2 $PostgreSQL: pgsql/doc/src/sgml/release.sgml,v 1.360 2005/08/24 19:34:33 momjian Exp $
6 [A-Z][A-Z ]+[A-Z] <command>
7 [A-Za-z_][A-Za-z0-9_]+() <function>
8 [A-Za-z_]/[A-Za-z_]+ <filename>
9 PostgreSQL <productname>
10 pg_[A-Za-z0-9_] <appication>
11 [A-Z][A-Z] <type>, <envar.
16 <appendix id="release">
17 <title>Release Notes</title>
19 <sect1 id="release-8-1">
20 <title>Release 8.1</title>
23 <title>Release date</title>
24 <simpara>2005-1?-??, Current as of 2005-08-24</simpara>
28 <title>Overview</title>
31 Major changes in this release:
38 Improve concurrent access to the shared buffer cache (Tom)
43 This was accomplished by eliminating global locks and using a
44 clock sweep algorithm to find free buffers. This increases
45 scalability on multi-CPU systems.
52 Automatically use indexes for <function>MIN()</> and
53 <function>MAX()</> (Tom)
58 In previous releases, the only way to use an index for
59 <function>MIN()</> or <function>MAX()</> was to rewrite the
60 query as <command>SELECT col FROM tab ORDER BY col LIMIT 1</>.
61 Index usage now happens automatically.
68 Add in-memory bitmaps which allows multiple indexes to be merged
69 in a single query (Tom)
74 In previous releases, only a single index could be used to do
75 lookups on a table. With this feature, if a query has
76 <command>WHERE tab.col1 = 4 and tab.col2 = 9</>, and there is no
77 multicolumn index on col1 and col2, but there is an index on
78 col1 and another on col2, it is possible to do lookups on the
79 col1 index and the col2 index and combine them in memory to do
80 heap lookups on rows matching both the col1 and col2
81 restrictions. This is very useful in environments that have a
82 lot of unstructured queries where it is impossible to create
83 indexes that match all possible access conditions.
90 Add two-phase commit (Heikki Linnakangas, Alvaro, Tom)
95 Two-phase commit allows transactions to be "prepared" on several
96 computers, and once all computers have successfully prepared
97 their transactions (none failed), all transactions can be
98 committed. Even if a machine crashes after a prepare, the
99 prepared transaction can be committed after it is restarted. New
100 syntax includes <command>PREPARE TRANSACTION</> and
101 <command>COMMIT/ROLLBACK PREPARED</>. A new system view
102 <literal>pg_prepared_xacts</> has also been added.
109 Create a new role system that replaces users and groups
115 Roles are a combination of users and groups. Like users, they
116 can have login capability, and like groups, a role can have
117 other roles as members. Roles basically remove the distinction
118 between users and groups. For example, a role can:
125 Have optionally have login capability
137 Inherit permission from other member roles
143 Switch to another member role using <command>SET ROLE</>
149 So, once a user logs into a role, she inherits capabilities of
150 the login role plus any inherited roles, and can use
151 <command>SET ROLE</> to switch to other member roles. This
152 change also replaces <literal>pg_shadow</> and
153 <literal>pg_group</> by with new role-capable catalogs
154 <literal>pg_authid</> and <literal>pg_auth_members</>. The old
155 tables are redefined as views on the new role tables.
162 Move <filename>/contrib/pgautovacuum</> into the main server
168 Moving pgautovacuum from <filename>/contrib</> allows it to be
169 automatically started and stoped in sync with the database
170 server, and allows for pgautovacuum to be configured from
178 Add shared row level locks using <command>SELECT ... FOR SHARE</>
184 While <productname>PostgreSQL</productname>'s MVCC locking
185 allows <command>SELECT</> to never be blocked by writers and
186 therefore does not need shared row locks for typical operations,
187 shared locks are useful for applications that require shared row
188 locking, and to reduce the locking requirements to maintain
189 referential integrity.
198 <title>Migration to version 8.1</title>
201 A dump/restore using <application>pg_dump</application> is required
202 for those wishing to migrate data from any previous release.
206 The 8.0 release announced that the <function>to_char()</> function
207 for intervals would be removed in 8.1. However, since no better API
208 has been suggested, to_char(interval) has been enhanced in 8.1 and
209 will remain in the server.
213 Observe the following incompatibilities:
220 Change add_missing_from to 'false'
223 Generate an error if a table used in a query without a
224 <command>FROM</> reference (Neil) This command, <command>SELECT
225 pg_class.*</>, now generates an error.
231 Cause input of a zero-length strings ('') for float4/float8/oid
232 to throw an error, rather than treat it as a zero (Neil)
235 This change is consistent with the current handling of
236 zero-length strings for integers. The schedule for this change
237 was announced in 8.0.
243 Change <varname>default_with_oids</> to default to false (Neil)
246 With this option set to false, user-created tables no longer have
247 an the usually-invisible OID column unless <command>WITH OIDS</>
248 is specified in <command>CREATE TABLE</>. Though OIDs have
249 existed in all previous releases of
250 <productname>PostgreSQL</productname>, their use is limited
251 because they are only four bytes long and the counter is unique
252 across all installed databases. The preferred way of uniquely
253 identifying rows is via sequences and <command>SERIAL</>, which
254 has been supported since <productname>PostgreSQL</productname>
261 Add <literal>E''</> syntax so eventually ordinary strings can
262 treat backslashes literally (Bruce)
265 Currently <productname>PostgreSQL</productname> processes a
266 backslash in a string as preceeding a character that requires
267 special processing, e.g. <literal>\n</> or <literal>\010</>.
268 While this allows easy entry of special values, it is
269 non-standard and makes porting of application from other
270 databases more difficult. For this reason, the
271 <productname>PostgreSQL</productname> project is planning to
272 remove the special meaning of backslashes in strings. For
273 backward compatibility and for users who want special backslash
274 procesing, a new string type will be created. This new string
275 type is formed by having an E preceed the single quote that
276 starts the string, e.g. <literal>E'hi\n'</>. While this release
277 does not change the handling of backslashes in strings, it does
278 add several new GUC variables to help users migrate applications
285 <varname>escape_string_warning</> - warn about backslashes in
286 ordinary (non-E) strings
292 <varname>escape_string_syntax</> - does this release support
299 <varname>standard_conforming_strings</> - does this release
300 treat backslashes literally in ordinary strings?
307 The last two values are read-only and should assist in the
308 porting of applications. Applications can retrieve these values
309 to know how backslashes are processed. In a later release,
310 standard_conforming_strings will be true, meaning backslashes
311 will be treated literally in non-E strings. To prepare for this
312 change, use E'' strings in places that need special backslash
313 processing, and turn on escape_string_warning to find additional
314 strings that need to be converted to use <literal>E''</>. Also,
315 use two single-quotes ('') to embed a literal single-quote
316 in a string, rather than the PostgreSQL-supported syntax of
317 backslash single-quote (\'). The former is standards-conforming
318 and does not require the use of the E'' string syntax. You can
319 also use the $$ string syntax, which does not treat backslashes
326 Make <command>REINDEX DATABASE</> reindex all indexes in the
330 The old behavior of <command>REINDEX</> database reindexed only
331 system tables. This new behavior seems more intuitive. A new
332 command <command>REINDEX SYSTEM</> allows for reindexing just the
339 In psql, treat unquoted <literal>\{digit}+</> sequences as octal
343 In previous releases, <literal>\{digit}+</> sequences were
344 treated as decimal, and only <literal>\0{digit}+</> were treated
345 as octal. This change was made for consistency.
351 Remove grammar productions for prefix and postfix <literal>%</>
352 and <literal>^</> operators
356 These have never been documented and complicated the use of the
357 modulus operator (<literal>%</>) with negative numbers.
363 Make <literal>&<</> and <literal>&></> for polygons
364 consistent with the box "over" operators (Tom)
372 <title>Additional Changes</title>
375 Below you will find a detailed account of the additional changes
376 between release 8.1 and the previous major release.
380 <title>Performance Improvements</title>
385 Improve rtree index capabilities and performance (Neil)
391 Improve the optimizer, including auto-resizing of hash joins
398 Overhaul internal API in several areas
404 Change WAL CRC records from 64bit to 32bit (Tom)
410 Prevent writing large empty gaps in WAL pages (Tom)
416 Allow non-consecutive index columns to be used in a multi-column
421 For example, this allows an index on columns a,b,c to be used in
422 a query with <command>WHERE a = 4 and c = 10</>.
428 Skip WAL logging for <command>CREATE TABLE AS</> /
429 <command>SELECT INTO</> (Simon)
432 Since a crash during <command>CREATE TABLE</> would cause the
433 table to be dropped during recovery, there is no reason to WAL
434 log as the table is loaded.
440 Allow concurrent GIST index access (Teodor, Oleg)
446 Add GUC <varname>full_page_writes</> to control writing full
451 To prevent partial disk writes from corrupting the database,
452 <productname>PostgreSQL</productname> writes a complete copy of
453 each database disk page to WAL the first time it is modified
454 after a checkpoint. This turns off that functionality for users
455 with battery-backed disk caches where partial page writes cannot
462 Add constraint_exclusion to restrict child table lookups based
463 on table constraints (Simon)
466 This allows for a type of table partitioning. If child table
467 placed in a different tablespaces using appropriate
468 <command>CHECK</> constraints, the optimizer will skip child
469 table accesses if the constraint guarantees no matching rows
470 exist in the child table.
476 Use <literal>O_DIRECT</> if available when using
477 <literal>O_SYNC</> for wal_sync_method (Itagaki Takahiro)
480 <literal>O_DIRECT</> causes disk writes to bypass the kernel
481 cache, and for WAL writes, this improves performance.
487 Improve <command>COPY FROM</> performance (Alon Goldshuv)
490 This was accomplished by reading <command>COPY</> input in
491 larger chunks, rather than character by character.
500 <title>Server Changes</title>
505 Prevent problems due to transaction ID (XID) wraparound (Tom)
508 This was accomplished by warning the transaction counter is near
509 the the earliest transaction id determined by the database whose
510 last vacuum is the oldest. If the limit is reached, the server
511 will no longer accept queries.
517 Fix problem of object ID (OID) wraparound conflicting with
518 existing system objects (Tom)
524 Add warning about the need to increase
525 <varname>max_fsm_relations</> and <varname>max_fsm_pages</>
526 during <command>VACUUM</> (Ron Mayer)
532 Add <varname>temp_buffers</> GUC variable to allow users to
533 determine the size of the local buffer area for temporary table
540 Add session start time and client IP address to
541 <literal>pg_stat_activity</> (Magnus)
547 Enhance <literal>pg_locks</> display (Tom)
553 Log queries for client-side <command>PREPARE</> and
554 <command>EXECUTE</> (Simon)
560 Allow Kerberos name and user name case sensitivity to be
561 specified in <filename>postgresql.conf</> (Magnus)
567 Add GUC <varname>krb_server_hostname</> so the server hostname
568 can be specified as part of service principal (Todd Kover)
571 If not set, any service principal matching an entry in the
572 keytab can be used. This is new Kerberos matching behavior in
579 Add <varname>log_line_prefix</> options for millisecond
580 timestamps (<literal>%m</>) and remote host (<literal>%h</>) (Ed
587 Add WAL logging for GIST indexes (Teodor, Oleg)
590 GIST indexes are now safe for crash and point-in-time recovery
596 Remove old <filename>*.backup</> files when we do
597 <function>pg_stop_backup()</> (Bruce)
600 This prevents a large number of <filename>*.backup</> files from
601 existing in <filename>/pg_xlog</>.
607 Add GUC variables to control TCP/IP keep-alive times for idle,
608 interval, and count (Oliver Jowett)
614 Add per-user and per-database connection limits (Petr Jelinek)
617 Using <command>ALTER USER</> and <command>ALTER DATABASE</>,
618 limits can now be enforced on the maximum number of users who
619 can connect as as a specific uesr or to a specific database.
620 Setting the limit to zero disables user or database connections.
626 Allow more than two gigabyes of shared memory and per-backend
627 work memory on 64-bit machines (Koichi Suzuki)
636 <title>Query Changes</title>
641 Add temporary views (Koju Iijima)
647 Fix <command>HAVING</> without any aggregate functions or
648 <command>GROUP BY</> so that the query returns a single group (Tom)
651 Previously, such a case would treat the <command>HAVING</>
652 clause the same as a <command>WHERE</> clause. This was not per spec.
658 Add <command>USING</> clause to allow additional tables to be
659 specified to <command>DELETE</> (Euler Taveira de Oliveira)
662 In prior releases, there was no clear method for specifying
663 additional tables to be used for joins in a <command>DELETE</>
664 statement. <command>UPDATE</> already has a <command>FROM</>
665 clause for this purpose.
671 Add support for <literal>\x</> hex escapes in backend and ecpg
675 This is just like the standard C <literal>\x</> escape syntax.
676 Octal escapes were already supported.
682 Add <command>BETWEEN SYMMETRIC</> query syntax (Pavel Stehule)
685 This feature allows <command>BETWEEN</> comparisons without
686 requiring the first value to be less than the second. For
687 example, <command>2 BETWEEN [ASYMMETRIC] 3 AND 1</> returns
688 false, while <command>2 BETWEEN SYMMETRIC 3 AND 1</> returns
689 true. <command>BETWEEN ASYMMETRIC</> was already supported.
695 Add <command>NOWAIT</> option to <command>SELECT ... FOR
696 UPDATE/SHARE</> (Hans-Juergen Schoenig)
699 While <command>SET</> statement_timeout allows a query taking
700 over a certain amount of time to be cancelled, the <command>NO
701 WAIT</> option allows a query to be canceled as soon as a
702 <command>SELECT ... FOR UPDATE/SHARE</> cannot immediately
712 <title>Object Manipulation Changes</title>
717 Track dependencies of shared objects (Alvaro)
720 <productname>PostgreSQL</productname> allows global tables
721 (users, databases, tablespaces) to reference information in
722 multiple databases. This addition adds dependency information
723 for global tables, so, for example, user ownership can be
724 tracked across databases, so a user who owns something in any
725 database can no longer be removed. Dependency tracking already
726 existed for database-local objects.
732 Allow limited <command>ALTER OWNER</> commands to be performed
733 by the object owner (Stephen Frost)
736 Prior releases allowed only the super-user to change object owners,
737 even if the current owner was executing the command and the new
738 owner was in the same group. With roles such transfers are now
745 Add <command>ALTER</> object <command>SET SCHEMA</> capability
746 for some object types (tables, functions, types) (Bernd Helmle)
749 This allows objects to be moved to different schemas.
755 Add ALTER TABLE ENABLE/DISABLE TRIGGER to disable triggers
765 <title>Utility Command Changes</title>
770 Allow <command>TRUNCATE</> to truncate multiple tables in a
771 single command (Alvaro)
774 Because of referential integrity checks, it is impossible to
775 truncate a table that is part of a referential integrity
776 constraint. Using this new functionality, <command>TRUNCATE</>
777 can be used to truncate all tables involved in referential
778 integrity if they are truncated in a single <command>TRUNCATE</>
785 Properly process carriage returns and line feeds in
786 <command>COPY CSV</> mode (Andrew Dunstan)
789 In release 8.0, carriage returns and line feeds in <command>CSV
790 COPY TO</> were processed in an inconsistent manner. (This was
791 documented on the TODO list.)
797 Add <command>COPY WITH CSV HEADER</> to allow a header line as
798 the first line in <command>COPY</> (Andrew)
801 This allows handling of the common <command>CSV</> usage of
802 placing the column names on the first line of the data file. For
803 <command>COPY TO</>, the first line contains the column names,
804 and for <command>COPY FROM</>, the first line is ignored.
810 On Win32, display better sub-second precision in
811 <command>EXPLAIN ANALYZE</> (Magnus)
817 Add trigger duration display to <command>EXPLAIN ANALYZE</>
821 Prior releases accumulated trigger execution time as part of the
822 total execution time.
828 Add support for <literal>\x</> hex escapes in <command>COPY</>
832 Previous releases only supported octal escapes.
838 Have <command>SHOW ALL</> include variable descriptions
842 <command>SHOW</> varname still only displays the variable's
843 value and does not include the description.
849 Have initdb create new standard database called
850 <literal>postgres</> and convert utilities to use
851 <literal>postgres</> rather than <literal>template1</> for
852 standard lookups (Dave)
855 In prior releases, template1 was used both as a default
856 connection for utilities like createuser, and as a template for
857 new databases. This caused <command>CREATE DATABASE</> to
858 sometimes fail because a new database cannot be created if
859 anyone else is in the template database. With this change, the
860 default connection database is now <literal>postgres</>, meaning
861 it is much less likely someone will be using
862 <literal>template1</> during <command>CREATE DATABASE</>.
868 Create new reindexdb command-line utility by moving
869 <filename>/contrib/reindexdb</> into the server (Euler Taveira
879 <title>Data Type and Function Changes</title>
884 Add <function>MAX()</> and <function>MIN()</> aggregates for
885 array types (Koju Iijima)
894 Fix <function>to_date()</> and <function>to_timestamp()</> to
895 behave reasonably when <literal>CC</> and <literal>YY</> fields
896 are both used (Karel Zak)
899 If the format specification contains <literal>CC</> and a year
900 specification is <literal>YYY</> or longer, ignore the
901 <literal>CC</>. If the year specification is <literal>YY</> or
902 shorter, interpret <literal>CC</> as the previous century. ?
908 Add <function>md5(bytea)</> (Abhijit Menon-Sen)
911 <function>md5(text)</> already existed.
917 Fix <type>CHAR()</> to properly pad spaces to the specified
918 length when using a multiple-byte character set (Yoshiyuki
922 In prior releases, the padding of <type>CHAR()</> was incorrect
923 because it only padded to the specified number of bytes without
924 considering how many characters were stored.
930 Add support for <command>NUMERIC ^ NUMERIC</> based on
931 <function>power(numeric, numeric)</>
934 The function already existed, but there was no operator assigned
941 Fix <type>NUMERIC</> modulus by properly truncating the quotient
942 during computation (Bruce)
945 In previous releases, modulus for large values sometimes
946 returned negative results due to the rounding of the quotient.
952 Add a function <function>lastval()</>(Dennis Björklund)
955 <function>lastval()</> is a simplified version of
956 <function>currval()</>. It automatically determines the proper
957 sequence name based on the most recent <function>nextval()</> or
958 <function>setval()</> call performed by the current session.
964 Add <function>to_timestamp(DOUBLE PRECISION)</>
967 Converts Unix seconds since 1970 to a <type>TIMESTAMP WITH
974 Add <function>pg_postmaster_start_time()</> function (Euler
975 Taveira de Oliveira, Matthias Schmidt)
981 Allow the full use of time zone names in <command>AT TIME
982 ZONE</>, not just the short list previously available (Magnus)
985 Previously, only a predefined list of time zone names were
986 supported by <command>AT TIME ZONE</>. Now any supported time
987 zone name can be used, e.g.:
989 SELECT CURRENT_TIMESTAMP AT TIME ZONE 'Europe/London';
991 In the above query, the time zone used is adjusted based on the
992 daylight savings time rules that were in effect on the supplied
999 Add <function>GREATEST()</> and <function>LEAST()</> functions
1003 These functions take a variable number of arguments and return
1004 the greatest or least value.
1010 Add <function>pg_column_size()</> (Mark Kirkwood)
1013 This returns storage size of a column, including any
1020 Add <function>regexp_replace()</> (Atsushi Ogawa)
1023 This allows regular expression replacement, like sed. A
1024 four-argument version also allows for global (replace all) and
1025 case-insensitive modes.
1031 Fix interval division and multiplication (Bruce)
1034 Previous versions sometimes returned unjustified results, like
1035 <command>'4 months'::interval / 5</> returning <command>'1 mon
1042 Add an internal day field to <type>INTERVAL</> so a one day
1043 interval can be distinguished from a 24 hour interval (Michael
1047 Days that contain a daylight savings time adjustment are not 24
1048 hours, but typically 23 or 25 hours. This change allows days
1049 (not fixed 24-hour periods) to be added to dates who's result
1050 includes a daylight savings time adjustment period. Therefore,
1051 while in previous releases <literal>1 day</> and <literal>24
1052 hours</> where interchangeable interval periods, in this release
1053 they are treated differently, e.g.
1055 '2005-05-03 00:00:00 EST' + '1 day' = '2005-05-04 00:00:00-04'
1056 '2005-05-03 00:00:00 EST' + '24 hours' = '2005-05-04 01:00:00-04'
1063 Add <function>justify_days()</> and <function>justify_hours()</>
1064 (Michael Glaesemann)
1067 These functions, respectively, adjust days to an appropriate
1068 number of full months and days, and adjust hours to an
1069 appropriate number of full days and hours.
1075 Move <filename>/contrib/dbsize</> into the backend, and rename
1076 some of the functions (Dave Page, Andreas Pflug)
1083 <function>pg_tablespace_size()</>
1089 <function>pg_database_size()</>
1095 <function>pg_relation_size()</>
1101 <function>pg_complete_relation_size()</>
1107 <function>pg_size_pretty()</>
1114 <function>pg_complete_relation_size()</> includes indexes and TOAST
1121 Add functions for read-only file access to the cluster directory
1122 (Dave Page, Andreas Pflug)
1129 <function>pg_stat_file()</>
1135 <function>pg_read_file()</>
1141 <function>pg_ls_dir()</>
1151 Add <function>pg_reload_conf()</> to force reloading of the
1152 configuration files (Dave Page, Andreas Pflug)
1159 Add <function>pg_rotate_logfile()</> to force rotation of the
1160 server log file (Dave Page,
1166 Change <literal>pg_stat_*</> views to show TOAST tables (Tom)
1175 <title>Encoding and Locale Changes</title>
1180 Rename some encodings to be more consistent and to follow
1181 international standards(Bruce)
1188 <literal>UNICODE</> is now <literal>UTF8</>
1194 <literal>ALT</> is now <literal>WIN866</>
1200 <literal>WIN</> is now <literal>WIN1251</>
1206 <literal>TCVN</> is now <literal>WIN1258</>
1213 The original names still work.
1219 Add support for <literal>WIN1252</> encoding (Roland Volkmann)
1225 Add support for 3 and 4-byte <literal>UTF8</> characters (John
1229 Previously only one and two-byte <literal>UTF8</> characters
1230 were supported. This is particularly important for support for
1231 some Chinese character.
1237 Allow direct conversion between <literal>EUC_JP</> and
1238 <literal>SJIS</> to improve performance (Atsushi
1245 Allow the UTF8 encoding to work on Win32 (Magnus)
1248 This is done by mapping UTF8 to the Win32-native UTF16
1257 <title>General Server-Side Language Changes</title>
1262 Fix <command>ALTER LANGUAGE RENAME</> (Sergey Yatskevich)
1268 Allow function characteristics, like strictness and volatility,
1269 to be modified via <command>ALTER FUNCTION</> (Neil)
1275 Increase the maximum number of function arguments to 100 (Tom)
1281 Allow SQL and PL/PgSQL functions to use <command>OUT</> and
1282 <command>INOUT</> parameters (Tom)
1285 <command>OUT</> is an alternate way for a function to return
1286 values. Instead of using <command>RETURN</>, values can be
1287 returned by assigning to parameters declared as <command>OUT</> or
1288 <command>INOUT</>. This is notationally simpler in some cases,
1289 particularly so when multiple values need to be returned.
1290 While returning multiple values from a function
1291 was possible in previous releases, this greatly simplifies the
1292 process. (The feature will be extended to other PLs in due course.)
1298 Move language handlers into the <literal>pg_catalog</> schema
1301 This makes it easier to drop the public schema if desired.
1309 <title>PL/PgSQL Server-Side Language Changes</title>
1314 Reduce memory usage of PL/PgSQL functions (Neil)
1317 Each function now has its own memory context that can be freed
1318 when the function exits.
1324 Check function syntax as <command>CREATE FUNCTION</> time,
1325 rather than at runtime (Neil)
1328 Previously, syntax errors were reported only when the function
1335 Allow <command>OPEN</> to open non-<command>SELECT</> queries
1336 like <command>EXPLAIN</> and <command>SHOW</> (Tom)
1342 No longer require functions to issue a <command>RETURN</>
1346 This is a byproduct of the newly added <command>OUT</> and
1347 <command>INOUT</> functionality.
1353 Add support for an optional <command>INTO</> clause to
1354 PL/PgSQL's <command>EXECUTE</> command (Pavel Stehule)
1360 Make <command>CREATE TABLE AS</> set <command>ROW_COUNT</> (Tom)
1366 Define <literal>SQLSTATE</> and <literal>SQLERRM</> to return
1367 the <literal>SQLSTATE</> and error message of the current
1368 exception (Pavel Stehule)
1371 These variables are only accessable inside exception blocks.
1377 Allow the parameters to the <command>RAISE</> statement to be
1378 expressions (Pavel Stehule)
1384 Add a loop <command>CONTINUE</> statement (Pavel Stehule)
1390 Allow block and loop labels (Pavel Stehule)
1399 <title>PL/Perl Server-Side Language Changes</title>
1404 Allow large result sets to be returned efficiently (Abhijit
1408 This allows functions to use <function>return_next()</> to avoid
1409 building the entire result set in memory.
1415 Allow one-row-at-a-time retrieval of query results (Abhijit)
1418 This allows functions to use <function>spi_query()</> and
1419 <function>spi_fetchrow()</> to avoid accumulating the entire
1420 result set in memory.
1426 Force PL/Perl to handle strings as <literal>UTF8</> if the
1427 server encoding is <literal>UTF8</> (David Kamholz)
1433 Add a validator function for PL/Perl (Andrew)
1436 This allows syntax errors to be reported at definition time,
1437 rather than execution time.
1443 Allow PL/Perl to return a Perl array when the function returns
1444 an array type (Andrew)
1447 This basically maps <productname>PostgreSQL</productname> arrays
1454 Allow Perl non-fatal warnings to generate <command>NOTICE</>
1461 Allow <literal>strict</> mode to be enabled (Andrew)
1470 <title><application>psql</> Changes</title>
1475 Add psql <command>\set ON_ERROR_ROLLBACK</> to allow statements
1476 in a transaction to error without affecting the rest of the
1477 transaction (Greg Sabino Mullane)
1480 This is basically implemented by wrapping every statement in a
1487 Add support for <literal>\x</> hex strings in psql variables
1491 Octal escapes were already supported.
1497 Add psql support for <command>troff -ms</> output format (Roger
1504 Allow psql's history file location to be controlled by
1505 <envar>HISTFILE</> (Andreas Seltenreich)
1508 This allows configuration of per-database history storage.
1514 Prevent psql <command>\x</> (expanded mode) from affecting other
1515 backslash output (Neil)
1521 Add <option>-L</> option to psql to log sessions (Lorne Sunley)
1524 This option was added because some operating systems do not have
1525 simple command-line activity logging functionality.
1531 Have psql <command>\d</> show tablespace of indexes (Qingqing
1538 Allow psql help (<command>\h</>) to make a best guess on the
1539 proper help information (Greg Sabino Mullane)
1542 This allows the user to just add <command>\h</> to the front of
1543 the syntax error query and get help on the supported syntax.
1544 Previously any additional query text beyond the help topics
1545 supported had to be removed to use <command>\h</>.
1551 Add psql <command>\pset numericlocale</> to allow numbers to be
1552 output in a locale-aware format (Eugen Nedelcu)
1555 For example, using <literal>C</> locale <literal>100000</> would
1556 be output as <literal>100,000.0</> and European locale might
1557 output this value as <literal>100.000,0</>.
1566 <title><application>pg_dump</> Changes</title>
1571 Add <option>-n</> / <option>--schema</> switch to
1572 <application>pg_restore</> (Richard van den Bergg)
1575 This allows only objects from a specified schema to be restored.
1581 Allow <application>pg_dump</> to dump a consistent snapshot of
1589 Allow <application>pg_dump</> to dump large objects even in
1596 Dump comments for large objects (Tom)
1602 Add <option>--encoding</> to <application>pg_dump</>
1606 This allows a database to be dumped in an encoding that is
1607 different from the server's encoding. This is valuable when
1608 transferring the dump to a machine with a different encoding.
1617 <title>libpq Changes</title>
1622 Add a <envar>PGPASSFILE</> environment variable to specify the
1623 password file's filename (Andrew Dunstan)
1629 Add <function>lo_create()</>, that is similar to
1630 <function>lo_creat()</> but allows the OID of the large object
1631 to be specified (Tom)
1640 <title>Source Code Changes</title>
1645 Add spinlock support for the Itanium processor using Intel
1653 Add Kerberos 5 support for Win32 (Magnus)
1659 Add Chinese FAQ (laser@pgsqldb.com)
1665 Rename Rendezvous to Bonjour to match OS/X feature renaming
1672 Add support for wal_fsync_writethrough for Darwin (Chris
1679 Streamline the passing of information within the server, the
1680 optimizer, and the lock system (Tom)
1686 Allow <application>pg_config</> to be compiled using MSVC (Andrew
1690 This is required to build DBD::Pg using <application>MSVC</>.
1696 Remove support for Kerberos V4 (Magnus)
1699 Kerberos 4 had security vulnerabilities and is no longer being
1700 maintained by the authors.
1706 Code cleanups (Coverity static analysis performed by
1713 Modify <filename>postgresql.conf</> to use documention defaults
1714 <literal>on</>/<literal>off</> rather than
1715 <literal>true</>/<literal>false</> (Bruce)
1721 Enhance <application>pg_config</> to be able to report more
1722 build-time values (Tom)
1728 Allow libpq to be build thread-safe on Win32 (Dave Page)
1737 <title>Contrib Changes</title>
1742 Add <filename>/contrib/pg_buffercache</> contrib module (Mark
1746 This displays the contents of the buffer cache, for debugging and
1747 performance tuning purposes.
1753 Remove <filename>/contrib/array</> because it is obsolete (Tom)
1759 Cleanup the <filename>/contrib/lo</> module (Tom)
1765 Move <filename>/contrib/findoidjoins</> to
1766 <filename>/src/tools</> (Tom)
1772 Remove the <literal><<</>, <literal>>></>,
1773 <literal>&<</>, and <literal>&></> operators from
1774 <filename>/contrib/cube</>
1777 These operators were not useful.
1783 Improve <filename>/contrib/btree_gist</> (Janko Richter)
1789 <filename>/contrib/pgcrypto</> (Marko Kreen)
1796 Implemention of OpenPGP symmetric-key and public-key encryption
1799 Supported are both RSA and Elgamal public-key algorithms.
1805 Standalone build: include SHA256/384/512 hashes, Fortuna PRNG
1811 OpenSSL build: support 3DES, use internal AES with OpenSSL < 0.9.7
1817 Take build parameters (OpenSSL, zlib) from <filename>./configure</> result
1820 No need to edit the <filename>Makefile</> anymore.
1826 Remove support for <filename>libmhash</> and <filename>libmcrypt</>
1839 <sect1 id="release-8-0-3">
1840 <title>Release 8.0.3</title>
1843 <title>Release date</title>
1844 <simpara>2005-05-09</simpara>
1848 This release contains a variety of fixes from 8.0.2, including several
1849 security-related issues.
1853 <title>Migration to version 8.0.3</title>
1856 A dump/restore is not required for those running 8.0.X. However,
1857 it is one possible way of handling two significant security problems
1858 that have been found in the initial contents of 8.0.X system
1859 catalogs. A dump/initdb/reload sequence using 8.0.3's initdb will
1860 automatically correct these problems.
1864 The larger security problem is that the built-in character set encoding
1865 conversion functions can be invoked from SQL commands by unprivileged
1866 users, but the functions were not designed for such use and are not
1867 secure against malicious choices of arguments. The fix involves changing
1868 the declared parameter list of these functions so that they can no longer
1869 be invoked from SQL commands. (This does not affect their normal use
1870 by the encoding conversion machinery.)
1874 The lesser problem is that the <filename>contrib/tsearch2</> module
1875 creates several functions that are misdeclared to return
1876 <type>internal</> when they do not accept <type>internal</> arguments.
1877 This breaks type safety for all functions using <type>internal</>
1882 It is strongly recommended that all installations repair these errors,
1883 either by initdb or by following the manual repair procedure given
1884 below. The errors at least allow unprivileged database users to crash
1885 their server process, and may allow unprivileged users to gain the
1886 privileges of a database superuser.
1890 If you wish not to do an initdb, perform the same manual repair
1891 procedures shown in the <link linkend="release-7-4-8">7.4.8 release
1897 <title>Changes</title>
1900 <listitem><para>Change encoding function signature to prevent
1901 misuse</para></listitem>
1902 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
1903 <type>INTERNAL</> function results</para></listitem>
1904 <listitem><para>Guard against incorrect second parameter to
1905 <function>record_out</></para></listitem>
1906 <listitem><para>Repair ancient race condition that allowed a transaction to be
1907 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
1908 than for other purposes</para>
1909 <para>This is an extremely serious bug since it could lead to apparent
1910 data inconsistencies being briefly visible to applications.</para></listitem>
1911 <listitem><para>Repair race condition between relation extension and
1913 <para>This could theoretically have caused loss of a page's worth of
1914 freshly-inserted data, although the scenario seems of very low probability.
1915 There are no known cases of it having caused more than an Assert failure.
1917 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
1919 The comparison code was wrong in the case where the
1920 <literal>--enable-integer-datetimes</> configuration switch had been used.
1921 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
1922 it will need to be <command>REINDEX</>ed after installing this update, because
1923 the fix corrects the sort order of column values.
1925 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
1926 <type>TIME WITH TIME ZONE</> values</para></listitem>
1927 <listitem><para>Fix mis-display of negative fractional seconds in
1928 <type>INTERVAL</> values</para>
1930 This error only occurred when the
1931 <literal>--enable-integer-datetimes</> configuration switch had been used.
1933 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
1934 correctly (Neil)</para></listitem>
1935 <listitem><para>Still more 64-bit fixes for
1936 <filename>contrib/intagg</></para></listitem>
1937 <listitem><para>Prevent incorrect optimization of functions returning
1938 <type>RECORD</></para></listitem>
1939 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
1940 <listitem><para>Fix Borland makefile for libpq</para></listitem>
1941 <listitem><para>Fix <filename>contrib/btree_gist</> for <type>timetz</> type
1942 (Teodor)</para></listitem>
1943 <listitem><para>Make <command>pg_ctl</> check the PID found in
1944 <filename>postmaster.pid</> to see if it is still a live
1945 process</para></listitem>
1946 <listitem><para>Fix <command>pg_dump</>/<command>pg_restore</> problems caused
1947 by addition of dump timestamps</para></listitem>
1948 <listitem><para>Fix interaction between materializing holdable cursors and
1949 firing deferred triggers during transaction commit</para></listitem>
1950 <listitem><para>Fix memory leak in SQL functions returning pass-by-reference
1951 data types</para></listitem>
1957 <sect1 id="release-8-0-2">
1958 <title>Release 8.0.2</title>
1961 <title>Release date</title>
1962 <simpara>2005-04-07</simpara>
1966 This release contains a variety of fixes from 8.0.1.
1970 <title>Migration to version 8.0.2</title>
1973 A dump/restore is not required for those running 8.0.*.
1974 This release updates the major version number of the
1975 <productname>PostgreSQL</productname> libraries, so it might be
1976 necessary to re-link some user applications if they cannot
1977 find the properly-numbered shared library.
1982 <title>Changes</title>
1985 <listitem><para>Increment the major version number of all interface
1986 libraries (Bruce)</para>
1988 This should have been done in 8.0.0. It is required so 7.4.X versions
1989 of PostgreSQL client applications, like <application>psql</>,
1990 can be used on the same machine as 8.0.X applications. This might require
1991 re-linking user applications that use these libraries.
1993 <listitem><para>Add Windows-only <varname>wal_sync_method</> setting of
1994 <option>fsync_writethrough</> (Magnus, Bruce)</para>
1996 This setting causes <productname>PostgreSQL</productname> to write through
1997 any disk-drive write cache when writing to WAL.
1998 This behavior was formerly called <option>fsync</>, but was
1999 renamed because it acts quite differently from <option>fsync</> on other
2003 <listitem><para>Enable the <varname>wal_sync_method</> setting of
2004 <option>open_datasync</> on Windows, and make it the default for that
2005 platform (Magnus, Bruce)</para>
2007 Because the default is no longer <option>fsync_writethrough</>,
2008 data loss is possible during a power failure if the disk drive has
2009 write caching enabled. To turn off the write cache on Windows,
2010 from the <application>Device Manager</>, choose the drive properties,
2011 then <literal>Policies</>.
2014 <listitem><para>New cache management algorithm <acronym>2Q</> replaces
2015 <acronym>ARC</> (Tom)</para>
2017 This was done to avoid a pending US patent on <acronym>ARC</>. The
2018 <acronym>2Q</> code might be a few percentage points slower than
2019 <acronym>ARC</> for some work loads. A better cache management algorithm
2022 <listitem><para>Planner adjustments to improve behavior on freshly-created
2023 tables (Tom)</para></listitem>
2024 <listitem><para>Allow plpgsql to assign to an element of an array that is
2025 initially <literal>NULL</> (Tom)</para>
2027 Formerly the array would remain <literal>NULL</>, but now it becomes a
2028 single-element array. The main SQL engine was changed to handle
2029 <command>UPDATE</> of a null array value this way in 8.0, but the similar
2030 case in plpgsql was overlooked.
2033 <listitem><para>Convert <literal>\r\n</> and <literal>\r</> to <literal>\n</>
2034 in plpython function bodies (Michael Fuhr)</para>
2036 This prevents syntax errors when plpython code is written on a Windows or
2040 <listitem><para>Allow SPI cursors to handle utility commands that return rows,
2041 such as <command>EXPLAIN</> (Tom)</para></listitem>
2042 <listitem><para>Fix <command>CLUSTER</> failure after <command>ALTER TABLE
2043 SET WITHOUT OIDS</> (Tom)</para></listitem>
2044 <listitem><para>Reduce memory usage of <command>ALTER TABLE ADD COLUMN</>
2045 (Neil)</para></listitem>
2046 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</> (Tom)</para></listitem>
2047 <listitem><para>Document the Windows-only <literal>register</> and
2048 <literal>unregister</> options of <application>pg_ctl</> (Magnus)</para></listitem>
2049 <listitem><para>Ensure operations done during backend shutdown are counted by
2050 statistics collector</para>
2052 This is expected to resolve reports of <application>pg_autovacuum</>
2053 not vacuuming the system catalogs often enough — it was not being
2054 told about catalog deletions caused by temporary table removal during
2057 <listitem><para>Change the Windows default for configuration parameter
2058 <varname>log_destination</> to <option>eventlog</> (Magnus)</para>
2060 By default, a server running on Windows will now send log output to the
2061 Windows event logger rather than standard error.
2063 <listitem><para>Make Kerberos authentication work on Windows (Magnus)</para></listitem>
2064 <listitem><para>Allow <command>ALTER DATABASE RENAME</> by superusers
2065 who aren't flagged as having CREATEDB privilege (Tom)</para></listitem>
2066 <listitem><para>Modify WAL log entries for <command>CREATE</> and
2067 <command>DROP DATABASE</> to not specify absolute paths (Tom)</para>
2068 <para>This allows point-in-time recovery on a different machine with possibly
2069 different database location. Note that <command>CREATE TABLESPACE</> still
2070 poses a hazard in such situations.
2072 <listitem><para>Fix crash from a backend exiting with an open transaction
2073 that created a table and opened a cursor on it (Tom)</para></listitem>
2074 <listitem><para>Fix <function>array_map()</> so it can call PL functions
2075 (Tom)</para></listitem>
2076 <listitem><para>Several <filename>contrib/tsearch2</> and
2077 <filename>contrib/btree_gist</> fixes (Teodor)
2079 <listitem><para>Fix crash of some <filename>contrib/pgcrypto</>
2080 functions on some platforms (Marko Kreen)</para></listitem>
2081 <listitem><para>Fix <filename>contrib/intagg</> for 64-bit platforms
2082 (Tom)</para></listitem>
2083 <listitem><para>Fix ecpg bugs in parsing of <command>CREATE</> statement
2084 (Michael)</para></listitem>
2085 <listitem><para>Work around gcc bug on powerpc and amd64 causing problems in
2086 ecpg (Christof Petig)</para></listitem>
2087 <listitem><para>Do not use locale-aware versions of <function>upper()</>,
2088 <function>lower()</>, and <function>initcap()</> when the locale is
2089 <literal>C</> (Bruce)</para>
2091 This allows these functions to work on platforms that generate errors
2092 for non-7-bit data when the locale is <literal>C</>.
2094 <listitem><para>Fix <function>quote_ident()</> to quote names that match keywords (Tom)</para></listitem>
2095 <listitem><para>Fix <function>to_date()</> to behave reasonably when
2096 <literal>CC</> and <literal>YY</> fields are both used (Karel)</para></listitem>
2097 <listitem><para>Prevent <function>to_char(interval)</> from failing
2098 when given a zero-month interval (Tom)</para></listitem>
2099 <listitem><para>Fix wrong week returned by <function>date_trunc('week')</>
2102 <function>date_trunc('week')</>
2103 returned the wrong year for the first few days of January in some years.
2105 <listitem><para>Use the correct default mask length for class <literal>D</>
2106 addresses in <type>INET</> data types (Tom)</para></listitem>
2112 <sect1 id="release-8-0-1">
2113 <title>Release 8.0.1</title>
2116 <title>Release date</title>
2117 <simpara>2005-01-31</simpara>
2121 This release contains a variety of fixes from 8.0.0, including several
2122 security-related issues.
2126 <title>Migration to version 8.0.1</title>
2129 A dump/restore is not required for those running 8.0.0.
2134 <title>Changes</title>
2137 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
2139 On platforms that will automatically execute initialization functions of a
2140 shared library (this includes at least Windows and ELF-based Unixen),
2141 <command>LOAD</> can be used to make the server execute arbitrary code.
2142 Thanks to NGS Software for reporting this.</para></listitem>
2143 <listitem><para>Check that creator of an aggregate function has the right to
2144 execute the specified transition functions</para>
2146 This oversight made it possible to bypass denial of EXECUTE
2147 permission on a function.</para></listitem>
2148 <listitem><para>Fix security and 64-bit issues in
2149 contrib/intagg</para></listitem>
2150 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
2151 Jurka)</para></listitem>
2152 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
2153 many parameters (Neil)</para></listitem>
2154 <listitem><para>Make <command>ALTER TABLE ADD COLUMN</> enforce domain
2155 constraints in all cases</para></listitem>
2156 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
2158 The result of the join was mistakenly supposed to be sorted the same as the
2159 left input. This could not only deliver mis-sorted output to the user, but
2160 in case of nested merge joins could give outright wrong answers.
2162 <listitem><para>Improve planning of grouped aggregate queries</para></listitem>
2163 <listitem><para><command>ROLLBACK TO <replaceable>savepoint</></command>
2164 closes cursors created since the savepoint</para></listitem>
2165 <listitem><para>Fix inadequate backend stack size on Windows</para></listitem>
2166 <listitem><para>Avoid SHGetSpecialFolderPath() on Windows
2167 (Magnus)</para></listitem>
2168 <listitem><para>Fix some problems in running pg_autovacuum as a Windows
2169 service (Dave Page)</para></listitem>
2170 <listitem><para>Multiple minor bug fixes in
2171 pg_dump/pg_restore</para></listitem>
2172 <listitem><para>Fix ecpg segfault with named structs used in
2173 typedefs (Michael)</para></listitem>
2179 <sect1 id="release-8-0">
2180 <title>Release 8.0</title>
2183 <title>Release date</title>
2184 <simpara>2005-01-19</simpara>
2188 <title>Overview</title>
2191 Major changes in this release:
2197 Microsoft Windows Native Server
2202 This is the first <productname>PostgreSQL</productname> release
2203 to run natively on <trademark class=registered>Microsoft Windows</> as
2204 a server. It can run as a <productname>Windows</> service. This
2205 release supports NT-based Windows releases like
2206 <productname>Windows 2000</>, <productname>Windows XP</>, and
2207 <productname>Windows 2003</>. Older releases like
2208 <productname>Windows 95</>, <productname>Windows 98</>, and
2209 <productname>Windows ME</> are not supported because these operating
2210 systems do not have the infrastructure to support
2211 <productname>PostgreSQL</productname>. A separate installer
2212 project has been created to ease installation on
2213 <productname>Windows</> — see <ulink
2214 url="http://www.postgresql.org/ftp/win32/"></ulink>.
2218 Although tested throughout our release cycle, the Windows port
2219 does not have the benefit of years of use in production
2220 environments that <productname>PostgreSQL</productname> has on
2221 Unix platforms. Therefore it should be treated with the same
2222 level of caution as you would a new product.
2226 Previous releases required the Unix emulation toolkit
2227 <productname>Cygwin</> in order to run the server on Windows
2228 operating systems. <productname>PostgreSQL</productname> has
2229 supported native clients on Windows for many years.
2241 Savepoints allow specific parts of a transaction to be aborted
2242 without affecting the remainder of the transaction. Prior
2243 releases had no such capability; there was no way to recover
2244 from a statement failure within a transaction except by
2245 aborting the whole transaction. This feature is valuable for
2246 application writers who require error recovery within a
2247 complex transaction.
2254 Point-In-Time Recovery
2259 In previous releases there was no way to recover from disk
2260 drive failure except to restore from a previous backup or use
2261 a standby replication server. Point-in-time recovery allows
2262 continuous backup of the server. You can recover either to
2263 the point of failure or to some transaction in the past.
2275 Tablespaces allow administrators to select different file systems
2276 for storage of individual tables, indexes, and databases.
2277 This improves performance and control over disk space
2278 usage. Prior releases used <application>initlocation</> and
2279 manual symlink management for such tasks.
2286 Improved Buffer Management, <command>CHECKPOINT</command>,
2287 <command>VACUUM</command>
2292 This release has a more intelligent buffer replacement strategy,
2293 which will make better use of available shared buffers and
2294 improve performance. The performance impact of vacuum and
2295 checkpoints is also lessened.
2307 A column's data type can now be changed with <command>ALTER
2315 New Perl Server-Side Language
2320 A new version of the <application>plperl</> server-side language now
2321 supports a persistent shared storage area, triggers, returning records
2322 and arrays of records, and SPI calls to access the database.
2329 Comma-separated-value (CSV) support in <command>COPY</command>
2334 <command>COPY</command> can now read and write
2335 comma-separated-value files. It has the flexibility to
2336 interpret non-standard quoting and separation characters too.
2345 <title>Migration to version 8.0</title>
2348 A dump/restore using <application>pg_dump</application> is
2349 required for those wishing to migrate data from any previous
2354 Observe the following incompatibilities:
2361 In <option>READ COMMITTED</> serialization mode, volatile functions
2362 now see the results of concurrent transactions committed up to the
2363 beginning of each statement within the function, rather than up to the
2364 beginning of the interactive command that called the function.
2370 Functions declared <option>STABLE</> or <option>IMMUTABLE</> always
2371 use the snapshot of the calling query, and therefore do not see the
2372 effects of actions taken after the calling query starts, whether in
2373 their own transaction or other transactions. Such a function must be
2374 read-only, too, meaning that it cannot use any SQL commands other than
2381 Non-deferred <option>AFTER</> triggers are now fired immediately
2382 after completion of the triggering query, rather than upon
2383 finishing the current interactive command. This makes a
2384 difference when the triggering query occurred within a function:
2385 the trigger is invoked before the function proceeds to its next
2392 Server configuration parameters <varname>virtual_host</> and
2393 <varname>tcpip_socket</> have been replaced with a more general
2394 parameter <varname>listen_addresses</>. Also, the server now listens on
2395 <literal>localhost</> by default, which eliminates the need for the
2396 <literal>-i</> postmaster switch in many scenarios.
2402 Server configuration parameters <varname>SortMem</> and
2403 <varname>VacuumMem</> have been renamed to <varname>work_mem</>
2404 and <varname>maintenance_work_mem</> to better reflect their
2405 use. The original names are still supported in
2406 <command>SET</command> and <command>SHOW</command>.
2412 Server configuration parameters <varname>log_pid</>,
2413 <varname>log_timestamp</>, and <varname>log_source_port</> have been
2414 replaced with a more general parameter <varname>log_line_prefix</>.
2420 Server configuration parameter <varname>syslog</> has been
2421 replaced with a more logical <varname>log_destination</> variable to
2422 control the log output destination.
2428 Server configuration parameter <varname>log_statement</> has been
2429 changed so it can selectively log just database modification or
2430 data definition statements. Server configuration parameter
2431 <varname>log_duration</> now prints only when <varname>log_statement</>
2438 Server configuration parameter <varname>max_expr_depth</> parameter has
2439 been replaced with <varname>max_stack_depth</> which measures the
2440 physical stack size rather than the expression nesting depth. This
2441 helps prevent session termination due to stack overflow caused by
2442 recursive functions.
2448 The <function>length()</> function no longer counts trailing spaces in
2449 <type>CHAR(n)</> values.
2455 Casting an integer to <type>BIT(N)</> selects the rightmost N bits of the
2456 integer, not the leftmost N bits as before.
2462 Updating an element or slice of a NULL array value now produces
2463 a non-NULL array result, namely an array containing
2464 just the assigned-to positions.
2470 Syntax checking of array input values has been tightened up
2471 considerably. Junk that was previously allowed in odd places with
2472 odd results now causes an error. Empty-string element values
2473 must now be written as <literal>""</>, rather than writing nothing.
2474 Also changed behavior with respect to whitespace surrounding
2475 array elements: trailing whitespace is now ignored, for symmetry
2476 with leading whitespace (which has always been ignored).
2482 Overflow in integer arithmetic operations is now detected and
2483 reported as an error.
2489 The arithmetic operators associated with the single-byte
2490 <type>"char"</> data type have been removed.
2496 The <function>extract()</> function (also called
2497 <function>date_part</>) now returns the proper year for BC dates.
2498 It previously returned one less than the correct year. The
2499 function now also returns the proper values for millennium and
2506 <type>CIDR</> values now must have their non-masked bits be zero.
2507 For example, we no longer allow
2508 <literal>204.248.199.1/31</literal> as a <type>CIDR</> value. Such
2509 values should never have been accepted by
2510 <productname>PostgreSQL</productname> and will now be rejected.
2516 <command>EXECUTE</command> now returns a completion tag that
2517 matches the executed statement.
2523 <application>psql</>'s <command>\copy</> command now reads or
2524 writes to the query's <literal>stdin/stdout</>, rather than
2525 <application>psql</>'s <literal>stdin/stdout</>. The previous
2526 behavior can be accessed via new
2527 <option>pstdin</>/<option>pstdout</> parameters.
2533 The JDBC client interface has been removed from the core
2534 distribution, and is now hosted at <ulink url=
2535 "http://jdbc.postgresql.org"></ulink>.
2541 The Tcl client interface has also been removed. There are several
2542 Tcl interfaces now hosted at <ulink url=
2543 "http://gborg.postgresql.org"></ulink>.
2549 The server now uses its own time zone database, rather than the
2550 one supplied by the operating system. This will provide consistent
2551 behavior across all platforms. In most cases, there should be
2552 little noticeable difference in time zone behavior, except that
2553 the time zone names used by <command>SET</>/<command>SHOW</>
2554 <varname>TimeZone</> may
2555 be different from what your platform provides.
2561 <application>Configure</>'s threading option no longer requires
2562 users to run tests or edit configuration files; threading options
2563 are now detected automatically.
2569 Now that tablespaces have been implemented,
2570 <application>initlocation</> has been removed.
2576 The API for user-defined GiST indexes has been changed. The
2577 Union and PickSplit methods are now passed a pointer to a
2578 special <structname>GistEntryVector</structname> structure,
2579 rather than a <type>bytea</type>.
2587 <title>Deprecated Features</title>
2590 Some aspects of <productname>PostgreSQL</productname>'s behavior
2591 have been determined to be suboptimal. For the sake of backward
2592 compatibility these have not been removed in 8.0, but they are
2593 considered deprecated and will be removed in the next major
2600 The 8.1 release will remove the <function>to_char()</> function
2607 The server now warns of empty strings passed to
2608 <type>oid</type>/<type>float4</type>/<type>float8</type> data
2609 types, but continues to interpret them as zeroes as before.
2610 In the next major release, empty strings will be considered
2611 invalid input for these data types.
2617 By default, tables in <productname>PostgreSQL</productname> 8.0
2618 and earlier are created with <type>OID</>s. In the next release,
2619 this will <emphasis>not</emphasis> be the case: to create a table
2620 that contains <type>OID</>s, the <option>WITH OIDS</> clause must
2621 be specified or the <varname>default_with_oids</varname>
2622 configuration parameter must be set. Users are encouraged to
2623 explicitly specify <option>WITH OIDS</> if their tables
2624 require OIDs for compatibility with future releases of
2625 <productname>PostgreSQL</productname>.
2633 <title>Changes</title>
2636 Below you will find a detailed account of the changes between
2637 release 8.0 and the previous major release.
2641 <title>Performance Improvements</title>
2646 Support cross-data-type index usage (Tom)
2649 Before this change, many queries would not use an index if the data
2650 types did not match exactly. This improvement makes index usage more
2651 intuitive and consistent.
2657 New buffer replacement strategy that improves caching (Jan)
2660 Prior releases used a least-recently-used (LRU) cache to keep
2661 recently referenced pages in memory. The LRU algorithm
2662 did not consider the number of times a specific cache entry was
2663 accessed, so large table scans could force out useful cache pages.
2664 The new cache algorithm uses four separate lists to track most
2665 recently used and most frequently used cache pages and dynamically
2666 optimize their replacement based on the work load. This should
2667 lead to much more efficient use of the shared buffer cache.
2668 Administrators who have tested shared buffer sizes in the past
2669 should retest with this new cache replacement policy.
2675 Add subprocess to write dirty buffers periodically to reduce
2676 checkpoint writes (Jan)
2679 In previous releases, the checkpoint process, which runs every few
2680 minutes, would write all dirty buffers to the operating system's
2681 buffer cache then flush all dirty operating system buffers to
2682 disk. This resulted in a periodic spike in disk usage that often
2683 hurt performance. The new code uses a background writer to trickle
2684 disk writes at a steady pace so checkpoints have far fewer dirty
2685 pages to write to disk. Also, the new code does not issue a global
2686 <function>sync()</> call, but instead <function>fsync()</>s just
2687 the files written since the last checkpoint. This should improve
2688 performance and minimize degradation during checkpoints.
2694 Add ability to prolong vacuum to reduce performance impact (Jan)
2697 On busy systems, <command>VACUUM</command> performs many I/O
2698 requests which can hurt performance for other users. This
2699 release allows you to slow down <command>VACUUM</command> to
2700 reduce its impact on other users, though this increases the
2701 total duration of <command>VACUUM</command>.
2707 Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
2710 This improves the way indexes are scanned when many duplicate
2711 values exist in the index.
2717 Use dynamically-generated table size estimates while planning (Tom)
2720 Formerly the planner estimated table sizes using the values seen
2721 by the last <command>VACUUM</command> or <command>ANALYZE</command>,
2722 both as to physical table size (number of pages) and number of rows.
2723 Now, the current physical table size is obtained from the kernel,
2724 and the number of rows is estimated by multiplying the table size
2725 by the row density (rows per page) seen by the last
2726 <command>VACUUM</command> or <command>ANALYZE</command>. This should
2727 produce more reliable estimates in cases where the table size has
2728 changed significantly since the last housekeeping command.
2734 Improved index usage with <literal>OR</> clauses (Tom)
2737 This allows the optimizer to use indexes in statements with many OR
2738 clauses that would not have been indexed in the past. It can also use
2739 multi-column indexes where the first column is specified and the second
2740 column is part of an <literal>OR</> clause.
2746 Improve matching of partial index clauses (Tom)
2749 The server is now smarter about using partial indexes in queries
2750 involving complex <option>WHERE</> clauses.
2756 Improve performance of the GEQO optimizer (Tom)
2759 The GEQO optimizer is used to plan queries involving many tables (by
2760 default, twelve or more). This release speeds up the way queries are
2761 analyzed to decrease time spent in optimization.
2767 Miscellaneous optimizer improvements
2770 There is not room here to list all the minor improvements made, but
2771 numerous special cases work better than in prior releases.
2777 Improve lookup speed for C functions (Tom)
2780 This release uses a hash table to lookup information for dynamically
2781 loaded C functions. This improves their speed so they perform nearly as
2782 quickly as functions that are built into the server executable.
2788 Add type-specific <command>ANALYZE</command> statistics
2789 capability (Mark Cave-Ayland)
2792 This feature allows more flexibility in generating statistics
2793 for non-standard data types.
2799 <command>ANALYZE</command> now collects statistics for
2800 expression indexes (Tom)
2803 Expression indexes (also called functional indexes) allow users to
2804 index not just columns but the results of expressions and function
2805 calls. With this release, the optimizer can gather and use statistics
2806 about the contents of expression indexes. This will greatly improve
2807 the quality of planning for queries in which an expression index is
2814 New two-stage sampling method for <command>ANALYZE</command>
2818 This gives better statistics when the density of valid rows is very
2819 different in different regions of a table.
2825 Speed up <command>TRUNCATE</command> (Tom)
2828 This buys back some of the performance loss observed in 7.4, while still
2829 keeping <command>TRUNCATE</command> transaction-safe.
2838 <title>Server Changes</title>
2843 Add WAL file archiving and point-in-time recovery (Simon Riggs)
2849 Add tablespaces so admins can control disk layout (Gavin)
2855 Add a built-in log rotation program (Andreas Pflug)
2858 It is now possible to log server messages conveniently without
2859 relying on either <application>syslog</> or an external log
2866 Add new read-only server configuration parameters to show server
2867 compile-time settings: <varname>block_size</>,
2868 <varname>integer_datetimes</>, <varname>max_function_args</>,
2869 <varname>max_identifier_length</>, <varname>max_index_keys</> (Joe)
2875 Make quoting of <literal>sameuser</>, <literal>samegroup</>, and
2876 <literal>all</> remove special meaning of these terms in
2877 <filename>pg_hba.conf</> (Andrew)
2883 Use clearer IPv6 name <literal>::1/128</> for
2884 <literal>localhost</> in default <filename>pg_hba.conf</> (Andrew)
2890 Use CIDR format in <filename>pg_hba.conf</> examples (Andrew)
2896 Rename server configuration parameters <varname>SortMem</> and
2897 <varname>VacuumMem</> to <varname>work_mem</> and
2898 <varname>maintenance_work_mem</> (Old names still supported) (Tom)
2901 This change was made to clarify that bulk operations such as index and
2902 foreign key creation use <varname>maintenance_work_mem</>, while
2903 <varname>work_mem</> is for workspaces used during query execution.
2909 Allow logging of session disconnections using server configuration
2910 <varname>log_disconnections</> (Andrew)
2916 Add new server configuration parameter <varname>log_line_prefix</> to
2917 allow control of information emitted in each log line (Andrew)
2920 Available information includes user name, database name, remote IP
2921 address, and session start time.
2927 Remove server configuration parameters <varname>log_pid</>,
2928 <varname>log_timestamp</>, <varname>log_source_port</>; functionality
2929 superseded by <varname>log_line_prefix</> (Andrew)
2935 Replace the <varname>virtual_host</> and <varname>tcpip_socket</>
2936 parameters with a unified <varname>listen_addresses</> parameter
2940 <varname>virtual_host</> could only specify a single IP address to
2941 listen on. <varname>listen_addresses</> allows multiple addresses
2948 Listen on localhost by default, which eliminates the need for the
2949 <option>-i</> postmaster switch in many scenarios (Andrew)
2952 Listening on localhost (<literal>127.0.0.1</>) opens no new
2953 security holes but allows configurations like Windows and JDBC,
2954 which do not support local sockets, to work without special
2961 Remove <varname>syslog</> server configuration parameter, and add more
2962 logical <varname>log_destination</> variable to control log output
2969 Change server configuration parameter <varname>log_statement</> to take
2970 values <varname>all</>, <varname>mod</>, <varname>ddl</>, or
2971 <varname>none</> to select which queries are logged (Bruce)
2974 This allows administrators to log only data definition changes or
2975 only data modification statements.
2981 Some logging-related configuration parameters could formerly be adjusted
2982 by ordinary users, but only in the <quote>more verbose</> direction.
2983 They are now treated more strictly: only superusers can set them.
2984 However, a superuser may use <command>ALTER USER</> to provide per-user
2985 settings of these values for non-superusers. Also, it is now possible
2986 for superusers to set values of superuser-only configuration parameters
2987 via <literal>PGOPTIONS</>.
2993 Allow configuration files to be placed outside the data directory (mlw)
2996 By default, configuration files are kept in the cluster's top directory.
2997 With this addition, configuration files can be placed outside the
2998 data directory, easing administration.
3004 Plan prepared queries only when first executed so constants can be
3005 used for statistics (Oliver Jowett)
3008 Prepared statements plan queries once and execute them many
3009 times. While prepared queries avoid the overhead of re-planning
3010 on each use, the quality of the plan suffers from not knowing the exact
3011 parameters to be used in the query. In this release, planning of
3012 unnamed prepared statements is delayed until the first execution,
3013 and the actual parameter values of that execution are used as
3014 optimization hints. This allows use of out-of-line parameter passing
3015 without incurring a performance penalty.
3021 Allow <command>DECLARE CURSOR</command> to take parameters
3025 It is now useful to issue <command>DECLARE CURSOR</command> in a
3026 <function>Parse</> message with parameters. The parameter values
3027 sent at <function>Bind</> time will be substituted into the
3028 execution of the cursor's query.
3034 Fix hash joins and aggregates of <type>inet</type> and
3035 <type>cidr</type> data types (Tom)
3038 Release 7.4 handled hashing of mixed <type>inet</type> and
3039 <type>cidr</type> values incorrectly. (This bug did not exist
3040 in prior releases because they wouldn't try to hash either
3047 Make <varname>log_duration</> print only when <varname>log_statement</>
3048 prints the query (Ed L.)
3057 <title>Query Changes</title>
3062 Add savepoints (nested transactions) (Alvaro)
3068 Unsupported isolation levels are now accepted and promoted to the
3069 nearest supported level (Peter)
3072 The SQL specification states that if a database doesn't support a
3073 specific isolation level, it should use the next more restrictive level.
3074 This change complies with that recommendation.
3080 Allow <command>BEGIN WORK</command> to specify transaction
3081 isolation levels like <command>START TRANSACTION</command> does
3088 Fix table permission checking for cases in which rules generate
3089 a query type different from the originally submitted query (Tom)
3095 Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
3099 In previous releases, because single quotes had to be used to
3100 quote a function's body, the use of single quotes inside the
3101 function text required use of two single quotes or other error-prone
3102 notations. With this release we add the ability to use "dollar
3103 quoting" to quote a block of text. The ability to use different
3104 quoting delimiters at different nesting levels greatly simplifies
3105 the task of quoting correctly, especially in complex functions.
3106 Dollar quoting can be used anywhere quoted text is needed.
3112 Make <literal>CASE val WHEN compval1 THEN ...</> evaluate <literal>val</> only once (Tom)
3115 <option>CASE</> no longer evaluates the tested expression multiple
3116 times. This has benefits when the expression is complex or is
3123 Test <option>HAVING</> before computing target list of an
3124 aggregate query (Tom)
3127 Fixes improper failure of cases such as <literal>SELECT SUM(win)/SUM(lose)
3128 ... GROUP BY ... HAVING SUM(lose) > 0</>. This should work but formerly
3129 could fail with divide-by-zero.
3135 Replace <varname>max_expr_depth</> parameter with
3136 <varname>max_stack_depth</> parameter, measured in kilobytes of stack
3140 This gives us a fairly bulletproof defense against crashing due to
3141 runaway recursive functions. Instead of measuring the depth of expression
3142 nesting, we now directly measure the size of the execution stack.
3148 Allow arbitrary row expressions (Tom)
3151 This release allows SQL expressions to contain arbitrary composite
3152 types, that is, row values. It also allows functions to more easily
3153 take rows as arguments and return row values.
3159 Allow <option>LIKE</>/<option>ILIKE</> to be used as the operator
3160 in row and subselect comparisons (Fabien Coelho)
3166 Avoid locale-specific case conversion of basic ASCII letters in
3167 identifiers and keywords (Tom)
3170 This solves the <quote>Turkish problem</> with mangling of words
3171 containing <literal>I</> and <literal>i</>. Folding of characters
3172 outside the 7-bit-ASCII set is still locale-aware.
3178 Improve syntax error reporting (Fabien, Tom)
3181 Syntax error reports are more useful than before.
3187 Change <command>EXECUTE</command> to return a completion tag
3188 matching the executed statement (Kris Jurka)
3191 Previous releases return an <command>EXECUTE</command> tag for
3192 any <command>EXECUTE</command> call. In this release, the tag
3193 returned will reflect the command executed.
3199 Avoid emitting <option>NATURAL CROSS JOIN</> in rule listings (Tom)
3202 Such a clause makes no logical sense, but in some cases the rule
3203 decompiler formerly produced this syntax.
3212 <title>Object Manipulation Changes</title>
3217 Add <command>COMMENT ON</> for casts, conversions, languages,
3218 operator classes, and large objects (Christopher)
3224 Add new server configuration parameter <varname>default_with_oids</> to
3225 control whether tables are created with <type>OID</>s by default (Neil)
3228 This allows administrators to control whether <command>CREATE
3229 TABLE</command> commands create tables with or without <type>OID</>
3230 columns by default. (Note: the current factory default setting for
3231 <varname>default_with_oids</> is <literal>TRUE</>, but the default
3232 will become <literal>FALSE</> in future releases.)
3238 Add <option>WITH</> / <option>WITHOUT OIDS</> clause to
3239 <command>CREATE TABLE AS</command> (Neil)
3245 Allow <command>ALTER TABLE DROP COLUMN</> to drop an <type>OID</>
3246 column (<command>ALTER TABLE SET WITHOUT OIDS</> still works)
3253 Allow composite types as table columns (Tom)
3259 Allow <command>ALTER ... ADD COLUMN</> with defaults and
3260 <option>NOT NULL</> constraints; works per SQL spec (Rod)
3263 It is now possible for <option>ADD COLUMN</> to create a column
3264 that is not initially filled with NULLs, but with a specified
3271 Add <command>ALTER COLUMN TYPE</> to change column's type (Rod)
3274 It is now possible to alter a column's data type without dropping
3275 and re-adding the column.
3281 Allow multiple <command>ALTER</> actions in a single <command>ALTER
3282 TABLE</command> command (Rod)
3285 This is particularly useful for <command>ALTER</> commands that
3286 rewrite the table (which include <option>ALTER COLUMN TYPE</> and
3287 <option>ADD COLUMN</> with a default). By grouping
3288 <command>ALTER</> commands together, the table need be rewritten
3295 Allow <command>ALTER TABLE</command> to add <type>SERIAL</type>
3299 This falls out from the new capability of specifying defaults for new
3306 Allow changing the owners of aggregates, conversions, databases,
3307 functions, operators, operator classes, schemas, types, and tablespaces
3308 (Christopher, Euler Taveira de Oliveira)
3311 Previously this required modifying the system tables directly.
3317 Allow temporary object creation to be limited to <option>SECURITY
3318 DEFINER</> functions (Sean Chittenden)
3324 Add <option>ALTER TABLE ... SET WITHOUT CLUSTER</> (Christopher)
3327 Prior to this release, there was no way to clear an auto-cluster
3328 specification except to modify the system tables.
3334 Constraint/Index/<type>SERIAL</> names are now
3335 <replaceable>table_column_type</>
3336 with numbers appended to guarantee uniqueness within the schema
3340 The SQL specification states that such names should be unique
3347 Add <function>pg_get_serial_sequence()</> to return a
3348 <type>SERIAL</> column's sequence name (Christopher)
3351 This allows automated scripts to reliably find the <type>SERIAL</>
3358 Warn when primary/foreign key data type mismatch requires costly lookup
3364 New <command>ALTER INDEX</> command to allow moving of indexes
3365 between tablespaces (Gavin)
3371 Make <command>ALTER TABLE OWNER</> change dependent sequence
3372 ownership too (Alvaro)
3382 <title>Utility Command Changes</title>
3387 Allow <command>CREATE SCHEMA</command> to create triggers,
3388 indexes, and sequences (Neil)
3394 Add <option>ALSO</> keyword to <command>CREATE RULE</command> (Fabien
3398 This allows <option>ALSO</> to be added to rule creation to contrast it with
3399 <option>INSTEAD</> rules.
3405 Add <option>NOWAIT</> option to <command>LOCK</command> (Tatsuo)
3408 This allows the <command>LOCK</command> command to fail if it
3409 would have to wait for the requested lock.
3415 Allow <command>COPY</command> to read and write
3416 comma-separated-value (CSV) files (Andrew, Bruce)
3422 Generate error if the <command>COPY</command> delimiter and NULL
3423 string conflict (Bruce)
3429 <command>GRANT</command>/<command>REVOKE</command> behavior
3430 follows the SQL spec more closely
3436 Avoid locking conflict between <command>CREATE INDEX</command>
3437 and <command>CHECKPOINT</command> (Tom)
3440 In 7.3 and 7.4, a long-running B-tree index build could block concurrent
3441 <command>CHECKPOINT</>s from completing, thereby causing WAL bloat because the
3442 WAL log could not be recycled.
3448 Database-wide <command>ANALYZE</command> does not hold locks
3452 This reduces the potential for deadlocks against other backends
3453 that want exclusive locks on tables. To get the benefit of this
3454 change, do not execute database-wide <command>ANALYZE</command>
3455 inside a transaction block (<command>BEGIN</command> block); it
3456 must be able to commit and start a new transaction for each
3463 <command>REINDEX</command> does not exclusively lock the index's
3464 parent table anymore
3467 The index itself is still exclusively locked, but readers of the
3468 table can continue if they are not using the particular index
3475 Erase MD5 user passwords when a user is renamed (Bruce)
3478 <productname>PostgreSQL</productname> uses the user name as salt
3479 when encrypting passwords via MD5. When a user's name is changed,
3480 the salt will no longer match the stored MD5 password, so the
3481 stored password becomes useless. In this release a notice is
3482 generated and the password is cleared. A new password must then
3483 be assigned if the user is to be able to log in with a password.
3489 New <application>pg_ctl</> <option>kill</> option for Windows (Andrew)
3492 Windows does not have a <literal>kill</> command to send signals to
3493 backends so this capability was added to <application>pg_ctl</>.
3499 Information schema improvements
3505 Add <option>--pwfile</> option to
3506 <application>initdb</application> so the initial password can be
3507 set by GUI tools (Magnus)
3513 Detect locale/encoding mismatch in
3514 <application>initdb</application> (Peter)
3520 Add <option>register</> command to <application>pg_ctl</> to
3521 register Windows operating system service (Dave Page)
3530 <title>Data Type and Function Changes</title>
3535 More complete support for composite types (row types) (Tom)
3538 Composite values can be used in many places where only scalar values
3545 Reject non-rectangular array values as erroneous (Joe)
3548 Formerly, <function>array_in</> would silently build a
3555 Overflow in integer arithmetic operations is now detected (Tom)
3561 The arithmetic operators associated with the single-byte
3562 <type>"char"</> data type have been removed.
3565 Formerly, the parser would select these operators in many situations
3566 where an <quote>unable to select an operator</> error would be more
3567 appropriate, such as <literal>null * null</>. If you actually want
3568 to do arithmetic on a <type>"char"</> column, you can cast it to
3575 Syntax checking of array input values considerably tightened up (Joe)
3578 Junk that was previously allowed in odd places with odd results
3579 now causes an <literal>ERROR</>, for example, non-whitespace
3580 after the closing right brace.
3586 Empty-string array element values must now be written as
3587 <literal>""</>, rather than writing nothing (Joe)
3590 Formerly, both ways of writing an empty-string element value were
3591 allowed, but now a quoted empty string is required. The case where
3592 nothing at all appears will probably be considered to be a NULL
3593 element value in some future release.
3599 Array element trailing whitespace is now ignored (Joe)
3602 Formerly leading whitespace was ignored, but trailing whitespace
3603 between an element value and the delimiter or right brace was
3604 significant. Now trailing whitespace is also ignored.
3610 Emit array values with explicit array bounds when lower bound is not one
3617 Accept <literal>YYYY-monthname-DD</> as a date string (Tom)
3623 Make <function>netmask</> and <function>hostmask</> functions
3624 return maximum-length mask length (Tom)
3630 Change factorial function to return <type>numeric</type> (Gavin)
3633 Returning <type>numeric</type> allows the factorial function to
3634 work for a wider range of input values.
3640 <function>to_char</>/<function>to_date()</> date conversion
3641 improvements (Kurt Roeckx, Fabien Coelho)
3647 Make <function>length()</> disregard trailing spaces in
3648 <type>CHAR(n)</> (Gavin)
3651 This change was made to improve consistency: trailing spaces are
3652 semantically insignificant in <type>CHAR(n)</> data, so they
3653 should not be counted by <function>length()</>.
3659 Warn about empty string being passed to
3660 <type>OID</>/<type>float4</>/<type>float8</> data types (Neil)
3663 8.1 will throw an error instead.
3669 Allow leading or trailing whitespace in
3670 <type>int2</>/<type>int4</>/<type>int8</>/<type>float4</>/<type>float8</>
3678 Better support for IEEE <literal>Infinity</> and <literal>NaN</>
3679 values in <type>float4</type>/<type>float8</type> (Neil)
3682 These should now work on all platforms that support IEEE-compliant
3683 floating point arithmetic.
3689 Add <option>week</> option to <function>date_trunc()</> (Robert Creager)
3695 Fix <function>to_char</function> for <literal>1 BC</>
3696 (previously it returned <literal>1 AD</>) (Bruce)
3702 Fix <function>date_part(year)</> for BC dates (previously it
3703 returned one less than the correct year) (Bruce)
3709 Fix <function>date_part()</> to return the proper millennium and
3710 century (Fabien Coelho)
3713 In previous versions, the century and millennium results had a wrong
3714 number and started in the wrong year, as compared to standard
3715 reckoning of such things.
3721 Add <function>ceiling()</> as an alias for <function>ceil()</>,
3722 and <function>power()</> as an alias for <function>pow()</> for
3723 standards compliance (Neil)
3729 Change <function>ln()</>, <function>log()</>,
3730 <function>power()</>, and <function>sqrt()</> to emit the correct
3731 <literal>SQLSTATE</> error codes for certain error conditions, as
3732 specified by SQL:2003 (Neil)
3738 Add <function>width_bucket()</> function as defined by SQL:2003 (Neil)
3744 Add <function>generate_series()</> functions to simplify working
3745 with numeric sets (Joe)
3751 Fix <function>upper/lower/initcap()</> functions to work with
3752 multibyte encodings (Tom)
3758 Add boolean and bitwise integer <option>AND</>/<option>OR</>
3759 aggregates (Fabien Coelho)
3765 New session information functions to return network addresses for client
3766 and server (Sean Chittenden)
3772 Add function to determine the area of a closed path (Sean Chittenden)
3778 Add function to send cancel request to other backends (Magnus)
3784 Add <type>interval</> plus <type>datetime</> operators (Tom)
3787 The reverse ordering, <type>datetime</> plus <type>interval</>,
3788 was already supported, but both are required by the SQL standard.
3794 Casting an integer to <type>BIT(N)</> selects the rightmost N bits
3799 In prior releases, the leftmost N bits were selected, but this was
3800 deemed unhelpful, not to mention inconsistent with casting from bit
3807 Require <type>CIDR</> values to have all non-masked bits be zero
3817 <title>Server-Side Language Changes</title>
3822 In <literal>READ COMMITTED</> serialization mode, volatile functions
3823 now see the results of concurrent transactions committed up to the
3824 beginning of each statement within the function, rather than up to the
3825 beginning of the interactive command that called the function.
3831 Functions declared <literal>STABLE</> or <literal>IMMUTABLE</> always
3832 use the snapshot of the calling query, and therefore do not see the
3833 effects of actions taken after the calling query starts, whether in
3834 their own transaction or other transactions. Such a function must be
3835 read-only, too, meaning that it cannot use any SQL commands other than
3836 <command>SELECT</>. There is a considerable performance gain from
3837 declaring a function <literal>STABLE</> or <literal>IMMUTABLE</>
3838 rather than <literal>VOLATILE</>.
3844 Non-deferred <option>AFTER</> triggers are now fired immediately
3845 after completion of the triggering query, rather than upon
3846 finishing the current interactive command. This makes a difference
3847 when the triggering query occurred within a function: the trigger
3848 is invoked before the function proceeds to its next operation. For
3849 example, if a function inserts a new row into a table, any
3850 non-deferred foreign key checks occur before proceeding with the
3857 Allow function parameters to be declared with names (Dennis Bjorklund)
3860 This allows better documentation of functions. Whether the names
3861 actually do anything depends on the specific function language
3868 Allow PL/pgSQL parameter names to be referenced in the function (Dennis Bjorklund)
3871 This basically creates an automatic alias for each named parameter.
3877 Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
3880 This allows us to catch simple syntax errors sooner.
3886 More support for composite types (row and record variables) in PL/pgSQL
3889 For example, it now works to pass a rowtype variable to another function
3890 as a single variable.
3896 Default values for PL/pgSQL variables can now reference previously
3903 Improve parsing of PL/pgSQL FOR loops (Tom)
3906 Parsing is now driven by presence of <literal>".."</> rather than
3907 data type of <option>FOR</> variable. This makes no difference for
3908 correct functions, but should result in more understandable error
3909 messages when a mistake is made.
3915 Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
3921 In PL/Tcl, SPI commands are now run in subtransactions. If an error
3922 occurs, the subtransaction is cleaned up and the error is reported
3923 as an ordinary Tcl error, which can be trapped with <literal>catch</>.
3924 Formerly, it was not possible to catch such errors.
3930 Accept <command>ELSEIF</> in PL/pgSQL (Neil)
3933 Previously PL/pgSQL only allowed <command>ELSIF</>, but many people
3934 are accustomed to spelling this keyword <command>ELSEIF</>.
3943 <title><application>psql</> Changes</title>
3948 Improve <application>psql</> information display about database
3949 objects (Christopher)
3955 Allow <application>psql</> to display group membership in
3956 <command>\du</> and <command>\dg</> (Markus Bertheau)
3962 Prevent <application>psql</> <command>\dn</command> from showing
3963 temporary schemas (Bruce)
3969 Allow <application>psql</> to handle tilde user expansion for file
3976 Allow <application>psql</> to display fancy prompts, including
3977 color, via <application>readline</> (Reece Hart, Chet Ramey)
3983 Make <application>psql</> <command>\copy</> match <command>COPY</command> command syntax
3990 Show the location of syntax errors (Fabien Coelho, Tom)
3996 Add <command>CLUSTER</command> information to <application>psql</>
3997 <command>\d</> display
4004 Change <application>psql</> <command>\copy stdin/stdout</> to read
4005 from command input/output (Bruce)
4011 Add <option>pstdin</>/<option>pstdout</> to read from
4012 <application>psql</>'s <literal>stdin</>/<literal>stdout</> (Mark
4019 Add global <application>psql</> configuration file, <filename>psqlrc.sample</filename>
4023 This allows a central file where global <application>psql</> startup commands can
4030 Have <application>psql</> <command>\d+</> indicate if the table
4031 has an <type>OID</> column (Neil)
4037 On Windows, use binary mode in <application>psql</> when reading files so control-Z
4038 is not seen as end-of-file
4044 Have <command>\dn+</> show permissions and description for schemas (Dennis
4051 Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
4057 Allow boolean settings to be set using upper or lower case (Michael Paesold)
4066 <title><application>pg_dump</> Changes</title>
4071 Use dependency information to improve the reliability of
4072 <application>pg_dump</> (Tom)
4075 This should solve the longstanding problems with related objects
4076 sometimes being dumped in the wrong order.
4082 Have <application>pg_dump</> output objects in alphabetical order if possible (Tom)
4085 This should make it easier to identify changes between
4092 Allow <application>pg_restore</> to ignore some SQL errors (Fabien Coelho)
4095 This makes <application>pg_restore</>'s behavior similar to the
4096 results of feeding a <application>pg_dump</> output script to
4097 <application>psql</>. In most cases, ignoring errors and plowing
4098 ahead is the most useful thing to do. Also added was a pg_restore
4099 option to give the old behavior of exiting on an error.
4105 <application>pg_restore</> <option>-l</> display now includes
4106 objects' schema names
4112 New begin/end markers in <application>pg_dump</> text output (Bruce)
4118 Add start/stop times for
4119 <application>pg_dump</>/<application>pg_dumpall</> in verbose mode
4126 Allow most <application>pg_dump</> options in
4127 <application>pg_dumpall</> (Christopher)
4133 Have <application>pg_dump</> use <command>ALTER OWNER</> rather
4134 than <command>SET SESSION AUTHORIZATION</> by default
4144 <title>libpq Changes</title>
4149 Make libpq's <option>SIGPIPE</> handling thread-safe (Bruce)
4155 Add <function>PQmbdsplen()</> which returns the display length
4156 of a character (Tatsuo)
4162 Add thread locking to <application>SSL</> and
4163 <application>Kerberos</> connections (Manfred Spraul)
4169 Allow <function>PQoidValue()</>, <function>PQcmdTuples()</>, and
4170 <function>PQoidStatus()</> to work on <command>EXECUTE</command>
4177 Add <function>PQserverVersion()</> to provide more convenient
4178 access to the server version number (Greg Sabino Mullane)
4184 Add <function>PQprepare/PQsendPrepared()</> functions to support
4185 preparing statements without necessarily specifying the data types
4186 of their parameters (Abhijit Menon-Sen)
4192 Many ECPG improvements, including <command>SET DESCRIPTOR</> (Michael)
4201 <title>Source Code Changes</title>
4206 Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
4212 Shell script commands converted to C versions for Windows support (Andrew)
4218 Create an extension makefile framework (Fabien Coelho, Peter)
4221 This simplifies the task of building extensions outside the original
4228 Support relocatable installations (Bruce)
4231 Directory paths for installed files (such as the
4232 <filename>/share</> directory) are now computed relative to the
4233 actual location of the executables, so that an installation tree
4234 can be moved to another place without reconfiguring and
4241 Use <option>--with-docdir</> to choose installation location of documentation; also
4242 allow <option>--infodir</> (Peter)
4248 Add <option>--without-docdir</> to prevent installation of documentation (Peter)
4254 Upgrade to <application>DocBook</> V4.2 SGML (Peter)
4260 New <literal>PostgreSQL</> <application>CVS</> tag (Marc)
4263 This was done to make it easier for organizations to manage their
4264 own copies of the <productname>PostgreSQL</productname>
4265 <application>CVS</> repository. File version stamps from the master
4266 repository will not get munged by checking into or out of a copied
4273 Clarify locking code (Manfred Koizar)
4279 Buffer manager cleanup (Neil)
4285 Decouple platform tests from CPU spinlock code (Bruce, Tom)
4291 Add inlined test-and-set code on PA-RISC for <application>gcc</>
4298 Improve i386 spinlock code (Manfred Spraul)
4304 Clean up spinlock assembly code to avoid warnings from newer
4305 <application>gcc</> releases (Tom)
4311 Remove JDBC from source tree; now a separate project
4317 Remove the libpgtcl client interface; now a separate project
4323 More accurately estimate memory and file descriptor usage (Tom)
4329 Improvements to the Mac OS X startup scripts (Ray A.)
4335 New <function>fsync()</> test program (Bruce)
4341 Major documentation improvements (Neil, Peter)
4347 Remove <application>pg_encoding</application>; not needed
4354 Remove <application>pg_id</application>; not needed anymore
4360 Remove <application>initlocation</application>; not needed
4367 Auto-detect thread flags (no more manual testing) (Bruce)
4373 Use Olson's public domain <application>timezone</> library (Magnus)
4379 With threading enabled, use thread flags on Unixware for
4380 backend executables too (Bruce)
4383 Unixware cannot mix threaded and non-threaded object files in the
4384 same executable, so everything must be compiled as threaded.
4390 <application>psql</> now uses a <application>flex</>-generated
4391 lexical analyzer to process command strings
4397 Reimplement the linked list data structure used throughout the
4401 This improves performance by allowing list append and length
4402 operations to be more efficient.
4408 Allow dynamically loaded modules to create their own server configuration
4409 parameters (Thomas Hallgren)
4415 New Brazilian version of FAQ (Euler Taveira de Oliveira)
4421 Add French FAQ (Guillaume Lelarge)
4427 New <application>pgevent</> for Windows logging
4433 Make libpq and ECPG build as proper shared libraries on OS X (Tom)
4442 <title>Contrib Changes</title>
4447 Overhaul of <filename>contrib/dblink</> (Joe)
4453 <filename>contrib/dbmirror</> improvements (Steven Singer)
4459 New <filename>contrib/xml2</> (John Gray, Torchbox)
4465 Updated <filename>contrib/mysql</filename>
4471 New version of <filename>contrib/btree_gist</> (Teodor)
4477 New <filename>contrib/trgm</>, trigram matching for
4478 <productname>PostgreSQL</productname> (Teodor)
4484 Many <filename>contrib/tsearch2</> improvements (Teodor)
4490 Add double metaphone to <filename>contrib/fuzzystrmatch</> (Andrew)
4496 Allow <filename>contrib/pg_autovacuum</> to run as a Windows service (Dave Page)
4502 Add functions to <filename>contrib/dbsize</> (Andreas Pflug)
4508 Removed <filename>contrib/pg_logger</>: obsoleted by integrated logging
4515 Removed <filename>contrib/rserv</>: obsoleted by various separate projects
4525 <sect1 id="release-7-4-8">
4526 <title>Release 7.4.8</title>
4529 <title>Release date</title>
4530 <simpara>2005-05-09</simpara>
4534 This release contains a variety of fixes from 7.4.7, including several
4535 security-related issues.
4539 <title>Migration to version 7.4.8</title>
4542 A dump/restore is not required for those running 7.4.X. However,
4543 it is one possible way of handling two significant security problems
4544 that have been found in the initial contents of 7.4.X system
4545 catalogs. A dump/initdb/reload sequence using 7.4.8's initdb will
4546 automatically correct these problems.
4550 The larger security problem is that the built-in character set encoding
4551 conversion functions can be invoked from SQL commands by unprivileged
4552 users, but the functions were not designed for such use and are not
4553 secure against malicious choices of arguments. The fix involves changing
4554 the declared parameter list of these functions so that they can no longer
4555 be invoked from SQL commands. (This does not affect their normal use
4556 by the encoding conversion machinery.)
4560 The lesser problem is that the <filename>contrib/tsearch2</> module
4561 creates several functions that are misdeclared to return
4562 <type>internal</> when they do not accept <type>internal</> arguments.
4563 This breaks type safety for all functions using <type>internal</>
4568 It is strongly recommended that all installations repair these errors,
4569 either by initdb or by following the manual repair procedures given
4570 below. The errors at least allow unprivileged database users to crash
4571 their server process, and may allow unprivileged users to gain the
4572 privileges of a database superuser.
4576 If you wish not to do an initdb, perform the following procedures instead.
4577 As the database superuser, do:
4581 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
4582 WHERE pronamespace = 11 AND pronargs = 5
4583 AND proargtypes[2] = 'cstring'::regtype;
4584 -- The command should report having updated 90 rows;
4585 -- if not, rollback and investigate instead of committing!
4589 Next, if you have installed <filename>contrib/tsearch2</>, do
4593 UPDATE pg_proc SET proargtypes[0] = 'internal'::regtype
4595 'dex_init(text)'::regprocedure,
4596 'snb_en_init(text)'::regprocedure,
4597 'snb_ru_init(text)'::regprocedure,
4598 'spell_init(text)'::regprocedure,
4599 'syn_init(text)'::regprocedure
4601 -- The command should report having updated 5 rows;
4602 -- if not, rollback and investigate instead of committing!
4606 If this command fails with a message like <quote>function
4607 "dex_init(text)" does not exist</>, then either <filename>tsearch2</>
4608 is not installed in this database, or you already did the update.
4612 The above procedures must be carried out in <emphasis>each</> database
4613 of an installation, including <literal>template1</>, and ideally
4614 including <literal>template0</> as well. If you do not fix the
4615 template databases then any subsequently created databases will contain
4616 the same errors. <literal>template1</> can be fixed in the same way
4617 as any other database, but fixing <literal>template0</> requires
4618 additional steps. First, from any database issue
4620 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
4622 Next connect to <literal>template0</> and perform the above repair
4623 procedures. Finally, do
4625 -- re-freeze template0:
4627 -- and protect it against future alterations:
4628 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
4634 <title>Changes</title>
4637 <listitem><para>Change encoding function signature to prevent
4638 misuse</para></listitem>
4639 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
4640 <type>INTERNAL</> function results</para></listitem>
4641 <listitem><para>Repair ancient race condition that allowed a transaction to be
4642 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
4643 than for other purposes</para>
4644 <para>This is an extremely serious bug since it could lead to apparent
4645 data inconsistencies being briefly visible to applications.</para></listitem>
4646 <listitem><para>Repair race condition between relation extension and
4648 <para>This could theoretically have caused loss of a page's worth of
4649 freshly-inserted data, although the scenario seems of very low probability.
4650 There are no known cases of it having caused more than an Assert failure.
4652 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
4654 The comparison code was wrong in the case where the
4655 <literal>--enable-integer-datetimes</> configuration switch had been used.
4656 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
4657 it will need to be <command>REINDEX</>ed after installing this update, because
4658 the fix corrects the sort order of column values.
4660 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
4661 <type>TIME WITH TIME ZONE</> values</para></listitem>
4662 <listitem><para>Fix mis-display of negative fractional seconds in
4663 <type>INTERVAL</> values</para>
4665 This error only occurred when the
4666 <literal>--enable-integer-datetimes</> configuration switch had been used.
4668 <listitem><para>Ensure operations done during backend shutdown are counted by
4669 statistics collector</para>
4671 This is expected to resolve reports of <application>pg_autovacuum</>
4672 not vacuuming the system catalogs often enough — it was not being
4673 told about catalog deletions caused by temporary table removal during
4676 <listitem><para>Additional buffer overrun checks in plpgsql
4677 (Neil)</para></listitem>
4678 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
4679 correctly (Neil)</para></listitem>
4680 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
4681 (Marko Kreen)</para></listitem>
4682 <listitem><para>Still more 64-bit fixes for
4683 <filename>contrib/intagg</></para></listitem>
4684 <listitem><para>Prevent incorrect optimization of functions returning
4685 <type>RECORD</></para></listitem>
4686 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
4687 month-related formats</para></listitem>
4688 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
4689 <listitem><para>Fix <function>array_map</> to call PL functions correctly</para></listitem>
4690 <listitem><para>Fix permission checking in <command>ALTER DATABASE RENAME</></para></listitem>
4691 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</></para></listitem>
4692 <listitem><para>Make <function>RemoveFromWaitQueue</> clean up after itself</para>
4694 This fixes a lock management error that would only be visible if a transaction
4695 was kicked out of a wait for a lock (typically by query cancel) and then the
4696 holder of the lock released it within a very narrow window.
4698 <listitem><para>Fix problem with untyped parameter appearing in
4699 <command>INSERT ... SELECT</></para></listitem>
4700 <listitem><para>Fix <command>CLUSTER</> failure after
4701 <command>ALTER TABLE SET WITHOUT OIDS</></para></listitem>
4707 <sect1 id="release-7-4-7">
4708 <title>Release 7.4.7</title>
4711 <title>Release date</title>
4712 <simpara>2005-01-31</simpara>
4716 This release contains a variety of fixes from 7.4.6, including several
4717 security-related issues.
4721 <title>Migration to version 7.4.7</title>
4724 A dump/restore is not required for those running 7.4.X.
4729 <title>Changes</title>
4732 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
4734 On platforms that will automatically execute initialization functions of a
4735 shared library (this includes at least Windows and ELF-based Unixen),
4736 <command>LOAD</> can be used to make the server execute arbitrary code.
4737 Thanks to NGS Software for reporting this.</para></listitem>
4738 <listitem><para>Check that creator of an aggregate function has the right to
4739 execute the specified transition functions</para>
4741 This oversight made it possible to bypass denial of EXECUTE
4742 permission on a function.</para></listitem>
4743 <listitem><para>Fix security and 64-bit issues in
4744 contrib/intagg</para></listitem>
4745 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
4746 Jurka)</para></listitem>
4747 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
4748 many parameters (Neil)</para></listitem>
4749 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
4751 The result of the join was mistakenly supposed to be sorted the same as the
4752 left input. This could not only deliver mis-sorted output to the user, but
4753 in case of nested merge joins could give outright wrong answers.
4755 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
4756 <listitem><para>Fix display of negative intervals in SQL and GERMAN
4757 datestyles</para></listitem>
4758 <listitem><para>Make age(timestamptz) do calculation in local timezone not
4759 GMT</para></listitem>
4765 <sect1 id="release-7-4-6">
4766 <title>Release 7.4.6</title>
4769 <title>Release date</title>
4770 <simpara>2004-10-22</simpara>
4774 This release contains a variety of fixes from 7.4.5.
4779 <title>Migration to version 7.4.6</title>
4782 A dump/restore is not required for those running 7.4.X.
4787 <title>Changes</title>
4790 <listitem><para>Repair possible failure to update hint bits on disk</para>
4792 Under rare circumstances this oversight could lead to
4793 <quote>could not access transaction status</> failures, which qualifies
4794 it as a potential-data-loss bug.
4796 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
4798 Very large left joins using a hash join plan could fail to output unmatched
4799 left-side rows given just the right data distribution.
4801 <listitem><para>Disallow running <application>pg_ctl</> as root</para>
4803 This is to guard against any possible security issues.
4805 <listitem><para>Avoid using temp files in <filename>/tmp</> in <command>make_oidjoins_check</command></para>
4807 This has been reported as a security issue, though it's hardly worthy of
4808 concern since there is no reason for non-developers to use this script anyway.
4810 <listitem><para>Prevent forced backend shutdown from re-emitting prior command
4813 In rare cases, a client might think that its last command had succeeded when
4814 it really had been aborted by forced database shutdown.
4816 <listitem><para>Repair bug in <function>pg_stat_get_backend_idset</function></para>
4818 This could lead to misbehavior in some of the system-statistics views.
4820 <listitem><para>Fix small memory leak in postmaster</para></listitem>
4821 <listitem><para>Fix <quote>expected both swapped tables to have TOAST
4822 tables</> bug</para>
4824 This could arise in cases such as CLUSTER after ALTER TABLE DROP COLUMN.
4826 <listitem><para>Prevent <literal>pg_ctl restart</> from adding <literal>-D</> multiple times</para></listitem>
4827 <listitem><para>Fix problem with NULL values in GiST indexes</para></listitem>
4828 <listitem><para><literal>::</> is no longer interpreted as a variable in an
4829 ECPG prepare statement</para></listitem>
4835 <sect1 id="release-7-4-5">
4836 <title>Release 7.4.5</title>
4839 <title>Release date</title>
4840 <simpara>2004-08-18</simpara>
4844 This release contains one serious bug fix over 7.4.4.
4849 <title>Migration to version 7.4.5</title>
4852 A dump/restore is not required for those running 7.4.X.
4857 <title>Changes</title>
4860 <listitem><para>Repair possible crash during concurrent B-tree index insertions</para>
4862 This patch fixes a rare case in which concurrent insertions into a B-tree index
4863 could result in a server panic. No permanent damage would result, but it's
4864 still worth a re-release. The bug does not exist in pre-7.4 releases.
4871 <sect1 id="release-7-4-4">
4872 <title>Release 7.4.4</title>
4875 <title>Release date</title>
4876 <simpara>2004-08-16</simpara>
4880 This release contains a variety of fixes from 7.4.3.
4885 <title>Migration to version 7.4.4</title>
4888 A dump/restore is not required for those running 7.4.X.
4893 <title>Changes</title>
4896 <listitem><para>Prevent possible loss of committed transactions during crash</para>
4898 Due to insufficient interlocking between transaction commit and checkpointing,
4899 it was possible for transactions committed just before the most recent
4900 checkpoint to be lost, in whole or in part, following a database crash and
4901 restart. This is a serious bug that has existed
4902 since <productname>PostgreSQL</productname> 7.1.
4904 <listitem><para>Check HAVING restriction before evaluating result list of an
4905 aggregate plan</para></listitem>
4906 <listitem><para>Avoid crash when session's current user ID is deleted</para></listitem>
4907 <listitem><para>Fix hashed crosstab for zero-rows case (Joe)</para></listitem>
4908 <listitem><para>Force cache update after renaming a column in a foreign key</para></listitem>
4909 <listitem><para>Pretty-print UNION queries correctly</para></listitem>
4910 <listitem><para>Make psql handle <literal>\r\n</> newlines properly in COPY IN</para></listitem>
4911 <listitem><para><application>pg_dump</> handled ACLs with grant options incorrectly</para></listitem>
4912 <listitem><para>Fix thread support for OS X and Solaris</para></listitem>
4913 <listitem><para>Updated JDBC driver (build 215) with various fixes</para></listitem>
4914 <listitem><para>ECPG fixes</para></listitem>
4915 <listitem><para>Translation updates (various contributors)</para></listitem>
4921 <sect1 id="release-7-4-3">
4922 <title>Release 7.4.3</title>
4925 <title>Release date</title>
4926 <simpara>2004-06-14</simpara>
4930 This release contains a variety of fixes from 7.4.2.
4935 <title>Migration to version 7.4.3</title>
4938 A dump/restore is not required for those running 7.4.X.
4943 <title>Changes</title>
4946 <listitem><para>Fix temporary memory leak when using non-hashed aggregates (Tom)</para></listitem>
4947 <listitem><para>ECPG fixes, including some for Informix compatibility (Michael)</para></listitem>
4948 <listitem><para>Fixes for compiling with thread-safety, particularly Solaris (Bruce)</para></listitem>
4949 <listitem><para>Fix error in COPY IN termination when using the old network protocol (ljb)</para></listitem>
4950 <listitem><para>Several important fixes in pg_autovacuum, including fixes for
4951 large tables, unsigned oids, stability, temp tables, and debug mode
4952 (Matthew T. O'Connor)</para></listitem>
4953 <listitem><para>Fix problem with reading tar-format dumps on NetBSD and BSD/OS (Bruce)</para></listitem>
4954 <listitem><para>Several JDBC fixes</para></listitem>
4955 <listitem><para>Fix ALTER SEQUENCE RESTART where last_value equals the restart value (Tom)</para></listitem>
4956 <listitem><para>Repair failure to recalculate nested sub-selects (Tom)</para></listitem>
4957 <listitem><para>Fix problems with non-constant expressions in LIMIT/OFFSET</para></listitem>
4958 <listitem><para>Support FULL JOIN with no join clause, such as X FULL JOIN Y ON TRUE (Tom)</para></listitem>
4959 <listitem><para>Fix another zero-column table bug (Tom)</para></listitem>
4960 <listitem><para>Improve handling of non-qualified identifiers in GROUP BY clauses in sub-selects (Tom)</para>
4962 Select-list aliases within the sub-select will now take precedence over
4963 names from outer query levels.
4965 <listitem><para>Do not generate <quote>NATURAL CROSS JOIN</> when decompiling rules (Tom)</para></listitem>
4966 <listitem><para>Add checks for invalid field length in binary COPY (Tom)</para>
4968 This fixes a difficult-to-exploit security hole.
4970 <listitem><para>Avoid locking conflict between <command>ANALYZE</command> and <command>LISTEN</command>/<command>NOTIFY</command></para></listitem>
4971 <listitem><para>Numerous translation updates (various contributors)</para></listitem>
4977 <sect1 id="release-7-4-2">
4978 <title>Release 7.4.2</title>
4981 <title>Release date</title>
4982 <simpara>2004-03-08</simpara>
4986 This release contains a variety of fixes from 7.4.1.
4991 <title>Migration to version 7.4.2</title>
4994 A dump/restore is not required for those running 7.4.X. However,
4995 it may be advisable as the easiest method of incorporating fixes for
4996 two errors that have been found in the initial contents of 7.4.X system
4997 catalogs. A dump/initdb/reload sequence using 7.4.2's initdb will
4998 automatically correct these problems.
5002 The more severe of the two errors is that data type <type>anyarray</>
5003 has the wrong alignment label; this is a problem because the
5004 <structname>pg_statistic</> system catalog uses <type>anyarray</>
5005 columns. The mislabeling can cause planner misestimations and even
5006 crashes when planning queries that involve <literal>WHERE</> clauses on
5007 double-aligned columns (such as <type>float8</> and <type>timestamp</>).
5008 It is strongly recommended that all installations repair this error,
5009 either by initdb or by following the manual repair procedure given
5014 The lesser error is that the system view <structname>pg_settings</>
5015 ought to be marked as having public update access, to allow
5016 <literal>UPDATE pg_settings</> to be used as a substitute for
5017 <command>SET</>. This can also be fixed either by initdb or manually,
5018 but it is not necessary to fix unless you want to use <literal>UPDATE
5023 If you wish not to do an initdb, the following procedure will work
5024 for fixing <structname>pg_statistic</>. As the database superuser,
5028 -- clear out old data in pg_statistic:
5029 DELETE FROM pg_statistic;
5030 VACUUM pg_statistic;
5031 -- this should update 1 row:
5032 UPDATE pg_type SET typalign = 'd' WHERE oid = 2277;
5033 -- this should update 6 rows:
5034 UPDATE pg_attribute SET attalign = 'd' WHERE atttypid = 2277;
5036 -- At this point you MUST start a fresh backend to avoid a crash!
5038 -- repopulate pg_statistic:
5042 This can be done in a live database, but beware that all backends
5043 running in the altered database must be restarted before it is safe to
5044 repopulate <structname>pg_statistic</>.
5048 To repair the <structname>pg_settings</> error, simply do:
5050 GRANT SELECT, UPDATE ON pg_settings TO PUBLIC;
5055 The above procedures must be carried out in <emphasis>each</> database
5056 of an installation, including <literal>template1</>, and ideally
5057 including <literal>template0</> as well. If you do not fix the
5058 template databases then any subsequently created databases will contain
5059 the same errors. <literal>template1</> can be fixed in the same way
5060 as any other database, but fixing <literal>template0</> requires
5061 additional steps. First, from any database issue
5063 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
5065 Next connect to <literal>template0</> and perform the above repair
5066 procedures. Finally, do
5068 -- re-freeze template0:
5070 -- and protect it against future alterations:
5071 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
5077 <title>Changes</title>
5080 Release 7.4.2 incorporates all the fixes included in release 7.3.6,
5081 plus the following fixes:
5085 <listitem><para>Fix <structname>pg_statistics</> alignment bug that could crash optimizer</para>
5086 <para>See above for details about this problem.</para></listitem>
5087 <listitem><para>Allow non-super users to update <structname>pg_settings</></para></listitem>
5088 <listitem><para>Fix several optimizer bugs, most of which led to
5089 <quote>variable not found in subplan target lists</> errors</para></listitem>
5090 <listitem><para>Avoid out-of-memory failure during startup of large multiple
5091 index scan</para></listitem>
5092 <listitem><para>Fix multibyte problem that could lead to <quote>out of
5093 memory</> error during <command>COPY IN</></para></listitem>
5094 <listitem><para>Fix problems with <command>SELECT INTO</> / <command>CREATE
5095 TABLE AS</> from tables without OIDs</para></listitem>
5096 <listitem><para>Fix problems with <filename>alter_table</> regression test
5097 during parallel testing</para></listitem>
5098 <listitem><para>Fix problems with hitting open file limit, especially on OS X (Tom)</para></listitem>
5099 <listitem><para>Partial fix for Turkish-locale issues</para>
5100 <para>initdb will succeed now in Turkish locale, but there are still some
5101 inconveniences associated with the <literal>i/I</> problem.</para></listitem>
5102 <listitem><para>Make pg_dump set client encoding on restore</para></listitem>
5103 <listitem><para>Other minor pg_dump fixes</para></listitem>
5104 <listitem><para>Allow ecpg to again use C keywords as column names (Michael)</para></listitem>
5105 <listitem><para>Added ecpg <literal>WHENEVER NOT_FOUND</> to
5106 <literal>SELECT/INSERT/UPDATE/DELETE</> (Michael)</para></listitem>
5107 <listitem><para>Fix ecpg crash for queries calling set-returning functions (Michael)</para></listitem>
5108 <listitem><para>Various other ecpg fixes (Michael)</para></listitem>
5109 <listitem><para>Fixes for Borland compiler</para></listitem>
5110 <listitem><para>Thread build improvements (Bruce)</para></listitem>
5111 <listitem><para>Various other build fixes</para></listitem>
5112 <listitem><para>Various JDBC fixes</para></listitem>
5118 <sect1 id="release-7-4-1">
5119 <title>Release 7.4.1</title>
5122 <title>Release date</title>
5123 <simpara>2003-12-22</simpara>
5127 This release contains a variety of fixes from 7.4.
5132 <title>Migration to version 7.4.1</title>
5135 A dump/restore is <emphasis>not</emphasis> required for those
5140 If you want to install the fixes in the information schema
5141 you need to reload it into the database.
5142 This is either accomplished by initializing a new cluster
5143 by running <command>initdb</command>, or by running the following
5144 sequence of SQL commands in each database (ideally including
5145 <literal>template1</literal>) as a superuser in
5146 <application>psql</application>, after installing the new release:
5148 DROP SCHEMA information_schema CASCADE;
5149 \i /usr/local/pgsql/share/information_schema.sql
5151 Substitute your installation path in the second command.
5157 <title>Changes</title>
5160 <listitem><para>Fixed bug in <command>CREATE SCHEMA</command> parsing in ECPG (Michael)</para></listitem>
5161 <listitem><para>Fix compile error when <option>--enable-thread-safety</option> and <option>--with-perl</option> are used together (Peter)</para></listitem>
5162 <listitem><para>Fix for subqueries that used hash joins (Tom)</para>
5164 Certain subqueries that used hash joins would crash because of
5165 improperly shared structures.
5167 <listitem><para>Fix free space map compaction bug (Tom)</para>
5169 This fixes a bug where compaction of the free space map could lead
5170 to a database server shutdown.
5173 <listitem><para>Fix for Borland compiler build of libpq (Bruce)</para></listitem>
5174 <listitem><para>Fix <function>netmask()</function> and <function>hostmask()</function> to return the maximum-length masklen (Tom)</para>
5176 Fix these functions to return values consistent with pre-7.4
5180 <listitem><para>Several <filename>contrib/pg_autovacuum</filename> fixes</para>
5182 Fixes include improper variable initialization, missing vacuum after
5183 <command>TRUNCATE</command>, and duration computation overflow for long vacuums.
5186 <listitem><para>Allow compile of <filename>contrib/cube</filename> under Cygwin (Jason Tishler)</para></listitem>
5187 <listitem><para>Fix Solaris use of password file when no passwords are defined (Tom)</para>
5189 Fix crash on Solaris caused by use of any type of password
5190 authentication when no passwords were defined.
5193 <listitem><para>JDBC fix for thread problems, other fixes</para></listitem>
5194 <listitem><para>Fix for <type>bytea</type> index lookups (Joe)</para></listitem>
5195 <listitem><para>Fix information schema for bit data types (Peter)</para></listitem>
5196 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
5197 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
5198 <listitem><para>Make <function>PQescapeBytea</function> and <function>byteaout</function> consistent with each other (Joe)</para></listitem>
5199 <listitem><para>Escape <type>bytea</type> output for bytes > 0x7e(Joe)</para>
5201 If different client encodings are used for <type>bytea</type> output and input, it
5202 is possible for <type>bytea</type> values to be corrupted by the differing
5203 encodings. This fix escapes all bytes that might be affected.
5206 <listitem><para>Added missing <function>SPI_finish()</function> calls to dblink's <function>get_tuple_of_interest()</function> (Joe)</para></listitem>
5207 <listitem><para>New Czech FAQ</para></listitem>
5208 <listitem><para>Fix information schema view <literal>constraint_column_usage</literal> for foreign keys (Peter)</para></listitem>
5209 <listitem><para>ECPG fixes (Michael)</para></listitem>
5210 <listitem><para>Fix bug with multiple <literal>IN</literal> subqueries and joins in the subqueries (Tom)</para></listitem>
5211 <listitem><para>Alllow <literal>COUNT('x')</literal> to work (Tom)</para></listitem>
5212 <listitem><para>Install ECPG include files for Informix compatibility into separate directory (Peter)</para>
5214 Some names of ECPG include files for Informix compatibility conflicted with operating system include files.
5215 By installing them in their own directory, name conflicts have been reduced.
5218 <listitem><para>Fix SSL memory leak (Neil)</para>
5220 This release fixes a bug in 7.4 where SSL didn't free all memory it allocated.
5223 <listitem><para>Prevent <filename>pg_service.conf</filename> from using service name as default dbname (Bruce)</para></listitem>
5224 <listitem><para>Fix local ident authentication on FreeBSD (Tom)</para></listitem>
5230 <sect1 id="release-7-4">
5231 <title>Release 7.4</title>
5234 <title>Release date</title>
5235 <simpara>2003-11-17</simpara>
5239 <title>Overview</title>
5242 Major changes in this release:
5248 <literal>IN</literal> / <literal>NOT IN</literal> subqueries are
5249 now much more efficient
5254 In previous releases, <literal>IN</literal>/<literal>NOT
5255 IN</literal> subqueries were joined to the upper query by
5256 sequentially scanning the subquery looking for a match. The
5257 7.4 code uses the same sophisticated techniques used by
5258 ordinary joins and so is much faster. An
5259 <literal>IN</literal> will now usually be as fast as or faster
5260 than an equivalent <literal>EXISTS</literal> subquery; this
5261 reverses the conventional wisdom that applied to previous
5269 Improved <literal>GROUP BY</literal> processing by using hash buckets
5274 In previous releases, rows to be grouped had to be sorted
5275 first. The 7.4 code can do <literal>GROUP BY</literal>
5276 without sorting, by accumulating results into a hash table
5277 with one entry per group. It will still use the sort
5278 technique, however, if the hash table is estimated to be too
5279 large to fit in <varname>sort_mem</>.
5286 New multikey hash join capability
5291 In previous releases, hash joins could only occur on single
5292 keys. This release allows multicolumn hash joins.
5299 Queries using the explicit <literal>JOIN</literal> syntax are
5300 now better optimized
5305 Prior releases evaluated queries using the explicit
5306 <literal>JOIN</literal> syntax only in the order implied by
5307 the syntax. 7.4 allows full optimization of these queries,
5308 meaning the optimizer considers all possible join orderings
5309 and chooses the most efficient. Outer joins, however, must
5310 still follow the declared ordering.
5317 Faster and more powerful regular expression code
5322 The entire regular expression module has been replaced with a
5323 new version by Henry Spencer, originally written for Tcl. The
5324 code greatly improves performance and supports several flavors
5325 of regular expressions.
5332 Function-inlining for simple SQL functions
5337 Simple SQL functions can now be inlined by including their SQL
5338 in the main query. This improves performance by eliminating
5339 per-call overhead. That means simple SQL functions now
5347 Full support for IPv6 connections and IPv6 address data types
5352 Previous releases allowed only IPv4 connections, and the IP
5353 data types only supported IPv4 addresses. This release adds
5354 full IPv6 support in both of these areas.
5361 Major improvements in SSL performance and reliability
5366 Several people very familiar with the SSL API have overhauled
5367 our SSL code to improve SSL key negotiation and error
5375 Make free space map efficiently reuse empty index pages,
5376 and other free space management improvements
5381 In previous releases, B-tree index pages that were left empty
5382 because of deleted rows could only be reused by rows with
5383 index values similar to the rows originally indexed on that
5384 page. In 7.4, <command>VACUUM</command> records empty index
5385 pages and allows them to be reused for any future index rows.
5392 SQL-standard information schema
5397 The information schema provides a standardized and stable way
5398 to access information about the schema objects defined in a
5406 Cursors conform more closely to the SQL standard
5411 The commands <command>FETCH</command> and
5412 <command>MOVE</command> have been overhauled to conform more
5413 closely to the SQL standard.
5420 Cursors can exist outside transactions
5425 These cursors are also called holdable cursors.
5432 New client-to-server protocol
5437 The new protocol adds error codes, more status information,
5438 faster startup, better support for binary data transmission,
5439 parameter values separated from SQL commands, prepared
5440 statements available at the protocol level, and cleaner
5441 recovery from <command>COPY</command> failures. The older
5442 protocol is still supported by both server and clients.
5449 <application>libpq</application> and
5450 <application>ECPG</application> applications are now fully
5456 While previous <application>libpq</application> releases
5457 already supported threads, this release improves thread safety
5458 by fixing some non-thread-safe code that was used during
5459 database connection startup. The <command>configure</command>
5460 option <option>--enable-thread-safety</option> must be used to
5461 enable this feature.
5468 New version of full-text indexing
5473 A new full-text indexing suite is available in
5474 <filename>contrib/tsearch2</filename>.
5486 The new autovacuum tool in
5487 <filename>contrib/autovacuum</filename> monitors the database
5488 statistics tables for
5489 <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
5490 activity and automatically vacuums tables when needed.
5497 Array handling has been improved and moved into the server core
5502 Many array limitations have been removed, and arrays behave
5503 more like fully-supported data types.
5512 <title>Migration to version 7.4</title>
5515 A dump/restore using <application>pg_dump</application> is
5516 required for those wishing to migrate data from any previous
5521 Observe the following incompatibilities:
5527 The server-side autocommit setting was removed and
5528 reimplemented in client applications and languages.
5529 Server-side autocommit was causing too many problems with
5530 languages and applications that wanted to control their own
5531 autocommit behavior, so autocommit was removed from the server
5532 and added to individual client APIs as appropriate.
5538 Error message wording has changed substantially in this
5539 release. Significant effort was invested to make the messages
5540 more consistent and user-oriented. If your applications try to
5541 detect different error conditions by parsing the error message,
5542 you are strongly encouraged to use the new error code facility instead.
5548 Inner joins using the explicit <literal>JOIN</literal> syntax
5549 may behave differently because they are now better
5556 A number of server configuration parameters have been renamed
5557 for clarity, primarily those related to
5564 <literal>FETCH 0</literal> or <literal>MOVE 0</literal> now
5565 does nothing. In prior releases, <literal>FETCH 0</literal>
5566 would fetch all remaining rows, and <literal>MOVE 0</literal>
5567 would move to the end of the cursor.
5573 <command>FETCH</command> and <command>MOVE</command> now return
5574 the actual number of rows fetched/moved, or zero if at the
5575 beginning/end of the cursor. Prior releases would return the
5576 row count passed to the command, not the number of rows
5577 actually fetched or moved.
5583 <command>COPY</command> now can process files that use
5584 carriage-return or carriage-return/line-feed end-of-line
5585 sequences. Literal carriage-returns and line-feeds are no
5586 longer accepted in data values; use <literal>\r</literal> and
5587 <literal>\n</literal> instead.
5593 Trailing spaces are now trimmed when converting from type
5594 <type>char(<replaceable>n</>)</type> to
5595 <type>varchar(<replaceable>n</>)</type> or <type>text</type>.
5596 This is what most people always expected to happen anyway.
5602 The data type <type>float(<replaceable>p</>)</type> now
5603 measures <replaceable>p</> in binary digits, not decimal
5604 digits. The new behavior follows the SQL standard.
5610 Ambiguous date values now must match the ordering specified by
5611 the <varname>datestyle</varname> setting. In prior releases, a
5612 date specification of <literal>10/20/03</> was interpreted as a
5613 date in October even if <varname>datestyle</> specified that
5614 the day should be first. 7.4 will throw an error if a date
5615 specification is invalid for the current setting of
5616 <varname>datestyle</>.
5622 The functions <function>oidrand</function>,
5623 <function>oidsrand</function>, and
5624 <function>userfntest</function> have been removed. These
5625 functions were determined to be no longer useful.
5631 String literals specifying time-varying date/time values, such
5632 as <literal>'now'</literal> or <literal>'today'</literal> will
5633 no longer work as expected in column default expressions; they
5634 now cause the time of the table creation to be the default, not
5635 the time of the insertion. Functions such as
5636 <function>now()</>, <function>current_timestamp</>, or
5637 <function>current_date</function> should be used instead.
5641 In previous releases, there was special code so that strings
5642 such as <literal>'now'</literal> were interpreted at
5643 <command>INSERT</> time and not at table creation time, but
5644 this work around didn't cover all cases. Release 7.4 now
5645 requires that defaults be defined properly using functions such
5646 as <function>now()</> or <function>current_timestamp</>. These
5647 will work in all situations.
5653 The dollar sign (<literal>$</>) is no longer allowed in
5654 operator names. It can instead be a non-first character in
5655 identifiers. This was done to improve compatibility with other
5656 database systems, and to avoid syntax problems when parameter
5657 placeholders (<literal>$<replaceable>n</></>) are written
5658 adjacent to operators.
5666 <title>Changes</title>
5669 Below you will find a detailed account of the changes between
5670 release 7.4 and the previous major release.
5674 <title>Server Operation Changes</title>
5679 Allow IPv6 server connections (Nigel Kukard, Johan Jordaan,
5680 Bruce, Tom, Kurt Roeckx, Andrew Dunstan)
5686 Fix SSL to handle errors cleanly (Nathan Mueller)
5689 In prior releases, certain SSL API error reports were not
5690 handled correctly. This release fixes those problems.
5696 SSL protocol security and performance improvements (Sean Chittenden)
5699 SSL key renegotiation was happening too frequently, causing poor
5700 SSL performance. Also, initial key handling was improved.
5706 Print lock information when a deadlock is detected (Tom)
5709 This allows easier debugging of deadlock situations.
5715 Update <filename>/tmp</filename> socket modification times
5716 regularly to avoid their removal (Tom)
5719 This should help prevent <filename>/tmp</filename> directory
5720 cleaner administration scripts from removing server socket
5725 <listitem><para>Enable PAM for Mac OS X (Aaron Hillegass)</para></listitem>
5728 <para>Make B-tree indexes fully WAL-safe (Tom)</para>
5730 In prior releases, under certain rare cases, a server crash
5731 could cause B-tree indexes to become corrupt. This release
5732 removes those last few rare cases.
5736 <listitem><para>Allow B-tree index compaction and empty page reuse (Tom)</para></listitem>
5740 Fix inconsistent index lookups during split of first root page (Tom)
5743 In prior releases, when a single-page index split into two
5744 pages, there was a brief period when another database session
5745 could miss seeing an index entry. This release fixes that rare
5750 <listitem><para>Improve free space map allocation logic (Tom)</para></listitem>
5753 <para>Preserve free space information between server restarts (Tom)</para>
5755 In prior releases, the free space map was not saved when the
5756 postmaster was stopped, so newly started servers had no free
5757 space information. This release saves the free space map, and
5758 reloads it when the server is restarted.
5762 <listitem><para>Add start time to <literal>pg_stat_activity</literal> (Neil)</para></listitem>
5763 <listitem><para>New code to detect corrupt disk pages; erase with <varname>zero_damaged_pages</varname> (Tom)</para></listitem>
5764 <listitem><para>New client/server protocol: faster, no username length limit, allow clean exit from <command>COPY</command> (Tom)</para></listitem>
5765 <listitem><para>Add transaction status, table ID, column ID to client/server protocol (Tom)</para></listitem>
5766 <listitem><para>Add binary I/O to client/server protocol (Tom)</para></listitem>
5767 <listitem><para>Remove autocommit server setting; move to client applications (Tom)</para></listitem>
5768 <listitem><para>New error message wording, error codes, and three levels of error detail (Tom, Joe, Peter)</para></listitem>
5773 <title>Performance Improvements</title>
5776 <listitem><para>Add hashing for <literal>GROUP BY</literal> aggregates (Tom)</para></listitem>
5777 <listitem><para>Make nested-loop joins be smarter about multicolumn indexes (Tom)</para></listitem>
5778 <listitem><para>Allow multikey hash joins (Tom)</para></listitem>
5779 <listitem><para>Improve constant folding (Tom)</para></listitem>
5780 <listitem><para>Add ability to inline simple SQL functions (Tom)</para></listitem>
5783 <para>Reduce memory usage for queries using complex functions (Tom)</para>
5785 In prior releases, functions returning allocated memory would
5786 not free it until the query completed. This release allows the
5787 freeing of function-allocated memory when the function call
5788 completes, reducing the total memory used by functions.
5793 <para>Improve GEQO optimizer performance (Tom)</para>
5795 This release fixes several inefficiencies in the way the GEQO optimizer
5796 manages potential query paths.
5802 Allow <literal>IN</>/<literal>NOT IN</> to be handled via hash
5809 Improve <literal>NOT IN (<replaceable>subquery</>)</literal>
5816 Allow most <literal>IN</literal> subqueries to be processed as
5823 Pattern matching operations can use indexes regardless of
5827 There is no way for non-ASCII locales to use the standard
5828 indexes for <literal>LIKE</literal> comparisons. This release
5829 adds a way to create a special index for
5830 <literal>LIKE</literal>.
5835 <para>Allow the postmaster to preload libraries using <varname>preload_libraries</varname> (Joe)</para>
5837 For shared libraries that require a long time to load, this
5838 option is available so the library can be preloaded in the
5839 postmaster and inherited by all database sessions.
5845 Improve optimizer cost computations, particularly for subqueries (Tom)
5851 Avoid sort when subquery <literal>ORDER BY</literal> matches upper query (Tom)
5857 Deduce that <literal>WHERE a.x = b.y AND b.y = 42</literal> also
5858 means <literal>a.x = 42</literal> (Tom)
5864 Allow hash/merge joins on complex joins (Tom)
5870 Allow hash joins for more data types (Tom)
5876 Allow join optimization of explicit inner joins, disable with
5877 <varname>join_collapse_limit</varname> (Tom)
5883 Add parameter <varname>from_collapse_limit</varname> to control
5884 conversion of subqueries to joins (Tom)
5890 Use faster and more powerful regular expression code from Tcl
5891 (Henry Spencer, Tom)
5897 Use bit-mapped relation sets in the optimizer (Tom)
5902 <para>Improve connection startup time (Tom)</para>
5904 The new client/server protocol requires fewer network packets to
5905 start a database session.
5911 Improve trigger/constraint performance (Stephan)
5917 Improve speed of <literal>col IN (const, const, const, ...)</literal> (Tom)
5923 Fix hash indexes which were broken in rare cases (Tom)
5927 <listitem><para>Improve hash index concurrency and speed (Tom)</para>
5929 Prior releases suffered from poor hash index performance,
5930 particularly for high concurrency situations. This release fixes
5931 that, and the development group is interested in reports
5932 comparing B-tree and hash index performance.
5937 <para>Align shared buffers on 32-byte boundary for copy speed improvement (Manfred Spraul)</para>
5939 Certain CPU's perform faster data copies when addresses are
5945 <para>Data type <type>numeric</type> reimplemented for better performance (Tom)</para>
5947 <type>numeric</type> used to be stored in base 100. The new code
5948 uses base 10000, for significantly better performance.
5955 <title>Server Configuration Changes</title>
5959 <para>Rename server parameter <varname>server_min_messages</> to <varname>log_min_messages</> (Bruce)</para>
5961 This was done so most parameters that control the server logs
5962 begin with <literal>log_</>.
5966 <listitem><para>Rename <varname>show_*_stats</> to <varname>log_*_stats</> (Bruce)</para></listitem>
5967 <listitem><para>Rename <varname>show_source_port</> to <varname>log_source_port</> (Bruce)</para></listitem>
5968 <listitem><para>Rename <varname>hostname_lookup</> to <varname>log_hostname</> (Bruce)</para></listitem>
5971 <para>Add <varname>checkpoint_warning</> to warn of excessive checkpointing (Bruce)</para>
5973 In prior releases, it was difficult to determine if checkpoint
5974 was happening too frequently. This feature adds a warning to the
5975 server logs when excessive checkpointing happens.
5979 <listitem><para>New read-only server parameters for localization (Tom)</para></listitem>
5983 Change debug server log messages to output as <literal>DEBUG</>
5984 rather than <literal>LOG</> (Bruce)
5989 <para>Prevent server log variables from being turned off by non-superusers (Bruce)</para>
5991 This is a security feature so non-superusers cannot disable
5992 logging that was enabled by the administrator.
5998 <varname>log_min_messages</>/<varname>client_min_messages</> now
5999 controls <varname>debug_*</> output (Bruce)
6002 This centralizes client debug information so all debug output
6003 can be sent to either the client or server logs.
6008 <para>Add Mac OS X Rendezvous server support (Chris Campbell)</para>
6010 This allows Mac OS X hosts to query the network for available
6011 <productname>PostgreSQL</productname> servers.
6017 Add ability to print only slow statements using
6018 <varname>log_min_duration_statement</varname>
6022 This is an often requested debugging feature that allows
6023 administrators to see only slow queries in their server logs.
6028 <para>Allow <filename>pg_hba.conf</filename> to accept netmasks in CIDR format (Andrew Dunstan)</para>
6030 This allows administrators to merge the host IP address and
6031 netmask fields into a single CIDR field in <filename>pg_hba.conf</filename>.
6035 <listitem><para>New read-only parameter <varname>is_superuser</varname> (Tom)</para></listitem>
6038 <para>New parameter <varname>log_error_verbosity</varname> to control error detail (Tom)</para>
6040 This works with the new error reporting feature to supply
6041 additional error information like hints, file names and line
6047 <para><literal>postgres --describe-config</literal> now dumps server config variables (Aizaz Ahmed, Peter)</para>
6049 This option is useful for administration tools that need to know
6050 the configuration variable names and their minimums, maximums,
6051 defaults, and descriptions.
6057 Add new columns in <literal>pg_settings</literal>:
6058 <literal>context</>, <literal>type</>, <literal>source</>,
6059 <literal>min_val</>, <literal>max_val</> (Joe)
6065 Make default <varname>shared_buffers</> 1000 and
6066 <varname>max_connections</> 100, if possible (Tom)
6069 Prior versions defaulted to 64 shared buffers so <productname>PostgreSQL</productname>
6070 would start on even very old systems. This release tests the
6071 amount of shared memory allowed by the platform and selects more
6072 reasonable default values if possible. Of course, users are
6073 still encouraged to evaluate their resource load and size
6074 <varname>shared_buffers</varname> accordingly.
6080 New <filename>pg_hba.conf</filename> record type
6081 <literal>hostnossl</> to prevent SSL connections (Jon
6085 In prior releases, there was no way to prevent SSL connections
6086 if both the client and server supported SSL. This option allows
6093 Remove parameter <varname>geqo_random_seed</varname>
6100 Add server parameter <varname>regex_flavor</varname> to control regular expression processing (Tom)
6106 Make <command>pg_ctl</command> better handle nonstandard ports (Greg)
6113 <title>Query Changes</title>
6116 <listitem><para>New SQL-standard information schema (Peter)</para></listitem>
6117 <listitem><para>Add read-only transactions (Peter)</para></listitem>
6118 <listitem><para>Print key name and value in foreign-key violation messages (Dmitry Tkach)</para></listitem>
6121 <para>Allow users to see their own queries in <literal>pg_stat_activity</literal> (Kevin Brown)</para>
6123 In prior releases, only the superuser could see query strings
6124 using <literal>pg_stat_activity</literal>. Now ordinary users
6125 can see their own query strings.
6130 <para>Fix aggregates in subqueries to match SQL standard (Tom)</para>
6132 The SQL standard says that an aggregate function appearing
6133 within a nested subquery belongs to the outer query if its
6134 argument contains only outer-query variables. Prior
6135 <productname>PostgreSQL</productname> releases did not handle
6136 this fine point correctly.
6141 <para>Add option to prevent auto-addition of tables referenced in query (Nigel J. Andrews)</para>
6143 By default, tables mentioned in the query are automatically
6144 added to the <literal>FROM</> clause if they are not already
6145 there. This is compatible with historic
6146 <productname>POSTGRES</productname> behavior but is contrary to
6147 the SQL standard. This option allows selecting
6148 standard-compatible behavior.
6153 <para>Allow <literal>UPDATE ... SET col = DEFAULT</literal> (Rod)</para>
6155 This allows <command>UPDATE</command> to set a column to its
6156 declared default value.
6161 <para>Allow expressions to be used in <literal>LIMIT</>/<literal>OFFSET</> (Tom)</para>
6163 In prior releases, <literal>LIMIT</>/<literal>OFFSET</> could
6164 only use constants, not expressions.
6169 <para>Implement <literal>CREATE TABLE AS EXECUTE</literal> (Neil, Peter)</para>
6175 <title>Object Manipulation Changes</title>
6179 <para>Make <command>CREATE SEQUENCE</command> grammar more conforming to SQL:2003 (Neil)</para>
6183 <para>Add statement-level triggers (Neil)</para>
6185 While this allows a trigger to fire at the end of a statement,
6186 it does not allow the trigger to access all rows modified by the
6187 statement. This capability is planned for a future release.
6192 <para>Add check constraints for domains (Rod)</para>
6194 This greatly increases the usefulness of domains by allowing
6195 them to use check constraints.
6200 <para>Add <command>ALTER DOMAIN</command> (Rod)</para>
6202 This allows manipulation of existing domains.
6207 <para>Fix several zero-column table bugs (Tom)</para>
6209 <productname>PostgreSQL</productname> supports zero-column tables. This fixes various bugs
6210 that occur when using such tables.
6215 <para>Have <literal>ALTER TABLE ... ADD PRIMARY KEY</literal> add not-null constraint (Rod)</para>
6217 In prior releases, <literal>ALTER TABLE ... ADD
6218 PRIMARY</literal> would add a unique index, but not a not-null
6219 constraint. That is fixed in this release.
6223 <listitem><para>Add <literal>ALTER TABLE ... WITHOUT OIDS</literal> (Rod)</para>
6225 This allows control over whether new and updated rows will have
6226 an OID column. This is most useful for saving storage space.
6232 Add <literal>ALTER SEQUENCE</literal> to modify minimum, maximum,
6233 increment, cache, cycle values (Rod)
6238 <para>Add <literal>ALTER TABLE ... CLUSTER ON</literal> (Alvaro Herrera)</para>
6240 This command is used by <command>pg_dump</command> to record the
6241 cluster column for each table previously clustered. This
6242 information is used by database-wide cluster to cluster all
6243 previously clustered tables.
6247 <listitem><para>Improve automatic type casting for domains (Rod, Tom)</para></listitem>
6248 <listitem><para>Allow dollar signs in identifiers, except as first character (Tom)</para></listitem>
6249 <listitem><para>Disallow dollar signs in operator names, so <literal>x=$1</> works (Tom)</para></listitem>
6253 Allow copying table schema using <literal>LIKE
6254 <replaceable>subtable</replaceable></literal>, also SQL:2003
6255 feature <literal>INCLUDING DEFAULTS</literal> (Rod)
6261 Add <literal>WITH GRANT OPTION</literal> clause to
6262 <command>GRANT</command> (Peter)
6265 This enabled <command>GRANT</command> to give other users the
6266 ability to grant privileges on a object.
6273 <title>Utility Command Changes</title>
6277 <para>Add <literal>ON COMMIT</literal> clause to <command>CREATE TABLE</command> for temporary tables (Gavin)</para>
6279 This adds the ability for a table to be dropped or all rows
6280 deleted on transaction commit.
6285 <para>Allow cursors outside transactions using <literal>WITH HOLD</literal> (Neil)</para>
6287 In previous releases, cursors were removed at the end of the
6288 transaction that created them. Cursors can now be created with
6289 the <literal>WITH HOLD</literal> option, which allows them to
6290 continue to be accessed after the creating transaction has
6296 <para><literal>FETCH 0</literal> and <literal>MOVE 0 </literal> now do nothing (Bruce)</para>
6298 In previous releases, <literal>FETCH 0</literal> fetched all
6299 remaining rows, and <literal>MOVE 0</literal> moved to the end
6306 Cause <command>FETCH</command> and <command>MOVE</command> to
6307 return the number of rows fetched/moved, or zero if at the
6308 beginning/end of cursor, per SQL standard (Bruce)
6311 In prior releases, the row count returned by
6312 <command>FETCH</command> and <command>MOVE</command> did not
6313 accurately reflect the number of rows processed.
6318 <para>Properly handle <literal>SCROLL</literal> with cursors, or
6319 report an error (Neil)</para>
6321 Allowing random access (both forward and backward scrolling) to
6322 some kinds of queries cannot be done without some additional
6323 work. If <literal>SCROLL</literal> is specified when the cursor
6324 is created, this additional work will be performed. Furthermore,
6325 if the cursor has been created with <literal>NO SCROLL</literal>,
6326 no random access is allowed.
6332 Implement SQL-compatible options <literal>FIRST</>,
6333 <literal>LAST</>, <literal>ABSOLUTE <replaceable>n</></>,
6334 <literal>RELATIVE <replaceable>n</></> for
6335 <command>FETCH</command> and <command>MOVE</command> (Tom)
6340 <para>Allow <command>EXPLAIN</command> on <command>DECLARE CURSOR</command> (Tom)</para>
6344 <para>Allow <command>CLUSTER</command> to use index marked as pre-clustered by default (Alvaro Herrera)</para>
6348 <para>Allow <command>CLUSTER</command> to cluster all tables (Alvaro Herrera)</para>
6350 This allows all previously clustered tables in a database to be
6351 reclustered with a single command.
6355 <listitem><para>Prevent <command>CLUSTER</command> on partial indexes (Tom)</para></listitem>
6357 <listitem><para>Allow DOS and Mac line-endings in <command>COPY</> files (Bruce)</para></listitem>
6361 Disallow literal carriage return as a data value,
6362 backslash-carriage-return and <literal>\r</> are still allowed
6368 <para><command>COPY</> changes (binary, <literal>\.</>) (Tom)</para>
6372 <para>Recover from <command>COPY</command> failure cleanly (Tom)</para>
6376 <para>Prevent possible memory leaks in <command>COPY</command> (Tom)</para>
6380 <para>Make <command>TRUNCATE</command> transaction-safe (Rod)</para>
6382 <command>TRUNCATE</command> can now be used inside a
6383 transaction. If the transaction aborts, the changes made by the
6384 <command>TRUNCATE</command> are automatically rolled back.
6390 Allow prepare/bind of utility commands like
6391 <command>FETCH</command> and <command>EXPLAIN</command> (Tom)
6396 <para>Add <command>EXPLAIN EXECUTE</command> (Neil)</para>
6400 <para>Improve <command>VACUUM</command> performance on indexes by reducing WAL traffic (Tom)</para>
6404 <para>Functional indexes have been generalized into indexes on expressions (Tom)</para>
6406 In prior releases, functional indexes only supported a simple
6407 function applied to one or more column names. This release
6408 allows any type of scalar expression.
6414 Have <command>SHOW TRANSACTION ISOLATION</command> match input
6415 to <command>SET TRANSACTION ISOLATION</command>
6422 Have <command>COMMENT ON DATABASE</command> on nonlocal
6423 database generate a warning, rather than an error (Rod)
6427 Database comments are stored in database-local tables so
6428 comments on a database have to be stored in each database.
6434 Improve reliability of <command>LISTEN</>/<command>NOTIFY</> (Tom)
6439 <para>Allow <command>REINDEX</command> to reliably reindex nonshared system catalog indexes (Tom)</para>
6441 This allows system tables to be reindexed without the
6442 requirement of a standalone session, which was necessary in
6443 previous releases. The only tables that now require a standalone
6444 session for reindexing are the global system tables
6445 <literal>pg_database</>, <literal>pg_shadow</>, and
6446 <literal>pg_group</>.
6453 <title>Data Type and Function Changes</title>
6458 New server parameter <varname>extra_float_digits</varname> to
6459 control precision display of floating-point numbers (Pedro
6463 This controls output precision which was causing regression
6468 <listitem><para>Allow <literal>+1300</literal> as a numeric time-zone specifier, for FJST (Tom)</para></listitem>
6472 Remove rarely used functions <function>oidrand</>,
6473 <function>oidsrand</>, and <function>userfntest</> functions
6479 <para>Add <function>md5()</> function to main server, already in <filename>contrib/pgcrypto</filename> (Joe)</para>
6481 An MD5 function was frequently requested. For more complex
6482 encryption capabilities, use
6483 <filename>contrib/pgcrypto</filename>.
6487 <listitem><para>Increase date range of <type>timestamp</type> (John Cochran)</para></listitem>
6491 Change <literal>EXTRACT(EPOCH FROM timestamp)</literal> so
6492 <type>timestamp without time zone</type> is assumed to be in
6493 local time, not GMT (Tom)
6497 <listitem><para>Trap division by zero in case the operating system doesn't prevent it (Tom)</para></listitem>
6498 <listitem><para>Change the <type>numeric</type> data type internally to base 10000 (Tom)</para></listitem>
6499 <listitem><para>New <function>hostmask()</function> function (Greg Wickham)</para></listitem>
6500 <listitem><para>Fixes for <function>to_char()</function> and <function>to_timestamp()</function> (Karel)</para></listitem>
6504 Allow functions that can take any argument data type and return
6505 any data type, using <type>anyelement</type> and
6506 <type>anyarray</type> (Joe)
6509 This allows the creation of functions that can work with any
6516 Arrays may now be specified as <literal>ARRAY[1,2,3]</literal>,
6517 <literal>ARRAY[['a','b'],['c','d']]</literal>, or
6518 <literal>ARRAY[ARRAY[ARRAY[2]]]</literal> (Joe)
6524 Allow proper comparisons for arrays, including <literal>ORDER
6525 BY</literal> and <literal>DISTINCT</literal> support
6530 <listitem><para>Allow indexes on array columns (Joe)</para></listitem>
6531 <listitem><para>Allow array concatenation with <literal>||</literal> (Joe)</para></listitem>
6535 Allow <literal>WHERE</literal> qualification
6536 <literal><replaceable>expr</> <replaceable>op</> ANY/SOME/ALL
6537 (<replaceable>array_expr</>)</literal> (Joe)
6540 This allows arrays to behave like a list of values, for purposes
6541 like <literal>SELECT * FROM tab WHERE col IN
6542 (array_val)</literal>.
6548 New array functions <function>array_append</>,
6549 <function>array_cat</>, <function>array_lower</>,
6550 <function>array_prepend</>, <function>array_to_string</>,
6551 <function>array_upper</>, <function>string_to_array</> (Joe)
6555 <listitem><para>Allow user defined aggregates to use polymorphic functions (Joe)</para></listitem>
6556 <listitem><para>Allow assignments to empty arrays (Joe)</para></listitem>
6560 Allow 60 in seconds fields of <type>time</type>,
6561 <type>timestamp</type>, and <type>interval</type> input values
6565 Sixty-second values are needed for leap seconds.
6569 <listitem><para>Allow <type>cidr</type> data type to be cast to <type>text</type> (Tom)</para></listitem>
6571 <listitem><para>Disallow invalid time zone names in SET TIMEZONE</para></listitem>
6575 Trim trailing spaces when <type>char</type> is cast to
6576 <type>varchar</> or <type>text</> (Tom)
6582 Make <type>float(<replaceable>p</>)</> measure the precision
6583 <replaceable>p</> in binary digits, not decimal digits
6589 <para>Add IPv6 support to the <type>inet</type> and <type>cidr</type> data types (Michael Graff)</para>
6593 <para>Add <function>family()</function> function to report whether address is IPv4 or IPv6 (Michael Graff)</para>
6598 Have <literal>SHOW datestyle</literal> generate output similar
6599 to that used by <literal>SET datestyle</literal> (Tom)
6605 Make <literal>EXTRACT(TIMEZONE)</literal> and <literal>SET/SHOW
6606 TIME ZONE</literal> follow the SQL convention for the sign of
6607 time zone offsets, i.e., positive is east from UTC (Tom)
6612 <para>Fix <literal>date_trunc('quarter', ...)</literal> (Böjthe Zoltán)</para>
6614 Prior releases returned an incorrect value for this function call.
6619 <para>Make <function>initcap()</function> more compatible with Oracle (Mike Nolan)</para>
6621 <function>initcap()</function> now uppercases a letter appearing
6622 after any non-alphanumeric character, rather than only after
6628 <para>Allow only <varname>datestyle</varname> field order for date values not in ISO-8601 format (Greg)</para>
6633 Add new <varname>datestyle</varname> values <literal>MDY</>,
6634 <literal>DMY</>, and <literal>YMD</> to set input field order;
6635 honor <literal>US</> and <literal>European</> for backward
6642 String literals like <literal>'now'</literal> or
6643 <literal>'today'</literal> will no longer work as a column
6644 default. Use functions such as <function>now()</function>,
6645 <function>current_timestamp</function> instead. (change
6646 required for prepared statements) (Tom)
6651 <para>Treat NaN as larger than any other value in <function>min()</>/<function>max()</> (Tom)</para>
6653 NaN was already sorted after ordinary numeric values for most
6654 purposes, but <function>min()</> and <function>max()</> didn't
6660 <para>Prevent interval from suppressing <literal>:00</literal>
6661 seconds display</para>
6666 New functions <function>pg_get_triggerdef(prettyprint)</function>
6667 and <function>pg_conversion_is_visible()</function> (Christopher)
6672 <para>Allow time to be specified as <literal>040506</> or <literal>0405</> (Tom)</para>
6677 Input date order must now be <literal>YYYY-MM-DD</literal> (with 4-digit year) or
6678 match <varname>datestyle</varname>
6684 Make <function>pg_get_constraintdef</function> support
6685 unique, primary-key, and check constraints (Christopher)
6692 <title>Server-Side Language Changes</title>
6697 Prevent PL/pgSQL crash when <literal>RETURN NEXT</literal> is
6698 used on a zero-row record variable (Tom)
6704 Make PL/Python's <function>spi_execute</function> interface
6705 handle null values properly (Andrew Bosma)
6710 <para>Allow PL/pgSQL to declare variables of composite types without <literal>%ROWTYPE</literal> (Tom)</para>
6714 <para>Fix PL/Python's <function>_quote()</function> function to handle big integers</para>
6718 <para>Make PL/Python an untrusted language, now called <literal>plpythonu</literal> (Kevin Jacobs, Tom)</para>
6720 The Python language no longer supports a restricted execution
6721 environment, so the trusted version of PL/Python was removed. If
6722 this situation changes, a version of PL/Python that can be used
6723 by non-superusers will be readded.
6728 <para>Allow polymorphic PL/pgSQL functions (Joe, Tom)</para>
6732 <para>Allow polymorphic SQL functions (Joe)</para>
6737 Improved compiled function caching mechanism in PL/pgSQL with
6738 full support for polymorphism (Joe)
6744 Add new parameter <literal>$0</> in PL/pgSQL representing the
6745 function's actual return type (Joe)
6751 Allow PL/Tcl and PL/Python to use the same trigger on multiple tables (Tom)
6757 Fixed PL/Tcl's <function>spi_prepare</function> to accept fully
6758 qualified type names in the parameter type list
6766 <title>psql Changes</title>
6770 <para>Add <literal>\pset pager always</literal> to always use pager (Greg)</para>
6772 This forces the pager to be used even if the number of rows is
6773 less than the screen height. This is valuable for rows that
6774 wrap across several screen rows.
6778 <listitem><para>Improve tab completion (Rod, Ross Reedstrom, Ian Barwick)</para></listitem>
6779 <listitem><para>Reorder <literal>\?</> help into groupings (Harald Armin Massa, Bruce)</para></listitem>
6780 <listitem><para>Add backslash commands for listing schemas, casts, and conversions (Christopher)</para></listitem>
6784 <command>\encoding</> now changes based on the server parameter
6785 <varname>client_encoding</varname> (Tom)
6788 In previous versions, <command>\encoding</command> was not aware
6789 of encoding changes made using <literal>SET
6790 client_encoding</literal>.
6795 <para>Save editor buffer into readline history (Ross)</para>
6797 When <command>\e</> is used to edit a query, the result is saved
6798 in the readline history for retrieval using the up arrow.
6802 <listitem><para>Improve <command>\d</command> display (Christopher)</para></listitem>
6803 <listitem><para>Enhance HTML mode to be more standards-conforming (Greg)</para></listitem>
6806 <para>New <command>\set AUTOCOMMIT off</command> capability (Tom)</para>
6808 This takes the place of the removed server parameter <varname>autocommit</varname>.
6813 <para>New <command>\set VERBOSITY</command> to control error detail (Tom)</para>
6815 This controls the new error reporting details.
6819 <listitem><para>New prompt escape sequence <literal>%x</literal> to show transaction status (Tom)</para></listitem>
6820 <listitem><para>Long options for <application>psql</application> are now available on all platforms</para></listitem>
6825 <title>pg_dump Changes</title>
6828 <listitem><para>Multiple pg_dump fixes, including tar format and large objects</para></listitem>
6829 <listitem><para>Allow pg_dump to dump specific schemas (Neil)</para></listitem>
6832 <para>Make pg_dump preserve column storage characteristics (Christopher)</para>
6834 This preserves <literal>ALTER TABLE ... SET STORAGE</literal> information.
6838 <listitem><para>Make pg_dump preserve <command>CLUSTER</command> characteristics (Christopher)</para></listitem>
6842 Have pg_dumpall use <command>GRANT</>/<command>REVOKE</> to dump database-level privleges (Tom)
6848 Allow pg_dumpall to support the options <option>-a</>,
6849 <option>-s</>, <option>-x</> of pg_dump (Tom)
6853 <listitem><para>Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)</para></listitem>
6857 pg_dump options <option>--use-set-session-authorization</option>
6858 and <option>--no-reconnect</option> now do nothing, all dumps
6859 use <command>SET SESSION AUTHORIZATION</command>
6862 pg_dump no longer reconnects to switch users, but instead always
6863 uses <command>SET SESSION AUTHORIZATION</command>. This will
6864 reduce password prompting during restores.
6869 <para>Long options for <application>pg_dump</application> are now available on all platforms</para>
6871 <productname>PostgreSQL</productname> now includes its own
6872 long-option processing routines.
6879 <title>libpq Changes</title>
6884 Add function <function>PQfreemem</function> for freeing memory on
6885 Windows, suggested for <command>NOTIFY</command> (Bruce)
6888 Windows requires that memory allocated in a library be freed by
6889 a function in the same library, hence
6890 <function>free()</function> doesn't work for freeing memory
6891 allocated by libpq. <function>PQfreemem</function> is the proper
6892 way to free libpq memory, especially on Windows, and is
6893 recommended for other platforms as well.
6898 <para>Document service capability, and add sample file (Bruce)</para>
6900 This allows clients to look up connection information in a
6901 central file on the client machine.
6907 Make <function>PQsetdbLogin</function> have the same defaults as
6908 <function>PQconnectdb</function> (Tom)
6912 <listitem><para>Allow libpq to cleanly fail when result sets are too large (Tom)</para></listitem>
6916 Improve performance of function <function>PQunescapeBytea</function> (Ben Lamb)
6922 Allow thread-safe libpq with <filename>configure</filename>
6923 option <option>--enable-thread-safety</option> (Lee Kindness,
6930 Allow function <function>pqInternalNotice</function> to accept a
6931 format string and arguments instead of just a preformatted
6932 message (Tom, Sean Chittenden)
6938 Control SSL negotiation with <literal>sslmode</literal> values
6939 <literal>disable</literal>, <literal>allow</literal>,
6940 <literal>prefer</literal>, and <literal>require</literal> (Jon
6946 <para>Allow new error codes and levels of text (Tom)</para>
6950 <para>Allow access to the underlying table and column of a query result (Tom)</para>
6952 This is helpful for query-builder applications that want to know
6953 the underlying table and column names associated with a specific
6958 <listitem><para>Allow access to the current transaction status (Tom)</para></listitem>
6959 <listitem><para>Add ability to pass binary data directly to the server (Tom)</para></listitem>
6963 Add function <function>PQexecPrepared</function> and
6964 <function>PQsendQueryPrepared</function> functions which perform
6965 bind/execute of previously prepared statements (Tom)
6972 <title>JDBC Changes</title>
6975 <listitem><para>Allow <function>setNull</function> on updateable result sets</para></listitem>
6976 <listitem><para>Allow <function>executeBatch</function> on a prepared statement (Barry)</para></listitem>
6977 <listitem><para>Support SSL connections (Barry)</para></listitem>
6978 <listitem><para>Handle schema names in result sets (Paul Sorenson)</para></listitem>
6979 <listitem><para>Add refcursor support (Nic Ferrier)</para></listitem>
6984 <title>Miscellaneous Interface Changes</title>
6988 <para>Prevent possible memory leak or core dump during libpgtcl shutdown (Tom)</para>
6991 <para>Add Informix compatibility to ECPG (Michael)</para>
6993 This allows ECPG to process embedded C programs that were
6994 written using certain Informix extensions.
6999 <para>Add type <type>decimal</type> to ECPG that is fixed length, for Informix (Michael)</para>
7004 Allow thread-safe embedded SQL programs with
7005 <filename>configure</filename> option
7006 <option>--enable-thread-safety</option> (Lee Kindness, Bruce)
7009 This allows multiple threads to access the database at the same
7015 <para>Moved Python client PyGreSQL to <ulink url="http://www.pygresql.org"></ulink> (Marc)</para>
7021 <title>Source Code Changes</title>
7024 <listitem><para>Prevent need for separate platform geometry regression result files (Tom)</para></listitem>
7025 <listitem><para>Improved PPC locking primitive (Reinhard Max)</para></listitem>
7026 <listitem><para>New function <function>palloc0</function> to allocate and clear memory (Bruce)</para></listitem>
7027 <listitem><para>Fix locking code for s390x CPU (64-bit) (Tom)</para></listitem>
7028 <listitem><para>Allow OpenBSD to use local ident credentials (William Ahern)</para></listitem>
7029 <listitem><para>Make query plan trees read-only to executor (Tom)</para></listitem>
7030 <listitem><para>Add Darwin startup scripts (David Wheeler)</para></listitem>
7031 <listitem><para>Allow libpq to compile with Borland C++ compiler (Lester Godwin, Karl Waclawek)</para></listitem>
7032 <listitem><para>Use our own version of <function>getopt_long()</function> if needed (Peter)</para></listitem>
7033 <listitem><para>Convert administration scripts to C (Peter)</para></listitem>
7034 <listitem><para> Bison >= 1.85 is now required to build the <productname>PostgreSQL</> grammar, if building from CVS</para></listitem>
7035 <listitem><para>Merge documentation into one book (Peter)</para></listitem>
7036 <listitem><para>Add Windows compatibility functions (Bruce)</para></listitem>
7037 <listitem><para>Allow client interfaces to compile under MinGW (Bruce)</para></listitem>
7038 <listitem><para>New <function>ereport()</function> function for error reporting (Tom)</para></listitem>
7039 <listitem><para>Support Intel compiler on Linux (Peter)</para></listitem>
7040 <listitem><para>Improve Linux startup scripts (Slawomir Sudnik, Darko Prenosil)</para></listitem>
7041 <listitem><para>Add support for AMD Opteron and Itanium (Jeffrey W. Baker, Bruce)</para></listitem>
7043 <para>Remove <option>--enable-recode</option> option from <command>configure</command></para>
7045 This was no longer needed now that we have <command>CREATE CONVERSION</command>.
7049 <para>Generate a compile error if spinlock code is not found (Bruce)</para>
7051 Platforms without spinlock code will now fail to compile, rather
7052 than silently using semaphores. This failure can be disabled
7053 with a new <command>configure</command> option.
7060 <title>Contrib Changes</title>
7063 <listitem><para>Change dbmirror license to BSD</para></listitem>
7064 <listitem><para>Improve earthdistance (Bruno Wolff III)</para></listitem>
7065 <listitem><para>Portability improvements to pgcrypto (Marko Kreen)</para></listitem>
7066 <listitem><para>Prevent crash in xml (John Gray, Michael Richards)</para></listitem>
7067 <listitem><para>Update oracle</para></listitem>
7068 <listitem><para>Update mysql</para></listitem>
7069 <listitem><para>Update cube (Bruno Wolff III)</para></listitem>
7070 <listitem><para>Update earthdistance to use cube (Bruno Wolff III)</para></listitem>
7071 <listitem><para>Update btree_gist (Oleg)</para></listitem>
7072 <listitem><para>New tsearch2 full-text search module (Oleg, Teodor)</para></listitem>
7073 <listitem><para>Add hash-based crosstab function to tablefuncs (Joe)</para></listitem>
7074 <listitem><para>Add serial column to order <function>connectby()</> siblings in tablefuncs (Nabil Sayegh,Joe)</para></listitem>
7075 <listitem><para>Add named persistent connections to dblink (Shridhar Daithanka)</para></listitem>
7076 <listitem><para>New pg_autovacuum allows automatic <command>VACUUM</command> (Matthew T. O'Connor)</para></listitem>
7077 <listitem><para>Make pgbench honor environment variables <envar>PGHOST</>, <envar>PGPORT</>, <envar>PGUSER</> (Tatsuo)</para></listitem>
7078 <listitem><para>Improve intarray (Teodor Sigaev)</para></listitem>
7079 <listitem><para>Improve pgstattuple (Rod)</para></listitem>
7080 <listitem><para>Fix bug in <function>metaphone()</function> in fuzzystrmatch</para></listitem>
7081 <listitem><para>Improve adddepend (Rod)</para></listitem>
7082 <listitem><para>Update spi/timetravel (Böjthe Zoltán)</para></listitem>
7083 <listitem><para>Fix dbase <option>-s</> option and improve non-ASCII handling (Thomas Behr, Márcio Smiderle)</para></listitem>
7084 <listitem><para>Remove array module because features now included by default (Joe)</para></listitem>
7090 <sect1 id="release-7-3-10">
7091 <title>Release 7.3.10</title>
7094 <title>Release date</title>
7095 <simpara>2005-05-09</simpara>
7099 This release contains a variety of fixes from 7.3.9, including several
7100 security-related issues.
7104 <title>Migration to version 7.3.10</title>
7107 A dump/restore is not required for those running 7.3.X. However,
7108 it is one possible way of handling a significant security problem
7109 that has been found in the initial contents of 7.3.X system
7110 catalogs. A dump/initdb/reload sequence using 7.3.10's initdb will
7111 automatically correct this problem.
7115 The security problem is that the built-in character set encoding
7116 conversion functions can be invoked from SQL commands by unprivileged
7117 users, but the functions were not designed for such use and are not
7118 secure against malicious choices of arguments. The fix involves changing
7119 the declared parameter list of these functions so that they can no longer
7120 be invoked from SQL commands. (This does not affect their normal use
7121 by the encoding conversion machinery.)
7122 It is strongly recommended that all installations repair this error,
7123 either by initdb or by following the manual repair procedure given
7124 below. The error at least allows unprivileged database users to crash
7125 their server process, and may allow unprivileged users to gain the
7126 privileges of a database superuser.
7130 If you wish not to do an initdb, perform the following procedure instead.
7131 As the database superuser, do:
7135 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
7136 WHERE pronamespace = 11 AND pronargs = 5
7137 AND proargtypes[2] = 'cstring'::regtype;
7138 -- The command should report having updated 90 rows;
7139 -- if not, rollback and investigate instead of committing!
7145 The above procedure must be carried out in <emphasis>each</> database
7146 of an installation, including <literal>template1</>, and ideally
7147 including <literal>template0</> as well. If you do not fix the
7148 template databases then any subsequently created databases will contain
7149 the same error. <literal>template1</> can be fixed in the same way
7150 as any other database, but fixing <literal>template0</> requires
7151 additional steps. First, from any database issue
7153 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
7155 Next connect to <literal>template0</> and perform the above repair
7156 procedure. Finally, do
7158 -- re-freeze template0:
7160 -- and protect it against future alterations:
7161 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
7167 <title>Changes</title>
7170 <listitem><para>Change encoding function signature to prevent
7171 misuse</para></listitem>
7172 <listitem><para>Repair ancient race condition that allowed a transaction to be
7173 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
7174 than for other purposes</para>
7175 <para>This is an extremely serious bug since it could lead to apparent
7176 data inconsistencies being briefly visible to applications.</para></listitem>
7177 <listitem><para>Repair race condition between relation extension and
7179 <para>This could theoretically have caused loss of a page's worth of
7180 freshly-inserted data, although the scenario seems of very low probability.
7181 There are no known cases of it having caused more than an Assert failure.
7183 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
7185 The comparison code was wrong in the case where the
7186 <literal>--enable-integer-datetimes</> configuration switch had been used.
7187 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
7188 it will need to be <command>REINDEX</>ed after installing this update, because
7189 the fix corrects the sort order of column values.
7191 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
7192 <type>TIME WITH TIME ZONE</> values</para></listitem>
7193 <listitem><para>Fix mis-display of negative fractional seconds in
7194 <type>INTERVAL</> values</para>
7196 This error only occurred when the
7197 <literal>--enable-integer-datetimes</> configuration switch had been used.
7199 <listitem><para>Additional buffer overrun checks in plpgsql
7200 (Neil)</para></listitem>
7201 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
7202 correctly (Neil)</para></listitem>
7203 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
7204 month-related formats</para></listitem>
7205 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
7206 (Marko Kreen)</para></listitem>
7207 <listitem><para>Still more 64-bit fixes for
7208 <filename>contrib/intagg</></para></listitem>
7209 <listitem><para>Prevent incorrect optimization of functions returning
7210 <type>RECORD</></para></listitem>
7216 <sect1 id="release-7-3-9">
7217 <title>Release 7.3.9</title>
7220 <title>Release date</title>
7221 <simpara>2005-01-31</simpara>
7225 This release contains a variety of fixes from 7.3.8, including several
7226 security-related issues.
7230 <title>Migration to version 7.3.9</title>
7233 A dump/restore is not required for those running 7.3.X.
7238 <title>Changes</title>
7241 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
7243 On platforms that will automatically execute initialization functions of a
7244 shared library (this includes at least Windows and ELF-based Unixen),
7245 <command>LOAD</> can be used to make the server execute arbitrary code.
7246 Thanks to NGS Software for reporting this.</para></listitem>
7247 <listitem><para>Check that creator of an aggregate function has the right to
7248 execute the specified transition functions</para>
7250 This oversight made it possible to bypass denial of EXECUTE
7251 permission on a function.</para></listitem>
7252 <listitem><para>Fix security and 64-bit issues in
7253 contrib/intagg</para></listitem>
7254 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
7255 Jurka)</para></listitem>
7256 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
7257 many parameters (Neil)</para></listitem>
7258 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
7260 The result of the join was mistakenly supposed to be sorted the same as the
7261 left input. This could not only deliver mis-sorted output to the user, but
7262 in case of nested merge joins could give outright wrong answers.
7264 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
7265 <listitem><para>Fix display of negative intervals in SQL and GERMAN
7266 datestyles</para></listitem>
7272 <sect1 id="release-7-3-8">
7273 <title>Release 7.3.8</title>
7276 <title>Release date</title>
7277 <simpara>2004-10-22</simpara>
7281 This release contains a variety of fixes from 7.3.7.
7286 <title>Migration to version 7.3.8</title>
7289 A dump/restore is not required for those running 7.3.X.
7294 <title>Changes</title>
7297 <listitem><para>Repair possible failure to update hint bits on disk</para>
7299 Under rare circumstances this oversight could lead to
7300 <quote>could not access transaction status</> failures, which qualifies
7301 it as a potential-data-loss bug.
7303 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
7305 Very large left joins using a hash join plan could fail to output unmatched
7306 left-side rows given just the right data distribution.
7308 <listitem><para>Disallow running pg_ctl as root</para>
7310 This is to guard against any possible security issues.
7312 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
7314 This has been reported as a security issue, though it's hardly worthy of
7315 concern since there is no reason for non-developers to use this script anyway.
7322 <sect1 id="release-7-3-7">
7323 <title>Release 7.3.7</title>
7326 <title>Release date</title>
7327 <simpara>2004-08-16</simpara>
7331 This release contains one critical fix over 7.3.6, and some minor items.
7336 <title>Migration to version 7.3.7</title>
7339 A dump/restore is not required for those running 7.3.X.
7344 <title>Changes</title>
7347 <listitem><para>Prevent possible loss of committed transactions during crash</para>
7349 Due to insufficient interlocking between transaction commit and checkpointing,
7350 it was possible for transactions committed just before the most recent
7351 checkpoint to be lost, in whole or in part, following a database crash and
7352 restart. This is a serious bug that has existed
7353 since <productname>PostgreSQL</productname> 7.1.
7355 <listitem><para>Remove asymmetrical word processing in tsearch (Teodor)</para></listitem>
7356 <listitem><para>Properly schema-qualify function names when pg_dump'ing a CAST</para></listitem>
7362 <sect1 id="release-7-3-6">
7363 <title>Release 7.3.6</title>
7366 <title>Release date</title>
7367 <simpara>2004-03-02</simpara>
7371 This release contains a variety of fixes from 7.3.5.
7376 <title>Migration to version 7.3.6</title>
7379 A dump/restore is <emphasis>not</emphasis> required for those
7386 <title>Changes</title>
7389 <listitem><para>Revert erroneous changes in rule permissions checking</para>
7390 <para>A patch applied in 7.3.3 to fix a corner case in rule permissions checks
7391 turns out to have disabled rule-related permissions checks in many
7392 not-so-corner cases. This would for example allow users to insert into views
7393 they weren't supposed to have permission to insert into. We have therefore
7394 reverted the 7.3.3 patch. The original bug will be fixed in 8.0.
7396 <listitem><para>Repair incorrect order of operations in
7397 GetNewTransactionId()</para>
7399 This bug could result in failure under out-of-disk-space conditions, including
7400 inability to restart even after disk space is freed.
7402 <listitem><para>Ensure configure selects -fno-strict-aliasing even when
7403 an external value for CFLAGS is supplied</para>
7405 On some platforms, building with -fstrict-aliasing causes bugs.
7407 <listitem><para>Make pg_restore handle 64-bit off_t correctly</para>
7409 This bug prevented proper restoration from archive files exceeding 4Gb.
7411 <listitem><para>Make contrib/dblink not assume that local and remote type OIDs
7412 match (Joe)</para></listitem>
7413 <listitem><para>Quote connectby()'s start_with argument properly (Joe)</para></listitem>
7414 <listitem><para>Don't crash when a rowtype argument to a plpgsql function is
7415 NULL</para></listitem>
7416 <listitem><para>Avoid generating invalid character encoding sequences in
7417 corner cases when planning LIKE operations</para></listitem>
7418 <listitem><para>Ensure text_position() cannot scan past end of source string
7419 in multibyte cases (Korea PostgreSQL Users' Group)</para></listitem>
7420 <listitem><para>Fix index optimization and selectivity estimates for LIKE
7421 operations on bytea columns (Joe)</para></listitem>
7427 <sect1 id="release-7-3-5">
7428 <title>Release 7.3.5</title>
7431 <title>Release date</title>
7432 <simpara>2003-12-03</simpara>
7436 This has a variety of fixes from 7.3.4.
7441 <title>Migration to version 7.3.5</title>
7444 A dump/restore is <emphasis>not</emphasis> required for those
7450 <title>Changes</title>
7453 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
7454 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
7455 <listitem><para>Force stats processes to detach from shared memory, ensuring cleaner shutdown</para></listitem>
7456 <listitem><para>Make PQescapeBytea and byteaout consistent with each other (Joe)</para></listitem>
7457 <listitem><para>Added missing SPI_finish() calls to dblink's get_tuple_of_interest() (Joe)</para></listitem>
7458 <listitem><para>Fix for possible foreign key violation when rule rewrites INSERT (Jan)</para></listitem>
7459 <listitem><para>Support qualified type names in PL/Tcl's spi_prepare command (Jan)</para></listitem>
7460 <listitem><para>Make pg_dump handle a procedural language handler located in pg_catalog</para></listitem>
7461 <listitem><para>Make pg_dump handle cases where a custom opclass is in another schema</para></listitem>
7462 <listitem><para>Make pg_dump dump binary-compatible casts correctly (Jan)</para></listitem>
7463 <listitem><para>Fix insertion of expressions containing subqueries into rule bodies</para></listitem>
7464 <listitem><para>Fix incorrect argument processing in clusterdb script (Anand Ranganathan)</para></listitem>
7465 <listitem><para>Fix problems with dropped columns in plpython triggers</para></listitem>
7466 <listitem><para>Repair problems with to_char() reading past end of its input string (Karel)</para></listitem>
7467 <listitem><para>Fix GB18030 mapping errors (Tatsuo)</para></listitem>
7468 <listitem><para>Fix several problems with SSL error handling and asynchronous SSL I/O</para></listitem>
7469 <listitem><para>Remove ability to bind a list of values to a single parameter in JDBC
7470 (prevents possible SQL-injection attacks)</para></listitem>
7471 <listitem><para>Fix some errors in HAVE_INT64_TIMESTAMP code paths</para></listitem>
7472 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
7478 <sect1 id="release-7-3-4">
7479 <title>Release 7.3.4</title>
7482 <title>Release date</title>
7483 <simpara>2003-07-24</simpara>
7487 This has a variety of fixes from 7.3.3.
7492 <title>Migration to version 7.3.4</title>
7495 A dump/restore is <emphasis>not</emphasis> required for those
7501 <title>Changes</title>
7504 <listitem><para>Repair breakage in timestamp-to-date conversion for dates before 2000</para></listitem>
7505 <listitem><para>Prevent rare possibility of server startup failure (Tom)</para></listitem>
7506 <listitem><para>Fix bugs in interval-to-time conversion (Tom)</para></listitem>
7507 <listitem><para>Add constraint names in a few places in pg_dump (Rod)</para></listitem>
7508 <listitem><para>Improve performance of functions with many parameters (Tom)</para></listitem>
7509 <listitem><para>Fix to_ascii() buffer overruns (Tom)</para></listitem>
7510 <listitem><para>Prevent restore of database comments from throwing an error (Tom)</para></listitem>
7511 <listitem><para>Work around buggy strxfrm() present in some Solaris releases (Tom)</para></listitem>
7512 <listitem><para>Properly escape jdbc setObject() strings to improve security (Barry)</para></listitem>
7518 <sect1 id="release-7-3-3">
7519 <title>Release 7.3.3</title>
7522 <title>Release date</title>
7523 <simpara>2003-05-22</simpara>
7527 This release contains a variety of fixes for version 7.3.2.
7531 <title>Migration to version 7.3.3</title>
7534 A dump/restore is <emphasis>not</emphasis> required for those
7535 running version 7.3.*.
7540 <title>Changes</title>
7543 <listitem><para>Repair sometimes-incorrect computation of StartUpID after a crash</para></listitem>
7544 <listitem><para>Avoid slowness with lots of deferred triggers in one transaction (Stephan)</para></listitem>
7545 <listitem><para>Don't lock referenced row when <command>UPDATE</command> doesn't change foreign key's value (Jan)</para></listitem>
7546 <listitem><para>Use <command>-fPIC</command> not <command>-fpic</command> on Sparc (Tom Callaway)</para></listitem>
7547 <listitem><para>Repair lack of schema-awareness in contrib/reindexdb</para></listitem>
7548 <listitem><para>Fix contrib/intarray error for zero-element result array (Teodor)</para></listitem>
7549 <listitem><para>Ensure createuser script will exit on control-C (Oliver)</para></listitem>
7550 <listitem><para>Fix errors when the type of a dropped column has itself been dropped</para></listitem>
7551 <listitem><para><command>CHECKPOINT</command> does not cause database panic on failure in noncritical steps</para></listitem>
7552 <listitem><para>Accept 60 in seconds fields of timestamp, time, interval input values</para></listitem>
7553 <listitem><para>Issue notice, not error, if <type>TIMESTAMP</type>,
7554 <type> TIME</type>, or <type>INTERVAL</type> precision too large</para></listitem>
7555 <listitem><para>Fix <function>abstime-to-time</function> cast function (fix is
7556 not applied unless you <application>initdb</application>)</para></listitem>
7557 <listitem><para>Fix <application>pg_proc</application> entry for
7558 <type>timestampt_izone</type> (fix is not applied unless you
7559 <application>initdb</application>)</para></listitem>
7560 <listitem><para>Make <function>EXTRACT(EPOCH FROM timestamp without time zone)</function> treat input as local time</para></listitem>
7561 <listitem><para><command>'now'::timestamptz</command> gave wrong answer if timezone changed earlier in transaction</para></listitem>
7562 <listitem><para><envar>HAVE_INT64_TIMESTAMP</envar> code for time with timezone overwrote its input</para></listitem>
7563 <listitem><para>Accept <command>GLOBAL TEMP/TEMPORARY</command> as a
7564 synonym for <command>TEMPORARY</command></para></listitem>
7565 <listitem><para>Avoid improper schema-privilege-check failure in foreign-key triggers</para></listitem>
7566 <listitem><para>Fix bugs in foreign-key triggers for <command>SET DEFAULT</command> action</para></listitem>
7567 <listitem><para>Fix incorrect time-qual check in row fetch for
7568 <command>UPDATE</command> and <command>DELETE</command> triggers</para></listitem>
7569 <listitem><para>Foreign-key clauses were parsed but ignored in
7570 <command>ALTER TABLE ADD COLUMN</command></para></listitem>
7571 <listitem><para>Fix createlang script breakage for case where handler function already exists</para></listitem>
7572 <listitem><para>Fix misbehavior on zero-column tables in <application>pg_dump</application>, COPY, ANALYZE, other places</para></listitem>
7573 <listitem><para>Fix misbehavior of <function>func_error()</function> on type names containing '%'</para></listitem>
7574 <listitem><para>Fix misbehavior of <function>replace()</function> on strings containing '%'</para></listitem>
7575 <listitem><para>Regular-expression patterns containing certain multibyte characters failed</para></listitem>
7576 <listitem><para>Account correctly for <command>NULL</command>s in more cases in join size estimation</para></listitem>
7577 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
7578 <listitem><para>Fix failure to convert large code point values in EUC_TW conversions (Tatsuo)</para></listitem>
7579 <listitem><para>Fix error recovery for <function>SSL_read</function>/<function>SSL_write</function> calls</para></listitem>
7580 <listitem><para>Don't do early constant-folding of type coercion expressions</para></listitem>
7581 <listitem><para>Validate page header fields immediately after reading in any page</para></listitem>
7582 <listitem><para>Repair incorrect check for ungrouped variables in unnamed joins</para></listitem>
7583 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
7584 <listitem><para>contrib/ltree fixes (Teodor)</para></listitem>
7585 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
7586 <listitem><para>Avoid running out of buffers in many-way indexscan (bug introduced in 7.3)</para></listitem>
7587 <listitem><para>Fix planner's selectivity estimation functions to handle domains properly</para></listitem>
7588 <listitem><para>Fix <application>dbmirror</application> memory-allocation bug (Steven Singer)</para></listitem>
7589 <listitem><para>Prevent infinite loop in <function>ln(numeric)</function> due to roundoff error</para></listitem>
7590 <listitem><para><command>GROUP BY</command> got confused if there were multiple equal GROUP BY items</para></listitem>
7591 <listitem><para>Fix bad plan when inherited <command>UPDATE</command>/<command>DELETE</command> references another inherited table</para></listitem>
7592 <listitem><para>Prevent clustering on incomplete (partial or non-NULL-storing) indexes</para></listitem>
7593 <listitem><para>Service shutdown request at proper time if it arrives while still starting up</para></listitem>
7594 <listitem><para>Fix left-links in temporary indexes (could make backwards scans miss entries)</para></listitem>
7595 <listitem><para>Fix incorrect handling of client_encoding setting in postgresql.conf (Tatsuo)</para></listitem>
7596 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
7597 <listitem><para>Fix SPI for case where rule contains multiple statements of the same type</para></listitem>
7598 <listitem><para>Fix problem with checking for wrong type of access privilege in rule query</para></listitem>
7599 <listitem><para>Fix problem with <command>EXCEPT</command> in <command>CREATE RULE</command></para></listitem>
7600 <listitem><para>Prevent problem with dropping temp tables having serial columns</para></listitem>
7601 <listitem><para>Fix replace_vars_with_subplan_refs failure in complex views</para></listitem>
7602 <listitem><para>Fix regexp slowness in single-byte encodings (Tatsuo)</para></listitem>
7603 <listitem><para>Allow qualified type names in <command>CREATE CAST</command>
7604 and <command> DROP CAST</command></para></listitem>
7605 <listitem><para>Accept <function>SETOF type[]</function>, which formerly had to
7606 be written <function>SETOF _type</function></para></listitem>
7607 <listitem><para>Fix <application>pg_dump</application> core dump in some cases with procedural languages</para></listitem>
7608 <listitem><para>Force ISO datestyle in <application>pg_dump</application> output, for portability (Oliver)</para></listitem>
7609 <listitem><para><application>pg_dump</application> failed to handle error return
7610 from <function>lo_read</function> (Oleg Drokin)</para></listitem>
7611 <listitem><para><application>pg_dumpall</application> failed with groups having no members (Nick Eskelinen)</para></listitem>
7612 <listitem><para><application>pg_dumpall</application> failed to recognize --globals-only switch</para></listitem>
7613 <listitem><para>pg_restore failed to restore blobs if -X disable-triggers is specified</para></listitem>
7614 <listitem><para>Repair intrafunction memory leak in plpgsql</para></listitem>
7615 <listitem><para>pltcl's <command>elog</command> command dumped core if given wrong parameters (Ian Harding)</para></listitem>
7616 <listitem><para>plpython used wrong value of <envar>atttypmod</envar> (Brad McLean)</para></listitem>
7617 <listitem><para>Fix improper quoting of boolean values in Python interface (D'Arcy)</para></listitem>
7618 <listitem><para>Added <function>addDataType()</function> method to PGConnection interface for JDBC</para></listitem>
7619 <listitem><para>Fixed various problems with updateable ResultSets for JDBC (Shawn Green)</para></listitem>
7620 <listitem><para>Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka, Peter Royal)</para></listitem>
7621 <listitem><para>Fixed problem with parsing table ACLs in JDBC</para></listitem>
7622 <listitem><para>Better error message for character set conversion problems in JDBC</para></listitem>
7628 <sect1 id="release-7-3-2">
7629 <title>Release 7.3.2</title>
7632 <title>Release date</title>
7633 <simpara>2003-02-04</simpara>
7637 This release contains a variety of fixes for version 7.3.1.
7642 <title>Migration to version 7.3.2</title>
7645 A dump/restore is <emphasis>not</emphasis> required for those
7646 running version 7.3.*.
7651 <title>Changes</title>
7654 <listitem><para>Restore creation of OID column in CREATE TABLE AS / SELECT INTO</para></listitem>
7655 <listitem><para>Fix <application>pg_dump</> core dump when dumping views having comments</para></listitem>
7656 <listitem><para>Dump DEFERRABLE/INITIALLY DEFERRED constraints properly</para></listitem>
7657 <listitem><para>Fix UPDATE when child table's column numbering differs from parent</para></listitem>
7658 <listitem><para>Increase default value of max_fsm_relations</para></listitem>
7659 <listitem><para>Fix problem when fetching backwards in a cursor for a single-row query</para></listitem>
7660 <listitem><para>Make backward fetch work properly with cursor on SELECT DISTINCT query</para></listitem>
7661 <listitem><para>Fix problems with loading <application>pg_dump</> files containing contrib/lo usage</para></listitem>
7662 <listitem><para>Fix problem with all-numeric user names</para></listitem>
7663 <listitem><para>Fix possible memory leak and core dump during disconnect in libpgtcl</para></listitem>
7664 <listitem><para>Make plpython's spi_execute command handle nulls properly (Andrew Bosma)</para></listitem>
7665 <listitem><para>Adjust plpython error reporting so that its regression test passes again</para></listitem>
7666 <listitem><para>Work with bison 1.875</para></listitem>
7667 <listitem><para>Handle mixed-case names properly in plpgsql's %type (Neil)</para></listitem>
7668 <listitem><para>Fix core dump in pltcl when executing a query rewritten by a rule</para></listitem>
7669 <listitem><para>Repair array subscript overruns (per report from Yichen Xie)</para></listitem>
7670 <listitem><para>Reduce MAX_TIME_PRECISION from 13 to 10 in floating-point case</para></listitem>
7671 <listitem><para>Correctly case-fold variable names in per-database and per-user settings</para></listitem>
7672 <listitem><para>Fix coredump in plpgsql's RETURN NEXT when SELECT into record returns no rows</para></listitem>
7673 <listitem><para>Fix outdated use of pg_type.typprtlen in python client interface</para></listitem>
7674 <listitem><para>Correctly handle fractional seconds in timestamps in JDBC driver</para></listitem>
7675 <listitem><para>Improve performance of getImportedKeys() in JDBC</para></listitem>
7676 <listitem><para>Make shared-library symlinks work standardly on HPUX (Giles)</para></listitem>
7677 <listitem><para>Repair inconsistent rounding behavior for timestamp, time, interval</para></listitem>
7678 <listitem><para>SSL negotiation fixes (Nathan Mueller)</para></listitem>
7679 <listitem><para>Make libpq's ~/.pgpass feature work when connecting with PQconnectDB</para></listitem>
7680 <listitem><para>Update my2pg, ora2pg</para></listitem>
7681 <listitem><para>Translation updates</para></listitem>
7682 <listitem><para>Add casts between types lo and oid in contrib/lo</para></listitem>
7683 <listitem><para>fastpath code now checks for privilege to call function</para></listitem>
7689 <sect1 id="release-7-3-1">
7690 <title>Release 7.3.1</title>
7693 <title>Release date</title>
7694 <simpara>2002-12-18</simpara>
7698 This release contains a variety of fixes for version 7.3.
7703 <title>Migration to version 7.3.1</title>
7706 A dump/restore is <emphasis>not</emphasis> required for those
7707 running version 7.3. However, it should be noted that the main
7708 <productname>PostgreSQL</productname> interface library, libpq,
7709 has a new major version number for this release, which may require
7710 recompilation of client code in certain cases.
7715 <title>Changes</title>
7718 <listitem><para>Fix a core dump of COPY TO when client/server encodings don't match (Tom)</para></listitem>
7719 <listitem><para>Allow <application>pg_dump</> to work with pre-7.2 servers (Philip)</para></listitem>
7720 <listitem><para>contrib/adddepend fixes (Tom)</para></listitem>
7721 <listitem><para>Fix problem with deletion of per-user/per-database config settings (Tom)</para></listitem>
7722 <listitem><para>contrib/vacuumlo fix (Tom)</para></listitem>
7723 <listitem><para>Allow 'password' encryption even when pg_shadow contains MD5 passwords (Bruce)</para></listitem>
7724 <listitem><para>contrib/dbmirror fix (Steven Singer)</para></listitem>
7725 <listitem><para>Optimizer fixes (Tom)</para></listitem>
7726 <listitem><para>contrib/tsearch fixes (Teodor Sigaev, Magnus)</para></listitem>
7727 <listitem><para>Allow locale names to be mixed case (Nicolai Tufar)</para></listitem>
7728 <listitem><para>Increment libpq library's major version number (Bruce)</para></listitem>
7729 <listitem><para>pg_hba.conf error reporting fixes (Bruce, Neil)</para></listitem>
7730 <listitem><para>Add SCO Openserver 5.0.4 as a supported platform (Bruce)</para></listitem>
7731 <listitem><para>Prevent EXPLAIN from crashing server (Tom)</para></listitem>
7732 <listitem><para>SSL fixes (Nathan Mueller)</para></listitem>
7733 <listitem><para>Prevent composite column creation via ALTER TABLE (Tom)</para></listitem>
7739 <sect1 id="release-7-3">
7740 <title>Release 7.3</title>
7743 <title>Release date</title>
7744 <simpara>2002-11-27</simpara>
7748 <title>Overview</title>
7751 Major changes in this release:
7756 <term>Schemas</term>
7759 Schemas allow users to create objects in separate namespaces,
7760 so two people or applications can have tables with the same
7761 name. There is also a public schema for shared tables.
7762 Table/index creation can be restricted by removing privileges
7763 on the public schema.
7769 <term>Drop Column</term>
7772 <productname>PostgreSQL</productname> now supports the
7773 <literal>ALTER TABLE ... DROP COLUMN</literal> functionality.
7779 <term>Table Functions</term>
7782 Functions returning multiple rows and/or multiple columns are
7783 now much easier to use than before. You can call such a
7784 <quote>table function</quote> in the <literal>SELECT</literal>
7785 <literal>FROM</literal> clause, treating its output like a
7786 table. Also, <application>PL/pgSQL</application> functions can
7793 <term>Prepared Queries</term>
7796 <productname>PostgreSQL</productname> now supports prepared
7797 queries, for improved performance.
7803 <term>Dependency Tracking</term>
7806 <productname>PostgreSQL</productname> now records object
7807 dependencies, which allows improvements in many areas.
7808 <command>DROP</command> statements now take either
7809 <literal>CASCADE</> or <literal>RESTRICT</> to control whether
7810 dependent objects are also dropped.
7816 <term>Privileges</term>
7819 Functions and procedural languages now have privileges, and
7820 functions can be defined to run with the privileges of their
7827 <term>Internationalization</term>
7830 Both multibyte and locale support are now always enabled.
7836 <term>Logging</term>
7839 A variety of logging options have been enhanced.
7845 <term>Interfaces</term>
7848 A large number of interfaces have been moved to <ulink
7849 url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
7850 where they can be developed and released independently.
7856 <term>Functions/Identifiers</term>
7859 By default, functions can now take up to 32 parameters, and
7860 identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
7861 is now deprecated: there are specific <quote>pseudo-datatypes</>
7862 to represent each of the former meanings of <literal>OPAQUE</>
7863 in function argument and result types.
7872 <title>Migration to version 7.3</title>
7875 A dump/restore using <application>pg_dump</> is required for those
7876 wishing to migrate data from any previous release. If your
7877 application examines the system catalogs, additional changes will
7878 be required due to the introduction of schemas in 7.3; for more
7879 information, see: <ulink
7880 url="http://developer.postgresql.org/~momjian/upgrade_tips_7.3"></>.
7884 Observe the following incompatibilities:
7890 Pre-6.3 clients are no longer supported.
7896 <filename>pg_hba.conf</filename> now has a column for the user
7897 name and additional features. Existing files need to be
7904 Several <filename>postgresql.conf</filename> logging parameters
7911 <literal>LIMIT #,#</literal> has been disabled; use
7912 <literal>LIMIT # OFFSET #</literal>.
7918 <command>INSERT</command> statements with column lists must
7919 specify a value for each specified column. For example,
7920 <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
7921 is now invalid. It's still allowed to supply fewer columns than
7922 expected if the <command>INSERT</command> does not have a column list.
7928 <type>serial</type> columns are no longer automatically
7929 <literal>UNIQUE</>; thus, an index will not automatically be
7936 A <command>SET</command> command inside an aborted transaction
7943 <command>COPY</command> no longer considers missing trailing
7944 columns to be null. All columns need to be specified.
7945 (However, one may achieve a similar effect by specifying a
7946 column list in the <command>COPY</command> command.)
7952 The data type <type>timestamp</type> is now equivalent to
7953 <type>timestamp without time zone</type>, instead of
7954 <type>timestamp with time zone</type>.
7960 Pre-7.3 databases loaded into 7.3 will not have the new object
7961 dependencies for <type>serial</type> columns, unique
7962 constraints, and foreign keys. See the directory
7963 <filename>contrib/adddepend/</filename> for a detailed
7964 description and a script that will add such dependencies.
7970 An empty string (<literal>''</literal>) is no longer allowed as
7971 the input into an integer field. Formerly, it was silently
7980 <title>Changes</title>
7983 <title>Server Operation</title>
7985 <listitem><para>Add pg_locks view to show locks (Neil)</para></listitem>
7986 <listitem><para>Security fixes for password negotiation memory allocation (Neil)</para></listitem>
7987 <listitem><para>Remove support for version 0 FE/BE protocol (<productname>PostgreSQL</productname> 6.2 and earlier) (Tom)</para></listitem>
7988 <listitem><para>Reserve the last few backend slots for superusers, add parameter superuser_reserved_connections to control this (Nigel J. Andrews)</para></listitem>
7993 <title>Performance</title>
7995 <listitem><para>Improve startup by calling localtime() only once (Tom)</para></listitem>
7996 <listitem><para>Cache system catalog information in flat files for faster startup (Tom)</para></listitem>
7997 <listitem><para>Improve caching of index information (Tom)</para></listitem>
7998 <listitem><para>Optimizer improvements (Tom, Fernando Nasser)</para></listitem>
7999 <listitem><para>Catalog caches now store failed lookups (Tom)</para></listitem>
8000 <listitem><para>Hash function improvements (Neil)</para></listitem>
8001 <listitem><para>Improve performance of query tokenization and network handling (Peter)</para></listitem>
8002 <listitem><para>Speed improvement for large object restore (Mario Weilguni)</para></listitem>
8003 <listitem><para>Mark expired index entries on first lookup, saving later heap fetches (Tom)</para></listitem>
8004 <listitem><para>Avoid excessive NULL bitmap padding (Manfred Koizar)</para></listitem>
8005 <listitem><para>Add BSD-licensed qsort() for Solaris, for performance (Bruce)</para></listitem>
8006 <listitem><para>Reduce per-row overhead by four bytes (Manfred Koizar)</para></listitem>
8007 <listitem><para>Fix GEQO optimizer bug (Neil Conway)</para></listitem>
8008 <listitem><para>Make WITHOUT OID actually save four bytes per row (Manfred Koizar)</para></listitem>
8009 <listitem><para>Add default_statistics_target variable to specify ANALYZE buckets (Neil)</para></listitem>
8010 <listitem><para>Use local buffer cache for temporary tables so no WAL overhead (Tom)</para></listitem>
8011 <listitem><para>Improve free space map performance on large tables (Stephen Marshall, Tom)</para></listitem>
8012 <listitem><para>Improved WAL write concurrency (Tom)</para></listitem>
8017 <title>Privileges</title>
8019 <listitem><para>Add privileges on functions and procedural languages (Peter)</para></listitem>
8020 <listitem><para>Add OWNER to CREATE DATABASE so superusers can create databases on behalf of unprivileged users (Gavin Sherry, Tom)</para></listitem>
8021 <listitem><para>Add new object privilege bits EXECUTE and USAGE (Tom)</para></listitem>
8022 <listitem><para>Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION (Tom)</para></listitem>
8023 <listitem><para>Allow functions to be executed with the privilege of the function owner (Peter)</para></listitem>
8028 <title>Server Configuration</title>
8030 <listitem><para>Server log messages now tagged with LOG, not DEBUG (Bruce)</para></listitem>
8031 <listitem><para>Add user column to pg_hba.conf (Bruce)</para></listitem>
8032 <listitem><para>Have log_connections output two lines in log file (Tom)</para></listitem>
8033 <listitem><para>Remove debug_level from postgresql.conf, now server_min_messages (Bruce)</para></listitem>
8034 <listitem><para>New ALTER DATABASE/USER ... SET command for per-user/database initialization (Peter)</para></listitem>
8035 <listitem><para>New parameters server_min_messages and client_min_messages to control which messages are sent to the server logs or client applications (Bruce)</para></listitem>
8036 <listitem><para>Allow pg_hba.conf to specify lists of users/databases separated by commas, group names prepended with +, and file names prepended with @ (Bruce)</para></listitem>
8037 <listitem><para>Remove secondary password file capability and pg_password utility (Bruce)</para></listitem>
8038 <listitem><para>Add variable db_user_namespace for database-local user names (Bruce)</para></listitem>
8039 <listitem><para>SSL improvements (Bear Giles)</para></listitem>
8040 <listitem><para>Make encryption of stored passwords the default (Bruce)</para></listitem>
8041 <listitem><para>Allow pg_statistics to be reset by calling pg_stat_reset() (Christopher)</para></listitem>
8042 <listitem><para>Add log_duration parameter (Bruce)</para></listitem>
8043 <listitem><para>Rename debug_print_query to log_statement (Bruce)</para></listitem>
8044 <listitem><para>Rename show_query_stats to show_statement_stats (Bruce)</para></listitem>
8045 <listitem><para>Add param log_min_error_statement to print commands to logs on error (Gavin)</para></listitem>
8050 <title>Queries</title>
8052 <listitem><para>Make cursors insensitive, meaning their contents do not change (Tom)</para></listitem>
8053 <listitem><para>Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)</para></listitem>
8054 <listitem><para>Increase identifier length to 63 (Neil, Bruce)</para></listitem>
8055 <listitem><para>UNION fixes for merging >= 3 columns of different lengths (Tom)</para></listitem>
8056 <listitem><para>Add DEFAULT key word to INSERT, e.g., INSERT ... (..., DEFAULT, ...) (Rod)</para></listitem>
8057 <listitem><para>Allow views to have default values using ALTER COLUMN ... SET DEFAULT (Neil)</para></listitem>
8058 <listitem><para>Fail on INSERTs with column lists that don't supply all column values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)</para></listitem>
8059 <listitem><para>Fix for join aliases (Tom)</para></listitem>
8060 <listitem><para>Fix for FULL OUTER JOINs (Tom)</para></listitem>
8061 <listitem><para>Improve reporting of invalid identifier and location (Tom, Gavin)</para></listitem>
8062 <listitem><para>Fix OPEN cursor(args) (Tom)</para></listitem>
8063 <listitem><para>Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)</para></listitem>
8064 <listitem><para>Fix for CREATE TABLE AS with UNION (Tom)</para></listitem>
8065 <listitem><para>SQL99 syntax improvements (Thomas)</para></listitem>
8066 <listitem><para>Add statement_timeout variable to cancel queries (Bruce)</para></listitem>
8067 <listitem><para>Allow prepared queries with PREPARE/EXECUTE (Neil)</para></listitem>
8068 <listitem><para>Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)</para></listitem>
8069 <listitem><para>Add variable autocommit (Tom, David Van Wie)</para></listitem>
8074 <title>Object Manipulation</title>
8076 <listitem><para>Make equals signs optional in CREATE DATABASE (Gavin Sherry)</para></listitem>
8077 <listitem><para>Make ALTER TABLE OWNER change index ownership too (Neil)</para></listitem>
8078 <listitem><para>New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls TOAST storage, compression (John Gray)</para></listitem>
8079 <listitem><para>Add schema support, CREATE/DROP SCHEMA (Tom)</para></listitem>
8080 <listitem><para>Create schema for temporary tables (Tom)</para></listitem>
8081 <listitem><para>Add variable search_path for schema search (Tom)</para></listitem>
8082 <listitem><para>Add ALTER TABLE SET/DROP NOT NULL (Christopher)</para></listitem>
8083 <listitem><para>New CREATE FUNCTION volatility levels (Tom)</para></listitem>
8084 <listitem><para>Make rule names unique only per table (Tom)</para></listitem>
8085 <listitem><para>Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)</para></listitem>
8086 <listitem><para>Add ALTER TRIGGER RENAME (Joe)</para></listitem>
8087 <listitem><para>New current_schema() and current_schemas() inquiry functions (Tom)</para></listitem>
8088 <listitem><para>Allow functions to return multiple rows (table functions) (Joe)</para></listitem>
8089 <listitem><para>Make WITH optional in CREATE DATABASE, for consistency (Bruce)</para></listitem>
8090 <listitem><para>Add object dependency tracking (Rod, Tom)</para></listitem>
8091 <listitem><para>Add RESTRICT/CASCADE to DROP commands (Rod)</para></listitem>
8092 <listitem><para>Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)</para></listitem>
8093 <listitem><para>Autodestroy sequence on DROP of table with SERIAL (Rod)</para></listitem>
8094 <listitem><para>Prevent column dropping if column is used by foreign key (Rod)</para></listitem>
8095 <listitem><para>Automatically drop constraints/functions when object is dropped (Rod)</para></listitem>
8096 <listitem><para>Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)</para></listitem>
8097 <listitem><para>Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)</para></listitem>
8098 <listitem><para>Prevent inherited columns from being removed or renamed (Alvaro Herrera)</para></listitem>
8099 <listitem><para>Fix foreign key constraints to not error on intermediate database states (Stephan)</para></listitem>
8100 <listitem><para>Propagate column or table renaming to foreign key constraints</para></listitem>
8101 <listitem><para>Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)</para></listitem>
8102 <listitem><para>Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)</para></listitem>
8103 <listitem><para>Have rules execute alphabetically, returning more predictable values (Tom)</para></listitem>
8104 <listitem><para>Triggers are now fired in alphabetical order (Tom)</para></listitem>
8105 <listitem><para>Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)</para></listitem>
8106 <listitem><para>Allow better casting when inserting/updating values (Tom)</para></listitem>
8111 <title>Utility Commands</title>
8113 <listitem><para>Have COPY TO output embedded carriage returns and newlines as \r and \n (Tom)</para></listitem>
8114 <listitem><para>Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)</para></listitem>
8115 <listitem><para>Make <application>pg_dump</> use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)</para></listitem>
8116 <listitem><para>Disable brackets in multistatement rules (Bruce)</para></listitem>
8117 <listitem><para>Disable VACUUM from being called inside a function (Bruce)</para></listitem>
8118 <listitem><para>Allow dropdb and other scripts to use identifiers with spaces (Bruce)</para></listitem>
8119 <listitem><para>Restrict database comment changes to the current database</para></listitem>
8120 <listitem><para>Allow comments on operators, independent of the underlying function (Rod)</para></listitem>
8121 <listitem><para>Rollback SET commands in aborted transactions (Tom)</para></listitem>
8122 <listitem><para>EXPLAIN now outputs as a query (Tom)</para></listitem>
8123 <listitem><para>Display condition expressions and sort keys in EXPLAIN (Tom)</para></listitem>
8124 <listitem><para>Add 'SET LOCAL var = value' to set configuration variables for a single transaction (Tom)</para></listitem>
8125 <listitem><para>Allow ANALYZE to run in a transaction (Bruce)</para></listitem>
8126 <listitem><para>Improve COPY syntax using new WITH clauses, keep backward compatibility (Bruce)</para></listitem>
8127 <listitem><para>Fix <application>pg_dump</> to consistently output tags in non-ASCII dumps (Bruce)</para></listitem>
8128 <listitem><para>Make foreign key constraints clearer in dump file (Rod)</para></listitem>
8129 <listitem><para>Add COMMENT ON CONSTRAINT (Rod)</para></listitem>
8130 <listitem><para>Allow COPY TO/FROM to specify column names (Brent Verner)</para></listitem>
8131 <listitem><para>Dump UNIQUE and PRIMARY KEY constraints as ALTER TABLE (Rod)</para></listitem>
8132 <listitem><para>Have SHOW output a query result (Joe)</para></listitem>
8133 <listitem><para>Generate failure on short COPY lines rather than pad NULLs (Neil)</para></listitem>
8134 <listitem><para>Fix CLUSTER to preserve all table attributes (Alvaro Herrera)</para></listitem>
8135 <listitem><para>New pg_settings table to view/modify GUC settings (Joe)</para></listitem>
8136 <listitem><para>Add smart quoting, portability improvements to <application>pg_dump</> output (Peter)</para></listitem>
8137 <listitem><para>Dump serial columns out as SERIAL (Tom)</para></listitem>
8138 <listitem><para>Enable large file support, >2G for <application>pg_dump</> (Peter, Philip Warner, Bruce)</para></listitem>
8139 <listitem><para>Disallow TRUNCATE on tables that are involved in referential constraints (Rod)</para></listitem>
8140 <listitem><para>Have TRUNCATE also auto-truncate the toast table of the relation (Tom)</para></listitem>
8141 <listitem><para>Add clusterdb utility that will auto-cluster an entire database based on previous CLUSTER operations (Alvaro Herrera)</para></listitem>
8142 <listitem><para>Overhaul pg_dumpall (Peter)</para></listitem>
8143 <listitem><para>Allow REINDEX of TOAST tables (Tom)</para></listitem>
8144 <listitem><para>Implemented START TRANSACTION, per SQL99 (Neil)</para></listitem>
8145 <listitem><para>Fix rare index corruption when a page split affects bulk delete (Tom)</para></listitem>
8146 <listitem><para>Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)</para></listitem>
8151 <title>Data Types and Functions</title>
8153 <listitem><para>Fix factorial(0) to return 1 (Bruce)</para></listitem>
8154 <listitem><para>Date/time/timezone improvements (Thomas)</para></listitem>
8155 <listitem><para>Fix for array slice extraction (Tom)</para></listitem>
8156 <listitem><para>Fix extract/date_part to report proper microseconds for timestamp (Tatsuo)</para></listitem>
8157 <listitem><para>Allow text_substr() and bytea_substr() to read TOAST values more efficiently (John Gray)</para></listitem>
8158 <listitem><para>Add domain support (Rod)</para></listitem>
8159 <listitem><para>Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types (Thomas)</para></listitem>
8160 <listitem><para>Allow alternate storage scheme of 64-bit integers for date/time types using --enable-integer-datetimes in configure (Thomas)</para></listitem>
8161 <listitem><para>Make timezone(timestamptz) return timestamp rather than a string (Thomas)</para></listitem>
8162 <listitem><para>Allow fractional seconds in date/time types for dates prior to 1BC (Thomas)</para></listitem>
8163 <listitem><para>Limit timestamp data types to 6 decimal places of precision (Thomas)</para></listitem>
8164 <listitem><para>Change timezone conversion functions from timetz() to timezone() (Thomas)</para></listitem>
8165 <listitem><para>Add configuration variables datestyle and timezone (Tom)</para></listitem>
8166 <listitem><para>Add OVERLAY(), which allows substitution of a substring in a string (Thomas)</para></listitem>
8167 <listitem><para>Add SIMILAR TO (Thomas, Tom)</para></listitem>
8168 <listitem><para>Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)</para></listitem>
8169 <listitem><para>Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)</para></listitem>
8170 <listitem><para>Add named composite types using CREATE TYPE typename AS (column) (Joe)</para></listitem>
8171 <listitem><para>Allow composite type definition in the table alias clause (Joe)</para></listitem>
8172 <listitem><para>Add new API to simplify creation of C language table functions (Joe)</para></listitem>
8173 <listitem><para>Remove ODBC-compatible empty parentheses from calls to SQL99 functions for which these parentheses do not match the standard (Thomas)</para></listitem>
8174 <listitem><para>Allow macaddr data type to accept 12 hex digits with no separators (Mike Wyer)</para></listitem>
8175 <listitem><para>Add CREATE/DROP CAST (Peter)</para></listitem>
8176 <listitem><para>Add IS DISTINCT FROM operator (Thomas)</para></listitem>
8177 <listitem><para>Add SQL99 TREAT() function, synonym for CAST() (Thomas)</para></listitem>
8178 <listitem><para>Add pg_backend_pid() to output backend pid (Bruce)</para></listitem>
8179 <listitem><para>Add IS OF / IS NOT OF type predicate (Thomas)</para></listitem>
8180 <listitem><para>Allow bit string constants without fully-specified length (Thomas)</para></listitem>
8181 <listitem><para>Allow conversion between 8-byte integers and bit strings (Thomas)</para></listitem>
8182 <listitem><para>Implement hex literal conversion to bit string literal (Thomas)</para></listitem>
8183 <listitem><para>Allow table functions to appear in the FROM clause (Joe)</para></listitem>
8184 <listitem><para>Increase maximum number of function parameters to 32 (Bruce)</para></listitem>
8185 <listitem><para>No longer automatically create index for SERIAL column (Tom)</para></listitem>
8186 <listitem><para>Add current_database() (Rod)</para></listitem>
8187 <listitem><para>Fix cash_words() to not overflow buffer (Tom)</para></listitem>
8188 <listitem><para>Add functions replace(), split_part(), to_hex() (Joe)</para></listitem>
8189 <listitem><para>Fix LIKE for bytea as a right-hand argument (Joe)</para></listitem>
8190 <listitem><para>Prevent crashes caused by SELECT cash_out(2) (Tom)</para></listitem>
8191 <listitem><para>Fix to_char(1,'FM999.99') to return a period (Karel)</para></listitem>
8192 <listitem><para>Fix trigger/type/language functions returning OPAQUE to return proper type (Tom)</para></listitem>
8197 <title>Internationalization</title>
8199 <listitem><para>Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC) (Eiji Tokuya)</para></listitem>
8200 <listitem><para>Enable locale support by default (Peter)</para></listitem>
8201 <listitem><para>Add locale variables (Peter)</para></listitem>
8202 <listitem><para>Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea (Tatsuo)</para></listitem>
8203 <listitem><para>Add locale awareness to regular expression character classes</para></listitem>
8204 <listitem><para>Enable multibyte support by default (Tatsuo)</para></listitem>
8205 <listitem><para>Add GB18030 multibyte support (Bill Huang)</para></listitem>
8206 <listitem><para>Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo, Kaori)</para></listitem>
8207 <listitem><para>Add pg_conversion table (Tatsuo)</para></listitem>
8208 <listitem><para>Add SQL99 CONVERT() function (Tatsuo)</para></listitem>
8209 <listitem><para>pg_dumpall, pg_controldata, and pg_resetxlog now national-language aware (Peter)</para></listitem>
8210 <listitem><para>New and updated translations</para></listitem>
8215 <title>Server-side Languages</title>
8217 <listitem><para>Allow recursive SQL function (Peter)</para></listitem>
8218 <listitem><para>Change PL/Tcl build to use configured compiler and Makefile.shlib (Peter)</para></listitem>
8219 <listitem><para>Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible (Neil, Tom)</para></listitem>
8220 <listitem><para>Allow PL/pgSQL to handle quoted identifiers (Tom)</para></listitem>
8221 <listitem><para>Allow set-returning PL/pgSQL functions (Neil)</para></listitem>
8222 <listitem><para>Make PL/pgSQL schema-aware (Joe)</para></listitem>
8223 <listitem><para>Remove some memory leaks (Nigel J. Andrews, Tom)</para></listitem>
8230 <listitem><para>Don't lowercase psql \connect database name for 7.2.0 compatibility (Tom)</para></listitem>
8231 <listitem><para>Add psql \timing to time user queries (Greg Sabino Mullane)</para></listitem>
8232 <listitem><para>Have psql \d show index information (Greg Sabino Mullane)</para></listitem>
8233 <listitem><para>New psql \dD shows domains (Jonathan Eisler)</para></listitem>
8234 <listitem><para>Allow psql to show rules on views (Paul ?)</para></listitem>
8235 <listitem><para>Fix for psql variable substitution (Tom)</para></listitem>
8236 <listitem><para>Allow psql \d to show temporary table structure (Tom)</para></listitem>
8237 <listitem><para>Allow psql \d to show foreign keys (Rod)</para></listitem>
8238 <listitem><para>Fix \? to honor \pset pager (Bruce)</para></listitem>
8239 <listitem><para>Have psql reports its version number on startup (Tom)</para></listitem>
8240 <listitem><para>Allow \copy to specify column names (Tom)</para></listitem>
8245 <title>libpq</title>
8247 <listitem><para>Add ~/.pgpass to store host/user password combinations (Alvaro Herrera)</para></listitem>
8248 <listitem><para>Add PQunescapeBytea() function to libpq (Patrick Welche)</para></listitem>
8249 <listitem><para>Fix for sending large queries over non-blocking connections (Bernhard Herzog)</para></listitem>
8250 <listitem><para>Fix for libpq using timers on Win9X (David Ford)</para></listitem>
8251 <listitem><para>Allow libpq notify to handle servers with different-length identifiers (Tom)</para></listitem>
8252 <listitem><para>Add libpq PQescapeString() and PQescapeBytea() to Windows (Bruce)</para></listitem>
8253 <listitem><para>Fix for SSL with non-blocking connections (Jack Bates)</para></listitem>
8254 <listitem><para>Add libpq connection timeout parameter (Denis A Ustimenko)</para></listitem>
8261 <listitem><para>Allow JDBC to compile with JDK 1.4 (Dave)</para></listitem>
8262 <listitem><para>Add JDBC 3 support (Barry)</para></listitem>
8263 <listitem><para>Allows JDBC to set loglevel by adding ?loglevel=X to the connection URL (Barry)</para></listitem>
8264 <listitem><para>Add Driver.info() message that prints out the version number (Barry)</para></listitem>
8265 <listitem><para>Add updateable result sets (Raghu Nidagal, Dave)</para></listitem>
8266 <listitem><para>Add support for callable statements (Paul Bethe)</para></listitem>
8267 <listitem><para>Add query cancel capability</para></listitem>
8268 <listitem><para>Add refresh row (Dave)</para></listitem>
8269 <listitem><para>Fix MD5 encryption handling for multibyte servers (Jun Kawai)</para></listitem>
8270 <listitem><para>Add support for prepared statements (Barry)</para></listitem>
8275 <title>Miscellaneous Interfaces</title>
8277 <listitem><para>Fixed ECPG bug concerning octal numbers in single quotes (Michael)</para></listitem>
8278 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8279 <listitem><para>Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)</para></listitem>
8280 <listitem><para>Add libpgtcl connection close event (Gerhard Hintermayer)</para></listitem>
8281 <listitem><para>Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8282 <listitem><para>Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)</para></listitem>
8283 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8284 <listitem><para>Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8285 <listitem><para>Remove src/bin/pgaccess from main tree, now at http://www.pgaccess.org (Bruce)</para></listitem>
8286 <listitem><para>Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer, Tom)</para></listitem>
8291 <title>Source Code</title>
8293 <listitem><para>Fix for parallel make (Peter)</para></listitem>
8294 <listitem><para>AIX fixes for linking Tcl (Andreas Zeugswetter)</para></listitem>
8295 <listitem><para>Allow PL/Perl to build under Cygwin (Jason Tishler)</para></listitem>
8296 <listitem><para>Improve MIPS compiles (Peter, Oliver Elphick)</para></listitem>
8297 <listitem><para>Require Autoconf version 2.53 (Peter)</para></listitem>
8298 <listitem><para>Require readline and zlib by default in configure (Peter)</para></listitem>
8299 <listitem><para>Allow Solaris to use Intimate Shared Memory (ISM), for performance (Scott Brunza, P.J. Josh Rovero)</para></listitem>
8300 <listitem><para>Always enable syslog in compile, remove --enable-syslog option (Tatsuo)</para></listitem>
8301 <listitem><para>Always enable multibyte in compile, remove --enable-multibyte option (Tatsuo)</para></listitem>
8302 <listitem><para>Always enable locale in compile, remove --enable-locale option (Peter)</para></listitem>
8303 <listitem><para>Fix for Win9x DLL creation (Magnus Naeslund)</para></listitem>
8304 <listitem><para>Fix for link() usage by WAL code on Windows, BeOS (Jason Tishler)</para></listitem>
8305 <listitem><para>Add sys/types.h to c.h, remove from main files (Peter, Bruce)</para></listitem>
8306 <listitem><para>Fix AIX hang on SMP machines (Tomoyuki Niijima)</para></listitem>
8307 <listitem><para>AIX SMP hang fix (Tomoyuki Niijima)</para></listitem>
8308 <listitem><para>Fix pre-1970 date handling on newer glibc libraries (Tom)</para></listitem>
8309 <listitem><para>Fix PowerPC SMP locking (Tom)</para></listitem>
8310 <listitem><para>Prevent gcc -ffast-math from being used (Peter, Tom)</para></listitem>
8311 <listitem><para>Bison >= 1.50 now required for developer builds</para></listitem>
8312 <listitem><para>Kerberos 5 support now builds with Heimdal (Peter)</para></listitem>
8313 <listitem><para>Add appendix in the User's Guide which lists SQL features (Thomas)</para></listitem>
8314 <listitem><para>Improve loadable module linking to use RTLD_NOW (Tom)</para></listitem>
8315 <listitem><para>New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)</para></listitem>
8316 <listitem><para>New src/port directory holds replaced libc functions (Peter, Bruce)</para></listitem>
8317 <listitem><para>New pg_namespace system catalog for schemas (Tom)</para></listitem>
8318 <listitem><para>Add pg_class.relnamespace for schemas (Tom)</para></listitem>
8319 <listitem><para>Add pg_type.typnamespace for schemas (Tom)</para></listitem>
8320 <listitem><para>Add pg_proc.pronamespace for schemas (Tom)</para></listitem>
8321 <listitem><para>Restructure aggregates to have pg_proc entries (Tom)</para></listitem>
8322 <listitem><para>System relations now have their own namespace, pg_* test not required (Fernando Nasser)</para></listitem>
8323 <listitem><para>Rename TOAST index names to be *_index rather than *_idx (Neil)</para></listitem>
8324 <listitem><para>Add namespaces for operators, opclasses (Tom)</para></listitem>
8325 <listitem><para>Add additional checks to server control file (Thomas)</para></listitem>
8326 <listitem><para>New Polish FAQ (Marcin Mazurek)</para></listitem>
8327 <listitem><para>Add Posix semaphore support (Tom)</para></listitem>
8328 <listitem><para>Document need for reindex (Bruce)</para></listitem>
8329 <listitem><para>Rename some internal identifiers to simplify Windows compile (Jan, Katherine Ward)</para></listitem>
8330 <listitem><para>Add documentation on computing disk space (Bruce)</para></listitem>
8331 <listitem><para>Remove KSQO from GUC (Bruce)</para></listitem>
8332 <listitem><para>Fix memory leak in rtree (Kenneth Been)</para></listitem>
8333 <listitem><para>Modify a few error messages for consistency (Bruce)</para></listitem>
8334 <listitem><para>Remove unused system table columns (Peter)</para></listitem>
8335 <listitem><para>Make system columns NOT NULL where appropriate (Tom)</para></listitem>
8336 <listitem><para>Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)</para></listitem>
8337 <listitem><para>Remove OPAQUE and create specific subtypes (Tom)</para></listitem>
8338 <listitem><para>Cleanups in array internal handling (Joe, Tom)</para></listitem>
8339 <listitem><para>Disallow pg_atoi('') (Bruce)</para></listitem>
8340 <listitem><para>Remove parameter wal_files because WAL files are now recycled (Bruce)</para></listitem>
8341 <listitem><para>Add version numbers to heap pages (Tom)</para></listitem>
8346 <title>Contrib</title>
8348 <listitem><para>Allow inet arrays in /contrib/array (Neil)</para></listitem>
8349 <listitem><para>GiST fixes (Teodor Sigaev, Neil)</para></listitem>
8350 <listitem><para>Upgrade /contrib/mysql</para></listitem>
8351 <listitem><para>Add /contrib/dbsize which shows table sizes without vacuum (Peter)</para></listitem>
8352 <listitem><para>Add /contrib/intagg, integer aggregator routines (mlw)</para></listitem>
8353 <listitem><para>Improve /contrib/oid2name (Neil, Bruce)</para></listitem>
8354 <listitem><para>Improve /contrib/tsearch (Oleg, Teodor Sigaev)</para></listitem>
8355 <listitem><para>Cleanups of /contrib/rserver (Alexey V. Borzov)</para></listitem>
8356 <listitem><para>Update /contrib/oracle conversion utility (Gilles Darold)</para></listitem>
8357 <listitem><para>Update /contrib/dblink (Joe)</para></listitem>
8358 <listitem><para>Improve options supported by /contrib/vacuumlo (Mario Weilguni)</para></listitem>
8359 <listitem><para>Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey Oktyabrski)</para></listitem>
8360 <listitem><para>Add /contrib/reindexdb utility (Shaun Thomas)</para></listitem>
8361 <listitem><para>Add indexing to /contrib/isbn_issn (Dan Weston)</para></listitem>
8362 <listitem><para>Add /contrib/dbmirror (Steven Singer)</para></listitem>
8363 <listitem><para>Improve /contrib/pgbench (Neil)</para></listitem>
8364 <listitem><para>Add /contrib/tablefunc table function examples (Joe)</para></listitem>
8365 <listitem><para>Add /contrib/ltree data type for tree structures (Teodor Sigaev, Oleg Bartunov)</para></listitem>
8366 <listitem><para>Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)</para></listitem>
8367 <listitem><para>Fixes to /contrib/cube (Bruno Wolff)</para></listitem>
8368 <listitem><para>Improve /contrib/fulltextindex (Christopher)</para></listitem>
8375 <sect1 id="release-7-2-8">
8376 <title>Release 7.2.8</title>
8379 <title>Release date</title>
8380 <simpara>2005-05-09</simpara>
8384 This release contains a variety of fixes from 7.2.7, including one
8385 security-related issue.
8389 <title>Migration to version 7.2.8</title>
8392 A dump/restore is not required for those running 7.2.X.
8397 <title>Changes</title>
8400 <listitem><para>Repair ancient race condition that allowed a transaction to be
8401 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
8402 than for other purposes</para>
8403 <para>This is an extremely serious bug since it could lead to apparent
8404 data inconsistencies being briefly visible to applications.</para></listitem>
8405 <listitem><para>Repair race condition between relation extension and
8407 <para>This could theoretically have caused loss of a page's worth of
8408 freshly-inserted data, although the scenario seems of very low probability.
8409 There are no known cases of it having caused more than an Assert failure.
8411 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
8412 <type>TIME WITH TIME ZONE</> values</para></listitem>
8413 <listitem><para>Additional buffer overrun checks in plpgsql
8414 (Neil)</para></listitem>
8415 <listitem><para>Fix pg_dump to dump index names and trigger names containing
8416 <literal>%</> correctly (Neil)</para></listitem>
8417 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
8418 month-related formats</para></listitem>
8419 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
8420 (Marko Kreen)</para></listitem>
8426 <sect1 id="release-7-2-7">
8427 <title>Release 7.2.7</title>
8430 <title>Release date</title>
8431 <simpara>2005-01-31</simpara>
8435 This release contains a variety of fixes from 7.2.6, including several
8436 security-related issues.
8440 <title>Migration to version 7.2.7</title>
8443 A dump/restore is not required for those running 7.2.X.
8448 <title>Changes</title>
8451 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
8453 On platforms that will automatically execute initialization functions of a
8454 shared library (this includes at least Windows and ELF-based Unixen),
8455 <command>LOAD</> can be used to make the server execute arbitrary code.
8456 Thanks to NGS Software for reporting this.</para></listitem>
8457 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
8458 Jurka)</para></listitem>
8459 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
8460 many parameters (Neil)</para></listitem>
8461 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
8463 The result of the join was mistakenly supposed to be sorted the same as the
8464 left input. This could not only deliver mis-sorted output to the user, but
8465 in case of nested merge joins could give outright wrong answers.
8467 <listitem><para>Fix display of negative intervals in SQL and GERMAN
8468 datestyles</para></listitem>
8474 <sect1 id="release-7-2-6">
8475 <title>Release 7.2.6</title>
8478 <title>Release date</title>
8479 <simpara>2004-10-22</simpara>
8483 This release contains a variety of fixes from 7.2.5.
8488 <title>Migration to version 7.2.6</title>
8491 A dump/restore is not required for those running 7.2.X.
8496 <title>Changes</title>
8499 <listitem><para>Repair possible failure to update hint bits on disk</para>
8501 Under rare circumstances this oversight could lead to
8502 <quote>could not access transaction status</> failures, which qualifies
8503 it as a potential-data-loss bug.
8505 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
8507 Very large left joins using a hash join plan could fail to output unmatched
8508 left-side rows given just the right data distribution.
8510 <listitem><para>Disallow running pg_ctl as root</para>
8512 This is to guard against any possible security issues.
8514 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
8516 This has been reported as a security issue, though it's hardly worthy of
8517 concern since there is no reason for non-developers to use this script anyway.
8519 <listitem><para>Update to newer versions of Bison</para></listitem>
8525 <sect1 id="release-7-2-5">
8526 <title>Release 7.2.5</title>
8529 <title>Release date</title>
8530 <simpara>2004-08-16</simpara>
8534 This release contains a variety of fixes from 7.2.4.
8539 <title>Migration to version 7.2.5</title>
8542 A dump/restore is not required for those running 7.2.X.
8547 <title>Changes</title>
8550 <listitem><para>Prevent possible loss of committed transactions during crash</para>
8552 Due to insufficient interlocking between transaction commit and checkpointing,
8553 it was possible for transactions committed just before the most recent
8554 checkpoint to be lost, in whole or in part, following a database crash and
8555 restart. This is a serious bug that has existed
8556 since <productname>PostgreSQL</productname> 7.1.
8558 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
8559 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
8560 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
8561 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
8562 <listitem><para>Repair memory leaks in pg_dump</para></listitem>
8563 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
8568 <sect1 id="release-7-2-4">
8569 <title>Release 7.2.4</title>
8572 <title>Release date</title>
8573 <simpara>2003-01-30</simpara>
8577 This release contains a variety of fixes for version 7.2.3,
8578 including fixes to prevent possible data loss.
8582 <title>Migration to version 7.2.4</title>
8585 A dump/restore is <emphasis>not</emphasis> required for those
8586 running version 7.2.*.
8591 <title>Changes</title>
8594 <listitem><para>Fix some additional cases of VACUUM "No one parent tuple was found" error</para></listitem>
8595 <listitem><para>Prevent VACUUM from being called inside a function (Bruce)</para></listitem>
8596 <listitem><para>Ensure pg_clog updates are sync'd to disk before marking checkpoint complete</para></listitem>
8597 <listitem><para>Avoid integer overflow during large hash joins</para></listitem>
8598 <listitem><para>Make GROUP commands work when pg_group.grolist is large enough to be toasted</para></listitem>
8599 <listitem><para>Fix errors in datetime tables; some timezone names weren't being recognized</para></listitem>
8600 <listitem><para>Fix integer overflows in circle_poly(), path_encode(), path_add() (Neil)</para></listitem>
8601 <listitem><para>Repair long-standing logic errors in lseg_eq(), lseg_ne(), lseg_center()</para></listitem>
8607 <sect1 id="release-7-2-3">
8608 <title>Release 7.2.3</title>
8611 <title>Release date</title>
8612 <simpara>2002-10-01</simpara>
8616 This release contains a variety of fixes for version 7.2.2,
8617 including fixes to prevent possible data loss.
8621 <title>Migration to version 7.2.3</title>
8624 A dump/restore is <emphasis>not</emphasis> required for those
8625 running version 7.2.*.
8630 <title>Changes</title>
8633 <listitem><para>Prevent possible compressed transaction log loss (Tom)</para></listitem>
8634 <listitem><para>Prevent non-superuser from increasing most recent vacuum info (Tom)</para></listitem>
8635 <listitem><para>Handle pre-1970 date values in newer versions of glibc (Tom)</para></listitem>
8636 <listitem><para>Fix possible hang during server shutdown</para></listitem>
8637 <listitem><para>Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)</para></listitem>
8638 <listitem><para>Fix <application>pg_dump</> to properly dump FULL JOIN USING (Tom)</para></listitem>
8644 <sect1 id="release-7-2-2">
8645 <title>Release 7.2.2</title>
8648 <title>Release date</title>
8649 <simpara>2002-08-23</simpara>
8653 This release contains a variety of fixes for version 7.2.1.
8657 <title>Migration to version 7.2.2</title>
8660 A dump/restore is <emphasis>not</emphasis> required for those
8661 running version 7.2.*.
8666 <title>Changes</title>
8669 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8670 <listitem><para>Fix for compressed transaction log id wraparound (Tom)</para></listitem>
8671 <listitem><para>Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Tatsuo)</para></listitem>
8672 <listitem><para>Fix for psql and <application>pg_dump</> crashing when invoked with non-existent long options (Tatsuo)</para></listitem>
8673 <listitem><para>Fix crash when invoking geometric operators (Tom)</para></listitem>
8674 <listitem><para>Allow OPEN cursor(args) (Tom)</para></listitem>
8675 <listitem><para>Fix for rtree_gist index build (Teodor)</para></listitem>
8676 <listitem><para>Fix for dumping user-defined aggregates (Tom)</para></listitem>
8677 <listitem><para>contrib/intarray fixes (Oleg)</para></listitem>
8678 <listitem><para>Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)</para></listitem>
8679 <listitem><para>Fix to pg_convert (Tatsuo)</para></listitem>
8680 <listitem><para>Fix for crash with long DATA strings (Thomas, Neil)</para></listitem>
8681 <listitem><para>Fix for repeat(), lpad(), rpad() and long strings (Neil)</para></listitem>
8687 <sect1 id="release-7-2-1">
8688 <title>Release 7.2.1</title>
8691 <title>Release date</title>
8692 <simpara>2002-03-21</simpara>
8696 This release contains a variety of fixes for version 7.2.
8700 <title>Migration to version 7.2.1</title>
8703 A dump/restore is <emphasis>not</emphasis> required for those
8704 running version 7.2.
8709 <title>Changes</title>
8712 <listitem><para>Ensure that sequence counters do not go backwards after a crash (Tom)</para></listitem>
8713 <listitem><para>Fix pgaccess kanji-conversion key binding (Tatsuo)</para></listitem>
8714 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8715 <listitem><para>Cash I/O improvements (Tom)</para></listitem>
8716 <listitem><para>New Russian FAQ</para></listitem>
8717 <listitem><para>Compile fix for missing AuthBlockSig (Heiko)</para></listitem>
8718 <listitem><para>Additional time zones and time zone fixes (Thomas)</para></listitem>
8719 <listitem><para>Allow psql \connect to handle mixed case database and user names (Tom)</para></listitem>
8720 <listitem><para>Return proper OID on command completion even with ON INSERT rules (Tom)</para></listitem>
8721 <listitem><para>Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)</para></listitem>
8722 <listitem><para>Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)</para></listitem>
8723 <listitem><para>Improve handling of multiple UNIONs with different lengths (Tom)</para></listitem>
8724 <listitem><para>contrib/btree_gist improvements (Teodor Sigaev)</para></listitem>
8725 <listitem><para>contrib/tsearch dictionary improvements, see README.tsearch for an additional installation step (Thomas T. Thai, Teodor Sigaev)</para></listitem>
8726 <listitem><para>Fix for array subscripts handling (Tom)</para></listitem>
8727 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8733 <sect1 id="release-7-2">
8734 <title>Release 7.2</title>
8737 <title>Release date</title>
8738 <simpara>2002-02-04</simpara>
8742 <title>Overview</title>
8745 This release improves <productname>PostgreSQL</> for use in
8746 high-volume applications.
8750 Major changes in this release:
8758 Vacuuming no longer locks tables, thus allowing normal user
8759 access during the vacuum. A new <command>VACUUM FULL</>
8760 command does old-style vacuum by locking the table and
8761 shrinking the on-disk copy of the table.
8767 <term>Transactions</term>
8770 There is no longer a problem with installations that exceed
8771 four billion transactions.
8780 OIDs are now optional. Users can now create tables without
8781 OIDs for cases where OID usage is excessive.
8787 <term>Optimizer</term>
8790 The system now computes histogram column statistics during
8791 <command>ANALYZE</>, allowing much better optimizer choices.
8797 <term>Security</term>
8800 A new MD5 encryption option allows more secure storage and
8801 transfer of passwords. A new Unix-domain socket
8802 authentication option is available on Linux and BSD systems.
8808 <term>Statistics</term>
8811 Administrators can use the new table access statistics module
8812 to get fine-grained information about table and index usage.
8818 <term>Internationalization</term>
8821 Program and library messages can now be displayed in several
8831 <title>Migration to version 7.2</title>
8834 A dump/restore using <command>pg_dump</command> is required for
8835 those wishing to migrate data from any previous release.
8839 Observe the following incompatibilities:
8845 The semantics of the <command>VACUUM</command> command have
8846 changed in this release. You may wish to update your
8847 maintenance procedures accordingly.
8853 In this release, comparisons using <literal>= NULL</literal>
8854 will always return false (or NULL, more precisely). Previous
8855 releases automatically transformed this syntax to <literal>IS
8856 NULL</literal>. The old behavior can be re-enabled using a
8857 <filename>postgresql.conf</filename> parameter.
8863 The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
8864 configuration is now only reloaded after receiving a
8865 <systemitem>SIGHUP</> signal, not with each connection.
8871 The function <filename>octet_length()</> now returns the uncompressed data length.
8877 The date/time value <literal>'current'</literal> is no longer
8878 available. You will need to rewrite your applications.
8884 The <literal>timestamp()</literal>, <literal>time()</literal>,
8885 and <literal>interval()</literal> functions are no longer
8886 available. Instead of <literal>timestamp()</literal>, use
8887 <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
8894 The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
8895 in the next release. You should change your queries to use
8896 separate LIMIT and OFFSET clauses, e.g. <literal>LIMIT 10 OFFSET
8902 <title>Changes</title>
8905 <title>Server Operation</title>
8907 <listitem><para>Create temporary files in a separate directory (Bruce)</para></listitem>
8908 <listitem><para>Delete orphaned temporary files on postmaster startup (Bruce)</para></listitem>
8909 <listitem><para>Added unique indexes to some system tables (Tom)</para></listitem>
8910 <listitem><para>System table operator reorganization (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
8911 <listitem><para>Renamed pg_log to pg_clog (Tom)</para></listitem>
8912 <listitem><para>Enable SIGTERM, SIGQUIT to kill backends (Jan)</para></listitem>
8913 <listitem><para>Removed compile-time limit on number of backends (Tom)</para></listitem>
8914 <listitem><para>Better cleanup for semaphore resource failure (Tatsuo, Tom)</para></listitem>
8915 <listitem><para>Allow safe transaction ID wraparound (Tom)</para></listitem>
8916 <listitem><para>Removed OIDs from some system tables (Tom)</para></listitem>
8917 <listitem><para>Removed "triggered data change violation" error check (Tom)</para></listitem>
8918 <listitem><para>SPI portal creation of prepared/saved plans (Jan)</para></listitem>
8919 <listitem><para>Allow SPI column functions to work for system columns (Tom)</para></listitem>
8920 <listitem><para>Long value compression improvement (Tom)</para></listitem>
8921 <listitem><para>Statistics collector for table, index access (Jan)</para></listitem>
8922 <listitem><para>Truncate extra-long sequence names to a reasonable value (Tom)</para></listitem>
8923 <listitem><para>Measure transaction times in milliseconds (Thomas)</para></listitem>
8924 <listitem><para>Fix TID sequential scans (Hiroshi)</para></listitem>
8925 <listitem><para>Superuser ID now fixed at 1 (Peter E)</para></listitem>
8926 <listitem><para>New pg_ctl "reload" option (Tom)</para></listitem>
8931 <title>Performance</title>
8933 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8934 <listitem><para>New histogram column statistics for optimizer (Tom)</para></listitem>
8935 <listitem><para>Reuse write-ahead log files rather than discarding them (Tom)</para></listitem>
8936 <listitem><para>Cache improvements (Tom)</para></listitem>
8937 <listitem><para>IS NULL, IS NOT NULL optimizer improvement (Tom)</para></listitem>
8938 <listitem><para>Improve lock manager to reduce lock contention (Tom)</para></listitem>
8939 <listitem><para>Keep relcache entries for index access support functions (Tom)</para></listitem>
8940 <listitem><para>Allow better selectivity with NaN and infinities in NUMERIC (Tom)</para></listitem>
8941 <listitem><para>R-tree performance improvements (Kenneth Been)</para></listitem>
8942 <listitem><para>B-tree splits more efficient (Tom)</para></listitem>
8947 <title>Privileges</title>
8949 <listitem><para>Change UPDATE, DELETE privileges to be distinct (Peter E)</para></listitem>
8950 <listitem><para>New REFERENCES, TRIGGER privileges (Peter E)</para></listitem>
8951 <listitem><para>Allow GRANT/REVOKE to/from more than one user at a time (Peter E)</para></listitem>
8952 <listitem><para>New has_table_privilege() function (Joe Conway)</para></listitem>
8953 <listitem><para>Allow non-superuser to vacuum database (Tom)</para></listitem>
8954 <listitem><para>New SET SESSION AUTHORIZATION command (Peter E)</para></listitem>
8955 <listitem><para>Fix bug in privilege modifications on newly created tables (Tom)</para></listitem>
8956 <listitem><para>Disallow access to pg_statistic for non-superuser, add user-accessible views (Tom)</para></listitem>
8961 <title>Client Authentication</title>
8963 <listitem><para>Fork postmaster before doing authentication to prevent hangs (Peter E)</para></listitem>
8964 <listitem><para>Add ident authentication over Unix domain sockets on Linux, *BSD (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)</para></listitem>
8965 <listitem><para>Add a password authentication method that uses MD5 encryption (Bruce)</para></listitem>
8966 <listitem><para>Allow encryption of stored passwords using MD5 (Bruce)</para></listitem>
8967 <listitem><para>PAM authentication (Dominic J. Eidson)</para></listitem>
8968 <listitem><para>Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)</para></listitem>
8973 <title>Server Configuration</title>
8975 <listitem><para>Interpretation of some time zone abbreviations as Australian rather than North American now settable at run time (Bruce)</para></listitem>
8976 <listitem><para>New parameter to set default transaction isolation level (Peter E)</para></listitem>
8977 <listitem><para>New parameter to enable conversion of "expr = NULL" into "expr IS NULL", off by default (Peter E)</para></listitem>
8978 <listitem><para>New parameter to control memory usage by VACUUM (Tom)</para></listitem>
8979 <listitem><para>New parameter to set client authentication timeout (Tom)</para></listitem>
8980 <listitem><para>New parameter to set maximum number of open files (Tom)</para></listitem>
8985 <title>Queries</title>
8987 <listitem><para>Statements added by INSERT rules now execute after the INSERT (Jan)</para></listitem>
8988 <listitem><para>Prevent unadorned relation names in target list (Bruce)</para></listitem>
8989 <listitem><para>NULLs now sort after all normal values in ORDER BY (Tom)</para></listitem>
8990 <listitem><para>New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)</para></listitem>
8991 <listitem><para>New SHARE UPDATE EXCLUSIVE lock mode (Tom)</para></listitem>
8992 <listitem><para>New EXPLAIN ANALYZE command that shows run times and row counts (Martijn van Oosterhout)</para></listitem>
8993 <listitem><para>Fix problem with LIMIT and subqueries (Tom)</para></listitem>
8994 <listitem><para>Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)</para></listitem>
8995 <listitem><para>Fix nested EXCEPT/INTERSECT (Tom)</para></listitem>
9000 <title>Schema Manipulation</title>
9002 <listitem><para>Fix SERIAL in temporary tables (Bruce)</para></listitem>
9003 <listitem><para>Allow temporary sequences (Bruce)</para></listitem>
9004 <listitem><para>Sequences now use int8 internally (Tom)</para></listitem>
9005 <listitem><para>New SERIAL8 creates int8 columns with sequences, default still SERIAL4 (Tom)</para></listitem>
9006 <listitem><para>Make OIDs optional using WITHOUT OIDS (Tom)</para></listitem>
9007 <listitem><para>Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)</para></listitem>
9008 <listitem><para>Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher Kings-Lynne)</para></listitem>
9009 <listitem><para>New CREATE OR REPLACE FUNCTION to alter existing function (preserving the function OID) (Gavin Sherry)</para></listitem>
9010 <listitem><para>Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)</para></listitem>
9011 <listitem><para>Allow column renaming in views</para></listitem>
9012 <listitem><para>Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent Verner)</para></listitem>
9013 <listitem><para>Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (Stephan Szabo)</para></listitem>
9014 <listitem><para>ALTER TABLE RENAME update foreign-key trigger arguments correctly (Brent Verner)</para></listitem>
9015 <listitem><para>DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)</para></listitem>
9016 <listitem><para>Add automatic return type data casting for SQL functions (Tom)</para></listitem>
9017 <listitem><para>Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
9018 <listitem><para>Enable partial indexes (Martijn van Oosterhout)</para></listitem>
9023 <title>Utility Commands</title>
9025 <listitem><para>Add RESET ALL, SHOW ALL (Marko Kreen)</para></listitem>
9026 <listitem><para>CREATE/ALTER USER/GROUP now allow options in any order (Vince)</para></listitem>
9027 <listitem><para>Add LOCK A, B, C functionality (Neil Padgett)</para></listitem>
9028 <listitem><para>New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)</para></listitem>
9029 <listitem><para>New light-weight VACUUM does not lock table; old semantics are available as VACUUM FULL (Tom)</para></listitem>
9030 <listitem><para>Disable COPY TO/FROM on views (Bruce)</para></listitem>
9031 <listitem><para>COPY DELIMITERS string must be exactly one character (Tom)</para></listitem>
9032 <listitem><para>VACUUM warning about index tuples fewer than heap now only appears when appropriate (Martijn van Oosterhout)</para></listitem>
9033 <listitem><para>Fix privilege checks for CREATE INDEX (Tom)</para></listitem>
9034 <listitem><para>Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)</para></listitem>
9039 <title>Data Types and Functions</title>
9041 <listitem><para>SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)</para></listitem>
9042 <listitem><para>Add convert(), convert2() (Tatsuo)</para></listitem>
9043 <listitem><para>New function bit_length() (Peter E)</para></listitem>
9044 <listitem><para>Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tatsuo)</para></listitem>
9045 <listitem><para>CHAR(), VARCHAR() now reject strings that are too long (Peter E)</para></listitem>
9046 <listitem><para>BIT VARYING now rejects bit strings that are too long (Peter E)</para></listitem>
9047 <listitem><para>BIT now rejects bit strings that do not match declared size (Peter E)</para></listitem>
9048 <listitem><para>INET, CIDR text conversion functions (Alex Pilosov)</para></listitem>
9049 <listitem><para>INET, CIDR operators << and <<= indexable (Alex Pilosov)</para></listitem>
9050 <listitem><para>Bytea \### now requires valid three digit octal number</para></listitem>
9051 <listitem><para>Bytea comparison improvements, now supports =, <>, >, >=, <, and <=</para></listitem>
9052 <listitem><para>Bytea now supports B-tree indexes</para></listitem>
9053 <listitem><para>Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE</para></listitem>
9054 <listitem><para>Bytea now supports concatenation</para></listitem>
9055 <listitem><para>New bytea functions: position, substring, trim, btrim, and length</para></listitem>
9056 <listitem><para>New encode() function mode, "escaped", converts minimally escaped bytea to/from text</para></listitem>
9057 <listitem><para>Add pg_database_encoding_max_length() (Tatsuo)</para></listitem>
9058 <listitem><para>Add pg_client_encoding() function (Tatsuo)</para></listitem>
9059 <listitem><para>now() returns time with millisecond precision (Thomas)</para></listitem>
9060 <listitem><para>New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)</para></listitem>
9061 <listitem><para>Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)</para></listitem>
9062 <listitem><para>New xid/int comparison functions (Hiroshi)</para></listitem>
9063 <listitem><para>Add precision to TIME, TIMESTAMP, and INTERVAL data types (Thomas)</para></listitem>
9064 <listitem><para>Modify type coercion logic to attempt binary-compatible functions first (Tom)</para></listitem>
9065 <listitem><para>New encode() function installed by default (Marko Kreen)</para></listitem>
9066 <listitem><para>Improved to_*() conversion functions (Karel Zak)</para></listitem>
9067 <listitem><para>Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)</para></listitem>
9068 <listitem><para>New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_salt() (Marko Kreen)</para></listitem>
9069 <listitem><para>Correct description of translate() function (Bruce)</para></listitem>
9070 <listitem><para>Add INTERVAL argument for SET TIME ZONE (Thomas)</para></listitem>
9071 <listitem><para>Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)</para></listitem>
9072 <listitem><para>Optimize length functions when using single-byte encodings (Tatsuo)</para></listitem>
9073 <listitem><para>Fix path_inter, path_distance, path_length, dist_ppath to handle closed paths (Curtis Barrett, Tom)</para></listitem>
9074 <listitem><para>octet_length(text) now returns non-compressed length (Tatsuo, Bruce)</para></listitem>
9075 <listitem><para>Handle "July" full name in date/time literals (Greg Sabino Mullane)</para></listitem>
9076 <listitem><para>Some datatype() function calls now evaluated differently</para></listitem>
9077 <listitem><para>Add support for Julian and ISO time specifications (Thomas)</para></listitem>
9082 <title>Internationalization</title>
9084 <listitem><para>National language support in psql, <application>pg_dump</>, libpq, and server (Peter E)</para></listitem>
9085 <listitem><para>Message translations in Chinese (simplified, traditional), Czech, French, German, Hungarian, Russian, Swedish (Peter E, Serguei A. Mokhov, Karel Zak, Weiping He, Zhenbang Wei, Kovacs Zoltan)</para></listitem>
9086 <listitem><para>Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware (Tatsuo)</para></listitem>
9087 <listitem><para>Add LATIN5,6,7,8,9,10 support (Tatsuo)</para></listitem>
9088 <listitem><para>Add ISO 8859-5,6,7,8 support (Tatsuo)</para></listitem>
9089 <listitem><para>Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)</para></listitem>
9090 <listitem><para>Make mic2ascii() non-ASCII aware (Tatsuo)</para></listitem>
9091 <listitem><para>Reject invalid multibyte character sequences (Tatsuo)</para></listitem>
9096 <title><application>PL/pgSQL</></title>
9098 <listitem><para>Now uses portals for SELECT loops, allowing huge result sets (Jan)</para></listitem>
9099 <listitem><para>CURSOR and REFCURSOR support (Jan)</para></listitem>
9100 <listitem><para>Can now return open cursors (Jan)</para></listitem>
9101 <listitem><para>Add ELSEIF (Klaus Reger)</para></listitem>
9102 <listitem><para>Improve PL/pgSQL error reporting, including location of error (Tom)</para></listitem>
9103 <listitem><para>Allow IS or FOR key words in cursor declaration, for compatibility (Bruce)</para></listitem>
9104 <listitem><para>Fix for SELECT ... FOR UPDATE (Tom)</para></listitem>
9105 <listitem><para>Fix for PERFORM returning multiple rows (Tom)</para></listitem>
9106 <listitem><para>Make PL/pgSQL use the server's type coercion code (Tom)</para></listitem>
9107 <listitem><para>Memory leak fix (Jan, Tom)</para></listitem>
9108 <listitem><para>Make trailing semicolon optional (Tom)</para></listitem>
9113 <title>PL/Perl</title>
9115 <listitem><para>New untrusted PL/Perl (Alex Pilosov)</para></listitem>
9116 <listitem><para>PL/Perl is now built on some platforms even if libperl is not shared (Peter E)</para></listitem>
9121 <title>PL/Tcl</title>
9123 <listitem><para>Now reports errorInfo (Vsevolod Lobko)</para></listitem>
9124 <listitem><para>Add spi_lastoid function (bob@redivi.com)</para></listitem>
9129 <title>PL/Python</title>
9131 <listitem><para>...is new (Andrew Bosma)</para></listitem>
9136 <title><application>psql</></title>
9138 <listitem><para>\d displays indexes in unique, primary groupings (Christopher Kings-Lynne)</para></listitem>
9139 <listitem><para>Allow trailing semicolons in backslash commands (Greg Sabino Mullane)</para></listitem>
9140 <listitem><para>Read password from /dev/tty if possible</para></listitem>
9141 <listitem><para>Force new password prompt when changing user and database (Tatsuo, Tom)</para></listitem>
9142 <listitem><para>Format the correct number of columns for Unicode (Patrice)</para></listitem>
9147 <title><application>libpq</></title>
9149 <listitem><para>New function PQescapeString() to escape quotes in command strings (Florian Weimer)</para></listitem>
9150 <listitem><para>New function PQescapeBytea() escapes binary strings for use as SQL string literals</para></listitem>
9157 <listitem><para>Return OID of INSERT (Ken K)</para></listitem>
9158 <listitem><para>Handle more data types (Ken K)</para></listitem>
9159 <listitem><para>Handle single quotes and newlines in strings (Ken K)</para></listitem>
9160 <listitem><para>Handle NULL variables (Ken K)</para></listitem>
9161 <listitem><para>Fix for time zone handling (Barry Lind)</para></listitem>
9162 <listitem><para>Improved Druid support</para></listitem>
9163 <listitem><para>Allow eight-bit characters with non-multibyte server (Barry Lind)</para></listitem>
9164 <listitem><para>Support BIT, BINARY types (Ned Wolpert)</para></listitem>
9165 <listitem><para>Reduce memory usage (Michael Stephens, Dave Cramer)</para></listitem>
9166 <listitem><para>Update DatabaseMetaData (Peter E)</para></listitem>
9167 <listitem><para>Add DatabaseMetaData.getCatalogs() (Peter E)</para></listitem>
9168 <listitem><para>Encoding fixes (Anders Bengtsson)</para></listitem>
9169 <listitem><para>Get/setCatalog methods (Jason Davies)</para></listitem>
9170 <listitem><para>DatabaseMetaData.getColumns() now returns column defaults (Jason Davies)</para></listitem>
9171 <listitem><para>DatabaseMetaData.getColumns() performance improvement (Jeroen van Vianen)</para></listitem>
9172 <listitem><para>Some JDBC1 and JDBC2 merging (Anders Bengtsson)</para></listitem>
9173 <listitem><para>Transaction performance improvements (Barry Lind)</para></listitem>
9174 <listitem><para>Array fixes (Greg Zoller)</para></listitem>
9175 <listitem><para>Serialize addition </para></listitem>
9176 <listitem><para>Fix batch processing (Rene Pijlman)</para></listitem>
9177 <listitem><para>ExecSQL method reorganization (Anders Bengtsson)</para></listitem>
9178 <listitem><para>GetColumn() fixes (Jeroen van Vianen)</para></listitem>
9179 <listitem><para>Fix isWriteable() function (Rene Pijlman)</para></listitem>
9180 <listitem><para>Improved passage of JDBC2 conformance tests (Rene Pijlman)</para></listitem>
9181 <listitem><para>Add bytea type capability (Barry Lind)</para></listitem>
9182 <listitem><para>Add isNullable() (Rene Pijlman)</para></listitem>
9183 <listitem><para>JDBC date/time test suite fixes (Liam Stewart)</para></listitem>
9184 <listitem><para>Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)</para></listitem>
9185 <listitem><para>Fix DatabaseMetaData to show precision properly (Mark Lillywhite)</para></listitem>
9186 <listitem><para>New getImported/getExported keys (Jason Davies)</para></listitem>
9187 <listitem><para>MD5 password encryption support (Jeremy Wohl)</para></listitem>
9188 <listitem><para>Fix to actually use type cache (Ned Wolpert)</para></listitem>
9195 <listitem><para>Remove query size limit (Hiroshi)</para></listitem>
9196 <listitem><para>Remove text field size limit (Hiroshi)</para></listitem>
9197 <listitem><para>Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)</para></listitem>
9198 <listitem><para>Allow ODBC procedure calls (Hiroshi)</para></listitem>
9199 <listitem><para>Improve boolean handing (Aidan Mountford)</para></listitem>
9200 <listitem><para>Most configuration options now settable via DSN (Hiroshi)</para></listitem>
9201 <listitem><para>Multibyte, performance fixes (Hiroshi)</para></listitem>
9202 <listitem><para>Allow driver to be used with iODBC or unixODBC (Peter E)</para></listitem>
9203 <listitem><para>MD5 password encryption support (Bruce)</para></listitem>
9204 <listitem><para>Add more compatibility functions to odbc.sql (Peter E)</para></listitem>
9209 <title><application>ECPG</></title>
9211 <listitem><para>EXECUTE ... INTO implemented (Christof Petig)</para></listitem>
9212 <listitem><para>Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)</para></listitem>
9213 <listitem><para>Fix for GRANT parameters (Lee Kindness)</para></listitem>
9214 <listitem><para>Fix INITIALLY DEFERRED bug</para></listitem>
9215 <listitem><para>Various bug fixes (Michael, Christof Petig)</para></listitem>
9216 <listitem><para>Auto allocation for indicator variable arrays (int *ind_p=NULL)</para></listitem>
9217 <listitem><para>Auto allocation for string arrays (char **foo_pp=NULL)</para></listitem>
9218 <listitem><para>ECPGfree_auto_mem fixed</para></listitem>
9219 <listitem><para>All function names with external linkage are now prefixed by ECPG</para></listitem>
9220 <listitem><para>Fixes for arrays of structures (Michael)</para></listitem>
9225 <title>Misc. Interfaces</title>
9227 <listitem><para>Python fix fetchone() (Gerhard Haring)</para></listitem>
9228 <listitem><para>Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Max)</para></listitem>
9229 <listitem><para>Add Tcl COPY TO/FROM (ljb)</para></listitem>
9230 <listitem><para>Prevent output of default index op class in <application>pg_dump</> (Tom)</para></listitem>
9231 <listitem><para>Fix libpgeasy memory leak (Bruce)</para></listitem>
9236 <title>Build and Install</title>
9238 <listitem><para>Configure, dynamic loader, and shared library fixes (Peter E)</para></listitem>
9239 <listitem><para>Fixes in QNX 4 port (Bernd Tegge)</para></listitem>
9240 <listitem><para>Fixes in Cygwin and Windows ports (Jason Tishler, Gerhard Haring, Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)</para></listitem>
9241 <listitem><para>Fix for Windows socket communication failures (Magnus, Mikhail Terekhov)</para></listitem>
9242 <listitem><para>Hurd compile fix (Oliver Elphick)</para></listitem>
9243 <listitem><para>BeOS fixes (Cyril Velter)</para></listitem>
9244 <listitem><para>Remove configure --enable-unicode-conversion, now enabled by multibyte (Tatsuo)</para></listitem>
9245 <listitem><para>AIX fixes (Tatsuo, Andreas)</para></listitem>
9246 <listitem><para>Fix parallel make (Peter E)</para></listitem>
9247 <listitem><para>Install SQL language manual pages into OS-specific directories (Peter E)</para></listitem>
9248 <listitem><para>Rename config.h to pg_config.h (Peter E)</para></listitem>
9249 <listitem><para>Reorganize installation layout of header files (Peter E)</para></listitem>
9254 <title>Source Code</title>
9256 <listitem><para>Remove SEP_CHAR (Bruce)</para></listitem>
9257 <listitem><para>New GUC hooks (Tom)</para></listitem>
9258 <listitem><para>Merge GUC and command line handling (Marko Kreen)</para></listitem>
9259 <listitem><para>Remove EXTEND INDEX (Martijn van Oosterhout, Tom)</para></listitem>
9260 <listitem><para>New pgjindent utility to indent java code (Bruce)</para></listitem>
9261 <listitem><para>Remove define of true/false when compiling under C++ (Leandro Fanzone, Tom)</para></listitem>
9262 <listitem><para>pgindent fixes (Bruce, Tom)</para></listitem>
9263 <listitem><para>Replace strcasecmp() with strcmp() where appropriate (Peter E)</para></listitem>
9264 <listitem><para>Dynahash portability improvements (Tom)</para></listitem>
9265 <listitem><para>Add 'volatile' usage in spinlock structures</para></listitem>
9266 <listitem><para>Improve signal handling logic (Tom)</para></listitem>
9271 <title>Contrib</title>
9273 <listitem><para>New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)</para></listitem>
9274 <listitem><para>New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)</para></listitem>
9275 <listitem><para>Add contrib/dblink for remote database access (Joe Conway)</para></listitem>
9276 <listitem><para>contrib/ora2pg Oracle conversion utility (Gilles Darold)</para></listitem>
9277 <listitem><para>contrib/xml XML conversion utility (John Gray)</para></listitem>
9278 <listitem><para>contrib/fulltextindex fixes (Christopher Kings-Lynne)</para></listitem>
9279 <listitem><para>New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merged (Joe Conway)</para></listitem>
9280 <listitem><para>Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartunov)</para></listitem>
9281 <listitem><para>New pg_upgrade utility (Bruce)</para></listitem>
9282 <listitem><para>Add new pg_resetxlog options (Bruce, Tom)</para></listitem>
9289 <sect1 id="release-7-1-3">
9290 <title>Release 7.1.3</title>
9293 <title>Release date</title>
9294 <simpara>2001-08-15</simpara>
9298 <title>Migration to version 7.1.3</title>
9301 A dump/restore is <emphasis>not</emphasis> required for those running
9307 <title>Changes</title>
9311 Remove unused WAL segements of large transactions (Tom)
9312 Multiaction rule fix (Tom)
9313 PL/pgSQL memory allocation fix (Jan)
9314 VACUUM buffer fix (Tom)
9315 Regression test fixes (Tom)
9316 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
9317 Fix subselects with DISTINCT ON or LIMIT (Tom)
9319 Disable COPY TO/FROM a view (Tom)
9320 Cygwin build (Jason Tishler)
9327 <sect1 id="release-7-1-2">
9328 <title>Release 7.1.2</title>
9331 <title>Release date</title>
9332 <simpara>2001-05-11</simpara>
9336 This has one fix from 7.1.1.
9341 <title>Migration to version 7.1.2</title>
9344 A dump/restore is <emphasis>not</emphasis> required for those running
9350 <title>Changes</title>
9354 Fix PL/pgSQL SELECTs when returning no rows
9355 Fix for psql backslash core dump
9356 Referential integrity privilege fix
9365 <sect1 id="release-7-1-1">
9366 <title>Release 7.1.1</title>
9369 <title>Release date</title>
9370 <simpara>2001-05-05</simpara>
9374 This has a variety of fixes from 7.1.
9379 <title>Migration to version 7.1.1</title>
9382 A dump/restore is <emphasis>not</emphasis> required for those running
9388 <title>Changes</title>
9392 Fix for numeric MODULO operator (Tom)
9393 pg_dump fixes (Philip)
9394 pg_dump can dump 7.0 databases (Philip)
9395 readline 4.2 fixes (Peter E)
9397 AIX, MSWIN, VAX, N32K fixes (Tom)
9398 Multibytes fixes (Tom)
9399 Unicode fixes (Tatsuo)
9400 Optimizer improvements (Tom)
9401 Fix for whole rows in functions (Tom)
9402 Fix for pg_ctl and option strings with spaces (Peter E)
9403 ODBC fixes (Hiroshi)
9404 EXTRACT can now take string argument (Thomas)
9405 Python fixes (Darcy)
9412 <sect1 id="release-7-1">
9413 <title>Release 7.1</title>
9416 <title>Release date</title>
9417 <simpara>2001-04-13</simpara>
9421 This release focuses on removing limitations that have existed in the
9422 <productname>PostgreSQL</productname> code for many years.
9426 Major changes in this release:
9432 Write-ahead Log (WAL)
9436 To maintain database consistency in case of an operating system crash,
9437 previous releases of <productname>PostgreSQL</productname> have forced
9438 all data modifications to disk before each transaction commit. With
9439 WAL, only one log file must be flushed to disk, greatly improving
9440 performance. If you have been using -F in previous releases to
9441 disable disk flushes, you may want to consider discontinuing its use.
9452 TOAST - Previous releases had a compiled-in row length limit,
9453 typically 8k - 32k. This limit made storage of long text fields
9454 difficult. With TOAST, long rows of any length can be stored with good
9466 We now support outer joins. The UNION/NOT IN
9467 workaround for outer joins is no longer required. We use the SQL92
9479 The previous C function manager did not
9480 handle null values properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
9481 function manager does. You can continue using your old custom
9482 functions, but you may want to rewrite them in the future to use the new
9483 function manager call interface.
9494 A large number of complex queries that were
9495 unsupported in previous releases now work. Many combinations of views,
9496 aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
9497 now work properly. Inherited tables are now accessed by default.
9498 Subqueries in FROM are now supported.
9506 <title>Migration to version 7.1</title>
9509 A dump/restore using pg_dump is required for those wishing to migrate
9510 data from any previous release.
9515 <title>Changes</title>
9521 Many multibyte/Unicode/locale fixes (Tatsuo and others)
9522 More reliable ALTER TABLE RENAME (Tom)
9523 Kerberos V fixes (David Wragg)
9524 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
9525 Prompt username/password on standard error (Bruce)
9526 Large objects inv_read/inv_write fixes (Tom)
9527 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
9529 Prevent query expressions from leaking memory (Tom)
9530 Allow UPDATE of arrays elements (Tom)
9531 Wake up lock waiters during cancel (Hiroshi)
9532 Fix rare cursor crash when using hash join (Tom)
9533 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
9534 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
9535 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
9536 Fix PL/perl (Alex Kapranoff)
9537 Disallow LOCK on views (Mark Hollomon)
9538 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
9539 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
9540 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
9541 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
9542 Fix rare failure with TRUNCATE command (Tom)
9543 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
9544 DISTINCT, ORDER BY, SELECT...INTO (Tom)
9545 Fix parser failures during aborted transactions (Tom)
9546 Allow temporary relations to properly clean up indexes (Bruce)
9547 Fix VACUUM problem with moving rows in same page (Tom)
9548 Modify pg_dump to better handle user-defined items in template1 (Philip)
9549 Allow LIMIT in VIEW (Tom)
9550 Require cursor FETCH to honor LIMIT (Tom)
9551 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
9552 Allow ORDER BY, LIMIT in subqueries (Tom)
9553 Allow UNION in CREATE RULE (Tom)
9554 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
9555 Store initdb collation in pg_control so collation cannot be changed (Tom)
9556 Fix INSERT...SELECT with rules (Tom)
9557 Fix FOR UPDATE inside views and subselects (Tom)
9558 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
9559 Fix lpad() and rpad() to handle length less than input string (Tom)
9560 Fix use of NOTIFY in some rules (Tom)
9561 Overhaul btree code (Tom)
9562 Fix NOT NULL use in Pl/pgSQL variables (Tom)
9563 Overhaul GIST code (Oleg)
9564 Fix CLUSTER to preserve constraints and column default (Tom)
9565 Improved deadlock detection handling (Tom)
9566 Allow multiple SERIAL columns in a table (Tom)
9567 Prevent occasional index corruption (Vadim)
9571 Add OUTER JOINs (Tom)
9572 Function manager overhaul (Tom)
9573 Allow ALTER TABLE RENAME on indexes (Tom)
9574 Improve CLUSTER (Tom)
9575 Improve ps status display for more platforms (Peter E, Marc)
9576 Improve CREATE FUNCTION failure message (Ross)
9577 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
9579 Grand Unified Configuration scheme/GUC. Many options can now be set in
9580 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
9581 Improved handling of file descriptor cache (Tom)
9582 New warning code about auto-created table alias entries (Bruce)
9583 Overhaul initdb process (Tom, Peter E)
9584 Overhaul of inherited tables; inherited tables now accessed by default;
9585 new ONLY key word prevents it (Chris Bitmead, Tom)
9586 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
9588 Allow renaming of temp tables (Tom)
9589 Overhaul memory manager contexts (Tom)
9590 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
9591 Overhaul pg_dump (Philip Warner)
9592 Allow pg_hba.conf secondary password file to specify only username (Peter E)
9593 Allow TEMPORARY or TEMP key word when creating temporary tables (Bruce)
9594 New memory leak checker (Karel)
9595 New SET SESSION CHARACTERISTICS (Thomas)
9596 Allow nested block comments (Thomas)
9597 Add WITHOUT TIME ZONE type qualifier (Thomas)
9598 New ALTER TABLE ADD CONSTRAINT (Stephan)
9599 Use NUMERIC accumulators for INTEGER aggregates (Tom)
9600 Overhaul aggregate code (Tom)
9601 New VARIANCE and STDDEV() aggregates
9602 Improve dependency ordering of pg_dump (Philip)
9603 New pg_restore command (Philip)
9604 New pg_dump tar output option (Philip)
9605 New pg_dump of large objects (Philip)
9606 New ESCAPE option to LIKE (Thomas)
9607 New case-insensitive LIKE - ILIKE (Thomas)
9608 Allow functional indexes to use binary-compatible type (Tom)
9609 Allow SQL functions to be used in more contexts (Tom)
9610 New pg_config utility (Peter E)
9611 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
9613 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
9614 New quote_identifiers() and quote_literal() functions (Jan)
9615 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
9616 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
9617 Update PyGreSQL to version 3.1 (D'Arcy)
9618 Store tables as files named by OID (Vadim)
9619 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
9620 Require DROP VIEW to remove views, no DROP TABLE (Mark)
9621 Allow DROP VIEW view1, view2 (Mark)
9622 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
9623 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
9624 New /contrib/pgcrypto hashing functions (Marko Kreen)
9625 New pg_dumpall --globals-only option (Peter E)
9626 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
9627 New AT TIME ZONE syntax (Thomas)
9628 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
9629 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
9630 Allow socket path name to be specified in hostname by using leading slash
9631 (David J. MacKenzie)
9632 Allow CREATE DATABASE to specify template database (Tom)
9633 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
9634 New /contrib/rserv replication toolkit (Vadim)
9635 New file format for COPY BINARY (Tom)
9636 New /contrib/oid2name to map numeric files to table names (B Palmer)
9637 New "idle in transaction" ps status message (Marc)
9638 Update to pgaccess 0.98.7 (Constantin Teodorescu)
9639 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
9640 Add rudimentary dependency checking to pg_dump (Philip)
9644 Fix INET/CIDR type ordering and add new functions (Tom)
9645 Make OID behave as an unsigned type (Tom)
9646 Allow BIGINT as synonym for INT8 (Peter E)
9647 New int2 and int8 comparison operators (Tom)
9648 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
9649 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
9650 New GIST seg/cube examples (Gene Selkov)
9651 Improved round(numeric) handling (Tom)
9652 Fix CIDR output formatting (Tom)
9653 New CIDR abbrev() function (Tom)
9657 Write-Ahead Log (WAL) to provide crash recovery with less performance
9659 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
9660 Reduced file seeks (Denis Perchine)
9661 Improve BTREE code for duplicate keys (Tom)
9662 Store all large objects in a single table (Denis Perchine, Tom)
9663 Improve memory allocation performance (Karel, Tom)
9667 New function manager call conventions (Tom)
9668 SGI portability fixes (David Kaelbling)
9669 New configure --enable-syslog option (Peter E)
9670 New BSDI README (Bruce)
9671 configure script moved to top level, not /src (Peter E)
9672 Makefile/configuration/compilation overhaul (Peter E)
9673 New configure --with-python option (Peter E)
9674 Solaris cleanups (Peter E)
9675 Overhaul /contrib Makefiles (Karel)
9676 New OpenSSL configuration option (Magnus, Peter E)
9678 QNX fixes (Maurizio)
9679 New heap_open(), heap_openr() API (Tom)
9680 Remove colon and semi-colon operators (Thomas)
9681 New pg_class.relkind value for views (Mark Hollomon)
9682 Rename ichar() to chr() (Karel)
9683 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
9684 Fixes for NT/Cygwin (Pete Forman)
9685 AIX port fixes (Andreas)
9686 New BeOS port (David Reid, Cyril Velter)
9687 Add proofreader's changes to docs (Addison-Wesley, Bruce)
9688 New Alpha spinlock code (Adriaan Joubert, Compaq)
9689 UnixWare port overhaul (Peter E)
9690 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
9691 New FreeBSD Alpha port (Alfred)
9692 Overhaul shared memory segments (Tom)
9693 Add IBM S/390 support (Neale Ferguson)
9694 Moved macmanuf to /contrib (Larry Rosenman)
9695 Syslog improvements (Larry Rosenman)
9696 New template0 database that contains no user additions (Tom)
9697 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
9698 Allow NetBSD's libedit instead of readline (Peter)
9699 Improved assembly language source code format (Bruce)
9700 New contrib/pg_logger
9701 New --template option to createdb
9702 New contrib/pg_control utility (Oliver)
9703 New FreeBSD tools ipc_check, start-scripts/freebsd
9710 <sect1 id="release-7-0-3">
9711 <title>Release 7.0.3</title>
9714 <title>Release date</title>
9715 <simpara>2000-11-11</simpara>
9719 This has a variety of fixes from 7.0.2.
9724 <title>Migration to version 7.0.3</title>
9727 A dump/restore is <emphasis>not</emphasis> required for those running
9733 <title>Changes</title>
9738 Large object fix (Tom)
9739 Fix lean in COPY WITH OIDS leak (Tom)
9740 Fix backwards-index-scan (Tom)
9741 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
9742 Add --enable-syslog to configure (Marc)
9743 Fix abort transaction at backend exit in rare cases (Tom)
9744 Fix for psql \l+ when multibyte enabled (Tatsuo)
9745 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
9746 Make vacuum always flush buffers (Tom)
9747 Fix to allow cancel while waiting for a lock (Hiroshi)
9748 Fix for memory aloocation problem in user authentication code (Tom)
9749 Remove bogus use of int4out() (Tom)
9750 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
9751 Fix for failure of triggers on heap open in certain cases (Jeroen van
9753 Fix for erroneous selectivity of not-equals (Tom)
9754 Fix for erroneous use of strcmp() (Tom)
9755 Fix for bug where storage manager accesses items beyond end of file
9757 Fix to include kernel errno message in all smgr elog messages (Tom)
9758 Fix for '.' not in PATH at build time (SL Baur)
9759 Fix for out-of-file-descriptors error (Tom)
9760 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
9761 Fix for subselect in targetlist of Append node (Tom)
9762 Fix for mergejoin plans (Tom)
9763 Fix TRUNCATE failure on relations with indexes (Tom)
9764 Avoid database-wide restart on write error (Hiroshi)
9765 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
9766 Fix VACUUM problem with moving chain of update row versions when source
9767 and destination of a row version lie on the same page (Tom)
9768 Fix user.c CommandCounterIncrement (Tom)
9769 Fix for AM/PM boundary problem in to_char() (Karel Zak)
9770 Fix TIME aggregate handling (Tom)
9771 Fix to_char() to avoid coredump on NULL input (Tom)
9773 Fix for inserting/copying longer multibyte strings into char() data
9775 Fix for crash of backend, on abort (Tom)
9782 <sect1 id="release-7-0-2">
9783 <title>Release 7.0.2</title>
9786 <title>Release date</title>
9787 <simpara>2000-06-05</simpara>
9791 This is a repackaging of 7.0.1 with added documentation.
9796 <title>Migration to version 7.0.2</title>
9799 A dump/restore is <emphasis>not</emphasis> required for those running
9805 <title>Changes</title>
9809 Added documentation to tarball.
9816 <sect1 id="release-7-0-1">
9817 <title>Release 7.0.1</title>
9820 <title>Release date</title>
9821 <simpara>2000-06-01</simpara>
9825 This is a cleanup release for 7.0.
9829 <title>Migration to version 7.0.1</title>
9832 A dump/restore is <emphasis>not</emphasis> required for those running
9838 <title>Changes</title>
9842 Fix many CLUSTER failures (Tom)
9843 Allow ALTER TABLE RENAME works on indexes (Tom)
9844 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
9845 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
9846 Fix the off by one errors in ResultSet from 6.5.3, and more.
9847 jdbc ResultSet fixes (Joseph Shraibman)
9848 optimizer tunings (Tom)
9849 Fix create user for pgaccess
9850 Fix for UNLISTEN failure
9851 IRIX fixes (David Kaelbling)
9852 QNX fixes (Andreas Kardos)
9853 Reduce COPY IN lock level (Tom)
9854 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
9855 Fix pg_dump to handle OID indexes (Tom)
9856 Fix small memory leak (Tom)
9857 Solaris fix for createdb/dropdb (Tatsuo)
9858 Fix for non-blocking connections (Alfred Perlstein)
9859 Fix improper recovery after RENAME TABLE failures (Tom)
9860 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
9861 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
9862 Fix too long syslog message (Tatsuo)
9863 Fix problem with quoted indexes that are too long (Tom)
9864 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
9865 ecpg changes (Michael)
9871 <sect1 id="release-7-0">
9872 <title>Release 7.0</title>
9875 <title>Release date</title>
9876 <simpara>2000-05-08</simpara>
9880 This release contains improvements in many areas, demonstrating
9881 the continued growth of <productname>PostgreSQL</productname>.
9882 There are more improvements and fixes in 7.0 than in any previous
9883 release. The developers have confidence that this is the best
9884 release yet; we do our best to put out only solid releases, and
9885 this one is no exception.
9889 Major changes in this release:
9899 Foreign keys are now implemented, with the exception of PARTIAL MATCH
9900 foreign keys. Many users have been asking for this feature, and we are
9901 pleased to offer it.
9912 Continuing on work started a year ago, the optimizer has been
9913 improved, allowing better query plan selection and faster performance
9914 with less memory usage.
9921 Updated <application>psql</application>
9925 <application>psql</application>, our interactive terminal monitor, has been
9926 updated with a variety of new features. See the <application>psql</application> manual page for details.
9937 SQL92 join syntax is now supported, though only as
9938 <literal>INNER JOIN</> for this release. <literal>JOIN</>,
9939 <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
9940 and <literal>JOIN</>/<literal>ON</> are available, as are
9941 column correlation names.
9949 <title>Migration to version 7.0</title>
9952 A dump/restore using <application>pg_dump</application>
9953 is required for those wishing to migrate data from any
9954 previous release of <productname>PostgreSQL</productname>.
9955 For those upgrading from 6.5.*, you may instead use
9956 <application>pg_upgrade</application> to upgrade to this
9957 release; however, a full dump/reload installation is always the
9958 most robust method for upgrades.
9962 Interface and compatibility issues to consider for the new
9969 The date/time types <type>datetime</type> and
9970 <type>timespan</type> have been superseded by the
9971 SQL92-defined types <type>timestamp</type> and
9972 <type>interval</type>. Although there has been some effort to
9973 ease the transition by allowing
9974 <productname>PostgreSQL</productname> to recognize
9975 the deprecated type names and translate them to the new type
9976 names, this mechanism may not be completely transparent to
9977 your existing application.
9983 The optimizer has been substantially improved in the area of
9984 query cost estimation. In some cases, this will result in
9985 decreased query times as the optimizer makes a better choice
9986 for the preferred plan. However, in a small number of cases,
9987 usually involving pathological distributions of data, your
9988 query times may go up. If you are dealing with large amounts
9989 of data, you may want to check your queries to verify
9996 The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
9997 interfaces have been upgraded and extended.
10003 The string function <function>CHAR_LENGTH</function> is now a
10004 native function. Previous versions translated this into a call
10005 to <function>LENGTH</function>, which could result in
10006 ambiguity with other types implementing
10007 <function>LENGTH</function> such as the geometric types.
10014 <title>Changes</title>
10020 Prevent function calls exceeding maximum number of arguments (Tom)
10021 Improve CASE construct (Tom)
10022 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
10023 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
10024 Fix GROUP BY scan bug (Tom)
10025 Improvements in SQL grammar processing (Tom)
10026 Fix for views involved in INSERT ... SELECT ... (Tom)
10027 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
10028 Fix for subselects in INSERT ... SELECT (Tom)
10029 Prevent INSERT ... SELECT ... ORDER BY (Tom)
10030 Fixes for relations greater than 2GB, including vacuum
10031 Improve propagating system table changes to other backends (Tom)
10032 Improve propagating user table changes to other backends (Tom)
10033 Fix handling of temp tables in complex situations (Bruce, Tom)
10034 Allow table locking at table open, improving concurrent reliability (Tom)
10035 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
10036 Prevent DROP DATABASE while others accessing
10037 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
10038 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
10039 Fix pg_upgrade so it works for MVCC (Tom)
10040 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
10041 Fix for "f1 datetime DEFAULT 'now'" (Tom)
10042 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
10043 Allow comment-only lines, and ;;; lines too. (Tom)
10044 Improve recovery after failed disk writes, disk full (Hiroshi)
10045 Fix cases where table is mentioned in FROM but not joined (Tom)
10046 Allow HAVING clause without aggregate functions (Tom)
10047 Fix for "--" comment and no trailing newline, as seen in perl interface
10048 Improve pg_dump failure error reports (Bruce)
10049 Allow sorts and hashes to exceed 2GB file sizes (Tom)
10050 Fix for pg_dump dumping of inherited rules (Tom)
10051 Fix for NULL handling comparisons (Tom)
10052 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
10053 Fix for dbname with dash
10054 Prevent DROP INDEX from interfering with other backends (Tom)
10055 Fix file descriptor leak in verify_password()
10056 Fix for "Unable to identify an operator =$" problem
10057 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
10058 Fix for recursive exit call (Massimo)
10059 Fix for extra-long timezones (Jeroen van Vianen)
10060 Make pg_dump preserve primary key information (Peter E)
10061 Prevent databases with single quotes (Peter E)
10062 Prevent DROP DATABASE inside transaction (Peter E)
10063 ecpg memory leak fixes (Stephen Birch)
10064 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
10065 Y2K timestamp fix (Massimo)
10066 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
10067 Fix for views with tables/columns containing spaces (Tom)
10068 Prevent privileges on indexes (Peter E)
10069 Fix for spinlock stuck problem when error is generated (Hiroshi)
10070 Fix ipcclean on Linux
10071 Fix handling of NULL constraint conditions (Tom)
10072 Fix memory leak in odbc driver (Nick Gorham)
10073 Fix for privilege check on UNION tables (Tom)
10074 Fix to allow SELECT 'a' LIKE 'a' (Tom)
10075 Fix for SELECT 1 + NULL (Tom)
10077 Fix log() on numeric type (Tom)
10078 Deprecate ':' and ';' operators
10079 Allow vacuum of temporary tables
10080 Disallow inherited columns with the same name as new columns
10081 Recover or force failure when disk space is exhausted (Hiroshi)
10082 Fix INSERT INTO ... SELECT with AS columns matching result columns
10083 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns (Tom)
10084 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
10085 Fix UNION with LIMIT
10086 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
10087 Fix CREATE TABLE test(col char(2) DEFAULT user)
10088 Fix mismatched types in CREATE TABLE ... DEFAULT
10089 Fix SELECT * FROM pg_class where oid in (0,-1)
10090 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
10091 Prevent user who can create databases can modifying pg_database table (Peter E)
10092 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
10093 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
10097 New CLI interface include file sqlcli.h, based on SQL3/SQL98
10098 Remove all limits on query length, row length limit still exists (Tom)
10099 Update jdbc protocol to 2.0 (Jens Glaser <email>jens@jens.de</email>)
10100 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
10101 Fix to give super user and createdb user proper update catalog rights (Peter E)
10102 Allow ecpg bool variables to have NULL values (Christof)
10103 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
10104 Allow ^C to cancel COPY command (Massimo)
10105 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
10106 Function name overloading for dynamically-loaded C functions (Frankpitt)
10107 Add CmdTuples() to libpq++(Vince)
10108 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
10109 Allow CREATE FUNCTION/WITH clause to be used for all language types
10110 configure --enable-debug adds -g (Peter E)
10111 configure --disable-debug removes -g (Peter E)
10112 Allow more complex default expressions (Tom)
10113 First real FOREIGN KEY constraint trigger functionality (Jan)
10114 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
10115 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
10116 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
10117 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
10118 Change pgeasy connectdb() parameter ordering (Bruce)
10119 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
10120 Add Oracle's COMMENT ON command (Mike Mascari <email>mascarim@yahoo.com</email>)
10121 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
10122 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
10123 Make USING in COPY optional (Bruce)
10124 Allow subselects in the target list (Tom)
10125 Allow subselects on the left side of comparison operators (Tom)
10126 New parallel regression test (Jan)
10127 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
10128 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
10129 Added psql LASTOID variable to return last inserted oid (Peter E)
10130 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
10131 Add privilege check for vacuum (Peter E)
10132 New libpq functions to allow asynchronous connections: PQconnectStart(),
10133 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
10134 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
10135 New libpq PQsetenv() function (Ewan Mellor)
10136 create/alter user extension (Peter E)
10137 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
10138 New scripts for create/drop user/db (Peter E)
10139 Major psql overhaul (Peter E)
10140 Add const to libpq interface (Peter E)
10141 New libpq function PQoidValue (Peter E)
10142 Show specific non-aggregate causing problem with GROUP BY (Tom)
10143 Make changes to pg_shadow recreate pg_pwd file (Peter E)
10144 Add aggregate(DISTINCT ...) (Tom)
10145 Allow flag to control COPY input/output of NULLs (Peter E)
10146 Make postgres user have a password by default (Peter E)
10147 Add CREATE/ALTER/DROP GROUP (Peter E)
10148 All administration scripts now support --long options (Peter E, Karel)
10149 Vacuumdb script now supports --all option (Peter E)
10150 ecpg new portable FETCH syntax
10151 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
10152 and EXEC SQL ENDIF directives
10153 Add pg_ctl script to control backend start-up (Tatsuo)
10154 Add postmaster.opts.default file to store start-up flags (Tatsuo)
10155 Allow --with-mb=SQL_ASCII
10156 Increase maximum number of index keys to 16 (Bruce)
10157 Increase maximum number of function arguments to 16 (Bruce)
10158 Allow configuration of maximum number of index keys and arguments (Bruce)
10159 Allow unprivileged users to change their passwords (Peter E)
10160 Password authentication enabled; required for new users (Peter E)
10161 Disallow dropping a user who owns a database (Peter E)
10162 Change initdb option --with-mb to --enable-multibyte
10163 Add option for initdb to prompts for superuser password (Peter E)
10164 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
10165 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
10166 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
10167 libpq non-blocking mode (Alfred Perlstein)
10168 Improve conversion of types in casts that don't specify a length
10169 New plperl internal programming language (Mark Hollomon)
10170 Allow COPY IN to read file that do not end with a newline (Tom)
10171 Indicate when long identifiers are truncated (Tom)
10172 Allow aggregates to use type equivalency (Peter E)
10173 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
10174 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
10175 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
10176 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
10177 Add NUMERIC and int8 types to ODBC
10178 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
10179 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
10180 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
10181 Enable backward sequential scan even after reaching EOF (Hiroshi)
10182 Add btree indexing of boolean values, >= and <= (Don Baccus)
10183 Print current line number when COPY FROM fails (Massimo)
10184 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
10185 Add DEC as synonym for DECIMAL (Thomas)
10186 Add SESSION_USER as SQL92 key word, same as CURRENT_USER (Thomas)
10187 Implement SQL92 column aliases (aka correlation names) (Thomas)
10188 Implement SQL92 join syntax (Thomas)
10189 Make INTERVAL reserved word allowed as a column identifier (Thomas)
10190 Implement REINDEX command (Hiroshi)
10191 Accept ALL in aggregate function SUM(ALL col) (Tom)
10192 Prevent GROUP BY from using column aliases (Tom)
10193 New psql \encoding option (Tatsuo)
10194 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
10195 Allow negation of a negative number in all cases
10196 Add ecpg descriptors (Christof, Michael)
10197 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
10198 Allow casts with length, like foo::char(8)
10199 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
10200 Add support for SJIS user defined characters (Tatsuo)
10201 Larger views/rules supported
10202 Make libpq's PQconndefaults() thread-safe (Tom)
10203 Disable // as comment to be ANSI conforming, should use -- (Tom)
10204 Allow column aliases on views CREATE VIEW name (collist)
10205 Fixes for views with subqueries (Tom)
10206 Allow UPDATE table SET fld = (SELECT ...) (Tom)
10207 SET command options no longer require quotes
10208 Update pgaccess to 0.98.6
10209 New SET SEED command
10210 New pg_options.sample file
10211 New SET FSYNC command (Massimo)
10212 Allow pg_descriptions when creating tables
10213 Allow pg_descriptions when creating types, columns, and functions
10214 Allow psql \copy to allow delimiters (Peter E)
10215 Allow psql to print nulls as distinct from "" [null] (Peter E)
10219 Many array fixes (Tom)
10220 Allow bare column names to be subscripted as arrays (Tom)
10221 Improve type casting of int and float constants (Tom)
10222 Cleanups for int8 inputs, range checking, and type conversion (Tom)
10223 Fix for SELECT timespan('21:11:26'::time) (Tom)
10224 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
10225 Add btree index on NUMERIC (Jan)
10226 Perl fix for large objects containing NUL characters (Douglas Thomson)
10227 ODBC fix for for large objects (free)
10228 Fix indexing of cidr data type
10229 Fix for Ethernet MAC addresses (macaddr type) comparisons
10230 Fix for date/time types when overflows happened in computations (Tom)
10231 Allow array on int8 (Peter E)
10232 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
10233 Allow NUMERIC arrays
10234 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
10235 Make char_length()/octet_length including trailing blanks (Tom)
10236 Made abstime/reltime use int4 instead of time_t (Peter E)
10237 New lztext data type for compressed text fields
10238 Revise code to handle coercion of int and float constants (Tom)
10239 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
10240 NUMERIC now accepts scientific notation (Tom)
10241 NUMERIC to int4 rounds (Tom)
10242 Convert float4/8 to NUMERIC properly (Tom)
10243 Allow type conversion with NUMERIC (Thomas)
10244 Make ISO date style (2000-02-16 09:33) the default (Thomas)
10245 Add NATIONAL CHAR [ VARYING ] (Thomas)
10246 Allow NUMERIC round and trunc to accept negative scales (Tom)
10247 New TIME WITH TIME ZONE type (Thomas)
10248 Add MAX()/MIN() on time type (Thomas)
10249 Add abs(), mod(), fac() for int8 (Thomas)
10250 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
10251 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
10252 Add exp() and ln() for NUMERIC type
10253 Rename NUMERIC power() to pow() (Thomas)
10254 Improved TRANSLATE() function (Edwin Ramirez, Tom)
10255 Allow X=-Y operators (Tom)
10256 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom)
10257 Allow LOCALE to use indexes in regular expression searches (Tom)
10258 Allow creation of functional indexes to use default types
10262 Prevent exponential space consumption with many AND's and OR's (Tom)
10263 Collect attribute selectivity values for system columns (Tom)
10264 Reduce memory usage of aggregates (Tom)
10265 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
10266 Fix r-tree index optimizer selectivity (Thomas)
10267 Improve optimizer selectivity computations and functions (Tom)
10268 Optimize btree searching for cases where many equal keys exist (Tom)
10269 Enable fast LIKE index processing only if index present (Tom)
10270 Re-use free space on index pages with duplicates (Tom)
10271 Improve hash join processing (Tom)
10272 Prevent descending sort if result is already sorted(Hiroshi)
10273 Allow commuting of index scan query qualifications (Tom)
10274 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
10275 Allocate large memory requests in fix-sized chunks for performance (Tom)
10276 Fix vacuum's performance by reducing memory allocation requests (Tom)
10277 Implement constant-expression simplification (Bernard Frankpitt, Tom)
10278 Use secondary columns to be used to determine start of index scan (Hiroshi)
10279 Prevent quadruple use of disk space when doing internal sorting (Tom)
10280 Faster sorting by calling fewer functions (Tom)
10281 Create system indexes to match all system caches (Bruce, Hiroshi)
10282 Make system caches use system indexes (Bruce)
10283 Make all system indexes unique (Bruce)
10284 Improve pg_statistics management for VACUUM speed improvement (Tom)
10285 Flush backend cache less frequently (Tom, Hiroshi)
10286 COPY now reuses previous memory allocation, improving performance (Tom)
10287 Improve optimization cost estimation (Tom)
10288 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom)
10289 Use DNF instead of CNF where appropriate (Tom, Taral)
10290 Further cleanup for OR-of-AND WHERE-clauses (Tom)
10291 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
10292 Smarter optimizer computations for random index page access (Tom)
10293 New SET variable to control optimizer costs (Tom)
10294 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
10295 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
10296 Major subquery speedup (Tom)
10297 Fewer fsync writes when fsync is not disabled (Tom)
10298 Improved LIKE optimizer estimates (Tom)
10299 Prevent fsync in SELECT-only queries (Vadim)
10300 Make index creation use psort code, because it is now faster (Tom)
10301 Allow creation of sort temp tables > 1 Gig
10303 Source Tree Changes
10304 -------------------
10305 Fix for linux PPC compile
10306 New generic expression-tree-walker subroutine (Tom)
10307 Change form() to varargform() to prevent portability problems
10308 Improved range checking for large integers on Alphas
10309 Clean up #include in /include directory (Bruce)
10310 Add scripts for checking includes (Bruce)
10311 Remove un-needed #include's from *.c files (Bruce)
10312 Change #include's to use <> and "" as appropriate (Bruce)
10313 Enable Windows compilation of libpq
10314 Alpha spinlock fix from Uncle George <email>gatgul@voicenet.com</email>
10315 Overhaul of optimizer data structures (Tom)
10316 Fix to cygipc library (Yutaka Tanida)
10317 Allow pgsql to work on newer Cygwin snapshots (Dan)
10318 New catalog version number (Tom)
10320 Rename heap_replace to heap_update
10321 Update for QNX (Dr. Andreas Kardos)
10322 New platform-specific regression handling (Tom)
10323 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
10324 Included all yacc and lex files into the distribution (Peter E.)
10325 Remove lextest, no longer needed (Peter E)
10326 Fix for libpq and psql on Windows (Magnus)
10327 Internally change datetime and timespan into timestamp and interval (Thomas)
10328 Fix for plpgsql on BSD/OS
10329 Add SQL_ASCII test case to the regression test (Tatsuo)
10330 configure --with-mb now deprecated (Tatsuo)
10332 NetBSD fixes (Johnny C. Lam <email>lamj@stat.cmu.edu</email>)
10333 Fixes for Alpha compiles
10334 New multibyte encodings
10340 <sect1 id="release-6-5-3">
10341 <title>Release 6.5.3</title>
10344 <title>Release date</title>
10345 <simpara>1999-10-13</simpara>
10349 This is basically a cleanup release for 6.5.2. We have added a new
10350 <application>PgAccess</> that was missing in 6.5.2, and installed an NT-specific fix.
10355 <title>Migration to version 6.5.3</title>
10358 A dump/restore is <emphasis>not</emphasis> required for those running
10363 <title>Changes</title>
10367 Updated version of pgaccess 0.98
10369 Fix dumping rules on inherited tables
10376 <sect1 id="release-6-5-2">
10377 <title>Release 6.5.2</title>
10380 <title>Release date</title>
10381 <simpara>1999-09-15</simpara>
10385 This is basically a cleanup release for 6.5.1. We have fixed a variety of
10386 problems reported by 6.5.1 users.
10391 <title>Migration to version 6.5.2</title>
10394 A dump/restore is <emphasis>not</emphasis> required for those running
10400 <title>Changes</title>
10404 subselect+CASE fixes(Tom)
10405 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
10406 Fixes for CASE in WHERE join clauses(Tom)
10407 Fix BTScan abort(Tom)
10408 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
10409 Improve it so that it checks for multicolumn constraints(Thomas)
10410 Fix for Windows making problem with MB enabled(Hiroki Kataoka)
10411 Allow BSD yacc and bison to compile pl code(Bruce)
10412 Fix SET NAMES working
10414 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
10415 Reduce the total memory consumption of vacuum(Tom)
10416 Fix for timestamp(datetime)
10417 Rule deparsing bugfixes(Tom)
10418 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(Tom)
10419 This is to re-use space on index pages freed by vacuum(Vadim)
10420 document -x for pg_dump(Bruce)
10421 Fix for unary operators in rule deparser(Tom)
10422 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
10423 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
10424 Repair logic error in LIKE: should not return LIKE_ABORT
10425 when reach end of pattern before end of text(Tom)
10426 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom)
10427 Updated version of pgaccess 0.98
10433 <sect1 id="release-6-5-1">
10434 <title>Release 6.5.1</title>
10437 <title>Release date</title>
10438 <simpara>1999-07-15</simpara>
10442 This is basically a cleanup release for 6.5. We have fixed a variety of
10443 problems reported by 6.5 users.
10447 <title>Migration to version 6.5.1</title>
10450 A dump/restore is <emphasis>not</emphasis> required for those running
10456 <title>Changes</title>
10461 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
10462 Remove QUERY_LIMIT, use SELECT...LIMIT
10463 Fix for EXPLAIN on inheritance(Tom)
10464 Patch to allow vacuum on multisegment tables(Hiroshi)
10465 R-Tree optimizer selectivity fix(Tom)
10466 ACL file descriptor leak fix(Atsushi Ogawa)
10467 New expresssion subtree code(Tom)
10468 Avoid disk writes for read-only transactions(Vadim)
10469 Fix for removal of temp tables if last transaction was aborted(Bruce)
10470 Fix to prevent too large row from being created(Bruce)
10472 Allow port numbers 32k - 64k(Bruce)
10473 Add ^ precidence(Bruce)
10474 Rename sort files called pg_temp to pg_sorttemp(Bruce)
10475 Fix for microseconds in time values(Tom)
10476 Tutorial source cleanup
10477 New linux_m68k port
10478 Fix for sorting of NULL's in some cases(Tom)
10479 Shared library dependencies fixed (Tom)
10480 Fixed glitches affecting GROUP BY in subselects(Tom)
10481 Fix some compiler warnings (Tomoaki Nishiyama)
10482 Add Win1250 (Czech) support (Pavel Behal)
10488 <sect1 id="release-6-5">
10489 <title>Release 6.5</title>
10492 <title>Release date</title>
10493 <simpara>1999-06-09</simpara>
10497 This release marks a major step in the development team's mastery of the source
10498 code we inherited from Berkeley. You will see we are now easily adding
10499 major features, thanks to the increasing size and experience of our
10500 world-wide development team.
10504 Here is a brief summary of the more notable changes:
10509 Multiversion concurrency control(MVCC)
10513 This removes our old table-level locking, and replaces it with
10514 a locking system that is superior to most commercial database
10515 systems. In a traditional system, each row that is modified
10516 is locked until committed, preventing reads by other users.
10517 MVCC uses the natural multiversion nature of
10518 <productname>PostgreSQL</productname> to allow readers to
10519 continue reading consistent data during writer activity.
10520 Writers continue to use the compact pg_log transaction system.
10521 This is all performed without having to allocate a lock for
10522 every row like traditional database systems. So, basically,
10523 we no longer are restricted by simple table-level locking; we
10524 have something better than row-level locking.
10531 Hot backups from <application>pg_dump</application>
10535 <application>pg_dump</application> takes advantage of the new
10536 MVCC features to give a consistent database dump/backup while
10537 the database stays online and available for queries.
10548 We now have a true numeric data type, with
10549 user-specified precision.
10560 Temporary tables are guaranteed to have unique names
10561 within a database session, and are destroyed on session exit.
10572 We now have CASE, INTERSECT, and EXCEPT statement
10573 support. We have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL,
10574 SELECT ... FOR UPDATE, and an improved LOCK TABLE command.
10585 We continue to speed up <productname>PostgreSQL</productname>,
10586 thanks to the variety of talents within our team. We have
10587 sped up memory allocation, optimization, table joins, and row
10599 We continue to expand our port list, this time including
10600 <systemitem class="osname">Windows NT</>/<systemitem>ix86</> and <systemitem class="osname">NetBSD</>/<systemitem>arm32</>.
10611 Most interfaces have new versions, and existing functionality
10623 New and updated material is present throughout the
10624 documentation. New <acronym>FAQ</acronym>s have been
10625 contributed for <systemitem class="osname">SGI</> and <systemitem class="osname">AIX</> platforms.
10626 The <citetitle>Tutorial</citetitle> has introductory information
10627 on <acronym>SQL</acronym> from Stefan Simkovics.
10628 For the <citetitle>User's Guide</citetitle>, there are
10629 reference pages covering the postmaster and more utility
10630 programs, and a new appendix
10631 contains details on date/time behavior.
10632 The <citetitle>Administrator's Guide</citetitle> has a new
10633 chapter on troubleshooting from Tom Lane.
10634 And the <citetitle>Programmer's Guide</citetitle> has a
10635 description of query processing, also from Stefan, and details
10636 on obtaining the <productname>PostgreSQL</productname> source
10637 tree via anonymous <productname>CVS</productname> and
10638 <productname>CVSup</productname>.
10646 <title>Migration to version 6.5</title>
10649 A dump/restore using <application>pg_dump</application>
10650 is required for those wishing to migrate data from any
10651 previous release of <productname>PostgreSQL</productname>.
10652 <application>pg_upgrade</application> can <emphasis>not</emphasis>
10653 be used to upgrade to this release because the on-disk structure
10654 of the tables has changed compared to previous releases.
10658 The new Multiversion Concurrency Control (MVCC) features can
10659 give somewhat different behaviors in multiuser
10660 environments. <emphasis>Read and understand the following section
10661 to ensure that your existing applications will give you the
10662 behavior you need.</emphasis>
10666 <title>Multiversion Concurrency Control</title>
10669 Because readers in 6.5 don't lock data, regardless of transaction
10670 isolation level, data read by one transaction can be overwritten by
10671 another. In other words, if a row is returned by
10672 <command>SELECT</command> it doesn't mean that this row really exists
10673 at the time it is returned (i.e. sometime after the statement or
10674 transaction began) nor that the row is protected from being deleted or
10675 updated by concurrent transactions before the current transaction does
10676 a commit or rollback.
10680 To ensure the actual existence of a row and protect it against
10681 concurrent updates one must use <command>SELECT FOR UPDATE</command> or
10682 an appropriate <command>LOCK TABLE</command> statement. This should be
10683 taken into account when porting applications from previous releases of
10684 <productname>PostgreSQL</productname> and other environments.
10688 Keep the above in mind if you are using
10689 <filename>contrib/refint.*</filename> triggers for
10690 referential integrity. Additional techniques are required now. One way is
10691 to use <command>LOCK parent_table IN SHARE ROW EXCLUSIVE MODE</command>
10692 command if a transaction is going to update/delete a primary key and
10693 use <command>LOCK parent_table IN SHARE MODE</command> command if a
10694 transaction is going to update/insert a foreign key.
10698 Note that if you run a transaction in SERIALIZABLE mode then you must
10699 execute the <command>LOCK</command> commands above before execution of any
10700 <acronym>DML</acronym> statement
10701 (<command>SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO</command>) in the
10708 These inconveniences will disappear in the future
10709 when the ability to read dirty
10710 (uncommitted) data (regardless of isolation level) and true referential
10711 integrity will be implemented.
10717 <title>Changes</title>
10723 Fix text<->float8 and text<->float4 conversion functions(Thomas)
10724 Fix for creating tables with mixed-case constraints(Billy)
10725 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
10726 Fix bug in pg_dump -z
10727 Memory overrun cleanups(Tatsuo)
10728 Fix for lo_import crash(Tatsuo)
10729 Adjust handling of data type names to suppress double quotes(Thomas)
10730 Use type coercion for matching columns and DEFAULT(Thomas)
10731 Fix deadlock so it only checks once after one second of sleep(Bruce)
10732 Fixes for aggregates and PL/pgsql(Hiroshi)
10733 Fix for subquery crash(Vadim)
10734 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
10735 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo)
10736 Fix for pg_dump -d or -D and quote special characters in INSERT
10737 Repair serious problems with dynahash(Tom)
10738 Fix INET/CIDR portability problems
10739 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
10740 Fix executor so mergejoin of different column types works(Tom)
10741 Fix for Alpha OR selectivity bug
10742 Fix OR index selectivity problem(Bruce)
10743 Fix so \d shows proper length for char()/varchar()(Ryan)
10744 Fix tutorial code(Clark)
10745 Improve destroyuser checking(Oliver)
10746 Fix for Kerberos(Rodney McDuff)
10747 Fix for dropping database while dirty buffers(Bruce)
10748 Fix so sequence nextval() can be case-sensitive(Bruce)
10750 Drop buffers before destroying database files(Bruce)
10751 Fix case where executor evaluates functions twice(Tatsuo)
10752 Allow sequence nextval actions to be case-sensitive(Bruce)
10753 Fix optimizer indexing not working for negative numbers(Bruce)
10754 Fix for memory leak in executor with fjIsNull
10755 Fix for aggregate memory leaks(Erik Riedel)
10756 Allow user name containing a dash to grant privileges
10757 Cleanup of NULL in inet types
10758 Clean up system table bugs(Tom)
10759 Fix problems of PAGER and \? command(Masaaki Sakaida)
10760 Reduce default multisegment file size limit to 1GB(Peter)
10761 Fix for dumping of CREATE OPERATOR(Tom)
10762 Fix for backward scanning of cursors(Hiroshi Inoue)
10763 Fix for COPY FROM STDIN when using \i(Tom)
10764 Fix for subselect is compared inside an expression(Jan)
10765 Fix handling of error reporting while returning rows(Tom)
10766 Fix problems with reference to array types(Tom,Jan)
10767 Prevent UPDATE SET oid(Jan)
10768 Fix pg_dump so -t option can handle case-sensitive tablenames
10769 Fixes for GROUP BY in special cases(Tom, Jan)
10770 Fix for memory leak in failed queries(Tom)
10771 DEFAULT now supports mixed-case identifiers(Tom)
10772 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
10773 Disable use of pg_dump with both -o and -d options(Bruce)
10774 Allow pg_dump to properly dump group privileges(Bruce)
10775 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
10776 Fix for computations in views(Jan)
10777 Fix for aggregates on array indexes(Tom)
10778 Fix for DEFAULT handles single quotes in value requiring too many quotes
10779 Fix security problem with non-super users importing/exporting large objects(Tom)
10780 Rollback of transaction that creates table cleaned up properly(Tom)
10781 Fix to allow long table and column names to generate proper serial names(Tom)
10785 Add "vacuumdb" utility
10786 Speed up libpq by allocating memory better(Tom)
10787 EXPLAIN all indexes used(Tom)
10788 Implement CASE, COALESCE, NULLIF expression(Thomas)
10789 New pg_dump table output format(Constantin)
10790 Add string min()/max() functions(Thomas)
10791 Extend new type coercion techniques to aggregates(Thomas)
10792 New moddatetime contrib(Terry)
10793 Update to pgaccess 0.96(Constantin)
10794 Add routines for single-byte "char" type(Thomas)
10795 Improved substr() function(Thomas)
10796 Improved multibyte handling(Tatsuo)
10797 Multiversion concurrency control/MVCC(Vadim)
10798 New Serialized mode(Vadim)
10799 Fix for tables over 2gigs(Peter)
10800 New SET TRANSACTION ISOLATION LEVEL(Vadim)
10801 New LOCK TABLE IN ... MODE(Vadim)
10802 Update ODBC driver(Byron)
10803 New NUMERIC data type(Jan)
10804 New SELECT FOR UPDATE(Vadim)
10805 Handle "NaN" and "Infinity" for input values(Jan)
10806 Improved date/year handling(Thomas)
10807 Improved handling of backend connections(Magnus)
10808 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
10809 New TCL_ARRAYS option(Massimo)
10810 New INTERSECT and EXCEPT(Stefan)
10811 New pg_index.indisprimary for primary key tracking(D'Arcy)
10812 New pg_dump option to allow dropping of tables before creation(Brook)
10813 Speedup of row output routines(Tom)
10814 New READ COMMITTED isolation level(Vadim)
10815 New TEMP tables/indexes(Bruce)
10816 Prevent sorting if result is already sorted(Jan)
10817 New memory allocation optimization(Jan)
10818 Allow psql to do \p\g(Bruce)
10819 Allow multiple rule actions(Jan)
10820 Added LIMIT/OFFSET functionality(Jan)
10821 Improve optimizer when joining a large number of tables(Bruce)
10822 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
10823 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
10824 Improved int8 support(Ryan Bradetich, Thomas, Tom)
10825 New routines to convert between int8 and text/varchar types(Thomas)
10826 New bushy plans, where meta-tables are joined(Bruce)
10827 Enable right-hand queries by default(Bruce)
10828 Allow reliable maximum number of backends to be set at configure time
10829 (--with-maxbackends and postmaster switch (-N backends))(Tom)
10830 GEQO default now 10 tables because of optimizer speedups(Tom)
10831 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
10832 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
10833 Allow psql \d on a view show query(Ryan)
10834 Speedup for LIKE(Bruce)
10835 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
10836 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
10837 Make % operator have precedence like /(Bruce)
10838 Add new postgres -O option to allow system table structure changes(Bruce)
10839 Update contrib/pginterface/findoidjoins script(Tom)
10840 Major speedup in vacuum of deleted rows with indexes(Vadim)
10841 Allow non-SQL functions to run different versions based on arguments(Tom)
10842 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida)
10843 Add version number in start-up banners for psql(Masaaki Sakaida)
10844 New contrib/vacuumlo removes large objects not referenced(Peter)
10845 New initialization for table sizes so non-vacuumed tables perform better(Tom)
10846 Improve error messages when a connection is rejected(Tom)
10847 Support for arrays of char() and varchar() fields(Massimo)
10848 Overhaul of hash code to increase reliability and performance(Tom)
10849 Update to PyGreSQL 2.4(D'Arcy)
10850 Changed debug options so -d4 and -d5 produce different node displays(Jan)
10851 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
10852 Better optimization statistics for system table access(Tom)
10853 Better handling of non-default block sizes(Massimo)
10854 Improve GEQO optimizer memory consumption(Tom)
10855 UNION now suppports ORDER BY of columns not in target list(Jan)
10856 Major libpq++ improvements(Vince Vielhaber)
10857 pg_dump now uses -z(ACL's) as default(Bruce)
10858 backend cache, memory speedups(Tom)
10859 have pg_dump do everything in one snapshot transaction(Vadim)
10860 fix for large object memory leakage, fix for pg_dumping(Tom)
10861 INET type now respects netmask for comparisons
10862 Make VACUUM ANALYZE only use a readlock(Vadim)
10863 Allow VIEWs on UNIONS(Jan)
10864 pg_dump now can generate consistent snapshots on active databases(Vadim)
10866 Source Tree Changes
10867 -------------------
10868 Improve port matching(Tom)
10869 Portability fixes for SunOS
10870 Add Windows NT backend port and enable dynamic loading(Magnus and Daniel Horak)
10871 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
10872 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
10873 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
10874 Port to NetBSD/macppc(Toshimi Aoki)
10875 Fix for tcl/tk configuration(Vince)
10876 Removed CURRENT key word for rule queries(Jan)
10877 NT dynamic loading now works(Daniel Horak)
10878 Add ARM32 support(Andrew McMurry)
10879 Better support for HP-UX 11 and UnixWare
10880 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
10881 New install commands for plpgsql(Jan)
10888 <sect1 id="release-6-4-2">
10889 <title>Release 6.4.2</title>
10892 <title>Release date</title>
10893 <simpara>1998-12-20</simpara>
10897 The 6.4.1 release was improperly packaged. This also has one additional
10903 <title>Migration to version 6.4.2</title>
10906 A dump/restore is <emphasis>not</emphasis> required for those running
10911 <title>Changes</title>
10915 Fix for datetime constant problem on some platforms(Thomas)
10923 <sect1 id="release-6-4-1">
10924 <title>Release 6.4.1</title>
10927 <title>Release date</title>
10928 <simpara>1998-12-18</simpara>
10932 This is basically a cleanup release for 6.4. We have fixed a variety of
10933 problems reported by 6.4 users.
10938 <title>Migration to version 6.4.1</title>
10941 A dump/restore is <emphasis>not</emphasis> required for those running
10946 <title>Changes</title>
10950 Add pg_dump -N flag to force double quotes around identifiers. This is
10951 the default(Thomas)
10952 Fix for NOT in where clause causing crash(Bruce)
10953 EXPLAIN VERBOSE coredump fix(Vadim)
10954 Fix shared-library problems on Linux
10955 Fix test for table existence to allow mixed-case and whitespace in
10956 the table name(Thomas)
10957 Fix a couple of pg_dump bugs
10958 Configure matches template/.similar entries better(Tom)
10959 Change builtin function names from SPI_* to spi_*
10960 OR WHERE clause fix(Vadim)
10961 Fixes for mixed-case table names(Billy)
10962 contrib/linux/postgres.init.csh/sh fix(Thomas)
10963 libpq memory overrun fix
10965 Change exp() behavior to generate error on underflow(Thomas)
10966 pg_dump fixes for memory leak, inheritance constraints, layout change
10967 update pgaccess to 0.93
10968 Fix prototype for 64-bit platforms
10969 Multibyte fixes(Tatsuo)
10971 Fix memory overruns(Tatsuo)
10972 Fix for lo_import() crash(Bruce)
10973 Better search for install program(Tom)
10974 Timezone fixes(Tom)
10976 Use implicit type coercion for matching DEFAULT values(Thomas)
10977 Add routines to help with single-byte (internal) character type(Thomas)
10978 Compilation of libpq for Windows fixes(Magnus)
10979 Upgrade to PyGreSQL 2.2(D'Arcy)
10987 <sect1 id="release-6-4">
10988 <title>Release 6.4</title>
10991 <title>Release date</title>
10992 <simpara>1998-10-30</simpara>
10996 There are <emphasis>many</emphasis> new features and improvements in this release.
10997 Thanks to our developers and maintainers, nearly every aspect of the system
10998 has received some attention since the previous release.
10999 Here is a brief, incomplete summary:
11004 Views and rules are now functional thanks to extensive new code in the
11005 rewrite rules system from Jan Wieck. He also wrote a chapter on it
11006 for the <citetitle>Programmer's Guide</citetitle>.
11011 Jan also contributed a second procedural language, <application>PL/pgSQL</application>, to go with the
11012 original <application>PL/pgTCL</application> procedural language he contributed last release.
11018 We have optional multiple-byte character set support from Tatsuo Ishii
11019 to complement our existing locale support.
11025 Client/server communications has been cleaned up, with better support for
11026 asynchronous messages and interrupts thanks to Tom Lane.
11032 The parser will now perform automatic type coercion to match arguments
11033 to available operators and functions, and to match columns and expressions
11034 with target columns. This uses a generic mechanism which supports
11035 the type extensibility features of <productname>PostgreSQL</productname>.
11036 There is a new chapter in the <citetitle>User's Guide</citetitle>
11037 which covers this topic.
11043 Three new data types have been added.
11044 Two types, <type>inet</type> and <type>cidr</type>, support various forms
11045 of IP network, subnet, and machine addressing. There is now an 8-byte integer
11046 type available on some platforms. See the chapter on data types
11047 in the <citetitle>User's Guide</citetitle> for details.
11048 A fourth type, <type>serial</type>, is now supported by the parser as an
11049 amalgam of the <type>int4</type> type, a sequence, and a unique index.
11055 Several more <acronym>SQL92</acronym>-compatible syntax features have been
11056 added, including <command>INSERT DEFAULT VALUES</command>
11062 The automatic configuration and installation system has received some
11063 attention, and should be more robust for more platforms than it has ever
11072 <title>Migration to version 6.4</title>
11075 A dump/restore using <application>pg_dump</application>
11076 or <application>pg_dumpall</application>
11077 is required for those wishing to migrate data from any
11078 previous release of <productname>PostgreSQL</productname>.
11083 <title>Changes</title>
11089 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
11090 Remove char2-16 data types, use char/varchar(Darren)
11091 Pqfn not handles a NOTICE message(Anders)
11092 Reduced busywaiting overhead for spinlocks with many backends (dg)
11093 Stuck spinlock detection (dg)
11094 Fix up "ISO-style" timespan decoding and encoding(Thomas)
11095 Fix problem with table drop after rollback of transaction(Vadim)
11096 Change error message and remove non-functional update message(Vadim)
11097 Fix for COPY array checking
11098 Fix for SELECT 1 UNION SELECT NULL
11099 Fix for buffer leaks in large object calls(Pascal)
11100 Change owner from oid to int4 type(Bruce)
11101 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
11102 Fix for shared invalidation cache overflow(Massimo)
11103 Prevent file descriptor leaks in failed COPY's(Bruce)
11104 Fix memory leak in libpgtcl's pg_select(Constantin)
11105 Fix problems with username/passwords over 8 characters(Tom)
11106 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
11107 Fix of many bad system table entries(Tom)
11111 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
11112 Show the index used in an EXPLAIN(Zeugswetter)
11113 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
11114 Multibyte awareness of many data types and functions, via configure(Tatsuo)
11115 New configure --with-mb option(Tatsuo)
11116 New initdb --pgencoding option(Tatsuo)
11117 New createdb -E multibyte option(Tatsuo)
11118 Select version(); now returns PostgreSQL version(Jeroen)
11119 libpq now allows asynchronous clients(Tom)
11120 Allow cancel from client of backend query(Tom)
11121 psql now cancels query with Control-C(Tom)
11122 libpq users need not issue dummy queries to get NOTIFY messages(Tom)
11123 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
11124 PGresult struct now includes associated error message, if any(Tom)
11125 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
11126 Add routines to convert between varchar and bpchar(Thomas)
11127 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
11128 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
11129 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas)
11130 Fixes for unary minus parsing with leading spaces(Thomas)
11131 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
11132 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
11133 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
11134 Enable HAVING clause but no fixes elsewhere yet.
11135 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thomas)
11136 Save string type if specified for DEFAULT clause handling(Thomas)
11137 Coerce operations involving different data types(Thomas)
11138 Allow some index use for columns of different types(Thomas)
11139 Add capabilities for automatic type conversion(Thomas)
11140 Cleanups for large objects, so file is truncated on open(Peter)
11141 Readline cleanups(Tom)
11142 Allow psql \f \ to make spaces as delimiter(Bruce)
11143 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
11144 Msql compatibility library in /contrib(Aldrin)
11145 Remove the requirement that ORDER/GROUP BY clause identifiers be
11146 included in the target list(David)
11147 Convert columns to match columns in UNION clauses(Thomas)
11148 Remove fork()/exec() and only do fork()(Bruce)
11149 Jdbc cleanups(Peter)
11150 Show backend status on ps command line(only works on some platforms)(Bruce)
11151 Pg_hba.conf now has a sameuser option in the database field
11152 Make lo_unlink take oid param, not int4
11153 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
11154 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
11155 libpgtcl cleanups(Tom)
11156 Add -error option to libpgtcl's pg_result command(Tom)
11157 New locale patch, see docs/README/locale(Oleg)
11158 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
11159 New contrib/lo code for large object orphan removal(Peter)
11160 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
11161 feature, see /doc/README.mb(Tatsuo)
11162 contrib/noupdate code to revoke update permission on a column
11163 libpq can now be compiled on Windows(Magnus)
11164 Add PQsetdbLogin() in libpq
11165 New 8-byte integer type, checked by configure for OS support(Thomas)
11166 Better support for quoted table/column names(Thomas)
11167 Surround table and column names with double-quotes in pg_dump(Thomas)
11168 PQreset() now works with passwords(Tom)
11169 Handle case of GROUP BY target list column number out of range(David)
11170 Allow UNION in subselects
11171 Add auto-size to screen to \d? commands(Bruce)
11172 Use UNION to show all \d? results in one query(Bruce)
11173 Add \d? field search feature(Bruce)
11174 Pg_dump issues fewer \connect requests(Tom)
11175 Make pg_dump -z flag work better, document it in manual page(Tom)
11176 Add HAVING clause with full support for subselects and unions(Stephan)
11177 Full text indexing routines in contrib/fulltextindex(Maarten)
11178 Transaction ids now stored in shared memory(Vadim)
11179 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
11180 Support for SQL92 syntax "SET NAMES"(Tatsuo)
11181 Support for LATIN2-5(Tatsuo)
11182 Add UNICODE regression test case(Tatsuo)
11183 Lock manager cleanup, new locking modes for LLL(Vadim)
11184 Allow index use with OR clauses(Bruce)
11185 Allows "SELECT NULL ORDER BY 1;"
11186 Explain VERBOSE prints the plan, and now pretty-prints the plan to
11187 the postmaster log file(Bruce)
11188 Add indexes display to \d command(Bruce)
11189 Allow GROUP BY on functions(David)
11190 New pg_class.relkind for large objects(Bruce)
11191 New way to send libpq NOTICE messages to a different location(Tom)
11192 New \w write command to psql(Bruce)
11193 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
11194 Allow binary-compatible indexes to be considered when checking for valid
11195 Indexes for restriction clauses containing a constant(Thomas)
11196 New ISBN/ISSN code in /contrib/isbn_issn
11197 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
11198 New rewrite system fixes many problems with rules and views(Jan)
11199 * Rules on relations work
11200 * Event qualifications on insert/update/delete work
11201 * New OLD variable to reference CURRENT, CURRENT will be remove in future
11202 * Update rules can reference NEW and OLD in rule qualifications/actions
11203 * Insert/update/delete rules on views work
11204 * Multiple rule actions are now supported, surrounded by parentheses
11205 * Regular users can create views/rules on tables they have RULE permits
11206 * Rules and views inherit the privileges of the creator
11207 * No rules at the column level
11208 * No UPDATE NEW/OLD rules
11209 * New pg_tables, pg_indexes, pg_rules and pg_views system views
11210 * Only a single action on SELECT rules
11211 * Total rewrite overhaul, perhaps for 6.5
11212 * handle subselects
11213 * handle aggregates on views
11214 * handle insert into select from view works
11215 System indexes are now multikey(Bruce)
11216 Oidint2, oidint4, and oidname types are removed(Bruce)
11217 Use system cache for more system table lookups(Bruce)
11218 New backend programming language PL/pgSQL in backend/pl(Jan)
11219 New SERIAL data type, auto-creates sequence/index(Thomas)
11220 Enable assert checking without a recompile(Massimo)
11221 User lock enhancements(Massimo)
11222 New setval() command to set sequence value(Massimo)
11223 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
11224 Conditional trace package(Massimo)
11225 New UNLISTEN command(Massimo)
11226 psql and libpq now compile under Windows using win32.mak(Magnus)
11227 Lo_read no longer stores trailing NULL(Bruce)
11228 Identifiers are now truncated to 31 characters internally(Bruce)
11229 Createuser options now availble on the command line
11230 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
11231 Prevent file descriptor leaf from failed COPY(Bruce)
11232 New pg_upgrade command(Bruce)
11233 Updated /contrib directories(Massimo)
11234 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
11235 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
11236 New DECLARE and FETCH feature(Thomas)
11237 libpq's internal structures now not exported(Tom)
11238 Allow up to 8 key indexes(Bruce)
11239 Remove ARCHIVE key word, that is no longer used(Thomas)
11240 pg_dump -n flag to supress quotes around indentifiers
11241 disable system columns for views(Jan)
11242 new INET and CIDR types for network addresses(TomH, Paul)
11243 no more double quotes in psql output
11244 pg_dump now dumps views(Terry)
11245 new SET QUERY_LIMIT(Tatsuo,Jan)
11247 Source Tree Changes
11248 -------------------
11249 /contrib cleanup(Jun)
11250 Inline some small functions called for every row(Bruce)
11252 HP-UX cleanups(Tom)
11253 Multibyte regression tests(Soonmyung.)
11254 Remove --disabled options from configure
11255 Define PGDOC to use POSTGRESDIR by default
11256 Make regression optional
11257 Remove extra braces code to pgindent(Bruce)
11258 Add bsdi shared library support(Bruce)
11259 New --without-CXX support configure option(Brook)
11261 Update backend flowchart in tools/backend(Bruce)
11262 Change atttypmod from int16 to int32(Bruce, Tom)
11263 Getrusage() fix for platforms that do not have it(Tom)
11264 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
11265 NS32K platform fixes(Phil Nelson, John Buller)
11266 SCO 7/UnixWare 2.x fixes(Billy,others)
11267 Sparc/Solaris 2.5 fixes(Ryan)
11268 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
11269 Even more documention(Thomas)
11270 Nextstep support(Jacek)
11272 pginterface manual page(Bruce)
11273 shared libraries all have version numbers
11274 merged all OS-specific shared library defines into one file
11275 smarter TCL/TK configuration checking(Billy)
11276 smarter perl configuration(Brook)
11277 configure uses supplied install-sh if no install script found(Tom)
11278 new Makefile.shlib for shared library configuration(Tom)
11284 <sect1 id="release-6-3-2">
11285 <title>Release 6.3.2</title>
11288 <title>Release date</title>
11289 <simpara>1998-04-07</simpara>
11293 This is a bug-fix release for 6.3.x.
11294 Refer to the release notes for version 6.3 for a more complete summary of new features.
11302 Repairs automatic configuration support for some platforms, including Linux,
11303 from breakage inadvertently introduced in version 6.3.1.
11309 Correctly handles function calls on the left side of BETWEEN and LIKE clauses.
11316 A dump/restore is NOT required for those running 6.3 or 6.3.1. A
11317 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11318 This last step should be performed while the postmaster is not running.
11319 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11322 For upgrades from pre-6.3 installations,
11323 refer to the installation and migration instructions for version 6.3.
11327 <title>Changes</title>
11331 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
11332 Manual page improvements(Bruce)
11333 BETWEEN and LIKE fix(Thomas)
11334 fix for psql \connect used by pg_dump(Oliver Elphick)
11336 pgaccess, version 0.86
11337 qsort removed, now uses libc version, cleanups(Jeroen)
11338 fix for buffer over-runs detected(Maurice Gittens)
11339 fix for buffer overrun in libpgtcl(Randy Kunkee)
11340 fix for UNION with DISTINCT or ORDER BY(Bruce)
11341 gettimeofday configure check(Doug Winterburn)
11342 Fix "indexes not used" bug(Vadim)
11343 docs additions(Thomas)
11344 Fix for backend memory leak(Bruce)
11345 libreadline cleanup(Erwan MAS)
11346 Remove DISTDIR(Bruce)
11347 Makefile dependency cleanup(Jeroen van Vianen)
11348 ASSERT fixes(Bruce)
11354 <sect1 id="release-6-3-1">
11355 <title>Release 6.3.1</title>
11358 <title>Release date</title>
11359 <simpara>1998-03-23</simpara>
11368 Additional support for multibyte character sets.
11374 Repair byte ordering for mixed-endian clients and servers.
11380 Minor updates to allowed SQL syntax.
11386 Improvements to the configuration autodetection for installation.
11393 A dump/restore is NOT required for those running 6.3. A
11394 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11395 This last step should be performed while the postmaster is not running.
11396 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11399 For upgrades from pre-6.3 installations,
11400 refer to the installation and migration instructions for version 6.3.
11404 <title>Changes</title>
11408 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
11409 pg_user cleanup(Bruce)
11410 large object fix for pg_dump and tclsh (alvin)
11411 LIKE fix for multiple adjacent underscores
11412 fix for redefining builtin functions(Thomas)
11414 upgrade to pg_access 0.83
11415 updated CLUSTER manual page
11416 multibyte character set support, see doc/README.mb(Tatsuo)
11417 configure --with-pgport fix
11419 big-endian fix for backend communications(Kataoka)
11420 SUBSTR() and substring() fix(Jan)
11421 several jdbc fixes(Peter)
11422 libpgtcl improvements, see libptcl/README(Randy Kunkee)
11423 Fix for "Datasize = 0" error(Vadim)
11424 Prevent \do from wrapping(Bruce)
11425 Remove duplicate Russian character set entries
11427 Allow optional TABLE key word in LOCK and SELECT INTO(Thomas)
11428 CREATE SEQUENCE options to allow a negative integer(Thomas)
11429 Add "PASSWORD" as an allowed column identifier(Thomas)
11430 Add checks for UNION target fields(Bruce)
11431 Fix Alpha port(Dwayne Bailey)
11432 Fix for text arrays containing quotes(Doug Gibson)
11433 Solaris compile fix(Albert Chin-A-Young)
11434 Better identify tcl and tk libs and includes(Bruce)
11440 <sect1 id="release-6-3">
11441 <title>Release 6.3</title>
11444 <title>Release date</title>
11445 <simpara>1998-03-01</simpara>
11449 There are <emphasis>many</emphasis> new features and improvements in this release.
11450 Here is a brief, incomplete summary:
11455 Many new SQL features, including
11456 full <acronym>SQL92</acronym> subselect capability
11457 (everything is here but target-list subselects).
11463 Support for client-side environment variables to specify time zone and date style.
11469 Socket interface for client/server connection. This is the default now
11470 so you may need to start <application>postmaster</application> with the
11471 <option>-i</option> flag.
11477 Better password authorization mechanisms. Default table privileges have changed.
11483 Old-style <firstterm>time travel</firstterm>
11484 has been removed. Performance has been improved.
11492 Bruce Momjian wrote the following notes to introduce the new release.
11497 There are some general 6.3 issues that I want to mention. These are
11498 only the big items that cannot be described in one sentence. A review
11499 of the detailed changes list is still needed.
11502 First, we now have subselects. Now that we have them, I would like to
11503 mention that without subselects, SQL is a very limited language.
11504 Subselects are a major feature, and you should review your code for
11505 places where subselects provide a better solution for your queries. I
11506 think you will find that there are more uses for subselects than you may
11507 think. Vadim has put us on the big SQL map with subselects, and fully
11508 functional ones too. The only thing you can't do with subselects is to
11509 use them in the target list.
11512 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
11513 enable connections from other machines, you have to use the new
11514 postmaster -i option, and of course edit <filename>pg_hba.conf</filename>. Also, for this
11515 reason, the format of <filename>pg_hba.conf</filename> has changed.
11518 Third, <type>char()</type> fields will now allow faster access than <type>varchar()</type> or
11519 <type>text</type>. Specifically, the <type>text</> and <type>varchar()</type> have a penalty for access to
11520 any columns after the first column of this type. <type>char()</type> used to also
11521 have this access penalty, but it no longer does. This may suggest that
11522 you redesign some of your tables, especially if you have short character
11523 columns that you have defined as <type>varchar()</type> or <type>text</type>. This and other
11524 changes make 6.3 even faster than earlier releases.
11527 We now have passwords definable independent of any Unix file. There are
11528 new SQL USER commands.
11529 See the <citetitle>Administrator's Guide</citetitle> for more
11530 information. There is a new table, pg_shadow, which is used to store
11531 user information and user passwords, and it by default only SELECT-able
11532 by the <systemitem>postgres</systemitem> super-user. pg_user is now a view of pg_shadow, and is
11533 SELECT-able by PUBLIC. You should keep using pg_user in your
11534 application without changes.
11537 User-created tables now no longer have SELECT privilege to PUBLIC by
11538 default. This was done because the ANSI standard requires it. You can
11539 of course GRANT any privileges you want after the table is created.
11540 System tables continue to be SELECT-able by PUBLIC.
11543 We also have real deadlock detection code. No more sixty-second
11544 timeouts. And the new locking code implements a <acronym>FIFO</acronym> better, so there
11545 should be less resource starvation during heavy use.
11548 Many complaints have been made about inadequate documentation in previous
11549 releases. Thomas has put much effort into many new manuals for this
11550 release. Check out the doc/ directory.
11553 For performance reasons, time travel is gone, but can be implemented
11554 using triggers (see <filename>pgsql/contrib/spi/README</filename>). Please check out the new
11555 \d command for types, operators, etc. Also, views have their own
11556 privileges now, not based on the underlying tables, so privileges on
11557 them have to be set separately. Check <filename>/pgsql/interfaces</filename> for some new
11558 ways to talk to <productname>PostgreSQL</productname>.
11561 This is the first release that really required an explanation for
11562 existing users. In many ways, this was necessary because the new
11563 release removes many limitations, and the work-arounds people were using
11564 are no longer needed.
11568 <title>Migration to version 6.3</title>
11571 A dump/restore using <application>pg_dump</application>
11572 or <application>pg_dumpall</application>
11573 is required for those wishing to migrate data from any
11574 previous release of <productname>PostgreSQL</productname>.
11579 <title>Changes</title>
11585 Fix binary cursors broken by MOVE implementation(Vadim)
11586 Fix for tcl library crash(Jan)
11587 Fix for array handling, from Gerhard Hintermayer
11588 Fix acl error, and remove duplicate pqtrace(Bruce)
11589 Fix psql \e for empty file(Bruce)
11590 Fix for textcat on varchar() fields(Bruce)
11591 Fix for DBT Sendproc (Zeugswetter Andres)
11592 Fix vacuum analyze syntax problem(Bruce)
11593 Fix for international identifiers(Tatsuo)
11594 Fix aggregates on inherited tables(Bruce)
11595 Fix substr() for out-of-bounds data
11596 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
11597 Fix notty output to show status result. -q option still turns it off(Bruce)
11598 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
11600 Fix for PQtrace start/stop several times(Bruce)
11601 Fix a variety of locking problems like newer lock waiters getting
11602 lock before older waiters, and having readlock people not share
11603 locks if a writer is waiting for a lock, and waiting writers not
11604 getting priority over waiting readers(Bruce)
11605 Fix crashes in psql when executing queries from external files(James)
11606 Fix problem with multiple order by columns, with the first one having
11607 NULL values(Jeroen)
11608 Use correct hash table support functions for float8 and int4(Thomas)
11609 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
11610 Change precedence for boolean operators to match expected behavior(Thomas)
11611 Generate elog(ERROR) on over-large integer(Bruce)
11612 Allow multiple-argument functions in constraint clauses(Thomas)
11613 Check boolean input literals for 'true','false','yes','no','1','0'
11614 and throw elog(ERROR) if unrecognized(Thomas)
11615 Major large objects fix
11616 Fix for GROUP BY showing duplicates(Vadim)
11617 Fix for index scans in MergeJion(Vadim)
11621 Subselects with EXISTS, IN, ALL, ANY key words (Vadim, Bruce, Thomas)
11622 New User Manual(Thomas, others)
11623 Speedup by inlining some frequently-called functions
11624 Real deadlock detection, no more timeouts(Bruce)
11625 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
11626 CURRENT_USER(Thomas)
11627 Modify constraint syntax to be SQL92-compliant(Thomas)
11628 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
11629 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
11630 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas)
11631 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
11632 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
11633 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
11634 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
11635 Allow SQL92 delimited identifiers(Thomas)
11636 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas)
11637 Support SQL92 syntax for type coercion of literal strings
11638 (e.g. "DATETIME 'now'")(Thomas)
11639 Add conversions for int2, int4, and OID types to and from text(Thomas)
11640 Use shared lock when building indexes(Vadim)
11641 Free memory allocated for an user query inside transaction block after
11642 this query is done, was turned off in <= 6.2.1(Vadim)
11643 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
11644 New <productname>PostgreSQL</productname> Procedural Language (PL) backend interface(Jan)
11645 Rename pg_dump -H option to -h(Bruce)
11646 Add Java support for passwords, European dates(Peter)
11647 Use indexes for LIKE and ~, !~ operations(Bruce)
11648 Add hash functions for datetime and timespan(Thomas)
11649 Time Travel removed(Vadim, Bruce)
11650 Add paging for \d and \z, and fix \i(Bruce)
11651 Add Unix domain socket support to backend and to frontend library(Goran)
11652 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
11653 Allow more SQL92 and/or <productname>PostgreSQL</productname> reserved words as column identifiers(Thomas)
11654 Augment support for SQL92 SET TIME ZONE...(Thomas)
11655 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
11656 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
11657 Enable SET TIME ZONE using TZ environment variable(Thomas)
11658 Add PGDATESTYLE environment variable to frontend and backend initialization(Thomas)
11659 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
11660 frontend library initialization environment variables(Thomas)
11661 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
11662 Add pg_description table for info on tables, columns, operators, types, and
11664 Increase 16 char limit on system table/index names to 32 characters(Bruce)
11665 Rename system indexes(Bruce)
11666 Add 'GERMAN' option to SET DATESTYLE(Thomas)
11667 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
11668 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
11669 Validate numeric input more carefully for delta times(Thomas)
11670 Implement day of year as possible input to date_part()(Thomas)
11671 Define timespan_finite() and text_timespan() functions(Thomas)
11672 Remove archive stuff(Bruce)
11673 Allow for a pg_password authentication database that is separate from
11674 the system password file(Todd)
11675 Dump ACLs, GRANT, REVOKE privileges(Matt)
11676 Define text, varchar, and bpchar string length functions(Thomas)
11677 Fix Query handling for inheritance, and cost computations(Bruce)
11678 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
11679 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
11680 Implement UNIONs for SELECT(Bruce)
11681 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
11682 varchar() stores only necessary bytes on disk(Bruce)
11683 Fix for BLOBs(Peter)
11684 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
11685 Remove unused "option" from PQconnectdb()
11686 New LOCK command and lock manual page describing deadlocks(Bruce)
11687 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
11688 Enhance psql \z to show sequences(Bruce)
11689 Show NOT NULL and DEFAULT in psql \d table(Bruce)
11690 New psql .psqlrc file start-up(Andrew)
11691 Modify sample start-up script in contrib/linux to show syslog(Thomas)
11692 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
11693 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
11694 Update of contrib stuff(Massimo)
11695 Add Unix socket support to DBD::Pg(Goran)
11696 New python interface (PyGreSQL 2.0)(D'Arcy)
11697 New frontend/backend protocol has a version number, network byte order(Phil)
11698 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
11699 CHAR() now faster access than VARCHAR() or TEXT
11700 ecpg embedded SQL preprocessor
11701 Reduce system column overhead(Vadmin)
11702 Remove pg_time table(Vadim)
11703 Add pg_type attribute to identify types that need length (bpchar, varchar)
11704 Add report of offending line when COPY command fails
11705 Allow VIEW privileges to be set separately from the underlying tables.
11706 For security, use GRANT/REVOKE on views as appropriate(Jan)
11707 Tables now have no default GRANT SELECT TO PUBLIC. You must
11708 explicitly grant such privileges.
11709 Clean up tutorial examples(Darren)
11711 Source Tree Changes
11712 -------------------
11713 Add new html development tools, and flow chart in /tools/backend
11714 Fix for SCO compiles
11715 Stratus computer port Robert Gillies
11716 Added support for shlib for BSD44_derived & i386_solaris
11717 Make configure more automated(Brook)
11718 Add script to check regression test results
11719 Break parser functions into smaller files, group together(Bruce)
11720 Rename heap_create to heap_create_and_catalog, rename heap_creatr
11721 to heap_create()(Bruce)
11722 Sparc/Linux patch for locking(TomS)
11723 Remove PORTNAME and reorganize port-specific stuff(Marc)
11724 Add optimizer README file(Bruce)
11725 Remove some recursion in optimizer and clean up some code there(Bruce)
11726 Fix for NetBSD locking(Henry)
11727 Fix for libptcl make(Tatsuo)
11729 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
11730 function calls to istrue() or isfalse() to allow optimization(Thomas)
11731 Various fixes NetBSD/Sparc related(TomH)
11732 Alpha linux locking(Travis,Ryan)
11733 Change elog(WARN) to elog(ERROR)(Bruce)
11734 FAQ for FreeBSD(Marc)
11735 Bring in the PostODBC source tree as part of our standard distribution(Marc)
11736 A minor patch for HP/UX 10 vs 9(Stan)
11737 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
11738 UnixWare patches(Billy)
11739 New i386 'lock' for spinlock asm(Billy)
11740 Support for multiplexed backends is removed
11741 Start an OpenBSD port
11743 Start a Cygnus port
11744 Add string functions to regression suite(Thomas)
11745 Expand a few function names formerly truncated to 16 characters(Thomas)
11746 Remove un-needed malloc() calls and replace with palloc()(Bruce)
11752 <sect1 id="release-6-2-1">
11753 <title>Release 6.2.1</title>
11756 <title>Release date</title>
11757 <simpara>1997-10-17</simpara>
11761 6.2.1 is a bug-fix and usability release on 6.2.
11769 Allow strings to span lines, per <acronym>SQL92</acronym>.
11775 Include example trigger function for inserting user names on table updates.
11782 This is a minor bug-fix release on 6.2.
11783 For upgrades from pre-6.2 systems, a full dump/reload is required.
11784 Refer to the 6.2 release notes for instructions.
11788 <title>Migration from version 6.2 to version 6.2.1</title>
11791 This is a minor bug-fix release. A dump/reload is not required from version 6.2,
11792 but is required from any release prior to 6.2.
11795 In upgrading from version 6.2, if you choose to dump/reload you will find that
11796 avg(money) is now calculated correctly. All other bug fixes take effect
11797 upon updating the executables.
11800 Another way to avoid dump/reload is to use the following SQL command
11801 from <command>psql</command> to update the existing system table:
11804 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
11805 where aggname = 'avg' and aggbasetype = 790;
11809 This will need to be done to every existing database, including template1.
11814 <title>Changes</title>
11818 Allow TIME and TYPE column names(Thomas)
11819 Allow larger range of true/false as boolean values(Thomas)
11820 Support output of "now" and "current"(Thomas)
11821 Handle DEFAULT with INSERT of NULL properly(Vadim)
11822 Fix for relation reference counts problem in buffer manager(Vadim)
11823 Allow strings to span lines, like ANSI(Thomas)
11824 Fix for backward cursor with ORDER BY(Vadim)
11825 Fix avg(cash) computation(Thomas)
11826 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
11827 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
11828 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
11834 <sect1 id="release-6-2">
11835 <title>Release 6.2</title>
11838 <title>Release date</title>
11839 <simpara>1997-10-02</simpara>
11843 A dump/restore is required for those wishing to migrate data from
11844 previous releases of <productname>PostgreSQL</productname>.
11848 <title>Migration from version 6.1 to version 6.2</title>
11851 This migration requires a complete dump of the 6.1 database and a
11852 restore of the database in 6.2.
11855 Note that the <command>pg_dump</command> and <command>pg_dumpall</command> utility from 6.2 should be used
11856 to dump the 6.1 database.
11861 <title>Migration from version 1.<replaceable>x</> to version 6.2</title>
11864 Those migrating from earlier 1.* releases should first upgrade to 1.09
11865 because the COPY output format was improved from the 1.02 release.
11870 <title>Changes</title>
11876 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
11877 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
11878 from Solaris(Diab Jerius)
11879 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
11880 Check for geometric intersections at endpoints to avoid rounding ugliness(Thomas)
11881 Catch non-functional delete attempts(Vadim)
11882 Change time function names to be more consistent(Michael Reifenberg)
11883 Check for zero divides(Michael Reifenberg)
11884 Fix very old bug which made rows changed/inserted by a command
11885 visible to the command itself (so we had multiple update of
11886 updated rows, etc.)(Vadim)
11887 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
11888 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
11889 Remove un-needed signal stuff from contrib/pginterface
11890 Fix OR (where x != 1 or x isnull didn't return rows with x NULL) (Vadim)
11891 Fix time_cmp function (Vadim)
11892 Fix handling of functions with non-attribute first argument in
11893 WHERE clauses (Vadim)
11894 Fix GROUP BY when order of entries is different from order
11895 in target list (Vadim)
11896 Fix pg_dump for aggregates without sfunc1 (Vadim)
11900 Default genetic optimizer GEQO parameter is now 8(Bruce)
11901 Allow use parameters in target list having aggregates in functions(Vadim)
11902 Added JDBC driver as an interface(Adrian & Peter)
11903 pg_password utility
11904 Return number of rows inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
11905 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
11906 SPI (Server Programming Interface) allows execution of queries inside
11907 C-functions (Vadim)
11908 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
11909 Include reserved words for string handling, outer joins, and unions(Thomas)
11910 Implement extended comments ("/* ... */") using exclusive states(Thomas)
11911 Add "//" single-line comments(Bruce)
11912 Remove some restrictions on characters in operator names(Thomas)
11913 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
11914 Add text concatenation operator and function (SQL92)(Thomas)
11915 Support WITH TIME ZONE syntax (SQL92)(Thomas)
11916 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
11917 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
11918 and CHARACTER VARYING (SQL92)(Thomas)
11919 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
11920 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
11921 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
11922 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
11923 Add more reserved words, mostly for SQL92 compliance(Thomas)
11924 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
11925 Add center() routines for lseg, path, polygon(Thomas)
11926 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
11927 Check explicitly for points and polygons contained within polygons
11928 using an axis-crossing algorithm(Thomas)
11929 Add routine to convert circle-box(Thomas)
11930 Merge conflicting operators for different geometric data types(Thomas)
11931 Replace distance operator "<===>" with "<->"(Thomas)
11932 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Thomas)
11933 Add routines for text trimming on both ends, substring, and string position(Thomas)
11934 Added conversion routines circle(box) and poly(circle)(Thomas)
11935 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
11936 Allow functions and operators on internally-identical types to succeed(Bruce)
11937 Speed up backend start-up after profiling analysis(Bruce)
11938 Inline frequently called functions for performance(Bruce)
11939 Reduce open() calls(Bruce)
11940 psql: Add PAGER for \h and \?,\C fix
11941 Fix for psql pager when no tty(Bruce)
11942 New entab utility(Bruce)
11943 General trigger functions for referential integrity (Vadim)
11944 General trigger functions for time travel (Vadim)
11945 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
11946 MOVE implementation (Vadim)
11948 Source Tree Changes
11949 -------------------
11950 HP-UX 10 patches (Vladimir Turin)
11951 Added SCO support, (Daniel Harris)
11952 MkLinux patches (Tatsuo Ishii)
11953 Change geometric box terminology from "length" to "width"(Thomas)
11954 Deprecate temporary unstored slope fields in geometric code(Thomas)
11955 Remove restart instructions from INSTALL(Bruce)
11956 Look in /usr/ucb first for install(Bruce)
11957 Fix c++ copy example code(Thomas)
11958 Add -o to psql manual page(Bruce)
11959 Prevent relname unallocated string length from being copied into database(Bruce)
11960 Cleanup for NAMEDATALEN use(Bruce)
11961 Fix pg_proc names over 15 chars in output(Bruce)
11962 Add strNcpy() function(Bruce)
11963 remove some (void) casts that are unnecessary(Bruce)
11964 new interfaces directory(Marc)
11965 Replace fopen() calls with calls to fd.c functions(Bruce)
11966 Make functions static where possible(Bruce)
11967 enclose unused functions in #ifdef NOT_USED(Bruce)
11968 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
11969 Changes for Digital Unix
11970 Portability fix for pg_dumpall(Bruce)
11971 Rename pg_attribute.attnvals to attdispersion(Bruce)
11972 "intro/unix" manual page now "pgintro"(Bruce)
11973 "built-in" manual page now "pgbuiltin"(Bruce)
11974 "drop" manual page now "drop_table"(Bruce)
11975 Add "create_trigger", "drop_trigger" manual pages(Thomas)
11976 Add constraints regression test(Vadim & Thomas)
11977 Add comments syntax regression test(Thomas)
11978 Add PGINDENT and support program(Bruce)
11979 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
11980 Files moved to /src/tools directory(Bruce)
11981 SPI and Trigger programming guides (Vadim & D'Arcy)
11987 <sect1 id="release-6-1-1">
11988 <title>Release 6.1.1</title>
11991 <title>Release date</title>
11992 <simpara>1997-07-22</simpara>
11996 <title>Migration from version 6.1 to version 6.1.1</title>
11999 This is a minor bug-fix release. A dump/reload is not required from version 6.1,
12000 but is required from any release prior to 6.1.
12001 Refer to the release notes for 6.1 for more details.
12006 <title>Changes</title>
12010 fix for SET with options (Thomas)
12011 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
12012 new psql \connect option allows changing usernames without changing databases
12013 fix for initdb --debug option(Yoshihiko Ichikawa))
12014 lextest cleanup(Bruce)
12016 fix date/time month boundary arithmetic(Thomas)
12017 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
12018 timestamp overhauled to use standard functions(Thomas)
12019 other code cleanup in date/time routines(Thomas)
12020 psql's \d now case-insensitive(Bruce)
12021 psql's backslash commands can now have trailing semicolon(Bruce)
12022 fix memory leak in psql when using \g(Bruce)
12023 major fix for endian handling of communication to server(Thomas, Tatsuo)
12024 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
12025 allow underscores in usernames(Bruce)
12026 pg_dumpall now returns proper status, portability fix(Bruce)
12032 <sect1 id="release-6-1">
12033 <title>Release 6.1</title>
12036 <title>Release date</title>
12037 <simpara>1997-06-08</simpara>
12041 The regression tests have been adapted and extensively modified for the
12042 6.1 release of <productname>PostgreSQL</productname>.
12046 Three new data types (<type>datetime</type>, <type>timespan</type>, and <type>circle</type>) have been added to
12047 the native set of <productname>PostgreSQL</productname> types. Points, boxes, paths, and polygons
12048 have had their output formats made consistent across the data types.
12049 The polygon output in misc.out has only been spot-checked for correctness
12050 relative to the original regression output.
12054 <productname>PostgreSQL</productname> 6.1 introduces a new, alternate
12055 optimizer which uses <firstterm>genetic</firstterm>
12056 algorithms. These algorithms introduce a random behavior in the ordering
12057 of query results when the query contains multiple qualifiers or multiple
12058 tables (giving the optimizer a choice on order of evaluation). Several
12059 regression tests have been modified to explicitly order the results, and
12060 hence are insensitive to optimizer choices. A few regression tests are
12061 for data types which are inherently unordered (e.g. points and time
12062 intervals) and tests involving those types are explicitly bracketed with
12063 <command>set geqo to 'off'</command> and <command>reset geqo</command>.
12067 The interpretation of array specifiers (the curly braces around atomic
12068 values) appears to have changed sometime after the original regression
12069 tests were generated. The current <filename>./expected/*.out</filename> files reflect this
12070 new interpretation, which may not be correct!
12074 The float8 regression test fails on at least some platforms. This is due
12075 to differences in implementations of <function>pow()</function> and <function>exp()</function> and the signaling
12076 mechanisms used for overflow and underflow conditions.
12080 The <quote>random</> results in the random test should cause the
12081 <quote>random</quote> test to be <quote>failed</quote>, since the
12082 regression tests are evaluated using a simple diff. However,
12083 <quote>random</> does not seem to produce random results on my test
12084 machine (Linux/<application>gcc</>/i686).
12088 <title>Migration to version 6.1</title>
12091 This migration requires a complete dump of the 6.0 database and a
12092 restore of the database in 6.1.
12095 Those migrating from earlier 1.* releases should first upgrade to 1.09
12096 because the COPY output format was improved from the 1.02 release.
12101 <title>Changes</title>
12107 packet length checking in library routines
12108 lock manager priority patch
12109 check for under/over flow of float8(Bruce)
12110 multitable join fix(Vadim)
12111 SIGPIPE crash fix(Darren)
12112 large object fixes(Sven)
12113 allow btree indexes to handle NULLs(Vadim)
12114 timezone fixes(D'Arcy)
12115 select SUM(x) can return NULL on no rows(Thomas)
12116 internal optimizer, executor bug fixes(Vadim)
12117 fix problem where inner loop in < or <= has no rows(Vadim)
12118 prevent re-commuting join index clauses(Vadim)
12119 fix join clauses for multiple tables(Vadim)
12120 fix hash, hashjoin for arrays(Vadim)
12121 fix btree for abstime type(Vadim)
12122 large object fixes(Raymond)
12123 fix buffer leak in hash indexes (Vadim)
12124 fix rtree for use in inner scan (Vadim)
12125 fix gist for use in inner scan, cleanups (Vadim, Andrea)
12126 avoid unnecessary local buffers allocation (Vadim, Massimo)
12127 fix local buffers leak in transaction aborts (Vadim)
12128 fix file manager memmory leaks, cleanups (Vadim, Massimo)
12129 fix storage manager memmory leaks (Vadim)
12130 fix btree duplicates handling (Vadim)
12131 fix deleted rows reincarnation caused by vacuum (Vadim)
12132 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
12133 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
12137 attribute optimization statistics(Bruce)
12138 much faster new btree bulk load code(Paul)
12139 BTREE UNIQUE added to bulk load code(Vadim)
12140 new lock debug code(Massimo)
12141 massive changes to libpg++(Leo)
12142 new GEQO optimizer speeds table multitable optimization(Martin)
12143 new WARN message for non-unique insert into unique key(Marc)
12144 update x=-3, no spaces, now valid(Bruce)
12145 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
12146 debug backend now pretty-prints tree(Darren)
12147 new Oracle character functions(Edmund)
12148 new plaintext password functions(Dan)
12149 no such class or insufficient privilege changed to distinct messages(Dan)
12150 new ANSI timestamp function(Dan)
12151 new ANSI Time and Date types (Thomas)
12152 move large chunks of data in backend(Martin)
12153 multicolumn btree indexes(Vadim)
12154 new SET var TO value command(Martin)
12155 update transaction status on reads(Dan)
12156 new locale settings for character types(Oleg)
12157 new SEQUENCE serial number generator(Vadim)
12158 GROUP BY function now possible(Vadim)
12159 re-organize regression test(Thomas,Marc)
12160 new optimizer operation weights(Vadim)
12161 new psql \z grant/permit option(Marc)
12162 new MONEY data type(D'Arcy,Thomas)
12163 tcp socket communication speed improved(Vadim)
12164 new VACUUM option for attribute statistics, and for certain columns (Vadim)
12165 many geometric type improvements(Thomas,Keith)
12166 additional regression tests(Thomas)
12167 new datestyle variable(Thomas,Vadim,Martin)
12168 more comparison operators for sorting types(Thomas)
12169 new conversion functions(Thomas)
12170 new more compact btree format(Vadim)
12171 allow pg_dumpall to preserve database ownership(Bruce)
12172 new SET GEQO=# and R_PLANS variable(Vadim)
12173 old (!GEQO) optimizer can use right-sided plans (Vadim)
12174 typechecking improvement in SQL parser(Bruce)
12175 new SET, SHOW, RESET commands(Thomas,Vadim)
12176 new \connect database USER option
12177 new destroydb -i option (Igor)
12178 new \dt and \di psql commands (Darren)
12179 SELECT "\n" now escapes newline (A. Duursma)
12180 new geometry conversion functions from old format (Thomas)
12182 Source tree changes
12183 -------------------
12184 new configuration script(Marc)
12185 readline configuration option added(Marc)
12186 OS-specific configuration options removed(Marc)
12187 new OS-specific template files(Marc)
12188 no more need to edit Makefile.global(Marc)
12189 re-arrange include files(Marc)
12190 nextstep patches (Gregor Hoffleit)
12191 removed Windows-specific code(Bruce)
12192 removed postmaster -e option, now only postgres -e option (Bruce)
12193 merge duplicate library code in front/backends(Martin)
12194 now works with eBones, international Kerberos(Jun)
12195 more shared library support
12196 c++ include file cleanup(Bruce)
12197 warn about buggy flex(Bruce)
12198 DG/UX, Ultrix, IRIX, AIX portability fixes
12204 <sect1 id="release-6-0">
12205 <title>Release 6.0</title>
12208 <title>Release date</title>
12209 <simpara>1997-01-29</simpara>
12213 A dump/restore is required for those wishing to migrate data from
12214 previous releases of <productname>PostgreSQL</productname>.
12218 <title>Migration from version 1.09 to version 6.0</title>
12221 This migration requires a complete dump of the 1.09 database and a
12222 restore of the database in 6.0.
12227 <title>Migration from pre-1.09 to version 6.0</title>
12230 Those migrating from earlier 1.* releases should first upgrade to 1.09
12231 because the COPY output format was improved from the 1.02 release.
12236 <title>Changes</title>
12242 ALTER TABLE bug - running postgress process needs to re-read table definition
12243 Allow vacuum to be run on one table or entire database(Bruce)
12245 Fix array over-runs of memory writes(Kurt)
12246 Fix elusive btree range/non-range bug(Dan)
12247 Fix for hash indexes on some types like time and date
12248 Fix for pg_log size explosion
12249 Fix permissions on lo_export()(Bruce)
12250 Fix unitialized reads of memory(Kurt)
12251 Fixed ALTER TABLE ... char(3) bug(Bruce)
12252 Fixed a few small memory leaks
12253 Fixed EXPLAIN handling of options and changed full_path option name
12254 Fixed output of group acl privileges
12255 Memory leaks (hunt and destroy with tools like Purify(Kurt)
12256 Minor improvements to rules system
12258 New asserts for run-checking
12259 Overhauled parser/analyze code to properly report errors and increase speed
12260 Pg_dump -d now handles NULL's properly(Bruce)
12261 Prevent SELECT NULL from crashing server (Bruce)
12262 Properly report errors when INSERT ... SELECT columns did not match
12263 Properly report errors when insert column names were not correct
12264 psql \g filename now works(Bruce)
12265 psql fixed problem with multiple statements on one line with multiple outputs
12266 Removed duplicate system OIDs
12267 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
12268 Several fixes for queries that crashed the backend
12269 Starting quote in insert string errors(Bruce)
12270 Submitting an empty query now returns empty status, not just " " query(Bruce)
12274 Add EXPLAIN manual page(Bruce)
12275 Add UNIQUE index capability(Dan)
12276 Add hostname/user level access control rather than just hostname and user
12277 Add synonym of != for <>(Bruce)
12278 Allow "select oid,* from table"
12279 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Bruce)
12280 Allow COPY from the frontend(Bryan)
12281 Allow GROUP BY to use alias column name(Bruce)
12282 Allow actual compression, not just reuse on the same page(Vadim)
12283 Allow installation-configuration option to auto-add all local users(Bryan)
12284 Allow libpq to distinguish between text value '' and null(Bruce)
12285 Allow non-postgres users with createdb privs to destroydb's
12286 Allow restriction on who can create C functions(Bryan)
12287 Allow restriction on who can do backend COPY(Bryan)
12288 Can shrink tables, pg_time and pg_log(Vadim & Erich)
12289 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
12290 Change default decimal constant representation from float4 to float8(Bruce)
12291 European date format now set when postmaster is started
12292 Execute lowercase function names if not found with exact case
12293 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z'
12294 Gist now included in the distrubution(Marc)
12295 Idend authentication of local users(Bryan)
12296 Implement BETWEEN qualifier(Bruce)
12297 Implement IN qualifier(Bruce)
12298 libpq has PQgetisnull()(Bruce)
12299 libpq++ improvements
12300 New options to initdb(Bryan)
12301 Pg_dump allow dump of OIDs(Bruce)
12302 Pg_dump create indexes after tables are loaded for speed(Bruce)
12303 Pg_dumpall dumps all databases, and the user table
12304 Pginterface additions for NULL values(Bruce)
12305 Prevent postmaster from being run as root
12306 psql \h and \? is now readable(Bruce)
12307 psql allow backslashed, semicolons anywhere on the line(Bruce)
12308 psql changed command prompt for lines in query or in quotes(Bruce)
12309 psql char(3) now displays as (bp)char in \d output(Bruce)
12310 psql return code now more accurate(Bryan?)
12311 psql updated help syntax(Bruce)
12312 Re-visit and fix vacuum(Vadim)
12313 Reduce size of regression diffs, remove timezone name difference(Bruce)
12314 Remove compile-time parameters to enable binary distributions(Bryan)
12315 Reverse meaning of HBA masks(Bryan)
12316 Secure Authentication of local users(Bryan)
12317 Speed up vacuum(Vadim)
12318 Vacuum now had VERBOSE option(Bruce)
12320 Source tree changes
12321 -------------------
12322 All functions now have prototypes that are compared against the calls
12323 Allow asserts to be disabled easly from Makefile.global(Bruce)
12324 Change oid constants used in code to #define names
12325 Decoupled sparc and solaris defines(Kurt)
12326 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
12327 Major include file reorganization/reduction(Marc)
12328 Make now stops on compile failure(Bryan)
12329 Makefile restructuring(Bryan, Marc)
12330 Merge bsdi_2_1 to bsdi(Bruce)
12331 Monitor program removed
12332 Name change from Postgres95 to PostgreSQL
12333 New config.h file(Marc, Bryan)
12334 PG_VERSION now set to 6.0 and used by postmaster
12335 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
12336 Reduced the number of #define's, centeralized #define's
12337 Remove duplicate OIDS in system tables(Dan)
12338 Remove duplicate system catalog info or report mismatches(Dan)
12339 Removed many os-specific #define's
12340 Restructured object file generation/location(Bryan, Marc)
12341 Restructured port-specific file locations(Bryan, Marc)
12342 Unused/uninialized variables corrected
12348 <sect1 id="release-1-09">
12349 <title>Release 1.09</title>
12352 <title>Release date</title>
12353 <simpara>1996-11-04</simpara>
12357 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of
12358 the changes listed in 6.0 were actually included in the 1.02.1 to 1.09
12363 <sect1 id="release-1-02">
12364 <title>Release 1.02</title>
12367 <title>Release date</title>
12368 <simpara>1996-08-01</simpara>
12372 <title>Migration from version 1.02 to version 1.02.1</title>
12375 Here is a new migration file for 1.02.1. It includes the 'copy' change
12376 and a script to convert old <acronym>ASCII</acronym> files.
12380 The following notes are for the benefit of users who want to migrate
12381 databases from <productname>Postgres95</> 1.01 and 1.02 to <productname>Postgres95</> 1.02.1.
12384 If you are starting afresh with <productname>Postgres95</> 1.02.1 and do not need
12385 to migrate old databases, you do not need to read any further.
12390 In order to upgrade older <productname>Postgres95</> version 1.01 or 1.02 databases to
12391 version 1.02.1, the following steps are required:
12396 Start up a new 1.02.1 postmaster
12401 Add the new built-in functions and operators of 1.02.1 to 1.01 or 1.02
12402 databases. This is done by running the new 1.02.1 server against
12403 your own 1.01 or 1.02 database and applying the queries attached at
12404 the end of the file. This can be done easily through <command>psql</>. If your
12405 1.01 or 1.02 database is named <literal>testdb</literal> and you have cut the commands
12406 from the end of this file and saved them in <filename>addfunc.sql</filename>:
12408 % psql testdb -f addfunc.sql
12411 Those upgrading 1.02 databases will get a warning when executing the
12412 last two statements in the file because they are already present in 1.02. This is
12413 not a cause for concern.
12420 <title>Dump/Reload Procedure</title>
12423 If you are trying to reload a pg_dump or text-mode, <literal>copy tablename to
12424 stdout</literal> generated with a previous version, you will need to run the
12425 attached <command>sed</command> script on the ASCII file before loading it into the
12426 database. The old format used '.' as end-of-data, while '\.' is now the
12427 end-of-data marker. Also, empty strings are now loaded in as '' rather
12428 than NULL. See the copy manual page for full details.
12431 sed 's/^\.$/\\./g' <in_file >out_file
12435 If you are loading an older binary copy or non-<systemitem>stdout</> copy, there is no
12436 end-of-data character, and hence no conversion necessary.
12439 -- following lines added by agc to reflect the case-insensitive
12440 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
12441 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexeq);
12442 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregexne);
12443 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregexeq);
12444 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticregexne);
12450 <title>Changes</title>
12454 Source code maintenance and development
12455 * worldwide team of volunteers
12456 * the source tree now in CVS at ftp.ki.net
12459 * psql (and underlying libpq library) now has many more options for
12460 formatting output, including HTML
12461 * pg_dump now output the schema and/or the data, with many fixes to
12462 enhance completeness.
12463 * psql used in place of monitor in administration shell scripts.
12464 monitor to be deprecated in next release.
12465 * date/time functions enhanced
12466 * NULL insert/update/comparison fixed/enhanced
12467 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
12469 Bug Fixes (almost too numerous to mention)
12471 * storage management
12472 * check for NULL pointer before dereferencing
12476 * added SolarisX86 port
12477 * added BSD/OS 2.1 port
12482 Contributors (appologies to any missed)
12483 * Kurt J. Lidl <lidl@va.pubnix.com>
12484 (missed in first run, but no less important)
12485 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12486 * Jason Wright <jason@shiloh.vnet.net>
12487 * Cees de Groot <C.deGroot@inter.NL.net>
12488 * ernst.molitor@uni-bonn.de
12489 * michael.siebenborn@ae3.Hypo.DE (Michael Siebenborn (6929))
12490 * Brian E. Gallew <geek+@cmu.edu>
12491 * Vadim B. Mikheev <vadim@sable.krasnoyarsk.su>
12492 * Adam Sussman <myddryn@vidya.com>
12493 * Chris Dunlop <chris@onthe.net.au>
12494 * Marc G. Fournier <scrappy@ki.net>
12495 * Dan McGuirk <mcguirk@indirect.com>
12496 * Dr_George_D_Detlefsen <drgeorge@ilt.com>
12497 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12498 * Massimo Dal Zotto <dz@cs.unitn.it>
12499 * Randy Kunkee <kunkee@Starbase.NeoSoft.COM>
12500 * Rick Weldon <rick@wisetech.com>
12501 * Thomas van Reimersdahl <reimersd@dali.techinfo.rwth-aachen.de>
12502 * david bennett <dave@bensoft.com>
12503 * ernst.molitor@uni-bonn.de
12504 * Julian Assange <proff@suburbia.net>
12505 * Bruce Momjian <pgman@candle.pha.pa.us>
12506 * Paul "Shag" Walmsley <ccshag@cclabs.missouri.edu>
12507 * "Alistair G. Crooks" <azcb0@sde.uts.amdahl.com>
12512 <sect1 id="release-1-01">
12513 <title>Release 1.01</title>
12516 <title>Release date</title>
12517 <simpara>1996-02-23</simpara>
12522 <title>Migration from version 1.0 to version 1.01</title>
12525 The following notes are for the benefit of users who want to migrate
12526 databases from <productname>Postgres95</> 1.0 to <productname>Postgres95</> 1.01.
12529 If you are starting afresh with <productname>Postgres95</> 1.01 and do not need
12530 to migrate old databases, you do not need to read any further.
12533 In order to <productname>Postgres95</> version 1.01 with databases created with
12534 <productname>Postgres95</> version 1.0, the following steps are required:
12539 Set the definition of <symbol>NAMEDATALEN</symbol> in <filename>src/Makefile.global</filename> to 16
12540 and <symbol>OIDNAMELEN</symbol> to 20.
12545 Decide whether you want to use Host based authentication.
12550 If you do, you must create a file name <literal>pg_hba</literal> in your top-level data
12551 directory (typically the value of your <envar>$PGDATA</envar>). <filename>src/libpq/pg_hba</filename>
12552 shows an example syntax.
12557 If you do not want host-based authentication, you can comment out
12562 in <filename>src/Makefile.global</filename>
12565 Note that host-based authentication is turned on by default, and if
12566 you do not take steps A or B above, the out-of-the-box 1.01 will
12567 not allow you to connect to 1.0 databases.
12575 Compile and install 1.01, but DO NOT do the <command>initdb</command> step.
12580 Before doing anything else, terminate your 1.0 postmaster, and
12581 backup your existing <envar>$PGDATA</envar> directory.
12586 Set your <envar>PGDATA</envar> environment variable to your 1.0 databases, but set up
12587 path up so that 1.01 binaries are being used.
12592 Modify the file <filename><envar>$PGDATA</envar>/PG_VERSION</filename> from 5.0 to 5.1
12597 Start up a new 1.01 postmaster
12602 Add the new built-in functions and operators of 1.01 to 1.0
12603 databases. This is done by running the new 1.01 server against
12604 your own 1.0 database and applying the queries attached and saving
12605 in the file 1.0_to_1.01.sql. This can be done easily through <command>psql</command>.
12606 If your 1.0 database is name <literal>testdb</literal>:
12609 % psql testdb -f 1.0_to_1.01.sql
12612 and then execute the following commands (cut and paste from here):
12615 -- add builtin functions that are new to 1.01
12617 create function int4eqoid (int4, oid) returns bool as 'foo'
12618 language 'internal';
12619 create function oideqint4 (oid, int4) returns bool as 'foo'
12620 language 'internal';
12621 create function char2icregexeq (char2, text) returns bool as 'foo'
12622 language 'internal';
12623 create function char2icregexne (char2, text) returns bool as 'foo'
12624 language 'internal';
12625 create function char4icregexeq (char4, text) returns bool as 'foo'
12626 language 'internal';
12627 create function char4icregexne (char4, text) returns bool as 'foo'
12628 language 'internal';
12629 create function char8icregexeq (char8, text) returns bool as 'foo'
12630 language 'internal';
12631 create function char8icregexne (char8, text) returns bool as 'foo'
12632 language 'internal';
12633 create function char16icregexeq (char16, text) returns bool as 'foo'
12634 language 'internal';
12635 create function char16icregexne (char16, text) returns bool as 'foo'
12636 language 'internal';
12637 create function texticregexeq (text, text) returns bool as 'foo'
12638 language 'internal';
12639 create function texticregexne (text, text) returns bool as 'foo'
12640 language 'internal';
12642 -- add builtin functions that are new to 1.01
12644 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
12645 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
12646 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexeq);
12647 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregexne);
12648 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexeq);
12649 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregexne);
12650 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexeq);
12651 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregexne);
12652 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icregexeq);
12653 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icregexne);
12654 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq);
12655 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne);
12663 <title>Changes</title>
12668 * 1.01 is backwards compatible with 1.0 database provided the user
12669 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
12670 If those steps are not taken, 1.01 is not compatible with 1.0 database.
12673 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
12674 * added NeXT port (requires SysVIPC implementation)
12675 * added CAST .. AS ... syntax
12676 * added ASC and DESC key words
12677 * added 'internal' as a possible language for CREATE FUNCTION
12678 internal functions are C functions which have been statically linked
12679 into the postgres backend.
12680 * a new type "name" has been added for system identifiers (table names,
12681 attribute names, etc.) This replaces the old char16 type. The
12682 of name is set by the NAMEDATALEN #define in src/Makefile.global
12683 * a readable reference manual that describes the query language.
12684 * added host-based access control. A configuration file ($PGDATA/pg_hba)
12685 is used to hold the configuration data. If host-based access control
12686 is not desired, comment out HBA=1 in src/Makefile.global.
12687 * changed regex handling to be uniform use of Henry Spencer's regex code
12688 regardless of platform. The regex code is included in the distribution
12689 * added functions and operators for case-insensitive regular expressions.
12690 The operators are ~* and !~*.
12691 * pg_dump uses COPY instead of SELECT loop for better performance
12694 * fixed an optimizer bug that was causing core dumps when
12695 functions calls were used in comparisons in the WHERE clause
12696 * changed all uses of getuid to geteuid so that effective uids are used
12697 * psql now returns non-zero status on errors when using -c
12698 * applied public patches 1-14
12704 <sect1 id="release-1-0">
12705 <title>Release 1.0</title>
12708 <title>Release date</title>
12709 <simpara>1995-09-05</simpara>
12713 <title>Changes</title>
12718 * The copyright of <productname>Postgres</productname> 1.0 has been loosened to be freely modifiable
12719 and modifiable for any purpose. Please read the COPYRIGHT file.
12720 Thanks to Professor Michael Stonebraker for making this possible.
12723 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
12724 EUROPEAN STYLE). This follows SQL-92 specs.
12725 * "delimiters" is now a key word
12728 * sql LIKE syntax has been added
12729 * copy command now takes an optional USING DELIMITER specification.
12730 delimiters can be any single-character string.
12731 * IRIX 5.3 port has been added.
12732 Thanks to Paul Walmsley and others.
12733 * updated pg_dump to work with new libpq
12734 * \d has been added psql
12735 Thanks to Keith Parks
12736 * regexp performance for architectures that use POSIX regex has been
12737 improved due to caching of precompiled patterns.
12738 Thanks to Alistair Crooks
12739 * a new version of libpq++
12740 Thanks to William Wanders
12743 * arbitrary userids can be specified in the createuser script
12744 * \c to connect to other databases in psql now works.
12745 * bad pg_proc entry for float4inc() is fixed
12746 * users with usecreatedb field set can now create databases without
12747 having to be usesuper
12748 * remove access control entries when the entry no longer has any
12750 * fixed non-portable datetimes implementation
12751 * added kerberos flags to the src/backend/Makefile
12752 * libpq now works with kerberos
12753 * typographic errors in the user manual have been corrected.
12754 * btrees with multiple index never worked, now we tell you they don't
12755 work when you try to use them
12761 <sect1 id="release-0-03">
12762 <title><productname>Postgres95</productname> Release 0.03</title>
12765 <title>Release date</title>
12766 <simpara>1995-07-21</simpara>
12770 <title>Changes</title>
12773 Incompatible changes:
12774 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
12775 (due to system catalog changes and indexing structure changes).
12776 * double-quote (") is deprecated as a quoting character for string literals;
12777 you need to convert them to single quotes ('). <!-- " -->
12778 * name of aggregates (eg. int4sum) are renamed in accordance with the
12779 SQL standard (eg. sum).
12780 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
12781 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
12782 previous releases); you might have to do typecasting if you depend on it
12783 being of type float8. If you neglect to do the typecasting and you assign
12784 a float literal to a field of type float8, you may get incorrect values
12786 * LIBPQ has been totally revamped so that frontend applications
12787 can connect to multiple backends
12788 * the usesysid field in pg_user has been changed from int2 to int4 to
12789 allow wider range of Unix user ids.
12790 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
12791 single BSD44_derived port. (thanks to Alistair Crooks)
12793 SQL standard-compliance (the following details changes that makes postgres95
12794 more compliant to the SQL-92 standard):
12795 * the following SQL types are now built-in: smallint, int(eger), float, real,
12796 char(N), varchar(N), date and time.
12798 The following are aliases to existing postgres types:
12799 smallint -> int2
12800 integer, int -> int4
12801 float, real -> float4
12802 char(N) and varchar(N) are implemented as truncated text types. In
12803 addition, char(N) does blank-padding.
12804 * single-quote (') is used for quoting string literals; '' (in addition to
12805 \') is supported as means of inserting a single quote in a string
12806 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
12807 (Also, aggregates can now be overloaded, i.e. you can define your
12808 own MAX aggregate to take in a user-defined type.)
12809 * CHANGE ACL removed. GRANT/REVOKE syntax added.
12810 - Privileges can be given to a group using the "GROUP" key word.
12812 GRANT SELECT ON foobar TO GROUP my_group;
12813 The key word 'PUBLIC' is also supported to mean all users.
12815 Privileges can only be granted or revoked to one user or group
12818 "WITH GRANT OPTION" is not supported. Only class owners can change
12820 - The default access control is to to grant users readonly access.
12821 You must explicitly grant insert/update access to users. To change
12822 this, modify the line in
12823 src/backend/utils/acl.h
12824 that defines ACL_WORLD_DEFAULT
12827 * the bug where aggregates of empty tables were not run has been fixed. Now,
12828 aggregates run on empty tables will return the initial conditions of the
12829 aggregates. Thus, COUNT of an empty table will now properly return 0.
12830 MAX/MIN of an empty table will return a row of value NULL.
12831 * allow the use of \; inside the monitor
12832 * the LISTEN/NOTIFY asynchronous notification mechanism now work
12833 * NOTIFY in rule action bodies now work
12834 * hash indexes work, and access methods in general should perform better.
12835 creation of large btree indexes should be much faster. (thanks to Paul
12838 Other changes and enhancements:
12839 * addition of an EXPLAIN statement used for explaining the query execution
12840 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
12842 * WARN and NOTICE messages no longer have timestamps on them. To turn on
12843 timestamps of error messages, uncomment the line in
12844 src/backend/utils/elog.h:
12845 /* define ELOG_TIMESTAMPS */
12846 * On an access control violation, the message
12847 "Either no such class or insufficient privilege"
12848 will be given. This is the same message that is returned when
12849 a class is not found. This dissuades non-privileged users from
12850 guessing the existence of privileged classes.
12851 * some additional system catalog changes have been made that are not
12852 visible to the user.
12855 * The -oid option has been added to the "pg_result" tcl command.
12856 pg_result -oid returns oid of the last row inserted. If the
12857 last command was not an INSERT, then pg_result -oid returns "".
12858 * the large object interface is available as pg_lo* tcl commands:
12859 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
12861 Portability enhancements and New Ports:
12862 * flex/lex problems have been cleared up. Now, you should be able to use
12863 flex instead of lex on any platforms. We no longer make assumptions of
12864 what lexer you use based on the platform you use.
12865 * The Linux-ELF port is now supported. Various configuration have been
12866 tested: The following configuration is known to work:
12867 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
12868 with everything in ELF format,
12871 * ipcclean added to the distribution
12872 ipcclean usually does not need to be run, but if your backend crashes
12873 and leaves shared memory segments hanging around, ipcclean will
12874 clean them up for you.
12877 * the user manual has been revised and libpq documentation added.
12883 <sect1 id="release-0-02">
12884 <title><productname>Postgres95</productname> Release 0.02</title>
12887 <title>Release date</title>
12888 <simpara>1995-05-25</simpara>
12892 <title>Changes</title>
12896 Incompatible changes:
12897 * The SQL statement for creating a database is 'CREATE DATABASE' instead
12898 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
12899 of 'DESTROYDB'. However, the names of the executables 'createdb' and
12900 'destroydb' remain the same.
12903 * pgperl - a Perl (4.036) interface to Postgres95
12904 * pg_dump - a utility for dumping out a postgres database into a
12905 script file containing query commands. The script files are in a ASCII
12906 format and can be used to reconstruct the database, even on other
12907 machines and other architectures. (Also good for converting
12908 a Postgres 4.2 database to Postgres95 database.)
12910 The following ports have been incorporated into postgres95-beta-0.02:
12911 * the NetBSD port by Alistair Crooks
12912 * the AIX port by Mike Tung
12913 * the Windows NT port by Jon Forrest (more stuff but not done yet)
12914 * the Linux ELF port by Brian Gallew
12916 The following bugs have been fixed in postgres95-beta-0.02:
12917 * new lines not escaped in COPY OUT and problem with COPY OUT when first
12919 * cannot type return to use the default user id in createuser
12920 * SELECT DISTINCT on big tables crashes
12921 * Linux installation problems
12922 * monitor doesn't allow use of 'localhost' as PGHOST
12923 * psql core dumps when doing \c or \l
12924 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
12925 * libpgtcl has a hard-wired default port number
12926 * SELECT DISTINCT INTO TABLE hangs
12927 * CREATE TYPE doesn't accept 'variable' as the internallength
12928 * wrong result using more than 1 aggregate in a SELECT
12934 <sect1 id="release-0-01">
12935 <title><productname>Postgres95</productname> Release 0.01</title>
12938 <title>Release date</title>
12939 <simpara>1995-05-01</simpara>
12948 <sect1 id="timing-results">
12949 <title>Timing Results</title>
12952 These timing results are from running the regression test with the commands
12955 % cd src/test/regress
12957 % time make runtest
12961 Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
12962 to run, presumably due to the scheduling vagaries of multitasking systems.
12966 <title>Version 6.5</title>
12969 As has been the case for previous releases, timing between
12970 releases is not directly comparable since new regression tests
12971 have been added. In general, 6.5 is faster than previous
12976 Timing with <function>fsync()</function> disabled:
12980 02:00 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
12981 04:38 Sparc Ultra 1 143MHz, 64MB, Solaris 2.6
12986 Timing with <function>fsync()</function> enabled:
12990 04:21 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
12993 For the <systemitem class="osname">Linux</systemitem> system above, using <acronym>UW-SCSI</acronym> disks rather than (older) <acronym>IDE</acronym>
12994 disks leads to a 50% improvement in speed on the regression test.
12999 <title>Version 6.4beta</title>
13002 The times for this release are not directly comparable to those for previous releases
13003 since some additional regression tests have been included.
13004 In general, however, 6.4 should be slightly faster than the previous release (thanks, Bruce!).
13009 02:26 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13015 <title>Version 6.3</title>
13018 The times for this release are not directly comparable to those for previous releases
13019 since some additional regression tests have been included and some obsolete tests involving
13020 time travel have been removed.
13021 In general, however, 6.3 is substantially faster than previous releases (thanks, Bruce!).
13026 02:30 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13027 04:12 Dual Pentium Pro 180, 96MB, EIDE, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13033 <title>Version 6.1</title>
13038 06:12 Pentium Pro 180, 32MB, EIDE, Linux 2.0.30, gcc 2.7.2 -O2 -m486
13039 12:06 P-100, 48MB, Linux 2.0.29, gcc
13040 39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
13048 <!-- Keep this comment at the end of the file
13053 sgml-minimize-attributes:nil
13054 sgml-always-quote-attributes:t
13057 sgml-parent-document:nil
13058 sgml-default-dtd-file:"./reference.ced"
13059 sgml-exposed-tags:nil
13060 sgml-local-catalogs:("/usr/lib/sgml/catalog")
13061 sgml-local-ecat-files:nil