column ID to client/server
301 * Add binary I/O to client/server protocol (Tom)
302 * Remove autocommit server setting; move to client applications
304 * New error message wording, error codes, and three levels of error
305 detail (Tom, Joe, Peter)
306 _________________________________________________________________
308 Performance Improvements
310 * Add hashing for GROUP BY aggregates (Tom)
311 * Make nested-loop joins be smarter about multicolumn indexes (Tom)
312 * Allow multikey hash joins (Tom)
313 * Improve constant folding (Tom)
314 * Add ability to inline simple SQL functions (Tom)
315 * Reduce memory usage for queries using complex functions (Tom)
316 In prior releases, functions returning allocated memory would not
317 free it until the query completed. This release allows the freeing
318 of function-allocated memory when the function call completes,
319 reducing the total memory used by functions.
320 * Improve GEQO optimizer performance (Tom)
321 This release fixes several inefficiencies in the way the GEQO
322 optimizer manages potential query paths.
323 * Allow IN/NOT IN to be handled via hash tables (Tom)
324 * Improve NOT IN (subquery) performance (Tom)
325 * Allow most IN subqueries to be processed as joins (Tom)
326 * Pattern matching operations can use indexes regardless of locale
328 There is no way for non-ASCII locales to use the standard indexes
329 for LIKE comparisons. This release adds a way to create a special
331 * Allow the postmaster to preload libraries using preload_libraries
333 For shared libraries that require a long time to load, this option
334 is available so the library can be preloaded in the postmaster and
335 inherited by all database sessions.
336 * Improve optimizer cost computations, particularly for subqueries
338 * Avoid sort when subquery ORDER BY matches upper query (Tom)
339 * Deduce that WHERE a.x = b.y AND b.y = 42 also means a.x = 42 (Tom)
340 * Allow hash/merge joins on complex joins (Tom)
341 * Allow hash joins for more data types (Tom)
342 * Allow join optimization of explicit inner joins, disable with
343 join_collapse_limit (Tom)
344 * Add parameter from_collapse_limit to control conversion of
345 subqueries to joins (Tom)
346 * Use faster and more powerful regular expression code from Tcl
348 * Use bit-mapped relation sets in the optimizer (Tom)
349 * Improve connection startup time (Tom)
350 The new client/server protocol requires fewer network packets to
351 start a database session.
352 * Improve trigger/constraint performance (Stephan)
353 * Improve speed of col IN (const, const, const, ...) (Tom)
354 * Fix hash indexes which were broken in rare cases (Tom)
355 * Improve hash index concurrency and speed (Tom)
356 Prior releases suffered from poor hash index performance,
357 particularly for high concurrency situations. This release fixes
358 that, and the development group is interested in reports comparing
359 B-tree and hash index performance.
360 * Align shared buffers on 32-byte boundary for copy speed
361 improvement (Manfred Spraul)
362 Certain CPU's perform faster data copies when addresses are
364 * Data type numeric reimplemented for better performance (Tom)
365 numeric used to be stored in base 100. The new code uses base
366 10000, for significantly better performance.
367 _________________________________________________________________
369 Server Configuration Changes
371 * Rename server parameter server_min_messages to log_min_messages
373 This was done so most parameters that control the server logs
375 * Rename show_*_stats to log_*_stats (Bruce)
376 * Rename show_source_port to log_source_port (Bruce)
377 * Rename hostname_lookup to log_hostname (Bruce)
378 * Add checkpoint_warning to warn of excessive checkpointing (Bruce)
379 In prior releases, it was difficult to determine if checkpoint was
380 happening too frequently. This feature adds a warning to the
381 server logs when excessive checkpointing happens.
382 * New read-only server parameters for localization (Tom)
383 * Change debug server log messages to output as DEBUG rather than
385 * Prevent server log variables from being turned off by
386 non-superusers (Bruce)
387 This is a security feature so non-superusers cannot disable
388 logging that was enabled by the administrator.
389 * log_min_messages/client_min_messages now controls debug_* output
391 This centralizes client debug information so all debug output can
392 be sent to either the client or server logs.
393 * Add Mac OS X Rendezvous server support (Chris Campbell)
394 This allows Mac OS X hosts to query the network for available
396 * Add ability to print only slow statements using
397 log_min_duration_statement (Christopher)
398 This is an often requested debugging feature that allows
399 administrators to see only slow queries in their server logs.
400 * Allow "pg_hba.conf" to accept netmasks in CIDR format (Andrew
402 This allows administrators to merge the host IP address and
403 netmask fields into a single CIDR field in "pg_hba.conf".
404 * New read-only parameter is_superuser (Tom)
405 * New parameter log_error_verbosity to control error detail (Tom)
406 This works with the new error reporting feature to supply
407 additional error information like hints, file names and line
409 * postgres --describe-config now dumps server config variables
411 This option is useful for administration tools that need to know
412 the configuration variable names and their minimums, maximums,
413 defaults, and descriptions.
414 * Add new columns in pg_settings: context, type, source, min_val,
416 * Make default shared_buffers 1000 and max_connections 100, if
418 Prior versions defaulted to 64 shared buffers so PostgreSQL would
419 start on even very old systems. This release tests the amount of
420 shared memory allowed by the platform and selects more reasonable
421 default values if possible. Of course, users are still encouraged
422 to evaluate their resource load and size shared_buffers
424 * New "pg_hba.conf" record type hostnossl to prevent SSL connections
426 In prior releases, there was no way to prevent SSL connections if
427 both the client and server supported SSL. This option allows that
429 * Remove parameter geqo_random_seed (Tom)
430 * Add server parameter regex_flavor to control regular expression
432 * Make "pg_ctl" better handle nonstandard ports (Greg)
433 _________________________________________________________________
437 * New SQL-standard information schema (Peter)
438 * Add read-only transactions (Peter)
439 * Print key name and value in foreign-key violation messages (Dmitry
441 * Allow users to see their own queries in pg_stat_activity (Kevin
443 In prior releases, only the superuser could see query strings
444 using pg_stat_activity. Now ordinary users can see their own query
446 * Fix aggregates in subqueries to match SQL standard (Tom)
447 The SQL standard says that an aggregate function appearing within
448 a nested subquery belongs to the outer query if its argument
449 contains only outer-query variables. Prior PostgreSQL releases did
450 not handle this fine point correctly.
451 * Add option to prevent auto-addition of tables referenced in query
453 By default, tables mentioned in the query are automatically added
454 to the FROM clause if they are not already there. This is
455 compatible with historic POSTGRES behavior but is contrary to the
456 SQL standard. This option allows selecting standard-compatible
458 * Allow UPDATE ... SET col = DEFAULT (Rod)
459 This allows "UPDATE" to set a column to its declared default
461 * Allow expressions to be used in LIMIT/OFFSET (Tom)
462 In prior releases, LIMIT/OFFSET could only use constants, not
464 * Implement CREATE TABLE AS EXECUTE (Neil, Peter)
465 _________________________________________________________________
467 Object Manipulation Changes
469 * Make "CREATE SEQUENCE" grammar more conforming to SQL 2003 (Neil)
470 * Add statement-level triggers (Neil)
471 While this allows a trigger to fire at the end of a statement, it
472 does not allow the trigger to access all rows modified by the
473 statement. This capability is planned for a future release.
474 * Add check constraints for domains (Rod)
475 This greatly increases the usefulness of domains by allowing them
476 to use check constraints.
477 * Add "ALTER DOMAIN" (Rod)
478 This allows manipulation of existing domains.
479 * Fix several zero-column table bugs (Tom)
480 PostgreSQL supports zero-column tables. This fixes various bugs
481 that occur when using such tables.
482 * Have ALTER TABLE ... ADD PRIMARY KEY add not-null constraint (Rod)
483 In prior releases, ALTER TABLE ... ADD PRIMARY would add a unique
484 index, but not a not-null constraint. That is fixed in this
486 * Add ALTER TABLE ... WITHOUT OIDS (Rod)
487 This allows control over whether new and updated rows will have an
488 OID column. This is most useful for saving storage space.
489 * Add ALTER SEQUENCE to modify minimum, maximum, increment, cache,
491 * Add ALTER TABLE ... CLUSTER ON (Alvaro Herrera)
492 This command is used by "pg_dump" to record the cluster column for
493 each table previously clustered. This information is used by
494 database-wide cluster to cluster all previously clustered tables.
495 * Improve automatic type casting for domains (Rod, Tom)
496 * Allow dollar signs in identifiers, except as first character (Tom)
497 * Disallow dollar signs in operator names, so x=$1 works (Tom)
498 * Allow copying table schema using LIKE subtable, also SQL 2003
499 feature INCLUDING DEFAULTS (Rod)
500 * Add WITH GRANT OPTION clause to "GRANT" (Peter)
501 This enabled "GRANT" to give other users the ability to grant
502 privileges on a object.
503 _________________________________________________________________
505 Utility Command Changes
507 * Add ON COMMIT clause to "CREATE TABLE" for temporary tables
509 This adds the ability for a table to be dropped or all rows
510 deleted on transaction commit.
511 * Allow cursors outside transactions using WITH HOLD (Neil)
512 In previous releases, cursors were removed at the end of the
513 transaction that created them. Cursors can now be created with the
514 WITH HOLD option, which allows them to continue to be accessed
515 after the creating transaction has committed.
516 * FETCH 0 and MOVE 0 now do nothing (Bruce)
517 In previous releases, FETCH 0 fetched all remaining rows, and MOVE
518 0 moved to the end of the cursor.
519 * Cause "FETCH" and "MOVE" to return the number of rows
520 fetched/moved, or zero if at the beginning/end of cursor, per SQL
522 In prior releases, the row count returned by "FETCH" and "MOVE"
523 did not accurately reflect the number of rows processed.
524 * Properly handle SCROLL with cursors, or report an error (Neil)
525 Allowing random access (both forward and backward scrolling) to
526 some kinds of queries cannot be done without some additional work.
527 If SCROLL is specified when the cursor is created, this additional
528 work will be performed. Furthermore, if the cursor has been
529 created with NO SCROLL, no random access is allowed.
530 * Implement SQL-compatible options FIRST, LAST, ABSOLUTE n, RELATIVE
531 n for "FETCH" and "MOVE" (Tom)
532 * Allow "EXPLAIN" on "DECLARE CURSOR" (Tom)
533 * Allow "CLUSTER" to use index marked as pre-clustered by default
535 * Allow "CLUSTER" to cluster all tables (Alvaro Herrera)
536 This allows all previously clustered tables in a database to be
537 reclustered with a single command.
538 * Prevent "CLUSTER" on partial indexes (Tom)
539 * Allow DOS and Mac line-endings in "COPY" files (Bruce)
540 * Disallow literal carriage return as a data value,
541 backslash-carriage-return and \r are still allowed (Bruce)
542 * "COPY" changes (binary, \.) (Tom)
543 * Recover from "COPY" failure cleanly (Tom)
544 * Prevent possible memory leaks in "COPY" (Tom)
545 * Make "TRUNCATE" transaction-safe (Rod)
546 "TRUNCATE" can now be used inside a transaction. If the
547 transaction aborts, the changes made by the "TRUNCATE" are
548 automatically rolled back.
549 * Allow prepare/bind of utility commands like "FETCH" and "EXPLAIN"
551 * Add "EXPLAIN EXECUTE" (Neil)
552 * Improve "VACUUM" performance on indexes by reducing WAL traffic
554 * Functional indexes have been generalized into indexes on
556 In prior releases, functional indexes only supported a simple
557 function applied to one or more column names. This release allows
558 any type of scalar expression.
559 * Have "SHOW TRANSACTION ISOLATION" match input to "SET TRANSACTION
561 * Have "COMMENT ON DATABASE" on nonlocal database generate a warning
563 Database comments are stored in database-local tables so comments
564 on a database have to be stored in each database.
565 * Improve reliability of "LISTEN"/"NOTIFY" (Tom)
566 * Allow "REINDEX" to reliably reindex nonshared system catalog
568 This allows system tables to be reindexed without the requirement
569 of a standalone session, which was necessary in previous releases.
570 The only tables that now require a standalone session for
571 reindexing are the global system tables pg_database, pg_shadow,
573 _________________________________________________________________
575 Data Type and Function Changes
577 * New server parameter extra_float_digits to control precision
578 display of floating-point numbers (Pedro Ferreira, Tom)
579 This controls output precision which was causing regression
581 * Allow +1300 as a numeric time-zone specifier, for FJST (Tom)
582 * Remove rarely used functions oidrand, oidsrand, and userfntest
584 * Add md5() function to main server, already in "contrib/pgcrypto"
586 An MD5 function was frequently requested. For more complex
587 encryption capabilities, use "contrib/pgcrypto".
588 * Increase date range of timestamp (John Cochran)
589 * Change EXTRACT(EPOCH FROM timestamp) so timestamp without time
590 zone is assumed to be in local time, not GMT (Tom)
591 * Trap division by zero in case the operating system doesn't prevent
593 * Change the numeric data type internally to base 10000 (Tom)
594 * New hostmask() function (Greg Wickham)
595 * Fixes for to_char() and to_timestamp() (Karel)
596 * Allow functions that can take any argument data type and return
597 any data type, using anyelement and anyarray (Joe)
598 This allows the creation of functions that can work with any data
600 * Arrays may now be specified as ARRAY[1,2,3],
601 ARRAY[['a','b'],['c','d']], or ARRAY[ARRAY[ARRAY[2]]] (Joe)
602 * Allow proper comparisons for arrays, including ORDER BY and
603 DISTINCT support (Joe)
604 * Allow indexes on array columns (Joe)
605 * Allow array concatenation with || (Joe)
606 * Allow WHERE qualification expr op ANY/SOME/ALL (array_expr) (Joe)
607 This allows arrays to behave like a list of values, for purposes
608 like SELECT * FROM tab WHERE col IN (array_val).
609 * New array functions array_append, array_cat, array_lower,
610 array_prepend, array_to_string, array_upper, string_to_array (Joe)
611 * Allow user defined aggregates to use polymorphic functions (Joe)
612 * Allow assignments to empty arrays (Joe)
613 * Allow 60 in seconds fields of time, timestamp, and interval input
615 Sixty-second values are needed for leap seconds.
616 * Allow cidr data type to be cast to text (Tom)
617 * Disallow invalid time zone names in SET TIMEZONE
618 * Trim trailing spaces when char is cast to varchar or text (Tom)
619 * Make float(p) measure the precision "p" in binary digits, not
621 * Add IPv6 support to the inet and cidr data types (Michael Graff)
622 * Add family() function to report whether address is IPv4 or IPv6
624 * Have SHOW datestyle generate output similar to that used by SET
626 * Make EXTRACT(TIMEZONE) and SET/SHOW TIME ZONE follow the SQL
627 convention for the sign of time zone offsets, i.e., positive is
629 * Fix date_trunc('quarter', ...) (Böjthe Zoltán)
630 Prior releases returned an incorrect value for this function call.
631 * Make initcap() more compatible with Oracle (Mike Nolan)
632 initcap() now uppercases a letter appearing after any
633 non-alphanumeric character, rather than only after whitespace.
634 * Allow only datestyle field order for date values not in ISO-8601
636 * Add new datestyle values MDY, DMY, and YMD to set input field
637 order; honor US and European for backward compatibility (Tom)
638 * String literals like 'now' or 'today' will no longer work as a
639 column default. Use functions such as now(), current_timestamp
640 instead. (change required for prepared statements) (Tom)
641 * Treat NaN as larger than any other value in min()/max() (Tom)
642 NaN was already sorted after ordinary numeric values for most
643 purposes, but min() and max() didn't get this right.
644 * Prevent interval from suppressing :00 seconds display
645 * New function pg_get_triggerdef(prettyprint) and
646 pg_constraint_is_visible()
647 * Allow time to be specified as 040506 or 0405 (Tom)
648 * Input date order must now be YYYY-MM-DD (with 4-digit year) or
650 * Make pg_get_constraintdef to support unique, primary-key, and
651 check constraints (Christopher)
652 _________________________________________________________________
654 Server-Side Language Changes
656 * Prevent PL/pgSQL crash when RETURN NEXT is used on a zero-row
657 record variable (Tom)
658 * Make PL/Python's spi_execute interface handle null values properly
660 * Allow PL/pgSQL to declare variables of composite types without
662 * Fix PL/Python's _quote() function to handle big integers
663 * Make PL/Python an untrusted language, now called plpythonu (Kevin
665 The Python language no longer supports a restricted execution
666 environment, so the trusted version of PL/Python was removed. If
667 this situation changes, a version of PL/python that can be used by
668 non-superusers will be readded.
669 * Allow polymorphic PL/pgSQL functions (Joe, Tom)
670 * Allow polymorphic SQL functions (Joe)
671 * Improved compiled function caching mechanism in PL/pgSQL with full
672 support for polymorphism (Joe)
673 * Add new parameter $0 in PL/pgSQL representing the function's
674 actual return type (Joe)
675 * Allow PL/Tcl and PL/Python to use the same trigger on multiple
677 * Fixed PL/Tcl's spi_prepare to accept fully qualified type names in
678 the parameter type list (Jan)
679 _________________________________________________________________
683 * Add \pset pager always to always use pager (Greg)
684 This forces the pager to be used even if the number of rows is
685 less than the screen height. This is valuable for rows that wrap
686 across several screen rows.
687 * Improve tab completion (Rod, Ross Reedstrom, Ian Barwick)
688 * Reorder \? help into groupings (Harald Armin Massa, Bruce)
689 * Add backslash commands for listing schemas, casts, and conversions
691 * "\encoding" now changes based on the server parameter
692 client_encoding server (Tom)
693 In previous versions, "\encoding" was not aware of encoding
694 changes made using SET client_encoding.
695 * Save editor buffer into readline history (Ross)
696 When "\e" is used to edit a query, the result is saved in the
697 readline history for retrieval using the up arrow.
698 * Improve "\d" display (Christopher)
699 * Enhance HTML mode to be more standards-conforming (Greg)
700 * New "\set AUTOCOMMIT off" capability (Tom)
701 This takes the place of the removed server parameter autocommit.
702 * New "\set VERBOSITY" to control error detail (Tom)
703 This controls the new error reporting details.
704 * New prompt escape sequence %x to show transaction status (Tom)
705 * Long options for psql are now available on all platforms
706 _________________________________________________________________
710 * Multiple pg_dump fixes, including tar format and large objects
711 * Allow pg_dump to dump specific schemas (Neil)
712 * Make pg_dump preserve column storage characteristics (Christopher)
713 This preserves ALTER TABLE ... SET STORAGE information.
714 * Make pg_dump preserve "CLUSTER" characteristics (Christopher)
715 * Have pg_dumpall use "GRANT"/"REVOKE" to dump database-level
717 * Allow pg_dumpall to support the options "-a", "-s", "-x" of
719 * Prevent pg_dump from lowercasing identifiers specified on the
721 * pg_dump options "--use-set-session-authorization" and
722 "--no-reconnect" now do nothing, all dumps use "SET SESSION
724 pg_dump no longer reconnects to switch users, but instead always
725 uses "SET SESSION AUTHORIZATION". This will reduce password
726 prompting during restores.
727 * Long options for pg_dump are now available on all platforms
728 PostgreSQL now includes its own long-option processing routines.
729 _________________________________________________________________
733 * Add function PQfreemem for freeing memory on Windows, suggested
735 Windows requires that memory allocated in a library be freed by a
736 function in the same library, hence free() doesn't work for
737 freeing memory allocated by libpq. PQfreemem is the proper way to
738 free libpq memory, especially on Windows, and is recommended for
739 other platforms as well.
740 * Document service capability, and add sample file (Bruce)
741 This allows clients to look up connection information in a central
742 file on the client machine.
743 * Make PQsetdbLogin have the same defaults as PQconnectdb (Tom)
744 * Allow libpq to cleanly fail when result sets are too large (Tom)
745 * Improve performance of function PGunescapeBytea (Ben Lamb)
746 * Allow thread-safe libpq with "configure" option
747 "--enable-thread-safety" (Lee Kindness, Philip Yarra)
748 * Allow function pqInternalNotice to accept a format string and
749 arguments instead of just a preformatted message (Tom, Sean
751 * Control SSL negotiation with sslmode values disable, allow,
752 prefer, and require (Jon Jensen)
753 * Allow new error codes and levels of text (Tom)
754 * Allow access to the underlying table and column of a query result
756 This is helpful for query-builder applications that want to know
757 the underlying table and column names associated with a specific
759 * Allow access to the current transaction status (Tom)
760 * Add ability to pass binary data directly to the server (Tom)
761 * Add function PQexecPrepared and PQsendQueryPrepared functions
762 which perform bind/execute of previously prepared statements (Tom)
763 _________________________________________________________________
767 * Allow setNull on updateable result sets
768 * Allow executeBatch on a prepared statement (Barry)
769 * Support SSL connections (Barry)
770 * Handle schema names in result sets (Paul Sorenson)
771 * Add refcursor support (Nic Ferrier)
772 _________________________________________________________________
774 Miscellaneous Interface Changes
776 * Prevent possible memory leak or core dump during libpgtcl shutdown
778 * Add Informix compatibility to ECPG (Michael)
779 This allows ECPG to process embedded C programs that were written
780 using certain Informix extensions.
781 * Add type decimal to ECPG that is fixed length, for Informix
783 * Allow thread-safe embedded SQL programs with "configure" option
784 "--enable-thread-safety" (Lee Kindness, Bruce)
785 This allows multiple threads to access the database at the same
787 * Moved Python client PyGreSQL to http://www.pygresql.org (Marc)
788 _________________________________________________________________
792 * Prevent need for separate platform geometry regression result
794 * Improved PPC locking primitive (Reinhard Max)
795 * New function palloc0 to allocate and clear memory (Bruce)
796 * Fix locking code for s390x CPU (64-bit) (Tom)
797 * Allow OpenBSD to use local ident credentials (William Ahern)
798 * Make query plan trees read-only to executor (Tom)
799 * Add Darwin startup scripts (David Wheeler)
800 * Allow libpq to compile with Borland C++ compiler (Lester Godwin,
802 * Use our own version of getopt_long() if needed (Peter)
803 * Convert administration scripts to C (Peter)
804 * Bison >= 1.85 is now required to build the PostgreSQL grammar, if
806 * Merge documentation into one book (Peter)
807 * Add Windows compatibility functions (Bruce)
808 * Allow client interfaces to compile under MinGW (Bruce)
809 * New ereport() function for error reporting (Tom)
810 * Support Intel compiler on Linux (Peter)
811 * Improve Linux startup scripts (Slawomir Sudnik, Darko Prenosil)
812 * Add support for AMD Opteron and Itanium (Jeffrey W. Baker, Bruce)
813 * Remove "--enable-recode" option from "configure"
814 This was no longer needed now that we have "CREATE CONVERSION".
815 * Generate a compile error if spinlock code is not found (Bruce)
816 Platforms without spinlock code will now fail to compile, rather
817 than silently using semaphores. This failure can be disabled with
818 a new "configure" option.
819 _________________________________________________________________
823 * Change dbmirror license to BSD
824 * Improve earthdistance (Bruno Wolff III)
825 * Portability improvements to pgcrypto (Marko Kreen)
826 * Prevent crash in xml (John Gray, Michael Richards)
829 * Update cube (Bruno Wolff III)
830 * Update earthdistance to use cube (Bruno Wolff III)
831 * Update btree_gist (Oleg)
832 * New tsearch2 full-text search module (Oleg, Teodor)
833 * Add hash-based crosstab function to tablefuncs (Joe)
834 * Add serial column to order connectby() siblings in tablefuncs
836 * Add named persistent connections to dblink (Shridhar Daithanka)
837 * New pg_autovacuum allows automatic "VACUUM" (Matthew T. O'Connor)
838 * Make pgbench honor environment variables PGHOST, PGPORT, PGUSER
840 * Improve intarray (Teodor Sigaev)
841 * Improve pgstattuple (Rod)
842 * Fix bug in metaphone() in fuzzystrmatch
843 * Improve adddepend (Rod)
844 * Update spi/timetravel (Böjthe Zoltán)
845 * Fix dbase "-s" option and improve non-ASCII handling (Thomas Behr,
847 * Remove array module because features now included by default (Joe)
848 _________________________________________________________________
852 Release date: 2003-12-03
854 This has a variety of fixes from 7.3.4.
855 _________________________________________________________________
857 Migration to version 7.3.5
859 A dump/restore is *not* required for those running 7.3.*.
860 _________________________________________________________________
864 * Force zero_damaged_pages to be on during recovery from WAL
865 * Prevent some obscure cases of "variable not in subplan target
867 * Force stats processes to detach from shared memory, ensuring
869 * Make PQescapeBytea and byteaout consistent with each other (Joe)
870 * Added missing SPI_finish() calls to dblink's
871 get_tuple_of_interest() (Joe)
872 * Fix for possible foreign key violation when rule rewrites INSERT
874 * Support qualified type names in PL/Tcl's spi_prepare command (Jan)
875 * Make pg_dump handle a procedural language handler located in
877 * Make pg_dump handle cases where a custom opclass is in another
879 * Make pg_dump dump binary-compatible casts correctly (Jan)
880 * Fix insertion of expressions containing subqueries into rule
882 * Fix incorrect argument processing in clusterdb script (Anand
884 * Fix problems with dropped columns in plpython triggers
885 * Repair problems with to_char() reading past end of its input
887 * Fix GB18030 mapping errors (Tatsuo)
888 * Fix several problems with SSL error handling and asynchronous SSL
890 * Remove ability to bind a list of values to a single parameter in
891 JDBC (prevents possible SQL-injection attacks)
892 * Fix some errors in HAVE_INT64_TIMESTAMP code paths
893 * Fix corner case for btree search in parallel with first root page
895 _________________________________________________________________
899 Release date: 2003-07-24
901 This has a variety of fixes from 7.3.3.
902 _________________________________________________________________
904 Migration to version 7.3.4
906 A dump/restore is *not* required for those running 7.3.*.
907 _________________________________________________________________
911 * Repair breakage in timestamp-to-date conversion for dates before
913 * Prevent rare possibility of server startup failure (Tom)
914 * Fix bugs in interval-to-time conversion (Tom)
915 * Add constraint names in a few places in pg_dump (Rod)
916 * Improve performance of functions with many parameters (Tom)
917 * Fix to_ascii() buffer overruns (Tom)
918 * Prevent restore of database comments from throwing an error (Tom)
919 * Work around buggy strxfrm() present in some Solaris releases (Tom)
920 * Properly escape jdbc setObject() strings to improve security
922 _________________________________________________________________
926 Release date: 2003-05-22
928 This release contains of variety of fixes for version 7.3.2.
929 _________________________________________________________________
931 Migration to version 7.3.3
933 A dump/restore is *not* required for those running version 7.3.*.
934 _________________________________________________________________
938 * Repair sometimes-incorrect computation of StartUpID after a crash
939 * Avoid slowness with lots of deferred triggers in one transaction
941 * Don't lock referenced row when "UPDATE" doesn't change foreign
943 * Use "-fPIC" not "-fpic" on Sparc (Tom Callaway)
944 * Repair lack of schema-awareness in contrib/reindexdb
945 * Fix contrib/intarray error for zero-element result array (Teodor)
946 * Ensure createuser script will exit on control-C (Oliver)
947 * Fix errors when the type of a dropped column has itself been
949 * "CHECKPOINT" does not cause database panic on failure in
951 * Accept 60 in seconds fields of timestamp, time, interval input
953 * Issue notice, not error, if TIMESTAMP, TIME, or INTERVAL precision
955 * Fix abstime-to-time cast function (fix is not applied unless you
957 * Fix pg_proc entry for timestampt_izone (fix is not applied unless
959 * Make EXTRACT(EPOCH FROM timestamp without time zone) treat input
961 * "'now'::timestamptz" gave wrong answer if timezone changed earlier
963 * HAVE_INT64_TIMESTAMP code for time with timezone overwrote its
965 * Accept "GLOBAL TEMP/TEMPORARY" as a synonym for "TEMPORARY"
966 * Avoid improper schema-privilege-check failure in foreign-key
968 * Fix bugs in foreign-key triggers for "SET DEFAULT" action
969 * Fix incorrect time-qual check in row fetch for "UPDATE" and
971 * Foreign-key clauses were parsed but ignored in "ALTER TABLE ADD
973 * Fix createlang script breakage for case where handler function
975 * Fix misbehavior on zero-column tables in pg_dump, COPY, ANALYZE,
977 * Fix misbehavior of func_error() on type names containing '%'
978 * Fix misbehavior of replace() on strings containing '%'
979 * Regular-expression patterns containing certain multibyte
981 * Account correctly for "NULL"s in more cases in join size
983 * Avoid conflict with system definition of isblank() function or
985 * Fix failure to convert large code point values in EUC_TW
987 * Fix error recovery for SSL_read/SSL_write calls
988 * Don't do early constant-folding of type coercion expressions
989 * Validate page header fields immediately after reading in any page
990 * Repair incorrect check for ungrouped variables in unnamed joins
991 * Fix buffer overrun in to_ascii (Guido Notari)
992 * contrib/ltree fixes (Teodor)
993 * Fix core dump in deadlock detection on machines where char is
995 * Avoid running out of buffers in many-way indexscan (bug introduced
997 * Fix planner's selectivity estimation functions to handle domains
999 * Fix dbmirror memory-allocation bug (Steven Singer)
1000 * Prevent infinite loop in ln(numeric) due to roundoff error
1001 * "GROUP BY" got confused if there were multiple equal GROUP BY
1003 * Fix bad plan when inherited "UPDATE"/"DELETE" references another
1005 * Prevent clustering on incomplete (partial or non-NULL-storing)
1007 * Service shutdown request at proper time if it arrives while still
1009 * Fix left-links in temporary indexes (could make backwards scans
1011 * Fix incorrect handling of client_encoding setting in
1012 postgresql.conf (Tatsuo)
1013 * Fix failure to respond to "pg_ctl stop -m fast" after
1014 Async_NotifyHandler runs
1015 * Fix SPI for case where rule contains multiple statements of the
1017 * Fix problem with checking for wrong type of access privilege in
1019 * Fix problem with "EXCEPT" in "CREATE RULE"
1020 * Prevent problem with dropping temp tables having serial columns
1021 * Fix replace_vars_with_subplan_refs failure in complex views
1022 * Fix regexp slowness in single-byte encodings (Tatsuo)
1023 * Allow qualified type names in "CREATE CAST" and " DROP CAST"
1024 * Accept SETOF type[], which formerly had to be written SETOF _type
1025 * Fix pg_dump core dump in some cases with procedural languages
1026 * Force ISO datestyle in pg_dump output, for portability (Oliver)
1027 * pg_dump failed to handle error return from lo_read (Oleg Drokin)
1028 * pg_dumpall failed with groups having no members (Nick Eskelinen)
1029 * pg_dumpall failed to recognize --globals-only switch
1030 * pg_restore failed to restore blobs if -X disable-triggers is
1032 * Repair intrafunction memory leak in plpgsql
1033 * pltcl's "elog" command dumped core if given wrong parameters (Ian
1035 * plpython used wrong value of atttypmod (Brad McLean)
1036 * Fix improper quoting of boolean values in Python interface
1038 * Added addDataType() method to PGConnection interface for JDBC
1039 * Fixed various problems with updateable ResultSets for JDBC (Shawn
1041 * Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka,
1043 * Fixed problem with parsing table ACLs in JDBC
1044 * Better error message for character set conversion problems in JDBC
1045 _________________________________________________________________
1049 Release date: 2003-02-04
1051 This release contains a variety of fixes for version 7.3.1.
1052 _________________________________________________________________
1054 Migration to version 7.3.2
1056 A dump/restore is *not* required for those running version 7.3.*.
1057 _________________________________________________________________
1061 * Restore creation of OID column in CREATE TABLE AS / SELECT INTO
1062 * Fix pg_dump core dump when dumping views having comments
1063 * Dump DEFERRABLE/INITIALLY DEFERRED constraints properly
1064 * Fix UPDATE when child table's column numbering differs from parent
1065 * Increase default value of max_fsm_relations
1066 * Fix problem when fetching backwards in a cursor for a single-row
1068 * Make backward fetch work properly with cursor on SELECT DISTINCT
1070 * Fix problems with loading pg_dump files containing contrib/lo
1072 * Fix problem with all-numeric user names
1073 * Fix possible memory leak and core dump during disconnect in
1075 * Make plpython's spi_execute command handle nulls properly (Andrew
1077 * Adjust plpython error reporting so that its regression test passes
1079 * Work with bison 1.875
1080 * Handle mixed-case names properly in plpgsql's %type (Neil)
1081 * Fix core dump in pltcl when executing a query rewritten by a rule
1082 * Repair array subscript overruns (per report from Yichen Xie)
1083 * Reduce MAX_TIME_PRECISION from 13 to 10 in floating-point case
1084 * Correctly case-fold variable names in per-database and per-user
1086 * Fix coredump in plpgsql's RETURN NEXT when SELECT into record
1088 * Fix outdated use of pg_type.typprtlen in python client interface
1089 * Correctly handle fractional seconds in timestamps in JDBC driver
1090 * Improve performance of getImportedKeys() in JDBC
1091 * Make shared-library symlinks work standardly on HPUX (Giles)
1092 * Repair inconsistent rounding behavior for timestamp, time,
1094 * SSL negotiation fixes (Nathan Mueller)
1095 * Make libpq's ~/.pgpass feature work when connecting with
1097 * Update my2pg, ora2pg
1098 * Translation updates
1099 * Add casts between types lo and oid in contrib/lo
1100 * fastpath code now checks for privilege to call function
1101 _________________________________________________________________
1105 Release date: 2002-12-18
1107 This release contains a variety of fixes for version 7.3.
1108 _________________________________________________________________
1110 Migration to version 7.3.1
1112 A dump/restore is *not* required for those running version 7.3.
1113 However, it should be noted that the main PostgreSQL interface
1114 library, libpq, has a new major version number for this release, which
1115 may require recompilation of client code in certain cases.
1116 _________________________________________________________________
1120 * Fix a core dump of COPY TO when client/server encodings don't
1122 * Allow pg_dump to work with pre-7.2 servers (Philip)
1123 * contrib/adddepend fixes (Tom)
1124 * Fix problem with deletion of per-user/per-database config settings
1126 * contrib/vacuumlo fix (Tom)
1127 * Allow 'password' encryption even when pg_shadow contains MD5
1129 * contrib/dbmirror fix (Steven Singer)
1130 * Optimizer fixes (Tom)
1131 * contrib/tsearch fixes (Teodor Sigaev, Magnus)
1132 * Allow locale names to be mixed case (Nicolai Tufar)
1133 * Increment libpq library's major version number (Bruce)
1134 * pg_hba.conf error reporting fixes (Bruce, Neil)
1135 * Add SCO Openserver 5.0.4 as a supported platform (Bruce)
1136 * Prevent EXPLAIN from crashing server (Tom)
1137 * SSL fixes (Nathan Mueller)
1138 * Prevent composite column creation via ALTER TABLE (Tom)
1139 _________________________________________________________________
1143 Release date: 2002-11-27
1144 _________________________________________________________________
1148 Major changes in this release:
1151 Schemas allow users to create objects in separate namespaces,
1152 so two people or applications can have tables with the same
1153 name. There is also a public schema for shared tables.
1154 Table/index creation can be restricted by removing privileges
1155 on the public schema.
1158 PostgreSQL now supports the ALTER TABLE ... DROP COLUMN
1162 Functions returning multiple rows and/or multiple columns are
1163 now much easier to use than before. You can call such a "table
1164 function" in the SELECT FROM clause, treating its output like a
1165 table. Also, PL/pgSQL functions can now return sets.
1168 PostgreSQL now supports prepared queries, for improved
1172 PostgreSQL now records object dependencies, which allows
1173 improvements in many areas. "DROP" statements now take either
1174 CASCADE or RESTRICT to control whether dependent objects are
1178 Functions and procedural languages now have privileges, and
1179 functions can be defined to run with the privileges of their
1182 Internationalization
1183 Both multibyte and locale support are now always enabled.
1186 A variety of logging options have been enhanced.
1189 A large number of interfaces have been moved to
1190 http://gborg.postgresql.org where they can be developed and
1191 released independently.
1193 Functions/Identifiers
1194 By default, functions can now take up to 32 parameters, and
1195 identifiers can be up to 63 bytes long. Also, OPAQUE is now
1196 deprecated: there are specific "pseudo-datatypes" to represent
1197 each of the former meanings of OPAQUE in function argument and
1199 _________________________________________________________________
1201 Migration to version 7.3
1203 A dump/restore using pg_dump is required for those wishing to migrate
1204 data from any previous release. If your application examines the
1205 system catalogs, additional changes will be required due to the
1206 introduction of schemas in 7.3; for more information, see:
1207 http://developer.postgresql.org/~momjian/upgrade_tips_7.3.
1209 Observe the following incompatibilities:
1211 * Pre-6.3 clients are no longer supported.
1212 * "pg_hba.conf" now has a column for the user name and additional
1213 features. Existing files need to be adjusted.
1214 * Several "postgresql.conf" logging parameters have been renamed.
1215 * LIMIT #,# has been disabled; use LIMIT # OFFSET #.
1216 * "INSERT" statements with column lists must specify a value for
1217 each specified column. For example, INSERT INTO tab (col1, col2)
1218 VALUES ('val1') is now invalid. It's still allowed to supply fewer
1219 columns than expected if the "INSERT" does not have a column list.
1220 * serial columns are no longer automatically UNIQUE; thus, an index
1221 will not automatically be created.
1222 * A "SET" command inside an aborted transaction is now rolled back.
1223 * "COPY" no longer considers missing trailing columns to be null.
1224 All columns need to be specified. (However, one may achieve a
1225 similar effect by specifying a column list in the "COPY" command.)
1226 * The data type timestamp is now equivalent to timestamp without
1227 time zone, instead of timestamp with time zone.
1228 * Pre-7.3 databases loaded into 7.3 will not have the new object
1229 dependencies for serial columns, unique constraints, and foreign
1230 keys. See the directory "contrib/adddepend/" for a detailed
1231 description and a script that will add such dependencies.
1232 * An empty string ('') is no longer allowed as the input into an
1233 integer field. Formerly, it was silently interpreted as 0.
1234 _________________________________________________________________
1240 * Add pg_locks view to show locks (Neil)
1241 * Security fixes for password negotiation memory allocation (Neil)
1242 * Remove support for version 0 FE/BE protocol (PostgreSQL 6.2 and
1244 * Reserve the last few backend slots for superusers, add parameter
1245 superuser_reserved_connections to control this (Nigel J. Andrews)
1246 _________________________________________________________________
1250 * Improve startup by calling localtime() only once (Tom)
1251 * Cache system catalog information in flat files for faster startup
1253 * Improve caching of index information (Tom)
1254 * Optimizer improvements (Tom, Fernando Nasser)
1255 * Catalog caches now store failed lookups (Tom)
1256 * Hash function improvements (Neil)
1257 * Improve performance of query tokenization and network handling
1259 * Speed improvement for large object restore (Mario Weilguni)
1260 * Mark expired index entries on first lookup, saving later heap
1262 * Avoid excessive NULL bitmap padding (Manfred Koizar)
1263 * Add BSD-licensed qsort() for Solaris, for performance (Bruce)
1264 * Reduce per-row overhead by four bytes (Manfred Koizar)
1265 * Fix GEQO optimizer bug (Neil Conway)
1266 * Make WITHOUT OID actually save four bytes per row (Manfred Koizar)
1267 * Add default_statistics_target variable to specify ANALYZE buckets
1269 * Use local buffer cache for temporary tables so no WAL overhead
1271 * Improve free space map performance on large tables (Stephen
1273 * Improved WAL write concurrency (Tom)
1274 _________________________________________________________________
1278 * Add privileges on functions and procedural languages (Peter)
1279 * Add OWNER to CREATE DATABASE so superusers can create databases on
1280 behalf of unprivileged users (Gavin Sherry, Tom)
1281 * Add new object privilege bits EXECUTE and USAGE (Tom)
1282 * Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION
1284 * Allow functions to be executed with the privilege of the function
1286 _________________________________________________________________
1288 Server Configuration
1290 * Server log messages now tagged with LOG, not DEBUG (Bruce)
1291 * Add user column to pg_hba.conf (Bruce)
1292 * Have log_connections output two lines in log file (Tom)
1293 * Remove debug_level from postgresql.conf, now server_min_messages
1295 * New ALTER DATABASE/USER ... SET command for per-user/database
1296 initialization (Peter)
1297 * New parameters server_min_messages and client_min_messages to
1298 control which messages are sent to the server logs or client
1299 applications (Bruce)
1300 * Allow pg_hba.conf to specify lists of users/databases separated by
1301 commas, group names prepended with +, and file names prepended
1303 * Remove secondary password file capability and pg_password utility
1305 * Add variable db_user_namespace for database-local user names
1307 * SSL improvements (Bear Giles)
1308 * Make encryption of stored passwords the default (Bruce)
1309 * Allow pg_statistics to be reset by calling pg_stat_reset()
1311 * Add log_duration parameter (Bruce)
1312 * Rename debug_print_query to log_statement (Bruce)
1313 * Rename show_query_stats to show_statement_stats (Bruce)
1314 * Add param log_min_error_statement to print commands to logs on
1316 _________________________________________________________________
1320 * Make cursors insensitive, meaning their contents do not change
1322 * Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported
1324 * Increase identifier length to 63 (Neil, Bruce)
1325 * UNION fixes for merging >= 3 columns of different lengths (Tom)
1326 * Add DEFAULT key word to INSERT, e.g., INSERT ... (..., DEFAULT,
1328 * Allow views to have default values using ALTER COLUMN ... SET
1330 * Fail on INSERTs with column lists that don't supply all column
1331 values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)
1332 * Fix for join aliases (Tom)
1333 * Fix for FULL OUTER JOINs (Tom)
1334 * Improve reporting of invalid identifier and location (Tom, Gavin)
1335 * Fix OPEN cursor(args) (Tom)
1336 * Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)
1337 * Fix for CREATE TABLE AS with UNION (Tom)
1338 * SQL99 syntax improvements (Thomas)
1339 * Add statement_timeout variable to cancel queries (Bruce)
1340 * Allow prepared queries with PREPARE/EXECUTE (Neil)
1341 * Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)
1342 * Add variable autocommit (Tom, David Van Wie)
1343 _________________________________________________________________
1347 * Make equals signs optional in CREATE DATABASE (Gavin Sherry)
1348 * Make ALTER TABLE OWNER change index ownership too (Neil)
1349 * New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls
1350 TOAST storage, compression (John Gray)
1351 * Add schema support, CREATE/DROP SCHEMA (Tom)
1352 * Create schema for temporary tables (Tom)
1353 * Add variable search_path for schema search (Tom)
1354 * Add ALTER TABLE SET/DROP NOT NULL (Christopher)
1355 * New CREATE FUNCTION volatility levels (Tom)
1356 * Make rule names unique only per table (Tom)
1357 * Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)
1358 * Add ALTER TRIGGER RENAME (Joe)
1359 * New current_schema() and current_schemas() inquiry functions (Tom)
1360 * Allow functions to return multiple rows (table functions) (Joe)
1361 * Make WITH optional in CREATE DATABASE, for consistency (Bruce)
1362 * Add object dependency tracking (Rod, Tom)
1363 * Add RESTRICT/CASCADE to DROP commands (Rod)
1364 * Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)
1365 * Autodestroy sequence on DROP of table with SERIAL (Rod)
1366 * Prevent column dropping if column is used by foreign key (Rod)
1367 * Automatically drop constraints/functions when object is dropped
1369 * Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)
1370 * Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)
1371 * Prevent inherited columns from being removed or renamed (Alvaro
1373 * Fix foreign key constraints to not error on intermediate database
1375 * Propagate column or table renaming to foreign key constraints
1376 * Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)
1377 * Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)
1378 * Have rules execute alphabetically, returning more predictable
1380 * Triggers are now fired in alphabetical order (Tom)
1381 * Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)
1382 * Allow better casting when inserting/updating values (Tom)
1383 _________________________________________________________________
1387 * Have COPY TO output embedded carriage returns and newlines as \r
1389 * Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)
1390 * Make pg_dump use ALTER TABLE ADD PRIMARY KEY, for performance
1392 * Disable brackets in multistatement rules (Bruce)
1393 * Disable VACUUM from being called inside a function (Bruce)
1394 * Allow dropdb and other scripts to use identifiers with spaces
1396 * Restrict database comment changes to the current database
1397 * Allow comments on operators, independent of the underlying
1399 * Rollback SET commands in aborted transactions (Tom)
1400 * EXPLAIN now outputs as a query (Tom)
1401 * Display condition expressions and sort keys in EXPLAIN (Tom)
1402 * Add 'SET LOCAL var = value' to set configuration variables for a
1403 single transaction (Tom)
1404 * Allow ANALYZE to run in a transaction (Bruce)
1405 * Improve COPY syntax using new WITH clauses, keep backward
1406 compatibility (Bruce)
1407 * Fix pg_dump to consistently output tags in non-ASCII dumps (Bruce)
1408 * Make foreign key constraints clearer in dump file (Rod)
1409 * Add COMMENT ON CONSTRAINT (Rod)
1410 * Allow COPY TO/FROM to specify column names (Brent Verner)
1411 * Dump UNIQUE and PRIMARY KEY constraints as ALTER TABLE (Rod)
1412 * Have SHOW output a query result (Joe)
1413 * Generate failure on short COPY lines rather than pad NULLs (Neil)
1414 * Fix CLUSTER to preserve all table attributes (Alvaro Herrera)
1415 * New pg_settings table to view/modify GUC settings (Joe)
1416 * Add smart quoting, portability improvements to pg_dump output
1418 * Dump serial columns out as SERIAL (Tom)
1419 * Enable large file support, >2G for pg_dump (Peter, Philip Warner,
1421 * Disallow TRUNCATE on tables that are involved in referential
1423 * Have TRUNCATE also auto-truncate the toast table of the relation
1425 * Add clusterdb utility that will auto-cluster an entire database
1426 based on previous CLUSTER operations (Alvaro Herrera)
1427 * Overhaul pg_dumpall (Peter)
1428 * Allow REINDEX of TOAST tables (Tom)
1429 * Implemented START TRANSACTION, per SQL99 (Neil)
1430 * Fix rare index corruption when a page split affects bulk delete
1432 * Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)
1433 _________________________________________________________________
1435 Data Types and Functions
1437 * Fix factorial(0) to return 1 (Bruce)
1438 * Date/time/timezone improvements (Thomas)
1439 * Fix for array slice extraction (Tom)
1440 * Fix extract/date_part to report proper microseconds for timestamp
1442 * Allow text_substr() and bytea_substr() to read TOAST values more
1443 efficiently (John Gray)
1444 * Add domain support (Rod)
1445 * Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data
1447 * Allow alternate storage scheme of 64-bit integers for date/time
1448 types using --enable-integer-datetimes in configure (Thomas)
1449 * Make timezone(timestamptz) return timestamp rather than a string
1451 * Allow fractional seconds in date/time types for dates prior to 1BC
1453 * Limit timestamp data types to 6 decimal places of precision
1455 * Change timezone conversion functions from timetz() to timezone()
1457 * Add configuration variables datestyle and timezone (Tom)
1458 * Add OVERLAY(), which allows substitution of a substring in a
1460 * Add SIMILAR TO (Thomas, Tom)
1461 * Add regular expression SUBSTRING(string FROM pat FOR escape)
1463 * Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)
1464 * Add named composite types using CREATE TYPE typename AS (column)
1466 * Allow composite type definition in the table alias clause (Joe)
1467 * Add new API to simplify creation of C language table functions
1469 * Remove ODBC-compatible empty parentheses from calls to SQL99
1470 functions for which these parentheses do not match the standard
1472 * Allow macaddr data type to accept 12 hex digits with no separators
1474 * Add CREATE/DROP CAST (Peter)
1475 * Add IS DISTINCT FROM operator (Thomas)
1476 * Add SQL99 TREAT() function, synonym for CAST() (Thomas)
1477 * Add pg_backend_pid() to output backend pid (Bruce)
1478 * Add IS OF / IS NOT OF type predicate (Thomas)
1479 * Allow bit string constants without fully-specified length (Thomas)
1480 * Allow conversion between 8-byte integers and bit strings (Thomas)
1481 * Implement hex literal conversion to bit string literal (Thomas)
1482 * Allow table functions to appear in the FROM clause (Joe)
1483 * Increase maximum number of function parameters to 32 (Bruce)
1484 * No longer automatically create index for SERIAL column (Tom)
1485 * Add current_database() (Rod)
1486 * Fix cash_words() to not overflow buffer (Tom)
1487 * Add functions replace(), split_part(), to_hex() (Joe)
1488 * Fix LIKE for bytea as a right-hand argument (Joe)
1489 * Prevent crashes caused by SELECT cash_out(2) (Tom)
1490 * Fix to_char(1,'FM999.99') to return a period (Karel)
1491 * Fix trigger/type/language functions returning OPAQUE to return
1493 _________________________________________________________________
1495 Internationalization
1497 * Add additional encodings: Korean (JOHAB), Thai (WIN874),
1498 Vietnamese (TCVN), Arabic (WIN1256), Simplified Chinese (GBK),
1499 Korean (UHC) (Eiji Tokuya)
1500 * Enable locale support by default (Peter)
1501 * Add locale variables (Peter)
1502 * Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea
1504 * Add locale awareness to regular expression character classes
1505 * Enable multibyte support by default (Tatsuo)
1506 * Add GB18030 multibyte support (Bill Huang)
1507 * Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo,
1509 * Add pg_conversion table (Tatsuo)
1510 * Add SQL99 CONVERT() function (Tatsuo)
1511 * pg_dumpall, pg_controldata, and pg_resetxlog now national-language
1513 * New and updated translations
1514 _________________________________________________________________
1516 Server-side Languages
1518 * Allow recursive SQL function (Peter)
1519 * Change PL/Tcl build to use configured compiler and Makefile.shlib
1521 * Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible
1523 * Allow PL/pgSQL to handle quoted identifiers (Tom)
1524 * Allow set-returning PL/pgSQL functions (Neil)
1525 * Make PL/pgSQL schema-aware (Joe)
1526 * Remove some memory leaks (Nigel J. Andrews, Tom)
1527 _________________________________________________________________
1531 * Don't lowercase psql \connect database name for 7.2.0
1533 * Add psql \timing to time user queries (Greg Sabino Mullane)
1534 * Have psql \d show index information (Greg Sabino Mullane)
1535 * New psql \dD shows domains (Jonathan Eisler)
1536 * Allow psql to show rules on views (Paul ?)
1537 * Fix for psql variable substitution (Tom)
1538 * Allow psql \d to show temporary table structure (Tom)
1539 * Allow psql \d to show foreign keys (Rod)
1540 * Fix \? to honor \pset pager (Bruce)
1541 * Have psql reports its version number on startup (Tom)
1542 * Allow \copy to specify column names (Tom)
1543 _________________________________________________________________
1547 * Add $HOME/.pgpass to store host/user password combinations (Alvaro
1549 * Add PQunescapeBytea() function to libpq (Patrick Welche)
1550 * Fix for sending large queries over non-blocking connections
1552 * Fix for libpq using timers on Win9X (David Ford)
1553 * Allow libpq notify to handle servers with different-length
1555 * Add libpq PQescapeString() and PQescapeBytea() to Windows (Bruce)
1556 * Fix for SSL with non-blocking connections (Jack Bates)
1557 * Add libpq connection timeout parameter (Denis A Ustimenko)
1558 _________________________________________________________________
1562 * Allow JDBC to compile with JDK 1.4 (Dave)
1563 * Add JDBC 3 support (Barry)
1564 * Allows JDBC to set loglevel by adding ?loglevel=X to the
1565 connection URL (Barry)
1566 * Add Driver.info() message that prints out the version number
1568 * Add updateable result sets (Raghu Nidagal, Dave)
1569 * Add support for callable statements (Paul Bethe)
1570 * Add query cancel capability
1571 * Add refresh row (Dave)
1572 * Fix MD5 encryption handling for multibyte servers (Jun Kawai)
1573 * Add support for prepared statements (Barry)
1574 _________________________________________________________________
1576 Miscellaneous Interfaces
1578 * Fixed ECPG bug concerning octal numbers in single quotes (Michael)
1579 * Move src/interfaces/libpgeasy to http://gborg.postgresql.org
1581 * Improve Python interface (Elliot Lee, Andrew Johnson, Greg
1583 * Add libpgtcl connection close event (Gerhard Hintermayer)
1584 * Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc,
1586 * Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)
1587 * Move src/interfaces/libpgeasy to http://gborg.postgresql.org
1589 * Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc,
1591 * Remove src/bin/pgaccess from main tree, now at
1592 http://www.pgaccess.org (Bruce)
1593 * Add pg_on_connection_loss command to libpgtcl (Gerhard
1595 _________________________________________________________________
1599 * Fix for parallel make (Peter)
1600 * AIX fixes for linking Tcl (Andreas Zeugswetter)
1601 * Allow PL/Perl to build under Cygwin (Jason Tishler)
1602 * Improve MIPS compiles (Peter, Oliver Elphick)
1603 * Require Autoconf version 2.53 (Peter)
1604 * Require readline and zlib by default in configure (Peter)
1605 * Allow Solaris to use Intimate Shared Memory (ISM), for performance
1606 (Scott Brunza, P.J. Josh Rovero)
1607 * Always enable syslog in compile, remove --enable-syslog option
1609 * Always enable multibyte in compile, remove --enable-multibyte
1611 * Always enable locale in compile, remove --enable-locale option
1613 * Fix for Win9x DLL creation (Magnus Naeslund)
1614 * Fix for link() usage by WAL code on Windows, BeOS (Jason Tishler)
1615 * Add sys/types.h to c.h, remove from main files (Peter, Bruce)
1616 * Fix AIX hang on SMP machines (Tomoyuki Niijima)
1617 * AIX SMP hang fix (Tomoyuki Niijima)
1618 * Fix pre-1970 date handling on newer glibc libraries (Tom)
1619 * Fix PowerPC SMP locking (Tom)
1620 * Prevent gcc -ffast-math from being used (Peter, Tom)
1621 * Bison >= 1.50 now required for developer builds
1622 * Kerberos 5 support now builds with Heimdal (Peter)
1623 * Add appendix in the User's Guide which lists SQL features (Thomas)
1624 * Improve loadable module linking to use RTLD_NOW (Tom)
1625 * New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)
1626 * New src/port directory holds replaced libc functions (Peter,
1628 * New pg_namespace system catalog for schemas (Tom)
1629 * Add pg_class.relnamespace for schemas (Tom)
1630 * Add pg_type.typnamespace for schemas (Tom)
1631 * Add pg_proc.pronamespace for schemas (Tom)
1632 * Restructure aggregates to have pg_proc entries (Tom)
1633 * System relations now have their own namespace, pg_* test not
1634 required (Fernando Nasser)
1635 * Rename TOAST index names to be *_index rather than *_idx (Neil)
1636 * Add namespaces for operators, opclasses (Tom)
1637 * Add additional checks to server control file (Thomas)
1638 * New Polish FAQ (Marcin Mazurek)
1639 * Add Posix semaphore support (Tom)
1640 * Document need for reindex (Bruce)
1641 * Rename some internal identifiers to simplify Windows compile (Jan,
1643 * Add documentation on computing disk space (Bruce)
1644 * Remove KSQO from GUC (Bruce)
1645 * Fix memory leak in rtree (Kenneth Been)
1646 * Modify a few error messages for consistency (Bruce)
1647 * Remove unused system table columns (Peter)
1648 * Make system columns NOT NULL where appropriate (Tom)
1649 * Clean up use of sprintf in favor of snprintf() (Neil, Jukka
1651 * Remove OPAQUE and create specific subtypes (Tom)
1652 * Cleanups in array internal handling (Joe, Tom)
1653 * Disallow pg_atoi('') (Bruce)
1654 * Remove parameter wal_files because WAL files are now recycled
1656 * Add version numbers to heap pages (Tom)
1657 _________________________________________________________________
1661 * Allow inet arrays in /contrib/array (Neil)
1662 * GiST fixes (Teodor Sigaev, Neil)
1663 * Upgrade /contrib/mysql
1664 * Add /contrib/dbsize which shows table sizes without vacuum (Peter)
1665 * Add /contrib/intagg, integer aggregator routines (mlw)
1666 * Improve /contrib/oid2name (Neil, Bruce)
1667 * Improve /contrib/tsearch (Oleg, Teodor Sigaev)
1668 * Cleanups of /contrib/rserver (Alexey V. Borzov)
1669 * Update /contrib/oracle conversion utility (Gilles Darold)
1670 * Update /contrib/dblink (Joe)
1671 * Improve options supported by /contrib/vacuumlo (Mario Weilguni)
1672 * Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey
1674 * Add /contrib/reindexdb utility (Shaun Thomas)
1675 * Add indexing to /contrib/isbn_issn (Dan Weston)
1676 * Add /contrib/dbmirror (Steven Singer)
1677 * Improve /contrib/pgbench (Neil)
1678 * Add /contrib/tablefunc table function examples (Joe)
1679 * Add /contrib/ltree data type for tree structures (Teodor Sigaev,
1681 * Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)
1682 * Fixes to /contrib/cube (Bruno Wolff)
1683 * Improve /contrib/fulltextindex (Christopher)
1684 _________________________________________________________________
1688 Release date: 2003-01-30
1690 This release contains a variety of fixes for version 7.2.3, including
1691 fixes to prevent possible data loss.
1692 _________________________________________________________________
1694 Migration to version 7.2.4
1696 A dump/restore is *not* required for those running version 7.2.*.
1697 _________________________________________________________________
1701 * Fix some additional cases of VACUUM "No one parent tuple was
1703 * Prevent VACUUM from being called inside a function (Bruce)
1704 * Ensure pg_clog updates are sync'd to disk before marking
1706 * Avoid integer overflow during large hash joins
1707 * Make GROUP commands work when pg_group.grolist is large enough to
1709 * Fix errors in datetime tables; some timezone names weren't being
1711 * Fix integer overflows in circle_poly(), path_encode(), path_add()
1713 * Repair long-standing logic errors in lseg_eq(), lseg_ne(),
1715 _________________________________________________________________
1719 Release date: 2002-10-01
1721 This release contains a variety of fixes for version 7.2.2, including
1722 fixes to prevent possible data loss.
1723 _________________________________________________________________
1725 Migration to version 7.2.3
1727 A dump/restore is *not* required for those running version 7.2.*.
1728 _________________________________________________________________
1732 * Prevent possible compressed transaction log loss (Tom)
1733 * Prevent non-superuser from increasing most recent vacuum info
1735 * Handle pre-1970 date values in newer versions of glibc (Tom)
1736 * Fix possible hang during server shutdown
1737 * Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)
1738 * Fix pg_dump to properly dump FULL JOIN USING (Tom)
1739 _________________________________________________________________
1743 Release date: 2002-08-23
1745 This release contains a variety of fixes for version 7.2.1.
1746 _________________________________________________________________
1748 Migration to version 7.2.2
1750 A dump/restore is *not* required for those running version 7.2.*.
1751 _________________________________________________________________
1755 * Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
1756 * Fix for compressed transaction log id wraparound (Tom)
1757 * Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f
1759 * Fix for psql and pg_dump crashing when invoked with non-existent
1760 long options (Tatsuo)
1761 * Fix crash when invoking geometric operators (Tom)
1762 * Allow OPEN cursor(args) (Tom)
1763 * Fix for rtree_gist index build (Teodor)
1764 * Fix for dumping user-defined aggregates (Tom)
1765 * contrib/intarray fixes (Oleg)
1766 * Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)
1767 * Fix to pg_convert (Tatsuo)
1768 * Fix for crash with long DATA strings (Thomas, Neil)
1769 * Fix for repeat(), lpad(), rpad() and long strings (Neil)
1770 _________________________________________________________________
1774 Release date: 2002-03-21
1776 This release contains a variety of fixes for version 7.2.
1777 _________________________________________________________________
1779 Migration to version 7.2.1
1781 A dump/restore is *not* required for those running version 7.2.
1782 _________________________________________________________________
1786 * Ensure that sequence counters do not go backwards after a crash
1788 * Fix pgaccess kanji-conversion key binding (Tatsuo)
1789 * Optimizer improvements (Tom)
1790 * Cash I/O improvements (Tom)
1792 * Compile fix for missing AuthBlockSig (Heiko)
1793 * Additional time zones and time zone fixes (Thomas)
1794 * Allow psql \connect to handle mixed case database and user names
1796 * Return proper OID on command completion even with ON INSERT rules
1798 * Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)
1799 * Fix bug in extract/date_part for milliseconds/microseconds
1801 * Improve handling of multiple UNIONs with different lengths (Tom)
1802 * contrib/btree_gist improvements (Teodor Sigaev)
1803 * contrib/tsearch dictionary improvements, see README.tsearch for an
1804 additional installation step (Thomas T. Thai, Teodor Sigaev)
1805 * Fix for array subscripts handling (Tom)
1806 * Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)
1807 _________________________________________________________________
1811 Release date: 2002-02-04
1812 _________________________________________________________________
1816 This release improves PostgreSQL for use in high-volume applications.
1818 Major changes in this release:
1821 Vacuuming no longer locks tables, thus allowing normal user
1822 access during the vacuum. A new "VACUUM FULL" command does
1823 old-style vacuum by locking the table and shrinking the on-disk
1827 There is no longer a problem with installations that exceed
1828 four billion transactions.
1831 OIDs are now optional. Users can now create tables without OIDs
1832 for cases where OID usage is excessive.
1835 The system now computes histogram column statistics during
1836 "ANALYZE", allowing much better optimizer choices.
1839 A new MD5 encryption option allows more secure storage and
1840 transfer of passwords. A new Unix-domain socket authentication
1841 option is available on Linux and BSD systems.
1844 Administrators can use the new table access statistics module
1845 to get fine-grained information about table and index usage.
1847 Internationalization
1848 Program and library messages can now be displayed in several
1850 _________________________________________________________________
1852 Migration to version 7.2
1854 A dump/restore using "pg_dump" is required for those wishing to
1855 migrate data from any previous release.
1857 Observe the following incompatibilities:
1859 * The semantics of the "VACUUM" command have changed in this
1860 release. You may wish to update your maintenance procedures
1862 * In this release, comparisons using = NULL will always return false
1863 (or NULL, more precisely). Previous releases automatically
1864 transformed this syntax to IS NULL. The old behavior can be
1865 re-enabled using a "postgresql.conf" parameter.
1866 * The "pg_hba.conf" and "pg_ident.conf" configuration is now only
1867 reloaded after receiving a SIGHUP signal, not with each
1869 * The function "octet_length()" now returns the uncompressed data
1871 * The date/time value 'current' is no longer available. You will
1872 need to rewrite your applications.
1873 * The timestamp(), time(), and interval() functions are no longer
1874 available. Instead of timestamp(), use timestamp 'string' or CAST.
1876 The SELECT ... LIMIT #,# syntax will be removed in the next release.
1877 You should change your queries to use separate LIMIT and OFFSET
1878 clauses, e.g. LIMIT 10 OFFSET 20.
1879 _________________________________________________________________
1885 * Create temporary files in a separate directory (Bruce)
1886 * Delete orphaned temporary files on postmaster startup (Bruce)
1887 * Added unique indexes to some system tables (Tom)
1888 * System table operator reorganization (Oleg Bartunov, Teodor
1890 * Renamed pg_log to pg_clog (Tom)
1891 * Enable SIGTERM, SIGQUIT to kill backends (Jan)
1892 * Removed compile-time limit on number of backends (Tom)
1893 * Better cleanup for semaphore resource failure (Tatsuo, Tom)
1894 * Allow safe transaction ID wraparound (Tom)
1895 * Removed OIDs from some system tables (Tom)
1896 * Removed "triggered data change violation" error check (Tom)
1897 * SPI portal creation of prepared/saved plans (Jan)
1898 * Allow SPI column functions to work for system columns (Tom)
1899 * Long value compression improvement (Tom)
1900 * Statistics collector for table, index access (Jan)
1901 * Truncate extra-long sequence names to a reasonable value (Tom)
1902 * Measure transaction times in milliseconds (Thomas)
1903 * Fix TID sequential scans (Hiroshi)
1904 * Superuser ID now fixed at 1 (Peter E)
1905 * New pg_ctl "reload" option (Tom)
1906 _________________________________________________________________
1910 * Optimizer improvements (Tom)
1911 * New histogram column statistics for optimizer (Tom)
1912 * Reuse write-ahead log files rather than discarding them (Tom)
1913 * Cache improvements (Tom)
1914 * IS NULL, IS NOT NULL optimizer improvement (Tom)
1915 * Improve lock manager to reduce lock contention (Tom)
1916 * Keep relcache entries for index access support functions (Tom)
1917 * Allow better selectivity with NaN and infinities in NUMERIC (Tom)
1918 * R-tree performance improvements (Kenneth Been)
1919 * B-tree splits more efficient (Tom)
1920 _________________________________________________________________
1924 * Change UPDATE, DELETE privileges to be distinct (Peter E)
1925 * New REFERENCES, TRIGGER privileges (Peter E)
1926 * Allow GRANT/REVOKE to/from more than one user at a time (Peter E)
1927 * New has_table_privilege() function (Joe Conway)
1928 * Allow non-superuser to vacuum database (Tom)
1929 * New SET SESSION AUTHORIZATION command (Peter E)
1930 * Fix bug in privilege modifications on newly created tables (Tom)
1931 * Disallow access to pg_statistic for non-superuser, add
1932 user-accessible views (Tom)
1933 _________________________________________________________________
1935 Client Authentication
1937 * Fork postmaster before doing authentication to prevent hangs
1939 * Add ident authentication over Unix domain sockets on Linux, *BSD
1940 (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)
1941 * Add a password authentication method that uses MD5 encryption
1943 * Allow encryption of stored passwords using MD5 (Bruce)
1944 * PAM authentication (Dominic J. Eidson)
1945 * Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP
1947 _________________________________________________________________
1949 Server Configuration
1951 * Interpretation of some time zone abbreviations as Australian
1952 rather than North American now settable at run time (Bruce)
1953 * New parameter to set default transaction isolation level (Peter E)
1954 * New parameter to enable conversion of "expr = NULL" into "expr IS
1955 NULL", off by default (Peter E)
1956 * New parameter to control memory usage by VACUUM (Tom)
1957 * New parameter to set client authentication timeout (Tom)
1958 * New parameter to set maximum number of open files (Tom)
1959 _________________________________________________________________
1963 * Statements added by INSERT rules now execute after the INSERT
1965 * Prevent unadorned relation names in target list (Bruce)
1966 * NULLs now sort after all normal values in ORDER BY (Tom)
1967 * New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)
1968 * New SHARE UPDATE EXCLUSIVE lock mode (Tom)
1969 * New EXPLAIN ANALYZE command that shows run times and row counts
1970 (Martijn van Oosterhout)
1971 * Fix problem with LIMIT and subqueries (Tom)
1972 * Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)
1973 * Fix nested EXCEPT/INTERSECT (Tom)
1974 _________________________________________________________________
1978 * Fix SERIAL in temporary tables (Bruce)
1979 * Allow temporary sequences (Bruce)
1980 * Sequences now use int8 internally (Tom)
1981 * New SERIAL8 creates int8 columns with sequences, default still
1983 * Make OIDs optional using WITHOUT OIDS (Tom)
1984 * Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)
1985 * Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints
1986 (Christopher Kings-Lynne)
1987 * New CREATE OR REPLACE FUNCTION to alter existing function
1988 (preserving the function OID) (Gavin Sherry)
1989 * Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher
1991 * Allow column renaming in views
1992 * Make ALTER TABLE / RENAME COLUMN update column names of indexes
1994 * Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited
1995 tables (Stephan Szabo)
1996 * ALTER TABLE RENAME update foreign-key trigger arguments correctly
1998 * DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype
2000 * Add automatic return type data casting for SQL functions (Tom)
2001 * Allow GiST indexes to handle NULLs and multikey indexes (Oleg
2002 Bartunov, Teodor Sigaev, Tom)
2003 * Enable partial indexes (Martijn van Oosterhout)
2004 _________________________________________________________________
2008 * Add RESET ALL, SHOW ALL (Marko Kreen)
2009 * CREATE/ALTER USER/GROUP now allow options in any order (Vince)
2010 * Add LOCK A, B, C functionality (Neil Padgett)
2011 * New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)
2012 * New light-weight VACUUM does not lock table; old semantics are
2013 available as VACUUM FULL (Tom)
2014 * Disable COPY TO/FROM on views (Bruce)
2015 * COPY DELIMITERS string must be exactly one character (Tom)
2016 * VACUUM warning about index tuples fewer than heap now only appears
2017 when appropriate (Martijn van Oosterhout)
2018 * Fix privilege checks for CREATE INDEX (Tom)
2019 * Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)
2020 _________________________________________________________________
2022 Data Types and Functions
2024 * SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)
2025 * Add convert(), convert2() (Tatsuo)
2026 * New function bit_length() (Peter E)
2027 * Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes
2029 * CHAR(), VARCHAR() now reject strings that are too long (Peter E)
2030 * BIT VARYING now rejects bit strings that are too long (Peter E)
2031 * BIT now rejects bit strings that do not match declared size (Peter
2033 * INET, CIDR text conversion functions (Alex Pilosov)
2034 * INET, CIDR operators << and <<= indexable (Alex Pilosov)
2035 * Bytea \### now requires valid three digit octal number
2036 * Bytea comparison improvements, now supports =, <>, >, >=, <, and
2038 * Bytea now supports B-tree indexes
2039 * Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT
2041 * Bytea now supports concatenation
2042 * New bytea functions: position, substring, trim, btrim, and length
2043 * New encode() function mode, "escaped", converts minimally escaped
2045 * Add pg_database_encoding_max_length() (Tatsuo)
2046 * Add pg_client_encoding() function (Tatsuo)
2047 * now() returns time with millisecond precision (Thomas)
2048 * New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)
2049 * Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss
2051 * New xid/int comparison functions (Hiroshi)
2052 * Add precision to TIME, TIMESTAMP, and INTERVAL data types (Thomas)
2053 * Modify type coercion logic to attempt binary-compatible functions
2055 * New encode() function installed by default (Marko Kreen)
2056 * Improved to_*() conversion functions (Karel Zak)
2057 * Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)
2058 * New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(),
2059 gen_salt() (Marko Kreen)
2060 * Correct description of translate() function (Bruce)
2061 * Add INTERVAL argument for SET TIME ZONE (Thomas)
2062 * Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)
2063 * Optimize length functions when using single-byte encodings
2065 * Fix path_inter, path_distance, path_length, dist_ppath to handle
2066 closed paths (Curtis Barrett, Tom)
2067 * octet_length(text) now returns non-compressed length (Tatsuo,
2069 * Handle "July" full name in date/time literals (Greg Sabino
2071 * Some datatype() function calls now evaluated differently
2072 * Add support for Julian and ISO time specifications (Thomas)
2073 _________________________________________________________________
2075 Internationalization
2077 * National language support in psql, pg_dump, libpq, and server
2079 * Message translations in Chinese (simplified, traditional), Czech,
2080 French, German, Hungarian, Russian, Swedish (Peter E, Serguei A.
2081 Mokhov, Karel Zak, Weiping He, Zhenbang Wei, Kovacs Zoltan)
2082 * Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte
2084 * Add LATIN5,6,7,8,9,10 support (Tatsuo)
2085 * Add ISO 8859-5,6,7,8 support (Tatsuo)
2086 * Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)
2087 * Make mic2ascii() non-ASCII aware (Tatsuo)
2088 * Reject invalid multibyte character sequences (Tatsuo)
2089 _________________________________________________________________
2093 * Now uses portals for SELECT loops, allowing huge result sets (Jan)
2094 * CURSOR and REFCURSOR support (Jan)
2095 * Can now return open cursors (Jan)
2096 * Add ELSEIF (Klaus Reger)
2097 * Improve PL/pgSQL error reporting, including location of error
2099 * Allow IS or FOR key words in cursor declaration, for compatibility
2101 * Fix for SELECT ... FOR UPDATE (Tom)
2102 * Fix for PERFORM returning multiple rows (Tom)
2103 * Make PL/pgSQL use the server's type coercion code (Tom)
2104 * Memory leak fix (Jan, Tom)
2105 * Make trailing semicolon optional (Tom)
2106 _________________________________________________________________
2110 * New untrusted PL/Perl (Alex Pilosov)
2111 * PL/Perl is now built on some platforms even if libperl is not
2113 _________________________________________________________________
2117 * Now reports errorInfo (Vsevolod Lobko)
2118 * Add spi_lastoid function (bob@redivi.com)
2119 _________________________________________________________________
2123 * ...is new (Andrew Bosma)
2124 _________________________________________________________________
2128 * \d displays indexes in unique, primary groupings (Christopher
2130 * Allow trailing semicolons in backslash commands (Greg Sabino
2132 * Read password from /dev/tty if possible
2133 * Force new password prompt when changing user and database (Tatsuo,
2135 * Format the correct number of columns for Unicode (Patrice)
2136 _________________________________________________________________
2140 * New function PQescapeString() to escape quotes in command strings
2142 * New function PQescapeBytea() escapes binary strings for use as SQL
2144 _________________________________________________________________
2148 * Return OID of INSERT (Ken K)
2149 * Handle more data types (Ken K)
2150 * Handle single quotes and newlines in strings (Ken K)
2151 * Handle NULL variables (Ken K)
2152 * Fix for time zone handling (Barry Lind)
2153 * Improved Druid support
2154 * Allow eight-bit characters with non-multibyte server (Barry Lind)
2155 * Support BIT, BINARY types (Ned Wolpert)
2156 * Reduce memory usage (Michael Stephens, Dave Cramer)
2157 * Update DatabaseMetaData (Peter E)
2158 * Add DatabaseMetaData.getCatalogs() (Peter E)
2159 * Encoding fixes (Anders Bengtsson)
2160 * Get/setCatalog methods (Jason Davies)
2161 * DatabaseMetaData.getColumns() now returns column defaults (Jason
2163 * DatabaseMetaData.getColumns() performance improvement (Jeroen van
2165 * Some JDBC1 and JDBC2 merging (Anders Bengtsson)
2166 * Transaction performance improvements (Barry Lind)
2167 * Array fixes (Greg Zoller)
2168 * Serialize addition
2169 * Fix batch processing (Rene Pijlman)
2170 * ExecSQL method reorganization (Anders Bengtsson)
2171 * GetColumn() fixes (Jeroen van Vianen)
2172 * Fix isWriteable() function (Rene Pijlman)
2173 * Improved passage of JDBC2 conformance tests (Rene Pijlman)
2174 * Add bytea type capability (Barry Lind)
2175 * Add isNullable() (Rene Pijlman)
2176 * JDBC date/time test suite fixes (Liam Stewart)
2177 * Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)
2178 * Fix DatabaseMetaData to show precision properly (Mark Lillywhite)
2179 * New getImported/getExported keys (Jason Davies)
2180 * MD5 password encryption support (Jeremy Wohl)
2181 * Fix to actually use type cache (Ned Wolpert)
2182 _________________________________________________________________
2186 * Remove query size limit (Hiroshi)
2187 * Remove text field size limit (Hiroshi)
2188 * Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)
2189 * Allow ODBC procedure calls (Hiroshi)
2190 * Improve boolean handing (Aidan Mountford)
2191 * Most configuration options now settable via DSN (Hiroshi)
2192 * Multibyte, performance fixes (Hiroshi)
2193 * Allow driver to be used with iODBC or unixODBC (Peter E)
2194 * MD5 password encryption support (Bruce)
2195 * Add more compatibility functions to odbc.sql (Peter E)
2196 _________________________________________________________________
2200 * EXECUTE ... INTO implemented (Christof Petig)
2201 * Multiple row descriptor support (e.g. CARDINALITY) (Christof
2203 * Fix for GRANT parameters (Lee Kindness)
2204 * Fix INITIALLY DEFERRED bug
2205 * Various bug fixes (Michael, Christof Petig)
2206 * Auto allocation for indicator variable arrays (int *ind_p=NULL)
2207 * Auto allocation for string arrays (char **foo_pp=NULL)
2208 * ECPGfree_auto_mem fixed
2209 * All function names with external linkage are now prefixed by ECPG
2210 * Fixes for arrays of structures (Michael)
2211 _________________________________________________________________
2215 * Python fix fetchone() (Gerhard Haring)
2216 * Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko,
2218 * Add Tcl COPY TO/FROM (ljb)
2219 * Prevent output of default index op class in pg_dump (Tom)
2220 * Fix libpgeasy memory leak (Bruce)
2221 _________________________________________________________________
2225 * Configure, dynamic loader, and shared library fixes (Peter E)
2226 * Fixes in QNX 4 port (Bernd Tegge)
2227 * Fixes in Cygwin and Windows ports (Jason Tishler, Gerhard Haring,
2228 Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)
2229 * Fix for Windows socket communication failures (Magnus, Mikhail
2231 * Hurd compile fix (Oliver Elphick)
2232 * BeOS fixes (Cyril Velter)
2233 * Remove configure --enable-unicode-conversion, now enabled by
2235 * AIX fixes (Tatsuo, Andreas)
2236 * Fix parallel make (Peter E)
2237 * Install SQL language manual pages into OS-specific directories
2239 * Rename config.h to pg_config.h (Peter E)
2240 * Reorganize installation layout of header files (Peter E)
2241 _________________________________________________________________
2245 * Remove SEP_CHAR (Bruce)
2246 * New GUC hooks (Tom)
2247 * Merge GUC and command line handling (Marko Kreen)
2248 * Remove EXTEND INDEX (Martijn van Oosterhout, Tom)
2249 * New pgjindent utility to indent java code (Bruce)
2250 * Remove define of true/false when compiling under C++ (Leandro
2252 * pgindent fixes (Bruce, Tom)
2253 * Replace strcasecmp() with strcmp() where appropriate (Peter E)
2254 * Dynahash portability improvements (Tom)
2255 * Add 'volatile' usage in spinlock structures
2256 * Improve signal handling logic (Tom)
2257 _________________________________________________________________
2261 * New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)
2262 * New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)
2263 * Add contrib/dblink for remote database access (Joe Conway)
2264 * contrib/ora2pg Oracle conversion utility (Gilles Darold)
2265 * contrib/xml XML conversion utility (John Gray)
2266 * contrib/fulltextindex fixes (Christopher Kings-Lynne)
2267 * New contrib/fuzzystrmatch with levenshtein and metaphone, soundex
2269 * Add contrib/intarray boolean queries, binary search, fixes (Oleg
2271 * New pg_upgrade utility (Bruce)
2272 * Add new pg_resetxlog options (Bruce, Tom)
2273 _________________________________________________________________
2277 Release date: 2001-08-15
2278 _________________________________________________________________
2280 Migration to version 7.1.3
2282 A dump/restore is *not* required for those running 7.1.X.
2283 _________________________________________________________________
2287 Remove unused WAL segements of large transactions (Tom)
2288 Multiaction rule fix (Tom)
2289 PL/pgSQL memory allocation fix (Jan)
2290 VACUUM buffer fix (Tom)
2291 Regression test fixes (Tom)
2292 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
2293 Fix subselects with DISTINCT ON or LIMIT (Tom)
2295 Disable COPY TO/FROM a view (Tom)
2296 Cygwin build (Jason Tishler)
2298 _________________________________________________________________
2302 Release date: 2001-05-11
2304 This has one fix from 7.1.1.
2305 _________________________________________________________________
2307 Migration to version 7.1.2
2309 A dump/restore is *not* required for those running 7.1.X.
2310 _________________________________________________________________
2314 Fix PL/pgSQL SELECTs when returning no rows
2315 Fix for psql backslash core dump
2316 Referential integrity privilege fix
2320 _________________________________________________________________
2324 Release date: 2001-05-05
2326 This has a variety of fixes from 7.1.
2327 _________________________________________________________________
2329 Migration to version 7.1.1
2331 A dump/restore is *not* required for those running 7.1.
2332 _________________________________________________________________
2336 Fix for numeric MODULO operator (Tom)
2337 pg_dump fixes (Philip)
2338 pg_dump can dump 7.0 databases (Philip)
2339 readline 4.2 fixes (Peter E)
2341 AIX, MSWIN, VAX, N32K fixes (Tom)
2342 Multibytes fixes (Tom)
2343 Unicode fixes (Tatsuo)
2344 Optimizer improvements (Tom)
2345 Fix for whole rows in functions (Tom)
2346 Fix for pg_ctl and option strings with spaces (Peter E)
2347 ODBC fixes (Hiroshi)
2348 EXTRACT can now take string argument (Thomas)
2349 Python fixes (Darcy)
2351 _________________________________________________________________
2355 Release date: 2001-04-13
2357 This release focuses on removing limitations that have existed in the
2358 PostgreSQL code for many years.
2360 Major changes in this release:
2362 Write-ahead Log (WAL)
2363 To maintain database consistency in case of an operating system
2364 crash, previous releases of PostgreSQL have forced all data
2365 modifications to disk before each transaction commit. With WAL,
2366 only one log file must be flushed to disk, greatly improving
2367 performance. If you have been using -F in previous releases to
2368 disable disk flushes, you may want to consider discontinuing
2372 TOAST - Previous releases had a compiled-in row length limit,
2373 typically 8k - 32k. This limit made storage of long text fields
2374 difficult. With TOAST, long rows of any length can be stored
2375 with good performance.
2378 We now support outer joins. The UNION/NOT IN workaround for
2379 outer joins is no longer required. We use the SQL92 outer join
2383 The previous C function manager did not handle null values
2384 properly, nor did it support 64-bit CPU's (Alpha). The new
2385 function manager does. You can continue using your old custom
2386 functions, but you may want to rewrite them in the future to
2387 use the new function manager call interface.
2390 A large number of complex queries that were unsupported in
2391 previous releases now work. Many combinations of views,
2392 aggregates, UNION, LIMIT, cursors, subqueries, and inherited
2393 tables now work properly. Inherited tables are now accessed by
2394 default. Subqueries in FROM are now supported.
2395 _________________________________________________________________
2397 Migration to version 7.1
2399 A dump/restore using pg_dump is required for those wishing to migrate
2400 data from any previous release.
2401 _________________________________________________________________
2407 Many multibyte/Unicode/locale fixes (Tatsuo and others)
2408 More reliable ALTER TABLE RENAME (Tom)
2409 Kerberos V fixes (David Wragg)
2410 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
2411 Prompt username/password on standard error (Bruce)
2412 Large objects inv_read/inv_write fixes (Tom)
2413 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
2415 Prevent query expressions from leaking memory (Tom)
2416 Allow UPDATE of arrays elements (Tom)
2417 Wake up lock waiters during cancel (Hiroshi)
2418 Fix rare cursor crash when using hash join (Tom)
2419 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
2420 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
2421 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
2422 Fix PL/perl (Alex Kapranoff)
2423 Disallow LOCK on views (Mark Hollomon)
2424 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
2425 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
2426 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
2427 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
2428 Fix rare failure with TRUNCATE command (Tom)
2429 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
2430 DISTINCT, ORDER BY, SELECT...INTO (Tom)
2431 Fix parser failures during aborted transactions (Tom)
2432 Allow temporary relations to properly clean up indexes (Bruce)
2433 Fix VACUUM problem with moving rows in same page (Tom)
2434 Modify pg_dump to better handle user-defined items in template1 (Philip)
2435 Allow LIMIT in VIEW (Tom)
2436 Require cursor FETCH to honor LIMIT (Tom)
2437 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
2438 Allow ORDER BY, LIMIT in subqueries (Tom)
2439 Allow UNION in CREATE RULE (Tom)
2440 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
2441 Store initdb collation in pg_control so collation cannot be changed (Tom)
2442 Fix INSERT...SELECT with rules (Tom)
2443 Fix FOR UPDATE inside views and subselects (Tom)
2444 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
2445 Fix lpad() and rpad() to handle length less than input string (Tom)
2446 Fix use of NOTIFY in some rules (Tom)
2447 Overhaul btree code (Tom)
2448 Fix NOT NULL use in Pl/pgSQL variables (Tom)
2449 Overhaul GIST code (Oleg)
2450 Fix CLUSTER to preserve constraints and column default (Tom)
2451 Improved deadlock detection handling (Tom)
2452 Allow multiple SERIAL columns in a table (Tom)
2453 Prevent occasional index corruption (Vadim)
2457 Add OUTER JOINs (Tom)
2458 Function manager overhaul (Tom)
2459 Allow ALTER TABLE RENAME on indexes (Tom)
2460 Improve CLUSTER (Tom)
2461 Improve ps status display for more platforms (Peter E, Marc)
2462 Improve CREATE FUNCTION failure message (Ross)
2463 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
2465 Grand Unified Configuration scheme/GUC. Many options can now be set in
2466 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
2467 Improved handling of file descriptor cache (Tom)
2468 New warning code about auto-created table alias entries (Bruce)
2469 Overhaul initdb process (Tom, Peter E)
2470 Overhaul of inherited tables; inherited tables now accessed by default;
2471 new ONLY key word prevents it (Chris Bitmead, Tom)
2472 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
2474 Allow renaming of temp tables (Tom)
2475 Overhaul memory manager contexts (Tom)
2476 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
2477 Overhaul pg_dump (Philip Warner)
2478 Allow pg_hba.conf secondary password file to specify only username (Peter E)
2479 Allow TEMPORARY or TEMP key word when creating temporary tables (Bruce)
2480 New memory leak checker (Karel)
2481 New SET SESSION CHARACTERISTICS (Thomas)
2482 Allow nested block comments (Thomas)
2483 Add WITHOUT TIME ZONE type qualifier (Thomas)
2484 New ALTER TABLE ADD CONSTRAINT (Stephan)
2485 Use NUMERIC accumulators for INTEGER aggregates (Tom)
2486 Overhaul aggregate code (Tom)
2487 New VARIANCE and STDDEV() aggregates
2488 Improve dependency ordering of pg_dump (Philip)
2489 New pg_restore command (Philip)
2490 New pg_dump tar output option (Philip)
2491 New pg_dump of large objects (Philip)
2492 New ESCAPE option to LIKE (Thomas)
2493 New case-insensitive LIKE - ILIKE (Thomas)
2494 Allow functional indexes to use binary-compatible type (Tom)
2495 Allow SQL functions to be used in more contexts (Tom)
2496 New pg_config utility (Peter E)
2497 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
2499 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
2500 New quote_identifiers() and quote_literal() functions (Jan)
2501 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
2502 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
2503 Update PyGreSQL to version 3.1 (D'Arcy)
2504 Store tables as files named by OID (Vadim)
2505 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
2506 Require DROP VIEW to remove views, no DROP TABLE (Mark)
2507 Allow DROP VIEW view1, view2 (Mark)
2508 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
2509 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
2510 New /contrib/pgcrypto hashing functions (Marko Kreen)
2511 New pg_dumpall --globals-only option (Peter E)
2512 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
2513 New AT TIME ZONE syntax (Thomas)
2514 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
2515 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
2516 Allow socket path name to be specified in hostname by using leading slash
2517 (David J. MacKenzie)
2518 Allow CREATE DATABASE to specify template database (Tom)
2519 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
2520 New /contrib/rserv replication toolkit (Vadim)
2521 New file format for COPY BINARY (Tom)
2522 New /contrib/oid2name to map numeric files to table names (B Palmer)
2523 New "idle in transaction" ps status message (Marc)
2524 Update to pgaccess 0.98.7 (Constantin Teodorescu)
2525 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
2526 Add rudimentary dependency checking to pg_dump (Philip)
2530 Fix INET/CIDR type ordering and add new functions (Tom)
2531 Make OID behave as an unsigned type (Tom)
2532 Allow BIGINT as synonym for INT8 (Peter E)
2533 New int2 and int8 comparison operators (Tom)
2534 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
2535 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
2536 New GIST seg/cube examples (Gene Selkov)
2537 Improved round(numeric) handling (Tom)
2538 Fix CIDR output formatting (Tom)
2539 New CIDR abbrev() function (Tom)
2543 Write-Ahead Log (WAL) to provide crash recovery with less performance
2545 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
2546 Reduced file seeks (Denis Perchine)
2547 Improve BTREE code for duplicate keys (Tom)
2548 Store all large objects in a single table (Denis Perchine, Tom)
2549 Improve memory allocation performance (Karel, Tom)
2553 New function manager call conventions (Tom)
2554 SGI portability fixes (David Kaelbling)
2555 New configure --enable-syslog option (Peter E)
2556 New BSDI README (Bruce)
2557 configure script moved to top level, not /src (Peter E)
2558 Makefile/configuration/compilation overhaul (Peter E)
2559 New configure --with-python option (Peter E)
2560 Solaris cleanups (Peter E)
2561 Overhaul /contrib Makefiles (Karel)
2562 New OpenSSL configuration option (Magnus, Peter E)
2564 QNX fixes (Maurizio)
2565 New heap_open(), heap_openr() API (Tom)
2566 Remove colon and semi-colon operators (Thomas)
2567 New pg_class.relkind value for views (Mark Hollomon)
2568 Rename ichar() to chr() (Karel)
2569 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
2570 Fixes for NT/Cygwin (Pete Forman)
2571 AIX port fixes (Andreas)
2572 New BeOS port (David Reid, Cyril Velter)
2573 Add proofreader's changes to docs (Addison-Wesley, Bruce)
2574 New Alpha spinlock code (Adriaan Joubert, Compaq)
2575 UnixWare port overhaul (Peter E)
2576 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
2577 New FreeBSD Alpha port (Alfred)
2578 Overhaul shared memory segments (Tom)
2579 Add IBM S/390 support (Neale Ferguson)
2580 Moved macmanuf to /contrib (Larry Rosenman)
2581 Syslog improvements (Larry Rosenman)
2582 New template0 database that contains no user additions (Tom)
2583 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
2584 Allow NetBSD's libedit instead of readline (Peter)
2585 Improved assembly language source code format (Bruce)
2586 New contrib/pg_logger
2587 New --template option to createdb
2588 New contrib/pg_control utility (Oliver)
2589 New FreeBSD tools ipc_check, start-scripts/freebsd
2591 _________________________________________________________________
2595 Release date: 2000-11-11
2597 This has a variety of fixes from 7.0.2.
2598 _________________________________________________________________
2600 Migration to version 7.0.3
2602 A dump/restore is *not* required for those running 7.0.*.
2603 _________________________________________________________________
2608 Large object fix (Tom)
2609 Fix lean in COPY WITH OIDS leak (Tom)
2610 Fix backwards-index-scan (Tom)
2611 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
2612 Add --enable-syslog to configure (Marc)
2613 Fix abort transaction at backend exit in rare cases (Tom)
2614 Fix for psql \l+ when multibyte enabled (Tatsuo)
2615 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
2616 Make vacuum always flush buffers (Tom)
2617 Fix to allow cancel while waiting for a lock (Hiroshi)
2618 Fix for memory aloocation problem in user authentication code (Tom)
2619 Remove bogus use of int4out() (Tom)
2620 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
2621 Fix for failure of triggers on heap open in certain cases (Jeroen van
2623 Fix for erroneous selectivity of not-equals (Tom)
2624 Fix for erroneous use of strcmp() (Tom)
2625 Fix for bug where storage manager accesses items beyond end of file
2627 Fix to include kernel errno message in all smgr elog messages (Tom)
2628 Fix for '.' not in PATH at build time (SL Baur)
2629 Fix for out-of-file-descriptors error (Tom)
2630 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
2631 Fix for subselect in targetlist of Append node (Tom)
2632 Fix for mergejoin plans (Tom)
2633 Fix TRUNCATE failure on relations with indexes (Tom)
2634 Avoid database-wide restart on write error (Hiroshi)
2635 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
2636 Fix VACUUM problem with moving chain of update row versions when source
2637 and destination of a row version lie on the same page (Tom)
2638 Fix user.c CommandCounterIncrement (Tom)
2639 Fix for AM/PM boundary problem in to_char() (Karel Zak)
2640 Fix TIME aggregate handling (Tom)
2641 Fix to_char() to avoid coredump on NULL input (Tom)
2643 Fix for inserting/copying longer multibyte strings into char() data
2645 Fix for crash of backend, on abort (Tom)
2647 _________________________________________________________________
2651 Release date: 2000-06-05
2653 This is a repackaging of 7.0.1 with added documentation.
2654 _________________________________________________________________
2656 Migration to version 7.0.2
2658 A dump/restore is *not* required for those running 7.*.
2659 _________________________________________________________________
2663 Added documentation to tarball.
2665 _________________________________________________________________
2669 Release date: 2000-06-01
2671 This is a cleanup release for 7.0.
2672 _________________________________________________________________
2674 Migration to version 7.0.1
2676 A dump/restore is *not* required for those running 7.0.
2677 _________________________________________________________________
2681 Fix many CLUSTER failures (Tom)
2682 Allow ALTER TABLE RENAME works on indexes (Tom)
2683 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
2684 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
2685 Fix the off by one errors in ResultSet from 6.5.3, and more.
2686 jdbc ResultSet fixes (Joseph Shraibman)
2687 optimizer tunings (Tom)
2688 Fix create user for pgaccess
2689 Fix for UNLISTEN failure
2690 IRIX fixes (David Kaelbling)
2691 QNX fixes (Andreas Kardos)
2692 Reduce COPY IN lock level (Tom)
2693 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
2694 Fix pg_dump to handle OID indexes (Tom)
2695 Fix small memory leak (Tom)
2696 Solaris fix for createdb/dropdb (Tatsuo)
2697 Fix for non-blocking connections (Alfred Perlstein)
2698 Fix improper recovery after RENAME TABLE failures (Tom)
2699 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
2700 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
2701 Fix too long syslog message (Tatsuo)
2702 Fix problem with quoted indexes that are too long (Tom)
2703 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
2704 ecpg changes (Michael)
2706 _________________________________________________________________
2710 Release date: 2000-05-08
2712 This release contains improvements in many areas, demonstrating the
2713 continued growth of PostgreSQL. There are more improvements and fixes
2714 in 7.0 than in any previous release. The developers have confidence
2715 that this is the best release yet; we do our best to put out only
2716 solid releases, and this one is no exception.
2718 Major changes in this release:
2721 Foreign keys are now implemented, with the exception of PARTIAL
2722 MATCH foreign keys. Many users have been asking for this
2723 feature, and we are pleased to offer it.
2726 Continuing on work started a year ago, the optimizer has been
2727 improved, allowing better query plan selection and faster
2728 performance with less memory usage.
2731 psql, our interactive terminal monitor, has been updated with a
2732 variety of new features. See the psql manual page for details.
2735 SQL92 join syntax is now supported, though only as INNER JOIN
2736 for this release. JOIN, NATURAL JOIN, JOIN/USING, and JOIN/ON
2737 are available, as are column correlation names.
2738 _________________________________________________________________
2740 Migration to version 7.0
2742 A dump/restore using pg_dump is required for those wishing to migrate
2743 data from any previous release of PostgreSQL. For those upgrading from
2744 6.5.*, you may instead use pg_upgrade to upgrade to this release;
2745 however, a full dump/reload installation is always the most robust
2746 method for upgrades.
2748 Interface and compatibility issues to consider for the new release
2751 * The date/time types datetime and timespan have been superseded by
2752 the SQL92-defined types timestamp and interval. Although there has
2753 been some effort to ease the transition by allowing PostgreSQL to
2754 recognize the deprecated type names and translate them to the new
2755 type names, this mechanism may not be completely transparent to
2756 your existing application.
2757 * The optimizer has been substantially improved in the area of query
2758 cost estimation. In some cases, this will result in decreased
2759 query times as the optimizer makes a better choice for the
2760 preferred plan. However, in a small number of cases, usually
2761 involving pathological distributions of data, your query times may
2762 go up. If you are dealing with large amounts of data, you may want
2763 to check your queries to verify performance.
2764 * The JDBC and ODBC interfaces have been upgraded and extended.
2765 * The string function CHAR_LENGTH is now a native function. Previous
2766 versions translated this into a call to LENGTH, which could result
2767 in ambiguity with other types implementing LENGTH such as the
2769 _________________________________________________________________
2775 Prevent function calls exceeding maximum number of arguments (Tom)
2776 Improve CASE construct (Tom)
2777 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
2778 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
2779 Fix GROUP BY scan bug (Tom)
2780 Improvements in SQL grammar processing (Tom)
2781 Fix for views involved in INSERT ... SELECT ... (Tom)
2782 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
2783 Fix for subselects in INSERT ... SELECT (Tom)
2784 Prevent INSERT ... SELECT ... ORDER BY (Tom)
2785 Fixes for relations greater than 2GB, including vacuum
2786 Improve propagating system table changes to other backends (Tom)
2787 Improve propagating user table changes to other backends (Tom)
2788 Fix handling of temp tables in complex situations (Bruce, Tom)
2789 Allow table locking at table open, improving concurrent reliability (Tom)
2790 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
2791 Prevent DROP DATABASE while others accessing
2792 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
2793 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
2794 Fix pg_upgrade so it works for MVCC (Tom)
2795 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
2796 Fix for "f1 datetime DEFAULT 'now'" (Tom)
2797 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
2798 Allow comment-only lines, and ;;; lines too. (Tom)
2799 Improve recovery after failed disk writes, disk full (Hiroshi)
2800 Fix cases where table is mentioned in FROM but not joined (Tom)
2801 Allow HAVING clause without aggregate functions (Tom)
2802 Fix for "--" comment and no trailing newline, as seen in perl interface
2803 Improve pg_dump failure error reports (Bruce)
2804 Allow sorts and hashes to exceed 2GB file sizes (Tom)
2805 Fix for pg_dump dumping of inherited rules (Tom)
2806 Fix for NULL handling comparisons (Tom)
2807 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
2808 Fix for dbname with dash
2809 Prevent DROP INDEX from interfering with other backends (Tom)
2810 Fix file descriptor leak in verify_password()
2811 Fix for "Unable to identify an operator =$" problem
2812 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
2813 Fix for recursive exit call (Massimo)
2814 Fix for extra-long timezones (Jeroen van Vianen)
2815 Make pg_dump preserve primary key information (Peter E)
2816 Prevent databases with single quotes (Peter E)
2817 Prevent DROP DATABASE inside transaction (Peter E)
2818 ecpg memory leak fixes (Stephen Birch)
2819 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
2820 Y2K timestamp fix (Massimo)
2821 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
2822 Fix for views with tables/columns containing spaces (Tom)
2823 Prevent privileges on indexes (Peter E)
2824 Fix for spinlock stuck problem when error is generated (Hiroshi)
2825 Fix ipcclean on Linux
2826 Fix handling of NULL constraint conditions (Tom)
2827 Fix memory leak in odbc driver (Nick Gorham)
2828 Fix for privilege check on UNION tables (Tom)
2829 Fix to allow SELECT 'a' LIKE 'a' (Tom)
2830 Fix for SELECT 1 + NULL (Tom)
2832 Fix log() on numeric type (Tom)
2833 Deprecate ':' and ';' operators
2834 Allow vacuum of temporary tables
2835 Disallow inherited columns with the same name as new columns
2836 Recover or force failure when disk space is exhausted (Hiroshi)
2837 Fix INSERT INTO ... SELECT with AS columns matching result columns
2838 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns
2840 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
2841 Fix UNION with LIMIT
2842 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
2843 Fix CREATE TABLE test(col char(2) DEFAULT user)
2844 Fix mismatched types in CREATE TABLE ... DEFAULT
2845 Fix SELECT * FROM pg_class where oid in (0,-1)
2846 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
2847 Prevent user who can create databases can modifying pg_database table (Peter E)
2848 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
2849 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
2853 New CLI interface include file sqlcli.h, based on SQL3/SQL98
2854 Remove all limits on query length, row length limit still exists (Tom)
2855 Update jdbc protocol to 2.0 (Jens Glaser <jens@jens.de>)
2856 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
2857 Fix to give super user and createdb user proper update catalog rights (Peter E)
2858 Allow ecpg bool variables to have NULL values (Christof)
2859 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
2860 Allow ^C to cancel COPY command (Massimo)
2861 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
2862 Function name overloading for dynamically-loaded C functions (Frankpitt)
2863 Add CmdTuples() to libpq++(Vince)
2864 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
2865 Allow CREATE FUNCTION/WITH clause to be used for all language types
2866 configure --enable-debug adds -g (Peter E)
2867 configure --disable-debug removes -g (Peter E)
2868 Allow more complex default expressions (Tom)
2869 First real FOREIGN KEY constraint trigger functionality (Jan)
2870 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
2871 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
2872 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
2873 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
2874 Change pgeasy connectdb() parameter ordering (Bruce)
2875 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
2876 Add Oracle's COMMENT ON command (Mike Mascari <mascarim@yahoo.com>)
2877 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
2878 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
2879 Make USING in COPY optional (Bruce)
2880 Allow subselects in the target list (Tom)
2881 Allow subselects on the left side of comparison operators (Tom)
2882 New parallel regression test (Jan)
2883 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
2884 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
2885 Added psql LASTOID variable to return last inserted oid (Peter E)
2886 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
2887 Add privilege check for vacuum (Peter E)
2888 New libpq functions to allow asynchronous connections: PQconnectStart(),
2889 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
2890 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
2891 New libpq PQsetenv() function (Ewan Mellor)
2892 create/alter user extension (Peter E)
2893 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
2894 New scripts for create/drop user/db (Peter E)
2895 Major psql overhaul (Peter E)
2896 Add const to libpq interface (Peter E)
2897 New libpq function PQoidValue (Peter E)
2898 Show specific non-aggregate causing problem with GROUP BY (Tom)
2899 Make changes to pg_shadow recreate pg_pwd file (Peter E)
2900 Add aggregate(DISTINCT ...) (Tom)
2901 Allow flag to control COPY input/output of NULLs (Peter E)
2902 Make postgres user have a password by default (Peter E)
2903 Add CREATE/ALTER/DROP GROUP (Peter E)
2904 All administration scripts now support --long options (Peter E, Karel)
2905 Vacuumdb script now supports --all option (Peter E)
2906 ecpg new portable FETCH syntax
2907 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
2908 and EXEC SQL ENDIF directives
2909 Add pg_ctl script to control backend start-up (Tatsuo)
2910 Add postmaster.opts.default file to store start-up flags (Tatsuo)
2911 Allow --with-mb=SQL_ASCII
2912 Increase maximum number of index keys to 16 (Bruce)
2913 Increase maximum number of function arguments to 16 (Bruce)
2914 Allow configuration of maximum number of index keys and arguments (Bruce)
2915 Allow unprivileged users to change their passwords (Peter E)
2916 Password authentication enabled; required for new users (Peter E)
2917 Disallow dropping a user who owns a database (Peter E)
2918 Change initdb option --with-mb to --enable-multibyte
2919 Add option for initdb to prompts for superuser password (Peter E)
2920 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
2921 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
2922 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
2923 libpq non-blocking mode (Alfred Perlstein)
2924 Improve conversion of types in casts that don't specify a length
2925 New plperl internal programming language (Mark Hollomon)
2926 Allow COPY IN to read file that do not end with a newline (Tom)
2927 Indicate when long identifiers are truncated (Tom)
2928 Allow aggregates to use type equivalency (Peter E)
2929 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
2930 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
2931 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
2932 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
2933 Add NUMERIC and int8 types to ODBC
2934 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
2935 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
2936 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
2937 Enable backward sequential scan even after reaching EOF (Hiroshi)
2938 Add btree indexing of boolean values, >= and <= (Don Baccus)
2939 Print current line number when COPY FROM fails (Massimo)
2940 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
2941 Add DEC as synonym for DECIMAL (Thomas)
2942 Add SESSION_USER as SQL92 key word, same as CURRENT_USER (Thomas)
2943 Implement SQL92 column aliases (aka correlation names) (Thomas)
2944 Implement SQL92 join syntax (Thomas)
2945 Make INTERVAL reserved word allowed as a column identifier (Thomas)
2946 Implement REINDEX command (Hiroshi)
2947 Accept ALL in aggregate function SUM(ALL col) (Tom)
2948 Prevent GROUP BY from using column aliases (Tom)
2949 New psql \encoding option (Tatsuo)
2950 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
2951 Allow negation of a negative number in all cases
2952 Add ecpg descriptors (Christof, Michael)
2953 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
2954 Allow casts with length, like foo::char(8)
2955 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
2956 Add support for SJIS user defined characters (Tatsuo)
2957 Larger views/rules supported
2958 Make libpq's PQconndefaults() thread-safe (Tom)
2959 Disable // as comment to be ANSI conforming, should use -- (Tom)
2960 Allow column aliases on views CREATE VIEW name (collist)
2961 Fixes for views with subqueries (Tom)
2962 Allow UPDATE table SET fld = (SELECT ...) (Tom)
2963 SET command options no longer require quotes
2964 Update pgaccess to 0.98.6
2965 New SET SEED command
2966 New pg_options.sample file
2967 New SET FSYNC command (Massimo)
2968 Allow pg_descriptions when creating tables
2969 Allow pg_descriptions when creating types, columns, and functions
2970 Allow psql \copy to allow delimiters (Peter E)
2971 Allow psql to print nulls as distinct from "" [null] (Peter E)
2975 Many array fixes (Tom)
2976 Allow bare column names to be subscripted as arrays (Tom)
2977 Improve type casting of int and float constants (Tom)
2978 Cleanups for int8 inputs, range checking, and type conversion (Tom)
2979 Fix for SELECT timespan('21:11:26'::time) (Tom)
2980 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
2981 Add btree index on NUMERIC (Jan)
2982 Perl fix for large objects containing NUL characters (Douglas Thomson)
2983 ODBC fix for for large objects (free)
2984 Fix indexing of cidr data type
2985 Fix for Ethernet MAC addresses (macaddr type) comparisons
2986 Fix for date/time types when overflows happened in computations (Tom)
2987 Allow array on int8 (Peter E)
2988 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
2989 Allow NUMERIC arrays
2990 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
2991 Make char_length()/octet_length including trailing blanks (Tom)
2992 Made abstime/reltime use int4 instead of time_t (Peter E)
2993 New lztext data type for compressed text fields
2994 Revise code to handle coercion of int and float constants (Tom)
2995 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
2996 NUMERIC now accepts scientific notation (Tom)
2997 NUMERIC to int4 rounds (Tom)
2998 Convert float4/8 to NUMERIC properly (Tom)
2999 Allow type conversion with NUMERIC (Thomas)
3000 Make ISO date style (2000-02-16 09:33) the default (Thomas)
3001 Add NATIONAL CHAR [ VARYING ] (Thomas)
3002 Allow NUMERIC round and trunc to accept negative scales (Tom)
3003 New TIME WITH TIME ZONE type (Thomas)
3004 Add MAX()/MIN() on time type (Thomas)
3005 Add abs(), mod(), fac() for int8 (Thomas)
3006 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
3007 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
3008 Add exp() and ln() for NUMERIC type
3009 Rename NUMERIC power() to pow() (Thomas)
3010 Improved TRANSLATE() function (Edwin Ramirez, Tom)
3011 Allow X=-Y operators (Tom)
3012 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom
3014 Allow LOCALE to use indexes in regular expression searches (Tom)
3015 Allow creation of functional indexes to use default types
3019 Prevent exponential space consumption with many AND's and OR's (Tom)
3020 Collect attribute selectivity values for system columns (Tom)
3021 Reduce memory usage of aggregates (Tom)
3022 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
3023 Fix r-tree index optimizer selectivity (Thomas)
3024 Improve optimizer selectivity computations and functions (Tom)
3025 Optimize btree searching for cases where many equal keys exist (Tom)
3026 Enable fast LIKE index processing only if index present (Tom)
3027 Re-use free space on index pages with duplicates (Tom)
3028 Improve hash join processing (Tom)
3029 Prevent descending sort if result is already sorted(Hiroshi)
3030 Allow commuting of index scan query qualifications (Tom)
3031 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
3032 Allocate large memory requests in fix-sized chunks for performance (Tom)
3033 Fix vacuum's performance by reducing memory allocation requests (Tom)
3034 Implement constant-expression simplification (Bernard Frankpitt, Tom)
3035 Use secondary columns to be used to determine start of index scan (Hiroshi)
3036 Prevent quadruple use of disk space when doing internal sorting (Tom)
3037 Faster sorting by calling fewer functions (Tom)
3038 Create system indexes to match all system caches (Bruce, Hiroshi)
3039 Make system caches use system indexes (Bruce)
3040 Make all system indexes unique (Bruce)
3041 Improve pg_statistics management for VACUUM speed improvement (Tom)
3042 Flush backend cache less frequently (Tom, Hiroshi)
3043 COPY now reuses previous memory allocation, improving performance (Tom)
3044 Improve optimization cost estimation (Tom)
3045 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom
3047 Use DNF instead of CNF where appropriate (Tom, Taral)
3048 Further cleanup for OR-of-AND WHERE-clauses (Tom)
3049 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
3050 Smarter optimizer computations for random index page access (Tom)
3051 New SET variable to control optimizer costs (Tom)
3052 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
3053 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
3054 Major subquery speedup (Tom)
3055 Fewer fsync writes when fsync is not disabled (Tom)
3056 Improved LIKE optimizer estimates (Tom)
3057 Prevent fsync in SELECT-only queries (Vadim)
3058 Make index creation use psort code, because it is now faster (Tom)
3059 Allow creation of sort temp tables > 1 Gig
3063 Fix for linux PPC compile
3064 New generic expression-tree-walker subroutine (Tom)
3065 Change form() to varargform() to prevent portability problems
3066 Improved range checking for large integers on Alphas
3067 Clean up #include in /include directory (Bruce)
3068 Add scripts for checking includes (Bruce)
3069 Remove un-needed #include's from *.c files (Bruce)
3070 Change #include's to use <> and "" as appropriate (Bruce)
3071 Enable Windows compilation of libpq
3072 Alpha spinlock fix from Uncle George <gatgul@voicenet.com>
3073 Overhaul of optimizer data structures (Tom)
3074 Fix to cygipc library (Yutaka Tanida)
3075 Allow pgsql to work on newer Cygwin snapshots (Dan)
3076 New catalog version number (Tom)
3078 Rename heap_replace to heap_update
3079 Update for QNX (Dr. Andreas Kardos)
3080 New platform-specific regression handling (Tom)
3081 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
3082 Included all yacc and lex files into the distribution (Peter E.)
3083 Remove lextest, no longer needed (Peter E)
3084 Fix for libpq and psql on Windows (Magnus)
3085 Internally change datetime and timespan into timestamp and interval (Thomas)
3086 Fix for plpgsql on BSD/OS
3087 Add SQL_ASCII test case to the regression test (Tatsuo)
3088 configure --with-mb now deprecated (Tatsuo)
3090 NetBSD fixes (Johnny C. Lam <lamj@stat.cmu.edu>)
3091 Fixes for Alpha compiles
3092 New multibyte encodings
3094 _________________________________________________________________
3098 Release date: 1999-10-13
3100 This is basically a cleanup release for 6.5.2. We have added a new
3101 PgAccess that was missing in 6.5.2, and installed an NT-specific fix.
3102 _________________________________________________________________
3104 Migration to version 6.5.3
3106 A dump/restore is *not* required for those running 6.5.*.
3107 _________________________________________________________________
3111 Updated version of pgaccess 0.98
3113 Fix dumping rules on inherited tables
3115 _________________________________________________________________
3119 Release date: 1999-09-15
3121 This is basically a cleanup release for 6.5.1. We have fixed a variety
3122 of problems reported by 6.5.1 users.
3123 _________________________________________________________________
3125 Migration to version 6.5.2
3127 A dump/restore is *not* required for those running 6.5.*.
3128 _________________________________________________________________
3132 subselect+CASE fixes(Tom)
3133 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
3134 Fixes for CASE in WHERE join clauses(Tom)
3135 Fix BTScan abort(Tom)
3136 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
3137 Improve it so that it checks for multicolumn constraints(Thomas)
3138 Fix for Windows making problem with MB enabled(Hiroki Kataoka)
3139 Allow BSD yacc and bison to compile pl code(Bruce)
3140 Fix SET NAMES working
3142 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
3143 Reduce the total memory consumption of vacuum(Tom)
3144 Fix for timestamp(datetime)
3145 Rule deparsing bugfixes(Tom)
3146 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(To
3148 This is to re-use space on index pages freed by vacuum(Vadim)
3149 document -x for pg_dump(Bruce)
3150 Fix for unary operators in rule deparser(Tom)
3151 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
3152 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
3153 Repair logic error in LIKE: should not return LIKE_ABORT
3154 when reach end of pattern before end of text(Tom)
3155 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom
3157 Updated version of pgaccess 0.98
3159 _________________________________________________________________
3163 Release date: 1999-07-15
3165 This is basically a cleanup release for 6.5. We have fixed a variety
3166 of problems reported by 6.5 users.
3167 _________________________________________________________________
3169 Migration to version 6.5.1
3171 A dump/restore is *not* required for those running 6.5.
3172 _________________________________________________________________
3177 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
3178 Remove QUERY_LIMIT, use SELECT...LIMIT
3179 Fix for EXPLAIN on inheritance(Tom)
3180 Patch to allow vacuum on multisegment tables(Hiroshi)
3181 R-Tree optimizer selectivity fix(Tom)
3182 ACL file descriptor leak fix(Atsushi Ogawa)
3183 New expresssion subtree code(Tom)
3184 Avoid disk writes for read-only transactions(Vadim)
3185 Fix for removal of temp tables if last transaction was aborted(Bruce)
3186 Fix to prevent too large row from being created(Bruce)
3188 Allow port numbers 32k - 64k(Bruce)
3189 Add ^ precidence(Bruce)
3190 Rename sort files called pg_temp to pg_sorttemp(Bruce)
3191 Fix for microseconds in time values(Tom)
3192 Tutorial source cleanup
3194 Fix for sorting of NULL's in some cases(Tom)
3195 Shared library dependencies fixed (Tom)
3196 Fixed glitches affecting GROUP BY in subselects(Tom)
3197 Fix some compiler warnings (Tomoaki Nishiyama)
3198 Add Win1250 (Czech) support (Pavel Behal)
3200 _________________________________________________________________
3204 Release date: 1999-06-09
3206 This release marks a major step in the development team's mastery of
3207 the source code we inherited from Berkeley. You will see we are now
3208 easily adding major features, thanks to the increasing size and
3209 experience of our world-wide development team.
3211 Here is a brief summary of the more notable changes:
3213 Multiversion concurrency control(MVCC)
3214 This removes our old table-level locking, and replaces it with
3215 a locking system that is superior to most commercial database
3216 systems. In a traditional system, each row that is modified is
3217 locked until committed, preventing reads by other users. MVCC
3218 uses the natural multiversion nature of PostgreSQL to allow
3219 readers to continue reading consistent data during writer
3220 activity. Writers continue to use the compact pg_log
3221 transaction system. This is all performed without having to
3222 allocate a lock for every row like traditional database
3223 systems. So, basically, we no longer are restricted by simple
3224 table-level locking; we have something better than row-level
3227 Hot backups from pg_dump
3228 pg_dump takes advantage of the new MVCC features to give a
3229 consistent database dump/backup while the database stays online
3230 and available for queries.
3233 We now have a true numeric data type, with user-specified
3237 Temporary tables are guaranteed to have unique names within a
3238 database session, and are destroyed on session exit.
3241 We now have CASE, INTERSECT, and EXCEPT statement support. We
3242 have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL, SELECT
3243 ... FOR UPDATE, and an improved LOCK TABLE command.
3246 We continue to speed up PostgreSQL, thanks to the variety of
3247 talents within our team. We have sped up memory allocation,
3248 optimization, table joins, and row transfer routines.
3251 We continue to expand our port list, this time including
3252 Windows NT/ix86 and NetBSD/arm32.
3255 Most interfaces have new versions, and existing functionality
3259 New and updated material is present throughout the
3260 documentation. New FAQs have been contributed for SGI and AIX
3261 platforms. The Tutorial has introductory information on SQL
3262 from Stefan Simkovics. For the User's Guide, there are
3263 reference pages covering the postmaster and more utility
3264 programs, and a new appendix contains details on date/time
3265 behavior. The Administrator's Guide has a new chapter on
3266 troubleshooting from Tom Lane. And the Programmer's Guide has a
3267 description of query processing, also from Stefan, and details
3268 on obtaining the PostgreSQL source tree via anonymous CVS and
3270 _________________________________________________________________
3272 Migration to version 6.5
3274 A dump/restore using pg_dump is required for those wishing to migrate
3275 data from any previous release of PostgreSQL. pg_upgrade can *not* be
3276 used to upgrade to this release because the on-disk structure of the
3277 tables has changed compared to previous releases.
3279 The new Multiversion Concurrency Control (MVCC) features can give
3280 somewhat different behaviors in multiuser environments. *Read and
3281 understand the following section to ensure that your existing
3282 applications will give you the behavior you need.*
3283 _________________________________________________________________
3285 Multiversion Concurrency Control
3287 Because readers in 6.5 don't lock data, regardless of transaction
3288 isolation level, data read by one transaction can be overwritten by
3289 another. In other words, if a row is returned by "SELECT" it doesn't
3290 mean that this row really exists at the time it is returned (i.e.
3291 sometime after the statement or transaction began) nor that the row is
3292 protected from being deleted or updated by concurrent transactions
3293 before the current transaction does a commit or rollback.
3295 To ensure the actual existence of a row and protect it against
3296 concurrent updates one must use "SELECT FOR UPDATE" or an appropriate
3297 "LOCK TABLE" statement. This should be taken into account when porting
3298 applications from previous releases of PostgreSQL and other
3301 Keep the above in mind if you are using "contrib/refint.*" triggers
3302 for referential integrity. Additional techniques are required now. One
3303 way is to use "LOCK parent_table IN SHARE ROW EXCLUSIVE MODE" command
3304 if a transaction is going to update/delete a primary key and use "LOCK
3305 parent_table IN SHARE MODE" command if a transaction is going to
3306 update/insert a foreign key.
3308 Note: Note that if you run a transaction in SERIALIZABLE mode then
3309 you must execute the "LOCK" commands above before execution of any
3310 DML statement ("SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO") in the
3313 These inconveniences will disappear in the future when the ability to
3314 read dirty (uncommitted) data (regardless of isolation level) and true
3315 referential integrity will be implemented.
3316 _________________________________________________________________
3322 Fix text<->float8 and text<->float4 conversion functions(Thomas)
3323 Fix for creating tables with mixed-case constraints(Billy)
3324 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
3325 Fix bug in pg_dump -z
3326 Memory overrun cleanups(Tatsuo)
3327 Fix for lo_import crash(Tatsuo)
3328 Adjust handling of data type names to suppress double quotes(Thomas)
3329 Use type coercion for matching columns and DEFAULT(Thomas)
3330 Fix deadlock so it only checks once after one second of sleep(Bruce)
3331 Fixes for aggregates and PL/pgsql(Hiroshi)
3332 Fix for subquery crash(Vadim)
3333 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
3334 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo
3336 Fix for pg_dump -d or -D and quote special characters in INSERT
3337 Repair serious problems with dynahash(Tom)
3338 Fix INET/CIDR portability problems
3339 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
3340 Fix executor so mergejoin of different column types works(Tom)
3341 Fix for Alpha OR selectivity bug
3342 Fix OR index selectivity problem(Bruce)
3343 Fix so \d shows proper length for char()/varchar()(Ryan)
3344 Fix tutorial code(Clark)
3345 Improve destroyuser checking(Oliver)
3346 Fix for Kerberos(Rodney McDuff)
3347 Fix for dropping database while dirty buffers(Bruce)
3348 Fix so sequence nextval() can be case-sensitive(Bruce)
3350 Drop buffers before destroying database files(Bruce)
3351 Fix case where executor evaluates functions twice(Tatsuo)
3352 Allow sequence nextval actions to be case-sensitive(Bruce)
3353 Fix optimizer indexing not working for negative numbers(Bruce)
3354 Fix for memory leak in executor with fjIsNull
3355 Fix for aggregate memory leaks(Erik Riedel)
3356 Allow user name containing a dash to grant privileges
3357 Cleanup of NULL in inet types
3358 Clean up system table bugs(Tom)
3359 Fix problems of PAGER and \? command(Masaaki Sakaida)
3360 Reduce default multisegment file size limit to 1GB(Peter)
3361 Fix for dumping of CREATE OPERATOR(Tom)
3362 Fix for backward scanning of cursors(Hiroshi Inoue)
3363 Fix for COPY FROM STDIN when using \i(Tom)
3364 Fix for subselect is compared inside an expression(Jan)
3365 Fix handling of error reporting while returning rows(Tom)
3366 Fix problems with reference to array types(Tom,Jan)
3367 Prevent UPDATE SET oid(Jan)
3368 Fix pg_dump so -t option can handle case-sensitive tablenames
3369 Fixes for GROUP BY in special cases(Tom, Jan)
3370 Fix for memory leak in failed queries(Tom)
3371 DEFAULT now supports mixed-case identifiers(Tom)
3372 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
3373 Disable use of pg_dump with both -o and -d options(Bruce)
3374 Allow pg_dump to properly dump group privileges(Bruce)
3375 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
3376 Fix for computations in views(Jan)
3377 Fix for aggregates on array indexes(Tom)
3378 Fix for DEFAULT handles single quotes in value requiring too many quotes
3379 Fix security problem with non-super users importing/exporting large objects(Tom
3381 Rollback of transaction that creates table cleaned up properly(Tom)
3382 Fix to allow long table and column names to generate proper serial names(Tom)
3386 Add "vacuumdb" utility
3387 Speed up libpq by allocating memory better(Tom)
3388 EXPLAIN all indexes used(Tom)
3389 Implement CASE, COALESCE, NULLIF expression(Thomas)
3390 New pg_dump table output format(Constantin)
3391 Add string min()/max() functions(Thomas)
3392 Extend new type coercion techniques to aggregates(Thomas)
3393 New moddatetime contrib(Terry)
3394 Update to pgaccess 0.96(Constantin)
3395 Add routines for single-byte "char" type(Thomas)
3396 Improved substr() function(Thomas)
3397 Improved multibyte handling(Tatsuo)
3398 Multiversion concurrency control/MVCC(Vadim)
3399 New Serialized mode(Vadim)
3400 Fix for tables over 2gigs(Peter)
3401 New SET TRANSACTION ISOLATION LEVEL(Vadim)
3402 New LOCK TABLE IN ... MODE(Vadim)
3403 Update ODBC driver(Byron)
3404 New NUMERIC data type(Jan)
3405 New SELECT FOR UPDATE(Vadim)
3406 Handle "NaN" and "Infinity" for input values(Jan)
3407 Improved date/year handling(Thomas)
3408 Improved handling of backend connections(Magnus)
3409 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
3410 New TCL_ARRAYS option(Massimo)
3411 New INTERSECT and EXCEPT(Stefan)
3412 New pg_index.indisprimary for primary key tracking(D'Arcy)
3413 New pg_dump option to allow dropping of tables before creation(Brook)
3414 Speedup of row output routines(Tom)
3415 New READ COMMITTED isolation level(Vadim)
3416 New TEMP tables/indexes(Bruce)
3417 Prevent sorting if result is already sorted(Jan)
3418 New memory allocation optimization(Jan)
3419 Allow psql to do \p\g(Bruce)
3420 Allow multiple rule actions(Jan)
3421 Added LIMIT/OFFSET functionality(Jan)
3422 Improve optimizer when joining a large number of tables(Bruce)
3423 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
3424 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
3425 Improved int8 support(Ryan Bradetich, Thomas, Tom)
3426 New routines to convert between int8 and text/varchar types(Thomas)
3427 New bushy plans, where meta-tables are joined(Bruce)
3428 Enable right-hand queries by default(Bruce)
3429 Allow reliable maximum number of backends to be set at configure time
3430 (--with-maxbackends and postmaster switch (-N backends))(Tom)
3431 GEQO default now 10 tables because of optimizer speedups(Tom)
3432 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
3433 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
3434 Allow psql \d on a view show query(Ryan)
3435 Speedup for LIKE(Bruce)
3436 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
3437 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
3438 Make % operator have precedence like /(Bruce)
3439 Add new postgres -O option to allow system table structure changes(Bruce)
3440 Update contrib/pginterface/findoidjoins script(Tom)
3441 Major speedup in vacuum of deleted rows with indexes(Vadim)
3442 Allow non-SQL functions to run different versions based on arguments(Tom)
3443 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida
3445 Add version number in start-up banners for psql(Masaaki Sakaida)
3446 New contrib/vacuumlo removes large objects not referenced(Peter)
3447 New initialization for table sizes so non-vacuumed tables perform better(Tom)
3448 Improve error messages when a connection is rejected(Tom)
3449 Support for arrays of char() and varchar() fields(Massimo)
3450 Overhaul of hash code to increase reliability and performance(Tom)
3451 Update to PyGreSQL 2.4(D'Arcy)
3452 Changed debug options so -d4 and -d5 produce different node displays(Jan)
3453 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
3454 Better optimization statistics for system table access(Tom)
3455 Better handling of non-default block sizes(Massimo)
3456 Improve GEQO optimizer memory consumption(Tom)
3457 UNION now suppports ORDER BY of columns not in target list(Jan)
3458 Major libpq++ improvements(Vince Vielhaber)
3459 pg_dump now uses -z(ACL's) as default(Bruce)
3460 backend cache, memory speedups(Tom)
3461 have pg_dump do everything in one snapshot transaction(Vadim)
3462 fix for large object memory leakage, fix for pg_dumping(Tom)
3463 INET type now respects netmask for comparisons
3464 Make VACUUM ANALYZE only use a readlock(Vadim)
3465 Allow VIEWs on UNIONS(Jan)
3466 pg_dump now can generate consistent snapshots on active databases(Vadim)
3470 Improve port matching(Tom)
3471 Portability fixes for SunOS
3472 Add Windows NT backend port and enable dynamic loading(Magnus and Daniel Horak)
3473 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
3474 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
3475 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
3476 Port to NetBSD/macppc(Toshimi Aoki)
3477 Fix for tcl/tk configuration(Vince)
3478 Removed CURRENT key word for rule queries(Jan)
3479 NT dynamic loading now works(Daniel Horak)
3480 Add ARM32 support(Andrew McMurry)
3481 Better support for HP-UX 11 and UnixWare
3482 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
3483 New install commands for plpgsql(Jan)
3485 _________________________________________________________________
3489 Release date: 1998-12-20
3491 The 6.4.1 release was improperly packaged. This also has one
3493 _________________________________________________________________
3495 Migration to version 6.4.2
3497 A dump/restore is *not* required for those running 6.4.*.
3498 _________________________________________________________________
3502 Fix for datetime constant problem on some platforms(Thomas)
3503 _________________________________________________________________
3507 Release date: 1998-12-18
3509 This is basically a cleanup release for 6.4. We have fixed a variety
3510 of problems reported by 6.4 users.
3511 _________________________________________________________________
3513 Migration to version 6.4.1
3515 A dump/restore is *not* required for those running 6.4.
3516 _________________________________________________________________
3520 Add pg_dump -N flag to force double quotes around identifiers. This is
3522 Fix for NOT in where clause causing crash(Bruce)
3523 EXPLAIN VERBOSE coredump fix(Vadim)
3524 Fix shared-library problems on Linux
3525 Fix test for table existence to allow mixed-case and whitespace in
3526 the table name(Thomas)
3527 Fix a couple of pg_dump bugs
3528 Configure matches template/.similar entries better(Tom)
3529 Change builtin function names from SPI_* to spi_*
3530 OR WHERE clause fix(Vadim)
3531 Fixes for mixed-case table names(Billy)
3532 contrib/linux/postgres.init.csh/sh fix(Thomas)
3533 libpq memory overrun fix
3535 Change exp() behavior to generate error on underflow(Thomas)
3536 pg_dump fixes for memory leak, inheritance constraints, layout change
3537 update pgaccess to 0.93
3538 Fix prototype for 64-bit platforms
3539 Multibyte fixes(Tatsuo)
3541 Fix memory overruns(Tatsuo)
3542 Fix for lo_import() crash(Bruce)
3543 Better search for install program(Tom)
3546 Use implicit type coercion for matching DEFAULT values(Thomas)
3547 Add routines to help with single-byte (internal) character type(Thomas)
3548 Compilation of libpq for Windows fixes(Magnus)
3549 Upgrade to PyGreSQL 2.2(D'Arcy)
3550 _________________________________________________________________
3554 Release date: 1998-10-30
3556 There are *many* new features and improvements in this release. Thanks
3557 to our developers and maintainers, nearly every aspect of the system
3558 has received some attention since the previous release. Here is a
3559 brief, incomplete summary:
3561 * Views and rules are now functional thanks to extensive new code in
3562 the rewrite rules system from Jan Wieck. He also wrote a chapter
3563 on it for the Programmer's Guide.
3564 * Jan also contributed a second procedural language, PL/pgSQL, to go
3565 with the original PL/pgTCL procedural language he contributed last
3567 * We have optional multiple-byte character set support from Tatsuo
3568 Ishii to complement our existing locale support.
3569 * Client/server communications has been cleaned up, with better
3570 support for asynchronous messages and interrupts thanks to Tom
3572 * The parser will now perform automatic type coercion to match
3573 arguments to available operators and functions, and to match
3574 columns and expressions with target columns. This uses a generic
3575 mechanism which supports the type extensibility features of
3576 PostgreSQL. There is a new chapter in the User's Guide which
3578 * Three new data types have been added. Two types, inet and cidr,
3579 support various forms of IP network, subnet, and machine
3580 addressing. There is now an 8-byte integer type available on some
3581 platforms. See the chapter on data types in the User's Guide for
3582 details. A fourth type, serial, is now supported by the parser as
3583 an amalgam of the int4 type, a sequence, and a unique index.
3584 * Several more SQL92-compatible syntax features have been added,
3585 including "INSERT DEFAULT VALUES"
3586 * The automatic configuration and installation system has received
3587 some attention, and should be more robust for more platforms than
3589 _________________________________________________________________
3591 Migration to version 6.4
3593 A dump/restore using pg_dump or pg_dumpall is required for those
3594 wishing to migrate data from any previous release of PostgreSQL.
3595 _________________________________________________________________
3601 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
3602 Remove char2-16 data types, use char/varchar(Darren)
3603 Pqfn not handles a NOTICE message(Anders)
3604 Reduced busywaiting overhead for spinlocks with many backends (dg)
3605 Stuck spinlock detection (dg)
3606 Fix up "ISO-style" timespan decoding and encoding(Thomas)
3607 Fix problem with table drop after rollback of transaction(Vadim)
3608 Change error message and remove non-functional update message(Vadim)
3609 Fix for COPY array checking
3610 Fix for SELECT 1 UNION SELECT NULL
3611 Fix for buffer leaks in large object calls(Pascal)
3612 Change owner from oid to int4 type(Bruce)
3613 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
3614 Fix for shared invalidation cache overflow(Massimo)
3615 Prevent file descriptor leaks in failed COPY's(Bruce)
3616 Fix memory leak in libpgtcl's pg_select(Constantin)
3617 Fix problems with username/passwords over 8 characters(Tom)
3618 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
3619 Fix of many bad system table entries(Tom)
3623 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
3624 Show the index used in an EXPLAIN(Zeugswetter)
3625 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
3626 Multibyte awareness of many data types and functions, via configure(Tatsuo)
3627 New configure --with-mb option(Tatsuo)
3628 New initdb --pgencoding option(Tatsuo)
3629 New createdb -E multibyte option(Tatsuo)
3630 Select version(); now returns PostgreSQL version(Jeroen)
3631 libpq now allows asynchronous clients(Tom)
3632 Allow cancel from client of backend query(Tom)
3633 psql now cancels query with Control-C(Tom)
3634 libpq users need not issue dummy queries to get NOTIFY messages(Tom)
3635 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
3636 PGresult struct now includes associated error message, if any(Tom)
3637 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
3638 Add routines to convert between varchar and bpchar(Thomas)
3639 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
3640 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
3641 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas
3643 Fixes for unary minus parsing with leading spaces(Thomas)
3644 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
3645 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
3646 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
3647 Enable HAVING clause but no fixes elsewhere yet.
3648 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thoma
3650 Save string type if specified for DEFAULT clause handling(Thomas)
3651 Coerce operations involving different data types(Thomas)
3652 Allow some index use for columns of different types(Thomas)
3653 Add capabilities for automatic type conversion(Thomas)
3654 Cleanups for large objects, so file is truncated on open(Peter)
3655 Readline cleanups(Tom)
3656 Allow psql \f \ to make spaces as delimiter(Bruce)
3657 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
3658 Msql compatibility library in /contrib(Aldrin)
3659 Remove the requirement that ORDER/GROUP BY clause identifiers be
3660 included in the target list(David)
3661 Convert columns to match columns in UNION clauses(Thomas)
3662 Remove fork()/exec() and only do fork()(Bruce)
3663 Jdbc cleanups(Peter)
3664 Show backend status on ps command line(only works on some platforms)(Bruce)
3665 Pg_hba.conf now has a sameuser option in the database field
3666 Make lo_unlink take oid param, not int4
3667 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
3668 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
3669 libpgtcl cleanups(Tom)
3670 Add -error option to libpgtcl's pg_result command(Tom)
3671 New locale patch, see docs/README/locale(Oleg)
3672 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
3673 New contrib/lo code for large object orphan removal(Peter)
3674 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
3675 feature, see /doc/README.mb(Tatsuo)
3676 contrib/noupdate code to revoke update permission on a column
3677 libpq can now be compiled on Windows(Magnus)
3678 Add PQsetdbLogin() in libpq
3679 New 8-byte integer type, checked by configure for OS support(Thomas)
3680 Better support for quoted table/column names(Thomas)
3681 Surround table and column names with double-quotes in pg_dump(Thomas)
3682 PQreset() now works with passwords(Tom)
3683 Handle case of GROUP BY target list column number out of range(David)
3684 Allow UNION in subselects
3685 Add auto-size to screen to \d? commands(Bruce)
3686 Use UNION to show all \d? results in one query(Bruce)
3687 Add \d? field search feature(Bruce)
3688 Pg_dump issues fewer \connect requests(Tom)
3689 Make pg_dump -z flag work better, document it in manual page(Tom)
3690 Add HAVING clause with full support for subselects and unions(Stephan)
3691 Full text indexing routines in contrib/fulltextindex(Maarten)
3692 Transaction ids now stored in shared memory(Vadim)
3693 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
3694 Support for SQL92 syntax "SET NAMES"(Tatsuo)
3695 Support for LATIN2-5(Tatsuo)
3696 Add UNICODE regression test case(Tatsuo)
3697 Lock manager cleanup, new locking modes for LLL(Vadim)
3698 Allow index use with OR clauses(Bruce)
3699 Allows "SELECT NULL ORDER BY 1;"
3700 Explain VERBOSE prints the plan, and now pretty-prints the plan to
3701 the postmaster log file(Bruce)
3702 Add indexes display to \d command(Bruce)
3703 Allow GROUP BY on functions(David)
3704 New pg_class.relkind for large objects(Bruce)
3705 New way to send libpq NOTICE messages to a different location(Tom)
3706 New \w write command to psql(Bruce)
3707 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
3708 Allow binary-compatible indexes to be considered when checking for valid
3709 Indexes for restriction clauses containing a constant(Thomas)
3710 New ISBN/ISSN code in /contrib/isbn_issn
3711 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
3712 New rewrite system fixes many problems with rules and views(Jan)
3713 * Rules on relations work
3714 * Event qualifications on insert/update/delete work
3715 * New OLD variable to reference CURRENT, CURRENT will be remove in futu
3717 * Update rules can reference NEW and OLD in rule qualifications/actions
3718 * Insert/update/delete rules on views work
3719 * Multiple rule actions are now supported, surrounded by parentheses
3720 * Regular users can create views/rules on tables they have RULE permits
3721 * Rules and views inherit the privileges of the creator
3722 * No rules at the column level
3723 * No UPDATE NEW/OLD rules
3724 * New pg_tables, pg_indexes, pg_rules and pg_views system views
3725 * Only a single action on SELECT rules
3726 * Total rewrite overhaul, perhaps for 6.5
3728 * handle aggregates on views
3729 * handle insert into select from view works
3730 System indexes are now multikey(Bruce)
3731 Oidint2, oidint4, and oidname types are removed(Bruce)
3732 Use system cache for more system table lookups(Bruce)
3733 New backend programming language PL/pgSQL in backend/pl(Jan)
3734 New SERIAL data type, auto-creates sequence/index(Thomas)
3735 Enable assert checking without a recompile(Massimo)
3736 User lock enhancements(Massimo)
3737 New setval() command to set sequence value(Massimo)
3738 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
3739 Conditional trace package(Massimo)
3740 New UNLISTEN command(Massimo)
3741 psql and libpq now compile under Windows using win32.mak(Magnus)
3742 Lo_read no longer stores trailing NULL(Bruce)
3743 Identifiers are now truncated to 31 characters internally(Bruce)
3744 Createuser options now availble on the command line
3745 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
3746 Prevent file descriptor leaf from failed COPY(Bruce)
3747 New pg_upgrade command(Bruce)
3748 Updated /contrib directories(Massimo)
3749 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
3750 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
3751 New DECLARE and FETCH feature(Thomas)
3752 libpq's internal structures now not exported(Tom)
3753 Allow up to 8 key indexes(Bruce)
3754 Remove ARCHIVE key word, that is no longer used(Thomas)
3755 pg_dump -n flag to supress quotes around indentifiers
3756 disable system columns for views(Jan)
3757 new INET and CIDR types for network addresses(TomH, Paul)
3758 no more double quotes in psql output
3759 pg_dump now dumps views(Terry)
3760 new SET QUERY_LIMIT(Tatsuo,Jan)
3764 /contrib cleanup(Jun)
3765 Inline some small functions called for every row(Bruce)
3768 Multibyte regression tests(Soonmyung.)
3769 Remove --disabled options from configure
3770 Define PGDOC to use POSTGRESDIR by default
3771 Make regression optional
3772 Remove extra braces code to pgindent(Bruce)
3773 Add bsdi shared library support(Bruce)
3774 New --without-CXX support configure option(Brook)
3776 Update backend flowchart in tools/backend(Bruce)
3777 Change atttypmod from int16 to int32(Bruce, Tom)
3778 Getrusage() fix for platforms that do not have it(Tom)
3779 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
3780 NS32K platform fixes(Phil Nelson, John Buller)
3781 SCO 7/UnixWare 2.x fixes(Billy,others)
3782 Sparc/Solaris 2.5 fixes(Ryan)
3783 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
3784 Even more documention(Thomas)
3785 Nextstep support(Jacek)
3787 pginterface manual page(Bruce)
3788 shared libraries all have version numbers
3789 merged all OS-specific shared library defines into one file
3790 smarter TCL/TK configuration checking(Billy)
3791 smarter perl configuration(Brook)
3792 configure uses supplied install-sh if no install script found(Tom)
3793 new Makefile.shlib for shared library configuration(Tom)
3794 _________________________________________________________________
3798 Release date: 1998-04-07
3800 This is a bug-fix release for 6.3.x. Refer to the release notes for
3801 version 6.3 for a more complete summary of new features.
3805 * Repairs automatic configuration support for some platforms,
3806 including Linux, from breakage inadvertently introduced in version
3808 * Correctly handles function calls on the left side of BETWEEN and
3811 A dump/restore is NOT required for those running 6.3 or 6.3.1. A make
3812 distclean, make, and make install is all that is required. This last
3813 step should be performed while the postmaster is not running. You
3814 should re-link any custom applications that use PostgreSQL libraries.
3816 For upgrades from pre-6.3 installations, refer to the installation and
3817 migration instructions for version 6.3.
3818 _________________________________________________________________
3822 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
3823 Manual page improvements(Bruce)
3824 BETWEEN and LIKE fix(Thomas)
3825 fix for psql \connect used by pg_dump(Oliver Elphick)
3827 pgaccess, version 0.86
3828 qsort removed, now uses libc version, cleanups(Jeroen)
3829 fix for buffer over-runs detected(Maurice Gittens)
3830 fix for buffer overrun in libpgtcl(Randy Kunkee)
3831 fix for UNION with DISTINCT or ORDER BY(Bruce)
3832 gettimeofday configure check(Doug Winterburn)
3833 Fix "indexes not used" bug(Vadim)
3834 docs additions(Thomas)
3835 Fix for backend memory leak(Bruce)
3836 libreadline cleanup(Erwan MAS)
3837 Remove DISTDIR(Bruce)
3838 Makefile dependency cleanup(Jeroen van Vianen)
3841 _________________________________________________________________
3845 Release date: 1998-03-23
3849 * Additional support for multibyte character sets.
3850 * Repair byte ordering for mixed-endian clients and servers.
3851 * Minor updates to allowed SQL syntax.
3852 * Improvements to the configuration autodetection for installation.
3854 A dump/restore is NOT required for those running 6.3. A make
3855 distclean, make, and make install is all that is required. This last
3856 step should be performed while the postmaster is not running. You
3857 should re-link any custom applications that use PostgreSQL libraries.
3859 For upgrades from pre-6.3 installations, refer to the installation and
3860 migration instructions for version 6.3.
3861 _________________________________________________________________
3865 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
3866 pg_user cleanup(Bruce)
3867 large object fix for pg_dump and tclsh (alvin)
3868 LIKE fix for multiple adjacent underscores
3869 fix for redefining builtin functions(Thomas)
3871 upgrade to pg_access 0.83
3872 updated CLUSTER manual page
3873 multibyte character set support, see doc/README.mb(Tatsuo)
3874 configure --with-pgport fix
3876 big-endian fix for backend communications(Kataoka)
3877 SUBSTR() and substring() fix(Jan)
3878 several jdbc fixes(Peter)
3879 libpgtcl improvements, see libptcl/README(Randy Kunkee)
3880 Fix for "Datasize = 0" error(Vadim)
3881 Prevent \do from wrapping(Bruce)
3882 Remove duplicate Russian character set entries
3884 Allow optional TABLE key word in LOCK and SELECT INTO(Thomas)
3885 CREATE SEQUENCE options to allow a negative integer(Thomas)
3886 Add "PASSWORD" as an allowed column identifier(Thomas)
3887 Add checks for UNION target fields(Bruce)
3888 Fix Alpha port(Dwayne Bailey)
3889 Fix for text arrays containing quotes(Doug Gibson)
3890 Solaris compile fix(Albert Chin-A-Young)
3891 Better identify tcl and tk libs and includes(Bruce)
3893 _________________________________________________________________
3897 Release date: 1998-03-01
3899 There are *many* new features and improvements in this release. Here
3900 is a brief, incomplete summary:
3902 * Many new SQL features, including full SQL92 subselect capability
3903 (everything is here but target-list subselects).
3904 * Support for client-side environment variables to specify time zone
3906 * Socket interface for client/server connection. This is the default
3907 now so you may need to start postmaster with the "-i" flag.
3908 * Better password authorization mechanisms. Default table privileges
3910 * Old-style time travel has been removed. Performance has been
3913 Note: Bruce Momjian wrote the following notes to introduce the new
3916 There are some general 6.3 issues that I want to mention. These are
3917 only the big items that can not be described in one sentence. A review
3918 of the detailed changes list is still needed.
3920 First, we now have subselects. Now that we have them, I would like to
3921 mention that without subselects, SQL is a very limited language.
3922 Subselects are a major feature, and you should review your code for
3923 places where subselects provide a better solution for your queries. I
3924 think you will find that there are more uses for subselects than you
3925 may think. Vadim has put us on the big SQL map with subselects, and
3926 fully functional ones too. The only thing you can't do with subselects
3927 is to use them in the target list.
3929 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
3930 enable connections from other machines, you have to use the new
3931 postmaster -i option, and of course edit "pg_hba.conf". Also, for this
3932 reason, the format of "pg_hba.conf" has changed.
3934 Third, char() fields will now allow faster access than varchar() or
3935 text. Specifically, the text and varchar() have a penalty for access
3936 to any columns after the first column of this type. char() used to
3937 also have this access penalty, but it no longer does. This may suggest
3938 that you redesign some of your tables, especially if you have short
3939 character columns that you have defined as varchar() or text. This and
3940 other changes make 6.3 even faster than earlier releases.
3942 We now have passwords definable independent of any Unix file. There
3943 are new SQL USER commands. See the Administrator's Guide for more
3944 information. There is a new table, pg_shadow, which is used to store
3945 user information and user passwords, and it by default only
3946 SELECT-able by the postgres super-user. pg_user is now a view of
3947 pg_shadow, and is SELECT-able by PUBLIC. You should keep using pg_user
3948 in your application without changes.
3950 User-created tables now no longer have SELECT privilege to PUBLIC by
3951 default. This was done because the ANSI standard requires it. You can
3952 of course GRANT any privileges you want after the table is created.
3953 System tables continue to be SELECT-able by PUBLIC.
3955 We also have real deadlock detection code. No more sixty-second
3956 timeouts. And the new locking code implements a FIFO better, so there
3957 should be less resource starvation during heavy use.
3959 Many complaints have been made about inadequate documentation in
3960 previous releases. Thomas has put much effort into many new manuals
3961 for this release. Check out the doc/ directory.
3963 For performance reasons, time travel is gone, but can be implemented
3964 using triggers (see "pgsql/contrib/spi/README"). Please check out the
3965 new \d command for types, operators, etc. Also, views have their own
3966 privileges now, not based on the underlying tables, so privileges on
3967 them have to be set separately. Check "/pgsql/interfaces" for some new
3968 ways to talk to PostgreSQL.
3970 This is the first release that really required an explanation for
3971 existing users. In many ways, this was necessary because the new
3972 release removes many limitations, and the work-arounds people were
3973 using are no longer needed.
3974 _________________________________________________________________
3976 Migration to version 6.3
3978 A dump/restore using pg_dump or pg_dumpall is required for those
3979 wishing to migrate data from any previous release of PostgreSQL.
3980 _________________________________________________________________
3986 Fix binary cursors broken by MOVE implementation(Vadim)
3987 Fix for tcl library crash(Jan)
3988 Fix for array handling, from Gerhard Hintermayer
3989 Fix acl error, and remove duplicate pqtrace(Bruce)
3990 Fix psql \e for empty file(Bruce)
3991 Fix for textcat on varchar() fields(Bruce)
3992 Fix for DBT Sendproc (Zeugswetter Andres)
3993 Fix vacuum analyze syntax problem(Bruce)
3994 Fix for international identifiers(Tatsuo)
3995 Fix aggregates on inherited tables(Bruce)
3996 Fix substr() for out-of-bounds data
3997 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
3998 Fix notty output to show status result. -q option still turns it off(Bruce)
3999 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
4001 Fix for PQtrace start/stop several times(Bruce)
4002 Fix a variety of locking problems like newer lock waiters getting
4003 lock before older waiters, and having readlock people not share
4004 locks if a writer is waiting for a lock, and waiting writers not
4005 getting priority over waiting readers(Bruce)
4006 Fix crashes in psql when executing queries from external files(James)
4007 Fix problem with multiple order by columns, with the first one having
4009 Use correct hash table support functions for float8 and int4(Thomas)
4010 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
4011 Change precedence for boolean operators to match expected behavior(Thomas)
4012 Generate elog(ERROR) on over-large integer(Bruce)
4013 Allow multiple-argument functions in constraint clauses(Thomas)
4014 Check boolean input literals for 'true','false','yes','no','1','0'
4015 and throw elog(ERROR) if unrecognized(Thomas)
4016 Major large objects fix
4017 Fix for GROUP BY showing duplicates(Vadim)
4018 Fix for index scans in MergeJion(Vadim)
4022 Subselects with EXISTS, IN, ALL, ANY key words (Vadim, Bruce, Thomas)
4023 New User Manual(Thomas, others)
4024 Speedup by inlining some frequently-called functions
4025 Real deadlock detection, no more timeouts(Bruce)
4026 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
4027 CURRENT_USER(Thomas)
4028 Modify constraint syntax to be SQL92-compliant(Thomas)
4029 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
4030 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
4031 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas
4033 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
4034 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
4035 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
4036 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
4037 Allow SQL92 delimited identifiers(Thomas)
4038 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas
4040 Support SQL92 syntax for type coercion of literal strings
4041 (e.g. "DATETIME 'now'")(Thomas)
4042 Add conversions for int2, int4, and OID types to and from text(Thomas)
4043 Use shared lock when building indexes(Vadim)
4044 Free memory allocated for an user query inside transaction block after
4045 this query is done, was turned off in <= 6.2.1(Vadim)
4046 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
4047 New PostgreSQL Procedural Language (PL) backend interface(Jan)
4048 Rename pg_dump -H option to -h(Bruce)
4049 Add Java support for passwords, European dates(Peter)
4050 Use indexes for LIKE and ~, !~ operations(Bruce)
4051 Add hash functions for datetime and timespan(Thomas)
4052 Time Travel removed(Vadim, Bruce)
4053 Add paging for \d and \z, and fix \i(Bruce)
4054 Add Unix domain socket support to backend and to frontend library(Goran)
4055 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
4056 Allow more SQL92 and/or PostgreSQL reserved words as column identifiers(Thomas)
4057 Augment support for SQL92 SET TIME ZONE...(Thomas)
4058 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
4059 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
4060 Enable SET TIME ZONE using TZ environment variable(Thomas)
4061 Add PGDATESTYLE environment variable to frontend and backend initialization(Tho
4063 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
4064 frontend library initialization environment variables(Thomas)
4065 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
4066 Add pg_description table for info on tables, columns, operators, types, and
4068 Increase 16 char limit on system table/index names to 32 characters(Bruce)
4069 Rename system indexes(Bruce)
4070 Add 'GERMAN' option to SET DATESTYLE(Thomas)
4071 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
4072 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
4073 Validate numeric input more carefully for delta times(Thomas)
4074 Implement day of year as possible input to date_part()(Thomas)
4075 Define timespan_finite() and text_timespan() functions(Thomas)
4076 Remove archive stuff(Bruce)
4077 Allow for a pg_password authentication database that is separate from
4078 the system password file(Todd)
4079 Dump ACLs, GRANT, REVOKE privileges(Matt)
4080 Define text, varchar, and bpchar string length functions(Thomas)
4081 Fix Query handling for inheritance, and cost computations(Bruce)
4082 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
4083 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
4084 Implement UNIONs for SELECT(Bruce)
4085 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
4086 varchar() stores only necessary bytes on disk(Bruce)
4087 Fix for BLOBs(Peter)
4088 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
4089 Remove unused "option" from PQconnectdb()
4090 New LOCK command and lock manual page describing deadlocks(Bruce)
4091 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
4092 Enhance psql \z to show sequences(Bruce)
4093 Show NOT NULL and DEFAULT in psql \d table(Bruce)
4094 New psql .psqlrc file start-up(Andrew)
4095 Modify sample start-up script in contrib/linux to show syslog(Thomas)
4096 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
4097 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
4098 Update of contrib stuff(Massimo)
4099 Add Unix socket support to DBD::Pg(Goran)
4100 New python interface (PyGreSQL 2.0)(D'Arcy)
4101 New frontend/backend protocol has a version number, network byte order(Phil)
4102 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
4103 CHAR() now faster access than VARCHAR() or TEXT
4104 ecpg embedded SQL preprocessor
4105 Reduce system column overhead(Vadmin)
4106 Remove pg_time table(Vadim)
4107 Add pg_type attribute to identify types that need length (bpchar, varchar)
4108 Add report of offending line when COPY command fails
4109 Allow VIEW privileges to be set separately from the underlying tables.
4110 For security, use GRANT/REVOKE on views as appropriate(Jan)
4111 Tables now have no default GRANT SELECT TO PUBLIC. You must
4112 explicitly grant such privileges.
4113 Clean up tutorial examples(Darren)
4117 Add new html development tools, and flow chart in /tools/backend
4118 Fix for SCO compiles
4119 Stratus computer port Robert Gillies
4120 Added support for shlib for BSD44_derived & i386_solaris
4121 Make configure more automated(Brook)
4122 Add script to check regression test results
4123 Break parser functions into smaller files, group together(Bruce)
4124 Rename heap_create to heap_create_and_catalog, rename heap_creatr
4125 to heap_create()(Bruce)
4126 Sparc/Linux patch for locking(TomS)
4127 Remove PORTNAME and reorganize port-specific stuff(Marc)
4128 Add optimizer README file(Bruce)
4129 Remove some recursion in optimizer and clean up some code there(Bruce)
4130 Fix for NetBSD locking(Henry)
4131 Fix for libptcl make(Tatsuo)
4133 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
4134 function calls to istrue() or isfalse() to allow optimization(Thomas)
4135 Various fixes NetBSD/Sparc related(TomH)
4136 Alpha linux locking(Travis,Ryan)
4137 Change elog(WARN) to elog(ERROR)(Bruce)
4138 FAQ for FreeBSD(Marc)
4139 Bring in the PostODBC source tree as part of our standard distribution(Marc)
4140 A minor patch for HP/UX 10 vs 9(Stan)
4141 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
4142 UnixWare patches(Billy)
4143 New i386 'lock' for spinlock asm(Billy)
4144 Support for multiplexed backends is removed
4145 Start an OpenBSD port
4148 Add string functions to regression suite(Thomas)
4149 Expand a few function names formerly truncated to 16 characters(Thomas)
4150 Remove un-needed malloc() calls and replace with palloc()(Bruce)
4151 _________________________________________________________________
4155 Release date: 1997-10-17
4157 6.2.1 is a bug-fix and usability release on 6.2.
4161 * Allow strings to span lines, per SQL92.
4162 * Include example trigger function for inserting user names on table
4165 This is a minor bug-fix release on 6.2. For upgrades from pre-6.2
4166 systems, a full dump/reload is required. Refer to the 6.2 release
4167 notes for instructions.
4168 _________________________________________________________________
4170 Migration from version 6.2 to version 6.2.1
4172 This is a minor bug-fix release. A dump/reload is not required from
4173 version 6.2, but is required from any release prior to 6.2.
4175 In upgrading from version 6.2, if you choose to dump/reload you will
4176 find that avg(money) is now calculated correctly. All other bug fixes
4177 take effect upon updating the executables.
4179 Another way to avoid dump/reload is to use the following SQL command
4180 from "psql" to update the existing system table:
4181 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
4182 where aggname = 'avg' and aggbasetype = 790;
4184 This will need to be done to every existing database, including
4186 _________________________________________________________________
4190 Allow TIME and TYPE column names(Thomas)
4191 Allow larger range of true/false as boolean values(Thomas)
4192 Support output of "now" and "current"(Thomas)
4193 Handle DEFAULT with INSERT of NULL properly(Vadim)
4194 Fix for relation reference counts problem in buffer manager(Vadim)
4195 Allow strings to span lines, like ANSI(Thomas)
4196 Fix for backward cursor with ORDER BY(Vadim)
4197 Fix avg(cash) computation(Thomas)
4198 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
4199 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
4200 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
4202 _________________________________________________________________
4206 Release date: 1997-10-02
4208 A dump/restore is required for those wishing to migrate data from
4209 previous releases of PostgreSQL.
4210 _________________________________________________________________
4212 Migration from version 6.1 to version 6.2
4214 This migration requires a complete dump of the 6.1 database and a
4215 restore of the database in 6.2.
4217 Note that the "pg_dump" and "pg_dumpall" utility from 6.2 should be
4218 used to dump the 6.1 database.
4219 _________________________________________________________________
4221 Migration from version 1.x to version 6.2
4223 Those migrating from earlier 1.* releases should first upgrade to 1.09
4224 because the COPY output format was improved from the 1.02 release.
4225 _________________________________________________________________
4231 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
4232 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
4233 from Solaris(Diab Jerius)
4234 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
4235 Check for geometric intersections at endpoints to avoid rounding ugliness(Thoma
4237 Catch non-functional delete attempts(Vadim)
4238 Change time function names to be more consistent(Michael Reifenberg)
4239 Check for zero divides(Michael Reifenberg)
4240 Fix very old bug which made rows changed/inserted by a command
4241 visible to the command itself (so we had multiple update of
4242 updated rows, etc.)(Vadim)
4243 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
4244 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
4245 Remove un-needed signal stuff from contrib/pginterface
4246 Fix OR (where x != 1 or x isnull didn't return rows with x NULL) (Vadim)
4247 Fix time_cmp function (Vadim)
4248 Fix handling of functions with non-attribute first argument in
4249 WHERE clauses (Vadim)
4250 Fix GROUP BY when order of entries is different from order
4251 in target list (Vadim)
4252 Fix pg_dump for aggregates without sfunc1 (Vadim)
4256 Default genetic optimizer GEQO parameter is now 8(Bruce)
4257 Allow use parameters in target list having aggregates in functions(Vadim)
4258 Added JDBC driver as an interface(Adrian & Peter)
4260 Return number of rows inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
4261 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
4262 SPI (Server Programming Interface) allows execution of queries inside
4264 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
4265 Include reserved words for string handling, outer joins, and unions(Thomas)
4266 Implement extended comments ("/* ... */") using exclusive states(Thomas)
4267 Add "//" single-line comments(Bruce)
4268 Remove some restrictions on characters in operator names(Thomas)
4269 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
4270 Add text concatenation operator and function (SQL92)(Thomas)
4271 Support WITH TIME ZONE syntax (SQL92)(Thomas)
4272 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
4273 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
4274 and CHARACTER VARYING (SQL92)(Thomas)
4275 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
4276 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
4277 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
4278 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
4279 Add more reserved words, mostly for SQL92 compliance(Thomas)
4280 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
4281 Add center() routines for lseg, path, polygon(Thomas)
4282 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
4283 Check explicitly for points and polygons contained within polygons
4284 using an axis-crossing algorithm(Thomas)
4285 Add routine to convert circle-box(Thomas)
4286 Merge conflicting operators for different geometric data types(Thomas)
4287 Replace distance operator "<===>" with "<->"(Thomas)
4288 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Tho
4290 Add routines for text trimming on both ends, substring, and string position(Tho
4292 Added conversion routines circle(box) and poly(circle)(Thomas)
4293 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
4294 Allow functions and operators on internally-identical types to succeed(Bruce)
4295 Speed up backend start-up after profiling analysis(Bruce)
4296 Inline frequently called functions for performance(Bruce)
4297 Reduce open() calls(Bruce)
4298 psql: Add PAGER for \h and \?,\C fix
4299 Fix for psql pager when no tty(Bruce)
4300 New entab utility(Bruce)
4301 General trigger functions for referential integrity (Vadim)
4302 General trigger functions for time travel (Vadim)
4303 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
4304 MOVE implementation (Vadim)
4308 HP-UX 10 patches (Vladimir Turin)
4309 Added SCO support, (Daniel Harris)
4310 MkLinux patches (Tatsuo Ishii)
4311 Change geometric box terminology from "length" to "width"(Thomas)
4312 Deprecate temporary unstored slope fields in geometric code(Thomas)
4313 Remove restart instructions from INSTALL(Bruce)
4314 Look in /usr/ucb first for install(Bruce)
4315 Fix c++ copy example code(Thomas)
4316 Add -o to psql manual page(Bruce)
4317 Prevent relname unallocated string length from being copied into database(Bruce
4319 Cleanup for NAMEDATALEN use(Bruce)
4320 Fix pg_proc names over 15 chars in output(Bruce)
4321 Add strNcpy() function(Bruce)
4322 remove some (void) casts that are unnecessary(Bruce)
4323 new interfaces directory(Marc)
4324 Replace fopen() calls with calls to fd.c functions(Bruce)
4325 Make functions static where possible(Bruce)
4326 enclose unused functions in #ifdef NOT_USED(Bruce)
4327 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
4328 Changes for Digital Unix
4329 Portability fix for pg_dumpall(Bruce)
4330 Rename pg_attribute.attnvals to attdispersion(Bruce)
4331 "intro/unix" manual page now "pgintro"(Bruce)
4332 "built-in" manual page now "pgbuiltin"(Bruce)
4333 "drop" manual page now "drop_table"(Bruce)
4334 Add "create_trigger", "drop_trigger" manual pages(Thomas)
4335 Add constraints regression test(Vadim & Thomas)
4336 Add comments syntax regression test(Thomas)
4337 Add PGINDENT and support program(Bruce)
4338 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
4339 Files moved to /src/tools directory(Bruce)
4340 SPI and Trigger programming guides (Vadim & D'Arcy)
4341 _________________________________________________________________
4345 Release date: 1997-07-22
4346 _________________________________________________________________
4348 Migration from version 6.1 to version 6.1.1
4350 This is a minor bug-fix release. A dump/reload is not required from
4351 version 6.1, but is required from any release prior to 6.1. Refer to
4352 the release notes for 6.1 for more details.
4353 _________________________________________________________________
4357 fix for SET with options (Thomas)
4358 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
4359 new psql \connect option allows changing usernames without changing databases
4360 fix for initdb --debug option(Yoshihiko Ichikawa))
4361 lextest cleanup(Bruce)
4363 fix date/time month boundary arithmetic(Thomas)
4364 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
4365 timestamp overhauled to use standard functions(Thomas)
4366 other code cleanup in date/time routines(Thomas)
4367 psql's \d now case-insensitive(Bruce)
4368 psql's backslash commands can now have trailing semicolon(Bruce)
4369 fix memory leak in psql when using \g(Bruce)
4370 major fix for endian handling of communication to server(Thomas, Tatsuo)
4371 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
4372 allow underscores in usernames(Bruce)
4373 pg_dumpall now returns proper status, portability fix(Bruce)
4375 _________________________________________________________________
4379 Release date: 1997-06-08
4381 The regression tests have been adapted and extensively modified for
4382 the 6.1 release of PostgreSQL.
4384 Three new data types (datetime, timespan, and circle) have been added
4385 to the native set of PostgreSQL types. Points, boxes, paths, and
4386 polygons have had their output formats made consistent across the data
4387 types. The polygon output in misc.out has only been spot-checked for
4388 correctness relative to the original regression output.
4390 PostgreSQL 6.1 introduces a new, alternate optimizer which uses
4391 genetic algorithms. These algorithms introduce a random behavior in
4392 the ordering of query results when the query contains multiple
4393 qualifiers or multiple tables (giving the optimizer a choice on order
4394 of evaluation). Several regression tests have been modified to
4395 explicitly order the results, and hence are insensitive to optimizer
4396 choices. A few regression tests are for data types which are
4397 inherently unordered (e.g. points and time intervals) and tests
4398 involving those types are explicitly bracketed with "set geqo to
4399 'off'" and "reset geqo".
4401 The interpretation of array specifiers (the curly braces around atomic
4402 values) appears to have changed sometime after the original regression
4403 tests were generated. The current "./expected/*.out" files reflect
4404 this new interpretation, which may not be correct!
4406 The float8 regression test fails on at least some platforms. This is
4407 due to differences in implementations of pow() and exp() and the
4408 signaling mechanisms used for overflow and underflow conditions.
4410 The "random" results in the random test should cause the "random" test
4411 to be "failed", since the regression tests are evaluated using a
4412 simple diff. However, "random" does not seem to produce random results
4413 on my test machine (Linux/gcc/i686).
4414 _________________________________________________________________
4416 Migration to version 6.1
4418 This migration requires a complete dump of the 6.0 database and a
4419 restore of the database in 6.1.
4421 Those migrating from earlier 1.* releases should first upgrade to 1.09
4422 because the COPY output format was improved from the 1.02 release.
4423 _________________________________________________________________
4429 packet length checking in library routines
4430 lock manager priority patch
4431 check for under/over flow of float8(Bruce)
4432 multitable join fix(Vadim)
4433 SIGPIPE crash fix(Darren)
4434 large object fixes(Sven)
4435 allow btree indexes to handle NULLs(Vadim)
4436 timezone fixes(D'Arcy)
4437 select SUM(x) can return NULL on no rows(Thomas)
4438 internal optimizer, executor bug fixes(Vadim)
4439 fix problem where inner loop in < or <= has no rows(Vadim)
4440 prevent re-commuting join index clauses(Vadim)
4441 fix join clauses for multiple tables(Vadim)
4442 fix hash, hashjoin for arrays(Vadim)
4443 fix btree for abstime type(Vadim)
4444 large object fixes(Raymond)
4445 fix buffer leak in hash indexes (Vadim)
4446 fix rtree for use in inner scan (Vadim)
4447 fix gist for use in inner scan, cleanups (Vadim, Andrea)
4448 avoid unnecessary local buffers allocation (Vadim, Massimo)
4449 fix local buffers leak in transaction aborts (Vadim)
4450 fix file manager memmory leaks, cleanups (Vadim, Massimo)
4451 fix storage manager memmory leaks (Vadim)
4452 fix btree duplicates handling (Vadim)
4453 fix deleted rows reincarnation caused by vacuum (Vadim)
4454 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
4455 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
4459 attribute optimization statistics(Bruce)
4460 much faster new btree bulk load code(Paul)
4461 BTREE UNIQUE added to bulk load code(Vadim)
4462 new lock debug code(Massimo)
4463 massive changes to libpg++(Leo)
4464 new GEQO optimizer speeds table multitable optimization(Martin)
4465 new WARN message for non-unique insert into unique key(Marc)
4466 update x=-3, no spaces, now valid(Bruce)
4467 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
4468 debug backend now pretty-prints tree(Darren)
4469 new Oracle character functions(Edmund)
4470 new plaintext password functions(Dan)
4471 no such class or insufficient privilege changed to distinct messages(Dan)
4472 new ANSI timestamp function(Dan)
4473 new ANSI Time and Date types (Thomas)
4474 move large chunks of data in backend(Martin)
4475 multicolumn btree indexes(Vadim)
4476 new SET var TO value command(Martin)
4477 update transaction status on reads(Dan)
4478 new locale settings for character types(Oleg)
4479 new SEQUENCE serial number generator(Vadim)
4480 GROUP BY function now possible(Vadim)
4481 re-organize regression test(Thomas,Marc)
4482 new optimizer operation weights(Vadim)
4483 new psql \z grant/permit option(Marc)
4484 new MONEY data type(D'Arcy,Thomas)
4485 tcp socket communication speed improved(Vadim)
4486 new VACUUM option for attribute statistics, and for certain columns (Vadim)
4487 many geometric type improvements(Thomas,Keith)
4488 additional regression tests(Thomas)
4489 new datestyle variable(Thomas,Vadim,Martin)
4490 more comparison operators for sorting types(Thomas)
4491 new conversion functions(Thomas)
4492 new more compact btree format(Vadim)
4493 allow pg_dumpall to preserve database ownership(Bruce)
4494 new SET GEQO=# and R_PLANS variable(Vadim)
4495 old (!GEQO) optimizer can use right-sided plans (Vadim)
4496 typechecking improvement in SQL parser(Bruce)
4497 new SET, SHOW, RESET commands(Thomas,Vadim)
4498 new \connect database USER option
4499 new destroydb -i option (Igor)
4500 new \dt and \di psql commands (Darren)
4501 SELECT "\n" now escapes newline (A. Duursma)
4502 new geometry conversion functions from old format (Thomas)
4506 new configuration script(Marc)
4507 readline configuration option added(Marc)
4508 OS-specific configuration options removed(Marc)
4509 new OS-specific template files(Marc)
4510 no more need to edit Makefile.global(Marc)
4511 re-arrange include files(Marc)
4512 nextstep patches (Gregor Hoffleit)
4513 removed Windows-specific code(Bruce)
4514 removed postmaster -e option, now only postgres -e option (Bruce)
4515 merge duplicate library code in front/backends(Martin)
4516 now works with eBones, international Kerberos(Jun)
4517 more shared library support
4518 c++ include file cleanup(Bruce)
4519 warn about buggy flex(Bruce)
4520 DG/UX, Ultrix, IRIX, AIX portability fixes
4521 _________________________________________________________________
4525 Release date: 1997-01-29
4527 A dump/restore is required for those wishing to migrate data from
4528 previous releases of PostgreSQL.
4529 _________________________________________________________________
4531 Migration from version 1.09 to version 6.0
4533 This migration requires a complete dump of the 1.09 database and a
4534 restore of the database in 6.0.
4535 _________________________________________________________________
4537 Migration from pre-1.09 to version 6.0
4539 Those migrating from earlier 1.* releases should first upgrade to 1.09
4540 because the COPY output format was improved from the 1.02 release.
4541 _________________________________________________________________
4547 ALTER TABLE bug - running postgress process needs to re-read table definition
4548 Allow vacuum to be run on one table or entire database(Bruce)
4550 Fix array over-runs of memory writes(Kurt)
4551 Fix elusive btree range/non-range bug(Dan)
4552 Fix for hash indexes on some types like time and date
4553 Fix for pg_log size explosion
4554 Fix permissions on lo_export()(Bruce)
4555 Fix unitialized reads of memory(Kurt)
4556 Fixed ALTER TABLE ... char(3) bug(Bruce)
4557 Fixed a few small memory leaks
4558 Fixed EXPLAIN handling of options and changed full_path option name
4559 Fixed output of group acl privileges
4560 Memory leaks (hunt and destroy with tools like Purify(Kurt)
4561 Minor improvements to rules system
4563 New asserts for run-checking
4564 Overhauled parser/analyze code to properly report errors and increase speed
4565 Pg_dump -d now handles NULL's properly(Bruce)
4566 Prevent SELECT NULL from crashing server (Bruce)
4567 Properly report errors when INSERT ... SELECT columns did not match
4568 Properly report errors when insert column names were not correct
4569 psql \g filename now works(Bruce)
4570 psql fixed problem with multiple statements on one line with multiple outputs
4571 Removed duplicate system OIDs
4572 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
4573 Several fixes for queries that crashed the backend
4574 Starting quote in insert string errors(Bruce)
4575 Submitting an empty query now returns empty status, not just " " query(Bruce)
4579 Add EXPLAIN manual page(Bruce)
4580 Add UNIQUE index capability(Dan)
4581 Add hostname/user level access control rather than just hostname and user
4582 Add synonym of != for <>(Bruce)
4583 Allow "select oid,* from table"
4584 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Br
4586 Allow COPY from the frontend(Bryan)
4587 Allow GROUP BY to use alias column name(Bruce)
4588 Allow actual compression, not just reuse on the same page(Vadim)
4589 Allow installation-configuration option to auto-add all local users(Bryan)
4590 Allow libpq to distinguish between text value '' and null(Bruce)
4591 Allow non-postgres users with createdb privs to destroydb's
4592 Allow restriction on who can create C functions(Bryan)
4593 Allow restriction on who can do backend COPY(Bryan)
4594 Can shrink tables, pg_time and pg_log(Vadim & Erich)
4595 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
4596 Change default decimal constant representation from float4 to float8(Bruce)
4597 European date format now set when postmaster is started
4598 Execute lowercase function names if not found with exact case
4599 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z
4601 Gist now included in the distrubution(Marc)
4602 Idend authentication of local users(Bryan)
4603 Implement BETWEEN qualifier(Bruce)
4604 Implement IN qualifier(Bruce)
4605 libpq has PQgetisnull()(Bruce)
4606 libpq++ improvements
4607 New options to initdb(Bryan)
4608 Pg_dump allow dump of OIDs(Bruce)
4609 Pg_dump create indexes after tables are loaded for speed(Bruce)
4610 Pg_dumpall dumps all databases, and the user table
4611 Pginterface additions for NULL values(Bruce)
4612 Prevent postmaster from being run as root
4613 psql \h and \? is now readable(Bruce)
4614 psql allow backslashed, semicolons anywhere on the line(Bruce)
4615 psql changed command prompt for lines in query or in quotes(Bruce)
4616 psql char(3) now displays as (bp)char in \d output(Bruce)
4617 psql return code now more accurate(Bryan?)
4618 psql updated help syntax(Bruce)
4619 Re-visit and fix vacuum(Vadim)
4620 Reduce size of regression diffs, remove timezone name difference(Bruce)
4621 Remove compile-time parameters to enable binary distributions(Bryan)
4622 Reverse meaning of HBA masks(Bryan)
4623 Secure Authentication of local users(Bryan)
4624 Speed up vacuum(Vadim)
4625 Vacuum now had VERBOSE option(Bruce)
4629 All functions now have prototypes that are compared against the calls
4630 Allow asserts to be disabled easly from Makefile.global(Bruce)
4631 Change oid constants used in code to #define names
4632 Decoupled sparc and solaris defines(Kurt)
4633 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
4634 Major include file reorganization/reduction(Marc)
4635 Make now stops on compile failure(Bryan)
4636 Makefile restructuring(Bryan, Marc)
4637 Merge bsdi_2_1 to bsdi(Bruce)
4638 Monitor program removed
4639 Name change from Postgres95 to PostgreSQL
4640 New config.h file(Marc, Bryan)
4641 PG_VERSION now set to 6.0 and used by postmaster
4642 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
4643 Reduced the number of #define's, centeralized #define's
4644 Remove duplicate OIDS in system tables(Dan)
4645 Remove duplicate system catalog info or report mismatches(Dan)
4646 Removed many os-specific #define's
4647 Restructured object file generation/location(Bryan, Marc)
4648 Restructured port-specific file locations(Bryan, Marc)
4649 Unused/uninialized variables corrected
4650 _________________________________________________________________
4654 Release date: 1996-11-04
4656 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of the
4657 changes listed in 6.0 were actually included in the 1.02.1 to 1.09
4659 _________________________________________________________________
4663 Release date: 1996-08-01
4664 _________________________________________________________________
4666 Migration from version 1.02 to version 1.02.1
4668 Here is a new migration file for 1.02.1. It includes the 'copy' change
4669 and a script to convert old ASCII files.
4671 Note: The following notes are for the benefit of users who want to
4672 migrate databases from Postgres95 1.01 and 1.02 to Postgres95
4675 If you are starting afresh with Postgres95 1.02.1 and do not need
4676 to migrate old databases, you do not need to read any further.
4678 In order to upgrade older Postgres95 version 1.01 or 1.02 databases to
4679 version 1.02.1, the following steps are required:
4680 1. Start up a new 1.02.1 postmaster
4681 2. Add the new built-in functions and operators of 1.02.1 to 1.01 or
4682 1.02 databases. This is done by running the new 1.02.1 server
4683 against your own 1.01 or 1.02 database and applying the queries
4684 attached at the end of the file. This can be done easily through
4685 "psql". If your 1.01 or 1.02 database is named testdb and you have
4686 cut the commands from the end of this file and saved them in
4688 % psql testdb -f addfunc.sql
4689 Those upgrading 1.02 databases will get a warning when executing
4690 the last two statements in the file because they are already
4691 present in 1.02. This is not a cause for concern.
4692 _________________________________________________________________
4694 Dump/Reload Procedure
4696 If you are trying to reload a pg_dump or text-mode, copy tablename to
4697 stdout generated with a previous version, you will need to run the
4698 attached "sed" script on the ASCII file before loading it into the
4699 database. The old format used '.' as end-of-data, while '\.' is now
4700 the end-of-data marker. Also, empty strings are now loaded in as ''
4701 rather than NULL. See the copy manual page for full details.
4702 sed 's/^\.$/\\./g' <in_file >out_file
4704 If you are loading an older binary copy or non-stdout copy, there is
4705 no end-of-data character, and hence no conversion necessary.
4706 -- following lines added by agc to reflect the case-insensitive
4707 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
4708 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexe
4710 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregex
4712 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregex
4714 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticrege
4716 _________________________________________________________________
4720 Source code maintenance and development
4721 * worldwide team of volunteers
4722 * the source tree now in CVS at ftp.ki.net
4725 * psql (and underlying libpq library) now has many more options for
4726 formatting output, including HTML
4727 * pg_dump now output the schema and/or the data, with many fixes to
4728 enhance completeness.
4729 * psql used in place of monitor in administration shell scripts.
4730 monitor to be deprecated in next release.
4731 * date/time functions enhanced
4732 * NULL insert/update/comparison fixed/enhanced
4733 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
4735 Bug Fixes (almost too numerous to mention)
4737 * storage management
4738 * check for NULL pointer before dereferencing
4742 * added SolarisX86 port
4743 * added BSD/OS 2.1 port
4745 _________________________________________________________________
4749 Release date: 1996-02-23
4750 _________________________________________________________________
4752 Migration from version 1.0 to version 1.01
4754 The following notes are for the benefit of users who want to migrate
4755 databases from Postgres95 1.0 to Postgres95 1.01.
4757 If you are starting afresh with Postgres95 1.01 and do not need to
4758 migrate old databases, you do not need to read any further.
4760 In order to Postgres95 version 1.01 with databases created with
4761 Postgres95 version 1.0, the following steps are required:
4762 1. Set the definition of NAMEDATALEN in "src/Makefile.global" to 16
4763 and OIDNAMELEN to 20.
4764 2. Decide whether you want to use Host based authentication.
4765 a. If you do, you must create a file name pg_hba in your
4766 top-level data directory (typically the value of your
4767 $PGDATA). "src/libpq/pg_hba" shows an example syntax.
4768 b. If you do not want host-based authentication, you can comment
4771 in "src/Makefile.global"
4772 Note that host-based authentication is turned on by default,
4773 and if you do not take steps A or B above, the out-of-the-box
4774 1.01 will not allow you to connect to 1.0 databases.
4775 3. Compile and install 1.01, but DO NOT do the "initdb" step.
4776 4. Before doing anything else, terminate your 1.0 postmaster, and
4777 backup your existing $PGDATA directory.
4778 5. Set your PGDATA environment variable to your 1.0 databases, but
4779 set up path up so that 1.01 binaries are being used.
4780 6. Modify the file "$PGDATA/PG_VERSION" from 5.0 to 5.1
4781 7. Start up a new 1.01 postmaster
4782 8. Add the new built-in functions and operators of 1.01 to 1.0
4783 databases. This is done by running the new 1.01 server against
4784 your own 1.0 database and applying the queries attached and saving
4785 in the file 1.0_to_1.01.sql. This can be done easily through
4786 "psql". If your 1.0 database is name testdb:
4787 % psql testdb -f 1.0_to_1.01.sql
4788 and then execute the following commands (cut and paste from here):
4789 -- add builtin functions that are new to 1.01
4791 create function int4eqoid (int4, oid) returns bool as 'foo'
4792 language 'internal';
4793 create function oideqint4 (oid, int4) returns bool as 'foo'
4794 language 'internal';
4795 create function char2icregexeq (char2, text) returns bool as 'foo'
4796 language 'internal';
4797 create function char2icregexne (char2, text) returns bool as 'foo'
4798 language 'internal';
4799 create function char4icregexeq (char4, text) returns bool as 'foo'
4800 language 'internal';
4801 create function char4icregexne (char4, text) returns bool as 'foo'
4802 language 'internal';
4803 create function char8icregexeq (char8, text) returns bool as 'foo'
4804 language 'internal';
4805 create function char8icregexne (char8, text) returns bool as 'foo'
4806 language 'internal';
4807 create function char16icregexeq (char16, text) returns bool as 'foo'
4808 language 'internal';
4809 create function char16icregexne (char16, text) returns bool as 'foo'
4810 language 'internal';
4811 create function texticregexeq (text, text) returns bool as 'foo'
4812 language 'internal';
4813 create function texticregexne (text, text) returns bool as 'foo'
4814 language 'internal';
4816 -- add builtin functions that are new to 1.01
4818 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
4819 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
4820 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexe
4822 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregex
4824 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexe
4826 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregex
4828 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexe
4830 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregex
4832 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icrege
4834 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icreg
4836 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq)
4838 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne
4840 _________________________________________________________________
4845 * 1.01 is backwards compatible with 1.0 database provided the user
4846 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
4847 If those steps are not taken, 1.01 is not compatible with 1.0 database.
4850 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
4851 * added NeXT port (requires SysVIPC implementation)
4852 * added CAST .. AS ... syntax
4853 * added ASC and DESC key words
4854 * added 'internal' as a possible language for CREATE FUNCTION
4855 internal functions are C functions which have been statically linked
4856 into the postgres backend.
4857 * a new type "name" has been added for system identifiers (table names,
4858 attribute names, etc.) This replaces the old char16 type. The
4859 of name is set by the NAMEDATALEN #define in src/Makefile.global
4860 * a readable reference manual that describes the query language.
4861 * added host-based access control. A configuration file ($PGDATA/pg_hba)
4862 is used to hold the configuration data. If host-based access control
4863 is not desired, comment out HBA=1 in src/Makefile.global.
4864 * changed regex handling to be uniform use of Henry Spencer's regex code
4865 regardless of platform. The regex code is included in the distribution
4866 * added functions and operators for case-insensitive regular expressions.
4867 The operators are ~* and !~*.
4868 * pg_dump uses COPY instead of SELECT loop for better performance
4871 * fixed an optimizer bug that was causing core dumps when
4872 functions calls were used in comparisons in the WHERE clause
4873 * changed all uses of getuid to geteuid so that effective uids are used
4874 * psql now returns non-zero status on errors when using -c
4875 * applied public patches 1-14
4876 _________________________________________________________________
4880 Release date: 1995-09-05
4881 _________________________________________________________________
4886 * The copyright of Postgres 1.0 has been loosened to be freely modifiable
4887 and modifiable for any purpose. Please read the COPYRIGHT file.
4888 Thanks to Professor Michael Stonebraker for making this possible.
4891 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
4892 EUROPEAN STYLE). This follows SQL-92 specs.
4893 * "delimiters" is now a key word
4896 * sql LIKE syntax has been added
4897 * copy command now takes an optional USING DELIMITER specification.
4898 delimiters can be any single-character string.
4899 * IRIX 5.3 port has been added.
4900 Thanks to Paul Walmsley and others.
4901 * updated pg_dump to work with new libpq
4902 * \d has been added psql
4903 Thanks to Keith Parks
4904 * regexp performance for architectures that use POSIX regex has been
4905 improved due to caching of precompiled patterns.
4906 Thanks to Alistair Crooks
4907 * a new version of libpq++
4908 Thanks to William Wanders
4911 * arbitrary userids can be specified in the createuser script
4912 * \c to connect to other databases in psql now works.
4913 * bad pg_proc entry for float4inc() is fixed
4914 * users with usecreatedb field set can now create databases without
4915 having to be usesuper
4916 * remove access control entries when the entry no longer has any
4918 * fixed non-portable datetimes implementation
4919 * added kerberos flags to the src/backend/Makefile
4920 * libpq now works with kerberos
4921 * typographic errors in the user manual have been corrected.
4922 * btrees with multiple index never worked, now we tell you they don't
4923 work when you try to use them
4924 _________________________________________________________________
4926 Postgres95 Release 0.03
4928 Release date: 1995-07-21
4929 _________________________________________________________________
4933 Incompatible changes:
4934 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
4935 (due to system catalog changes and indexing structure changes).
4936 * double-quote (") is deprecated as a quoting character for string literals;
4937 you need to convert them to single quotes (').
4938 * name of aggregates (eg. int4sum) are renamed in accordance with the
4939 SQL standard (eg. sum).
4940 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
4941 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
4942 previous releases); you might have to do typecasting if you depend on it
4943 being of type float8. If you neglect to do the typecasting and you assign
4944 a float literal to a field of type float8, you may get incorrect values
4946 * LIBPQ has been totally revamped so that frontend applications
4947 can connect to multiple backends
4948 * the usesysid field in pg_user has been changed from int2 to int4 to
4949 allow wider range of Unix user ids.
4950 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
4951 single BSD44_derived port. (thanks to Alistair Crooks)
4953 SQL standard-compliance (the following details changes that makes postgres95
4954 more compliant to the SQL-92 standard):
4955 * the following SQL types are now built-in: smallint, int(eger), float, real,
4956 char(N), varchar(N), date and time.
4958 The following are aliases to existing postgres types:
4960 integer, int -> int4
4961 float, real -> float4
4962 char(N) and varchar(N) are implemented as truncated text types. In
4963 addition, char(N) does blank-padding.
4964 * single-quote (') is used for quoting string literals; '' (in addition to
4965 \') is supported as means of inserting a single quote in a string
4966 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
4967 (Also, aggregates can now be overloaded, i.e. you can define your
4968 own MAX aggregate to take in a user-defined type.)
4969 * CHANGE ACL removed. GRANT/REVOKE syntax added.
4970 - Privileges can be given to a group using the "GROUP" key word.
4972 GRANT SELECT ON foobar TO GROUP my_group;
4973 The key word 'PUBLIC' is also supported to mean all users.
4975 Privileges can only be granted or revoked to one user or group
4978 "WITH GRANT OPTION" is not supported. Only class owners can change
4980 - The default access control is to to grant users readonly access.
4981 You must explicitly grant insert/update access to users. To change
4982 this, modify the line in
4983 src/backend/utils/acl.h
4984 that defines ACL_WORLD_DEFAULT
4987 * the bug where aggregates of empty tables were not run has been fixed. Now,
4988 aggregates run on empty tables will return the initial conditions of the
4989 aggregates. Thus, COUNT of an empty table will now properly return 0.
4990 MAX/MIN of an empty table will return a row of value NULL.
4991 * allow the use of \; inside the monitor
4992 * the LISTEN/NOTIFY asynchronous notification mechanism now work
4993 * NOTIFY in rule action bodies now work
4994 * hash indexes work, and access methods in general should perform better.
4995 creation of large btree indexes should be much faster. (thanks to Paul
4998 Other changes and enhancements:
4999 * addition of an EXPLAIN statement used for explaining the query execution
5000 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
5002 * WARN and NOTICE messages no longer have timestamps on them. To turn on
5003 timestamps of error messages, uncomment the line in
5004 src/backend/utils/elog.h:
5005 /* define ELOG_TIMESTAMPS */
5006 * On an access control violation, the message
5007 "Either no such class or insufficient privilege"
5008 will be given. This is the same message that is returned when
5009 a class is not found. This dissuades non-privileged users from
5010 guessing the existence of privileged classes.
5011 * some additional system catalog changes have been made that are not
5012 visible to the user.
5015 * The -oid option has been added to the "pg_result" tcl command.
5016 pg_result -oid returns oid of the last row inserted. If the
5017 last command was not an INSERT, then pg_result -oid returns "".
5018 * the large object interface is available as pg_lo* tcl commands:
5019 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
5021 Portability enhancements and New Ports:
5022 * flex/lex problems have been cleared up. Now, you should be able to use
5023 flex instead of lex on any platforms. We no longer make assumptions of
5024 what lexer you use based on the platform you use.
5025 * The Linux-ELF port is now supported. Various configuration have been
5026 tested: The following configuration is known to work:
5027 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
5028 with everything in ELF format,
5031 * ipcclean added to the distribution
5032 ipcclean usually does not need to be run, but if your backend crashes
5033 and leaves shared memory segments hanging around, ipcclean will
5034 clean them up for you.
5037 * the user manual has been revised and libpq documentation added.
5038 _________________________________________________________________
5040 Postgres95 Release 0.02
5042 Release date: 1995-05-25
5043 _________________________________________________________________
5047 Incompatible changes:
5048 * The SQL statement for creating a database is 'CREATE DATABASE' instead
5049 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
5050 of 'DESTROYDB'. However, the names of the executables 'createdb' and
5051 'destroydb' remain the same.
5054 * pgperl - a Perl (4.036) interface to Postgres95
5055 * pg_dump - a utility for dumping out a postgres database into a
5056 script file containing query commands. The script files are in a ASCII
5057 format and can be used to reconstruct the database, even on other
5058 machines and other architectures. (Also good for converting
5059 a Postgres 4.2 database to Postgres95 database.)
5061 The following ports have been incorporated into postgres95-beta-0.02:
5062 * the NetBSD port by Alistair Crooks
5063 * the AIX port by Mike Tung
5064 * the Windows NT port by Jon Forrest (more stuff but not done yet)
5065 * the Linux ELF port by Brian Gallew
5067 The following bugs have been fixed in postgres95-beta-0.02:
5068 * new lines not escaped in COPY OUT and problem with COPY OUT when first
5070 * cannot type return to use the default user id in createuser
5071 * SELECT DISTINCT on big tables crashes
5072 * Linux installation problems
5073 * monitor doesn't allow use of 'localhost' as PGHOST
5074 * psql core dumps when doing \c or \l
5075 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
5076 * libpgtcl has a hard-wired default port number
5077 * SELECT DISTINCT INTO TABLE hangs
5078 * CREATE TYPE doesn't accept 'variable' as the internallength
5079 * wrong result using more than 1 aggregate in a SELECT
5080 _________________________________________________________________
5082 Postgres95 Release 0.01
5084 Release date: 1995-05-01