2 $PostgreSQL: pgsql/doc/src/sgml/release.sgml,v 1.365 2005/09/04 13:40:24 alvherre 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_] <application>
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 Allow index scans to use an intermediate in-memory bitmap (Tom)
57 In previous releases, only a single index could be used to do
58 lookups on a table. With this feature, if a query has
59 <command>WHERE tab.col1 = 4 and tab.col2 = 9</>, and there is no
60 multicolumn index on col1 and col2, but there is an index on
61 col1 and another on col2, it is possible to search both indexes
62 and combine the results in memory, then do heap fetches for only
63 the rows matching both the col1 and col2
64 restrictions. This is very useful in environments that have a
65 lot of unstructured queries where it is impossible to create
66 indexes that match all possible access conditions. Bitmap scans
67 are useful even with a single index, as they reduce the amount
68 of random access needed; a bitmap index scan is efficient for
69 retrieving fairly large fractions of the complete table, whereas
70 plain index scans are not.
77 Add two-phase commit (Heikki Linnakangas, Alvaro, Tom)
82 Two-phase commit allows transactions to be "prepared" on several
83 computers, and once all computers have successfully prepared
84 their transactions (none failed), all transactions can be
85 committed. Even if a machine crashes after a prepare, the
86 prepared transaction can be committed after the machine is
87 restarted. New syntax includes <command>PREPARE TRANSACTION</> and
88 <command>COMMIT/ROLLBACK PREPARED</>. A new system view
89 <literal>pg_prepared_xacts</> has also been added.
96 Create a new role system that replaces users and groups
102 Roles are a combination of users and groups. Like users, they
103 can have login capability, and like groups, a role can have
104 other roles as members. Roles basically remove the distinction
105 between users and groups. For example, a role can:
112 Have login capability (optionally)
124 Hold access permissions for database objects
130 Inherit permissions from other roles it is a member of
136 Once a user logs into a role, she obtains capabilities of
137 the login role plus any inherited roles, and can use
138 <command>SET ROLE</> to switch to other roles she is a member of.
139 This feature is a generalization of the SQL standard's concept of
141 This change also replaces <structname>pg_shadow</> and
142 <structname>pg_group</> by new role-capable catalogs
143 <structname>pg_authid</> and <structname>pg_auth_members</>. The old
144 tables are redefined as read-only views on the new role tables.
151 Automatically use indexes for <function>MIN()</> and
152 <function>MAX()</> (Tom)
157 In previous releases, the only way to use an index for
158 <function>MIN()</> or <function>MAX()</> was to rewrite the
159 query as <command>SELECT col FROM tab ORDER BY col LIMIT 1</>.
160 Index usage now happens automatically.
167 Move <filename>/contrib/pg_autovacuum</> into the main server
173 Integrating autovacuum into the server allows it to be
174 automatically started and stopped in sync with the database
175 server, and allows autovacuum to be configured from
176 <filename>postgresql.conf</>.
183 Add shared row level locks using <command>SELECT ... FOR SHARE</>
189 While <productname>PostgreSQL</productname>'s MVCC locking
190 allows <command>SELECT</> to never be blocked by writers and
191 therefore does not need shared row locks for typical operations,
192 shared locks are useful for applications that require shared row
193 locking. In particular this reduces the locking requirements
194 imposed by referential integrity checks.
201 Add dependencies on shared objects, specifically roles
207 This extension of the dependency mechanism prevents roles from
208 being dropped while there are still database objects they own.
209 Formerly it was possible to accidentally <quote>orphan</> objects by
210 deleting their owner. While this could be recovered from, it
211 was messy and unpleasant.
220 <title>Migration to version 8.1</title>
223 A dump/restore using <application>pg_dump</application> is required
224 for those wishing to migrate data from any previous release.
228 The 8.0 release announced that the <function>to_char()</> function
229 for intervals would be removed in 8.1. However, since no better API
230 has been suggested, <function>to_char(interval)</> has been enhanced in
231 8.1 and will remain in the server.
235 Observe the following incompatibilities:
242 <varname>add_missing_from</> is now false by default (Neil)
245 By default, we now generate an error if a table is used in a query
246 without a <command>FROM</> reference. The old behavior is still
247 available, but the parameter must be set to 'true' to obtain it.
253 Cause input of a zero-length string ('') for float4/float8/oid
254 to throw an error, rather than treating it as a zero (Neil)
257 This change is consistent with the current handling of
258 zero-length strings for integers. The schedule for this change
259 was announced in 8.0.
265 <varname>default_with_oids</> is now false by default (Neil)
268 With this option set to false, user-created tables no longer have
269 the usually-invisible OID column unless <command>WITH OIDS</>
270 is specified in <command>CREATE TABLE</>. Though OIDs have
271 existed in all releases of
272 <productname>PostgreSQL</productname>, their use is limited
273 because they are only four bytes long and the counter is shared
274 across all installed databases. The preferred way of uniquely
275 identifying rows is via sequences and <command>SERIAL</>, which
276 have been supported since <productname>PostgreSQL</productname>
283 Add <literal>E''</> syntax so eventually ordinary strings can
284 treat backslashes literally (Bruce)
287 Currently <productname>PostgreSQL</productname> processes a
288 backslash in a string literal as introducing a special escape sequence,
289 e.g. <literal>\n</> or <literal>\010</>.
290 While this allows easy entry of special values, it is
291 non-standard and makes porting of applications from other
292 databases more difficult. For this reason, the
293 <productname>PostgreSQL</productname> project is planning to
294 remove the special meaning of backslashes in strings. For
295 backward compatibility and for users who want special backslash
296 processing, a new string syntax has been created. This new string
297 syntax is formed by writing an <literal>E</> immediately preceding the
298 single quote that starts the string, e.g. <literal>E'hi\n'</>. While
299 this release does not change the handling of backslashes in strings, it
300 does add new GUC variables to help users migrate applications
307 <varname>standard_conforming_strings</> - does this release
308 treat backslashes literally in ordinary strings?
314 <varname>escape_string_warning</> - warn about backslashes in
315 ordinary (non-E) strings
322 The <varname>standard_conforming_strings</> value is read-only.
323 Applications can retrieve the value to know how backslashes are
324 processed. (Presence of the parameter can also be taken as an
325 indication that <literal>E''</> string syntax is supported.) In a
326 future release, <varname>standard_conforming_strings</> will be true,
327 meaning backslashes will be treated literally in non-E strings. To
328 prepare for this change, use <literal>E''</> strings in places that
329 need special backslash processing, and turn on
330 <varname>escape_string_warning</> to find additional strings that need
331 to be converted to use <literal>E''</>. Also, use two single-quotes
332 (<literal>''</>) to embed a literal single-quote in a string, rather
333 than the PostgreSQL-supported syntax of backslash single-quote
334 (<literal>\'</>). The former is standards-conforming and does not
335 require the use of the <literal>E''</> string syntax. You can also use
336 the <literal>$$</> string syntax, which does not treat backslashes
343 Make <command>REINDEX DATABASE</> reindex all indexes in the
347 The old behavior of <command>REINDEX</> database reindexed only
348 system tables. This new behavior seems more intuitive. A new
349 command <command>REINDEX SYSTEM</> allows for reindexing just the
356 In psql, treat unquoted <literal>\{digit}+</> sequences as octal
360 In previous releases, <literal>\{digit}+</> sequences were
361 treated as decimal, and only <literal>\0{digit}+</> were treated
362 as octal. This change was made for consistency.
368 Remove grammar productions for prefix and postfix <literal>%</>
369 and <literal>^</> operators
373 These have never been documented and complicated the use of the
374 modulus operator (<literal>%</>) with negative numbers.
380 Make <literal>&<</> and <literal>&></> for polygons
381 consistent with the box "over" operators (Tom)
389 <title>Additional Changes</title>
392 Below you will find a detailed account of the additional changes
393 between release 8.1 and the previous major release.
397 <title>Performance Improvements</title>
402 Improve rtree index capabilities and performance (Neil)
408 Improve the optimizer, including auto-resizing of hash joins
415 Overhaul internal API in several areas
421 Change WAL record CRCs from 64-bit to 32-bit (Tom)
424 We determined that the extra cost of computing 64-bit CRCs was
425 significant, and the gain in reliability too marginal to justify it.
431 Prevent writing large empty gaps in WAL pages (Tom)
437 Allow non-consecutive index columns to be used in a multi-column
442 For example, this allows an index on columns a,b,c to be used in
443 a query with <command>WHERE a = 4 and c = 10</>.
449 Skip WAL logging for <command>CREATE TABLE AS</> /
450 <command>SELECT INTO</> (Simon)
453 Since a crash during <command>CREATE TABLE</> would cause the
454 table to be dropped during recovery, there is no reason to WAL
455 log as the table is loaded. (Logging still happens if WAL
456 archiving is enabled, however.)
462 Allow concurrent GIST index access (Teodor, Oleg)
468 Add GUC <varname>full_page_writes</> to control writing full
473 To prevent partial disk writes from corrupting the database,
474 <productname>PostgreSQL</productname> writes a complete copy of
475 each database disk page to WAL the first time it is modified
476 after a checkpoint. This option turns off that functionality for more
477 speed. This is safe to use with battery-backed disk caches where
478 partial page writes cannot happen.
484 Add constraint_exclusion to restrict child table lookups based
485 on table constraints (Simon)
488 This allows for a type of table partitioning. If child tables
489 store separate key ranges and this is enforced using appropriate
490 <command>CHECK</> constraints, the optimizer will skip child
491 table accesses when the constraint guarantees no matching rows
492 exist in the child table.
498 Use <literal>O_DIRECT</> if available when using
499 <literal>O_SYNC</> for wal_sync_method (Itagaki Takahiro)
502 <literal>O_DIRECT</> causes disk writes to bypass the kernel
503 cache, and for WAL writes, this improves performance.
509 Improve <command>COPY FROM</> performance (Alon Goldshuv)
512 This was accomplished by reading <command>COPY</> input in
513 larger chunks, rather than character by character.
522 <title>Server Changes</title>
527 Prevent problems due to transaction ID (XID) wraparound (Tom)
530 The server will now warn when the transaction counter approaches
531 the wraparound point. If the counter becomes too close to wraparound,
532 the server will stop accepting queries. This ensures that data is
533 not lost before needed vacuuming is performed.
539 Fix problems with object IDs (OIDs) conflicting with existing system
540 objects after the OID counter has wrapped around (Tom)
546 Add warning about the need to increase
547 <varname>max_fsm_relations</> and <varname>max_fsm_pages</>
548 during <command>VACUUM</> (Ron Mayer)
554 Add <varname>temp_buffers</> GUC variable to allow users to
555 determine the size of the local buffer area for temporary table
562 Add session start time and client IP address to
563 <literal>pg_stat_activity</> (Magnus)
569 Enhance <literal>pg_locks</> display (Tom)
575 Log queries for client-side <command>PREPARE</> and
576 <command>EXECUTE</> (Simon)
582 Allow Kerberos name and user name case sensitivity to be
583 specified in <filename>postgresql.conf</> (Magnus)
589 Add GUC <varname>krb_server_hostname</> so the server hostname
590 can be specified as part of service principal (Todd Kover)
593 If not set, any service principal matching an entry in the
594 keytab can be used. This is new Kerberos matching behavior in
601 Add <varname>log_line_prefix</> options for millisecond
602 timestamps (<literal>%m</>) and remote host (<literal>%h</>) (Ed
609 Add WAL logging for GIST indexes (Teodor, Oleg)
612 GIST indexes are now safe for crash and point-in-time recovery
618 Remove old <filename>*.backup</> files when we do
619 <function>pg_stop_backup()</> (Bruce)
622 This prevents a large number of <filename>*.backup</> files from
623 existing in <filename>/pg_xlog</>.
629 Add GUC variables to control TCP/IP keep-alive times for idle,
630 interval, and count (Oliver Jowett)
634 These values can be changed to allow more rapid detection of
635 lost client connections.
641 Add per-user and per-database connection limits (Petr Jelinek)
644 Using <command>ALTER USER</> and <command>ALTER DATABASE</>,
645 limits can now be enforced on the maximum number of sessions that
646 can concurrently connect as a specific user or to a specific database.
647 Setting the limit to zero disables user or database connections.
653 Allow more than two gigabyes of shared memory and per-backend
654 work memory on 64-bit machines (Koichi Suzuki)
663 <title>Query Changes</title>
668 Add temporary views (Koju Iijima)
674 Fix <command>HAVING</> without any aggregate functions or
675 <command>GROUP BY</> so that the query returns a single group (Tom)
678 Previously, such a case would treat the <command>HAVING</>
679 clause the same as a <command>WHERE</> clause. This was not per spec.
685 Add <command>USING</> clause to allow additional tables to be
686 specified to <command>DELETE</> (Euler Taveira de Oliveira)
689 In prior releases, there was no clear method for specifying
690 additional tables to be used for joins in a <command>DELETE</>
691 statement. <command>UPDATE</> already has a <command>FROM</>
692 clause for this purpose.
698 Add support for <literal>\x</> hex escapes in backend and ecpg
702 This is just like the standard C <literal>\x</> escape syntax.
703 Octal escapes were already supported.
709 Add <command>BETWEEN SYMMETRIC</> query syntax (Pavel Stehule)
712 This feature allows <command>BETWEEN</> comparisons without
713 requiring the first value to be less than the second. For
714 example, <command>2 BETWEEN [ASYMMETRIC] 3 AND 1</> returns
715 false, while <command>2 BETWEEN SYMMETRIC 3 AND 1</> returns
716 true. <command>BETWEEN ASYMMETRIC</> was already supported.
722 Add <command>NOWAIT</> option to <command>SELECT ... FOR
723 UPDATE/SHARE</> (Hans-Juergen Schoenig)
726 While <command>SET</> <varname>statement_timeout</> allows a query
727 taking over a certain amount of time to be cancelled, the
728 <command>NOWAIT</> option allows a query to be canceled as soon as a
729 <command>SELECT ... FOR UPDATE/SHARE</> command cannot immediately
739 <title>Object Manipulation Changes</title>
744 Track dependencies of shared objects (Alvaro)
747 <productname>PostgreSQL</productname> allows global tables
748 (users, databases, tablespaces) to reference information in
749 multiple databases. This addition adds dependency information
750 for global tables, so, for example, user ownership can be
751 tracked across databases, so a user who owns something in any
752 database can no longer be removed. Dependency tracking already
753 existed for database-local objects.
759 Allow limited <command>ALTER OWNER</> commands to be performed
760 by the object owner (Stephen Frost)
763 Prior releases allowed only superusers to change object owners.
764 Now, ownership can be transferred if the user executing the command
765 owns the object and would be able to create it as the new owner
766 (that is, the user is a member of the new owning role and that role
767 has the CREATE permission that would be needed to create the object
774 Add <command>ALTER</> object <command>SET SCHEMA</> capability
775 for some object types (tables, functions, types) (Bernd Helmle)
778 This allows objects to be moved to different schemas.
784 Add ALTER TABLE ENABLE/DISABLE TRIGGER to disable triggers
794 <title>Utility Command Changes</title>
799 Allow <command>TRUNCATE</> to truncate multiple tables in a
800 single command (Alvaro)
803 Because of referential integrity checks, it is not allowed to
804 truncate a table that is part of a referential integrity
805 constraint. Using this new functionality, <command>TRUNCATE</>
806 can be used to truncate such tables, if both tables involved in
807 a referential integrity constraint are truncated in a single
808 <command>TRUNCATE</> command.
814 Properly process carriage returns and line feeds in
815 <command>COPY CSV</> mode (Andrew Dunstan)
818 In release 8.0, carriage returns and line feeds in <command>CSV
819 COPY TO</> were processed in an inconsistent manner. (This was
820 documented on the TODO list.)
826 Add <command>COPY WITH CSV HEADER</> to allow a header line as
827 the first line in <command>COPY</> (Andrew)
830 This allows handling of the common <command>CSV</> usage of
831 placing the column names on the first line of the data file. For
832 <command>COPY TO</>, the first line contains the column names,
833 and for <command>COPY FROM</>, the first line is ignored.
839 On Windows, display better sub-second precision in
840 <command>EXPLAIN ANALYZE</> (Magnus)
846 Add trigger duration display to <command>EXPLAIN ANALYZE</>
850 Prior releases included trigger execution time as part of the
851 total execution time, but did not show it separately. It is now
852 possible to see how much time is spent in each trigger.
858 Add support for <literal>\x</> hex escapes in <command>COPY</>
862 Previous releases only supported octal escapes.
868 Make <command>SHOW ALL</> include variable descriptions
872 <command>SHOW</> varname still only displays the variable's
873 value and does not include the description.
879 Make initdb create a new standard database called
880 <literal>postgres</>, and convert utilities to use
881 <literal>postgres</> rather than <literal>template1</> for
882 standard lookups (Dave)
885 In prior releases, <literal>template1</> was used both as a default
886 connection for utilities like createuser, and as a template for
887 new databases. This caused <command>CREATE DATABASE</> to
888 sometimes fail, because a new database cannot be created if
889 anyone else is in the template database. With this change, the
890 default connection database is now <literal>postgres</>, meaning
891 it is much less likely someone will be using
892 <literal>template1</> during <command>CREATE DATABASE</>.
898 Create new reindexdb command-line utility by moving
899 <filename>/contrib/reindexdb</> into the server (Euler Taveira
909 <title>Data Type and Function Changes</title>
914 Add <function>MAX()</> and <function>MIN()</> aggregates for
915 array types (Koju Iijima)
921 Fix <function>to_date()</> and <function>to_timestamp()</> to
922 behave reasonably when <literal>CC</> and <literal>YY</> fields
923 are both used (Karel Zak)
926 If the format specification contains <literal>CC</> and a year
927 specification is <literal>YYY</> or longer, ignore the
928 <literal>CC</>. If the year specification is <literal>YY</> or
929 shorter, interpret <literal>CC</> as the previous century.
935 Add <function>md5(bytea)</> (Abhijit Menon-Sen)
938 <function>md5(text)</> already existed.
944 Fix <type>CHAR()</> to properly pad spaces to the specified
945 length when using a multiple-byte character set (Yoshiyuki
949 In prior releases, the padding of <type>CHAR()</> was incorrect
950 because it only padded to the specified number of bytes without
951 considering how many characters were stored.
957 Add support for <command>numeric ^ numeric</> based on
958 <function>power(numeric, numeric)</>
961 The function already existed, but there was no operator assigned
968 Fix <type>NUMERIC</> modulus by properly truncating the quotient
969 during computation (Bruce)
972 In previous releases, modulus for large values sometimes
973 returned negative results due to rounding of the quotient.
979 Add a function <function>lastval()</>(Dennis Björklund)
982 <function>lastval()</> is a simplified version of
983 <function>currval()</>. It automatically determines the proper
984 sequence name based on the most recent <function>nextval()</> or
985 <function>setval()</> call performed by the current session.
991 Add <function>to_timestamp(DOUBLE PRECISION)</>
994 Converts Unix seconds since 1970 to a <type>TIMESTAMP WITH
1001 Add <function>pg_postmaster_start_time()</> function (Euler
1002 Taveira de Oliveira, Matthias Schmidt)
1008 Allow the full use of time zone names in <command>AT TIME
1009 ZONE</>, not just the short list previously available (Magnus)
1012 Previously, only a predefined list of time zone names were
1013 supported by <command>AT TIME ZONE</>. Now any supported time
1014 zone name can be used, e.g.:
1016 SELECT CURRENT_TIMESTAMP AT TIME ZONE 'Europe/London';
1018 In the above query, the time zone used is adjusted based on the
1019 daylight savings time rules that were in effect on the supplied
1026 Add <function>GREATEST()</> and <function>LEAST()</> functions
1030 These functions take a variable number of arguments and return
1031 the greatest or least value among the arguments.
1037 Add <function>pg_column_size()</> (Mark Kirkwood)
1040 This returns storage size of a column, including any
1047 Add <function>regexp_replace()</> (Atsushi Ogawa)
1050 This allows regular expression replacement, like sed. An optional
1051 flag argument allows selection of global (replace all) and
1052 case-insensitive modes.
1058 Fix interval division and multiplication (Bruce)
1061 Previous versions sometimes returned unjustified results, like
1062 <command>'4 months'::interval / 5</> returning <command>'1 mon
1069 Add an internal day field to <type>INTERVAL</> so a one day
1070 interval can be distinguished from a 24 hour interval (Michael
1074 Days that contain a daylight savings time adjustment are not 24
1075 hours, but typically 23 or 25 hours. This change allows days
1076 (not fixed 24-hour periods) to be added to dates who's result
1077 includes a daylight savings time adjustment period. Therefore,
1078 while in previous releases <literal>1 day</> and <literal>24
1079 hours</> were interchangeable interval values, in this release
1080 they are treated differently, e.g.
1082 '2005-05-03 00:00:00 EST' + '1 day' = '2005-05-04 00:00:00-04'
1083 '2005-05-03 00:00:00 EST' + '24 hours' = '2005-05-04 01:00:00-04'
1090 Add <function>justify_days()</> and <function>justify_hours()</>
1091 (Michael Glaesemann)
1094 These functions, respectively, adjust days to an appropriate
1095 number of full months and days, and adjust hours to an
1096 appropriate number of full days and hours.
1102 Move <filename>/contrib/dbsize</> into the backend, and rename
1103 some of the functions (Dave Page, Andreas Pflug)
1110 <function>pg_tablespace_size()</>
1116 <function>pg_database_size()</>
1122 <function>pg_relation_size()</>
1128 <function>pg_complete_relation_size()</>
1134 <function>pg_size_pretty()</>
1141 <function>pg_complete_relation_size()</> includes indexes and TOAST
1148 Add functions for read-only file access to the cluster directory
1149 (Dave Page, Andreas Pflug)
1156 <function>pg_stat_file()</>
1162 <function>pg_read_file()</>
1168 <function>pg_ls_dir()</>
1178 Add <function>pg_reload_conf()</> to force reloading of the
1179 configuration files (Dave Page, Andreas Pflug)
1186 Add <function>pg_rotate_logfile()</> to force rotation of the
1187 server log file (Dave Page, Andreas Pflug)
1193 Change <literal>pg_stat_*</> views to include TOAST tables (Tom)
1202 <title>Encoding and Locale Changes</title>
1207 Rename some encodings to be more consistent and to follow
1208 international standards (Bruce)
1215 <literal>UNICODE</> is now <literal>UTF8</>
1221 <literal>ALT</> is now <literal>WIN866</>
1227 <literal>WIN</> is now <literal>WIN1251</>
1233 <literal>TCVN</> is now <literal>WIN1258</>
1240 The original names still work.
1246 Add support for <literal>WIN1252</> encoding (Roland Volkmann)
1252 Add support for 3 and 4-byte <literal>UTF8</> characters (John
1256 Previously only one and two-byte <literal>UTF8</> characters
1257 were supported. This is particularly important for support for
1258 some Chinese characters.
1264 Allow direct conversion between <literal>EUC_JP</> and
1265 <literal>SJIS</> to improve performance (Atsushi
1272 Allow the UTF8 encoding to work on Windows (Magnus)
1275 This is done by mapping UTF8 to the Windows-native UTF16
1285 <title>General Server-Side Language Changes</title>
1290 Fix <command>ALTER LANGUAGE RENAME</> (Sergey Yatskevich)
1296 Allow function characteristics, like strictness and volatility,
1297 to be modified via <command>ALTER FUNCTION</> (Neil)
1303 Increase the maximum number of function arguments to 100 (Tom)
1309 Allow SQL and PL/PgSQL functions to use <command>OUT</> and
1310 <command>INOUT</> parameters (Tom)
1313 <command>OUT</> is an alternate way for a function to return
1314 values. Instead of using <command>RETURN</>, values can be
1315 returned by assigning to parameters declared as <command>OUT</> or
1316 <command>INOUT</>. This is notationally simpler in some cases,
1317 particularly so when multiple values need to be returned.
1318 While returning multiple values from a function
1319 was possible in previous releases, this greatly simplifies the
1320 process. (The feature will be extended to other server-side
1321 languages in future releases.)
1327 Move language handler functions into the <literal>pg_catalog</> schema
1330 This makes it easier to drop the public schema if desired.
1338 <title>PL/PgSQL Server-Side Language Changes</title>
1343 Reduce memory usage of PL/PgSQL functions (Neil)
1346 Each function now has its own memory context that can be freed
1347 when the function exits.
1353 Check function syntax at <command>CREATE FUNCTION</> time,
1354 rather than at runtime (Neil)
1357 Previously, syntax errors were reported only when the function
1364 Allow <command>OPEN</> to open non-<command>SELECT</> queries
1365 like <command>EXPLAIN</> and <command>SHOW</> (Tom)
1371 No longer require functions to issue a <command>RETURN</>
1375 This is a byproduct of the newly added <command>OUT</> and
1376 <command>INOUT</> functionality. <command>RETURN</> can
1377 be omitted when it is not needed to provide the function result.
1383 Add support for an optional <command>INTO</> clause to
1384 PL/PgSQL's <command>EXECUTE</> command (Pavel Stehule)
1390 Make <command>CREATE TABLE AS</> set <command>ROW_COUNT</> (Tom)
1396 Define <literal>SQLSTATE</> and <literal>SQLERRM</> to return
1397 the <literal>SQLSTATE</> and error message of the current
1398 exception (Pavel Stehule)
1401 These variables are only accessible inside exception blocks.
1407 Allow the parameters to the <command>RAISE</> statement to be
1408 expressions (Pavel Stehule)
1414 Add a loop <command>CONTINUE</> statement (Pavel Stehule)
1420 Allow block and loop labels (Pavel Stehule)
1429 <title>PL/Perl Server-Side Language Changes</title>
1434 Allow large result sets to be returned efficiently (Abhijit
1438 This allows functions to use <function>return_next()</> to avoid
1439 building the entire result set in memory.
1445 Allow one-row-at-a-time retrieval of query results (Abhijit)
1448 This allows functions to use <function>spi_query()</> and
1449 <function>spi_fetchrow()</> to avoid accumulating the entire
1450 result set in memory.
1456 Force PL/Perl to handle strings as <literal>UTF8</> if the
1457 server encoding is <literal>UTF8</> (David Kamholz)
1463 Add a validator function for PL/Perl (Andrew)
1466 This allows syntax errors to be reported at definition time,
1467 rather than execution time.
1473 Allow PL/Perl to return a Perl array when the function returns
1474 an array type (Andrew)
1477 This basically maps <productname>PostgreSQL</productname> arrays
1484 Allow Perl non-fatal warnings to generate <command>NOTICE</>
1491 Allow <literal>strict</> mode to be enabled (Andrew)
1500 <title><application>psql</> Changes</title>
1505 Add psql <command>\set ON_ERROR_ROLLBACK</> to allow statements
1506 in a transaction to error without affecting the rest of the
1507 transaction (Greg Sabino Mullane)
1510 This is basically implemented by wrapping every statement in a
1517 Add support for <literal>\x</> hex strings in psql variables
1521 Octal escapes were already supported.
1527 Add psql support for <command>troff -ms</> output format (Roger
1534 Allow psql's history file location to be controlled by
1535 <envar>HISTFILE</> (Andreas Seltenreich)
1538 This allows configuration of per-database history storage.
1544 Prevent psql <command>\x</> (expanded mode) from affecting
1545 backslash-command displays (Neil)
1551 Add <option>-L</> option to psql to log sessions (Lorne Sunley)
1554 This option was added because some operating systems do not have
1555 simple command-line activity logging functionality.
1561 Make psql <command>\d</> show tablespaces of indexes (Qingqing
1568 Allow psql help (<command>\h</>) to make a best guess on the
1569 proper help information (Greg Sabino Mullane)
1572 This allows the user to just add <command>\h</> to the front of
1573 the syntax error query and get help on the supported syntax.
1574 Previously any additional query text beyond the command name
1575 had to be removed to use <command>\h</>.
1581 Add psql <command>\pset numericlocale</> to allow numbers to be
1582 output in a locale-aware format (Eugen Nedelcu)
1585 For example, using <literal>C</> locale <literal>100000</> would
1586 be output as <literal>100,000.0</> while a European locale might
1587 output this value as <literal>100.000,0</>.
1596 <title><application>pg_dump</> Changes</title>
1601 Add <option>-n</> / <option>--schema</> switch to
1602 <application>pg_restore</> (Richard van den Berg)
1605 This allows only objects from a specified schema to be restored.
1611 Allow <application>pg_dump</> to dump a consistent snapshot of
1619 Allow <application>pg_dump</> to dump large objects even in
1626 Dump comments for large objects (Tom)
1632 Add <option>--encoding</> to <application>pg_dump</>
1636 This allows a database to be dumped in an encoding that is
1637 different from the server's encoding. This is valuable when
1638 transferring the dump to a machine with a different encoding.
1647 <title>libpq Changes</title>
1652 Add a <envar>PGPASSFILE</> environment variable to specify the
1653 password file's filename (Andrew Dunstan)
1659 Add <function>lo_create()</>, that is similar to
1660 <function>lo_creat()</> but allows the OID of the large object
1661 to be specified (Tom)
1670 <title>Source Code Changes</title>
1675 Add spinlock support for the Itanium processor using Intel
1683 Add Kerberos 5 support for Windows (Magnus)
1689 Add Chinese FAQ (laser@pgsqldb.com)
1695 Rename Rendezvous to Bonjour to match OS/X feature renaming
1702 Add support for wal_fsync_writethrough for Darwin (Chris
1709 Streamline the passing of information within the server, the
1710 optimizer, and the lock system (Tom)
1716 Allow <application>pg_config</> to be compiled using MSVC (Andrew
1720 This is required to build DBD::Pg using <application>MSVC</>.
1726 Remove support for Kerberos V4 (Magnus)
1729 Kerberos 4 had security vulnerabilities and is no longer being
1730 maintained by the authors.
1736 Code cleanups (Coverity static analysis performed by
1743 Modify <filename>postgresql.conf</> to use documentation defaults
1744 <literal>on</>/<literal>off</> rather than
1745 <literal>true</>/<literal>false</> (Bruce)
1751 Enhance <application>pg_config</> to be able to report more
1752 build-time values (Tom)
1758 Allow libpq to be built thread-safe on Windows (Dave Page)
1764 Allow IPv6 connections to be used on Win32 (Andrew)
1773 <title>Contrib Changes</title>
1778 Add <filename>/contrib/pg_buffercache</> contrib module (Mark
1782 This displays the contents of the buffer cache, for debugging and
1783 performance tuning purposes.
1789 Remove <filename>/contrib/array</> because it is obsolete (Tom)
1795 Clean up the <filename>/contrib/lo</> module (Tom)
1801 Move <filename>/contrib/findoidjoins</> to
1802 <filename>/src/tools</> (Tom)
1808 Remove the <literal><<</>, <literal>>></>,
1809 <literal>&<</>, and <literal>&></> operators from
1810 <filename>/contrib/cube</>
1813 These operators were not useful.
1819 Improve <filename>/contrib/btree_gist</> (Janko Richter)
1825 Improve <filename>/contrib/pgcrypto</> (Marko Kreen)
1832 implementation of OpenPGP symmetric-key and public-key encryption
1835 Both RSA and Elgamal public-key algorithms are supported.
1841 Stand alone build: include SHA256/384/512 hashes, Fortuna PRNG
1847 OpenSSL build: support 3DES, use internal AES with OpenSSL < 0.9.7
1853 Take build parameters (OpenSSL, zlib) from <filename>./configure</> result
1856 No need to edit the <filename>Makefile</> anymore.
1862 Remove support for <filename>libmhash</> and <filename>libmcrypt</>
1875 <sect1 id="release-8-0-3">
1876 <title>Release 8.0.3</title>
1879 <title>Release date</title>
1880 <simpara>2005-05-09</simpara>
1884 This release contains a variety of fixes from 8.0.2, including several
1885 security-related issues.
1889 <title>Migration to version 8.0.3</title>
1892 A dump/restore is not required for those running 8.0.X. However,
1893 it is one possible way of handling two significant security problems
1894 that have been found in the initial contents of 8.0.X system
1895 catalogs. A dump/initdb/reload sequence using 8.0.3's initdb will
1896 automatically correct these problems.
1900 The larger security problem is that the built-in character set encoding
1901 conversion functions can be invoked from SQL commands by unprivileged
1902 users, but the functions were not designed for such use and are not
1903 secure against malicious choices of arguments. The fix involves changing
1904 the declared parameter list of these functions so that they can no longer
1905 be invoked from SQL commands. (This does not affect their normal use
1906 by the encoding conversion machinery.)
1910 The lesser problem is that the <filename>contrib/tsearch2</> module
1911 creates several functions that are improperly declared to return
1912 <type>internal</> when they do not accept <type>internal</> arguments.
1913 This breaks type safety for all functions using <type>internal</>
1918 It is strongly recommended that all installations repair these errors,
1919 either by initdb or by following the manual repair procedure given
1920 below. The errors at least allow unprivileged database users to crash
1921 their server process, and may allow unprivileged users to gain the
1922 privileges of a database superuser.
1926 If you wish not to do an initdb, perform the same manual repair
1927 procedures shown in the <link linkend="release-7-4-8">7.4.8 release
1933 <title>Changes</title>
1936 <listitem><para>Change encoding function signature to prevent
1937 misuse</para></listitem>
1938 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
1939 <type>INTERNAL</> function results</para></listitem>
1940 <listitem><para>Guard against incorrect second parameter to
1941 <function>record_out</></para></listitem>
1942 <listitem><para>Repair ancient race condition that allowed a transaction to be
1943 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
1944 than for other purposes</para>
1945 <para>This is an extremely serious bug since it could lead to apparent
1946 data inconsistencies being briefly visible to applications.</para></listitem>
1947 <listitem><para>Repair race condition between relation extension and
1949 <para>This could theoretically have caused loss of a page's worth of
1950 freshly-inserted data, although the scenario seems of very low probability.
1951 There are no known cases of it having caused more than an Assert failure.
1953 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
1955 The comparison code was wrong in the case where the
1956 <literal>--enable-integer-datetimes</> configuration switch had been used.
1957 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
1958 it will need to be <command>REINDEX</>ed after installing this update, because
1959 the fix corrects the sort order of column values.
1961 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
1962 <type>TIME WITH TIME ZONE</> values</para></listitem>
1963 <listitem><para>Fix mis-display of negative fractional seconds in
1964 <type>INTERVAL</> values</para>
1966 This error only occurred when the
1967 <literal>--enable-integer-datetimes</> configuration switch had been used.
1969 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
1970 correctly (Neil)</para></listitem>
1971 <listitem><para>Still more 64-bit fixes for
1972 <filename>contrib/intagg</></para></listitem>
1973 <listitem><para>Prevent incorrect optimization of functions returning
1974 <type>RECORD</></para></listitem>
1975 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
1976 <listitem><para>Fix Borland makefile for libpq</para></listitem>
1977 <listitem><para>Fix <filename>contrib/btree_gist</> for <type>timetz</> type
1978 (Teodor)</para></listitem>
1979 <listitem><para>Make <command>pg_ctl</> check the PID found in
1980 <filename>postmaster.pid</> to see if it is still a live
1981 process</para></listitem>
1982 <listitem><para>Fix <command>pg_dump</>/<command>pg_restore</> problems caused
1983 by addition of dump timestamps</para></listitem>
1984 <listitem><para>Fix interaction between materializing holdable cursors and
1985 firing deferred triggers during transaction commit</para></listitem>
1986 <listitem><para>Fix memory leak in SQL functions returning pass-by-reference
1987 data types</para></listitem>
1993 <sect1 id="release-8-0-2">
1994 <title>Release 8.0.2</title>
1997 <title>Release date</title>
1998 <simpara>2005-04-07</simpara>
2002 This release contains a variety of fixes from 8.0.1.
2006 <title>Migration to version 8.0.2</title>
2009 A dump/restore is not required for those running 8.0.*.
2010 This release updates the major version number of the
2011 <productname>PostgreSQL</productname> libraries, so it might be
2012 necessary to re-link some user applications if they cannot
2013 find the properly-numbered shared library.
2018 <title>Changes</title>
2021 <listitem><para>Increment the major version number of all interface
2022 libraries (Bruce)</para>
2024 This should have been done in 8.0.0. It is required so 7.4.X versions
2025 of PostgreSQL client applications, like <application>psql</>,
2026 can be used on the same machine as 8.0.X applications. This might require
2027 re-linking user applications that use these libraries.
2029 <listitem><para>Add Windows-only <varname>wal_sync_method</> setting of
2030 <option>fsync_writethrough</> (Magnus, Bruce)</para>
2032 This setting causes <productname>PostgreSQL</productname> to write through
2033 any disk-drive write cache when writing to WAL.
2034 This behavior was formerly called <option>fsync</>, but was
2035 renamed because it acts quite differently from <option>fsync</> on other
2039 <listitem><para>Enable the <varname>wal_sync_method</> setting of
2040 <option>open_datasync</> on Windows, and make it the default for that
2041 platform (Magnus, Bruce)</para>
2043 Because the default is no longer <option>fsync_writethrough</>,
2044 data loss is possible during a power failure if the disk drive has
2045 write caching enabled. To turn off the write cache on Windows,
2046 from the <application>Device Manager</>, choose the drive properties,
2047 then <literal>Policies</>.
2050 <listitem><para>New cache management algorithm <acronym>2Q</> replaces
2051 <acronym>ARC</> (Tom)</para>
2053 This was done to avoid a pending US patent on <acronym>ARC</>. The
2054 <acronym>2Q</> code might be a few percentage points slower than
2055 <acronym>ARC</> for some work loads. A better cache management algorithm
2058 <listitem><para>Planner adjustments to improve behavior on freshly-created
2059 tables (Tom)</para></listitem>
2060 <listitem><para>Allow plpgsql to assign to an element of an array that is
2061 initially <literal>NULL</> (Tom)</para>
2063 Formerly the array would remain <literal>NULL</>, but now it becomes a
2064 single-element array. The main SQL engine was changed to handle
2065 <command>UPDATE</> of a null array value this way in 8.0, but the similar
2066 case in plpgsql was overlooked.
2069 <listitem><para>Convert <literal>\r\n</> and <literal>\r</> to <literal>\n</>
2070 in plpython function bodies (Michael Fuhr)</para>
2072 This prevents syntax errors when plpython code is written on a Windows or
2076 <listitem><para>Allow SPI cursors to handle utility commands that return rows,
2077 such as <command>EXPLAIN</> (Tom)</para></listitem>
2078 <listitem><para>Fix <command>CLUSTER</> failure after <command>ALTER TABLE
2079 SET WITHOUT OIDS</> (Tom)</para></listitem>
2080 <listitem><para>Reduce memory usage of <command>ALTER TABLE ADD COLUMN</>
2081 (Neil)</para></listitem>
2082 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</> (Tom)</para></listitem>
2083 <listitem><para>Document the Windows-only <literal>register</> and
2084 <literal>unregister</> options of <application>pg_ctl</> (Magnus)</para></listitem>
2085 <listitem><para>Ensure operations done during backend shutdown are counted by
2086 statistics collector</para>
2088 This is expected to resolve reports of <application>pg_autovacuum</>
2089 not vacuuming the system catalogs often enough — it was not being
2090 told about catalog deletions caused by temporary table removal during
2093 <listitem><para>Change the Windows default for configuration parameter
2094 <varname>log_destination</> to <option>eventlog</> (Magnus)</para>
2096 By default, a server running on Windows will now send log output to the
2097 Windows event logger rather than standard error.
2099 <listitem><para>Make Kerberos authentication work on Windows (Magnus)</para></listitem>
2100 <listitem><para>Allow <command>ALTER DATABASE RENAME</> by superusers
2101 who aren't flagged as having CREATEDB privilege (Tom)</para></listitem>
2102 <listitem><para>Modify WAL log entries for <command>CREATE</> and
2103 <command>DROP DATABASE</> to not specify absolute paths (Tom)</para>
2104 <para>This allows point-in-time recovery on a different machine with possibly
2105 different database location. Note that <command>CREATE TABLESPACE</> still
2106 poses a hazard in such situations.
2108 <listitem><para>Fix crash from a backend exiting with an open transaction
2109 that created a table and opened a cursor on it (Tom)</para></listitem>
2110 <listitem><para>Fix <function>array_map()</> so it can call PL functions
2111 (Tom)</para></listitem>
2112 <listitem><para>Several <filename>contrib/tsearch2</> and
2113 <filename>contrib/btree_gist</> fixes (Teodor)
2115 <listitem><para>Fix crash of some <filename>contrib/pgcrypto</>
2116 functions on some platforms (Marko Kreen)</para></listitem>
2117 <listitem><para>Fix <filename>contrib/intagg</> for 64-bit platforms
2118 (Tom)</para></listitem>
2119 <listitem><para>Fix ecpg bugs in parsing of <command>CREATE</> statement
2120 (Michael)</para></listitem>
2121 <listitem><para>Work around gcc bug on powerpc and amd64 causing problems in
2122 ecpg (Christof Petig)</para></listitem>
2123 <listitem><para>Do not use locale-aware versions of <function>upper()</>,
2124 <function>lower()</>, and <function>initcap()</> when the locale is
2125 <literal>C</> (Bruce)</para>
2127 This allows these functions to work on platforms that generate errors
2128 for non-7-bit data when the locale is <literal>C</>.
2130 <listitem><para>Fix <function>quote_ident()</> to quote names that match keywords (Tom)</para></listitem>
2131 <listitem><para>Fix <function>to_date()</> to behave reasonably when
2132 <literal>CC</> and <literal>YY</> fields are both used (Karel)</para></listitem>
2133 <listitem><para>Prevent <function>to_char(interval)</> from failing
2134 when given a zero-month interval (Tom)</para></listitem>
2135 <listitem><para>Fix wrong week returned by <function>date_trunc('week')</>
2138 <function>date_trunc('week')</>
2139 returned the wrong year for the first few days of January in some years.
2141 <listitem><para>Use the correct default mask length for class <literal>D</>
2142 addresses in <type>INET</> data types (Tom)</para></listitem>
2148 <sect1 id="release-8-0-1">
2149 <title>Release 8.0.1</title>
2152 <title>Release date</title>
2153 <simpara>2005-01-31</simpara>
2157 This release contains a variety of fixes from 8.0.0, including several
2158 security-related issues.
2162 <title>Migration to version 8.0.1</title>
2165 A dump/restore is not required for those running 8.0.0.
2170 <title>Changes</title>
2173 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
2175 On platforms that will automatically execute initialization functions of a
2176 shared library (this includes at least Windows and ELF-based Unixen),
2177 <command>LOAD</> can be used to make the server execute arbitrary code.
2178 Thanks to NGS Software for reporting this.</para></listitem>
2179 <listitem><para>Check that creator of an aggregate function has the right to
2180 execute the specified transition functions</para>
2182 This oversight made it possible to bypass denial of EXECUTE
2183 permission on a function.</para></listitem>
2184 <listitem><para>Fix security and 64-bit issues in
2185 contrib/intagg</para></listitem>
2186 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
2187 Jurka)</para></listitem>
2188 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
2189 many parameters (Neil)</para></listitem>
2190 <listitem><para>Make <command>ALTER TABLE ADD COLUMN</> enforce domain
2191 constraints in all cases</para></listitem>
2192 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
2194 The result of the join was mistakenly supposed to be sorted the same as the
2195 left input. This could not only deliver mis-sorted output to the user, but
2196 in case of nested merge joins could give outright wrong answers.
2198 <listitem><para>Improve planning of grouped aggregate queries</para></listitem>
2199 <listitem><para><command>ROLLBACK TO <replaceable>savepoint</></command>
2200 closes cursors created since the savepoint</para></listitem>
2201 <listitem><para>Fix inadequate backend stack size on Windows</para></listitem>
2202 <listitem><para>Avoid SHGetSpecialFolderPath() on Windows
2203 (Magnus)</para></listitem>
2204 <listitem><para>Fix some problems in running pg_autovacuum as a Windows
2205 service (Dave Page)</para></listitem>
2206 <listitem><para>Multiple minor bug fixes in
2207 pg_dump/pg_restore</para></listitem>
2208 <listitem><para>Fix ecpg segfault with named structs used in
2209 typedefs (Michael)</para></listitem>
2215 <sect1 id="release-8-0">
2216 <title>Release 8.0</title>
2219 <title>Release date</title>
2220 <simpara>2005-01-19</simpara>
2224 <title>Overview</title>
2227 Major changes in this release:
2233 Microsoft Windows Native Server
2238 This is the first <productname>PostgreSQL</productname> release
2239 to run natively on <trademark class=registered>Microsoft Windows</> as
2240 a server. It can run as a <productname>Windows</> service. This
2241 release supports NT-based Windows releases like
2242 <productname>Windows 2000</>, <productname>Windows XP</>, and
2243 <productname>Windows 2003</>. Older releases like
2244 <productname>Windows 95</>, <productname>Windows 98</>, and
2245 <productname>Windows ME</> are not supported because these operating
2246 systems do not have the infrastructure to support
2247 <productname>PostgreSQL</productname>. A separate installer
2248 project has been created to ease installation on
2249 <productname>Windows</> — see <ulink
2250 url="http://www.postgresql.org/ftp/win32/"></ulink>.
2254 Although tested throughout our release cycle, the Windows port
2255 does not have the benefit of years of use in production
2256 environments that <productname>PostgreSQL</productname> has on
2257 Unix platforms. Therefore it should be treated with the same
2258 level of caution as you would a new product.
2262 Previous releases required the Unix emulation toolkit
2263 <productname>Cygwin</> in order to run the server on Windows
2264 operating systems. <productname>PostgreSQL</productname> has
2265 supported native clients on Windows for many years.
2277 Savepoints allow specific parts of a transaction to be aborted
2278 without affecting the remainder of the transaction. Prior
2279 releases had no such capability; there was no way to recover
2280 from a statement failure within a transaction except by
2281 aborting the whole transaction. This feature is valuable for
2282 application writers who require error recovery within a
2283 complex transaction.
2290 Point-In-Time Recovery
2295 In previous releases there was no way to recover from disk
2296 drive failure except to restore from a previous backup or use
2297 a standby replication server. Point-in-time recovery allows
2298 continuous backup of the server. You can recover either to
2299 the point of failure or to some transaction in the past.
2311 Tablespaces allow administrators to select different file systems
2312 for storage of individual tables, indexes, and databases.
2313 This improves performance and control over disk space
2314 usage. Prior releases used <application>initlocation</> and
2315 manual symlink management for such tasks.
2322 Improved Buffer Management, <command>CHECKPOINT</command>,
2323 <command>VACUUM</command>
2328 This release has a more intelligent buffer replacement strategy,
2329 which will make better use of available shared buffers and
2330 improve performance. The performance impact of vacuum and
2331 checkpoints is also lessened.
2343 A column's data type can now be changed with <command>ALTER
2351 New Perl Server-Side Language
2356 A new version of the <application>plperl</> server-side language now
2357 supports a persistent shared storage area, triggers, returning records
2358 and arrays of records, and SPI calls to access the database.
2365 Comma-separated-value (CSV) support in <command>COPY</command>
2370 <command>COPY</command> can now read and write
2371 comma-separated-value files. It has the flexibility to
2372 interpret non-standard quoting and separation characters too.
2381 <title>Migration to version 8.0</title>
2384 A dump/restore using <application>pg_dump</application> is
2385 required for those wishing to migrate data from any previous
2390 Observe the following incompatibilities:
2397 In <option>READ COMMITTED</> serialization mode, volatile functions
2398 now see the results of concurrent transactions committed up to the
2399 beginning of each statement within the function, rather than up to the
2400 beginning of the interactive command that called the function.
2406 Functions declared <option>STABLE</> or <option>IMMUTABLE</> always
2407 use the snapshot of the calling query, and therefore do not see the
2408 effects of actions taken after the calling query starts, whether in
2409 their own transaction or other transactions. Such a function must be
2410 read-only, too, meaning that it cannot use any SQL commands other than
2417 Non-deferred <option>AFTER</> triggers are now fired immediately
2418 after completion of the triggering query, rather than upon
2419 finishing the current interactive command. This makes a
2420 difference when the triggering query occurred within a function:
2421 the trigger is invoked before the function proceeds to its next
2428 Server configuration parameters <varname>virtual_host</> and
2429 <varname>tcpip_socket</> have been replaced with a more general
2430 parameter <varname>listen_addresses</>. Also, the server now listens on
2431 <literal>localhost</> by default, which eliminates the need for the
2432 <literal>-i</> postmaster switch in many scenarios.
2438 Server configuration parameters <varname>SortMem</> and
2439 <varname>VacuumMem</> have been renamed to <varname>work_mem</>
2440 and <varname>maintenance_work_mem</> to better reflect their
2441 use. The original names are still supported in
2442 <command>SET</command> and <command>SHOW</command>.
2448 Server configuration parameters <varname>log_pid</>,
2449 <varname>log_timestamp</>, and <varname>log_source_port</> have been
2450 replaced with a more general parameter <varname>log_line_prefix</>.
2456 Server configuration parameter <varname>syslog</> has been
2457 replaced with a more logical <varname>log_destination</> variable to
2458 control the log output destination.
2464 Server configuration parameter <varname>log_statement</> has been
2465 changed so it can selectively log just database modification or
2466 data definition statements. Server configuration parameter
2467 <varname>log_duration</> now prints only when <varname>log_statement</>
2474 Server configuration parameter <varname>max_expr_depth</> parameter has
2475 been replaced with <varname>max_stack_depth</> which measures the
2476 physical stack size rather than the expression nesting depth. This
2477 helps prevent session termination due to stack overflow caused by
2478 recursive functions.
2484 The <function>length()</> function no longer counts trailing spaces in
2485 <type>CHAR(n)</> values.
2491 Casting an integer to <type>BIT(N)</> selects the rightmost N bits of the
2492 integer, not the leftmost N bits as before.
2498 Updating an element or slice of a NULL array value now produces
2499 a non-NULL array result, namely an array containing
2500 just the assigned-to positions.
2506 Syntax checking of array input values has been tightened up
2507 considerably. Junk that was previously allowed in odd places with
2508 odd results now causes an error. Empty-string element values
2509 must now be written as <literal>""</>, rather than writing nothing.
2510 Also changed behavior with respect to whitespace surrounding
2511 array elements: trailing whitespace is now ignored, for symmetry
2512 with leading whitespace (which has always been ignored).
2518 Overflow in integer arithmetic operations is now detected and
2519 reported as an error.
2525 The arithmetic operators associated with the single-byte
2526 <type>"char"</> data type have been removed.
2532 The <function>extract()</> function (also called
2533 <function>date_part</>) now returns the proper year for BC dates.
2534 It previously returned one less than the correct year. The
2535 function now also returns the proper values for millennium and
2542 <type>CIDR</> values now must have their non-masked bits be zero.
2543 For example, we no longer allow
2544 <literal>204.248.199.1/31</literal> as a <type>CIDR</> value. Such
2545 values should never have been accepted by
2546 <productname>PostgreSQL</productname> and will now be rejected.
2552 <command>EXECUTE</command> now returns a completion tag that
2553 matches the executed statement.
2559 <application>psql</>'s <command>\copy</> command now reads or
2560 writes to the query's <literal>stdin/stdout</>, rather than
2561 <application>psql</>'s <literal>stdin/stdout</>. The previous
2562 behavior can be accessed via new
2563 <option>pstdin</>/<option>pstdout</> parameters.
2569 The JDBC client interface has been removed from the core
2570 distribution, and is now hosted at <ulink url=
2571 "http://jdbc.postgresql.org"></ulink>.
2577 The Tcl client interface has also been removed. There are several
2578 Tcl interfaces now hosted at <ulink url=
2579 "http://gborg.postgresql.org"></ulink>.
2585 The server now uses its own time zone database, rather than the
2586 one supplied by the operating system. This will provide consistent
2587 behavior across all platforms. In most cases, there should be
2588 little noticeable difference in time zone behavior, except that
2589 the time zone names used by <command>SET</>/<command>SHOW</>
2590 <varname>TimeZone</> may
2591 be different from what your platform provides.
2597 <application>Configure</>'s threading option no longer requires
2598 users to run tests or edit configuration files; threading options
2599 are now detected automatically.
2605 Now that tablespaces have been implemented,
2606 <application>initlocation</> has been removed.
2612 The API for user-defined GiST indexes has been changed. The
2613 Union and PickSplit methods are now passed a pointer to a
2614 special <structname>GistEntryVector</structname> structure,
2615 rather than a <type>bytea</type>.
2623 <title>Deprecated Features</title>
2626 Some aspects of <productname>PostgreSQL</productname>'s behavior
2627 have been determined to be suboptimal. For the sake of backward
2628 compatibility these have not been removed in 8.0, but they are
2629 considered deprecated and will be removed in the next major
2636 The 8.1 release will remove the <function>to_char()</> function
2643 The server now warns of empty strings passed to
2644 <type>oid</type>/<type>float4</type>/<type>float8</type> data
2645 types, but continues to interpret them as zeroes as before.
2646 In the next major release, empty strings will be considered
2647 invalid input for these data types.
2653 By default, tables in <productname>PostgreSQL</productname> 8.0
2654 and earlier are created with <type>OID</>s. In the next release,
2655 this will <emphasis>not</emphasis> be the case: to create a table
2656 that contains <type>OID</>s, the <option>WITH OIDS</> clause must
2657 be specified or the <varname>default_with_oids</varname>
2658 configuration parameter must be set. Users are encouraged to
2659 explicitly specify <option>WITH OIDS</> if their tables
2660 require OIDs for compatibility with future releases of
2661 <productname>PostgreSQL</productname>.
2669 <title>Changes</title>
2672 Below you will find a detailed account of the changes between
2673 release 8.0 and the previous major release.
2677 <title>Performance Improvements</title>
2682 Support cross-data-type index usage (Tom)
2685 Before this change, many queries would not use an index if the data
2686 types did not match exactly. This improvement makes index usage more
2687 intuitive and consistent.
2693 New buffer replacement strategy that improves caching (Jan)
2696 Prior releases used a least-recently-used (LRU) cache to keep
2697 recently referenced pages in memory. The LRU algorithm
2698 did not consider the number of times a specific cache entry was
2699 accessed, so large table scans could force out useful cache pages.
2700 The new cache algorithm uses four separate lists to track most
2701 recently used and most frequently used cache pages and dynamically
2702 optimize their replacement based on the work load. This should
2703 lead to much more efficient use of the shared buffer cache.
2704 Administrators who have tested shared buffer sizes in the past
2705 should retest with this new cache replacement policy.
2711 Add subprocess to write dirty buffers periodically to reduce
2712 checkpoint writes (Jan)
2715 In previous releases, the checkpoint process, which runs every few
2716 minutes, would write all dirty buffers to the operating system's
2717 buffer cache then flush all dirty operating system buffers to
2718 disk. This resulted in a periodic spike in disk usage that often
2719 hurt performance. The new code uses a background writer to trickle
2720 disk writes at a steady pace so checkpoints have far fewer dirty
2721 pages to write to disk. Also, the new code does not issue a global
2722 <function>sync()</> call, but instead <function>fsync()</>s just
2723 the files written since the last checkpoint. This should improve
2724 performance and minimize degradation during checkpoints.
2730 Add ability to prolong vacuum to reduce performance impact (Jan)
2733 On busy systems, <command>VACUUM</command> performs many I/O
2734 requests which can hurt performance for other users. This
2735 release allows you to slow down <command>VACUUM</command> to
2736 reduce its impact on other users, though this increases the
2737 total duration of <command>VACUUM</command>.
2743 Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
2746 This improves the way indexes are scanned when many duplicate
2747 values exist in the index.
2753 Use dynamically-generated table size estimates while planning (Tom)
2756 Formerly the planner estimated table sizes using the values seen
2757 by the last <command>VACUUM</command> or <command>ANALYZE</command>,
2758 both as to physical table size (number of pages) and number of rows.
2759 Now, the current physical table size is obtained from the kernel,
2760 and the number of rows is estimated by multiplying the table size
2761 by the row density (rows per page) seen by the last
2762 <command>VACUUM</command> or <command>ANALYZE</command>. This should
2763 produce more reliable estimates in cases where the table size has
2764 changed significantly since the last housekeeping command.
2770 Improved index usage with <literal>OR</> clauses (Tom)
2773 This allows the optimizer to use indexes in statements with many OR
2774 clauses that would not have been indexed in the past. It can also use
2775 multi-column indexes where the first column is specified and the second
2776 column is part of an <literal>OR</> clause.
2782 Improve matching of partial index clauses (Tom)
2785 The server is now smarter about using partial indexes in queries
2786 involving complex <option>WHERE</> clauses.
2792 Improve performance of the GEQO optimizer (Tom)
2795 The GEQO optimizer is used to plan queries involving many tables (by
2796 default, twelve or more). This release speeds up the way queries are
2797 analyzed to decrease time spent in optimization.
2803 Miscellaneous optimizer improvements
2806 There is not room here to list all the minor improvements made, but
2807 numerous special cases work better than in prior releases.
2813 Improve lookup speed for C functions (Tom)
2816 This release uses a hash table to lookup information for dynamically
2817 loaded C functions. This improves their speed so they perform nearly as
2818 quickly as functions that are built into the server executable.
2824 Add type-specific <command>ANALYZE</command> statistics
2825 capability (Mark Cave-Ayland)
2828 This feature allows more flexibility in generating statistics
2829 for non-standard data types.
2835 <command>ANALYZE</command> now collects statistics for
2836 expression indexes (Tom)
2839 Expression indexes (also called functional indexes) allow users to
2840 index not just columns but the results of expressions and function
2841 calls. With this release, the optimizer can gather and use statistics
2842 about the contents of expression indexes. This will greatly improve
2843 the quality of planning for queries in which an expression index is
2850 New two-stage sampling method for <command>ANALYZE</command>
2854 This gives better statistics when the density of valid rows is very
2855 different in different regions of a table.
2861 Speed up <command>TRUNCATE</command> (Tom)
2864 This buys back some of the performance loss observed in 7.4, while still
2865 keeping <command>TRUNCATE</command> transaction-safe.
2874 <title>Server Changes</title>
2879 Add WAL file archiving and point-in-time recovery (Simon Riggs)
2885 Add tablespaces so admins can control disk layout (Gavin)
2891 Add a built-in log rotation program (Andreas Pflug)
2894 It is now possible to log server messages conveniently without
2895 relying on either <application>syslog</> or an external log
2902 Add new read-only server configuration parameters to show server
2903 compile-time settings: <varname>block_size</>,
2904 <varname>integer_datetimes</>, <varname>max_function_args</>,
2905 <varname>max_identifier_length</>, <varname>max_index_keys</> (Joe)
2911 Make quoting of <literal>sameuser</>, <literal>samegroup</>, and
2912 <literal>all</> remove special meaning of these terms in
2913 <filename>pg_hba.conf</> (Andrew)
2919 Use clearer IPv6 name <literal>::1/128</> for
2920 <literal>localhost</> in default <filename>pg_hba.conf</> (Andrew)
2926 Use CIDR format in <filename>pg_hba.conf</> examples (Andrew)
2932 Rename server configuration parameters <varname>SortMem</> and
2933 <varname>VacuumMem</> to <varname>work_mem</> and
2934 <varname>maintenance_work_mem</> (Old names still supported) (Tom)
2937 This change was made to clarify that bulk operations such as index and
2938 foreign key creation use <varname>maintenance_work_mem</>, while
2939 <varname>work_mem</> is for workspaces used during query execution.
2945 Allow logging of session disconnections using server configuration
2946 <varname>log_disconnections</> (Andrew)
2952 Add new server configuration parameter <varname>log_line_prefix</> to
2953 allow control of information emitted in each log line (Andrew)
2956 Available information includes user name, database name, remote IP
2957 address, and session start time.
2963 Remove server configuration parameters <varname>log_pid</>,
2964 <varname>log_timestamp</>, <varname>log_source_port</>; functionality
2965 superseded by <varname>log_line_prefix</> (Andrew)
2971 Replace the <varname>virtual_host</> and <varname>tcpip_socket</>
2972 parameters with a unified <varname>listen_addresses</> parameter
2976 <varname>virtual_host</> could only specify a single IP address to
2977 listen on. <varname>listen_addresses</> allows multiple addresses
2984 Listen on localhost by default, which eliminates the need for the
2985 <option>-i</> postmaster switch in many scenarios (Andrew)
2988 Listening on localhost (<literal>127.0.0.1</>) opens no new
2989 security holes but allows configurations like Windows and JDBC,
2990 which do not support local sockets, to work without special
2997 Remove <varname>syslog</> server configuration parameter, and add more
2998 logical <varname>log_destination</> variable to control log output
3005 Change server configuration parameter <varname>log_statement</> to take
3006 values <varname>all</>, <varname>mod</>, <varname>ddl</>, or
3007 <varname>none</> to select which queries are logged (Bruce)
3010 This allows administrators to log only data definition changes or
3011 only data modification statements.
3017 Some logging-related configuration parameters could formerly be adjusted
3018 by ordinary users, but only in the <quote>more verbose</> direction.
3019 They are now treated more strictly: only superusers can set them.
3020 However, a superuser may use <command>ALTER USER</> to provide per-user
3021 settings of these values for non-superusers. Also, it is now possible
3022 for superusers to set values of superuser-only configuration parameters
3023 via <literal>PGOPTIONS</>.
3029 Allow configuration files to be placed outside the data directory (mlw)
3032 By default, configuration files are kept in the cluster's top directory.
3033 With this addition, configuration files can be placed outside the
3034 data directory, easing administration.
3040 Plan prepared queries only when first executed so constants can be
3041 used for statistics (Oliver Jowett)
3044 Prepared statements plan queries once and execute them many
3045 times. While prepared queries avoid the overhead of re-planning
3046 on each use, the quality of the plan suffers from not knowing the exact
3047 parameters to be used in the query. In this release, planning of
3048 unnamed prepared statements is delayed until the first execution,
3049 and the actual parameter values of that execution are used as
3050 optimization hints. This allows use of out-of-line parameter passing
3051 without incurring a performance penalty.
3057 Allow <command>DECLARE CURSOR</command> to take parameters
3061 It is now useful to issue <command>DECLARE CURSOR</command> in a
3062 <function>Parse</> message with parameters. The parameter values
3063 sent at <function>Bind</> time will be substituted into the
3064 execution of the cursor's query.
3070 Fix hash joins and aggregates of <type>inet</type> and
3071 <type>cidr</type> data types (Tom)
3074 Release 7.4 handled hashing of mixed <type>inet</type> and
3075 <type>cidr</type> values incorrectly. (This bug did not exist
3076 in prior releases because they wouldn't try to hash either
3083 Make <varname>log_duration</> print only when <varname>log_statement</>
3084 prints the query (Ed L.)
3093 <title>Query Changes</title>
3098 Add savepoints (nested transactions) (Alvaro)
3104 Unsupported isolation levels are now accepted and promoted to the
3105 nearest supported level (Peter)
3108 The SQL specification states that if a database doesn't support a
3109 specific isolation level, it should use the next more restrictive level.
3110 This change complies with that recommendation.
3116 Allow <command>BEGIN WORK</command> to specify transaction
3117 isolation levels like <command>START TRANSACTION</command> does
3124 Fix table permission checking for cases in which rules generate
3125 a query type different from the originally submitted query (Tom)
3131 Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
3135 In previous releases, because single quotes had to be used to
3136 quote a function's body, the use of single quotes inside the
3137 function text required use of two single quotes or other error-prone
3138 notations. With this release we add the ability to use "dollar
3139 quoting" to quote a block of text. The ability to use different
3140 quoting delimiters at different nesting levels greatly simplifies
3141 the task of quoting correctly, especially in complex functions.
3142 Dollar quoting can be used anywhere quoted text is needed.
3148 Make <literal>CASE val WHEN compval1 THEN ...</> evaluate <literal>val</> only once (Tom)
3151 <option>CASE</> no longer evaluates the tested expression multiple
3152 times. This has benefits when the expression is complex or is
3159 Test <option>HAVING</> before computing target list of an
3160 aggregate query (Tom)
3163 Fixes improper failure of cases such as <literal>SELECT SUM(win)/SUM(lose)
3164 ... GROUP BY ... HAVING SUM(lose) > 0</>. This should work but formerly
3165 could fail with divide-by-zero.
3171 Replace <varname>max_expr_depth</> parameter with
3172 <varname>max_stack_depth</> parameter, measured in kilobytes of stack
3176 This gives us a fairly bulletproof defense against crashing due to
3177 runaway recursive functions. Instead of measuring the depth of expression
3178 nesting, we now directly measure the size of the execution stack.
3184 Allow arbitrary row expressions (Tom)
3187 This release allows SQL expressions to contain arbitrary composite
3188 types, that is, row values. It also allows functions to more easily
3189 take rows as arguments and return row values.
3195 Allow <option>LIKE</>/<option>ILIKE</> to be used as the operator
3196 in row and subselect comparisons (Fabien Coelho)
3202 Avoid locale-specific case conversion of basic ASCII letters in
3203 identifiers and keywords (Tom)
3206 This solves the <quote>Turkish problem</> with mangling of words
3207 containing <literal>I</> and <literal>i</>. Folding of characters
3208 outside the 7-bit-ASCII set is still locale-aware.
3214 Improve syntax error reporting (Fabien, Tom)
3217 Syntax error reports are more useful than before.
3223 Change <command>EXECUTE</command> to return a completion tag
3224 matching the executed statement (Kris Jurka)
3227 Previous releases return an <command>EXECUTE</command> tag for
3228 any <command>EXECUTE</command> call. In this release, the tag
3229 returned will reflect the command executed.
3235 Avoid emitting <option>NATURAL CROSS JOIN</> in rule listings (Tom)
3238 Such a clause makes no logical sense, but in some cases the rule
3239 decompiler formerly produced this syntax.
3248 <title>Object Manipulation Changes</title>
3253 Add <command>COMMENT ON</> for casts, conversions, languages,
3254 operator classes, and large objects (Christopher)
3260 Add new server configuration parameter <varname>default_with_oids</> to
3261 control whether tables are created with <type>OID</>s by default (Neil)
3264 This allows administrators to control whether <command>CREATE
3265 TABLE</command> commands create tables with or without <type>OID</>
3266 columns by default. (Note: the current factory default setting for
3267 <varname>default_with_oids</> is <literal>TRUE</>, but the default
3268 will become <literal>FALSE</> in future releases.)
3274 Add <option>WITH</> / <option>WITHOUT OIDS</> clause to
3275 <command>CREATE TABLE AS</command> (Neil)
3281 Allow <command>ALTER TABLE DROP COLUMN</> to drop an <type>OID</>
3282 column (<command>ALTER TABLE SET WITHOUT OIDS</> still works)
3289 Allow composite types as table columns (Tom)
3295 Allow <command>ALTER ... ADD COLUMN</> with defaults and
3296 <option>NOT NULL</> constraints; works per SQL spec (Rod)
3299 It is now possible for <option>ADD COLUMN</> to create a column
3300 that is not initially filled with NULLs, but with a specified
3307 Add <command>ALTER COLUMN TYPE</> to change column's type (Rod)
3310 It is now possible to alter a column's data type without dropping
3311 and re-adding the column.
3317 Allow multiple <command>ALTER</> actions in a single <command>ALTER
3318 TABLE</command> command (Rod)
3321 This is particularly useful for <command>ALTER</> commands that
3322 rewrite the table (which include <option>ALTER COLUMN TYPE</> and
3323 <option>ADD COLUMN</> with a default). By grouping
3324 <command>ALTER</> commands together, the table need be rewritten
3331 Allow <command>ALTER TABLE</command> to add <type>SERIAL</type>
3335 This falls out from the new capability of specifying defaults for new
3342 Allow changing the owners of aggregates, conversions, databases,
3343 functions, operators, operator classes, schemas, types, and tablespaces
3344 (Christopher, Euler Taveira de Oliveira)
3347 Previously this required modifying the system tables directly.
3353 Allow temporary object creation to be limited to <option>SECURITY
3354 DEFINER</> functions (Sean Chittenden)
3360 Add <option>ALTER TABLE ... SET WITHOUT CLUSTER</> (Christopher)
3363 Prior to this release, there was no way to clear an auto-cluster
3364 specification except to modify the system tables.
3370 Constraint/Index/<type>SERIAL</> names are now
3371 <replaceable>table_column_type</>
3372 with numbers appended to guarantee uniqueness within the schema
3376 The SQL specification states that such names should be unique
3383 Add <function>pg_get_serial_sequence()</> to return a
3384 <type>SERIAL</> column's sequence name (Christopher)
3387 This allows automated scripts to reliably find the <type>SERIAL</>
3394 Warn when primary/foreign key data type mismatch requires costly lookup
3400 New <command>ALTER INDEX</> command to allow moving of indexes
3401 between tablespaces (Gavin)
3407 Make <command>ALTER TABLE OWNER</> change dependent sequence
3408 ownership too (Alvaro)
3418 <title>Utility Command Changes</title>
3423 Allow <command>CREATE SCHEMA</command> to create triggers,
3424 indexes, and sequences (Neil)
3430 Add <option>ALSO</> keyword to <command>CREATE RULE</command> (Fabien
3434 This allows <option>ALSO</> to be added to rule creation to contrast it with
3435 <option>INSTEAD</> rules.
3441 Add <option>NOWAIT</> option to <command>LOCK</command> (Tatsuo)
3444 This allows the <command>LOCK</command> command to fail if it
3445 would have to wait for the requested lock.
3451 Allow <command>COPY</command> to read and write
3452 comma-separated-value (CSV) files (Andrew, Bruce)
3458 Generate error if the <command>COPY</command> delimiter and NULL
3459 string conflict (Bruce)
3465 <command>GRANT</command>/<command>REVOKE</command> behavior
3466 follows the SQL spec more closely
3472 Avoid locking conflict between <command>CREATE INDEX</command>
3473 and <command>CHECKPOINT</command> (Tom)
3476 In 7.3 and 7.4, a long-running B-tree index build could block concurrent
3477 <command>CHECKPOINT</>s from completing, thereby causing WAL bloat because the
3478 WAL log could not be recycled.
3484 Database-wide <command>ANALYZE</command> does not hold locks
3488 This reduces the potential for deadlocks against other backends
3489 that want exclusive locks on tables. To get the benefit of this
3490 change, do not execute database-wide <command>ANALYZE</command>
3491 inside a transaction block (<command>BEGIN</command> block); it
3492 must be able to commit and start a new transaction for each
3499 <command>REINDEX</command> does not exclusively lock the index's
3500 parent table anymore
3503 The index itself is still exclusively locked, but readers of the
3504 table can continue if they are not using the particular index
3511 Erase MD5 user passwords when a user is renamed (Bruce)
3514 <productname>PostgreSQL</productname> uses the user name as salt
3515 when encrypting passwords via MD5. When a user's name is changed,
3516 the salt will no longer match the stored MD5 password, so the
3517 stored password becomes useless. In this release a notice is
3518 generated and the password is cleared. A new password must then
3519 be assigned if the user is to be able to log in with a password.
3525 New <application>pg_ctl</> <option>kill</> option for Windows (Andrew)
3528 Windows does not have a <literal>kill</> command to send signals to
3529 backends so this capability was added to <application>pg_ctl</>.
3535 Information schema improvements
3541 Add <option>--pwfile</> option to
3542 <application>initdb</application> so the initial password can be
3543 set by GUI tools (Magnus)
3549 Detect locale/encoding mismatch in
3550 <application>initdb</application> (Peter)
3556 Add <option>register</> command to <application>pg_ctl</> to
3557 register Windows operating system service (Dave Page)
3566 <title>Data Type and Function Changes</title>
3571 More complete support for composite types (row types) (Tom)
3574 Composite values can be used in many places where only scalar values
3581 Reject non-rectangular array values as erroneous (Joe)
3584 Formerly, <function>array_in</> would silently build a
3591 Overflow in integer arithmetic operations is now detected (Tom)
3597 The arithmetic operators associated with the single-byte
3598 <type>"char"</> data type have been removed.
3601 Formerly, the parser would select these operators in many situations
3602 where an <quote>unable to select an operator</> error would be more
3603 appropriate, such as <literal>null * null</>. If you actually want
3604 to do arithmetic on a <type>"char"</> column, you can cast it to
3611 Syntax checking of array input values considerably tightened up (Joe)
3614 Junk that was previously allowed in odd places with odd results
3615 now causes an <literal>ERROR</>, for example, non-whitespace
3616 after the closing right brace.
3622 Empty-string array element values must now be written as
3623 <literal>""</>, rather than writing nothing (Joe)
3626 Formerly, both ways of writing an empty-string element value were
3627 allowed, but now a quoted empty string is required. The case where
3628 nothing at all appears will probably be considered to be a NULL
3629 element value in some future release.
3635 Array element trailing whitespace is now ignored (Joe)
3638 Formerly leading whitespace was ignored, but trailing whitespace
3639 between an element value and the delimiter or right brace was
3640 significant. Now trailing whitespace is also ignored.
3646 Emit array values with explicit array bounds when lower bound is not one
3653 Accept <literal>YYYY-monthname-DD</> as a date string (Tom)
3659 Make <function>netmask</> and <function>hostmask</> functions
3660 return maximum-length mask length (Tom)
3666 Change factorial function to return <type>numeric</type> (Gavin)
3669 Returning <type>numeric</type> allows the factorial function to
3670 work for a wider range of input values.
3676 <function>to_char</>/<function>to_date()</> date conversion
3677 improvements (Kurt Roeckx, Fabien Coelho)
3683 Make <function>length()</> disregard trailing spaces in
3684 <type>CHAR(n)</> (Gavin)
3687 This change was made to improve consistency: trailing spaces are
3688 semantically insignificant in <type>CHAR(n)</> data, so they
3689 should not be counted by <function>length()</>.
3695 Warn about empty string being passed to
3696 <type>OID</>/<type>float4</>/<type>float8</> data types (Neil)
3699 8.1 will throw an error instead.
3705 Allow leading or trailing whitespace in
3706 <type>int2</>/<type>int4</>/<type>int8</>/<type>float4</>/<type>float8</>
3714 Better support for IEEE <literal>Infinity</> and <literal>NaN</>
3715 values in <type>float4</type>/<type>float8</type> (Neil)
3718 These should now work on all platforms that support IEEE-compliant
3719 floating point arithmetic.
3725 Add <option>week</> option to <function>date_trunc()</> (Robert Creager)
3731 Fix <function>to_char</function> for <literal>1 BC</>
3732 (previously it returned <literal>1 AD</>) (Bruce)
3738 Fix <function>date_part(year)</> for BC dates (previously it
3739 returned one less than the correct year) (Bruce)
3745 Fix <function>date_part()</> to return the proper millennium and
3746 century (Fabien Coelho)
3749 In previous versions, the century and millennium results had a wrong
3750 number and started in the wrong year, as compared to standard
3751 reckoning of such things.
3757 Add <function>ceiling()</> as an alias for <function>ceil()</>,
3758 and <function>power()</> as an alias for <function>pow()</> for
3759 standards compliance (Neil)
3765 Change <function>ln()</>, <function>log()</>,
3766 <function>power()</>, and <function>sqrt()</> to emit the correct
3767 <literal>SQLSTATE</> error codes for certain error conditions, as
3768 specified by SQL:2003 (Neil)
3774 Add <function>width_bucket()</> function as defined by SQL:2003 (Neil)
3780 Add <function>generate_series()</> functions to simplify working
3781 with numeric sets (Joe)
3787 Fix <function>upper/lower/initcap()</> functions to work with
3788 multibyte encodings (Tom)
3794 Add boolean and bitwise integer <option>AND</>/<option>OR</>
3795 aggregates (Fabien Coelho)
3801 New session information functions to return network addresses for client
3802 and server (Sean Chittenden)
3808 Add function to determine the area of a closed path (Sean Chittenden)
3814 Add function to send cancel request to other backends (Magnus)
3820 Add <type>interval</> plus <type>datetime</> operators (Tom)
3823 The reverse ordering, <type>datetime</> plus <type>interval</>,
3824 was already supported, but both are required by the SQL standard.
3830 Casting an integer to <type>BIT(N)</> selects the rightmost N bits
3835 In prior releases, the leftmost N bits were selected, but this was
3836 deemed unhelpful, not to mention inconsistent with casting from bit
3843 Require <type>CIDR</> values to have all non-masked bits be zero
3853 <title>Server-Side Language Changes</title>
3858 In <literal>READ COMMITTED</> serialization mode, volatile functions
3859 now see the results of concurrent transactions committed up to the
3860 beginning of each statement within the function, rather than up to the
3861 beginning of the interactive command that called the function.
3867 Functions declared <literal>STABLE</> or <literal>IMMUTABLE</> always
3868 use the snapshot of the calling query, and therefore do not see the
3869 effects of actions taken after the calling query starts, whether in
3870 their own transaction or other transactions. Such a function must be
3871 read-only, too, meaning that it cannot use any SQL commands other than
3872 <command>SELECT</>. There is a considerable performance gain from
3873 declaring a function <literal>STABLE</> or <literal>IMMUTABLE</>
3874 rather than <literal>VOLATILE</>.
3880 Non-deferred <option>AFTER</> triggers are now fired immediately
3881 after completion of the triggering query, rather than upon
3882 finishing the current interactive command. This makes a difference
3883 when the triggering query occurred within a function: the trigger
3884 is invoked before the function proceeds to its next operation. For
3885 example, if a function inserts a new row into a table, any
3886 non-deferred foreign key checks occur before proceeding with the
3893 Allow function parameters to be declared with names (Dennis Bjorklund)
3896 This allows better documentation of functions. Whether the names
3897 actually do anything depends on the specific function language
3904 Allow PL/pgSQL parameter names to be referenced in the function (Dennis Bjorklund)
3907 This basically creates an automatic alias for each named parameter.
3913 Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
3916 This allows us to catch simple syntax errors sooner.
3922 More support for composite types (row and record variables) in PL/pgSQL
3925 For example, it now works to pass a rowtype variable to another function
3926 as a single variable.
3932 Default values for PL/pgSQL variables can now reference previously
3939 Improve parsing of PL/pgSQL FOR loops (Tom)
3942 Parsing is now driven by presence of <literal>".."</> rather than
3943 data type of <option>FOR</> variable. This makes no difference for
3944 correct functions, but should result in more understandable error
3945 messages when a mistake is made.
3951 Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
3957 In PL/Tcl, SPI commands are now run in subtransactions. If an error
3958 occurs, the subtransaction is cleaned up and the error is reported
3959 as an ordinary Tcl error, which can be trapped with <literal>catch</>.
3960 Formerly, it was not possible to catch such errors.
3966 Accept <command>ELSEIF</> in PL/pgSQL (Neil)
3969 Previously PL/pgSQL only allowed <command>ELSIF</>, but many people
3970 are accustomed to spelling this keyword <command>ELSEIF</>.
3979 <title><application>psql</> Changes</title>
3984 Improve <application>psql</> information display about database
3985 objects (Christopher)
3991 Allow <application>psql</> to display group membership in
3992 <command>\du</> and <command>\dg</> (Markus Bertheau)
3998 Prevent <application>psql</> <command>\dn</command> from showing
3999 temporary schemas (Bruce)
4005 Allow <application>psql</> to handle tilde user expansion for file
4012 Allow <application>psql</> to display fancy prompts, including
4013 color, via <application>readline</> (Reece Hart, Chet Ramey)
4019 Make <application>psql</> <command>\copy</> match <command>COPY</command> command syntax
4026 Show the location of syntax errors (Fabien Coelho, Tom)
4032 Add <command>CLUSTER</command> information to <application>psql</>
4033 <command>\d</> display
4040 Change <application>psql</> <command>\copy stdin/stdout</> to read
4041 from command input/output (Bruce)
4047 Add <option>pstdin</>/<option>pstdout</> to read from
4048 <application>psql</>'s <literal>stdin</>/<literal>stdout</> (Mark
4055 Add global <application>psql</> configuration file, <filename>psqlrc.sample</filename>
4059 This allows a central file where global <application>psql</> startup commands can
4066 Have <application>psql</> <command>\d+</> indicate if the table
4067 has an <type>OID</> column (Neil)
4073 On Windows, use binary mode in <application>psql</> when reading files so control-Z
4074 is not seen as end-of-file
4080 Have <command>\dn+</> show permissions and description for schemas (Dennis
4087 Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
4093 Allow boolean settings to be set using upper or lower case (Michael Paesold)
4102 <title><application>pg_dump</> Changes</title>
4107 Use dependency information to improve the reliability of
4108 <application>pg_dump</> (Tom)
4111 This should solve the longstanding problems with related objects
4112 sometimes being dumped in the wrong order.
4118 Have <application>pg_dump</> output objects in alphabetical order if possible (Tom)
4121 This should make it easier to identify changes between
4128 Allow <application>pg_restore</> to ignore some SQL errors (Fabien Coelho)
4131 This makes <application>pg_restore</>'s behavior similar to the
4132 results of feeding a <application>pg_dump</> output script to
4133 <application>psql</>. In most cases, ignoring errors and plowing
4134 ahead is the most useful thing to do. Also added was a pg_restore
4135 option to give the old behavior of exiting on an error.
4141 <application>pg_restore</> <option>-l</> display now includes
4142 objects' schema names
4148 New begin/end markers in <application>pg_dump</> text output (Bruce)
4154 Add start/stop times for
4155 <application>pg_dump</>/<application>pg_dumpall</> in verbose mode
4162 Allow most <application>pg_dump</> options in
4163 <application>pg_dumpall</> (Christopher)
4169 Have <application>pg_dump</> use <command>ALTER OWNER</> rather
4170 than <command>SET SESSION AUTHORIZATION</> by default
4180 <title>libpq Changes</title>
4185 Make libpq's <option>SIGPIPE</> handling thread-safe (Bruce)
4191 Add <function>PQmbdsplen()</> which returns the display length
4192 of a character (Tatsuo)
4198 Add thread locking to <application>SSL</> and
4199 <application>Kerberos</> connections (Manfred Spraul)
4205 Allow <function>PQoidValue()</>, <function>PQcmdTuples()</>, and
4206 <function>PQoidStatus()</> to work on <command>EXECUTE</command>
4213 Add <function>PQserverVersion()</> to provide more convenient
4214 access to the server version number (Greg Sabino Mullane)
4220 Add <function>PQprepare/PQsendPrepared()</> functions to support
4221 preparing statements without necessarily specifying the data types
4222 of their parameters (Abhijit Menon-Sen)
4228 Many ECPG improvements, including <command>SET DESCRIPTOR</> (Michael)
4237 <title>Source Code Changes</title>
4242 Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
4248 Shell script commands converted to C versions for Windows support (Andrew)
4254 Create an extension makefile framework (Fabien Coelho, Peter)
4257 This simplifies the task of building extensions outside the original
4264 Support relocatable installations (Bruce)
4267 Directory paths for installed files (such as the
4268 <filename>/share</> directory) are now computed relative to the
4269 actual location of the executables, so that an installation tree
4270 can be moved to another place without reconfiguring and
4277 Use <option>--with-docdir</> to choose installation location of documentation; also
4278 allow <option>--infodir</> (Peter)
4284 Add <option>--without-docdir</> to prevent installation of documentation (Peter)
4290 Upgrade to <application>DocBook</> V4.2 SGML (Peter)
4296 New <literal>PostgreSQL</> <application>CVS</> tag (Marc)
4299 This was done to make it easier for organizations to manage their
4300 own copies of the <productname>PostgreSQL</productname>
4301 <application>CVS</> repository. File version stamps from the master
4302 repository will not get munged by checking into or out of a copied
4309 Clarify locking code (Manfred Koizar)
4315 Buffer manager cleanup (Neil)
4321 Decouple platform tests from CPU spinlock code (Bruce, Tom)
4327 Add inlined test-and-set code on PA-RISC for <application>gcc</>
4334 Improve i386 spinlock code (Manfred Spraul)
4340 Clean up spinlock assembly code to avoid warnings from newer
4341 <application>gcc</> releases (Tom)
4347 Remove JDBC from source tree; now a separate project
4353 Remove the libpgtcl client interface; now a separate project
4359 More accurately estimate memory and file descriptor usage (Tom)
4365 Improvements to the Mac OS X startup scripts (Ray A.)
4371 New <function>fsync()</> test program (Bruce)
4377 Major documentation improvements (Neil, Peter)
4383 Remove <application>pg_encoding</application>; not needed
4390 Remove <application>pg_id</application>; not needed anymore
4396 Remove <application>initlocation</application>; not needed
4403 Auto-detect thread flags (no more manual testing) (Bruce)
4409 Use Olson's public domain <application>timezone</> library (Magnus)
4415 With threading enabled, use thread flags on Unixware for
4416 backend executables too (Bruce)
4419 Unixware cannot mix threaded and non-threaded object files in the
4420 same executable, so everything must be compiled as threaded.
4426 <application>psql</> now uses a <application>flex</>-generated
4427 lexical analyzer to process command strings
4433 Reimplement the linked list data structure used throughout the
4437 This improves performance by allowing list append and length
4438 operations to be more efficient.
4444 Allow dynamically loaded modules to create their own server configuration
4445 parameters (Thomas Hallgren)
4451 New Brazilian version of FAQ (Euler Taveira de Oliveira)
4457 Add French FAQ (Guillaume Lelarge)
4463 New <application>pgevent</> for Windows logging
4469 Make libpq and ECPG build as proper shared libraries on OS X (Tom)
4478 <title>Contrib Changes</title>
4483 Overhaul of <filename>contrib/dblink</> (Joe)
4489 <filename>contrib/dbmirror</> improvements (Steven Singer)
4495 New <filename>contrib/xml2</> (John Gray, Torchbox)
4501 Updated <filename>contrib/mysql</filename>
4507 New version of <filename>contrib/btree_gist</> (Teodor)
4513 New <filename>contrib/trgm</>, trigram matching for
4514 <productname>PostgreSQL</productname> (Teodor)
4520 Many <filename>contrib/tsearch2</> improvements (Teodor)
4526 Add double metaphone to <filename>contrib/fuzzystrmatch</> (Andrew)
4532 Allow <filename>contrib/pg_autovacuum</> to run as a Windows service (Dave Page)
4538 Add functions to <filename>contrib/dbsize</> (Andreas Pflug)
4544 Removed <filename>contrib/pg_logger</>: obsoleted by integrated logging
4551 Removed <filename>contrib/rserv</>: obsoleted by various separate projects
4561 <sect1 id="release-7-4-8">
4562 <title>Release 7.4.8</title>
4565 <title>Release date</title>
4566 <simpara>2005-05-09</simpara>
4570 This release contains a variety of fixes from 7.4.7, including several
4571 security-related issues.
4575 <title>Migration to version 7.4.8</title>
4578 A dump/restore is not required for those running 7.4.X. However,
4579 it is one possible way of handling two significant security problems
4580 that have been found in the initial contents of 7.4.X system
4581 catalogs. A dump/initdb/reload sequence using 7.4.8's initdb will
4582 automatically correct these problems.
4586 The larger security problem is that the built-in character set encoding
4587 conversion functions can be invoked from SQL commands by unprivileged
4588 users, but the functions were not designed for such use and are not
4589 secure against malicious choices of arguments. The fix involves changing
4590 the declared parameter list of these functions so that they can no longer
4591 be invoked from SQL commands. (This does not affect their normal use
4592 by the encoding conversion machinery.)
4596 The lesser problem is that the <filename>contrib/tsearch2</> module
4597 creates several functions that are misdeclared to return
4598 <type>internal</> when they do not accept <type>internal</> arguments.
4599 This breaks type safety for all functions using <type>internal</>
4604 It is strongly recommended that all installations repair these errors,
4605 either by initdb or by following the manual repair procedures given
4606 below. The errors at least allow unprivileged database users to crash
4607 their server process, and may allow unprivileged users to gain the
4608 privileges of a database superuser.
4612 If you wish not to do an initdb, perform the following procedures instead.
4613 As the database superuser, do:
4617 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
4618 WHERE pronamespace = 11 AND pronargs = 5
4619 AND proargtypes[2] = 'cstring'::regtype;
4620 -- The command should report having updated 90 rows;
4621 -- if not, rollback and investigate instead of committing!
4625 Next, if you have installed <filename>contrib/tsearch2</>, do
4629 UPDATE pg_proc SET proargtypes[0] = 'internal'::regtype
4631 'dex_init(text)'::regprocedure,
4632 'snb_en_init(text)'::regprocedure,
4633 'snb_ru_init(text)'::regprocedure,
4634 'spell_init(text)'::regprocedure,
4635 'syn_init(text)'::regprocedure
4637 -- The command should report having updated 5 rows;
4638 -- if not, rollback and investigate instead of committing!
4642 If this command fails with a message like <quote>function
4643 "dex_init(text)" does not exist</>, then either <filename>tsearch2</>
4644 is not installed in this database, or you already did the update.
4648 The above procedures must be carried out in <emphasis>each</> database
4649 of an installation, including <literal>template1</>, and ideally
4650 including <literal>template0</> as well. If you do not fix the
4651 template databases then any subsequently created databases will contain
4652 the same errors. <literal>template1</> can be fixed in the same way
4653 as any other database, but fixing <literal>template0</> requires
4654 additional steps. First, from any database issue
4656 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
4658 Next connect to <literal>template0</> and perform the above repair
4659 procedures. Finally, do
4661 -- re-freeze template0:
4663 -- and protect it against future alterations:
4664 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
4670 <title>Changes</title>
4673 <listitem><para>Change encoding function signature to prevent
4674 misuse</para></listitem>
4675 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
4676 <type>INTERNAL</> function results</para></listitem>
4677 <listitem><para>Repair ancient race condition that allowed a transaction to be
4678 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
4679 than for other purposes</para>
4680 <para>This is an extremely serious bug since it could lead to apparent
4681 data inconsistencies being briefly visible to applications.</para></listitem>
4682 <listitem><para>Repair race condition between relation extension and
4684 <para>This could theoretically have caused loss of a page's worth of
4685 freshly-inserted data, although the scenario seems of very low probability.
4686 There are no known cases of it having caused more than an Assert failure.
4688 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
4690 The comparison code was wrong in the case where the
4691 <literal>--enable-integer-datetimes</> configuration switch had been used.
4692 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
4693 it will need to be <command>REINDEX</>ed after installing this update, because
4694 the fix corrects the sort order of column values.
4696 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
4697 <type>TIME WITH TIME ZONE</> values</para></listitem>
4698 <listitem><para>Fix mis-display of negative fractional seconds in
4699 <type>INTERVAL</> values</para>
4701 This error only occurred when the
4702 <literal>--enable-integer-datetimes</> configuration switch had been used.
4704 <listitem><para>Ensure operations done during backend shutdown are counted by
4705 statistics collector</para>
4707 This is expected to resolve reports of <application>pg_autovacuum</>
4708 not vacuuming the system catalogs often enough — it was not being
4709 told about catalog deletions caused by temporary table removal during
4712 <listitem><para>Additional buffer overrun checks in plpgsql
4713 (Neil)</para></listitem>
4714 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
4715 correctly (Neil)</para></listitem>
4716 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
4717 (Marko Kreen)</para></listitem>
4718 <listitem><para>Still more 64-bit fixes for
4719 <filename>contrib/intagg</></para></listitem>
4720 <listitem><para>Prevent incorrect optimization of functions returning
4721 <type>RECORD</></para></listitem>
4722 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
4723 month-related formats</para></listitem>
4724 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
4725 <listitem><para>Fix <function>array_map</> to call PL functions correctly</para></listitem>
4726 <listitem><para>Fix permission checking in <command>ALTER DATABASE RENAME</></para></listitem>
4727 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</></para></listitem>
4728 <listitem><para>Make <function>RemoveFromWaitQueue</> clean up after itself</para>
4730 This fixes a lock management error that would only be visible if a transaction
4731 was kicked out of a wait for a lock (typically by query cancel) and then the
4732 holder of the lock released it within a very narrow window.
4734 <listitem><para>Fix problem with untyped parameter appearing in
4735 <command>INSERT ... SELECT</></para></listitem>
4736 <listitem><para>Fix <command>CLUSTER</> failure after
4737 <command>ALTER TABLE SET WITHOUT OIDS</></para></listitem>
4743 <sect1 id="release-7-4-7">
4744 <title>Release 7.4.7</title>
4747 <title>Release date</title>
4748 <simpara>2005-01-31</simpara>
4752 This release contains a variety of fixes from 7.4.6, including several
4753 security-related issues.
4757 <title>Migration to version 7.4.7</title>
4760 A dump/restore is not required for those running 7.4.X.
4765 <title>Changes</title>
4768 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
4770 On platforms that will automatically execute initialization functions of a
4771 shared library (this includes at least Windows and ELF-based Unixen),
4772 <command>LOAD</> can be used to make the server execute arbitrary code.
4773 Thanks to NGS Software for reporting this.</para></listitem>
4774 <listitem><para>Check that creator of an aggregate function has the right to
4775 execute the specified transition functions</para>
4777 This oversight made it possible to bypass denial of EXECUTE
4778 permission on a function.</para></listitem>
4779 <listitem><para>Fix security and 64-bit issues in
4780 contrib/intagg</para></listitem>
4781 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
4782 Jurka)</para></listitem>
4783 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
4784 many parameters (Neil)</para></listitem>
4785 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
4787 The result of the join was mistakenly supposed to be sorted the same as the
4788 left input. This could not only deliver mis-sorted output to the user, but
4789 in case of nested merge joins could give outright wrong answers.
4791 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
4792 <listitem><para>Fix display of negative intervals in SQL and GERMAN
4793 datestyles</para></listitem>
4794 <listitem><para>Make age(timestamptz) do calculation in local timezone not
4795 GMT</para></listitem>
4801 <sect1 id="release-7-4-6">
4802 <title>Release 7.4.6</title>
4805 <title>Release date</title>
4806 <simpara>2004-10-22</simpara>
4810 This release contains a variety of fixes from 7.4.5.
4815 <title>Migration to version 7.4.6</title>
4818 A dump/restore is not required for those running 7.4.X.
4823 <title>Changes</title>
4826 <listitem><para>Repair possible failure to update hint bits on disk</para>
4828 Under rare circumstances this oversight could lead to
4829 <quote>could not access transaction status</> failures, which qualifies
4830 it as a potential-data-loss bug.
4832 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
4834 Very large left joins using a hash join plan could fail to output unmatched
4835 left-side rows given just the right data distribution.
4837 <listitem><para>Disallow running <application>pg_ctl</> as root</para>
4839 This is to guard against any possible security issues.
4841 <listitem><para>Avoid using temp files in <filename>/tmp</> in <command>make_oidjoins_check</command></para>
4843 This has been reported as a security issue, though it's hardly worthy of
4844 concern since there is no reason for non-developers to use this script anyway.
4846 <listitem><para>Prevent forced backend shutdown from re-emitting prior command
4849 In rare cases, a client might think that its last command had succeeded when
4850 it really had been aborted by forced database shutdown.
4852 <listitem><para>Repair bug in <function>pg_stat_get_backend_idset</function></para>
4854 This could lead to misbehavior in some of the system-statistics views.
4856 <listitem><para>Fix small memory leak in postmaster</para></listitem>
4857 <listitem><para>Fix <quote>expected both swapped tables to have TOAST
4858 tables</> bug</para>
4860 This could arise in cases such as CLUSTER after ALTER TABLE DROP COLUMN.
4862 <listitem><para>Prevent <literal>pg_ctl restart</> from adding <literal>-D</> multiple times</para></listitem>
4863 <listitem><para>Fix problem with NULL values in GiST indexes</para></listitem>
4864 <listitem><para><literal>::</> is no longer interpreted as a variable in an
4865 ECPG prepare statement</para></listitem>
4871 <sect1 id="release-7-4-5">
4872 <title>Release 7.4.5</title>
4875 <title>Release date</title>
4876 <simpara>2004-08-18</simpara>
4880 This release contains one serious bug fix over 7.4.4.
4885 <title>Migration to version 7.4.5</title>
4888 A dump/restore is not required for those running 7.4.X.
4893 <title>Changes</title>
4896 <listitem><para>Repair possible crash during concurrent B-tree index insertions</para>
4898 This patch fixes a rare case in which concurrent insertions into a B-tree index
4899 could result in a server panic. No permanent damage would result, but it's
4900 still worth a re-release. The bug does not exist in pre-7.4 releases.
4907 <sect1 id="release-7-4-4">
4908 <title>Release 7.4.4</title>
4911 <title>Release date</title>
4912 <simpara>2004-08-16</simpara>
4916 This release contains a variety of fixes from 7.4.3.
4921 <title>Migration to version 7.4.4</title>
4924 A dump/restore is not required for those running 7.4.X.
4929 <title>Changes</title>
4932 <listitem><para>Prevent possible loss of committed transactions during crash</para>
4934 Due to insufficient interlocking between transaction commit and checkpointing,
4935 it was possible for transactions committed just before the most recent
4936 checkpoint to be lost, in whole or in part, following a database crash and
4937 restart. This is a serious bug that has existed
4938 since <productname>PostgreSQL</productname> 7.1.
4940 <listitem><para>Check HAVING restriction before evaluating result list of an
4941 aggregate plan</para></listitem>
4942 <listitem><para>Avoid crash when session's current user ID is deleted</para></listitem>
4943 <listitem><para>Fix hashed crosstab for zero-rows case (Joe)</para></listitem>
4944 <listitem><para>Force cache update after renaming a column in a foreign key</para></listitem>
4945 <listitem><para>Pretty-print UNION queries correctly</para></listitem>
4946 <listitem><para>Make psql handle <literal>\r\n</> newlines properly in COPY IN</para></listitem>
4947 <listitem><para><application>pg_dump</> handled ACLs with grant options incorrectly</para></listitem>
4948 <listitem><para>Fix thread support for OS X and Solaris</para></listitem>
4949 <listitem><para>Updated JDBC driver (build 215) with various fixes</para></listitem>
4950 <listitem><para>ECPG fixes</para></listitem>
4951 <listitem><para>Translation updates (various contributors)</para></listitem>
4957 <sect1 id="release-7-4-3">
4958 <title>Release 7.4.3</title>
4961 <title>Release date</title>
4962 <simpara>2004-06-14</simpara>
4966 This release contains a variety of fixes from 7.4.2.
4971 <title>Migration to version 7.4.3</title>
4974 A dump/restore is not required for those running 7.4.X.
4979 <title>Changes</title>
4982 <listitem><para>Fix temporary memory leak when using non-hashed aggregates (Tom)</para></listitem>
4983 <listitem><para>ECPG fixes, including some for Informix compatibility (Michael)</para></listitem>
4984 <listitem><para>Fixes for compiling with thread-safety, particularly Solaris (Bruce)</para></listitem>
4985 <listitem><para>Fix error in COPY IN termination when using the old network protocol (ljb)</para></listitem>
4986 <listitem><para>Several important fixes in pg_autovacuum, including fixes for
4987 large tables, unsigned oids, stability, temp tables, and debug mode
4988 (Matthew T. O'Connor)</para></listitem>
4989 <listitem><para>Fix problem with reading tar-format dumps on NetBSD and BSD/OS (Bruce)</para></listitem>
4990 <listitem><para>Several JDBC fixes</para></listitem>
4991 <listitem><para>Fix ALTER SEQUENCE RESTART where last_value equals the restart value (Tom)</para></listitem>
4992 <listitem><para>Repair failure to recalculate nested sub-selects (Tom)</para></listitem>
4993 <listitem><para>Fix problems with non-constant expressions in LIMIT/OFFSET</para></listitem>
4994 <listitem><para>Support FULL JOIN with no join clause, such as X FULL JOIN Y ON TRUE (Tom)</para></listitem>
4995 <listitem><para>Fix another zero-column table bug (Tom)</para></listitem>
4996 <listitem><para>Improve handling of non-qualified identifiers in GROUP BY clauses in sub-selects (Tom)</para>
4998 Select-list aliases within the sub-select will now take precedence over
4999 names from outer query levels.
5001 <listitem><para>Do not generate <quote>NATURAL CROSS JOIN</> when decompiling rules (Tom)</para></listitem>
5002 <listitem><para>Add checks for invalid field length in binary COPY (Tom)</para>
5004 This fixes a difficult-to-exploit security hole.
5006 <listitem><para>Avoid locking conflict between <command>ANALYZE</command> and <command>LISTEN</command>/<command>NOTIFY</command></para></listitem>
5007 <listitem><para>Numerous translation updates (various contributors)</para></listitem>
5013 <sect1 id="release-7-4-2">
5014 <title>Release 7.4.2</title>
5017 <title>Release date</title>
5018 <simpara>2004-03-08</simpara>
5022 This release contains a variety of fixes from 7.4.1.
5027 <title>Migration to version 7.4.2</title>
5030 A dump/restore is not required for those running 7.4.X. However,
5031 it may be advisable as the easiest method of incorporating fixes for
5032 two errors that have been found in the initial contents of 7.4.X system
5033 catalogs. A dump/initdb/reload sequence using 7.4.2's initdb will
5034 automatically correct these problems.
5038 The more severe of the two errors is that data type <type>anyarray</>
5039 has the wrong alignment label; this is a problem because the
5040 <structname>pg_statistic</> system catalog uses <type>anyarray</>
5041 columns. The mislabeling can cause planner misestimations and even
5042 crashes when planning queries that involve <literal>WHERE</> clauses on
5043 double-aligned columns (such as <type>float8</> and <type>timestamp</>).
5044 It is strongly recommended that all installations repair this error,
5045 either by initdb or by following the manual repair procedure given
5050 The lesser error is that the system view <structname>pg_settings</>
5051 ought to be marked as having public update access, to allow
5052 <literal>UPDATE pg_settings</> to be used as a substitute for
5053 <command>SET</>. This can also be fixed either by initdb or manually,
5054 but it is not necessary to fix unless you want to use <literal>UPDATE
5059 If you wish not to do an initdb, the following procedure will work
5060 for fixing <structname>pg_statistic</>. As the database superuser,
5064 -- clear out old data in pg_statistic:
5065 DELETE FROM pg_statistic;
5066 VACUUM pg_statistic;
5067 -- this should update 1 row:
5068 UPDATE pg_type SET typalign = 'd' WHERE oid = 2277;
5069 -- this should update 6 rows:
5070 UPDATE pg_attribute SET attalign = 'd' WHERE atttypid = 2277;
5072 -- At this point you MUST start a fresh backend to avoid a crash!
5074 -- repopulate pg_statistic:
5078 This can be done in a live database, but beware that all backends
5079 running in the altered database must be restarted before it is safe to
5080 repopulate <structname>pg_statistic</>.
5084 To repair the <structname>pg_settings</> error, simply do:
5086 GRANT SELECT, UPDATE ON pg_settings TO PUBLIC;
5091 The above procedures must be carried out in <emphasis>each</> database
5092 of an installation, including <literal>template1</>, and ideally
5093 including <literal>template0</> as well. If you do not fix the
5094 template databases then any subsequently created databases will contain
5095 the same errors. <literal>template1</> can be fixed in the same way
5096 as any other database, but fixing <literal>template0</> requires
5097 additional steps. First, from any database issue
5099 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
5101 Next connect to <literal>template0</> and perform the above repair
5102 procedures. Finally, do
5104 -- re-freeze template0:
5106 -- and protect it against future alterations:
5107 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
5113 <title>Changes</title>
5116 Release 7.4.2 incorporates all the fixes included in release 7.3.6,
5117 plus the following fixes:
5121 <listitem><para>Fix <structname>pg_statistics</> alignment bug that could crash optimizer</para>
5122 <para>See above for details about this problem.</para></listitem>
5123 <listitem><para>Allow non-super users to update <structname>pg_settings</></para></listitem>
5124 <listitem><para>Fix several optimizer bugs, most of which led to
5125 <quote>variable not found in subplan target lists</> errors</para></listitem>
5126 <listitem><para>Avoid out-of-memory failure during startup of large multiple
5127 index scan</para></listitem>
5128 <listitem><para>Fix multibyte problem that could lead to <quote>out of
5129 memory</> error during <command>COPY IN</></para></listitem>
5130 <listitem><para>Fix problems with <command>SELECT INTO</> / <command>CREATE
5131 TABLE AS</> from tables without OIDs</para></listitem>
5132 <listitem><para>Fix problems with <filename>alter_table</> regression test
5133 during parallel testing</para></listitem>
5134 <listitem><para>Fix problems with hitting open file limit, especially on OS X (Tom)</para></listitem>
5135 <listitem><para>Partial fix for Turkish-locale issues</para>
5136 <para>initdb will succeed now in Turkish locale, but there are still some
5137 inconveniences associated with the <literal>i/I</> problem.</para></listitem>
5138 <listitem><para>Make pg_dump set client encoding on restore</para></listitem>
5139 <listitem><para>Other minor pg_dump fixes</para></listitem>
5140 <listitem><para>Allow ecpg to again use C keywords as column names (Michael)</para></listitem>
5141 <listitem><para>Added ecpg <literal>WHENEVER NOT_FOUND</> to
5142 <literal>SELECT/INSERT/UPDATE/DELETE</> (Michael)</para></listitem>
5143 <listitem><para>Fix ecpg crash for queries calling set-returning functions (Michael)</para></listitem>
5144 <listitem><para>Various other ecpg fixes (Michael)</para></listitem>
5145 <listitem><para>Fixes for Borland compiler</para></listitem>
5146 <listitem><para>Thread build improvements (Bruce)</para></listitem>
5147 <listitem><para>Various other build fixes</para></listitem>
5148 <listitem><para>Various JDBC fixes</para></listitem>
5154 <sect1 id="release-7-4-1">
5155 <title>Release 7.4.1</title>
5158 <title>Release date</title>
5159 <simpara>2003-12-22</simpara>
5163 This release contains a variety of fixes from 7.4.
5168 <title>Migration to version 7.4.1</title>
5171 A dump/restore is <emphasis>not</emphasis> required for those
5176 If you want to install the fixes in the information schema
5177 you need to reload it into the database.
5178 This is either accomplished by initializing a new cluster
5179 by running <command>initdb</command>, or by running the following
5180 sequence of SQL commands in each database (ideally including
5181 <literal>template1</literal>) as a superuser in
5182 <application>psql</application>, after installing the new release:
5184 DROP SCHEMA information_schema CASCADE;
5185 \i /usr/local/pgsql/share/information_schema.sql
5187 Substitute your installation path in the second command.
5193 <title>Changes</title>
5196 <listitem><para>Fixed bug in <command>CREATE SCHEMA</command> parsing in ECPG (Michael)</para></listitem>
5197 <listitem><para>Fix compile error when <option>--enable-thread-safety</option> and <option>--with-perl</option> are used together (Peter)</para></listitem>
5198 <listitem><para>Fix for subqueries that used hash joins (Tom)</para>
5200 Certain subqueries that used hash joins would crash because of
5201 improperly shared structures.
5203 <listitem><para>Fix free space map compaction bug (Tom)</para>
5205 This fixes a bug where compaction of the free space map could lead
5206 to a database server shutdown.
5209 <listitem><para>Fix for Borland compiler build of libpq (Bruce)</para></listitem>
5210 <listitem><para>Fix <function>netmask()</function> and <function>hostmask()</function> to return the maximum-length masklen (Tom)</para>
5212 Fix these functions to return values consistent with pre-7.4
5216 <listitem><para>Several <filename>contrib/pg_autovacuum</filename> fixes</para>
5218 Fixes include improper variable initialization, missing vacuum after
5219 <command>TRUNCATE</command>, and duration computation overflow for long vacuums.
5222 <listitem><para>Allow compile of <filename>contrib/cube</filename> under Cygwin (Jason Tishler)</para></listitem>
5223 <listitem><para>Fix Solaris use of password file when no passwords are defined (Tom)</para>
5225 Fix crash on Solaris caused by use of any type of password
5226 authentication when no passwords were defined.
5229 <listitem><para>JDBC fix for thread problems, other fixes</para></listitem>
5230 <listitem><para>Fix for <type>bytea</type> index lookups (Joe)</para></listitem>
5231 <listitem><para>Fix information schema for bit data types (Peter)</para></listitem>
5232 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
5233 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
5234 <listitem><para>Make <function>PQescapeBytea</function> and <function>byteaout</function> consistent with each other (Joe)</para></listitem>
5235 <listitem><para>Escape <type>bytea</type> output for bytes > 0x7e(Joe)</para>
5237 If different client encodings are used for <type>bytea</type> output and input, it
5238 is possible for <type>bytea</type> values to be corrupted by the differing
5239 encodings. This fix escapes all bytes that might be affected.
5242 <listitem><para>Added missing <function>SPI_finish()</function> calls to dblink's <function>get_tuple_of_interest()</function> (Joe)</para></listitem>
5243 <listitem><para>New Czech FAQ</para></listitem>
5244 <listitem><para>Fix information schema view <literal>constraint_column_usage</literal> for foreign keys (Peter)</para></listitem>
5245 <listitem><para>ECPG fixes (Michael)</para></listitem>
5246 <listitem><para>Fix bug with multiple <literal>IN</literal> subqueries and joins in the subqueries (Tom)</para></listitem>
5247 <listitem><para>Alllow <literal>COUNT('x')</literal> to work (Tom)</para></listitem>
5248 <listitem><para>Install ECPG include files for Informix compatibility into separate directory (Peter)</para>
5250 Some names of ECPG include files for Informix compatibility conflicted with operating system include files.
5251 By installing them in their own directory, name conflicts have been reduced.
5254 <listitem><para>Fix SSL memory leak (Neil)</para>
5256 This release fixes a bug in 7.4 where SSL didn't free all memory it allocated.
5259 <listitem><para>Prevent <filename>pg_service.conf</filename> from using service name as default dbname (Bruce)</para></listitem>
5260 <listitem><para>Fix local ident authentication on FreeBSD (Tom)</para></listitem>
5266 <sect1 id="release-7-4">
5267 <title>Release 7.4</title>
5270 <title>Release date</title>
5271 <simpara>2003-11-17</simpara>
5275 <title>Overview</title>
5278 Major changes in this release:
5284 <literal>IN</literal> / <literal>NOT IN</literal> subqueries are
5285 now much more efficient
5290 In previous releases, <literal>IN</literal>/<literal>NOT
5291 IN</literal> subqueries were joined to the upper query by
5292 sequentially scanning the subquery looking for a match. The
5293 7.4 code uses the same sophisticated techniques used by
5294 ordinary joins and so is much faster. An
5295 <literal>IN</literal> will now usually be as fast as or faster
5296 than an equivalent <literal>EXISTS</literal> subquery; this
5297 reverses the conventional wisdom that applied to previous
5305 Improved <literal>GROUP BY</literal> processing by using hash buckets
5310 In previous releases, rows to be grouped had to be sorted
5311 first. The 7.4 code can do <literal>GROUP BY</literal>
5312 without sorting, by accumulating results into a hash table
5313 with one entry per group. It will still use the sort
5314 technique, however, if the hash table is estimated to be too
5315 large to fit in <varname>sort_mem</>.
5322 New multikey hash join capability
5327 In previous releases, hash joins could only occur on single
5328 keys. This release allows multicolumn hash joins.
5335 Queries using the explicit <literal>JOIN</literal> syntax are
5336 now better optimized
5341 Prior releases evaluated queries using the explicit
5342 <literal>JOIN</literal> syntax only in the order implied by
5343 the syntax. 7.4 allows full optimization of these queries,
5344 meaning the optimizer considers all possible join orderings
5345 and chooses the most efficient. Outer joins, however, must
5346 still follow the declared ordering.
5353 Faster and more powerful regular expression code
5358 The entire regular expression module has been replaced with a
5359 new version by Henry Spencer, originally written for Tcl. The
5360 code greatly improves performance and supports several flavors
5361 of regular expressions.
5368 Function-inlining for simple SQL functions
5373 Simple SQL functions can now be inlined by including their SQL
5374 in the main query. This improves performance by eliminating
5375 per-call overhead. That means simple SQL functions now
5383 Full support for IPv6 connections and IPv6 address data types
5388 Previous releases allowed only IPv4 connections, and the IP
5389 data types only supported IPv4 addresses. This release adds
5390 full IPv6 support in both of these areas.
5397 Major improvements in SSL performance and reliability
5402 Several people very familiar with the SSL API have overhauled
5403 our SSL code to improve SSL key negotiation and error
5411 Make free space map efficiently reuse empty index pages,
5412 and other free space management improvements
5417 In previous releases, B-tree index pages that were left empty
5418 because of deleted rows could only be reused by rows with
5419 index values similar to the rows originally indexed on that
5420 page. In 7.4, <command>VACUUM</command> records empty index
5421 pages and allows them to be reused for any future index rows.
5428 SQL-standard information schema
5433 The information schema provides a standardized and stable way
5434 to access information about the schema objects defined in a
5442 Cursors conform more closely to the SQL standard
5447 The commands <command>FETCH</command> and
5448 <command>MOVE</command> have been overhauled to conform more
5449 closely to the SQL standard.
5456 Cursors can exist outside transactions
5461 These cursors are also called holdable cursors.
5468 New client-to-server protocol
5473 The new protocol adds error codes, more status information,
5474 faster startup, better support for binary data transmission,
5475 parameter values separated from SQL commands, prepared
5476 statements available at the protocol level, and cleaner
5477 recovery from <command>COPY</command> failures. The older
5478 protocol is still supported by both server and clients.
5485 <application>libpq</application> and
5486 <application>ECPG</application> applications are now fully
5492 While previous <application>libpq</application> releases
5493 already supported threads, this release improves thread safety
5494 by fixing some non-thread-safe code that was used during
5495 database connection startup. The <command>configure</command>
5496 option <option>--enable-thread-safety</option> must be used to
5497 enable this feature.
5504 New version of full-text indexing
5509 A new full-text indexing suite is available in
5510 <filename>contrib/tsearch2</filename>.
5522 The new autovacuum tool in
5523 <filename>contrib/autovacuum</filename> monitors the database
5524 statistics tables for
5525 <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
5526 activity and automatically vacuums tables when needed.
5533 Array handling has been improved and moved into the server core
5538 Many array limitations have been removed, and arrays behave
5539 more like fully-supported data types.
5548 <title>Migration to version 7.4</title>
5551 A dump/restore using <application>pg_dump</application> is
5552 required for those wishing to migrate data from any previous
5557 Observe the following incompatibilities:
5563 The server-side autocommit setting was removed and
5564 reimplemented in client applications and languages.
5565 Server-side autocommit was causing too many problems with
5566 languages and applications that wanted to control their own
5567 autocommit behavior, so autocommit was removed from the server
5568 and added to individual client APIs as appropriate.
5574 Error message wording has changed substantially in this
5575 release. Significant effort was invested to make the messages
5576 more consistent and user-oriented. If your applications try to
5577 detect different error conditions by parsing the error message,
5578 you are strongly encouraged to use the new error code facility instead.
5584 Inner joins using the explicit <literal>JOIN</literal> syntax
5585 may behave differently because they are now better
5592 A number of server configuration parameters have been renamed
5593 for clarity, primarily those related to
5600 <literal>FETCH 0</literal> or <literal>MOVE 0</literal> now
5601 does nothing. In prior releases, <literal>FETCH 0</literal>
5602 would fetch all remaining rows, and <literal>MOVE 0</literal>
5603 would move to the end of the cursor.
5609 <command>FETCH</command> and <command>MOVE</command> now return
5610 the actual number of rows fetched/moved, or zero if at the
5611 beginning/end of the cursor. Prior releases would return the
5612 row count passed to the command, not the number of rows
5613 actually fetched or moved.
5619 <command>COPY</command> now can process files that use
5620 carriage-return or carriage-return/line-feed end-of-line
5621 sequences. Literal carriage-returns and line-feeds are no
5622 longer accepted in data values; use <literal>\r</literal> and
5623 <literal>\n</literal> instead.
5629 Trailing spaces are now trimmed when converting from type
5630 <type>char(<replaceable>n</>)</type> to
5631 <type>varchar(<replaceable>n</>)</type> or <type>text</type>.
5632 This is what most people always expected to happen anyway.
5638 The data type <type>float(<replaceable>p</>)</type> now
5639 measures <replaceable>p</> in binary digits, not decimal
5640 digits. The new behavior follows the SQL standard.
5646 Ambiguous date values now must match the ordering specified by
5647 the <varname>datestyle</varname> setting. In prior releases, a
5648 date specification of <literal>10/20/03</> was interpreted as a
5649 date in October even if <varname>datestyle</> specified that
5650 the day should be first. 7.4 will throw an error if a date
5651 specification is invalid for the current setting of
5652 <varname>datestyle</>.
5658 The functions <function>oidrand</function>,
5659 <function>oidsrand</function>, and
5660 <function>userfntest</function> have been removed. These
5661 functions were determined to be no longer useful.
5667 String literals specifying time-varying date/time values, such
5668 as <literal>'now'</literal> or <literal>'today'</literal> will
5669 no longer work as expected in column default expressions; they
5670 now cause the time of the table creation to be the default, not
5671 the time of the insertion. Functions such as
5672 <function>now()</>, <function>current_timestamp</>, or
5673 <function>current_date</function> should be used instead.
5677 In previous releases, there was special code so that strings
5678 such as <literal>'now'</literal> were interpreted at
5679 <command>INSERT</> time and not at table creation time, but
5680 this work around didn't cover all cases. Release 7.4 now
5681 requires that defaults be defined properly using functions such
5682 as <function>now()</> or <function>current_timestamp</>. These
5683 will work in all situations.
5689 The dollar sign (<literal>$</>) is no longer allowed in
5690 operator names. It can instead be a non-first character in
5691 identifiers. This was done to improve compatibility with other
5692 database systems, and to avoid syntax problems when parameter
5693 placeholders (<literal>$<replaceable>n</></>) are written
5694 adjacent to operators.
5702 <title>Changes</title>
5705 Below you will find a detailed account of the changes between
5706 release 7.4 and the previous major release.
5710 <title>Server Operation Changes</title>
5715 Allow IPv6 server connections (Nigel Kukard, Johan Jordaan,
5716 Bruce, Tom, Kurt Roeckx, Andrew Dunstan)
5722 Fix SSL to handle errors cleanly (Nathan Mueller)
5725 In prior releases, certain SSL API error reports were not
5726 handled correctly. This release fixes those problems.
5732 SSL protocol security and performance improvements (Sean Chittenden)
5735 SSL key renegotiation was happening too frequently, causing poor
5736 SSL performance. Also, initial key handling was improved.
5742 Print lock information when a deadlock is detected (Tom)
5745 This allows easier debugging of deadlock situations.
5751 Update <filename>/tmp</filename> socket modification times
5752 regularly to avoid their removal (Tom)
5755 This should help prevent <filename>/tmp</filename> directory
5756 cleaner administration scripts from removing server socket
5761 <listitem><para>Enable PAM for Mac OS X (Aaron Hillegass)</para></listitem>
5764 <para>Make B-tree indexes fully WAL-safe (Tom)</para>
5766 In prior releases, under certain rare cases, a server crash
5767 could cause B-tree indexes to become corrupt. This release
5768 removes those last few rare cases.
5772 <listitem><para>Allow B-tree index compaction and empty page reuse (Tom)</para></listitem>
5776 Fix inconsistent index lookups during split of first root page (Tom)
5779 In prior releases, when a single-page index split into two
5780 pages, there was a brief period when another database session
5781 could miss seeing an index entry. This release fixes that rare
5786 <listitem><para>Improve free space map allocation logic (Tom)</para></listitem>
5789 <para>Preserve free space information between server restarts (Tom)</para>
5791 In prior releases, the free space map was not saved when the
5792 postmaster was stopped, so newly started servers had no free
5793 space information. This release saves the free space map, and
5794 reloads it when the server is restarted.
5798 <listitem><para>Add start time to <literal>pg_stat_activity</literal> (Neil)</para></listitem>
5799 <listitem><para>New code to detect corrupt disk pages; erase with <varname>zero_damaged_pages</varname> (Tom)</para></listitem>
5800 <listitem><para>New client/server protocol: faster, no username length limit, allow clean exit from <command>COPY</command> (Tom)</para></listitem>
5801 <listitem><para>Add transaction status, table ID, column ID to client/server protocol (Tom)</para></listitem>
5802 <listitem><para>Add binary I/O to client/server protocol (Tom)</para></listitem>
5803 <listitem><para>Remove autocommit server setting; move to client applications (Tom)</para></listitem>
5804 <listitem><para>New error message wording, error codes, and three levels of error detail (Tom, Joe, Peter)</para></listitem>
5809 <title>Performance Improvements</title>
5812 <listitem><para>Add hashing for <literal>GROUP BY</literal> aggregates (Tom)</para></listitem>
5813 <listitem><para>Make nested-loop joins be smarter about multicolumn indexes (Tom)</para></listitem>
5814 <listitem><para>Allow multikey hash joins (Tom)</para></listitem>
5815 <listitem><para>Improve constant folding (Tom)</para></listitem>
5816 <listitem><para>Add ability to inline simple SQL functions (Tom)</para></listitem>
5819 <para>Reduce memory usage for queries using complex functions (Tom)</para>
5821 In prior releases, functions returning allocated memory would
5822 not free it until the query completed. This release allows the
5823 freeing of function-allocated memory when the function call
5824 completes, reducing the total memory used by functions.
5829 <para>Improve GEQO optimizer performance (Tom)</para>
5831 This release fixes several inefficiencies in the way the GEQO optimizer
5832 manages potential query paths.
5838 Allow <literal>IN</>/<literal>NOT IN</> to be handled via hash
5845 Improve <literal>NOT IN (<replaceable>subquery</>)</literal>
5852 Allow most <literal>IN</literal> subqueries to be processed as
5859 Pattern matching operations can use indexes regardless of
5863 There is no way for non-ASCII locales to use the standard
5864 indexes for <literal>LIKE</literal> comparisons. This release
5865 adds a way to create a special index for
5866 <literal>LIKE</literal>.
5871 <para>Allow the postmaster to preload libraries using <varname>preload_libraries</varname> (Joe)</para>
5873 For shared libraries that require a long time to load, this
5874 option is available so the library can be preloaded in the
5875 postmaster and inherited by all database sessions.
5881 Improve optimizer cost computations, particularly for subqueries (Tom)
5887 Avoid sort when subquery <literal>ORDER BY</literal> matches upper query (Tom)
5893 Deduce that <literal>WHERE a.x = b.y AND b.y = 42</literal> also
5894 means <literal>a.x = 42</literal> (Tom)
5900 Allow hash/merge joins on complex joins (Tom)
5906 Allow hash joins for more data types (Tom)
5912 Allow join optimization of explicit inner joins, disable with
5913 <varname>join_collapse_limit</varname> (Tom)
5919 Add parameter <varname>from_collapse_limit</varname> to control
5920 conversion of subqueries to joins (Tom)
5926 Use faster and more powerful regular expression code from Tcl
5927 (Henry Spencer, Tom)
5933 Use bit-mapped relation sets in the optimizer (Tom)
5938 <para>Improve connection startup time (Tom)</para>
5940 The new client/server protocol requires fewer network packets to
5941 start a database session.
5947 Improve trigger/constraint performance (Stephan)
5953 Improve speed of <literal>col IN (const, const, const, ...)</literal> (Tom)
5959 Fix hash indexes which were broken in rare cases (Tom)
5963 <listitem><para>Improve hash index concurrency and speed (Tom)</para>
5965 Prior releases suffered from poor hash index performance,
5966 particularly for high concurrency situations. This release fixes
5967 that, and the development group is interested in reports
5968 comparing B-tree and hash index performance.
5973 <para>Align shared buffers on 32-byte boundary for copy speed improvement (Manfred Spraul)</para>
5975 Certain CPU's perform faster data copies when addresses are
5981 <para>Data type <type>numeric</type> reimplemented for better performance (Tom)</para>
5983 <type>numeric</type> used to be stored in base 100. The new code
5984 uses base 10000, for significantly better performance.
5991 <title>Server Configuration Changes</title>
5995 <para>Rename server parameter <varname>server_min_messages</> to <varname>log_min_messages</> (Bruce)</para>
5997 This was done so most parameters that control the server logs
5998 begin with <literal>log_</>.
6002 <listitem><para>Rename <varname>show_*_stats</> to <varname>log_*_stats</> (Bruce)</para></listitem>
6003 <listitem><para>Rename <varname>show_source_port</> to <varname>log_source_port</> (Bruce)</para></listitem>
6004 <listitem><para>Rename <varname>hostname_lookup</> to <varname>log_hostname</> (Bruce)</para></listitem>
6007 <para>Add <varname>checkpoint_warning</> to warn of excessive checkpointing (Bruce)</para>
6009 In prior releases, it was difficult to determine if checkpoint
6010 was happening too frequently. This feature adds a warning to the
6011 server logs when excessive checkpointing happens.
6015 <listitem><para>New read-only server parameters for localization (Tom)</para></listitem>
6019 Change debug server log messages to output as <literal>DEBUG</>
6020 rather than <literal>LOG</> (Bruce)
6025 <para>Prevent server log variables from being turned off by non-superusers (Bruce)</para>
6027 This is a security feature so non-superusers cannot disable
6028 logging that was enabled by the administrator.
6034 <varname>log_min_messages</>/<varname>client_min_messages</> now
6035 controls <varname>debug_*</> output (Bruce)
6038 This centralizes client debug information so all debug output
6039 can be sent to either the client or server logs.
6044 <para>Add Mac OS X Rendezvous server support (Chris Campbell)</para>
6046 This allows Mac OS X hosts to query the network for available
6047 <productname>PostgreSQL</productname> servers.
6053 Add ability to print only slow statements using
6054 <varname>log_min_duration_statement</varname>
6058 This is an often requested debugging feature that allows
6059 administrators to see only slow queries in their server logs.
6064 <para>Allow <filename>pg_hba.conf</filename> to accept netmasks in CIDR format (Andrew Dunstan)</para>
6066 This allows administrators to merge the host IP address and
6067 netmask fields into a single CIDR field in <filename>pg_hba.conf</filename>.
6071 <listitem><para>New read-only parameter <varname>is_superuser</varname> (Tom)</para></listitem>
6074 <para>New parameter <varname>log_error_verbosity</varname> to control error detail (Tom)</para>
6076 This works with the new error reporting feature to supply
6077 additional error information like hints, file names and line
6083 <para><literal>postgres --describe-config</literal> now dumps server config variables (Aizaz Ahmed, Peter)</para>
6085 This option is useful for administration tools that need to know
6086 the configuration variable names and their minimums, maximums,
6087 defaults, and descriptions.
6093 Add new columns in <literal>pg_settings</literal>:
6094 <literal>context</>, <literal>type</>, <literal>source</>,
6095 <literal>min_val</>, <literal>max_val</> (Joe)
6101 Make default <varname>shared_buffers</> 1000 and
6102 <varname>max_connections</> 100, if possible (Tom)
6105 Prior versions defaulted to 64 shared buffers so <productname>PostgreSQL</productname>
6106 would start on even very old systems. This release tests the
6107 amount of shared memory allowed by the platform and selects more
6108 reasonable default values if possible. Of course, users are
6109 still encouraged to evaluate their resource load and size
6110 <varname>shared_buffers</varname> accordingly.
6116 New <filename>pg_hba.conf</filename> record type
6117 <literal>hostnossl</> to prevent SSL connections (Jon
6121 In prior releases, there was no way to prevent SSL connections
6122 if both the client and server supported SSL. This option allows
6129 Remove parameter <varname>geqo_random_seed</varname>
6136 Add server parameter <varname>regex_flavor</varname> to control regular expression processing (Tom)
6142 Make <command>pg_ctl</command> better handle nonstandard ports (Greg)
6149 <title>Query Changes</title>
6152 <listitem><para>New SQL-standard information schema (Peter)</para></listitem>
6153 <listitem><para>Add read-only transactions (Peter)</para></listitem>
6154 <listitem><para>Print key name and value in foreign-key violation messages (Dmitry Tkach)</para></listitem>
6157 <para>Allow users to see their own queries in <literal>pg_stat_activity</literal> (Kevin Brown)</para>
6159 In prior releases, only the superuser could see query strings
6160 using <literal>pg_stat_activity</literal>. Now ordinary users
6161 can see their own query strings.
6166 <para>Fix aggregates in subqueries to match SQL standard (Tom)</para>
6168 The SQL standard says that an aggregate function appearing
6169 within a nested subquery belongs to the outer query if its
6170 argument contains only outer-query variables. Prior
6171 <productname>PostgreSQL</productname> releases did not handle
6172 this fine point correctly.
6177 <para>Add option to prevent auto-addition of tables referenced in query (Nigel J. Andrews)</para>
6179 By default, tables mentioned in the query are automatically
6180 added to the <literal>FROM</> clause if they are not already
6181 there. This is compatible with historic
6182 <productname>POSTGRES</productname> behavior but is contrary to
6183 the SQL standard. This option allows selecting
6184 standard-compatible behavior.
6189 <para>Allow <literal>UPDATE ... SET col = DEFAULT</literal> (Rod)</para>
6191 This allows <command>UPDATE</command> to set a column to its
6192 declared default value.
6197 <para>Allow expressions to be used in <literal>LIMIT</>/<literal>OFFSET</> (Tom)</para>
6199 In prior releases, <literal>LIMIT</>/<literal>OFFSET</> could
6200 only use constants, not expressions.
6205 <para>Implement <literal>CREATE TABLE AS EXECUTE</literal> (Neil, Peter)</para>
6211 <title>Object Manipulation Changes</title>
6215 <para>Make <command>CREATE SEQUENCE</command> grammar more conforming to SQL:2003 (Neil)</para>
6219 <para>Add statement-level triggers (Neil)</para>
6221 While this allows a trigger to fire at the end of a statement,
6222 it does not allow the trigger to access all rows modified by the
6223 statement. This capability is planned for a future release.
6228 <para>Add check constraints for domains (Rod)</para>
6230 This greatly increases the usefulness of domains by allowing
6231 them to use check constraints.
6236 <para>Add <command>ALTER DOMAIN</command> (Rod)</para>
6238 This allows manipulation of existing domains.
6243 <para>Fix several zero-column table bugs (Tom)</para>
6245 <productname>PostgreSQL</productname> supports zero-column tables. This fixes various bugs
6246 that occur when using such tables.
6251 <para>Have <literal>ALTER TABLE ... ADD PRIMARY KEY</literal> add not-null constraint (Rod)</para>
6253 In prior releases, <literal>ALTER TABLE ... ADD
6254 PRIMARY</literal> would add a unique index, but not a not-null
6255 constraint. That is fixed in this release.
6259 <listitem><para>Add <literal>ALTER TABLE ... WITHOUT OIDS</literal> (Rod)</para>
6261 This allows control over whether new and updated rows will have
6262 an OID column. This is most useful for saving storage space.
6268 Add <literal>ALTER SEQUENCE</literal> to modify minimum, maximum,
6269 increment, cache, cycle values (Rod)
6274 <para>Add <literal>ALTER TABLE ... CLUSTER ON</literal> (Alvaro Herrera)</para>
6276 This command is used by <command>pg_dump</command> to record the
6277 cluster column for each table previously clustered. This
6278 information is used by database-wide cluster to cluster all
6279 previously clustered tables.
6283 <listitem><para>Improve automatic type casting for domains (Rod, Tom)</para></listitem>
6284 <listitem><para>Allow dollar signs in identifiers, except as first character (Tom)</para></listitem>
6285 <listitem><para>Disallow dollar signs in operator names, so <literal>x=$1</> works (Tom)</para></listitem>
6289 Allow copying table schema using <literal>LIKE
6290 <replaceable>subtable</replaceable></literal>, also SQL:2003
6291 feature <literal>INCLUDING DEFAULTS</literal> (Rod)
6297 Add <literal>WITH GRANT OPTION</literal> clause to
6298 <command>GRANT</command> (Peter)
6301 This enabled <command>GRANT</command> to give other users the
6302 ability to grant privileges on a object.
6309 <title>Utility Command Changes</title>
6313 <para>Add <literal>ON COMMIT</literal> clause to <command>CREATE TABLE</command> for temporary tables (Gavin)</para>
6315 This adds the ability for a table to be dropped or all rows
6316 deleted on transaction commit.
6321 <para>Allow cursors outside transactions using <literal>WITH HOLD</literal> (Neil)</para>
6323 In previous releases, cursors were removed at the end of the
6324 transaction that created them. Cursors can now be created with
6325 the <literal>WITH HOLD</literal> option, which allows them to
6326 continue to be accessed after the creating transaction has
6332 <para><literal>FETCH 0</literal> and <literal>MOVE 0 </literal> now do nothing (Bruce)</para>
6334 In previous releases, <literal>FETCH 0</literal> fetched all
6335 remaining rows, and <literal>MOVE 0</literal> moved to the end
6342 Cause <command>FETCH</command> and <command>MOVE</command> to
6343 return the number of rows fetched/moved, or zero if at the
6344 beginning/end of cursor, per SQL standard (Bruce)
6347 In prior releases, the row count returned by
6348 <command>FETCH</command> and <command>MOVE</command> did not
6349 accurately reflect the number of rows processed.
6354 <para>Properly handle <literal>SCROLL</literal> with cursors, or
6355 report an error (Neil)</para>
6357 Allowing random access (both forward and backward scrolling) to
6358 some kinds of queries cannot be done without some additional
6359 work. If <literal>SCROLL</literal> is specified when the cursor
6360 is created, this additional work will be performed. Furthermore,
6361 if the cursor has been created with <literal>NO SCROLL</literal>,
6362 no random access is allowed.
6368 Implement SQL-compatible options <literal>FIRST</>,
6369 <literal>LAST</>, <literal>ABSOLUTE <replaceable>n</></>,
6370 <literal>RELATIVE <replaceable>n</></> for
6371 <command>FETCH</command> and <command>MOVE</command> (Tom)
6376 <para>Allow <command>EXPLAIN</command> on <command>DECLARE CURSOR</command> (Tom)</para>
6380 <para>Allow <command>CLUSTER</command> to use index marked as pre-clustered by default (Alvaro Herrera)</para>
6384 <para>Allow <command>CLUSTER</command> to cluster all tables (Alvaro Herrera)</para>
6386 This allows all previously clustered tables in a database to be
6387 reclustered with a single command.
6391 <listitem><para>Prevent <command>CLUSTER</command> on partial indexes (Tom)</para></listitem>
6393 <listitem><para>Allow DOS and Mac line-endings in <command>COPY</> files (Bruce)</para></listitem>
6397 Disallow literal carriage return as a data value,
6398 backslash-carriage-return and <literal>\r</> are still allowed
6404 <para><command>COPY</> changes (binary, <literal>\.</>) (Tom)</para>
6408 <para>Recover from <command>COPY</command> failure cleanly (Tom)</para>
6412 <para>Prevent possible memory leaks in <command>COPY</command> (Tom)</para>
6416 <para>Make <command>TRUNCATE</command> transaction-safe (Rod)</para>
6418 <command>TRUNCATE</command> can now be used inside a
6419 transaction. If the transaction aborts, the changes made by the
6420 <command>TRUNCATE</command> are automatically rolled back.
6426 Allow prepare/bind of utility commands like
6427 <command>FETCH</command> and <command>EXPLAIN</command> (Tom)
6432 <para>Add <command>EXPLAIN EXECUTE</command> (Neil)</para>
6436 <para>Improve <command>VACUUM</command> performance on indexes by reducing WAL traffic (Tom)</para>
6440 <para>Functional indexes have been generalized into indexes on expressions (Tom)</para>
6442 In prior releases, functional indexes only supported a simple
6443 function applied to one or more column names. This release
6444 allows any type of scalar expression.
6450 Have <command>SHOW TRANSACTION ISOLATION</command> match input
6451 to <command>SET TRANSACTION ISOLATION</command>
6458 Have <command>COMMENT ON DATABASE</command> on nonlocal
6459 database generate a warning, rather than an error (Rod)
6463 Database comments are stored in database-local tables so
6464 comments on a database have to be stored in each database.
6470 Improve reliability of <command>LISTEN</>/<command>NOTIFY</> (Tom)
6475 <para>Allow <command>REINDEX</command> to reliably reindex nonshared system catalog indexes (Tom)</para>
6477 This allows system tables to be reindexed without the
6478 requirement of a standalone session, which was necessary in
6479 previous releases. The only tables that now require a standalone
6480 session for reindexing are the global system tables
6481 <literal>pg_database</>, <literal>pg_shadow</>, and
6482 <literal>pg_group</>.
6489 <title>Data Type and Function Changes</title>
6494 New server parameter <varname>extra_float_digits</varname> to
6495 control precision display of floating-point numbers (Pedro
6499 This controls output precision which was causing regression
6504 <listitem><para>Allow <literal>+1300</literal> as a numeric time-zone specifier, for FJST (Tom)</para></listitem>
6508 Remove rarely used functions <function>oidrand</>,
6509 <function>oidsrand</>, and <function>userfntest</> functions
6515 <para>Add <function>md5()</> function to main server, already in <filename>contrib/pgcrypto</filename> (Joe)</para>
6517 An MD5 function was frequently requested. For more complex
6518 encryption capabilities, use
6519 <filename>contrib/pgcrypto</filename>.
6523 <listitem><para>Increase date range of <type>timestamp</type> (John Cochran)</para></listitem>
6527 Change <literal>EXTRACT(EPOCH FROM timestamp)</literal> so
6528 <type>timestamp without time zone</type> is assumed to be in
6529 local time, not GMT (Tom)
6533 <listitem><para>Trap division by zero in case the operating system doesn't prevent it (Tom)</para></listitem>
6534 <listitem><para>Change the <type>numeric</type> data type internally to base 10000 (Tom)</para></listitem>
6535 <listitem><para>New <function>hostmask()</function> function (Greg Wickham)</para></listitem>
6536 <listitem><para>Fixes for <function>to_char()</function> and <function>to_timestamp()</function> (Karel)</para></listitem>
6540 Allow functions that can take any argument data type and return
6541 any data type, using <type>anyelement</type> and
6542 <type>anyarray</type> (Joe)
6545 This allows the creation of functions that can work with any
6552 Arrays may now be specified as <literal>ARRAY[1,2,3]</literal>,
6553 <literal>ARRAY[['a','b'],['c','d']]</literal>, or
6554 <literal>ARRAY[ARRAY[ARRAY[2]]]</literal> (Joe)
6560 Allow proper comparisons for arrays, including <literal>ORDER
6561 BY</literal> and <literal>DISTINCT</literal> support
6566 <listitem><para>Allow indexes on array columns (Joe)</para></listitem>
6567 <listitem><para>Allow array concatenation with <literal>||</literal> (Joe)</para></listitem>
6571 Allow <literal>WHERE</literal> qualification
6572 <literal><replaceable>expr</> <replaceable>op</> ANY/SOME/ALL
6573 (<replaceable>array_expr</>)</literal> (Joe)
6576 This allows arrays to behave like a list of values, for purposes
6577 like <literal>SELECT * FROM tab WHERE col IN
6578 (array_val)</literal>.
6584 New array functions <function>array_append</>,
6585 <function>array_cat</>, <function>array_lower</>,
6586 <function>array_prepend</>, <function>array_to_string</>,
6587 <function>array_upper</>, <function>string_to_array</> (Joe)
6591 <listitem><para>Allow user defined aggregates to use polymorphic functions (Joe)</para></listitem>
6592 <listitem><para>Allow assignments to empty arrays (Joe)</para></listitem>
6596 Allow 60 in seconds fields of <type>time</type>,
6597 <type>timestamp</type>, and <type>interval</type> input values
6601 Sixty-second values are needed for leap seconds.
6605 <listitem><para>Allow <type>cidr</type> data type to be cast to <type>text</type> (Tom)</para></listitem>
6607 <listitem><para>Disallow invalid time zone names in SET TIMEZONE</para></listitem>
6611 Trim trailing spaces when <type>char</type> is cast to
6612 <type>varchar</> or <type>text</> (Tom)
6618 Make <type>float(<replaceable>p</>)</> measure the precision
6619 <replaceable>p</> in binary digits, not decimal digits
6625 <para>Add IPv6 support to the <type>inet</type> and <type>cidr</type> data types (Michael Graff)</para>
6629 <para>Add <function>family()</function> function to report whether address is IPv4 or IPv6 (Michael Graff)</para>
6634 Have <literal>SHOW datestyle</literal> generate output similar
6635 to that used by <literal>SET datestyle</literal> (Tom)
6641 Make <literal>EXTRACT(TIMEZONE)</literal> and <literal>SET/SHOW
6642 TIME ZONE</literal> follow the SQL convention for the sign of
6643 time zone offsets, i.e., positive is east from UTC (Tom)
6648 <para>Fix <literal>date_trunc('quarter', ...)</literal> (Böjthe Zoltán)</para>
6650 Prior releases returned an incorrect value for this function call.
6655 <para>Make <function>initcap()</function> more compatible with Oracle (Mike Nolan)</para>
6657 <function>initcap()</function> now uppercases a letter appearing
6658 after any non-alphanumeric character, rather than only after
6664 <para>Allow only <varname>datestyle</varname> field order for date values not in ISO-8601 format (Greg)</para>
6669 Add new <varname>datestyle</varname> values <literal>MDY</>,
6670 <literal>DMY</>, and <literal>YMD</> to set input field order;
6671 honor <literal>US</> and <literal>European</> for backward
6678 String literals like <literal>'now'</literal> or
6679 <literal>'today'</literal> will no longer work as a column
6680 default. Use functions such as <function>now()</function>,
6681 <function>current_timestamp</function> instead. (change
6682 required for prepared statements) (Tom)
6687 <para>Treat NaN as larger than any other value in <function>min()</>/<function>max()</> (Tom)</para>
6689 NaN was already sorted after ordinary numeric values for most
6690 purposes, but <function>min()</> and <function>max()</> didn't
6696 <para>Prevent interval from suppressing <literal>:00</literal>
6697 seconds display</para>
6702 New functions <function>pg_get_triggerdef(prettyprint)</function>
6703 and <function>pg_conversion_is_visible()</function> (Christopher)
6708 <para>Allow time to be specified as <literal>040506</> or <literal>0405</> (Tom)</para>
6713 Input date order must now be <literal>YYYY-MM-DD</literal> (with 4-digit year) or
6714 match <varname>datestyle</varname>
6720 Make <function>pg_get_constraintdef</function> support
6721 unique, primary-key, and check constraints (Christopher)
6728 <title>Server-Side Language Changes</title>
6733 Prevent PL/pgSQL crash when <literal>RETURN NEXT</literal> is
6734 used on a zero-row record variable (Tom)
6740 Make PL/Python's <function>spi_execute</function> interface
6741 handle null values properly (Andrew Bosma)
6746 <para>Allow PL/pgSQL to declare variables of composite types without <literal>%ROWTYPE</literal> (Tom)</para>
6750 <para>Fix PL/Python's <function>_quote()</function> function to handle big integers</para>
6754 <para>Make PL/Python an untrusted language, now called <literal>plpythonu</literal> (Kevin Jacobs, Tom)</para>
6756 The Python language no longer supports a restricted execution
6757 environment, so the trusted version of PL/Python was removed. If
6758 this situation changes, a version of PL/Python that can be used
6759 by non-superusers will be readded.
6764 <para>Allow polymorphic PL/pgSQL functions (Joe, Tom)</para>
6768 <para>Allow polymorphic SQL functions (Joe)</para>
6773 Improved compiled function caching mechanism in PL/pgSQL with
6774 full support for polymorphism (Joe)
6780 Add new parameter <literal>$0</> in PL/pgSQL representing the
6781 function's actual return type (Joe)
6787 Allow PL/Tcl and PL/Python to use the same trigger on multiple tables (Tom)
6793 Fixed PL/Tcl's <function>spi_prepare</function> to accept fully
6794 qualified type names in the parameter type list
6802 <title>psql Changes</title>
6806 <para>Add <literal>\pset pager always</literal> to always use pager (Greg)</para>
6808 This forces the pager to be used even if the number of rows is
6809 less than the screen height. This is valuable for rows that
6810 wrap across several screen rows.
6814 <listitem><para>Improve tab completion (Rod, Ross Reedstrom, Ian Barwick)</para></listitem>
6815 <listitem><para>Reorder <literal>\?</> help into groupings (Harald Armin Massa, Bruce)</para></listitem>
6816 <listitem><para>Add backslash commands for listing schemas, casts, and conversions (Christopher)</para></listitem>
6820 <command>\encoding</> now changes based on the server parameter
6821 <varname>client_encoding</varname> (Tom)
6824 In previous versions, <command>\encoding</command> was not aware
6825 of encoding changes made using <literal>SET
6826 client_encoding</literal>.
6831 <para>Save editor buffer into readline history (Ross)</para>
6833 When <command>\e</> is used to edit a query, the result is saved
6834 in the readline history for retrieval using the up arrow.
6838 <listitem><para>Improve <command>\d</command> display (Christopher)</para></listitem>
6839 <listitem><para>Enhance HTML mode to be more standards-conforming (Greg)</para></listitem>
6842 <para>New <command>\set AUTOCOMMIT off</command> capability (Tom)</para>
6844 This takes the place of the removed server parameter <varname>autocommit</varname>.
6849 <para>New <command>\set VERBOSITY</command> to control error detail (Tom)</para>
6851 This controls the new error reporting details.
6855 <listitem><para>New prompt escape sequence <literal>%x</literal> to show transaction status (Tom)</para></listitem>
6856 <listitem><para>Long options for <application>psql</application> are now available on all platforms</para></listitem>
6861 <title>pg_dump Changes</title>
6864 <listitem><para>Multiple pg_dump fixes, including tar format and large objects</para></listitem>
6865 <listitem><para>Allow pg_dump to dump specific schemas (Neil)</para></listitem>
6868 <para>Make pg_dump preserve column storage characteristics (Christopher)</para>
6870 This preserves <literal>ALTER TABLE ... SET STORAGE</literal> information.
6874 <listitem><para>Make pg_dump preserve <command>CLUSTER</command> characteristics (Christopher)</para></listitem>
6878 Have pg_dumpall use <command>GRANT</>/<command>REVOKE</> to dump database-level privleges (Tom)
6884 Allow pg_dumpall to support the options <option>-a</>,
6885 <option>-s</>, <option>-x</> of pg_dump (Tom)
6889 <listitem><para>Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)</para></listitem>
6893 pg_dump options <option>--use-set-session-authorization</option>
6894 and <option>--no-reconnect</option> now do nothing, all dumps
6895 use <command>SET SESSION AUTHORIZATION</command>
6898 pg_dump no longer reconnects to switch users, but instead always
6899 uses <command>SET SESSION AUTHORIZATION</command>. This will
6900 reduce password prompting during restores.
6905 <para>Long options for <application>pg_dump</application> are now available on all platforms</para>
6907 <productname>PostgreSQL</productname> now includes its own
6908 long-option processing routines.
6915 <title>libpq Changes</title>
6920 Add function <function>PQfreemem</function> for freeing memory on
6921 Windows, suggested for <command>NOTIFY</command> (Bruce)
6924 Windows requires that memory allocated in a library be freed by
6925 a function in the same library, hence
6926 <function>free()</function> doesn't work for freeing memory
6927 allocated by libpq. <function>PQfreemem</function> is the proper
6928 way to free libpq memory, especially on Windows, and is
6929 recommended for other platforms as well.
6934 <para>Document service capability, and add sample file (Bruce)</para>
6936 This allows clients to look up connection information in a
6937 central file on the client machine.
6943 Make <function>PQsetdbLogin</function> have the same defaults as
6944 <function>PQconnectdb</function> (Tom)
6948 <listitem><para>Allow libpq to cleanly fail when result sets are too large (Tom)</para></listitem>
6952 Improve performance of function <function>PQunescapeBytea</function> (Ben Lamb)
6958 Allow thread-safe libpq with <filename>configure</filename>
6959 option <option>--enable-thread-safety</option> (Lee Kindness,
6966 Allow function <function>pqInternalNotice</function> to accept a
6967 format string and arguments instead of just a preformatted
6968 message (Tom, Sean Chittenden)
6974 Control SSL negotiation with <literal>sslmode</literal> values
6975 <literal>disable</literal>, <literal>allow</literal>,
6976 <literal>prefer</literal>, and <literal>require</literal> (Jon
6982 <para>Allow new error codes and levels of text (Tom)</para>
6986 <para>Allow access to the underlying table and column of a query result (Tom)</para>
6988 This is helpful for query-builder applications that want to know
6989 the underlying table and column names associated with a specific
6994 <listitem><para>Allow access to the current transaction status (Tom)</para></listitem>
6995 <listitem><para>Add ability to pass binary data directly to the server (Tom)</para></listitem>
6999 Add function <function>PQexecPrepared</function> and
7000 <function>PQsendQueryPrepared</function> functions which perform
7001 bind/execute of previously prepared statements (Tom)
7008 <title>JDBC Changes</title>
7011 <listitem><para>Allow <function>setNull</function> on updateable result sets</para></listitem>
7012 <listitem><para>Allow <function>executeBatch</function> on a prepared statement (Barry)</para></listitem>
7013 <listitem><para>Support SSL connections (Barry)</para></listitem>
7014 <listitem><para>Handle schema names in result sets (Paul Sorenson)</para></listitem>
7015 <listitem><para>Add refcursor support (Nic Ferrier)</para></listitem>
7020 <title>Miscellaneous Interface Changes</title>
7024 <para>Prevent possible memory leak or core dump during libpgtcl shutdown (Tom)</para>
7027 <para>Add Informix compatibility to ECPG (Michael)</para>
7029 This allows ECPG to process embedded C programs that were
7030 written using certain Informix extensions.
7035 <para>Add type <type>decimal</type> to ECPG that is fixed length, for Informix (Michael)</para>
7040 Allow thread-safe embedded SQL programs with
7041 <filename>configure</filename> option
7042 <option>--enable-thread-safety</option> (Lee Kindness, Bruce)
7045 This allows multiple threads to access the database at the same
7051 <para>Moved Python client PyGreSQL to <ulink url="http://www.pygresql.org"></ulink> (Marc)</para>
7057 <title>Source Code Changes</title>
7060 <listitem><para>Prevent need for separate platform geometry regression result files (Tom)</para></listitem>
7061 <listitem><para>Improved PPC locking primitive (Reinhard Max)</para></listitem>
7062 <listitem><para>New function <function>palloc0</function> to allocate and clear memory (Bruce)</para></listitem>
7063 <listitem><para>Fix locking code for s390x CPU (64-bit) (Tom)</para></listitem>
7064 <listitem><para>Allow OpenBSD to use local ident credentials (William Ahern)</para></listitem>
7065 <listitem><para>Make query plan trees read-only to executor (Tom)</para></listitem>
7066 <listitem><para>Add Darwin startup scripts (David Wheeler)</para></listitem>
7067 <listitem><para>Allow libpq to compile with Borland C++ compiler (Lester Godwin, Karl Waclawek)</para></listitem>
7068 <listitem><para>Use our own version of <function>getopt_long()</function> if needed (Peter)</para></listitem>
7069 <listitem><para>Convert administration scripts to C (Peter)</para></listitem>
7070 <listitem><para> Bison >= 1.85 is now required to build the <productname>PostgreSQL</> grammar, if building from CVS</para></listitem>
7071 <listitem><para>Merge documentation into one book (Peter)</para></listitem>
7072 <listitem><para>Add Windows compatibility functions (Bruce)</para></listitem>
7073 <listitem><para>Allow client interfaces to compile under MinGW (Bruce)</para></listitem>
7074 <listitem><para>New <function>ereport()</function> function for error reporting (Tom)</para></listitem>
7075 <listitem><para>Support Intel compiler on Linux (Peter)</para></listitem>
7076 <listitem><para>Improve Linux startup scripts (Slawomir Sudnik, Darko Prenosil)</para></listitem>
7077 <listitem><para>Add support for AMD Opteron and Itanium (Jeffrey W. Baker, Bruce)</para></listitem>
7079 <para>Remove <option>--enable-recode</option> option from <command>configure</command></para>
7081 This was no longer needed now that we have <command>CREATE CONVERSION</command>.
7085 <para>Generate a compile error if spinlock code is not found (Bruce)</para>
7087 Platforms without spinlock code will now fail to compile, rather
7088 than silently using semaphores. This failure can be disabled
7089 with a new <command>configure</command> option.
7096 <title>Contrib Changes</title>
7099 <listitem><para>Change dbmirror license to BSD</para></listitem>
7100 <listitem><para>Improve earthdistance (Bruno Wolff III)</para></listitem>
7101 <listitem><para>Portability improvements to pgcrypto (Marko Kreen)</para></listitem>
7102 <listitem><para>Prevent crash in xml (John Gray, Michael Richards)</para></listitem>
7103 <listitem><para>Update oracle</para></listitem>
7104 <listitem><para>Update mysql</para></listitem>
7105 <listitem><para>Update cube (Bruno Wolff III)</para></listitem>
7106 <listitem><para>Update earthdistance to use cube (Bruno Wolff III)</para></listitem>
7107 <listitem><para>Update btree_gist (Oleg)</para></listitem>
7108 <listitem><para>New tsearch2 full-text search module (Oleg, Teodor)</para></listitem>
7109 <listitem><para>Add hash-based crosstab function to tablefuncs (Joe)</para></listitem>
7110 <listitem><para>Add serial column to order <function>connectby()</> siblings in tablefuncs (Nabil Sayegh,Joe)</para></listitem>
7111 <listitem><para>Add named persistent connections to dblink (Shridhar Daithanka)</para></listitem>
7112 <listitem><para>New pg_autovacuum allows automatic <command>VACUUM</command> (Matthew T. O'Connor)</para></listitem>
7113 <listitem><para>Make pgbench honor environment variables <envar>PGHOST</>, <envar>PGPORT</>, <envar>PGUSER</> (Tatsuo)</para></listitem>
7114 <listitem><para>Improve intarray (Teodor Sigaev)</para></listitem>
7115 <listitem><para>Improve pgstattuple (Rod)</para></listitem>
7116 <listitem><para>Fix bug in <function>metaphone()</function> in fuzzystrmatch</para></listitem>
7117 <listitem><para>Improve adddepend (Rod)</para></listitem>
7118 <listitem><para>Update spi/timetravel (Böjthe Zoltán)</para></listitem>
7119 <listitem><para>Fix dbase <option>-s</> option and improve non-ASCII handling (Thomas Behr, Márcio Smiderle)</para></listitem>
7120 <listitem><para>Remove array module because features now included by default (Joe)</para></listitem>
7126 <sect1 id="release-7-3-10">
7127 <title>Release 7.3.10</title>
7130 <title>Release date</title>
7131 <simpara>2005-05-09</simpara>
7135 This release contains a variety of fixes from 7.3.9, including several
7136 security-related issues.
7140 <title>Migration to version 7.3.10</title>
7143 A dump/restore is not required for those running 7.3.X. However,
7144 it is one possible way of handling a significant security problem
7145 that has been found in the initial contents of 7.3.X system
7146 catalogs. A dump/initdb/reload sequence using 7.3.10's initdb will
7147 automatically correct this problem.
7151 The security problem is that the built-in character set encoding
7152 conversion functions can be invoked from SQL commands by unprivileged
7153 users, but the functions were not designed for such use and are not
7154 secure against malicious choices of arguments. The fix involves changing
7155 the declared parameter list of these functions so that they can no longer
7156 be invoked from SQL commands. (This does not affect their normal use
7157 by the encoding conversion machinery.)
7158 It is strongly recommended that all installations repair this error,
7159 either by initdb or by following the manual repair procedure given
7160 below. The error at least allows unprivileged database users to crash
7161 their server process, and may allow unprivileged users to gain the
7162 privileges of a database superuser.
7166 If you wish not to do an initdb, perform the following procedure instead.
7167 As the database superuser, do:
7171 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
7172 WHERE pronamespace = 11 AND pronargs = 5
7173 AND proargtypes[2] = 'cstring'::regtype;
7174 -- The command should report having updated 90 rows;
7175 -- if not, rollback and investigate instead of committing!
7181 The above procedure must be carried out in <emphasis>each</> database
7182 of an installation, including <literal>template1</>, and ideally
7183 including <literal>template0</> as well. If you do not fix the
7184 template databases then any subsequently created databases will contain
7185 the same error. <literal>template1</> can be fixed in the same way
7186 as any other database, but fixing <literal>template0</> requires
7187 additional steps. First, from any database issue
7189 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
7191 Next connect to <literal>template0</> and perform the above repair
7192 procedure. Finally, do
7194 -- re-freeze template0:
7196 -- and protect it against future alterations:
7197 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
7203 <title>Changes</title>
7206 <listitem><para>Change encoding function signature to prevent
7207 misuse</para></listitem>
7208 <listitem><para>Repair ancient race condition that allowed a transaction to be
7209 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
7210 than for other purposes</para>
7211 <para>This is an extremely serious bug since it could lead to apparent
7212 data inconsistencies being briefly visible to applications.</para></listitem>
7213 <listitem><para>Repair race condition between relation extension and
7215 <para>This could theoretically have caused loss of a page's worth of
7216 freshly-inserted data, although the scenario seems of very low probability.
7217 There are no known cases of it having caused more than an Assert failure.
7219 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
7221 The comparison code was wrong in the case where the
7222 <literal>--enable-integer-datetimes</> configuration switch had been used.
7223 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
7224 it will need to be <command>REINDEX</>ed after installing this update, because
7225 the fix corrects the sort order of column values.
7227 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
7228 <type>TIME WITH TIME ZONE</> values</para></listitem>
7229 <listitem><para>Fix mis-display of negative fractional seconds in
7230 <type>INTERVAL</> values</para>
7232 This error only occurred when the
7233 <literal>--enable-integer-datetimes</> configuration switch had been used.
7235 <listitem><para>Additional buffer overrun checks in plpgsql
7236 (Neil)</para></listitem>
7237 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
7238 correctly (Neil)</para></listitem>
7239 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
7240 month-related formats</para></listitem>
7241 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
7242 (Marko Kreen)</para></listitem>
7243 <listitem><para>Still more 64-bit fixes for
7244 <filename>contrib/intagg</></para></listitem>
7245 <listitem><para>Prevent incorrect optimization of functions returning
7246 <type>RECORD</></para></listitem>
7252 <sect1 id="release-7-3-9">
7253 <title>Release 7.3.9</title>
7256 <title>Release date</title>
7257 <simpara>2005-01-31</simpara>
7261 This release contains a variety of fixes from 7.3.8, including several
7262 security-related issues.
7266 <title>Migration to version 7.3.9</title>
7269 A dump/restore is not required for those running 7.3.X.
7274 <title>Changes</title>
7277 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
7279 On platforms that will automatically execute initialization functions of a
7280 shared library (this includes at least Windows and ELF-based Unixen),
7281 <command>LOAD</> can be used to make the server execute arbitrary code.
7282 Thanks to NGS Software for reporting this.</para></listitem>
7283 <listitem><para>Check that creator of an aggregate function has the right to
7284 execute the specified transition functions</para>
7286 This oversight made it possible to bypass denial of EXECUTE
7287 permission on a function.</para></listitem>
7288 <listitem><para>Fix security and 64-bit issues in
7289 contrib/intagg</para></listitem>
7290 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
7291 Jurka)</para></listitem>
7292 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
7293 many parameters (Neil)</para></listitem>
7294 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
7296 The result of the join was mistakenly supposed to be sorted the same as the
7297 left input. This could not only deliver mis-sorted output to the user, but
7298 in case of nested merge joins could give outright wrong answers.
7300 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
7301 <listitem><para>Fix display of negative intervals in SQL and GERMAN
7302 datestyles</para></listitem>
7308 <sect1 id="release-7-3-8">
7309 <title>Release 7.3.8</title>
7312 <title>Release date</title>
7313 <simpara>2004-10-22</simpara>
7317 This release contains a variety of fixes from 7.3.7.
7322 <title>Migration to version 7.3.8</title>
7325 A dump/restore is not required for those running 7.3.X.
7330 <title>Changes</title>
7333 <listitem><para>Repair possible failure to update hint bits on disk</para>
7335 Under rare circumstances this oversight could lead to
7336 <quote>could not access transaction status</> failures, which qualifies
7337 it as a potential-data-loss bug.
7339 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
7341 Very large left joins using a hash join plan could fail to output unmatched
7342 left-side rows given just the right data distribution.
7344 <listitem><para>Disallow running pg_ctl as root</para>
7346 This is to guard against any possible security issues.
7348 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
7350 This has been reported as a security issue, though it's hardly worthy of
7351 concern since there is no reason for non-developers to use this script anyway.
7358 <sect1 id="release-7-3-7">
7359 <title>Release 7.3.7</title>
7362 <title>Release date</title>
7363 <simpara>2004-08-16</simpara>
7367 This release contains one critical fix over 7.3.6, and some minor items.
7372 <title>Migration to version 7.3.7</title>
7375 A dump/restore is not required for those running 7.3.X.
7380 <title>Changes</title>
7383 <listitem><para>Prevent possible loss of committed transactions during crash</para>
7385 Due to insufficient interlocking between transaction commit and checkpointing,
7386 it was possible for transactions committed just before the most recent
7387 checkpoint to be lost, in whole or in part, following a database crash and
7388 restart. This is a serious bug that has existed
7389 since <productname>PostgreSQL</productname> 7.1.
7391 <listitem><para>Remove asymmetrical word processing in tsearch (Teodor)</para></listitem>
7392 <listitem><para>Properly schema-qualify function names when pg_dump'ing a CAST</para></listitem>
7398 <sect1 id="release-7-3-6">
7399 <title>Release 7.3.6</title>
7402 <title>Release date</title>
7403 <simpara>2004-03-02</simpara>
7407 This release contains a variety of fixes from 7.3.5.
7412 <title>Migration to version 7.3.6</title>
7415 A dump/restore is <emphasis>not</emphasis> required for those
7422 <title>Changes</title>
7425 <listitem><para>Revert erroneous changes in rule permissions checking</para>
7426 <para>A patch applied in 7.3.3 to fix a corner case in rule permissions checks
7427 turns out to have disabled rule-related permissions checks in many
7428 not-so-corner cases. This would for example allow users to insert into views
7429 they weren't supposed to have permission to insert into. We have therefore
7430 reverted the 7.3.3 patch. The original bug will be fixed in 8.0.
7432 <listitem><para>Repair incorrect order of operations in
7433 GetNewTransactionId()</para>
7435 This bug could result in failure under out-of-disk-space conditions, including
7436 inability to restart even after disk space is freed.
7438 <listitem><para>Ensure configure selects -fno-strict-aliasing even when
7439 an external value for CFLAGS is supplied</para>
7441 On some platforms, building with -fstrict-aliasing causes bugs.
7443 <listitem><para>Make pg_restore handle 64-bit off_t correctly</para>
7445 This bug prevented proper restoration from archive files exceeding 4Gb.
7447 <listitem><para>Make contrib/dblink not assume that local and remote type OIDs
7448 match (Joe)</para></listitem>
7449 <listitem><para>Quote connectby()'s start_with argument properly (Joe)</para></listitem>
7450 <listitem><para>Don't crash when a rowtype argument to a plpgsql function is
7451 NULL</para></listitem>
7452 <listitem><para>Avoid generating invalid character encoding sequences in
7453 corner cases when planning LIKE operations</para></listitem>
7454 <listitem><para>Ensure text_position() cannot scan past end of source string
7455 in multibyte cases (Korea PostgreSQL Users' Group)</para></listitem>
7456 <listitem><para>Fix index optimization and selectivity estimates for LIKE
7457 operations on bytea columns (Joe)</para></listitem>
7463 <sect1 id="release-7-3-5">
7464 <title>Release 7.3.5</title>
7467 <title>Release date</title>
7468 <simpara>2003-12-03</simpara>
7472 This has a variety of fixes from 7.3.4.
7477 <title>Migration to version 7.3.5</title>
7480 A dump/restore is <emphasis>not</emphasis> required for those
7486 <title>Changes</title>
7489 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
7490 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
7491 <listitem><para>Force stats processes to detach from shared memory, ensuring cleaner shutdown</para></listitem>
7492 <listitem><para>Make PQescapeBytea and byteaout consistent with each other (Joe)</para></listitem>
7493 <listitem><para>Added missing SPI_finish() calls to dblink's get_tuple_of_interest() (Joe)</para></listitem>
7494 <listitem><para>Fix for possible foreign key violation when rule rewrites INSERT (Jan)</para></listitem>
7495 <listitem><para>Support qualified type names in PL/Tcl's spi_prepare command (Jan)</para></listitem>
7496 <listitem><para>Make pg_dump handle a procedural language handler located in pg_catalog</para></listitem>
7497 <listitem><para>Make pg_dump handle cases where a custom opclass is in another schema</para></listitem>
7498 <listitem><para>Make pg_dump dump binary-compatible casts correctly (Jan)</para></listitem>
7499 <listitem><para>Fix insertion of expressions containing subqueries into rule bodies</para></listitem>
7500 <listitem><para>Fix incorrect argument processing in clusterdb script (Anand Ranganathan)</para></listitem>
7501 <listitem><para>Fix problems with dropped columns in plpython triggers</para></listitem>
7502 <listitem><para>Repair problems with to_char() reading past end of its input string (Karel)</para></listitem>
7503 <listitem><para>Fix GB18030 mapping errors (Tatsuo)</para></listitem>
7504 <listitem><para>Fix several problems with SSL error handling and asynchronous SSL I/O</para></listitem>
7505 <listitem><para>Remove ability to bind a list of values to a single parameter in JDBC
7506 (prevents possible SQL-injection attacks)</para></listitem>
7507 <listitem><para>Fix some errors in HAVE_INT64_TIMESTAMP code paths</para></listitem>
7508 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
7514 <sect1 id="release-7-3-4">
7515 <title>Release 7.3.4</title>
7518 <title>Release date</title>
7519 <simpara>2003-07-24</simpara>
7523 This has a variety of fixes from 7.3.3.
7528 <title>Migration to version 7.3.4</title>
7531 A dump/restore is <emphasis>not</emphasis> required for those
7537 <title>Changes</title>
7540 <listitem><para>Repair breakage in timestamp-to-date conversion for dates before 2000</para></listitem>
7541 <listitem><para>Prevent rare possibility of server startup failure (Tom)</para></listitem>
7542 <listitem><para>Fix bugs in interval-to-time conversion (Tom)</para></listitem>
7543 <listitem><para>Add constraint names in a few places in pg_dump (Rod)</para></listitem>
7544 <listitem><para>Improve performance of functions with many parameters (Tom)</para></listitem>
7545 <listitem><para>Fix to_ascii() buffer overruns (Tom)</para></listitem>
7546 <listitem><para>Prevent restore of database comments from throwing an error (Tom)</para></listitem>
7547 <listitem><para>Work around buggy strxfrm() present in some Solaris releases (Tom)</para></listitem>
7548 <listitem><para>Properly escape jdbc setObject() strings to improve security (Barry)</para></listitem>
7554 <sect1 id="release-7-3-3">
7555 <title>Release 7.3.3</title>
7558 <title>Release date</title>
7559 <simpara>2003-05-22</simpara>
7563 This release contains a variety of fixes for version 7.3.2.
7567 <title>Migration to version 7.3.3</title>
7570 A dump/restore is <emphasis>not</emphasis> required for those
7571 running version 7.3.*.
7576 <title>Changes</title>
7579 <listitem><para>Repair sometimes-incorrect computation of StartUpID after a crash</para></listitem>
7580 <listitem><para>Avoid slowness with lots of deferred triggers in one transaction (Stephan)</para></listitem>
7581 <listitem><para>Don't lock referenced row when <command>UPDATE</command> doesn't change foreign key's value (Jan)</para></listitem>
7582 <listitem><para>Use <command>-fPIC</command> not <command>-fpic</command> on Sparc (Tom Callaway)</para></listitem>
7583 <listitem><para>Repair lack of schema-awareness in contrib/reindexdb</para></listitem>
7584 <listitem><para>Fix contrib/intarray error for zero-element result array (Teodor)</para></listitem>
7585 <listitem><para>Ensure createuser script will exit on control-C (Oliver)</para></listitem>
7586 <listitem><para>Fix errors when the type of a dropped column has itself been dropped</para></listitem>
7587 <listitem><para><command>CHECKPOINT</command> does not cause database panic on failure in noncritical steps</para></listitem>
7588 <listitem><para>Accept 60 in seconds fields of timestamp, time, interval input values</para></listitem>
7589 <listitem><para>Issue notice, not error, if <type>TIMESTAMP</type>,
7590 <type> TIME</type>, or <type>INTERVAL</type> precision too large</para></listitem>
7591 <listitem><para>Fix <function>abstime-to-time</function> cast function (fix is
7592 not applied unless you <application>initdb</application>)</para></listitem>
7593 <listitem><para>Fix <application>pg_proc</application> entry for
7594 <type>timestampt_izone</type> (fix is not applied unless you
7595 <application>initdb</application>)</para></listitem>
7596 <listitem><para>Make <function>EXTRACT(EPOCH FROM timestamp without time zone)</function> treat input as local time</para></listitem>
7597 <listitem><para><command>'now'::timestamptz</command> gave wrong answer if timezone changed earlier in transaction</para></listitem>
7598 <listitem><para><envar>HAVE_INT64_TIMESTAMP</envar> code for time with timezone overwrote its input</para></listitem>
7599 <listitem><para>Accept <command>GLOBAL TEMP/TEMPORARY</command> as a
7600 synonym for <command>TEMPORARY</command></para></listitem>
7601 <listitem><para>Avoid improper schema-privilege-check failure in foreign-key triggers</para></listitem>
7602 <listitem><para>Fix bugs in foreign-key triggers for <command>SET DEFAULT</command> action</para></listitem>
7603 <listitem><para>Fix incorrect time-qual check in row fetch for
7604 <command>UPDATE</command> and <command>DELETE</command> triggers</para></listitem>
7605 <listitem><para>Foreign-key clauses were parsed but ignored in
7606 <command>ALTER TABLE ADD COLUMN</command></para></listitem>
7607 <listitem><para>Fix createlang script breakage for case where handler function already exists</para></listitem>
7608 <listitem><para>Fix misbehavior on zero-column tables in <application>pg_dump</application>, COPY, ANALYZE, other places</para></listitem>
7609 <listitem><para>Fix misbehavior of <function>func_error()</function> on type names containing '%'</para></listitem>
7610 <listitem><para>Fix misbehavior of <function>replace()</function> on strings containing '%'</para></listitem>
7611 <listitem><para>Regular-expression patterns containing certain multibyte characters failed</para></listitem>
7612 <listitem><para>Account correctly for <command>NULL</command>s in more cases in join size estimation</para></listitem>
7613 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
7614 <listitem><para>Fix failure to convert large code point values in EUC_TW conversions (Tatsuo)</para></listitem>
7615 <listitem><para>Fix error recovery for <function>SSL_read</function>/<function>SSL_write</function> calls</para></listitem>
7616 <listitem><para>Don't do early constant-folding of type coercion expressions</para></listitem>
7617 <listitem><para>Validate page header fields immediately after reading in any page</para></listitem>
7618 <listitem><para>Repair incorrect check for ungrouped variables in unnamed joins</para></listitem>
7619 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
7620 <listitem><para>contrib/ltree fixes (Teodor)</para></listitem>
7621 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
7622 <listitem><para>Avoid running out of buffers in many-way indexscan (bug introduced in 7.3)</para></listitem>
7623 <listitem><para>Fix planner's selectivity estimation functions to handle domains properly</para></listitem>
7624 <listitem><para>Fix <application>dbmirror</application> memory-allocation bug (Steven Singer)</para></listitem>
7625 <listitem><para>Prevent infinite loop in <function>ln(numeric)</function> due to roundoff error</para></listitem>
7626 <listitem><para><command>GROUP BY</command> got confused if there were multiple equal GROUP BY items</para></listitem>
7627 <listitem><para>Fix bad plan when inherited <command>UPDATE</command>/<command>DELETE</command> references another inherited table</para></listitem>
7628 <listitem><para>Prevent clustering on incomplete (partial or non-NULL-storing) indexes</para></listitem>
7629 <listitem><para>Service shutdown request at proper time if it arrives while still starting up</para></listitem>
7630 <listitem><para>Fix left-links in temporary indexes (could make backwards scans miss entries)</para></listitem>
7631 <listitem><para>Fix incorrect handling of client_encoding setting in postgresql.conf (Tatsuo)</para></listitem>
7632 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
7633 <listitem><para>Fix SPI for case where rule contains multiple statements of the same type</para></listitem>
7634 <listitem><para>Fix problem with checking for wrong type of access privilege in rule query</para></listitem>
7635 <listitem><para>Fix problem with <command>EXCEPT</command> in <command>CREATE RULE</command></para></listitem>
7636 <listitem><para>Prevent problem with dropping temp tables having serial columns</para></listitem>
7637 <listitem><para>Fix replace_vars_with_subplan_refs failure in complex views</para></listitem>
7638 <listitem><para>Fix regexp slowness in single-byte encodings (Tatsuo)</para></listitem>
7639 <listitem><para>Allow qualified type names in <command>CREATE CAST</command>
7640 and <command> DROP CAST</command></para></listitem>
7641 <listitem><para>Accept <function>SETOF type[]</function>, which formerly had to
7642 be written <function>SETOF _type</function></para></listitem>
7643 <listitem><para>Fix <application>pg_dump</application> core dump in some cases with procedural languages</para></listitem>
7644 <listitem><para>Force ISO datestyle in <application>pg_dump</application> output, for portability (Oliver)</para></listitem>
7645 <listitem><para><application>pg_dump</application> failed to handle error return
7646 from <function>lo_read</function> (Oleg Drokin)</para></listitem>
7647 <listitem><para><application>pg_dumpall</application> failed with groups having no members (Nick Eskelinen)</para></listitem>
7648 <listitem><para><application>pg_dumpall</application> failed to recognize --globals-only switch</para></listitem>
7649 <listitem><para>pg_restore failed to restore blobs if -X disable-triggers is specified</para></listitem>
7650 <listitem><para>Repair intrafunction memory leak in plpgsql</para></listitem>
7651 <listitem><para>pltcl's <command>elog</command> command dumped core if given wrong parameters (Ian Harding)</para></listitem>
7652 <listitem><para>plpython used wrong value of <envar>atttypmod</envar> (Brad McLean)</para></listitem>
7653 <listitem><para>Fix improper quoting of boolean values in Python interface (D'Arcy)</para></listitem>
7654 <listitem><para>Added <function>addDataType()</function> method to PGConnection interface for JDBC</para></listitem>
7655 <listitem><para>Fixed various problems with updateable ResultSets for JDBC (Shawn Green)</para></listitem>
7656 <listitem><para>Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka, Peter Royal)</para></listitem>
7657 <listitem><para>Fixed problem with parsing table ACLs in JDBC</para></listitem>
7658 <listitem><para>Better error message for character set conversion problems in JDBC</para></listitem>
7664 <sect1 id="release-7-3-2">
7665 <title>Release 7.3.2</title>
7668 <title>Release date</title>
7669 <simpara>2003-02-04</simpara>
7673 This release contains a variety of fixes for version 7.3.1.
7678 <title>Migration to version 7.3.2</title>
7681 A dump/restore is <emphasis>not</emphasis> required for those
7682 running version 7.3.*.
7687 <title>Changes</title>
7690 <listitem><para>Restore creation of OID column in CREATE TABLE AS / SELECT INTO</para></listitem>
7691 <listitem><para>Fix <application>pg_dump</> core dump when dumping views having comments</para></listitem>
7692 <listitem><para>Dump DEFERRABLE/INITIALLY DEFERRED constraints properly</para></listitem>
7693 <listitem><para>Fix UPDATE when child table's column numbering differs from parent</para></listitem>
7694 <listitem><para>Increase default value of max_fsm_relations</para></listitem>
7695 <listitem><para>Fix problem when fetching backwards in a cursor for a single-row query</para></listitem>
7696 <listitem><para>Make backward fetch work properly with cursor on SELECT DISTINCT query</para></listitem>
7697 <listitem><para>Fix problems with loading <application>pg_dump</> files containing contrib/lo usage</para></listitem>
7698 <listitem><para>Fix problem with all-numeric user names</para></listitem>
7699 <listitem><para>Fix possible memory leak and core dump during disconnect in libpgtcl</para></listitem>
7700 <listitem><para>Make plpython's spi_execute command handle nulls properly (Andrew Bosma)</para></listitem>
7701 <listitem><para>Adjust plpython error reporting so that its regression test passes again</para></listitem>
7702 <listitem><para>Work with bison 1.875</para></listitem>
7703 <listitem><para>Handle mixed-case names properly in plpgsql's %type (Neil)</para></listitem>
7704 <listitem><para>Fix core dump in pltcl when executing a query rewritten by a rule</para></listitem>
7705 <listitem><para>Repair array subscript overruns (per report from Yichen Xie)</para></listitem>
7706 <listitem><para>Reduce MAX_TIME_PRECISION from 13 to 10 in floating-point case</para></listitem>
7707 <listitem><para>Correctly case-fold variable names in per-database and per-user settings</para></listitem>
7708 <listitem><para>Fix coredump in plpgsql's RETURN NEXT when SELECT into record returns no rows</para></listitem>
7709 <listitem><para>Fix outdated use of pg_type.typprtlen in python client interface</para></listitem>
7710 <listitem><para>Correctly handle fractional seconds in timestamps in JDBC driver</para></listitem>
7711 <listitem><para>Improve performance of getImportedKeys() in JDBC</para></listitem>
7712 <listitem><para>Make shared-library symlinks work standardly on HPUX (Giles)</para></listitem>
7713 <listitem><para>Repair inconsistent rounding behavior for timestamp, time, interval</para></listitem>
7714 <listitem><para>SSL negotiation fixes (Nathan Mueller)</para></listitem>
7715 <listitem><para>Make libpq's ~/.pgpass feature work when connecting with PQconnectDB</para></listitem>
7716 <listitem><para>Update my2pg, ora2pg</para></listitem>
7717 <listitem><para>Translation updates</para></listitem>
7718 <listitem><para>Add casts between types lo and oid in contrib/lo</para></listitem>
7719 <listitem><para>fastpath code now checks for privilege to call function</para></listitem>
7725 <sect1 id="release-7-3-1">
7726 <title>Release 7.3.1</title>
7729 <title>Release date</title>
7730 <simpara>2002-12-18</simpara>
7734 This release contains a variety of fixes for version 7.3.
7739 <title>Migration to version 7.3.1</title>
7742 A dump/restore is <emphasis>not</emphasis> required for those
7743 running version 7.3. However, it should be noted that the main
7744 <productname>PostgreSQL</productname> interface library, libpq,
7745 has a new major version number for this release, which may require
7746 recompilation of client code in certain cases.
7751 <title>Changes</title>
7754 <listitem><para>Fix a core dump of COPY TO when client/server encodings don't match (Tom)</para></listitem>
7755 <listitem><para>Allow <application>pg_dump</> to work with pre-7.2 servers (Philip)</para></listitem>
7756 <listitem><para>contrib/adddepend fixes (Tom)</para></listitem>
7757 <listitem><para>Fix problem with deletion of per-user/per-database config settings (Tom)</para></listitem>
7758 <listitem><para>contrib/vacuumlo fix (Tom)</para></listitem>
7759 <listitem><para>Allow 'password' encryption even when pg_shadow contains MD5 passwords (Bruce)</para></listitem>
7760 <listitem><para>contrib/dbmirror fix (Steven Singer)</para></listitem>
7761 <listitem><para>Optimizer fixes (Tom)</para></listitem>
7762 <listitem><para>contrib/tsearch fixes (Teodor Sigaev, Magnus)</para></listitem>
7763 <listitem><para>Allow locale names to be mixed case (Nicolai Tufar)</para></listitem>
7764 <listitem><para>Increment libpq library's major version number (Bruce)</para></listitem>
7765 <listitem><para>pg_hba.conf error reporting fixes (Bruce, Neil)</para></listitem>
7766 <listitem><para>Add SCO Openserver 5.0.4 as a supported platform (Bruce)</para></listitem>
7767 <listitem><para>Prevent EXPLAIN from crashing server (Tom)</para></listitem>
7768 <listitem><para>SSL fixes (Nathan Mueller)</para></listitem>
7769 <listitem><para>Prevent composite column creation via ALTER TABLE (Tom)</para></listitem>
7775 <sect1 id="release-7-3">
7776 <title>Release 7.3</title>
7779 <title>Release date</title>
7780 <simpara>2002-11-27</simpara>
7784 <title>Overview</title>
7787 Major changes in this release:
7792 <term>Schemas</term>
7795 Schemas allow users to create objects in separate namespaces,
7796 so two people or applications can have tables with the same
7797 name. There is also a public schema for shared tables.
7798 Table/index creation can be restricted by removing privileges
7799 on the public schema.
7805 <term>Drop Column</term>
7808 <productname>PostgreSQL</productname> now supports the
7809 <literal>ALTER TABLE ... DROP COLUMN</literal> functionality.
7815 <term>Table Functions</term>
7818 Functions returning multiple rows and/or multiple columns are
7819 now much easier to use than before. You can call such a
7820 <quote>table function</quote> in the <literal>SELECT</literal>
7821 <literal>FROM</literal> clause, treating its output like a
7822 table. Also, <application>PL/pgSQL</application> functions can
7829 <term>Prepared Queries</term>
7832 <productname>PostgreSQL</productname> now supports prepared
7833 queries, for improved performance.
7839 <term>Dependency Tracking</term>
7842 <productname>PostgreSQL</productname> now records object
7843 dependencies, which allows improvements in many areas.
7844 <command>DROP</command> statements now take either
7845 <literal>CASCADE</> or <literal>RESTRICT</> to control whether
7846 dependent objects are also dropped.
7852 <term>Privileges</term>
7855 Functions and procedural languages now have privileges, and
7856 functions can be defined to run with the privileges of their
7863 <term>Internationalization</term>
7866 Both multibyte and locale support are now always enabled.
7872 <term>Logging</term>
7875 A variety of logging options have been enhanced.
7881 <term>Interfaces</term>
7884 A large number of interfaces have been moved to <ulink
7885 url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
7886 where they can be developed and released independently.
7892 <term>Functions/Identifiers</term>
7895 By default, functions can now take up to 32 parameters, and
7896 identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
7897 is now deprecated: there are specific <quote>pseudo-datatypes</>
7898 to represent each of the former meanings of <literal>OPAQUE</>
7899 in function argument and result types.
7908 <title>Migration to version 7.3</title>
7911 A dump/restore using <application>pg_dump</> is required for those
7912 wishing to migrate data from any previous release. If your
7913 application examines the system catalogs, additional changes will
7914 be required due to the introduction of schemas in 7.3; for more
7915 information, see: <ulink
7916 url="http://developer.postgresql.org/~momjian/upgrade_tips_7.3"></>.
7920 Observe the following incompatibilities:
7926 Pre-6.3 clients are no longer supported.
7932 <filename>pg_hba.conf</filename> now has a column for the user
7933 name and additional features. Existing files need to be
7940 Several <filename>postgresql.conf</filename> logging parameters
7947 <literal>LIMIT #,#</literal> has been disabled; use
7948 <literal>LIMIT # OFFSET #</literal>.
7954 <command>INSERT</command> statements with column lists must
7955 specify a value for each specified column. For example,
7956 <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
7957 is now invalid. It's still allowed to supply fewer columns than
7958 expected if the <command>INSERT</command> does not have a column list.
7964 <type>serial</type> columns are no longer automatically
7965 <literal>UNIQUE</>; thus, an index will not automatically be
7972 A <command>SET</command> command inside an aborted transaction
7979 <command>COPY</command> no longer considers missing trailing
7980 columns to be null. All columns need to be specified.
7981 (However, one may achieve a similar effect by specifying a
7982 column list in the <command>COPY</command> command.)
7988 The data type <type>timestamp</type> is now equivalent to
7989 <type>timestamp without time zone</type>, instead of
7990 <type>timestamp with time zone</type>.
7996 Pre-7.3 databases loaded into 7.3 will not have the new object
7997 dependencies for <type>serial</type> columns, unique
7998 constraints, and foreign keys. See the directory
7999 <filename>contrib/adddepend/</filename> for a detailed
8000 description and a script that will add such dependencies.
8006 An empty string (<literal>''</literal>) is no longer allowed as
8007 the input into an integer field. Formerly, it was silently
8016 <title>Changes</title>
8019 <title>Server Operation</title>
8021 <listitem><para>Add pg_locks view to show locks (Neil)</para></listitem>
8022 <listitem><para>Security fixes for password negotiation memory allocation (Neil)</para></listitem>
8023 <listitem><para>Remove support for version 0 FE/BE protocol (<productname>PostgreSQL</productname> 6.2 and earlier) (Tom)</para></listitem>
8024 <listitem><para>Reserve the last few backend slots for superusers, add parameter superuser_reserved_connections to control this (Nigel J. Andrews)</para></listitem>
8029 <title>Performance</title>
8031 <listitem><para>Improve startup by calling localtime() only once (Tom)</para></listitem>
8032 <listitem><para>Cache system catalog information in flat files for faster startup (Tom)</para></listitem>
8033 <listitem><para>Improve caching of index information (Tom)</para></listitem>
8034 <listitem><para>Optimizer improvements (Tom, Fernando Nasser)</para></listitem>
8035 <listitem><para>Catalog caches now store failed lookups (Tom)</para></listitem>
8036 <listitem><para>Hash function improvements (Neil)</para></listitem>
8037 <listitem><para>Improve performance of query tokenization and network handling (Peter)</para></listitem>
8038 <listitem><para>Speed improvement for large object restore (Mario Weilguni)</para></listitem>
8039 <listitem><para>Mark expired index entries on first lookup, saving later heap fetches (Tom)</para></listitem>
8040 <listitem><para>Avoid excessive NULL bitmap padding (Manfred Koizar)</para></listitem>
8041 <listitem><para>Add BSD-licensed qsort() for Solaris, for performance (Bruce)</para></listitem>
8042 <listitem><para>Reduce per-row overhead by four bytes (Manfred Koizar)</para></listitem>
8043 <listitem><para>Fix GEQO optimizer bug (Neil Conway)</para></listitem>
8044 <listitem><para>Make WITHOUT OID actually save four bytes per row (Manfred Koizar)</para></listitem>
8045 <listitem><para>Add default_statistics_target variable to specify ANALYZE buckets (Neil)</para></listitem>
8046 <listitem><para>Use local buffer cache for temporary tables so no WAL overhead (Tom)</para></listitem>
8047 <listitem><para>Improve free space map performance on large tables (Stephen Marshall, Tom)</para></listitem>
8048 <listitem><para>Improved WAL write concurrency (Tom)</para></listitem>
8053 <title>Privileges</title>
8055 <listitem><para>Add privileges on functions and procedural languages (Peter)</para></listitem>
8056 <listitem><para>Add OWNER to CREATE DATABASE so superusers can create databases on behalf of unprivileged users (Gavin Sherry, Tom)</para></listitem>
8057 <listitem><para>Add new object privilege bits EXECUTE and USAGE (Tom)</para></listitem>
8058 <listitem><para>Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION (Tom)</para></listitem>
8059 <listitem><para>Allow functions to be executed with the privilege of the function owner (Peter)</para></listitem>
8064 <title>Server Configuration</title>
8066 <listitem><para>Server log messages now tagged with LOG, not DEBUG (Bruce)</para></listitem>
8067 <listitem><para>Add user column to pg_hba.conf (Bruce)</para></listitem>
8068 <listitem><para>Have log_connections output two lines in log file (Tom)</para></listitem>
8069 <listitem><para>Remove debug_level from postgresql.conf, now server_min_messages (Bruce)</para></listitem>
8070 <listitem><para>New ALTER DATABASE/USER ... SET command for per-user/database initialization (Peter)</para></listitem>
8071 <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>
8072 <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>
8073 <listitem><para>Remove secondary password file capability and pg_password utility (Bruce)</para></listitem>
8074 <listitem><para>Add variable db_user_namespace for database-local user names (Bruce)</para></listitem>
8075 <listitem><para>SSL improvements (Bear Giles)</para></listitem>
8076 <listitem><para>Make encryption of stored passwords the default (Bruce)</para></listitem>
8077 <listitem><para>Allow pg_statistics to be reset by calling pg_stat_reset() (Christopher)</para></listitem>
8078 <listitem><para>Add log_duration parameter (Bruce)</para></listitem>
8079 <listitem><para>Rename debug_print_query to log_statement (Bruce)</para></listitem>
8080 <listitem><para>Rename show_query_stats to show_statement_stats (Bruce)</para></listitem>
8081 <listitem><para>Add param log_min_error_statement to print commands to logs on error (Gavin)</para></listitem>
8086 <title>Queries</title>
8088 <listitem><para>Make cursors insensitive, meaning their contents do not change (Tom)</para></listitem>
8089 <listitem><para>Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)</para></listitem>
8090 <listitem><para>Increase identifier length to 63 (Neil, Bruce)</para></listitem>
8091 <listitem><para>UNION fixes for merging >= 3 columns of different lengths (Tom)</para></listitem>
8092 <listitem><para>Add DEFAULT key word to INSERT, e.g., INSERT ... (..., DEFAULT, ...) (Rod)</para></listitem>
8093 <listitem><para>Allow views to have default values using ALTER COLUMN ... SET DEFAULT (Neil)</para></listitem>
8094 <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>
8095 <listitem><para>Fix for join aliases (Tom)</para></listitem>
8096 <listitem><para>Fix for FULL OUTER JOINs (Tom)</para></listitem>
8097 <listitem><para>Improve reporting of invalid identifier and location (Tom, Gavin)</para></listitem>
8098 <listitem><para>Fix OPEN cursor(args) (Tom)</para></listitem>
8099 <listitem><para>Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)</para></listitem>
8100 <listitem><para>Fix for CREATE TABLE AS with UNION (Tom)</para></listitem>
8101 <listitem><para>SQL99 syntax improvements (Thomas)</para></listitem>
8102 <listitem><para>Add statement_timeout variable to cancel queries (Bruce)</para></listitem>
8103 <listitem><para>Allow prepared queries with PREPARE/EXECUTE (Neil)</para></listitem>
8104 <listitem><para>Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)</para></listitem>
8105 <listitem><para>Add variable autocommit (Tom, David Van Wie)</para></listitem>
8110 <title>Object Manipulation</title>
8112 <listitem><para>Make equals signs optional in CREATE DATABASE (Gavin Sherry)</para></listitem>
8113 <listitem><para>Make ALTER TABLE OWNER change index ownership too (Neil)</para></listitem>
8114 <listitem><para>New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls TOAST storage, compression (John Gray)</para></listitem>
8115 <listitem><para>Add schema support, CREATE/DROP SCHEMA (Tom)</para></listitem>
8116 <listitem><para>Create schema for temporary tables (Tom)</para></listitem>
8117 <listitem><para>Add variable search_path for schema search (Tom)</para></listitem>
8118 <listitem><para>Add ALTER TABLE SET/DROP NOT NULL (Christopher)</para></listitem>
8119 <listitem><para>New CREATE FUNCTION volatility levels (Tom)</para></listitem>
8120 <listitem><para>Make rule names unique only per table (Tom)</para></listitem>
8121 <listitem><para>Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)</para></listitem>
8122 <listitem><para>Add ALTER TRIGGER RENAME (Joe)</para></listitem>
8123 <listitem><para>New current_schema() and current_schemas() inquiry functions (Tom)</para></listitem>
8124 <listitem><para>Allow functions to return multiple rows (table functions) (Joe)</para></listitem>
8125 <listitem><para>Make WITH optional in CREATE DATABASE, for consistency (Bruce)</para></listitem>
8126 <listitem><para>Add object dependency tracking (Rod, Tom)</para></listitem>
8127 <listitem><para>Add RESTRICT/CASCADE to DROP commands (Rod)</para></listitem>
8128 <listitem><para>Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)</para></listitem>
8129 <listitem><para>Autodestroy sequence on DROP of table with SERIAL (Rod)</para></listitem>
8130 <listitem><para>Prevent column dropping if column is used by foreign key (Rod)</para></listitem>
8131 <listitem><para>Automatically drop constraints/functions when object is dropped (Rod)</para></listitem>
8132 <listitem><para>Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)</para></listitem>
8133 <listitem><para>Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)</para></listitem>
8134 <listitem><para>Prevent inherited columns from being removed or renamed (Alvaro Herrera)</para></listitem>
8135 <listitem><para>Fix foreign key constraints to not error on intermediate database states (Stephan)</para></listitem>
8136 <listitem><para>Propagate column or table renaming to foreign key constraints</para></listitem>
8137 <listitem><para>Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)</para></listitem>
8138 <listitem><para>Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)</para></listitem>
8139 <listitem><para>Have rules execute alphabetically, returning more predictable values (Tom)</para></listitem>
8140 <listitem><para>Triggers are now fired in alphabetical order (Tom)</para></listitem>
8141 <listitem><para>Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)</para></listitem>
8142 <listitem><para>Allow better casting when inserting/updating values (Tom)</para></listitem>
8147 <title>Utility Commands</title>
8149 <listitem><para>Have COPY TO output embedded carriage returns and newlines as \r and \n (Tom)</para></listitem>
8150 <listitem><para>Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)</para></listitem>
8151 <listitem><para>Make <application>pg_dump</> use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)</para></listitem>
8152 <listitem><para>Disable brackets in multistatement rules (Bruce)</para></listitem>
8153 <listitem><para>Disable VACUUM from being called inside a function (Bruce)</para></listitem>
8154 <listitem><para>Allow dropdb and other scripts to use identifiers with spaces (Bruce)</para></listitem>
8155 <listitem><para>Restrict database comment changes to the current database</para></listitem>
8156 <listitem><para>Allow comments on operators, independent of the underlying function (Rod)</para></listitem>
8157 <listitem><para>Rollback SET commands in aborted transactions (Tom)</para></listitem>
8158 <listitem><para>EXPLAIN now outputs as a query (Tom)</para></listitem>
8159 <listitem><para>Display condition expressions and sort keys in EXPLAIN (Tom)</para></listitem>
8160 <listitem><para>Add 'SET LOCAL var = value' to set configuration variables for a single transaction (Tom)</para></listitem>
8161 <listitem><para>Allow ANALYZE to run in a transaction (Bruce)</para></listitem>
8162 <listitem><para>Improve COPY syntax using new WITH clauses, keep backward compatibility (Bruce)</para></listitem>
8163 <listitem><para>Fix <application>pg_dump</> to consistently output tags in non-ASCII dumps (Bruce)</para></listitem>
8164 <listitem><para>Make foreign key constraints clearer in dump file (Rod)</para></listitem>
8165 <listitem><para>Add COMMENT ON CONSTRAINT (Rod)</para></listitem>
8166 <listitem><para>Allow COPY TO/FROM to specify column names (Brent Verner)</para></listitem>
8167 <listitem><para>Dump UNIQUE and PRIMARY KEY constraints as ALTER TABLE (Rod)</para></listitem>
8168 <listitem><para>Have SHOW output a query result (Joe)</para></listitem>
8169 <listitem><para>Generate failure on short COPY lines rather than pad NULLs (Neil)</para></listitem>
8170 <listitem><para>Fix CLUSTER to preserve all table attributes (Alvaro Herrera)</para></listitem>
8171 <listitem><para>New pg_settings table to view/modify GUC settings (Joe)</para></listitem>
8172 <listitem><para>Add smart quoting, portability improvements to <application>pg_dump</> output (Peter)</para></listitem>
8173 <listitem><para>Dump serial columns out as SERIAL (Tom)</para></listitem>
8174 <listitem><para>Enable large file support, >2G for <application>pg_dump</> (Peter, Philip Warner, Bruce)</para></listitem>
8175 <listitem><para>Disallow TRUNCATE on tables that are involved in referential constraints (Rod)</para></listitem>
8176 <listitem><para>Have TRUNCATE also auto-truncate the toast table of the relation (Tom)</para></listitem>
8177 <listitem><para>Add clusterdb utility that will auto-cluster an entire database based on previous CLUSTER operations (Alvaro Herrera)</para></listitem>
8178 <listitem><para>Overhaul pg_dumpall (Peter)</para></listitem>
8179 <listitem><para>Allow REINDEX of TOAST tables (Tom)</para></listitem>
8180 <listitem><para>Implemented START TRANSACTION, per SQL99 (Neil)</para></listitem>
8181 <listitem><para>Fix rare index corruption when a page split affects bulk delete (Tom)</para></listitem>
8182 <listitem><para>Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)</para></listitem>
8187 <title>Data Types and Functions</title>
8189 <listitem><para>Fix factorial(0) to return 1 (Bruce)</para></listitem>
8190 <listitem><para>Date/time/timezone improvements (Thomas)</para></listitem>
8191 <listitem><para>Fix for array slice extraction (Tom)</para></listitem>
8192 <listitem><para>Fix extract/date_part to report proper microseconds for timestamp (Tatsuo)</para></listitem>
8193 <listitem><para>Allow text_substr() and bytea_substr() to read TOAST values more efficiently (John Gray)</para></listitem>
8194 <listitem><para>Add domain support (Rod)</para></listitem>
8195 <listitem><para>Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types (Thomas)</para></listitem>
8196 <listitem><para>Allow alternate storage scheme of 64-bit integers for date/time types using --enable-integer-datetimes in configure (Thomas)</para></listitem>
8197 <listitem><para>Make timezone(timestamptz) return timestamp rather than a string (Thomas)</para></listitem>
8198 <listitem><para>Allow fractional seconds in date/time types for dates prior to 1BC (Thomas)</para></listitem>
8199 <listitem><para>Limit timestamp data types to 6 decimal places of precision (Thomas)</para></listitem>
8200 <listitem><para>Change timezone conversion functions from timetz() to timezone() (Thomas)</para></listitem>
8201 <listitem><para>Add configuration variables datestyle and timezone (Tom)</para></listitem>
8202 <listitem><para>Add OVERLAY(), which allows substitution of a substring in a string (Thomas)</para></listitem>
8203 <listitem><para>Add SIMILAR TO (Thomas, Tom)</para></listitem>
8204 <listitem><para>Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)</para></listitem>
8205 <listitem><para>Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)</para></listitem>
8206 <listitem><para>Add named composite types using CREATE TYPE typename AS (column) (Joe)</para></listitem>
8207 <listitem><para>Allow composite type definition in the table alias clause (Joe)</para></listitem>
8208 <listitem><para>Add new API to simplify creation of C language table functions (Joe)</para></listitem>
8209 <listitem><para>Remove ODBC-compatible empty parentheses from calls to SQL99 functions for which these parentheses do not match the standard (Thomas)</para></listitem>
8210 <listitem><para>Allow macaddr data type to accept 12 hex digits with no separators (Mike Wyer)</para></listitem>
8211 <listitem><para>Add CREATE/DROP CAST (Peter)</para></listitem>
8212 <listitem><para>Add IS DISTINCT FROM operator (Thomas)</para></listitem>
8213 <listitem><para>Add SQL99 TREAT() function, synonym for CAST() (Thomas)</para></listitem>
8214 <listitem><para>Add pg_backend_pid() to output backend pid (Bruce)</para></listitem>
8215 <listitem><para>Add IS OF / IS NOT OF type predicate (Thomas)</para></listitem>
8216 <listitem><para>Allow bit string constants without fully-specified length (Thomas)</para></listitem>
8217 <listitem><para>Allow conversion between 8-byte integers and bit strings (Thomas)</para></listitem>
8218 <listitem><para>Implement hex literal conversion to bit string literal (Thomas)</para></listitem>
8219 <listitem><para>Allow table functions to appear in the FROM clause (Joe)</para></listitem>
8220 <listitem><para>Increase maximum number of function parameters to 32 (Bruce)</para></listitem>
8221 <listitem><para>No longer automatically create index for SERIAL column (Tom)</para></listitem>
8222 <listitem><para>Add current_database() (Rod)</para></listitem>
8223 <listitem><para>Fix cash_words() to not overflow buffer (Tom)</para></listitem>
8224 <listitem><para>Add functions replace(), split_part(), to_hex() (Joe)</para></listitem>
8225 <listitem><para>Fix LIKE for bytea as a right-hand argument (Joe)</para></listitem>
8226 <listitem><para>Prevent crashes caused by SELECT cash_out(2) (Tom)</para></listitem>
8227 <listitem><para>Fix to_char(1,'FM999.99') to return a period (Karel)</para></listitem>
8228 <listitem><para>Fix trigger/type/language functions returning OPAQUE to return proper type (Tom)</para></listitem>
8233 <title>Internationalization</title>
8235 <listitem><para>Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC) (Eiji Tokuya)</para></listitem>
8236 <listitem><para>Enable locale support by default (Peter)</para></listitem>
8237 <listitem><para>Add locale variables (Peter)</para></listitem>
8238 <listitem><para>Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea (Tatsuo)</para></listitem>
8239 <listitem><para>Add locale awareness to regular expression character classes</para></listitem>
8240 <listitem><para>Enable multibyte support by default (Tatsuo)</para></listitem>
8241 <listitem><para>Add GB18030 multibyte support (Bill Huang)</para></listitem>
8242 <listitem><para>Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo, Kaori)</para></listitem>
8243 <listitem><para>Add pg_conversion table (Tatsuo)</para></listitem>
8244 <listitem><para>Add SQL99 CONVERT() function (Tatsuo)</para></listitem>
8245 <listitem><para>pg_dumpall, pg_controldata, and pg_resetxlog now national-language aware (Peter)</para></listitem>
8246 <listitem><para>New and updated translations</para></listitem>
8251 <title>Server-side Languages</title>
8253 <listitem><para>Allow recursive SQL function (Peter)</para></listitem>
8254 <listitem><para>Change PL/Tcl build to use configured compiler and Makefile.shlib (Peter)</para></listitem>
8255 <listitem><para>Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible (Neil, Tom)</para></listitem>
8256 <listitem><para>Allow PL/pgSQL to handle quoted identifiers (Tom)</para></listitem>
8257 <listitem><para>Allow set-returning PL/pgSQL functions (Neil)</para></listitem>
8258 <listitem><para>Make PL/pgSQL schema-aware (Joe)</para></listitem>
8259 <listitem><para>Remove some memory leaks (Nigel J. Andrews, Tom)</para></listitem>
8266 <listitem><para>Don't lowercase psql \connect database name for 7.2.0 compatibility (Tom)</para></listitem>
8267 <listitem><para>Add psql \timing to time user queries (Greg Sabino Mullane)</para></listitem>
8268 <listitem><para>Have psql \d show index information (Greg Sabino Mullane)</para></listitem>
8269 <listitem><para>New psql \dD shows domains (Jonathan Eisler)</para></listitem>
8270 <listitem><para>Allow psql to show rules on views (Paul ?)</para></listitem>
8271 <listitem><para>Fix for psql variable substitution (Tom)</para></listitem>
8272 <listitem><para>Allow psql \d to show temporary table structure (Tom)</para></listitem>
8273 <listitem><para>Allow psql \d to show foreign keys (Rod)</para></listitem>
8274 <listitem><para>Fix \? to honor \pset pager (Bruce)</para></listitem>
8275 <listitem><para>Have psql reports its version number on startup (Tom)</para></listitem>
8276 <listitem><para>Allow \copy to specify column names (Tom)</para></listitem>
8281 <title>libpq</title>
8283 <listitem><para>Add ~/.pgpass to store host/user password combinations (Alvaro Herrera)</para></listitem>
8284 <listitem><para>Add PQunescapeBytea() function to libpq (Patrick Welche)</para></listitem>
8285 <listitem><para>Fix for sending large queries over non-blocking connections (Bernhard Herzog)</para></listitem>
8286 <listitem><para>Fix for libpq using timers on Win9X (David Ford)</para></listitem>
8287 <listitem><para>Allow libpq notify to handle servers with different-length identifiers (Tom)</para></listitem>
8288 <listitem><para>Add libpq PQescapeString() and PQescapeBytea() to Windows (Bruce)</para></listitem>
8289 <listitem><para>Fix for SSL with non-blocking connections (Jack Bates)</para></listitem>
8290 <listitem><para>Add libpq connection timeout parameter (Denis A Ustimenko)</para></listitem>
8297 <listitem><para>Allow JDBC to compile with JDK 1.4 (Dave)</para></listitem>
8298 <listitem><para>Add JDBC 3 support (Barry)</para></listitem>
8299 <listitem><para>Allows JDBC to set loglevel by adding ?loglevel=X to the connection URL (Barry)</para></listitem>
8300 <listitem><para>Add Driver.info() message that prints out the version number (Barry)</para></listitem>
8301 <listitem><para>Add updateable result sets (Raghu Nidagal, Dave)</para></listitem>
8302 <listitem><para>Add support for callable statements (Paul Bethe)</para></listitem>
8303 <listitem><para>Add query cancel capability</para></listitem>
8304 <listitem><para>Add refresh row (Dave)</para></listitem>
8305 <listitem><para>Fix MD5 encryption handling for multibyte servers (Jun Kawai)</para></listitem>
8306 <listitem><para>Add support for prepared statements (Barry)</para></listitem>
8311 <title>Miscellaneous Interfaces</title>
8313 <listitem><para>Fixed ECPG bug concerning octal numbers in single quotes (Michael)</para></listitem>
8314 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8315 <listitem><para>Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)</para></listitem>
8316 <listitem><para>Add libpgtcl connection close event (Gerhard Hintermayer)</para></listitem>
8317 <listitem><para>Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8318 <listitem><para>Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)</para></listitem>
8319 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8320 <listitem><para>Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8321 <listitem><para>Remove src/bin/pgaccess from main tree, now at http://www.pgaccess.org (Bruce)</para></listitem>
8322 <listitem><para>Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer, Tom)</para></listitem>
8327 <title>Source Code</title>
8329 <listitem><para>Fix for parallel make (Peter)</para></listitem>
8330 <listitem><para>AIX fixes for linking Tcl (Andreas Zeugswetter)</para></listitem>
8331 <listitem><para>Allow PL/Perl to build under Cygwin (Jason Tishler)</para></listitem>
8332 <listitem><para>Improve MIPS compiles (Peter, Oliver Elphick)</para></listitem>
8333 <listitem><para>Require Autoconf version 2.53 (Peter)</para></listitem>
8334 <listitem><para>Require readline and zlib by default in configure (Peter)</para></listitem>
8335 <listitem><para>Allow Solaris to use Intimate Shared Memory (ISM), for performance (Scott Brunza, P.J. Josh Rovero)</para></listitem>
8336 <listitem><para>Always enable syslog in compile, remove --enable-syslog option (Tatsuo)</para></listitem>
8337 <listitem><para>Always enable multibyte in compile, remove --enable-multibyte option (Tatsuo)</para></listitem>
8338 <listitem><para>Always enable locale in compile, remove --enable-locale option (Peter)</para></listitem>
8339 <listitem><para>Fix for Win9x DLL creation (Magnus Naeslund)</para></listitem>
8340 <listitem><para>Fix for link() usage by WAL code on Windows, BeOS (Jason Tishler)</para></listitem>
8341 <listitem><para>Add sys/types.h to c.h, remove from main files (Peter, Bruce)</para></listitem>
8342 <listitem><para>Fix AIX hang on SMP machines (Tomoyuki Niijima)</para></listitem>
8343 <listitem><para>AIX SMP hang fix (Tomoyuki Niijima)</para></listitem>
8344 <listitem><para>Fix pre-1970 date handling on newer glibc libraries (Tom)</para></listitem>
8345 <listitem><para>Fix PowerPC SMP locking (Tom)</para></listitem>
8346 <listitem><para>Prevent gcc -ffast-math from being used (Peter, Tom)</para></listitem>
8347 <listitem><para>Bison >= 1.50 now required for developer builds</para></listitem>
8348 <listitem><para>Kerberos 5 support now builds with Heimdal (Peter)</para></listitem>
8349 <listitem><para>Add appendix in the User's Guide which lists SQL features (Thomas)</para></listitem>
8350 <listitem><para>Improve loadable module linking to use RTLD_NOW (Tom)</para></listitem>
8351 <listitem><para>New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)</para></listitem>
8352 <listitem><para>New src/port directory holds replaced libc functions (Peter, Bruce)</para></listitem>
8353 <listitem><para>New pg_namespace system catalog for schemas (Tom)</para></listitem>
8354 <listitem><para>Add pg_class.relnamespace for schemas (Tom)</para></listitem>
8355 <listitem><para>Add pg_type.typnamespace for schemas (Tom)</para></listitem>
8356 <listitem><para>Add pg_proc.pronamespace for schemas (Tom)</para></listitem>
8357 <listitem><para>Restructure aggregates to have pg_proc entries (Tom)</para></listitem>
8358 <listitem><para>System relations now have their own namespace, pg_* test not required (Fernando Nasser)</para></listitem>
8359 <listitem><para>Rename TOAST index names to be *_index rather than *_idx (Neil)</para></listitem>
8360 <listitem><para>Add namespaces for operators, opclasses (Tom)</para></listitem>
8361 <listitem><para>Add additional checks to server control file (Thomas)</para></listitem>
8362 <listitem><para>New Polish FAQ (Marcin Mazurek)</para></listitem>
8363 <listitem><para>Add Posix semaphore support (Tom)</para></listitem>
8364 <listitem><para>Document need for reindex (Bruce)</para></listitem>
8365 <listitem><para>Rename some internal identifiers to simplify Windows compile (Jan, Katherine Ward)</para></listitem>
8366 <listitem><para>Add documentation on computing disk space (Bruce)</para></listitem>
8367 <listitem><para>Remove KSQO from GUC (Bruce)</para></listitem>
8368 <listitem><para>Fix memory leak in rtree (Kenneth Been)</para></listitem>
8369 <listitem><para>Modify a few error messages for consistency (Bruce)</para></listitem>
8370 <listitem><para>Remove unused system table columns (Peter)</para></listitem>
8371 <listitem><para>Make system columns NOT NULL where appropriate (Tom)</para></listitem>
8372 <listitem><para>Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)</para></listitem>
8373 <listitem><para>Remove OPAQUE and create specific subtypes (Tom)</para></listitem>
8374 <listitem><para>Cleanups in array internal handling (Joe, Tom)</para></listitem>
8375 <listitem><para>Disallow pg_atoi('') (Bruce)</para></listitem>
8376 <listitem><para>Remove parameter wal_files because WAL files are now recycled (Bruce)</para></listitem>
8377 <listitem><para>Add version numbers to heap pages (Tom)</para></listitem>
8382 <title>Contrib</title>
8384 <listitem><para>Allow inet arrays in /contrib/array (Neil)</para></listitem>
8385 <listitem><para>GiST fixes (Teodor Sigaev, Neil)</para></listitem>
8386 <listitem><para>Upgrade /contrib/mysql</para></listitem>
8387 <listitem><para>Add /contrib/dbsize which shows table sizes without vacuum (Peter)</para></listitem>
8388 <listitem><para>Add /contrib/intagg, integer aggregator routines (mlw)</para></listitem>
8389 <listitem><para>Improve /contrib/oid2name (Neil, Bruce)</para></listitem>
8390 <listitem><para>Improve /contrib/tsearch (Oleg, Teodor Sigaev)</para></listitem>
8391 <listitem><para>Cleanups of /contrib/rserver (Alexey V. Borzov)</para></listitem>
8392 <listitem><para>Update /contrib/oracle conversion utility (Gilles Darold)</para></listitem>
8393 <listitem><para>Update /contrib/dblink (Joe)</para></listitem>
8394 <listitem><para>Improve options supported by /contrib/vacuumlo (Mario Weilguni)</para></listitem>
8395 <listitem><para>Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey Oktyabrski)</para></listitem>
8396 <listitem><para>Add /contrib/reindexdb utility (Shaun Thomas)</para></listitem>
8397 <listitem><para>Add indexing to /contrib/isbn_issn (Dan Weston)</para></listitem>
8398 <listitem><para>Add /contrib/dbmirror (Steven Singer)</para></listitem>
8399 <listitem><para>Improve /contrib/pgbench (Neil)</para></listitem>
8400 <listitem><para>Add /contrib/tablefunc table function examples (Joe)</para></listitem>
8401 <listitem><para>Add /contrib/ltree data type for tree structures (Teodor Sigaev, Oleg Bartunov)</para></listitem>
8402 <listitem><para>Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)</para></listitem>
8403 <listitem><para>Fixes to /contrib/cube (Bruno Wolff)</para></listitem>
8404 <listitem><para>Improve /contrib/fulltextindex (Christopher)</para></listitem>
8411 <sect1 id="release-7-2-8">
8412 <title>Release 7.2.8</title>
8415 <title>Release date</title>
8416 <simpara>2005-05-09</simpara>
8420 This release contains a variety of fixes from 7.2.7, including one
8421 security-related issue.
8425 <title>Migration to version 7.2.8</title>
8428 A dump/restore is not required for those running 7.2.X.
8433 <title>Changes</title>
8436 <listitem><para>Repair ancient race condition that allowed a transaction to be
8437 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
8438 than for other purposes</para>
8439 <para>This is an extremely serious bug since it could lead to apparent
8440 data inconsistencies being briefly visible to applications.</para></listitem>
8441 <listitem><para>Repair race condition between relation extension and
8443 <para>This could theoretically have caused loss of a page's worth of
8444 freshly-inserted data, although the scenario seems of very low probability.
8445 There are no known cases of it having caused more than an Assert failure.
8447 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
8448 <type>TIME WITH TIME ZONE</> values</para></listitem>
8449 <listitem><para>Additional buffer overrun checks in plpgsql
8450 (Neil)</para></listitem>
8451 <listitem><para>Fix pg_dump to dump index names and trigger names containing
8452 <literal>%</> correctly (Neil)</para></listitem>
8453 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
8454 month-related formats</para></listitem>
8455 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
8456 (Marko Kreen)</para></listitem>
8462 <sect1 id="release-7-2-7">
8463 <title>Release 7.2.7</title>
8466 <title>Release date</title>
8467 <simpara>2005-01-31</simpara>
8471 This release contains a variety of fixes from 7.2.6, including several
8472 security-related issues.
8476 <title>Migration to version 7.2.7</title>
8479 A dump/restore is not required for those running 7.2.X.
8484 <title>Changes</title>
8487 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
8489 On platforms that will automatically execute initialization functions of a
8490 shared library (this includes at least Windows and ELF-based Unixen),
8491 <command>LOAD</> can be used to make the server execute arbitrary code.
8492 Thanks to NGS Software for reporting this.</para></listitem>
8493 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
8494 Jurka)</para></listitem>
8495 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
8496 many parameters (Neil)</para></listitem>
8497 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
8499 The result of the join was mistakenly supposed to be sorted the same as the
8500 left input. This could not only deliver mis-sorted output to the user, but
8501 in case of nested merge joins could give outright wrong answers.
8503 <listitem><para>Fix display of negative intervals in SQL and GERMAN
8504 datestyles</para></listitem>
8510 <sect1 id="release-7-2-6">
8511 <title>Release 7.2.6</title>
8514 <title>Release date</title>
8515 <simpara>2004-10-22</simpara>
8519 This release contains a variety of fixes from 7.2.5.
8524 <title>Migration to version 7.2.6</title>
8527 A dump/restore is not required for those running 7.2.X.
8532 <title>Changes</title>
8535 <listitem><para>Repair possible failure to update hint bits on disk</para>
8537 Under rare circumstances this oversight could lead to
8538 <quote>could not access transaction status</> failures, which qualifies
8539 it as a potential-data-loss bug.
8541 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
8543 Very large left joins using a hash join plan could fail to output unmatched
8544 left-side rows given just the right data distribution.
8546 <listitem><para>Disallow running pg_ctl as root</para>
8548 This is to guard against any possible security issues.
8550 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
8552 This has been reported as a security issue, though it's hardly worthy of
8553 concern since there is no reason for non-developers to use this script anyway.
8555 <listitem><para>Update to newer versions of Bison</para></listitem>
8561 <sect1 id="release-7-2-5">
8562 <title>Release 7.2.5</title>
8565 <title>Release date</title>
8566 <simpara>2004-08-16</simpara>
8570 This release contains a variety of fixes from 7.2.4.
8575 <title>Migration to version 7.2.5</title>
8578 A dump/restore is not required for those running 7.2.X.
8583 <title>Changes</title>
8586 <listitem><para>Prevent possible loss of committed transactions during crash</para>
8588 Due to insufficient interlocking between transaction commit and checkpointing,
8589 it was possible for transactions committed just before the most recent
8590 checkpoint to be lost, in whole or in part, following a database crash and
8591 restart. This is a serious bug that has existed
8592 since <productname>PostgreSQL</productname> 7.1.
8594 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
8595 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
8596 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
8597 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
8598 <listitem><para>Repair memory leaks in pg_dump</para></listitem>
8599 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
8604 <sect1 id="release-7-2-4">
8605 <title>Release 7.2.4</title>
8608 <title>Release date</title>
8609 <simpara>2003-01-30</simpara>
8613 This release contains a variety of fixes for version 7.2.3,
8614 including fixes to prevent possible data loss.
8618 <title>Migration to version 7.2.4</title>
8621 A dump/restore is <emphasis>not</emphasis> required for those
8622 running version 7.2.*.
8627 <title>Changes</title>
8630 <listitem><para>Fix some additional cases of VACUUM "No one parent tuple was found" error</para></listitem>
8631 <listitem><para>Prevent VACUUM from being called inside a function (Bruce)</para></listitem>
8632 <listitem><para>Ensure pg_clog updates are sync'd to disk before marking checkpoint complete</para></listitem>
8633 <listitem><para>Avoid integer overflow during large hash joins</para></listitem>
8634 <listitem><para>Make GROUP commands work when pg_group.grolist is large enough to be toasted</para></listitem>
8635 <listitem><para>Fix errors in datetime tables; some timezone names weren't being recognized</para></listitem>
8636 <listitem><para>Fix integer overflows in circle_poly(), path_encode(), path_add() (Neil)</para></listitem>
8637 <listitem><para>Repair long-standing logic errors in lseg_eq(), lseg_ne(), lseg_center()</para></listitem>
8643 <sect1 id="release-7-2-3">
8644 <title>Release 7.2.3</title>
8647 <title>Release date</title>
8648 <simpara>2002-10-01</simpara>
8652 This release contains a variety of fixes for version 7.2.2,
8653 including fixes to prevent possible data loss.
8657 <title>Migration to version 7.2.3</title>
8660 A dump/restore is <emphasis>not</emphasis> required for those
8661 running version 7.2.*.
8666 <title>Changes</title>
8669 <listitem><para>Prevent possible compressed transaction log loss (Tom)</para></listitem>
8670 <listitem><para>Prevent non-superuser from increasing most recent vacuum info (Tom)</para></listitem>
8671 <listitem><para>Handle pre-1970 date values in newer versions of glibc (Tom)</para></listitem>
8672 <listitem><para>Fix possible hang during server shutdown</para></listitem>
8673 <listitem><para>Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)</para></listitem>
8674 <listitem><para>Fix <application>pg_dump</> to properly dump FULL JOIN USING (Tom)</para></listitem>
8680 <sect1 id="release-7-2-2">
8681 <title>Release 7.2.2</title>
8684 <title>Release date</title>
8685 <simpara>2002-08-23</simpara>
8689 This release contains a variety of fixes for version 7.2.1.
8693 <title>Migration to version 7.2.2</title>
8696 A dump/restore is <emphasis>not</emphasis> required for those
8697 running version 7.2.*.
8702 <title>Changes</title>
8705 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8706 <listitem><para>Fix for compressed transaction log id wraparound (Tom)</para></listitem>
8707 <listitem><para>Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Tatsuo)</para></listitem>
8708 <listitem><para>Fix for psql and <application>pg_dump</> crashing when invoked with non-existent long options (Tatsuo)</para></listitem>
8709 <listitem><para>Fix crash when invoking geometric operators (Tom)</para></listitem>
8710 <listitem><para>Allow OPEN cursor(args) (Tom)</para></listitem>
8711 <listitem><para>Fix for rtree_gist index build (Teodor)</para></listitem>
8712 <listitem><para>Fix for dumping user-defined aggregates (Tom)</para></listitem>
8713 <listitem><para>contrib/intarray fixes (Oleg)</para></listitem>
8714 <listitem><para>Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)</para></listitem>
8715 <listitem><para>Fix to pg_convert (Tatsuo)</para></listitem>
8716 <listitem><para>Fix for crash with long DATA strings (Thomas, Neil)</para></listitem>
8717 <listitem><para>Fix for repeat(), lpad(), rpad() and long strings (Neil)</para></listitem>
8723 <sect1 id="release-7-2-1">
8724 <title>Release 7.2.1</title>
8727 <title>Release date</title>
8728 <simpara>2002-03-21</simpara>
8732 This release contains a variety of fixes for version 7.2.
8736 <title>Migration to version 7.2.1</title>
8739 A dump/restore is <emphasis>not</emphasis> required for those
8740 running version 7.2.
8745 <title>Changes</title>
8748 <listitem><para>Ensure that sequence counters do not go backwards after a crash (Tom)</para></listitem>
8749 <listitem><para>Fix pgaccess kanji-conversion key binding (Tatsuo)</para></listitem>
8750 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8751 <listitem><para>Cash I/O improvements (Tom)</para></listitem>
8752 <listitem><para>New Russian FAQ</para></listitem>
8753 <listitem><para>Compile fix for missing AuthBlockSig (Heiko)</para></listitem>
8754 <listitem><para>Additional time zones and time zone fixes (Thomas)</para></listitem>
8755 <listitem><para>Allow psql \connect to handle mixed case database and user names (Tom)</para></listitem>
8756 <listitem><para>Return proper OID on command completion even with ON INSERT rules (Tom)</para></listitem>
8757 <listitem><para>Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)</para></listitem>
8758 <listitem><para>Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)</para></listitem>
8759 <listitem><para>Improve handling of multiple UNIONs with different lengths (Tom)</para></listitem>
8760 <listitem><para>contrib/btree_gist improvements (Teodor Sigaev)</para></listitem>
8761 <listitem><para>contrib/tsearch dictionary improvements, see README.tsearch for an additional installation step (Thomas T. Thai, Teodor Sigaev)</para></listitem>
8762 <listitem><para>Fix for array subscripts handling (Tom)</para></listitem>
8763 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8769 <sect1 id="release-7-2">
8770 <title>Release 7.2</title>
8773 <title>Release date</title>
8774 <simpara>2002-02-04</simpara>
8778 <title>Overview</title>
8781 This release improves <productname>PostgreSQL</> for use in
8782 high-volume applications.
8786 Major changes in this release:
8794 Vacuuming no longer locks tables, thus allowing normal user
8795 access during the vacuum. A new <command>VACUUM FULL</>
8796 command does old-style vacuum by locking the table and
8797 shrinking the on-disk copy of the table.
8803 <term>Transactions</term>
8806 There is no longer a problem with installations that exceed
8807 four billion transactions.
8816 OIDs are now optional. Users can now create tables without
8817 OIDs for cases where OID usage is excessive.
8823 <term>Optimizer</term>
8826 The system now computes histogram column statistics during
8827 <command>ANALYZE</>, allowing much better optimizer choices.
8833 <term>Security</term>
8836 A new MD5 encryption option allows more secure storage and
8837 transfer of passwords. A new Unix-domain socket
8838 authentication option is available on Linux and BSD systems.
8844 <term>Statistics</term>
8847 Administrators can use the new table access statistics module
8848 to get fine-grained information about table and index usage.
8854 <term>Internationalization</term>
8857 Program and library messages can now be displayed in several
8867 <title>Migration to version 7.2</title>
8870 A dump/restore using <command>pg_dump</command> is required for
8871 those wishing to migrate data from any previous release.
8875 Observe the following incompatibilities:
8881 The semantics of the <command>VACUUM</command> command have
8882 changed in this release. You may wish to update your
8883 maintenance procedures accordingly.
8889 In this release, comparisons using <literal>= NULL</literal>
8890 will always return false (or NULL, more precisely). Previous
8891 releases automatically transformed this syntax to <literal>IS
8892 NULL</literal>. The old behavior can be re-enabled using a
8893 <filename>postgresql.conf</filename> parameter.
8899 The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
8900 configuration is now only reloaded after receiving a
8901 <systemitem>SIGHUP</> signal, not with each connection.
8907 The function <filename>octet_length()</> now returns the uncompressed data length.
8913 The date/time value <literal>'current'</literal> is no longer
8914 available. You will need to rewrite your applications.
8920 The <literal>timestamp()</literal>, <literal>time()</literal>,
8921 and <literal>interval()</literal> functions are no longer
8922 available. Instead of <literal>timestamp()</literal>, use
8923 <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
8930 The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
8931 in the next release. You should change your queries to use
8932 separate LIMIT and OFFSET clauses, e.g. <literal>LIMIT 10 OFFSET
8938 <title>Changes</title>
8941 <title>Server Operation</title>
8943 <listitem><para>Create temporary files in a separate directory (Bruce)</para></listitem>
8944 <listitem><para>Delete orphaned temporary files on postmaster startup (Bruce)</para></listitem>
8945 <listitem><para>Added unique indexes to some system tables (Tom)</para></listitem>
8946 <listitem><para>System table operator reorganization (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
8947 <listitem><para>Renamed pg_log to pg_clog (Tom)</para></listitem>
8948 <listitem><para>Enable SIGTERM, SIGQUIT to kill backends (Jan)</para></listitem>
8949 <listitem><para>Removed compile-time limit on number of backends (Tom)</para></listitem>
8950 <listitem><para>Better cleanup for semaphore resource failure (Tatsuo, Tom)</para></listitem>
8951 <listitem><para>Allow safe transaction ID wraparound (Tom)</para></listitem>
8952 <listitem><para>Removed OIDs from some system tables (Tom)</para></listitem>
8953 <listitem><para>Removed "triggered data change violation" error check (Tom)</para></listitem>
8954 <listitem><para>SPI portal creation of prepared/saved plans (Jan)</para></listitem>
8955 <listitem><para>Allow SPI column functions to work for system columns (Tom)</para></listitem>
8956 <listitem><para>Long value compression improvement (Tom)</para></listitem>
8957 <listitem><para>Statistics collector for table, index access (Jan)</para></listitem>
8958 <listitem><para>Truncate extra-long sequence names to a reasonable value (Tom)</para></listitem>
8959 <listitem><para>Measure transaction times in milliseconds (Thomas)</para></listitem>
8960 <listitem><para>Fix TID sequential scans (Hiroshi)</para></listitem>
8961 <listitem><para>Superuser ID now fixed at 1 (Peter E)</para></listitem>
8962 <listitem><para>New pg_ctl "reload" option (Tom)</para></listitem>
8967 <title>Performance</title>
8969 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8970 <listitem><para>New histogram column statistics for optimizer (Tom)</para></listitem>
8971 <listitem><para>Reuse write-ahead log files rather than discarding them (Tom)</para></listitem>
8972 <listitem><para>Cache improvements (Tom)</para></listitem>
8973 <listitem><para>IS NULL, IS NOT NULL optimizer improvement (Tom)</para></listitem>
8974 <listitem><para>Improve lock manager to reduce lock contention (Tom)</para></listitem>
8975 <listitem><para>Keep relcache entries for index access support functions (Tom)</para></listitem>
8976 <listitem><para>Allow better selectivity with NaN and infinities in NUMERIC (Tom)</para></listitem>
8977 <listitem><para>R-tree performance improvements (Kenneth Been)</para></listitem>
8978 <listitem><para>B-tree splits more efficient (Tom)</para></listitem>
8983 <title>Privileges</title>
8985 <listitem><para>Change UPDATE, DELETE privileges to be distinct (Peter E)</para></listitem>
8986 <listitem><para>New REFERENCES, TRIGGER privileges (Peter E)</para></listitem>
8987 <listitem><para>Allow GRANT/REVOKE to/from more than one user at a time (Peter E)</para></listitem>
8988 <listitem><para>New has_table_privilege() function (Joe Conway)</para></listitem>
8989 <listitem><para>Allow non-superuser to vacuum database (Tom)</para></listitem>
8990 <listitem><para>New SET SESSION AUTHORIZATION command (Peter E)</para></listitem>
8991 <listitem><para>Fix bug in privilege modifications on newly created tables (Tom)</para></listitem>
8992 <listitem><para>Disallow access to pg_statistic for non-superuser, add user-accessible views (Tom)</para></listitem>
8997 <title>Client Authentication</title>
8999 <listitem><para>Fork postmaster before doing authentication to prevent hangs (Peter E)</para></listitem>
9000 <listitem><para>Add ident authentication over Unix domain sockets on Linux, *BSD (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)</para></listitem>
9001 <listitem><para>Add a password authentication method that uses MD5 encryption (Bruce)</para></listitem>
9002 <listitem><para>Allow encryption of stored passwords using MD5 (Bruce)</para></listitem>
9003 <listitem><para>PAM authentication (Dominic J. Eidson)</para></listitem>
9004 <listitem><para>Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)</para></listitem>
9009 <title>Server Configuration</title>
9011 <listitem><para>Interpretation of some time zone abbreviations as Australian rather than North American now settable at run time (Bruce)</para></listitem>
9012 <listitem><para>New parameter to set default transaction isolation level (Peter E)</para></listitem>
9013 <listitem><para>New parameter to enable conversion of "expr = NULL" into "expr IS NULL", off by default (Peter E)</para></listitem>
9014 <listitem><para>New parameter to control memory usage by VACUUM (Tom)</para></listitem>
9015 <listitem><para>New parameter to set client authentication timeout (Tom)</para></listitem>
9016 <listitem><para>New parameter to set maximum number of open files (Tom)</para></listitem>
9021 <title>Queries</title>
9023 <listitem><para>Statements added by INSERT rules now execute after the INSERT (Jan)</para></listitem>
9024 <listitem><para>Prevent unadorned relation names in target list (Bruce)</para></listitem>
9025 <listitem><para>NULLs now sort after all normal values in ORDER BY (Tom)</para></listitem>
9026 <listitem><para>New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)</para></listitem>
9027 <listitem><para>New SHARE UPDATE EXCLUSIVE lock mode (Tom)</para></listitem>
9028 <listitem><para>New EXPLAIN ANALYZE command that shows run times and row counts (Martijn van Oosterhout)</para></listitem>
9029 <listitem><para>Fix problem with LIMIT and subqueries (Tom)</para></listitem>
9030 <listitem><para>Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)</para></listitem>
9031 <listitem><para>Fix nested EXCEPT/INTERSECT (Tom)</para></listitem>
9036 <title>Schema Manipulation</title>
9038 <listitem><para>Fix SERIAL in temporary tables (Bruce)</para></listitem>
9039 <listitem><para>Allow temporary sequences (Bruce)</para></listitem>
9040 <listitem><para>Sequences now use int8 internally (Tom)</para></listitem>
9041 <listitem><para>New SERIAL8 creates int8 columns with sequences, default still SERIAL4 (Tom)</para></listitem>
9042 <listitem><para>Make OIDs optional using WITHOUT OIDS (Tom)</para></listitem>
9043 <listitem><para>Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)</para></listitem>
9044 <listitem><para>Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher Kings-Lynne)</para></listitem>
9045 <listitem><para>New CREATE OR REPLACE FUNCTION to alter existing function (preserving the function OID) (Gavin Sherry)</para></listitem>
9046 <listitem><para>Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)</para></listitem>
9047 <listitem><para>Allow column renaming in views</para></listitem>
9048 <listitem><para>Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent Verner)</para></listitem>
9049 <listitem><para>Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (Stephan Szabo)</para></listitem>
9050 <listitem><para>ALTER TABLE RENAME update foreign-key trigger arguments correctly (Brent Verner)</para></listitem>
9051 <listitem><para>DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)</para></listitem>
9052 <listitem><para>Add automatic return type data casting for SQL functions (Tom)</para></listitem>
9053 <listitem><para>Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
9054 <listitem><para>Enable partial indexes (Martijn van Oosterhout)</para></listitem>
9059 <title>Utility Commands</title>
9061 <listitem><para>Add RESET ALL, SHOW ALL (Marko Kreen)</para></listitem>
9062 <listitem><para>CREATE/ALTER USER/GROUP now allow options in any order (Vince)</para></listitem>
9063 <listitem><para>Add LOCK A, B, C functionality (Neil Padgett)</para></listitem>
9064 <listitem><para>New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)</para></listitem>
9065 <listitem><para>New light-weight VACUUM does not lock table; old semantics are available as VACUUM FULL (Tom)</para></listitem>
9066 <listitem><para>Disable COPY TO/FROM on views (Bruce)</para></listitem>
9067 <listitem><para>COPY DELIMITERS string must be exactly one character (Tom)</para></listitem>
9068 <listitem><para>VACUUM warning about index tuples fewer than heap now only appears when appropriate (Martijn van Oosterhout)</para></listitem>
9069 <listitem><para>Fix privilege checks for CREATE INDEX (Tom)</para></listitem>
9070 <listitem><para>Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)</para></listitem>
9075 <title>Data Types and Functions</title>
9077 <listitem><para>SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)</para></listitem>
9078 <listitem><para>Add convert(), convert2() (Tatsuo)</para></listitem>
9079 <listitem><para>New function bit_length() (Peter E)</para></listitem>
9080 <listitem><para>Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tatsuo)</para></listitem>
9081 <listitem><para>CHAR(), VARCHAR() now reject strings that are too long (Peter E)</para></listitem>
9082 <listitem><para>BIT VARYING now rejects bit strings that are too long (Peter E)</para></listitem>
9083 <listitem><para>BIT now rejects bit strings that do not match declared size (Peter E)</para></listitem>
9084 <listitem><para>INET, CIDR text conversion functions (Alex Pilosov)</para></listitem>
9085 <listitem><para>INET, CIDR operators << and <<= indexable (Alex Pilosov)</para></listitem>
9086 <listitem><para>Bytea \### now requires valid three digit octal number</para></listitem>
9087 <listitem><para>Bytea comparison improvements, now supports =, <>, >, >=, <, and <=</para></listitem>
9088 <listitem><para>Bytea now supports B-tree indexes</para></listitem>
9089 <listitem><para>Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE</para></listitem>
9090 <listitem><para>Bytea now supports concatenation</para></listitem>
9091 <listitem><para>New bytea functions: position, substring, trim, btrim, and length</para></listitem>
9092 <listitem><para>New encode() function mode, "escaped", converts minimally escaped bytea to/from text</para></listitem>
9093 <listitem><para>Add pg_database_encoding_max_length() (Tatsuo)</para></listitem>
9094 <listitem><para>Add pg_client_encoding() function (Tatsuo)</para></listitem>
9095 <listitem><para>now() returns time with millisecond precision (Thomas)</para></listitem>
9096 <listitem><para>New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)</para></listitem>
9097 <listitem><para>Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)</para></listitem>
9098 <listitem><para>New xid/int comparison functions (Hiroshi)</para></listitem>
9099 <listitem><para>Add precision to TIME, TIMESTAMP, and INTERVAL data types (Thomas)</para></listitem>
9100 <listitem><para>Modify type coercion logic to attempt binary-compatible functions first (Tom)</para></listitem>
9101 <listitem><para>New encode() function installed by default (Marko Kreen)</para></listitem>
9102 <listitem><para>Improved to_*() conversion functions (Karel Zak)</para></listitem>
9103 <listitem><para>Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)</para></listitem>
9104 <listitem><para>New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_salt() (Marko Kreen)</para></listitem>
9105 <listitem><para>Correct description of translate() function (Bruce)</para></listitem>
9106 <listitem><para>Add INTERVAL argument for SET TIME ZONE (Thomas)</para></listitem>
9107 <listitem><para>Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)</para></listitem>
9108 <listitem><para>Optimize length functions when using single-byte encodings (Tatsuo)</para></listitem>
9109 <listitem><para>Fix path_inter, path_distance, path_length, dist_ppath to handle closed paths (Curtis Barrett, Tom)</para></listitem>
9110 <listitem><para>octet_length(text) now returns non-compressed length (Tatsuo, Bruce)</para></listitem>
9111 <listitem><para>Handle "July" full name in date/time literals (Greg Sabino Mullane)</para></listitem>
9112 <listitem><para>Some datatype() function calls now evaluated differently</para></listitem>
9113 <listitem><para>Add support for Julian and ISO time specifications (Thomas)</para></listitem>
9118 <title>Internationalization</title>
9120 <listitem><para>National language support in psql, <application>pg_dump</>, libpq, and server (Peter E)</para></listitem>
9121 <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>
9122 <listitem><para>Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware (Tatsuo)</para></listitem>
9123 <listitem><para>Add LATIN5,6,7,8,9,10 support (Tatsuo)</para></listitem>
9124 <listitem><para>Add ISO 8859-5,6,7,8 support (Tatsuo)</para></listitem>
9125 <listitem><para>Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)</para></listitem>
9126 <listitem><para>Make mic2ascii() non-ASCII aware (Tatsuo)</para></listitem>
9127 <listitem><para>Reject invalid multibyte character sequences (Tatsuo)</para></listitem>
9132 <title><application>PL/pgSQL</></title>
9134 <listitem><para>Now uses portals for SELECT loops, allowing huge result sets (Jan)</para></listitem>
9135 <listitem><para>CURSOR and REFCURSOR support (Jan)</para></listitem>
9136 <listitem><para>Can now return open cursors (Jan)</para></listitem>
9137 <listitem><para>Add ELSEIF (Klaus Reger)</para></listitem>
9138 <listitem><para>Improve PL/pgSQL error reporting, including location of error (Tom)</para></listitem>
9139 <listitem><para>Allow IS or FOR key words in cursor declaration, for compatibility (Bruce)</para></listitem>
9140 <listitem><para>Fix for SELECT ... FOR UPDATE (Tom)</para></listitem>
9141 <listitem><para>Fix for PERFORM returning multiple rows (Tom)</para></listitem>
9142 <listitem><para>Make PL/pgSQL use the server's type coercion code (Tom)</para></listitem>
9143 <listitem><para>Memory leak fix (Jan, Tom)</para></listitem>
9144 <listitem><para>Make trailing semicolon optional (Tom)</para></listitem>
9149 <title>PL/Perl</title>
9151 <listitem><para>New untrusted PL/Perl (Alex Pilosov)</para></listitem>
9152 <listitem><para>PL/Perl is now built on some platforms even if libperl is not shared (Peter E)</para></listitem>
9157 <title>PL/Tcl</title>
9159 <listitem><para>Now reports errorInfo (Vsevolod Lobko)</para></listitem>
9160 <listitem><para>Add spi_lastoid function (bob@redivi.com)</para></listitem>
9165 <title>PL/Python</title>
9167 <listitem><para>...is new (Andrew Bosma)</para></listitem>
9172 <title><application>psql</></title>
9174 <listitem><para>\d displays indexes in unique, primary groupings (Christopher Kings-Lynne)</para></listitem>
9175 <listitem><para>Allow trailing semicolons in backslash commands (Greg Sabino Mullane)</para></listitem>
9176 <listitem><para>Read password from /dev/tty if possible</para></listitem>
9177 <listitem><para>Force new password prompt when changing user and database (Tatsuo, Tom)</para></listitem>
9178 <listitem><para>Format the correct number of columns for Unicode (Patrice)</para></listitem>
9183 <title><application>libpq</></title>
9185 <listitem><para>New function PQescapeString() to escape quotes in command strings (Florian Weimer)</para></listitem>
9186 <listitem><para>New function PQescapeBytea() escapes binary strings for use as SQL string literals</para></listitem>
9193 <listitem><para>Return OID of INSERT (Ken K)</para></listitem>
9194 <listitem><para>Handle more data types (Ken K)</para></listitem>
9195 <listitem><para>Handle single quotes and newlines in strings (Ken K)</para></listitem>
9196 <listitem><para>Handle NULL variables (Ken K)</para></listitem>
9197 <listitem><para>Fix for time zone handling (Barry Lind)</para></listitem>
9198 <listitem><para>Improved Druid support</para></listitem>
9199 <listitem><para>Allow eight-bit characters with non-multibyte server (Barry Lind)</para></listitem>
9200 <listitem><para>Support BIT, BINARY types (Ned Wolpert)</para></listitem>
9201 <listitem><para>Reduce memory usage (Michael Stephens, Dave Cramer)</para></listitem>
9202 <listitem><para>Update DatabaseMetaData (Peter E)</para></listitem>
9203 <listitem><para>Add DatabaseMetaData.getCatalogs() (Peter E)</para></listitem>
9204 <listitem><para>Encoding fixes (Anders Bengtsson)</para></listitem>
9205 <listitem><para>Get/setCatalog methods (Jason Davies)</para></listitem>
9206 <listitem><para>DatabaseMetaData.getColumns() now returns column defaults (Jason Davies)</para></listitem>
9207 <listitem><para>DatabaseMetaData.getColumns() performance improvement (Jeroen van Vianen)</para></listitem>
9208 <listitem><para>Some JDBC1 and JDBC2 merging (Anders Bengtsson)</para></listitem>
9209 <listitem><para>Transaction performance improvements (Barry Lind)</para></listitem>
9210 <listitem><para>Array fixes (Greg Zoller)</para></listitem>
9211 <listitem><para>Serialize addition </para></listitem>
9212 <listitem><para>Fix batch processing (Rene Pijlman)</para></listitem>
9213 <listitem><para>ExecSQL method reorganization (Anders Bengtsson)</para></listitem>
9214 <listitem><para>GetColumn() fixes (Jeroen van Vianen)</para></listitem>
9215 <listitem><para>Fix isWriteable() function (Rene Pijlman)</para></listitem>
9216 <listitem><para>Improved passage of JDBC2 conformance tests (Rene Pijlman)</para></listitem>
9217 <listitem><para>Add bytea type capability (Barry Lind)</para></listitem>
9218 <listitem><para>Add isNullable() (Rene Pijlman)</para></listitem>
9219 <listitem><para>JDBC date/time test suite fixes (Liam Stewart)</para></listitem>
9220 <listitem><para>Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)</para></listitem>
9221 <listitem><para>Fix DatabaseMetaData to show precision properly (Mark Lillywhite)</para></listitem>
9222 <listitem><para>New getImported/getExported keys (Jason Davies)</para></listitem>
9223 <listitem><para>MD5 password encryption support (Jeremy Wohl)</para></listitem>
9224 <listitem><para>Fix to actually use type cache (Ned Wolpert)</para></listitem>
9231 <listitem><para>Remove query size limit (Hiroshi)</para></listitem>
9232 <listitem><para>Remove text field size limit (Hiroshi)</para></listitem>
9233 <listitem><para>Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)</para></listitem>
9234 <listitem><para>Allow ODBC procedure calls (Hiroshi)</para></listitem>
9235 <listitem><para>Improve boolean handing (Aidan Mountford)</para></listitem>
9236 <listitem><para>Most configuration options now settable via DSN (Hiroshi)</para></listitem>
9237 <listitem><para>Multibyte, performance fixes (Hiroshi)</para></listitem>
9238 <listitem><para>Allow driver to be used with iODBC or unixODBC (Peter E)</para></listitem>
9239 <listitem><para>MD5 password encryption support (Bruce)</para></listitem>
9240 <listitem><para>Add more compatibility functions to odbc.sql (Peter E)</para></listitem>
9245 <title><application>ECPG</></title>
9247 <listitem><para>EXECUTE ... INTO implemented (Christof Petig)</para></listitem>
9248 <listitem><para>Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)</para></listitem>
9249 <listitem><para>Fix for GRANT parameters (Lee Kindness)</para></listitem>
9250 <listitem><para>Fix INITIALLY DEFERRED bug</para></listitem>
9251 <listitem><para>Various bug fixes (Michael, Christof Petig)</para></listitem>
9252 <listitem><para>Auto allocation for indicator variable arrays (int *ind_p=NULL)</para></listitem>
9253 <listitem><para>Auto allocation for string arrays (char **foo_pp=NULL)</para></listitem>
9254 <listitem><para>ECPGfree_auto_mem fixed</para></listitem>
9255 <listitem><para>All function names with external linkage are now prefixed by ECPG</para></listitem>
9256 <listitem><para>Fixes for arrays of structures (Michael)</para></listitem>
9261 <title>Misc. Interfaces</title>
9263 <listitem><para>Python fix fetchone() (Gerhard Haring)</para></listitem>
9264 <listitem><para>Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Max)</para></listitem>
9265 <listitem><para>Add Tcl COPY TO/FROM (ljb)</para></listitem>
9266 <listitem><para>Prevent output of default index op class in <application>pg_dump</> (Tom)</para></listitem>
9267 <listitem><para>Fix libpgeasy memory leak (Bruce)</para></listitem>
9272 <title>Build and Install</title>
9274 <listitem><para>Configure, dynamic loader, and shared library fixes (Peter E)</para></listitem>
9275 <listitem><para>Fixes in QNX 4 port (Bernd Tegge)</para></listitem>
9276 <listitem><para>Fixes in Cygwin and Windows ports (Jason Tishler, Gerhard Haring, Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)</para></listitem>
9277 <listitem><para>Fix for Windows socket communication failures (Magnus, Mikhail Terekhov)</para></listitem>
9278 <listitem><para>Hurd compile fix (Oliver Elphick)</para></listitem>
9279 <listitem><para>BeOS fixes (Cyril Velter)</para></listitem>
9280 <listitem><para>Remove configure --enable-unicode-conversion, now enabled by multibyte (Tatsuo)</para></listitem>
9281 <listitem><para>AIX fixes (Tatsuo, Andreas)</para></listitem>
9282 <listitem><para>Fix parallel make (Peter E)</para></listitem>
9283 <listitem><para>Install SQL language manual pages into OS-specific directories (Peter E)</para></listitem>
9284 <listitem><para>Rename config.h to pg_config.h (Peter E)</para></listitem>
9285 <listitem><para>Reorganize installation layout of header files (Peter E)</para></listitem>
9290 <title>Source Code</title>
9292 <listitem><para>Remove SEP_CHAR (Bruce)</para></listitem>
9293 <listitem><para>New GUC hooks (Tom)</para></listitem>
9294 <listitem><para>Merge GUC and command line handling (Marko Kreen)</para></listitem>
9295 <listitem><para>Remove EXTEND INDEX (Martijn van Oosterhout, Tom)</para></listitem>
9296 <listitem><para>New pgjindent utility to indent java code (Bruce)</para></listitem>
9297 <listitem><para>Remove define of true/false when compiling under C++ (Leandro Fanzone, Tom)</para></listitem>
9298 <listitem><para>pgindent fixes (Bruce, Tom)</para></listitem>
9299 <listitem><para>Replace strcasecmp() with strcmp() where appropriate (Peter E)</para></listitem>
9300 <listitem><para>Dynahash portability improvements (Tom)</para></listitem>
9301 <listitem><para>Add 'volatile' usage in spinlock structures</para></listitem>
9302 <listitem><para>Improve signal handling logic (Tom)</para></listitem>
9307 <title>Contrib</title>
9309 <listitem><para>New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)</para></listitem>
9310 <listitem><para>New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)</para></listitem>
9311 <listitem><para>Add contrib/dblink for remote database access (Joe Conway)</para></listitem>
9312 <listitem><para>contrib/ora2pg Oracle conversion utility (Gilles Darold)</para></listitem>
9313 <listitem><para>contrib/xml XML conversion utility (John Gray)</para></listitem>
9314 <listitem><para>contrib/fulltextindex fixes (Christopher Kings-Lynne)</para></listitem>
9315 <listitem><para>New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merged (Joe Conway)</para></listitem>
9316 <listitem><para>Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartunov)</para></listitem>
9317 <listitem><para>New pg_upgrade utility (Bruce)</para></listitem>
9318 <listitem><para>Add new pg_resetxlog options (Bruce, Tom)</para></listitem>
9325 <sect1 id="release-7-1-3">
9326 <title>Release 7.1.3</title>
9329 <title>Release date</title>
9330 <simpara>2001-08-15</simpara>
9334 <title>Migration to version 7.1.3</title>
9337 A dump/restore is <emphasis>not</emphasis> required for those running
9343 <title>Changes</title>
9347 Remove unused WAL segements of large transactions (Tom)
9348 Multiaction rule fix (Tom)
9349 PL/pgSQL memory allocation fix (Jan)
9350 VACUUM buffer fix (Tom)
9351 Regression test fixes (Tom)
9352 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
9353 Fix subselects with DISTINCT ON or LIMIT (Tom)
9355 Disable COPY TO/FROM a view (Tom)
9356 Cygwin build (Jason Tishler)
9363 <sect1 id="release-7-1-2">
9364 <title>Release 7.1.2</title>
9367 <title>Release date</title>
9368 <simpara>2001-05-11</simpara>
9372 This has one fix from 7.1.1.
9377 <title>Migration to version 7.1.2</title>
9380 A dump/restore is <emphasis>not</emphasis> required for those running
9386 <title>Changes</title>
9390 Fix PL/pgSQL SELECTs when returning no rows
9391 Fix for psql backslash core dump
9392 Referential integrity privilege fix
9401 <sect1 id="release-7-1-1">
9402 <title>Release 7.1.1</title>
9405 <title>Release date</title>
9406 <simpara>2001-05-05</simpara>
9410 This has a variety of fixes from 7.1.
9415 <title>Migration to version 7.1.1</title>
9418 A dump/restore is <emphasis>not</emphasis> required for those running
9424 <title>Changes</title>
9428 Fix for numeric MODULO operator (Tom)
9429 pg_dump fixes (Philip)
9430 pg_dump can dump 7.0 databases (Philip)
9431 readline 4.2 fixes (Peter E)
9433 AIX, MSWIN, VAX, N32K fixes (Tom)
9434 Multibytes fixes (Tom)
9435 Unicode fixes (Tatsuo)
9436 Optimizer improvements (Tom)
9437 Fix for whole rows in functions (Tom)
9438 Fix for pg_ctl and option strings with spaces (Peter E)
9439 ODBC fixes (Hiroshi)
9440 EXTRACT can now take string argument (Thomas)
9441 Python fixes (Darcy)
9448 <sect1 id="release-7-1">
9449 <title>Release 7.1</title>
9452 <title>Release date</title>
9453 <simpara>2001-04-13</simpara>
9457 This release focuses on removing limitations that have existed in the
9458 <productname>PostgreSQL</productname> code for many years.
9462 Major changes in this release:
9468 Write-ahead Log (WAL)
9472 To maintain database consistency in case of an operating system crash,
9473 previous releases of <productname>PostgreSQL</productname> have forced
9474 all data modifications to disk before each transaction commit. With
9475 WAL, only one log file must be flushed to disk, greatly improving
9476 performance. If you have been using -F in previous releases to
9477 disable disk flushes, you may want to consider discontinuing its use.
9488 TOAST - Previous releases had a compiled-in row length limit,
9489 typically 8k - 32k. This limit made storage of long text fields
9490 difficult. With TOAST, long rows of any length can be stored with good
9502 We now support outer joins. The UNION/NOT IN
9503 workaround for outer joins is no longer required. We use the SQL92
9515 The previous C function manager did not
9516 handle null values properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
9517 function manager does. You can continue using your old custom
9518 functions, but you may want to rewrite them in the future to use the new
9519 function manager call interface.
9530 A large number of complex queries that were
9531 unsupported in previous releases now work. Many combinations of views,
9532 aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
9533 now work properly. Inherited tables are now accessed by default.
9534 Subqueries in FROM are now supported.
9542 <title>Migration to version 7.1</title>
9545 A dump/restore using pg_dump is required for those wishing to migrate
9546 data from any previous release.
9551 <title>Changes</title>
9557 Many multibyte/Unicode/locale fixes (Tatsuo and others)
9558 More reliable ALTER TABLE RENAME (Tom)
9559 Kerberos V fixes (David Wragg)
9560 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
9561 Prompt username/password on standard error (Bruce)
9562 Large objects inv_read/inv_write fixes (Tom)
9563 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
9565 Prevent query expressions from leaking memory (Tom)
9566 Allow UPDATE of arrays elements (Tom)
9567 Wake up lock waiters during cancel (Hiroshi)
9568 Fix rare cursor crash when using hash join (Tom)
9569 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
9570 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
9571 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
9572 Fix PL/perl (Alex Kapranoff)
9573 Disallow LOCK on views (Mark Hollomon)
9574 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
9575 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
9576 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
9577 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
9578 Fix rare failure with TRUNCATE command (Tom)
9579 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
9580 DISTINCT, ORDER BY, SELECT...INTO (Tom)
9581 Fix parser failures during aborted transactions (Tom)
9582 Allow temporary relations to properly clean up indexes (Bruce)
9583 Fix VACUUM problem with moving rows in same page (Tom)
9584 Modify pg_dump to better handle user-defined items in template1 (Philip)
9585 Allow LIMIT in VIEW (Tom)
9586 Require cursor FETCH to honor LIMIT (Tom)
9587 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
9588 Allow ORDER BY, LIMIT in subqueries (Tom)
9589 Allow UNION in CREATE RULE (Tom)
9590 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
9591 Store initdb collation in pg_control so collation cannot be changed (Tom)
9592 Fix INSERT...SELECT with rules (Tom)
9593 Fix FOR UPDATE inside views and subselects (Tom)
9594 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
9595 Fix lpad() and rpad() to handle length less than input string (Tom)
9596 Fix use of NOTIFY in some rules (Tom)
9597 Overhaul btree code (Tom)
9598 Fix NOT NULL use in Pl/pgSQL variables (Tom)
9599 Overhaul GIST code (Oleg)
9600 Fix CLUSTER to preserve constraints and column default (Tom)
9601 Improved deadlock detection handling (Tom)
9602 Allow multiple SERIAL columns in a table (Tom)
9603 Prevent occasional index corruption (Vadim)
9607 Add OUTER JOINs (Tom)
9608 Function manager overhaul (Tom)
9609 Allow ALTER TABLE RENAME on indexes (Tom)
9610 Improve CLUSTER (Tom)
9611 Improve ps status display for more platforms (Peter E, Marc)
9612 Improve CREATE FUNCTION failure message (Ross)
9613 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
9615 Grand Unified Configuration scheme/GUC. Many options can now be set in
9616 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
9617 Improved handling of file descriptor cache (Tom)
9618 New warning code about auto-created table alias entries (Bruce)
9619 Overhaul initdb process (Tom, Peter E)
9620 Overhaul of inherited tables; inherited tables now accessed by default;
9621 new ONLY key word prevents it (Chris Bitmead, Tom)
9622 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
9624 Allow renaming of temp tables (Tom)
9625 Overhaul memory manager contexts (Tom)
9626 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
9627 Overhaul pg_dump (Philip Warner)
9628 Allow pg_hba.conf secondary password file to specify only username (Peter E)
9629 Allow TEMPORARY or TEMP key word when creating temporary tables (Bruce)
9630 New memory leak checker (Karel)
9631 New SET SESSION CHARACTERISTICS (Thomas)
9632 Allow nested block comments (Thomas)
9633 Add WITHOUT TIME ZONE type qualifier (Thomas)
9634 New ALTER TABLE ADD CONSTRAINT (Stephan)
9635 Use NUMERIC accumulators for INTEGER aggregates (Tom)
9636 Overhaul aggregate code (Tom)
9637 New VARIANCE and STDDEV() aggregates
9638 Improve dependency ordering of pg_dump (Philip)
9639 New pg_restore command (Philip)
9640 New pg_dump tar output option (Philip)
9641 New pg_dump of large objects (Philip)
9642 New ESCAPE option to LIKE (Thomas)
9643 New case-insensitive LIKE - ILIKE (Thomas)
9644 Allow functional indexes to use binary-compatible type (Tom)
9645 Allow SQL functions to be used in more contexts (Tom)
9646 New pg_config utility (Peter E)
9647 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
9649 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
9650 New quote_identifiers() and quote_literal() functions (Jan)
9651 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
9652 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
9653 Update PyGreSQL to version 3.1 (D'Arcy)
9654 Store tables as files named by OID (Vadim)
9655 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
9656 Require DROP VIEW to remove views, no DROP TABLE (Mark)
9657 Allow DROP VIEW view1, view2 (Mark)
9658 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
9659 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
9660 New /contrib/pgcrypto hashing functions (Marko Kreen)
9661 New pg_dumpall --globals-only option (Peter E)
9662 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
9663 New AT TIME ZONE syntax (Thomas)
9664 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
9665 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
9666 Allow socket path name to be specified in hostname by using leading slash
9667 (David J. MacKenzie)
9668 Allow CREATE DATABASE to specify template database (Tom)
9669 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
9670 New /contrib/rserv replication toolkit (Vadim)
9671 New file format for COPY BINARY (Tom)
9672 New /contrib/oid2name to map numeric files to table names (B Palmer)
9673 New "idle in transaction" ps status message (Marc)
9674 Update to pgaccess 0.98.7 (Constantin Teodorescu)
9675 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
9676 Add rudimentary dependency checking to pg_dump (Philip)
9680 Fix INET/CIDR type ordering and add new functions (Tom)
9681 Make OID behave as an unsigned type (Tom)
9682 Allow BIGINT as synonym for INT8 (Peter E)
9683 New int2 and int8 comparison operators (Tom)
9684 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
9685 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
9686 New GIST seg/cube examples (Gene Selkov)
9687 Improved round(numeric) handling (Tom)
9688 Fix CIDR output formatting (Tom)
9689 New CIDR abbrev() function (Tom)
9693 Write-Ahead Log (WAL) to provide crash recovery with less performance
9695 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
9696 Reduced file seeks (Denis Perchine)
9697 Improve BTREE code for duplicate keys (Tom)
9698 Store all large objects in a single table (Denis Perchine, Tom)
9699 Improve memory allocation performance (Karel, Tom)
9703 New function manager call conventions (Tom)
9704 SGI portability fixes (David Kaelbling)
9705 New configure --enable-syslog option (Peter E)
9706 New BSDI README (Bruce)
9707 configure script moved to top level, not /src (Peter E)
9708 Makefile/configuration/compilation overhaul (Peter E)
9709 New configure --with-python option (Peter E)
9710 Solaris cleanups (Peter E)
9711 Overhaul /contrib Makefiles (Karel)
9712 New OpenSSL configuration option (Magnus, Peter E)
9714 QNX fixes (Maurizio)
9715 New heap_open(), heap_openr() API (Tom)
9716 Remove colon and semi-colon operators (Thomas)
9717 New pg_class.relkind value for views (Mark Hollomon)
9718 Rename ichar() to chr() (Karel)
9719 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
9720 Fixes for NT/Cygwin (Pete Forman)
9721 AIX port fixes (Andreas)
9722 New BeOS port (David Reid, Cyril Velter)
9723 Add proofreader's changes to docs (Addison-Wesley, Bruce)
9724 New Alpha spinlock code (Adriaan Joubert, Compaq)
9725 UnixWare port overhaul (Peter E)
9726 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
9727 New FreeBSD Alpha port (Alfred)
9728 Overhaul shared memory segments (Tom)
9729 Add IBM S/390 support (Neale Ferguson)
9730 Moved macmanuf to /contrib (Larry Rosenman)
9731 Syslog improvements (Larry Rosenman)
9732 New template0 database that contains no user additions (Tom)
9733 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
9734 Allow NetBSD's libedit instead of readline (Peter)
9735 Improved assembly language source code format (Bruce)
9736 New contrib/pg_logger
9737 New --template option to createdb
9738 New contrib/pg_control utility (Oliver)
9739 New FreeBSD tools ipc_check, start-scripts/freebsd
9746 <sect1 id="release-7-0-3">
9747 <title>Release 7.0.3</title>
9750 <title>Release date</title>
9751 <simpara>2000-11-11</simpara>
9755 This has a variety of fixes from 7.0.2.
9760 <title>Migration to version 7.0.3</title>
9763 A dump/restore is <emphasis>not</emphasis> required for those running
9769 <title>Changes</title>
9774 Large object fix (Tom)
9775 Fix lean in COPY WITH OIDS leak (Tom)
9776 Fix backwards-index-scan (Tom)
9777 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
9778 Add --enable-syslog to configure (Marc)
9779 Fix abort transaction at backend exit in rare cases (Tom)
9780 Fix for psql \l+ when multibyte enabled (Tatsuo)
9781 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
9782 Make vacuum always flush buffers (Tom)
9783 Fix to allow cancel while waiting for a lock (Hiroshi)
9784 Fix for memory aloocation problem in user authentication code (Tom)
9785 Remove bogus use of int4out() (Tom)
9786 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
9787 Fix for failure of triggers on heap open in certain cases (Jeroen van
9789 Fix for erroneous selectivity of not-equals (Tom)
9790 Fix for erroneous use of strcmp() (Tom)
9791 Fix for bug where storage manager accesses items beyond end of file
9793 Fix to include kernel errno message in all smgr elog messages (Tom)
9794 Fix for '.' not in PATH at build time (SL Baur)
9795 Fix for out-of-file-descriptors error (Tom)
9796 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
9797 Fix for subselect in targetlist of Append node (Tom)
9798 Fix for mergejoin plans (Tom)
9799 Fix TRUNCATE failure on relations with indexes (Tom)
9800 Avoid database-wide restart on write error (Hiroshi)
9801 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
9802 Fix VACUUM problem with moving chain of update row versions when source
9803 and destination of a row version lie on the same page (Tom)
9804 Fix user.c CommandCounterIncrement (Tom)
9805 Fix for AM/PM boundary problem in to_char() (Karel Zak)
9806 Fix TIME aggregate handling (Tom)
9807 Fix to_char() to avoid coredump on NULL input (Tom)
9809 Fix for inserting/copying longer multibyte strings into char() data
9811 Fix for crash of backend, on abort (Tom)
9818 <sect1 id="release-7-0-2">
9819 <title>Release 7.0.2</title>
9822 <title>Release date</title>
9823 <simpara>2000-06-05</simpara>
9827 This is a repackaging of 7.0.1 with added documentation.
9832 <title>Migration to version 7.0.2</title>
9835 A dump/restore is <emphasis>not</emphasis> required for those running
9841 <title>Changes</title>
9845 Added documentation to tarball.
9852 <sect1 id="release-7-0-1">
9853 <title>Release 7.0.1</title>
9856 <title>Release date</title>
9857 <simpara>2000-06-01</simpara>
9861 This is a cleanup release for 7.0.
9865 <title>Migration to version 7.0.1</title>
9868 A dump/restore is <emphasis>not</emphasis> required for those running
9874 <title>Changes</title>
9878 Fix many CLUSTER failures (Tom)
9879 Allow ALTER TABLE RENAME works on indexes (Tom)
9880 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
9881 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
9882 Fix the off by one errors in ResultSet from 6.5.3, and more.
9883 jdbc ResultSet fixes (Joseph Shraibman)
9884 optimizer tunings (Tom)
9885 Fix create user for pgaccess
9886 Fix for UNLISTEN failure
9887 IRIX fixes (David Kaelbling)
9888 QNX fixes (Andreas Kardos)
9889 Reduce COPY IN lock level (Tom)
9890 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
9891 Fix pg_dump to handle OID indexes (Tom)
9892 Fix small memory leak (Tom)
9893 Solaris fix for createdb/dropdb (Tatsuo)
9894 Fix for non-blocking connections (Alfred Perlstein)
9895 Fix improper recovery after RENAME TABLE failures (Tom)
9896 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
9897 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
9898 Fix too long syslog message (Tatsuo)
9899 Fix problem with quoted indexes that are too long (Tom)
9900 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
9901 ecpg changes (Michael)
9907 <sect1 id="release-7-0">
9908 <title>Release 7.0</title>
9911 <title>Release date</title>
9912 <simpara>2000-05-08</simpara>
9916 This release contains improvements in many areas, demonstrating
9917 the continued growth of <productname>PostgreSQL</productname>.
9918 There are more improvements and fixes in 7.0 than in any previous
9919 release. The developers have confidence that this is the best
9920 release yet; we do our best to put out only solid releases, and
9921 this one is no exception.
9925 Major changes in this release:
9935 Foreign keys are now implemented, with the exception of PARTIAL MATCH
9936 foreign keys. Many users have been asking for this feature, and we are
9937 pleased to offer it.
9948 Continuing on work started a year ago, the optimizer has been
9949 improved, allowing better query plan selection and faster performance
9950 with less memory usage.
9957 Updated <application>psql</application>
9961 <application>psql</application>, our interactive terminal monitor, has been
9962 updated with a variety of new features. See the <application>psql</application> manual page for details.
9973 SQL92 join syntax is now supported, though only as
9974 <literal>INNER JOIN</> for this release. <literal>JOIN</>,
9975 <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
9976 and <literal>JOIN</>/<literal>ON</> are available, as are
9977 column correlation names.
9985 <title>Migration to version 7.0</title>
9988 A dump/restore using <application>pg_dump</application>
9989 is required for those wishing to migrate data from any
9990 previous release of <productname>PostgreSQL</productname>.
9991 For those upgrading from 6.5.*, you may instead use
9992 <application>pg_upgrade</application> to upgrade to this
9993 release; however, a full dump/reload installation is always the
9994 most robust method for upgrades.
9998 Interface and compatibility issues to consider for the new
10005 The date/time types <type>datetime</type> and
10006 <type>timespan</type> have been superseded by the
10007 SQL92-defined types <type>timestamp</type> and
10008 <type>interval</type>. Although there has been some effort to
10009 ease the transition by allowing
10010 <productname>PostgreSQL</productname> to recognize
10011 the deprecated type names and translate them to the new type
10012 names, this mechanism may not be completely transparent to
10013 your existing application.
10019 The optimizer has been substantially improved in the area of
10020 query cost estimation. In some cases, this will result in
10021 decreased query times as the optimizer makes a better choice
10022 for the preferred plan. However, in a small number of cases,
10023 usually involving pathological distributions of data, your
10024 query times may go up. If you are dealing with large amounts
10025 of data, you may want to check your queries to verify
10032 The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
10033 interfaces have been upgraded and extended.
10039 The string function <function>CHAR_LENGTH</function> is now a
10040 native function. Previous versions translated this into a call
10041 to <function>LENGTH</function>, which could result in
10042 ambiguity with other types implementing
10043 <function>LENGTH</function> such as the geometric types.
10050 <title>Changes</title>
10056 Prevent function calls exceeding maximum number of arguments (Tom)
10057 Improve CASE construct (Tom)
10058 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
10059 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
10060 Fix GROUP BY scan bug (Tom)
10061 Improvements in SQL grammar processing (Tom)
10062 Fix for views involved in INSERT ... SELECT ... (Tom)
10063 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
10064 Fix for subselects in INSERT ... SELECT (Tom)
10065 Prevent INSERT ... SELECT ... ORDER BY (Tom)
10066 Fixes for relations greater than 2GB, including vacuum
10067 Improve propagating system table changes to other backends (Tom)
10068 Improve propagating user table changes to other backends (Tom)
10069 Fix handling of temp tables in complex situations (Bruce, Tom)
10070 Allow table locking at table open, improving concurrent reliability (Tom)
10071 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
10072 Prevent DROP DATABASE while others accessing
10073 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
10074 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
10075 Fix pg_upgrade so it works for MVCC (Tom)
10076 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
10077 Fix for "f1 datetime DEFAULT 'now'" (Tom)
10078 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
10079 Allow comment-only lines, and ;;; lines too. (Tom)
10080 Improve recovery after failed disk writes, disk full (Hiroshi)
10081 Fix cases where table is mentioned in FROM but not joined (Tom)
10082 Allow HAVING clause without aggregate functions (Tom)
10083 Fix for "--" comment and no trailing newline, as seen in perl interface
10084 Improve pg_dump failure error reports (Bruce)
10085 Allow sorts and hashes to exceed 2GB file sizes (Tom)
10086 Fix for pg_dump dumping of inherited rules (Tom)
10087 Fix for NULL handling comparisons (Tom)
10088 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
10089 Fix for dbname with dash
10090 Prevent DROP INDEX from interfering with other backends (Tom)
10091 Fix file descriptor leak in verify_password()
10092 Fix for "Unable to identify an operator =$" problem
10093 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
10094 Fix for recursive exit call (Massimo)
10095 Fix for extra-long timezones (Jeroen van Vianen)
10096 Make pg_dump preserve primary key information (Peter E)
10097 Prevent databases with single quotes (Peter E)
10098 Prevent DROP DATABASE inside transaction (Peter E)
10099 ecpg memory leak fixes (Stephen Birch)
10100 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
10101 Y2K timestamp fix (Massimo)
10102 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
10103 Fix for views with tables/columns containing spaces (Tom)
10104 Prevent privileges on indexes (Peter E)
10105 Fix for spinlock stuck problem when error is generated (Hiroshi)
10106 Fix ipcclean on Linux
10107 Fix handling of NULL constraint conditions (Tom)
10108 Fix memory leak in odbc driver (Nick Gorham)
10109 Fix for privilege check on UNION tables (Tom)
10110 Fix to allow SELECT 'a' LIKE 'a' (Tom)
10111 Fix for SELECT 1 + NULL (Tom)
10113 Fix log() on numeric type (Tom)
10114 Deprecate ':' and ';' operators
10115 Allow vacuum of temporary tables
10116 Disallow inherited columns with the same name as new columns
10117 Recover or force failure when disk space is exhausted (Hiroshi)
10118 Fix INSERT INTO ... SELECT with AS columns matching result columns
10119 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns (Tom)
10120 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
10121 Fix UNION with LIMIT
10122 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
10123 Fix CREATE TABLE test(col char(2) DEFAULT user)
10124 Fix mismatched types in CREATE TABLE ... DEFAULT
10125 Fix SELECT * FROM pg_class where oid in (0,-1)
10126 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
10127 Prevent user who can create databases can modifying pg_database table (Peter E)
10128 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
10129 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
10133 New CLI interface include file sqlcli.h, based on SQL3/SQL98
10134 Remove all limits on query length, row length limit still exists (Tom)
10135 Update jdbc protocol to 2.0 (Jens Glaser <email>jens@jens.de</email>)
10136 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
10137 Fix to give super user and createdb user proper update catalog rights (Peter E)
10138 Allow ecpg bool variables to have NULL values (Christof)
10139 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
10140 Allow ^C to cancel COPY command (Massimo)
10141 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
10142 Function name overloading for dynamically-loaded C functions (Frankpitt)
10143 Add CmdTuples() to libpq++(Vince)
10144 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
10145 Allow CREATE FUNCTION/WITH clause to be used for all language types
10146 configure --enable-debug adds -g (Peter E)
10147 configure --disable-debug removes -g (Peter E)
10148 Allow more complex default expressions (Tom)
10149 First real FOREIGN KEY constraint trigger functionality (Jan)
10150 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
10151 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
10152 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
10153 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
10154 Change pgeasy connectdb() parameter ordering (Bruce)
10155 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
10156 Add Oracle's COMMENT ON command (Mike Mascari <email>mascarim@yahoo.com</email>)
10157 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
10158 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
10159 Make USING in COPY optional (Bruce)
10160 Allow subselects in the target list (Tom)
10161 Allow subselects on the left side of comparison operators (Tom)
10162 New parallel regression test (Jan)
10163 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
10164 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
10165 Added psql LASTOID variable to return last inserted oid (Peter E)
10166 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
10167 Add privilege check for vacuum (Peter E)
10168 New libpq functions to allow asynchronous connections: PQconnectStart(),
10169 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
10170 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
10171 New libpq PQsetenv() function (Ewan Mellor)
10172 create/alter user extension (Peter E)
10173 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
10174 New scripts for create/drop user/db (Peter E)
10175 Major psql overhaul (Peter E)
10176 Add const to libpq interface (Peter E)
10177 New libpq function PQoidValue (Peter E)
10178 Show specific non-aggregate causing problem with GROUP BY (Tom)
10179 Make changes to pg_shadow recreate pg_pwd file (Peter E)
10180 Add aggregate(DISTINCT ...) (Tom)
10181 Allow flag to control COPY input/output of NULLs (Peter E)
10182 Make postgres user have a password by default (Peter E)
10183 Add CREATE/ALTER/DROP GROUP (Peter E)
10184 All administration scripts now support --long options (Peter E, Karel)
10185 Vacuumdb script now supports --all option (Peter E)
10186 ecpg new portable FETCH syntax
10187 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
10188 and EXEC SQL ENDIF directives
10189 Add pg_ctl script to control backend start-up (Tatsuo)
10190 Add postmaster.opts.default file to store start-up flags (Tatsuo)
10191 Allow --with-mb=SQL_ASCII
10192 Increase maximum number of index keys to 16 (Bruce)
10193 Increase maximum number of function arguments to 16 (Bruce)
10194 Allow configuration of maximum number of index keys and arguments (Bruce)
10195 Allow unprivileged users to change their passwords (Peter E)
10196 Password authentication enabled; required for new users (Peter E)
10197 Disallow dropping a user who owns a database (Peter E)
10198 Change initdb option --with-mb to --enable-multibyte
10199 Add option for initdb to prompts for superuser password (Peter E)
10200 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
10201 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
10202 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
10203 libpq non-blocking mode (Alfred Perlstein)
10204 Improve conversion of types in casts that don't specify a length
10205 New plperl internal programming language (Mark Hollomon)
10206 Allow COPY IN to read file that do not end with a newline (Tom)
10207 Indicate when long identifiers are truncated (Tom)
10208 Allow aggregates to use type equivalency (Peter E)
10209 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
10210 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
10211 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
10212 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
10213 Add NUMERIC and int8 types to ODBC
10214 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
10215 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
10216 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
10217 Enable backward sequential scan even after reaching EOF (Hiroshi)
10218 Add btree indexing of boolean values, >= and <= (Don Baccus)
10219 Print current line number when COPY FROM fails (Massimo)
10220 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
10221 Add DEC as synonym for DECIMAL (Thomas)
10222 Add SESSION_USER as SQL92 key word, same as CURRENT_USER (Thomas)
10223 Implement SQL92 column aliases (aka correlation names) (Thomas)
10224 Implement SQL92 join syntax (Thomas)
10225 Make INTERVAL reserved word allowed as a column identifier (Thomas)
10226 Implement REINDEX command (Hiroshi)
10227 Accept ALL in aggregate function SUM(ALL col) (Tom)
10228 Prevent GROUP BY from using column aliases (Tom)
10229 New psql \encoding option (Tatsuo)
10230 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
10231 Allow negation of a negative number in all cases
10232 Add ecpg descriptors (Christof, Michael)
10233 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
10234 Allow casts with length, like foo::char(8)
10235 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
10236 Add support for SJIS user defined characters (Tatsuo)
10237 Larger views/rules supported
10238 Make libpq's PQconndefaults() thread-safe (Tom)
10239 Disable // as comment to be ANSI conforming, should use -- (Tom)
10240 Allow column aliases on views CREATE VIEW name (collist)
10241 Fixes for views with subqueries (Tom)
10242 Allow UPDATE table SET fld = (SELECT ...) (Tom)
10243 SET command options no longer require quotes
10244 Update pgaccess to 0.98.6
10245 New SET SEED command
10246 New pg_options.sample file
10247 New SET FSYNC command (Massimo)
10248 Allow pg_descriptions when creating tables
10249 Allow pg_descriptions when creating types, columns, and functions
10250 Allow psql \copy to allow delimiters (Peter E)
10251 Allow psql to print nulls as distinct from "" [null] (Peter E)
10255 Many array fixes (Tom)
10256 Allow bare column names to be subscripted as arrays (Tom)
10257 Improve type casting of int and float constants (Tom)
10258 Cleanups for int8 inputs, range checking, and type conversion (Tom)
10259 Fix for SELECT timespan('21:11:26'::time) (Tom)
10260 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
10261 Add btree index on NUMERIC (Jan)
10262 Perl fix for large objects containing NUL characters (Douglas Thomson)
10263 ODBC fix for for large objects (free)
10264 Fix indexing of cidr data type
10265 Fix for Ethernet MAC addresses (macaddr type) comparisons
10266 Fix for date/time types when overflows happened in computations (Tom)
10267 Allow array on int8 (Peter E)
10268 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
10269 Allow NUMERIC arrays
10270 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
10271 Make char_length()/octet_length including trailing blanks (Tom)
10272 Made abstime/reltime use int4 instead of time_t (Peter E)
10273 New lztext data type for compressed text fields
10274 Revise code to handle coercion of int and float constants (Tom)
10275 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
10276 NUMERIC now accepts scientific notation (Tom)
10277 NUMERIC to int4 rounds (Tom)
10278 Convert float4/8 to NUMERIC properly (Tom)
10279 Allow type conversion with NUMERIC (Thomas)
10280 Make ISO date style (2000-02-16 09:33) the default (Thomas)
10281 Add NATIONAL CHAR [ VARYING ] (Thomas)
10282 Allow NUMERIC round and trunc to accept negative scales (Tom)
10283 New TIME WITH TIME ZONE type (Thomas)
10284 Add MAX()/MIN() on time type (Thomas)
10285 Add abs(), mod(), fac() for int8 (Thomas)
10286 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
10287 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
10288 Add exp() and ln() for NUMERIC type
10289 Rename NUMERIC power() to pow() (Thomas)
10290 Improved TRANSLATE() function (Edwin Ramirez, Tom)
10291 Allow X=-Y operators (Tom)
10292 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom)
10293 Allow LOCALE to use indexes in regular expression searches (Tom)
10294 Allow creation of functional indexes to use default types
10298 Prevent exponential space consumption with many AND's and OR's (Tom)
10299 Collect attribute selectivity values for system columns (Tom)
10300 Reduce memory usage of aggregates (Tom)
10301 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
10302 Fix r-tree index optimizer selectivity (Thomas)
10303 Improve optimizer selectivity computations and functions (Tom)
10304 Optimize btree searching for cases where many equal keys exist (Tom)
10305 Enable fast LIKE index processing only if index present (Tom)
10306 Re-use free space on index pages with duplicates (Tom)
10307 Improve hash join processing (Tom)
10308 Prevent descending sort if result is already sorted(Hiroshi)
10309 Allow commuting of index scan query qualifications (Tom)
10310 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
10311 Allocate large memory requests in fix-sized chunks for performance (Tom)
10312 Fix vacuum's performance by reducing memory allocation requests (Tom)
10313 Implement constant-expression simplification (Bernard Frankpitt, Tom)
10314 Use secondary columns to be used to determine start of index scan (Hiroshi)
10315 Prevent quadruple use of disk space when doing internal sorting (Tom)
10316 Faster sorting by calling fewer functions (Tom)
10317 Create system indexes to match all system caches (Bruce, Hiroshi)
10318 Make system caches use system indexes (Bruce)
10319 Make all system indexes unique (Bruce)
10320 Improve pg_statistics management for VACUUM speed improvement (Tom)
10321 Flush backend cache less frequently (Tom, Hiroshi)
10322 COPY now reuses previous memory allocation, improving performance (Tom)
10323 Improve optimization cost estimation (Tom)
10324 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom)
10325 Use DNF instead of CNF where appropriate (Tom, Taral)
10326 Further cleanup for OR-of-AND WHERE-clauses (Tom)
10327 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
10328 Smarter optimizer computations for random index page access (Tom)
10329 New SET variable to control optimizer costs (Tom)
10330 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
10331 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
10332 Major subquery speedup (Tom)
10333 Fewer fsync writes when fsync is not disabled (Tom)
10334 Improved LIKE optimizer estimates (Tom)
10335 Prevent fsync in SELECT-only queries (Vadim)
10336 Make index creation use psort code, because it is now faster (Tom)
10337 Allow creation of sort temp tables > 1 Gig
10339 Source Tree Changes
10340 -------------------
10341 Fix for linux PPC compile
10342 New generic expression-tree-walker subroutine (Tom)
10343 Change form() to varargform() to prevent portability problems
10344 Improved range checking for large integers on Alphas
10345 Clean up #include in /include directory (Bruce)
10346 Add scripts for checking includes (Bruce)
10347 Remove un-needed #include's from *.c files (Bruce)
10348 Change #include's to use <> and "" as appropriate (Bruce)
10349 Enable Windows compilation of libpq
10350 Alpha spinlock fix from Uncle George <email>gatgul@voicenet.com</email>
10351 Overhaul of optimizer data structures (Tom)
10352 Fix to cygipc library (Yutaka Tanida)
10353 Allow pgsql to work on newer Cygwin snapshots (Dan)
10354 New catalog version number (Tom)
10356 Rename heap_replace to heap_update
10357 Update for QNX (Dr. Andreas Kardos)
10358 New platform-specific regression handling (Tom)
10359 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
10360 Included all yacc and lex files into the distribution (Peter E.)
10361 Remove lextest, no longer needed (Peter E)
10362 Fix for libpq and psql on Windows (Magnus)
10363 Internally change datetime and timespan into timestamp and interval (Thomas)
10364 Fix for plpgsql on BSD/OS
10365 Add SQL_ASCII test case to the regression test (Tatsuo)
10366 configure --with-mb now deprecated (Tatsuo)
10368 NetBSD fixes (Johnny C. Lam <email>lamj@stat.cmu.edu</email>)
10369 Fixes for Alpha compiles
10370 New multibyte encodings
10376 <sect1 id="release-6-5-3">
10377 <title>Release 6.5.3</title>
10380 <title>Release date</title>
10381 <simpara>1999-10-13</simpara>
10385 This is basically a cleanup release for 6.5.2. We have added a new
10386 <application>PgAccess</> that was missing in 6.5.2, and installed an NT-specific fix.
10391 <title>Migration to version 6.5.3</title>
10394 A dump/restore is <emphasis>not</emphasis> required for those running
10399 <title>Changes</title>
10403 Updated version of pgaccess 0.98
10405 Fix dumping rules on inherited tables
10412 <sect1 id="release-6-5-2">
10413 <title>Release 6.5.2</title>
10416 <title>Release date</title>
10417 <simpara>1999-09-15</simpara>
10421 This is basically a cleanup release for 6.5.1. We have fixed a variety of
10422 problems reported by 6.5.1 users.
10427 <title>Migration to version 6.5.2</title>
10430 A dump/restore is <emphasis>not</emphasis> required for those running
10436 <title>Changes</title>
10440 subselect+CASE fixes(Tom)
10441 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
10442 Fixes for CASE in WHERE join clauses(Tom)
10443 Fix BTScan abort(Tom)
10444 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
10445 Improve it so that it checks for multicolumn constraints(Thomas)
10446 Fix for Windows making problem with MB enabled(Hiroki Kataoka)
10447 Allow BSD yacc and bison to compile pl code(Bruce)
10448 Fix SET NAMES working
10450 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
10451 Reduce the total memory consumption of vacuum(Tom)
10452 Fix for timestamp(datetime)
10453 Rule deparsing bugfixes(Tom)
10454 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(Tom)
10455 This is to re-use space on index pages freed by vacuum(Vadim)
10456 document -x for pg_dump(Bruce)
10457 Fix for unary operators in rule deparser(Tom)
10458 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
10459 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
10460 Repair logic error in LIKE: should not return LIKE_ABORT
10461 when reach end of pattern before end of text(Tom)
10462 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom)
10463 Updated version of pgaccess 0.98
10469 <sect1 id="release-6-5-1">
10470 <title>Release 6.5.1</title>
10473 <title>Release date</title>
10474 <simpara>1999-07-15</simpara>
10478 This is basically a cleanup release for 6.5. We have fixed a variety of
10479 problems reported by 6.5 users.
10483 <title>Migration to version 6.5.1</title>
10486 A dump/restore is <emphasis>not</emphasis> required for those running
10492 <title>Changes</title>
10497 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
10498 Remove QUERY_LIMIT, use SELECT...LIMIT
10499 Fix for EXPLAIN on inheritance(Tom)
10500 Patch to allow vacuum on multisegment tables(Hiroshi)
10501 R-Tree optimizer selectivity fix(Tom)
10502 ACL file descriptor leak fix(Atsushi Ogawa)
10503 New expresssion subtree code(Tom)
10504 Avoid disk writes for read-only transactions(Vadim)
10505 Fix for removal of temp tables if last transaction was aborted(Bruce)
10506 Fix to prevent too large row from being created(Bruce)
10508 Allow port numbers 32k - 64k(Bruce)
10509 Add ^ precidence(Bruce)
10510 Rename sort files called pg_temp to pg_sorttemp(Bruce)
10511 Fix for microseconds in time values(Tom)
10512 Tutorial source cleanup
10513 New linux_m68k port
10514 Fix for sorting of NULL's in some cases(Tom)
10515 Shared library dependencies fixed (Tom)
10516 Fixed glitches affecting GROUP BY in subselects(Tom)
10517 Fix some compiler warnings (Tomoaki Nishiyama)
10518 Add Win1250 (Czech) support (Pavel Behal)
10524 <sect1 id="release-6-5">
10525 <title>Release 6.5</title>
10528 <title>Release date</title>
10529 <simpara>1999-06-09</simpara>
10533 This release marks a major step in the development team's mastery of the source
10534 code we inherited from Berkeley. You will see we are now easily adding
10535 major features, thanks to the increasing size and experience of our
10536 world-wide development team.
10540 Here is a brief summary of the more notable changes:
10545 Multiversion concurrency control(MVCC)
10549 This removes our old table-level locking, and replaces it with
10550 a locking system that is superior to most commercial database
10551 systems. In a traditional system, each row that is modified
10552 is locked until committed, preventing reads by other users.
10553 MVCC uses the natural multiversion nature of
10554 <productname>PostgreSQL</productname> to allow readers to
10555 continue reading consistent data during writer activity.
10556 Writers continue to use the compact pg_log transaction system.
10557 This is all performed without having to allocate a lock for
10558 every row like traditional database systems. So, basically,
10559 we no longer are restricted by simple table-level locking; we
10560 have something better than row-level locking.
10567 Hot backups from <application>pg_dump</application>
10571 <application>pg_dump</application> takes advantage of the new
10572 MVCC features to give a consistent database dump/backup while
10573 the database stays online and available for queries.
10584 We now have a true numeric data type, with
10585 user-specified precision.
10596 Temporary tables are guaranteed to have unique names
10597 within a database session, and are destroyed on session exit.
10608 We now have CASE, INTERSECT, and EXCEPT statement
10609 support. We have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL,
10610 SELECT ... FOR UPDATE, and an improved LOCK TABLE command.
10621 We continue to speed up <productname>PostgreSQL</productname>,
10622 thanks to the variety of talents within our team. We have
10623 sped up memory allocation, optimization, table joins, and row
10635 We continue to expand our port list, this time including
10636 <systemitem class="osname">Windows NT</>/<systemitem>ix86</> and <systemitem class="osname">NetBSD</>/<systemitem>arm32</>.
10647 Most interfaces have new versions, and existing functionality
10659 New and updated material is present throughout the
10660 documentation. New <acronym>FAQ</acronym>s have been
10661 contributed for <systemitem class="osname">SGI</> and <systemitem class="osname">AIX</> platforms.
10662 The <citetitle>Tutorial</citetitle> has introductory information
10663 on <acronym>SQL</acronym> from Stefan Simkovics.
10664 For the <citetitle>User's Guide</citetitle>, there are
10665 reference pages covering the postmaster and more utility
10666 programs, and a new appendix
10667 contains details on date/time behavior.
10668 The <citetitle>Administrator's Guide</citetitle> has a new
10669 chapter on troubleshooting from Tom Lane.
10670 And the <citetitle>Programmer's Guide</citetitle> has a
10671 description of query processing, also from Stefan, and details
10672 on obtaining the <productname>PostgreSQL</productname> source
10673 tree via anonymous <productname>CVS</productname> and
10674 <productname>CVSup</productname>.
10682 <title>Migration to version 6.5</title>
10685 A dump/restore using <application>pg_dump</application>
10686 is required for those wishing to migrate data from any
10687 previous release of <productname>PostgreSQL</productname>.
10688 <application>pg_upgrade</application> can <emphasis>not</emphasis>
10689 be used to upgrade to this release because the on-disk structure
10690 of the tables has changed compared to previous releases.
10694 The new Multiversion Concurrency Control (MVCC) features can
10695 give somewhat different behaviors in multiuser
10696 environments. <emphasis>Read and understand the following section
10697 to ensure that your existing applications will give you the
10698 behavior you need.</emphasis>
10702 <title>Multiversion Concurrency Control</title>
10705 Because readers in 6.5 don't lock data, regardless of transaction
10706 isolation level, data read by one transaction can be overwritten by
10707 another. In other words, if a row is returned by
10708 <command>SELECT</command> it doesn't mean that this row really exists
10709 at the time it is returned (i.e. sometime after the statement or
10710 transaction began) nor that the row is protected from being deleted or
10711 updated by concurrent transactions before the current transaction does
10712 a commit or rollback.
10716 To ensure the actual existence of a row and protect it against
10717 concurrent updates one must use <command>SELECT FOR UPDATE</command> or
10718 an appropriate <command>LOCK TABLE</command> statement. This should be
10719 taken into account when porting applications from previous releases of
10720 <productname>PostgreSQL</productname> and other environments.
10724 Keep the above in mind if you are using
10725 <filename>contrib/refint.*</filename> triggers for
10726 referential integrity. Additional techniques are required now. One way is
10727 to use <command>LOCK parent_table IN SHARE ROW EXCLUSIVE MODE</command>
10728 command if a transaction is going to update/delete a primary key and
10729 use <command>LOCK parent_table IN SHARE MODE</command> command if a
10730 transaction is going to update/insert a foreign key.
10734 Note that if you run a transaction in SERIALIZABLE mode then you must
10735 execute the <command>LOCK</command> commands above before execution of any
10736 <acronym>DML</acronym> statement
10737 (<command>SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO</command>) in the
10744 These inconveniences will disappear in the future
10745 when the ability to read dirty
10746 (uncommitted) data (regardless of isolation level) and true referential
10747 integrity will be implemented.
10753 <title>Changes</title>
10759 Fix text<->float8 and text<->float4 conversion functions(Thomas)
10760 Fix for creating tables with mixed-case constraints(Billy)
10761 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
10762 Fix bug in pg_dump -z
10763 Memory overrun cleanups(Tatsuo)
10764 Fix for lo_import crash(Tatsuo)
10765 Adjust handling of data type names to suppress double quotes(Thomas)
10766 Use type coercion for matching columns and DEFAULT(Thomas)
10767 Fix deadlock so it only checks once after one second of sleep(Bruce)
10768 Fixes for aggregates and PL/pgsql(Hiroshi)
10769 Fix for subquery crash(Vadim)
10770 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
10771 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo)
10772 Fix for pg_dump -d or -D and quote special characters in INSERT
10773 Repair serious problems with dynahash(Tom)
10774 Fix INET/CIDR portability problems
10775 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
10776 Fix executor so mergejoin of different column types works(Tom)
10777 Fix for Alpha OR selectivity bug
10778 Fix OR index selectivity problem(Bruce)
10779 Fix so \d shows proper length for char()/varchar()(Ryan)
10780 Fix tutorial code(Clark)
10781 Improve destroyuser checking(Oliver)
10782 Fix for Kerberos(Rodney McDuff)
10783 Fix for dropping database while dirty buffers(Bruce)
10784 Fix so sequence nextval() can be case-sensitive(Bruce)
10786 Drop buffers before destroying database files(Bruce)
10787 Fix case where executor evaluates functions twice(Tatsuo)
10788 Allow sequence nextval actions to be case-sensitive(Bruce)
10789 Fix optimizer indexing not working for negative numbers(Bruce)
10790 Fix for memory leak in executor with fjIsNull
10791 Fix for aggregate memory leaks(Erik Riedel)
10792 Allow user name containing a dash to grant privileges
10793 Cleanup of NULL in inet types
10794 Clean up system table bugs(Tom)
10795 Fix problems of PAGER and \? command(Masaaki Sakaida)
10796 Reduce default multisegment file size limit to 1GB(Peter)
10797 Fix for dumping of CREATE OPERATOR(Tom)
10798 Fix for backward scanning of cursors(Hiroshi Inoue)
10799 Fix for COPY FROM STDIN when using \i(Tom)
10800 Fix for subselect is compared inside an expression(Jan)
10801 Fix handling of error reporting while returning rows(Tom)
10802 Fix problems with reference to array types(Tom,Jan)
10803 Prevent UPDATE SET oid(Jan)
10804 Fix pg_dump so -t option can handle case-sensitive tablenames
10805 Fixes for GROUP BY in special cases(Tom, Jan)
10806 Fix for memory leak in failed queries(Tom)
10807 DEFAULT now supports mixed-case identifiers(Tom)
10808 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
10809 Disable use of pg_dump with both -o and -d options(Bruce)
10810 Allow pg_dump to properly dump group privileges(Bruce)
10811 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
10812 Fix for computations in views(Jan)
10813 Fix for aggregates on array indexes(Tom)
10814 Fix for DEFAULT handles single quotes in value requiring too many quotes
10815 Fix security problem with non-super users importing/exporting large objects(Tom)
10816 Rollback of transaction that creates table cleaned up properly(Tom)
10817 Fix to allow long table and column names to generate proper serial names(Tom)
10821 Add "vacuumdb" utility
10822 Speed up libpq by allocating memory better(Tom)
10823 EXPLAIN all indexes used(Tom)
10824 Implement CASE, COALESCE, NULLIF expression(Thomas)
10825 New pg_dump table output format(Constantin)
10826 Add string min()/max() functions(Thomas)
10827 Extend new type coercion techniques to aggregates(Thomas)
10828 New moddatetime contrib(Terry)
10829 Update to pgaccess 0.96(Constantin)
10830 Add routines for single-byte "char" type(Thomas)
10831 Improved substr() function(Thomas)
10832 Improved multibyte handling(Tatsuo)
10833 Multiversion concurrency control/MVCC(Vadim)
10834 New Serialized mode(Vadim)
10835 Fix for tables over 2gigs(Peter)
10836 New SET TRANSACTION ISOLATION LEVEL(Vadim)
10837 New LOCK TABLE IN ... MODE(Vadim)
10838 Update ODBC driver(Byron)
10839 New NUMERIC data type(Jan)
10840 New SELECT FOR UPDATE(Vadim)
10841 Handle "NaN" and "Infinity" for input values(Jan)
10842 Improved date/year handling(Thomas)
10843 Improved handling of backend connections(Magnus)
10844 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
10845 New TCL_ARRAYS option(Massimo)
10846 New INTERSECT and EXCEPT(Stefan)
10847 New pg_index.indisprimary for primary key tracking(D'Arcy)
10848 New pg_dump option to allow dropping of tables before creation(Brook)
10849 Speedup of row output routines(Tom)
10850 New READ COMMITTED isolation level(Vadim)
10851 New TEMP tables/indexes(Bruce)
10852 Prevent sorting if result is already sorted(Jan)
10853 New memory allocation optimization(Jan)
10854 Allow psql to do \p\g(Bruce)
10855 Allow multiple rule actions(Jan)
10856 Added LIMIT/OFFSET functionality(Jan)
10857 Improve optimizer when joining a large number of tables(Bruce)
10858 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
10859 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
10860 Improved int8 support(Ryan Bradetich, Thomas, Tom)
10861 New routines to convert between int8 and text/varchar types(Thomas)
10862 New bushy plans, where meta-tables are joined(Bruce)
10863 Enable right-hand queries by default(Bruce)
10864 Allow reliable maximum number of backends to be set at configure time
10865 (--with-maxbackends and postmaster switch (-N backends))(Tom)
10866 GEQO default now 10 tables because of optimizer speedups(Tom)
10867 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
10868 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
10869 Allow psql \d on a view show query(Ryan)
10870 Speedup for LIKE(Bruce)
10871 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
10872 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
10873 Make % operator have precedence like /(Bruce)
10874 Add new postgres -O option to allow system table structure changes(Bruce)
10875 Update contrib/pginterface/findoidjoins script(Tom)
10876 Major speedup in vacuum of deleted rows with indexes(Vadim)
10877 Allow non-SQL functions to run different versions based on arguments(Tom)
10878 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida)
10879 Add version number in start-up banners for psql(Masaaki Sakaida)
10880 New contrib/vacuumlo removes large objects not referenced(Peter)
10881 New initialization for table sizes so non-vacuumed tables perform better(Tom)
10882 Improve error messages when a connection is rejected(Tom)
10883 Support for arrays of char() and varchar() fields(Massimo)
10884 Overhaul of hash code to increase reliability and performance(Tom)
10885 Update to PyGreSQL 2.4(D'Arcy)
10886 Changed debug options so -d4 and -d5 produce different node displays(Jan)
10887 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
10888 Better optimization statistics for system table access(Tom)
10889 Better handling of non-default block sizes(Massimo)
10890 Improve GEQO optimizer memory consumption(Tom)
10891 UNION now suppports ORDER BY of columns not in target list(Jan)
10892 Major libpq++ improvements(Vince Vielhaber)
10893 pg_dump now uses -z(ACL's) as default(Bruce)
10894 backend cache, memory speedups(Tom)
10895 have pg_dump do everything in one snapshot transaction(Vadim)
10896 fix for large object memory leakage, fix for pg_dumping(Tom)
10897 INET type now respects netmask for comparisons
10898 Make VACUUM ANALYZE only use a readlock(Vadim)
10899 Allow VIEWs on UNIONS(Jan)
10900 pg_dump now can generate consistent snapshots on active databases(Vadim)
10902 Source Tree Changes
10903 -------------------
10904 Improve port matching(Tom)
10905 Portability fixes for SunOS
10906 Add Windows NT backend port and enable dynamic loading(Magnus and Daniel Horak)
10907 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
10908 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
10909 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
10910 Port to NetBSD/macppc(Toshimi Aoki)
10911 Fix for tcl/tk configuration(Vince)
10912 Removed CURRENT key word for rule queries(Jan)
10913 NT dynamic loading now works(Daniel Horak)
10914 Add ARM32 support(Andrew McMurry)
10915 Better support for HP-UX 11 and UnixWare
10916 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
10917 New install commands for plpgsql(Jan)
10924 <sect1 id="release-6-4-2">
10925 <title>Release 6.4.2</title>
10928 <title>Release date</title>
10929 <simpara>1998-12-20</simpara>
10933 The 6.4.1 release was improperly packaged. This also has one additional
10939 <title>Migration to version 6.4.2</title>
10942 A dump/restore is <emphasis>not</emphasis> required for those running
10947 <title>Changes</title>
10951 Fix for datetime constant problem on some platforms(Thomas)
10959 <sect1 id="release-6-4-1">
10960 <title>Release 6.4.1</title>
10963 <title>Release date</title>
10964 <simpara>1998-12-18</simpara>
10968 This is basically a cleanup release for 6.4. We have fixed a variety of
10969 problems reported by 6.4 users.
10974 <title>Migration to version 6.4.1</title>
10977 A dump/restore is <emphasis>not</emphasis> required for those running
10982 <title>Changes</title>
10986 Add pg_dump -N flag to force double quotes around identifiers. This is
10987 the default(Thomas)
10988 Fix for NOT in where clause causing crash(Bruce)
10989 EXPLAIN VERBOSE coredump fix(Vadim)
10990 Fix shared-library problems on Linux
10991 Fix test for table existence to allow mixed-case and whitespace in
10992 the table name(Thomas)
10993 Fix a couple of pg_dump bugs
10994 Configure matches template/.similar entries better(Tom)
10995 Change builtin function names from SPI_* to spi_*
10996 OR WHERE clause fix(Vadim)
10997 Fixes for mixed-case table names(Billy)
10998 contrib/linux/postgres.init.csh/sh fix(Thomas)
10999 libpq memory overrun fix
11001 Change exp() behavior to generate error on underflow(Thomas)
11002 pg_dump fixes for memory leak, inheritance constraints, layout change
11003 update pgaccess to 0.93
11004 Fix prototype for 64-bit platforms
11005 Multibyte fixes(Tatsuo)
11007 Fix memory overruns(Tatsuo)
11008 Fix for lo_import() crash(Bruce)
11009 Better search for install program(Tom)
11010 Timezone fixes(Tom)
11012 Use implicit type coercion for matching DEFAULT values(Thomas)
11013 Add routines to help with single-byte (internal) character type(Thomas)
11014 Compilation of libpq for Windows fixes(Magnus)
11015 Upgrade to PyGreSQL 2.2(D'Arcy)
11023 <sect1 id="release-6-4">
11024 <title>Release 6.4</title>
11027 <title>Release date</title>
11028 <simpara>1998-10-30</simpara>
11032 There are <emphasis>many</emphasis> new features and improvements in this release.
11033 Thanks to our developers and maintainers, nearly every aspect of the system
11034 has received some attention since the previous release.
11035 Here is a brief, incomplete summary:
11040 Views and rules are now functional thanks to extensive new code in the
11041 rewrite rules system from Jan Wieck. He also wrote a chapter on it
11042 for the <citetitle>Programmer's Guide</citetitle>.
11047 Jan also contributed a second procedural language, <application>PL/pgSQL</application>, to go with the
11048 original <application>PL/pgTCL</application> procedural language he contributed last release.
11054 We have optional multiple-byte character set support from Tatsuo Ishii
11055 to complement our existing locale support.
11061 Client/server communications has been cleaned up, with better support for
11062 asynchronous messages and interrupts thanks to Tom Lane.
11068 The parser will now perform automatic type coercion to match arguments
11069 to available operators and functions, and to match columns and expressions
11070 with target columns. This uses a generic mechanism which supports
11071 the type extensibility features of <productname>PostgreSQL</productname>.
11072 There is a new chapter in the <citetitle>User's Guide</citetitle>
11073 which covers this topic.
11079 Three new data types have been added.
11080 Two types, <type>inet</type> and <type>cidr</type>, support various forms
11081 of IP network, subnet, and machine addressing. There is now an 8-byte integer
11082 type available on some platforms. See the chapter on data types
11083 in the <citetitle>User's Guide</citetitle> for details.
11084 A fourth type, <type>serial</type>, is now supported by the parser as an
11085 amalgam of the <type>int4</type> type, a sequence, and a unique index.
11091 Several more <acronym>SQL92</acronym>-compatible syntax features have been
11092 added, including <command>INSERT DEFAULT VALUES</command>
11098 The automatic configuration and installation system has received some
11099 attention, and should be more robust for more platforms than it has ever
11108 <title>Migration to version 6.4</title>
11111 A dump/restore using <application>pg_dump</application>
11112 or <application>pg_dumpall</application>
11113 is required for those wishing to migrate data from any
11114 previous release of <productname>PostgreSQL</productname>.
11119 <title>Changes</title>
11125 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
11126 Remove char2-16 data types, use char/varchar(Darren)
11127 Pqfn not handles a NOTICE message(Anders)
11128 Reduced busywaiting overhead for spinlocks with many backends (dg)
11129 Stuck spinlock detection (dg)
11130 Fix up "ISO-style" timespan decoding and encoding(Thomas)
11131 Fix problem with table drop after rollback of transaction(Vadim)
11132 Change error message and remove non-functional update message(Vadim)
11133 Fix for COPY array checking
11134 Fix for SELECT 1 UNION SELECT NULL
11135 Fix for buffer leaks in large object calls(Pascal)
11136 Change owner from oid to int4 type(Bruce)
11137 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
11138 Fix for shared invalidation cache overflow(Massimo)
11139 Prevent file descriptor leaks in failed COPY's(Bruce)
11140 Fix memory leak in libpgtcl's pg_select(Constantin)
11141 Fix problems with username/passwords over 8 characters(Tom)
11142 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
11143 Fix of many bad system table entries(Tom)
11147 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
11148 Show the index used in an EXPLAIN(Zeugswetter)
11149 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
11150 Multibyte awareness of many data types and functions, via configure(Tatsuo)
11151 New configure --with-mb option(Tatsuo)
11152 New initdb --pgencoding option(Tatsuo)
11153 New createdb -E multibyte option(Tatsuo)
11154 Select version(); now returns PostgreSQL version(Jeroen)
11155 libpq now allows asynchronous clients(Tom)
11156 Allow cancel from client of backend query(Tom)
11157 psql now cancels query with Control-C(Tom)
11158 libpq users need not issue dummy queries to get NOTIFY messages(Tom)
11159 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
11160 PGresult struct now includes associated error message, if any(Tom)
11161 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
11162 Add routines to convert between varchar and bpchar(Thomas)
11163 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
11164 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
11165 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas)
11166 Fixes for unary minus parsing with leading spaces(Thomas)
11167 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
11168 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
11169 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
11170 Enable HAVING clause but no fixes elsewhere yet.
11171 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thomas)
11172 Save string type if specified for DEFAULT clause handling(Thomas)
11173 Coerce operations involving different data types(Thomas)
11174 Allow some index use for columns of different types(Thomas)
11175 Add capabilities for automatic type conversion(Thomas)
11176 Cleanups for large objects, so file is truncated on open(Peter)
11177 Readline cleanups(Tom)
11178 Allow psql \f \ to make spaces as delimiter(Bruce)
11179 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
11180 Msql compatibility library in /contrib(Aldrin)
11181 Remove the requirement that ORDER/GROUP BY clause identifiers be
11182 included in the target list(David)
11183 Convert columns to match columns in UNION clauses(Thomas)
11184 Remove fork()/exec() and only do fork()(Bruce)
11185 Jdbc cleanups(Peter)
11186 Show backend status on ps command line(only works on some platforms)(Bruce)
11187 Pg_hba.conf now has a sameuser option in the database field
11188 Make lo_unlink take oid param, not int4
11189 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
11190 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
11191 libpgtcl cleanups(Tom)
11192 Add -error option to libpgtcl's pg_result command(Tom)
11193 New locale patch, see docs/README/locale(Oleg)
11194 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
11195 New contrib/lo code for large object orphan removal(Peter)
11196 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
11197 feature, see /doc/README.mb(Tatsuo)
11198 contrib/noupdate code to revoke update permission on a column
11199 libpq can now be compiled on Windows(Magnus)
11200 Add PQsetdbLogin() in libpq
11201 New 8-byte integer type, checked by configure for OS support(Thomas)
11202 Better support for quoted table/column names(Thomas)
11203 Surround table and column names with double-quotes in pg_dump(Thomas)
11204 PQreset() now works with passwords(Tom)
11205 Handle case of GROUP BY target list column number out of range(David)
11206 Allow UNION in subselects
11207 Add auto-size to screen to \d? commands(Bruce)
11208 Use UNION to show all \d? results in one query(Bruce)
11209 Add \d? field search feature(Bruce)
11210 Pg_dump issues fewer \connect requests(Tom)
11211 Make pg_dump -z flag work better, document it in manual page(Tom)
11212 Add HAVING clause with full support for subselects and unions(Stephan)
11213 Full text indexing routines in contrib/fulltextindex(Maarten)
11214 Transaction ids now stored in shared memory(Vadim)
11215 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
11216 Support for SQL92 syntax "SET NAMES"(Tatsuo)
11217 Support for LATIN2-5(Tatsuo)
11218 Add UNICODE regression test case(Tatsuo)
11219 Lock manager cleanup, new locking modes for LLL(Vadim)
11220 Allow index use with OR clauses(Bruce)
11221 Allows "SELECT NULL ORDER BY 1;"
11222 Explain VERBOSE prints the plan, and now pretty-prints the plan to
11223 the postmaster log file(Bruce)
11224 Add indexes display to \d command(Bruce)
11225 Allow GROUP BY on functions(David)
11226 New pg_class.relkind for large objects(Bruce)
11227 New way to send libpq NOTICE messages to a different location(Tom)
11228 New \w write command to psql(Bruce)
11229 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
11230 Allow binary-compatible indexes to be considered when checking for valid
11231 Indexes for restriction clauses containing a constant(Thomas)
11232 New ISBN/ISSN code in /contrib/isbn_issn
11233 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
11234 New rewrite system fixes many problems with rules and views(Jan)
11235 * Rules on relations work
11236 * Event qualifications on insert/update/delete work
11237 * New OLD variable to reference CURRENT, CURRENT will be remove in future
11238 * Update rules can reference NEW and OLD in rule qualifications/actions
11239 * Insert/update/delete rules on views work
11240 * Multiple rule actions are now supported, surrounded by parentheses
11241 * Regular users can create views/rules on tables they have RULE permits
11242 * Rules and views inherit the privileges of the creator
11243 * No rules at the column level
11244 * No UPDATE NEW/OLD rules
11245 * New pg_tables, pg_indexes, pg_rules and pg_views system views
11246 * Only a single action on SELECT rules
11247 * Total rewrite overhaul, perhaps for 6.5
11248 * handle subselects
11249 * handle aggregates on views
11250 * handle insert into select from view works
11251 System indexes are now multikey(Bruce)
11252 Oidint2, oidint4, and oidname types are removed(Bruce)
11253 Use system cache for more system table lookups(Bruce)
11254 New backend programming language PL/pgSQL in backend/pl(Jan)
11255 New SERIAL data type, auto-creates sequence/index(Thomas)
11256 Enable assert checking without a recompile(Massimo)
11257 User lock enhancements(Massimo)
11258 New setval() command to set sequence value(Massimo)
11259 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
11260 Conditional trace package(Massimo)
11261 New UNLISTEN command(Massimo)
11262 psql and libpq now compile under Windows using win32.mak(Magnus)
11263 Lo_read no longer stores trailing NULL(Bruce)
11264 Identifiers are now truncated to 31 characters internally(Bruce)
11265 Createuser options now availble on the command line
11266 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
11267 Prevent file descriptor leaf from failed COPY(Bruce)
11268 New pg_upgrade command(Bruce)
11269 Updated /contrib directories(Massimo)
11270 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
11271 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
11272 New DECLARE and FETCH feature(Thomas)
11273 libpq's internal structures now not exported(Tom)
11274 Allow up to 8 key indexes(Bruce)
11275 Remove ARCHIVE key word, that is no longer used(Thomas)
11276 pg_dump -n flag to supress quotes around indentifiers
11277 disable system columns for views(Jan)
11278 new INET and CIDR types for network addresses(TomH, Paul)
11279 no more double quotes in psql output
11280 pg_dump now dumps views(Terry)
11281 new SET QUERY_LIMIT(Tatsuo,Jan)
11283 Source Tree Changes
11284 -------------------
11285 /contrib cleanup(Jun)
11286 Inline some small functions called for every row(Bruce)
11288 HP-UX cleanups(Tom)
11289 Multibyte regression tests(Soonmyung.)
11290 Remove --disabled options from configure
11291 Define PGDOC to use POSTGRESDIR by default
11292 Make regression optional
11293 Remove extra braces code to pgindent(Bruce)
11294 Add bsdi shared library support(Bruce)
11295 New --without-CXX support configure option(Brook)
11297 Update backend flowchart in tools/backend(Bruce)
11298 Change atttypmod from int16 to int32(Bruce, Tom)
11299 Getrusage() fix for platforms that do not have it(Tom)
11300 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
11301 NS32K platform fixes(Phil Nelson, John Buller)
11302 SCO 7/UnixWare 2.x fixes(Billy,others)
11303 Sparc/Solaris 2.5 fixes(Ryan)
11304 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
11305 Even more documention(Thomas)
11306 Nextstep support(Jacek)
11308 pginterface manual page(Bruce)
11309 shared libraries all have version numbers
11310 merged all OS-specific shared library defines into one file
11311 smarter TCL/TK configuration checking(Billy)
11312 smarter perl configuration(Brook)
11313 configure uses supplied install-sh if no install script found(Tom)
11314 new Makefile.shlib for shared library configuration(Tom)
11320 <sect1 id="release-6-3-2">
11321 <title>Release 6.3.2</title>
11324 <title>Release date</title>
11325 <simpara>1998-04-07</simpara>
11329 This is a bug-fix release for 6.3.x.
11330 Refer to the release notes for version 6.3 for a more complete summary of new features.
11338 Repairs automatic configuration support for some platforms, including Linux,
11339 from breakage inadvertently introduced in version 6.3.1.
11345 Correctly handles function calls on the left side of BETWEEN and LIKE clauses.
11352 A dump/restore is NOT required for those running 6.3 or 6.3.1. A
11353 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11354 This last step should be performed while the postmaster is not running.
11355 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11358 For upgrades from pre-6.3 installations,
11359 refer to the installation and migration instructions for version 6.3.
11363 <title>Changes</title>
11367 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
11368 Manual page improvements(Bruce)
11369 BETWEEN and LIKE fix(Thomas)
11370 fix for psql \connect used by pg_dump(Oliver Elphick)
11372 pgaccess, version 0.86
11373 qsort removed, now uses libc version, cleanups(Jeroen)
11374 fix for buffer over-runs detected(Maurice Gittens)
11375 fix for buffer overrun in libpgtcl(Randy Kunkee)
11376 fix for UNION with DISTINCT or ORDER BY(Bruce)
11377 gettimeofday configure check(Doug Winterburn)
11378 Fix "indexes not used" bug(Vadim)
11379 docs additions(Thomas)
11380 Fix for backend memory leak(Bruce)
11381 libreadline cleanup(Erwan MAS)
11382 Remove DISTDIR(Bruce)
11383 Makefile dependency cleanup(Jeroen van Vianen)
11384 ASSERT fixes(Bruce)
11390 <sect1 id="release-6-3-1">
11391 <title>Release 6.3.1</title>
11394 <title>Release date</title>
11395 <simpara>1998-03-23</simpara>
11404 Additional support for multibyte character sets.
11410 Repair byte ordering for mixed-endian clients and servers.
11416 Minor updates to allowed SQL syntax.
11422 Improvements to the configuration autodetection for installation.
11429 A dump/restore is NOT required for those running 6.3. A
11430 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11431 This last step should be performed while the postmaster is not running.
11432 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11435 For upgrades from pre-6.3 installations,
11436 refer to the installation and migration instructions for version 6.3.
11440 <title>Changes</title>
11444 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
11445 pg_user cleanup(Bruce)
11446 large object fix for pg_dump and tclsh (alvin)
11447 LIKE fix for multiple adjacent underscores
11448 fix for redefining builtin functions(Thomas)
11450 upgrade to pg_access 0.83
11451 updated CLUSTER manual page
11452 multibyte character set support, see doc/README.mb(Tatsuo)
11453 configure --with-pgport fix
11455 big-endian fix for backend communications(Kataoka)
11456 SUBSTR() and substring() fix(Jan)
11457 several jdbc fixes(Peter)
11458 libpgtcl improvements, see libptcl/README(Randy Kunkee)
11459 Fix for "Datasize = 0" error(Vadim)
11460 Prevent \do from wrapping(Bruce)
11461 Remove duplicate Russian character set entries
11463 Allow optional TABLE key word in LOCK and SELECT INTO(Thomas)
11464 CREATE SEQUENCE options to allow a negative integer(Thomas)
11465 Add "PASSWORD" as an allowed column identifier(Thomas)
11466 Add checks for UNION target fields(Bruce)
11467 Fix Alpha port(Dwayne Bailey)
11468 Fix for text arrays containing quotes(Doug Gibson)
11469 Solaris compile fix(Albert Chin-A-Young)
11470 Better identify tcl and tk libs and includes(Bruce)
11476 <sect1 id="release-6-3">
11477 <title>Release 6.3</title>
11480 <title>Release date</title>
11481 <simpara>1998-03-01</simpara>
11485 There are <emphasis>many</emphasis> new features and improvements in this release.
11486 Here is a brief, incomplete summary:
11491 Many new SQL features, including
11492 full <acronym>SQL92</acronym> subselect capability
11493 (everything is here but target-list subselects).
11499 Support for client-side environment variables to specify time zone and date style.
11505 Socket interface for client/server connection. This is the default now
11506 so you may need to start <application>postmaster</application> with the
11507 <option>-i</option> flag.
11513 Better password authorization mechanisms. Default table privileges have changed.
11519 Old-style <firstterm>time travel</firstterm>
11520 has been removed. Performance has been improved.
11528 Bruce Momjian wrote the following notes to introduce the new release.
11533 There are some general 6.3 issues that I want to mention. These are
11534 only the big items that cannot be described in one sentence. A review
11535 of the detailed changes list is still needed.
11538 First, we now have subselects. Now that we have them, I would like to
11539 mention that without subselects, SQL is a very limited language.
11540 Subselects are a major feature, and you should review your code for
11541 places where subselects provide a better solution for your queries. I
11542 think you will find that there are more uses for subselects than you may
11543 think. Vadim has put us on the big SQL map with subselects, and fully
11544 functional ones too. The only thing you can't do with subselects is to
11545 use them in the target list.
11548 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
11549 enable connections from other machines, you have to use the new
11550 postmaster -i option, and of course edit <filename>pg_hba.conf</filename>. Also, for this
11551 reason, the format of <filename>pg_hba.conf</filename> has changed.
11554 Third, <type>char()</type> fields will now allow faster access than <type>varchar()</type> or
11555 <type>text</type>. Specifically, the <type>text</> and <type>varchar()</type> have a penalty for access to
11556 any columns after the first column of this type. <type>char()</type> used to also
11557 have this access penalty, but it no longer does. This may suggest that
11558 you redesign some of your tables, especially if you have short character
11559 columns that you have defined as <type>varchar()</type> or <type>text</type>. This and other
11560 changes make 6.3 even faster than earlier releases.
11563 We now have passwords definable independent of any Unix file. There are
11564 new SQL USER commands.
11565 See the <citetitle>Administrator's Guide</citetitle> for more
11566 information. There is a new table, pg_shadow, which is used to store
11567 user information and user passwords, and it by default only SELECT-able
11568 by the <systemitem>postgres</systemitem> super-user. pg_user is now a view of pg_shadow, and is
11569 SELECT-able by PUBLIC. You should keep using pg_user in your
11570 application without changes.
11573 User-created tables now no longer have SELECT privilege to PUBLIC by
11574 default. This was done because the ANSI standard requires it. You can
11575 of course GRANT any privileges you want after the table is created.
11576 System tables continue to be SELECT-able by PUBLIC.
11579 We also have real deadlock detection code. No more sixty-second
11580 timeouts. And the new locking code implements a <acronym>FIFO</acronym> better, so there
11581 should be less resource starvation during heavy use.
11584 Many complaints have been made about inadequate documentation in previous
11585 releases. Thomas has put much effort into many new manuals for this
11586 release. Check out the doc/ directory.
11589 For performance reasons, time travel is gone, but can be implemented
11590 using triggers (see <filename>pgsql/contrib/spi/README</filename>). Please check out the new
11591 \d command for types, operators, etc. Also, views have their own
11592 privileges now, not based on the underlying tables, so privileges on
11593 them have to be set separately. Check <filename>/pgsql/interfaces</filename> for some new
11594 ways to talk to <productname>PostgreSQL</productname>.
11597 This is the first release that really required an explanation for
11598 existing users. In many ways, this was necessary because the new
11599 release removes many limitations, and the work-arounds people were using
11600 are no longer needed.
11604 <title>Migration to version 6.3</title>
11607 A dump/restore using <application>pg_dump</application>
11608 or <application>pg_dumpall</application>
11609 is required for those wishing to migrate data from any
11610 previous release of <productname>PostgreSQL</productname>.
11615 <title>Changes</title>
11621 Fix binary cursors broken by MOVE implementation(Vadim)
11622 Fix for tcl library crash(Jan)
11623 Fix for array handling, from Gerhard Hintermayer
11624 Fix acl error, and remove duplicate pqtrace(Bruce)
11625 Fix psql \e for empty file(Bruce)
11626 Fix for textcat on varchar() fields(Bruce)
11627 Fix for DBT Sendproc (Zeugswetter Andres)
11628 Fix vacuum analyze syntax problem(Bruce)
11629 Fix for international identifiers(Tatsuo)
11630 Fix aggregates on inherited tables(Bruce)
11631 Fix substr() for out-of-bounds data
11632 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
11633 Fix notty output to show status result. -q option still turns it off(Bruce)
11634 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
11636 Fix for PQtrace start/stop several times(Bruce)
11637 Fix a variety of locking problems like newer lock waiters getting
11638 lock before older waiters, and having readlock people not share
11639 locks if a writer is waiting for a lock, and waiting writers not
11640 getting priority over waiting readers(Bruce)
11641 Fix crashes in psql when executing queries from external files(James)
11642 Fix problem with multiple order by columns, with the first one having
11643 NULL values(Jeroen)
11644 Use correct hash table support functions for float8 and int4(Thomas)
11645 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
11646 Change precedence for boolean operators to match expected behavior(Thomas)
11647 Generate elog(ERROR) on over-large integer(Bruce)
11648 Allow multiple-argument functions in constraint clauses(Thomas)
11649 Check boolean input literals for 'true','false','yes','no','1','0'
11650 and throw elog(ERROR) if unrecognized(Thomas)
11651 Major large objects fix
11652 Fix for GROUP BY showing duplicates(Vadim)
11653 Fix for index scans in MergeJion(Vadim)
11657 Subselects with EXISTS, IN, ALL, ANY key words (Vadim, Bruce, Thomas)
11658 New User Manual(Thomas, others)
11659 Speedup by inlining some frequently-called functions
11660 Real deadlock detection, no more timeouts(Bruce)
11661 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
11662 CURRENT_USER(Thomas)
11663 Modify constraint syntax to be SQL92-compliant(Thomas)
11664 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
11665 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
11666 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas)
11667 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
11668 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
11669 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
11670 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
11671 Allow SQL92 delimited identifiers(Thomas)
11672 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas)
11673 Support SQL92 syntax for type coercion of literal strings
11674 (e.g. "DATETIME 'now'")(Thomas)
11675 Add conversions for int2, int4, and OID types to and from text(Thomas)
11676 Use shared lock when building indexes(Vadim)
11677 Free memory allocated for an user query inside transaction block after
11678 this query is done, was turned off in <= 6.2.1(Vadim)
11679 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
11680 New <productname>PostgreSQL</productname> Procedural Language (PL) backend interface(Jan)
11681 Rename pg_dump -H option to -h(Bruce)
11682 Add Java support for passwords, European dates(Peter)
11683 Use indexes for LIKE and ~, !~ operations(Bruce)
11684 Add hash functions for datetime and timespan(Thomas)
11685 Time Travel removed(Vadim, Bruce)
11686 Add paging for \d and \z, and fix \i(Bruce)
11687 Add Unix domain socket support to backend and to frontend library(Goran)
11688 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
11689 Allow more SQL92 and/or <productname>PostgreSQL</productname> reserved words as column identifiers(Thomas)
11690 Augment support for SQL92 SET TIME ZONE...(Thomas)
11691 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
11692 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
11693 Enable SET TIME ZONE using TZ environment variable(Thomas)
11694 Add PGDATESTYLE environment variable to frontend and backend initialization(Thomas)
11695 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
11696 frontend library initialization environment variables(Thomas)
11697 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
11698 Add pg_description table for info on tables, columns, operators, types, and
11700 Increase 16 char limit on system table/index names to 32 characters(Bruce)
11701 Rename system indexes(Bruce)
11702 Add 'GERMAN' option to SET DATESTYLE(Thomas)
11703 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
11704 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
11705 Validate numeric input more carefully for delta times(Thomas)
11706 Implement day of year as possible input to date_part()(Thomas)
11707 Define timespan_finite() and text_timespan() functions(Thomas)
11708 Remove archive stuff(Bruce)
11709 Allow for a pg_password authentication database that is separate from
11710 the system password file(Todd)
11711 Dump ACLs, GRANT, REVOKE privileges(Matt)
11712 Define text, varchar, and bpchar string length functions(Thomas)
11713 Fix Query handling for inheritance, and cost computations(Bruce)
11714 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
11715 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
11716 Implement UNIONs for SELECT(Bruce)
11717 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
11718 varchar() stores only necessary bytes on disk(Bruce)
11719 Fix for BLOBs(Peter)
11720 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
11721 Remove unused "option" from PQconnectdb()
11722 New LOCK command and lock manual page describing deadlocks(Bruce)
11723 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
11724 Enhance psql \z to show sequences(Bruce)
11725 Show NOT NULL and DEFAULT in psql \d table(Bruce)
11726 New psql .psqlrc file start-up(Andrew)
11727 Modify sample start-up script in contrib/linux to show syslog(Thomas)
11728 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
11729 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
11730 Update of contrib stuff(Massimo)
11731 Add Unix socket support to DBD::Pg(Goran)
11732 New python interface (PyGreSQL 2.0)(D'Arcy)
11733 New frontend/backend protocol has a version number, network byte order(Phil)
11734 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
11735 CHAR() now faster access than VARCHAR() or TEXT
11736 ecpg embedded SQL preprocessor
11737 Reduce system column overhead(Vadmin)
11738 Remove pg_time table(Vadim)
11739 Add pg_type attribute to identify types that need length (bpchar, varchar)
11740 Add report of offending line when COPY command fails
11741 Allow VIEW privileges to be set separately from the underlying tables.
11742 For security, use GRANT/REVOKE on views as appropriate(Jan)
11743 Tables now have no default GRANT SELECT TO PUBLIC. You must
11744 explicitly grant such privileges.
11745 Clean up tutorial examples(Darren)
11747 Source Tree Changes
11748 -------------------
11749 Add new html development tools, and flow chart in /tools/backend
11750 Fix for SCO compiles
11751 Stratus computer port Robert Gillies
11752 Added support for shlib for BSD44_derived & i386_solaris
11753 Make configure more automated(Brook)
11754 Add script to check regression test results
11755 Break parser functions into smaller files, group together(Bruce)
11756 Rename heap_create to heap_create_and_catalog, rename heap_creatr
11757 to heap_create()(Bruce)
11758 Sparc/Linux patch for locking(TomS)
11759 Remove PORTNAME and reorganize port-specific stuff(Marc)
11760 Add optimizer README file(Bruce)
11761 Remove some recursion in optimizer and clean up some code there(Bruce)
11762 Fix for NetBSD locking(Henry)
11763 Fix for libptcl make(Tatsuo)
11765 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
11766 function calls to istrue() or isfalse() to allow optimization(Thomas)
11767 Various fixes NetBSD/Sparc related(TomH)
11768 Alpha linux locking(Travis,Ryan)
11769 Change elog(WARN) to elog(ERROR)(Bruce)
11770 FAQ for FreeBSD(Marc)
11771 Bring in the PostODBC source tree as part of our standard distribution(Marc)
11772 A minor patch for HP/UX 10 vs 9(Stan)
11773 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
11774 UnixWare patches(Billy)
11775 New i386 'lock' for spinlock asm(Billy)
11776 Support for multiplexed backends is removed
11777 Start an OpenBSD port
11779 Start a Cygnus port
11780 Add string functions to regression suite(Thomas)
11781 Expand a few function names formerly truncated to 16 characters(Thomas)
11782 Remove un-needed malloc() calls and replace with palloc()(Bruce)
11788 <sect1 id="release-6-2-1">
11789 <title>Release 6.2.1</title>
11792 <title>Release date</title>
11793 <simpara>1997-10-17</simpara>
11797 6.2.1 is a bug-fix and usability release on 6.2.
11805 Allow strings to span lines, per <acronym>SQL92</acronym>.
11811 Include example trigger function for inserting user names on table updates.
11818 This is a minor bug-fix release on 6.2.
11819 For upgrades from pre-6.2 systems, a full dump/reload is required.
11820 Refer to the 6.2 release notes for instructions.
11824 <title>Migration from version 6.2 to version 6.2.1</title>
11827 This is a minor bug-fix release. A dump/reload is not required from version 6.2,
11828 but is required from any release prior to 6.2.
11831 In upgrading from version 6.2, if you choose to dump/reload you will find that
11832 avg(money) is now calculated correctly. All other bug fixes take effect
11833 upon updating the executables.
11836 Another way to avoid dump/reload is to use the following SQL command
11837 from <command>psql</command> to update the existing system table:
11840 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
11841 where aggname = 'avg' and aggbasetype = 790;
11845 This will need to be done to every existing database, including template1.
11850 <title>Changes</title>
11854 Allow TIME and TYPE column names(Thomas)
11855 Allow larger range of true/false as boolean values(Thomas)
11856 Support output of "now" and "current"(Thomas)
11857 Handle DEFAULT with INSERT of NULL properly(Vadim)
11858 Fix for relation reference counts problem in buffer manager(Vadim)
11859 Allow strings to span lines, like ANSI(Thomas)
11860 Fix for backward cursor with ORDER BY(Vadim)
11861 Fix avg(cash) computation(Thomas)
11862 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
11863 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
11864 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
11870 <sect1 id="release-6-2">
11871 <title>Release 6.2</title>
11874 <title>Release date</title>
11875 <simpara>1997-10-02</simpara>
11879 A dump/restore is required for those wishing to migrate data from
11880 previous releases of <productname>PostgreSQL</productname>.
11884 <title>Migration from version 6.1 to version 6.2</title>
11887 This migration requires a complete dump of the 6.1 database and a
11888 restore of the database in 6.2.
11891 Note that the <command>pg_dump</command> and <command>pg_dumpall</command> utility from 6.2 should be used
11892 to dump the 6.1 database.
11897 <title>Migration from version 1.<replaceable>x</> to version 6.2</title>
11900 Those migrating from earlier 1.* releases should first upgrade to 1.09
11901 because the COPY output format was improved from the 1.02 release.
11906 <title>Changes</title>
11912 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
11913 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
11914 from Solaris(Diab Jerius)
11915 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
11916 Check for geometric intersections at endpoints to avoid rounding ugliness(Thomas)
11917 Catch non-functional delete attempts(Vadim)
11918 Change time function names to be more consistent(Michael Reifenberg)
11919 Check for zero divides(Michael Reifenberg)
11920 Fix very old bug which made rows changed/inserted by a command
11921 visible to the command itself (so we had multiple update of
11922 updated rows, etc.)(Vadim)
11923 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
11924 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
11925 Remove un-needed signal stuff from contrib/pginterface
11926 Fix OR (where x != 1 or x isnull didn't return rows with x NULL) (Vadim)
11927 Fix time_cmp function (Vadim)
11928 Fix handling of functions with non-attribute first argument in
11929 WHERE clauses (Vadim)
11930 Fix GROUP BY when order of entries is different from order
11931 in target list (Vadim)
11932 Fix pg_dump for aggregates without sfunc1 (Vadim)
11936 Default genetic optimizer GEQO parameter is now 8(Bruce)
11937 Allow use parameters in target list having aggregates in functions(Vadim)
11938 Added JDBC driver as an interface(Adrian & Peter)
11939 pg_password utility
11940 Return number of rows inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
11941 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
11942 SPI (Server Programming Interface) allows execution of queries inside
11943 C-functions (Vadim)
11944 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
11945 Include reserved words for string handling, outer joins, and unions(Thomas)
11946 Implement extended comments ("/* ... */") using exclusive states(Thomas)
11947 Add "//" single-line comments(Bruce)
11948 Remove some restrictions on characters in operator names(Thomas)
11949 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
11950 Add text concatenation operator and function (SQL92)(Thomas)
11951 Support WITH TIME ZONE syntax (SQL92)(Thomas)
11952 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
11953 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
11954 and CHARACTER VARYING (SQL92)(Thomas)
11955 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
11956 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
11957 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
11958 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
11959 Add more reserved words, mostly for SQL92 compliance(Thomas)
11960 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
11961 Add center() routines for lseg, path, polygon(Thomas)
11962 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
11963 Check explicitly for points and polygons contained within polygons
11964 using an axis-crossing algorithm(Thomas)
11965 Add routine to convert circle-box(Thomas)
11966 Merge conflicting operators for different geometric data types(Thomas)
11967 Replace distance operator "<===>" with "<->"(Thomas)
11968 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Thomas)
11969 Add routines for text trimming on both ends, substring, and string position(Thomas)
11970 Added conversion routines circle(box) and poly(circle)(Thomas)
11971 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
11972 Allow functions and operators on internally-identical types to succeed(Bruce)
11973 Speed up backend start-up after profiling analysis(Bruce)
11974 Inline frequently called functions for performance(Bruce)
11975 Reduce open() calls(Bruce)
11976 psql: Add PAGER for \h and \?,\C fix
11977 Fix for psql pager when no tty(Bruce)
11978 New entab utility(Bruce)
11979 General trigger functions for referential integrity (Vadim)
11980 General trigger functions for time travel (Vadim)
11981 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
11982 MOVE implementation (Vadim)
11984 Source Tree Changes
11985 -------------------
11986 HP-UX 10 patches (Vladimir Turin)
11987 Added SCO support, (Daniel Harris)
11988 MkLinux patches (Tatsuo Ishii)
11989 Change geometric box terminology from "length" to "width"(Thomas)
11990 Deprecate temporary unstored slope fields in geometric code(Thomas)
11991 Remove restart instructions from INSTALL(Bruce)
11992 Look in /usr/ucb first for install(Bruce)
11993 Fix c++ copy example code(Thomas)
11994 Add -o to psql manual page(Bruce)
11995 Prevent relname unallocated string length from being copied into database(Bruce)
11996 Cleanup for NAMEDATALEN use(Bruce)
11997 Fix pg_proc names over 15 chars in output(Bruce)
11998 Add strNcpy() function(Bruce)
11999 remove some (void) casts that are unnecessary(Bruce)
12000 new interfaces directory(Marc)
12001 Replace fopen() calls with calls to fd.c functions(Bruce)
12002 Make functions static where possible(Bruce)
12003 enclose unused functions in #ifdef NOT_USED(Bruce)
12004 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
12005 Changes for Digital Unix
12006 Portability fix for pg_dumpall(Bruce)
12007 Rename pg_attribute.attnvals to attdispersion(Bruce)
12008 "intro/unix" manual page now "pgintro"(Bruce)
12009 "built-in" manual page now "pgbuiltin"(Bruce)
12010 "drop" manual page now "drop_table"(Bruce)
12011 Add "create_trigger", "drop_trigger" manual pages(Thomas)
12012 Add constraints regression test(Vadim & Thomas)
12013 Add comments syntax regression test(Thomas)
12014 Add PGINDENT and support program(Bruce)
12015 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
12016 Files moved to /src/tools directory(Bruce)
12017 SPI and Trigger programming guides (Vadim & D'Arcy)
12023 <sect1 id="release-6-1-1">
12024 <title>Release 6.1.1</title>
12027 <title>Release date</title>
12028 <simpara>1997-07-22</simpara>
12032 <title>Migration from version 6.1 to version 6.1.1</title>
12035 This is a minor bug-fix release. A dump/reload is not required from version 6.1,
12036 but is required from any release prior to 6.1.
12037 Refer to the release notes for 6.1 for more details.
12042 <title>Changes</title>
12046 fix for SET with options (Thomas)
12047 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
12048 new psql \connect option allows changing usernames without changing databases
12049 fix for initdb --debug option(Yoshihiko Ichikawa))
12050 lextest cleanup(Bruce)
12052 fix date/time month boundary arithmetic(Thomas)
12053 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
12054 timestamp overhauled to use standard functions(Thomas)
12055 other code cleanup in date/time routines(Thomas)
12056 psql's \d now case-insensitive(Bruce)
12057 psql's backslash commands can now have trailing semicolon(Bruce)
12058 fix memory leak in psql when using \g(Bruce)
12059 major fix for endian handling of communication to server(Thomas, Tatsuo)
12060 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
12061 allow underscores in usernames(Bruce)
12062 pg_dumpall now returns proper status, portability fix(Bruce)
12068 <sect1 id="release-6-1">
12069 <title>Release 6.1</title>
12072 <title>Release date</title>
12073 <simpara>1997-06-08</simpara>
12077 The regression tests have been adapted and extensively modified for the
12078 6.1 release of <productname>PostgreSQL</productname>.
12082 Three new data types (<type>datetime</type>, <type>timespan</type>, and <type>circle</type>) have been added to
12083 the native set of <productname>PostgreSQL</productname> types. Points, boxes, paths, and polygons
12084 have had their output formats made consistent across the data types.
12085 The polygon output in misc.out has only been spot-checked for correctness
12086 relative to the original regression output.
12090 <productname>PostgreSQL</productname> 6.1 introduces a new, alternate
12091 optimizer which uses <firstterm>genetic</firstterm>
12092 algorithms. These algorithms introduce a random behavior in the ordering
12093 of query results when the query contains multiple qualifiers or multiple
12094 tables (giving the optimizer a choice on order of evaluation). Several
12095 regression tests have been modified to explicitly order the results, and
12096 hence are insensitive to optimizer choices. A few regression tests are
12097 for data types which are inherently unordered (e.g. points and time
12098 intervals) and tests involving those types are explicitly bracketed with
12099 <command>set geqo to 'off'</command> and <command>reset geqo</command>.
12103 The interpretation of array specifiers (the curly braces around atomic
12104 values) appears to have changed sometime after the original regression
12105 tests were generated. The current <filename>./expected/*.out</filename> files reflect this
12106 new interpretation, which may not be correct!
12110 The float8 regression test fails on at least some platforms. This is due
12111 to differences in implementations of <function>pow()</function> and <function>exp()</function> and the signaling
12112 mechanisms used for overflow and underflow conditions.
12116 The <quote>random</> results in the random test should cause the
12117 <quote>random</quote> test to be <quote>failed</quote>, since the
12118 regression tests are evaluated using a simple diff. However,
12119 <quote>random</> does not seem to produce random results on my test
12120 machine (Linux/<application>gcc</>/i686).
12124 <title>Migration to version 6.1</title>
12127 This migration requires a complete dump of the 6.0 database and a
12128 restore of the database in 6.1.
12131 Those migrating from earlier 1.* releases should first upgrade to 1.09
12132 because the COPY output format was improved from the 1.02 release.
12137 <title>Changes</title>
12143 packet length checking in library routines
12144 lock manager priority patch
12145 check for under/over flow of float8(Bruce)
12146 multitable join fix(Vadim)
12147 SIGPIPE crash fix(Darren)
12148 large object fixes(Sven)
12149 allow btree indexes to handle NULLs(Vadim)
12150 timezone fixes(D'Arcy)
12151 select SUM(x) can return NULL on no rows(Thomas)
12152 internal optimizer, executor bug fixes(Vadim)
12153 fix problem where inner loop in < or <= has no rows(Vadim)
12154 prevent re-commuting join index clauses(Vadim)
12155 fix join clauses for multiple tables(Vadim)
12156 fix hash, hashjoin for arrays(Vadim)
12157 fix btree for abstime type(Vadim)
12158 large object fixes(Raymond)
12159 fix buffer leak in hash indexes (Vadim)
12160 fix rtree for use in inner scan (Vadim)
12161 fix gist for use in inner scan, cleanups (Vadim, Andrea)
12162 avoid unnecessary local buffers allocation (Vadim, Massimo)
12163 fix local buffers leak in transaction aborts (Vadim)
12164 fix file manager memmory leaks, cleanups (Vadim, Massimo)
12165 fix storage manager memmory leaks (Vadim)
12166 fix btree duplicates handling (Vadim)
12167 fix deleted rows reincarnation caused by vacuum (Vadim)
12168 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
12169 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
12173 attribute optimization statistics(Bruce)
12174 much faster new btree bulk load code(Paul)
12175 BTREE UNIQUE added to bulk load code(Vadim)
12176 new lock debug code(Massimo)
12177 massive changes to libpg++(Leo)
12178 new GEQO optimizer speeds table multitable optimization(Martin)
12179 new WARN message for non-unique insert into unique key(Marc)
12180 update x=-3, no spaces, now valid(Bruce)
12181 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
12182 debug backend now pretty-prints tree(Darren)
12183 new Oracle character functions(Edmund)
12184 new plaintext password functions(Dan)
12185 no such class or insufficient privilege changed to distinct messages(Dan)
12186 new ANSI timestamp function(Dan)
12187 new ANSI Time and Date types (Thomas)
12188 move large chunks of data in backend(Martin)
12189 multicolumn btree indexes(Vadim)
12190 new SET var TO value command(Martin)
12191 update transaction status on reads(Dan)
12192 new locale settings for character types(Oleg)
12193 new SEQUENCE serial number generator(Vadim)
12194 GROUP BY function now possible(Vadim)
12195 re-organize regression test(Thomas,Marc)
12196 new optimizer operation weights(Vadim)
12197 new psql \z grant/permit option(Marc)
12198 new MONEY data type(D'Arcy,Thomas)
12199 tcp socket communication speed improved(Vadim)
12200 new VACUUM option for attribute statistics, and for certain columns (Vadim)
12201 many geometric type improvements(Thomas,Keith)
12202 additional regression tests(Thomas)
12203 new datestyle variable(Thomas,Vadim,Martin)
12204 more comparison operators for sorting types(Thomas)
12205 new conversion functions(Thomas)
12206 new more compact btree format(Vadim)
12207 allow pg_dumpall to preserve database ownership(Bruce)
12208 new SET GEQO=# and R_PLANS variable(Vadim)
12209 old (!GEQO) optimizer can use right-sided plans (Vadim)
12210 typechecking improvement in SQL parser(Bruce)
12211 new SET, SHOW, RESET commands(Thomas,Vadim)
12212 new \connect database USER option
12213 new destroydb -i option (Igor)
12214 new \dt and \di psql commands (Darren)
12215 SELECT "\n" now escapes newline (A. Duursma)
12216 new geometry conversion functions from old format (Thomas)
12218 Source tree changes
12219 -------------------
12220 new configuration script(Marc)
12221 readline configuration option added(Marc)
12222 OS-specific configuration options removed(Marc)
12223 new OS-specific template files(Marc)
12224 no more need to edit Makefile.global(Marc)
12225 re-arrange include files(Marc)
12226 nextstep patches (Gregor Hoffleit)
12227 removed Windows-specific code(Bruce)
12228 removed postmaster -e option, now only postgres -e option (Bruce)
12229 merge duplicate library code in front/backends(Martin)
12230 now works with eBones, international Kerberos(Jun)
12231 more shared library support
12232 c++ include file cleanup(Bruce)
12233 warn about buggy flex(Bruce)
12234 DG/UX, Ultrix, IRIX, AIX portability fixes
12240 <sect1 id="release-6-0">
12241 <title>Release 6.0</title>
12244 <title>Release date</title>
12245 <simpara>1997-01-29</simpara>
12249 A dump/restore is required for those wishing to migrate data from
12250 previous releases of <productname>PostgreSQL</productname>.
12254 <title>Migration from version 1.09 to version 6.0</title>
12257 This migration requires a complete dump of the 1.09 database and a
12258 restore of the database in 6.0.
12263 <title>Migration from pre-1.09 to version 6.0</title>
12266 Those migrating from earlier 1.* releases should first upgrade to 1.09
12267 because the COPY output format was improved from the 1.02 release.
12272 <title>Changes</title>
12278 ALTER TABLE bug - running postgress process needs to re-read table definition
12279 Allow vacuum to be run on one table or entire database(Bruce)
12281 Fix array over-runs of memory writes(Kurt)
12282 Fix elusive btree range/non-range bug(Dan)
12283 Fix for hash indexes on some types like time and date
12284 Fix for pg_log size explosion
12285 Fix permissions on lo_export()(Bruce)
12286 Fix unitialized reads of memory(Kurt)
12287 Fixed ALTER TABLE ... char(3) bug(Bruce)
12288 Fixed a few small memory leaks
12289 Fixed EXPLAIN handling of options and changed full_path option name
12290 Fixed output of group acl privileges
12291 Memory leaks (hunt and destroy with tools like Purify(Kurt)
12292 Minor improvements to rules system
12294 New asserts for run-checking
12295 Overhauled parser/analyze code to properly report errors and increase speed
12296 Pg_dump -d now handles NULL's properly(Bruce)
12297 Prevent SELECT NULL from crashing server (Bruce)
12298 Properly report errors when INSERT ... SELECT columns did not match
12299 Properly report errors when insert column names were not correct
12300 psql \g filename now works(Bruce)
12301 psql fixed problem with multiple statements on one line with multiple outputs
12302 Removed duplicate system OIDs
12303 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
12304 Several fixes for queries that crashed the backend
12305 Starting quote in insert string errors(Bruce)
12306 Submitting an empty query now returns empty status, not just " " query(Bruce)
12310 Add EXPLAIN manual page(Bruce)
12311 Add UNIQUE index capability(Dan)
12312 Add hostname/user level access control rather than just hostname and user
12313 Add synonym of != for <>(Bruce)
12314 Allow "select oid,* from table"
12315 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Bruce)
12316 Allow COPY from the frontend(Bryan)
12317 Allow GROUP BY to use alias column name(Bruce)
12318 Allow actual compression, not just reuse on the same page(Vadim)
12319 Allow installation-configuration option to auto-add all local users(Bryan)
12320 Allow libpq to distinguish between text value '' and null(Bruce)
12321 Allow non-postgres users with createdb privs to destroydb's
12322 Allow restriction on who can create C functions(Bryan)
12323 Allow restriction on who can do backend COPY(Bryan)
12324 Can shrink tables, pg_time and pg_log(Vadim & Erich)
12325 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
12326 Change default decimal constant representation from float4 to float8(Bruce)
12327 European date format now set when postmaster is started
12328 Execute lowercase function names if not found with exact case
12329 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z'
12330 Gist now included in the distrubution(Marc)
12331 Idend authentication of local users(Bryan)
12332 Implement BETWEEN qualifier(Bruce)
12333 Implement IN qualifier(Bruce)
12334 libpq has PQgetisnull()(Bruce)
12335 libpq++ improvements
12336 New options to initdb(Bryan)
12337 Pg_dump allow dump of OIDs(Bruce)
12338 Pg_dump create indexes after tables are loaded for speed(Bruce)
12339 Pg_dumpall dumps all databases, and the user table
12340 Pginterface additions for NULL values(Bruce)
12341 Prevent postmaster from being run as root
12342 psql \h and \? is now readable(Bruce)
12343 psql allow backslashed, semicolons anywhere on the line(Bruce)
12344 psql changed command prompt for lines in query or in quotes(Bruce)
12345 psql char(3) now displays as (bp)char in \d output(Bruce)
12346 psql return code now more accurate(Bryan?)
12347 psql updated help syntax(Bruce)
12348 Re-visit and fix vacuum(Vadim)
12349 Reduce size of regression diffs, remove timezone name difference(Bruce)
12350 Remove compile-time parameters to enable binary distributions(Bryan)
12351 Reverse meaning of HBA masks(Bryan)
12352 Secure Authentication of local users(Bryan)
12353 Speed up vacuum(Vadim)
12354 Vacuum now had VERBOSE option(Bruce)
12356 Source tree changes
12357 -------------------
12358 All functions now have prototypes that are compared against the calls
12359 Allow asserts to be disabled easly from Makefile.global(Bruce)
12360 Change oid constants used in code to #define names
12361 Decoupled sparc and solaris defines(Kurt)
12362 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
12363 Major include file reorganization/reduction(Marc)
12364 Make now stops on compile failure(Bryan)
12365 Makefile restructuring(Bryan, Marc)
12366 Merge bsdi_2_1 to bsdi(Bruce)
12367 Monitor program removed
12368 Name change from Postgres95 to PostgreSQL
12369 New config.h file(Marc, Bryan)
12370 PG_VERSION now set to 6.0 and used by postmaster
12371 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
12372 Reduced the number of #define's, centeralized #define's
12373 Remove duplicate OIDS in system tables(Dan)
12374 Remove duplicate system catalog info or report mismatches(Dan)
12375 Removed many os-specific #define's
12376 Restructured object file generation/location(Bryan, Marc)
12377 Restructured port-specific file locations(Bryan, Marc)
12378 Unused/uninialized variables corrected
12384 <sect1 id="release-1-09">
12385 <title>Release 1.09</title>
12388 <title>Release date</title>
12389 <simpara>1996-11-04</simpara>
12393 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of
12394 the changes listed in 6.0 were actually included in the 1.02.1 to 1.09
12399 <sect1 id="release-1-02">
12400 <title>Release 1.02</title>
12403 <title>Release date</title>
12404 <simpara>1996-08-01</simpara>
12408 <title>Migration from version 1.02 to version 1.02.1</title>
12411 Here is a new migration file for 1.02.1. It includes the 'copy' change
12412 and a script to convert old <acronym>ASCII</acronym> files.
12416 The following notes are for the benefit of users who want to migrate
12417 databases from <productname>Postgres95</> 1.01 and 1.02 to <productname>Postgres95</> 1.02.1.
12420 If you are starting afresh with <productname>Postgres95</> 1.02.1 and do not need
12421 to migrate old databases, you do not need to read any further.
12426 In order to upgrade older <productname>Postgres95</> version 1.01 or 1.02 databases to
12427 version 1.02.1, the following steps are required:
12432 Start up a new 1.02.1 postmaster
12437 Add the new built-in functions and operators of 1.02.1 to 1.01 or 1.02
12438 databases. This is done by running the new 1.02.1 server against
12439 your own 1.01 or 1.02 database and applying the queries attached at
12440 the end of the file. This can be done easily through <command>psql</>. If your
12441 1.01 or 1.02 database is named <literal>testdb</literal> and you have cut the commands
12442 from the end of this file and saved them in <filename>addfunc.sql</filename>:
12444 % psql testdb -f addfunc.sql
12447 Those upgrading 1.02 databases will get a warning when executing the
12448 last two statements in the file because they are already present in 1.02. This is
12449 not a cause for concern.
12456 <title>Dump/Reload Procedure</title>
12459 If you are trying to reload a pg_dump or text-mode, <literal>copy tablename to
12460 stdout</literal> generated with a previous version, you will need to run the
12461 attached <command>sed</command> script on the ASCII file before loading it into the
12462 database. The old format used '.' as end-of-data, while '\.' is now the
12463 end-of-data marker. Also, empty strings are now loaded in as '' rather
12464 than NULL. See the copy manual page for full details.
12467 sed 's/^\.$/\\./g' <in_file >out_file
12471 If you are loading an older binary copy or non-<systemitem>stdout</> copy, there is no
12472 end-of-data character, and hence no conversion necessary.
12475 -- following lines added by agc to reflect the case-insensitive
12476 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
12477 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexeq);
12478 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregexne);
12479 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregexeq);
12480 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticregexne);
12486 <title>Changes</title>
12490 Source code maintenance and development
12491 * worldwide team of volunteers
12492 * the source tree now in CVS at ftp.ki.net
12495 * psql (and underlying libpq library) now has many more options for
12496 formatting output, including HTML
12497 * pg_dump now output the schema and/or the data, with many fixes to
12498 enhance completeness.
12499 * psql used in place of monitor in administration shell scripts.
12500 monitor to be deprecated in next release.
12501 * date/time functions enhanced
12502 * NULL insert/update/comparison fixed/enhanced
12503 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
12505 Bug Fixes (almost too numerous to mention)
12507 * storage management
12508 * check for NULL pointer before dereferencing
12512 * added SolarisX86 port
12513 * added BSD/OS 2.1 port
12518 Contributors (appologies to any missed)
12519 * Kurt J. Lidl <lidl@va.pubnix.com>
12520 (missed in first run, but no less important)
12521 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12522 * Jason Wright <jason@shiloh.vnet.net>
12523 * Cees de Groot <C.deGroot@inter.NL.net>
12524 * ernst.molitor@uni-bonn.de
12525 * michael.siebenborn@ae3.Hypo.DE (Michael Siebenborn (6929))
12526 * Brian E. Gallew <geek+@cmu.edu>
12527 * Vadim B. Mikheev <vadim@sable.krasnoyarsk.su>
12528 * Adam Sussman <myddryn@vidya.com>
12529 * Chris Dunlop <chris@onthe.net.au>
12530 * Marc G. Fournier <scrappy@ki.net>
12531 * Dan McGuirk <mcguirk@indirect.com>
12532 * Dr_George_D_Detlefsen <drgeorge@ilt.com>
12533 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12534 * Massimo Dal Zotto <dz@cs.unitn.it>
12535 * Randy Kunkee <kunkee@Starbase.NeoSoft.COM>
12536 * Rick Weldon <rick@wisetech.com>
12537 * Thomas van Reimersdahl <reimersd@dali.techinfo.rwth-aachen.de>
12538 * david bennett <dave@bensoft.com>
12539 * ernst.molitor@uni-bonn.de
12540 * Julian Assange <proff@suburbia.net>
12541 * Bruce Momjian <pgman@candle.pha.pa.us>
12542 * Paul "Shag" Walmsley <ccshag@cclabs.missouri.edu>
12543 * "Alistair G. Crooks" <azcb0@sde.uts.amdahl.com>
12548 <sect1 id="release-1-01">
12549 <title>Release 1.01</title>
12552 <title>Release date</title>
12553 <simpara>1996-02-23</simpara>
12558 <title>Migration from version 1.0 to version 1.01</title>
12561 The following notes are for the benefit of users who want to migrate
12562 databases from <productname>Postgres95</> 1.0 to <productname>Postgres95</> 1.01.
12565 If you are starting afresh with <productname>Postgres95</> 1.01 and do not need
12566 to migrate old databases, you do not need to read any further.
12569 In order to <productname>Postgres95</> version 1.01 with databases created with
12570 <productname>Postgres95</> version 1.0, the following steps are required:
12575 Set the definition of <symbol>NAMEDATALEN</symbol> in <filename>src/Makefile.global</filename> to 16
12576 and <symbol>OIDNAMELEN</symbol> to 20.
12581 Decide whether you want to use Host based authentication.
12586 If you do, you must create a file name <literal>pg_hba</literal> in your top-level data
12587 directory (typically the value of your <envar>$PGDATA</envar>). <filename>src/libpq/pg_hba</filename>
12588 shows an example syntax.
12593 If you do not want host-based authentication, you can comment out
12598 in <filename>src/Makefile.global</filename>
12601 Note that host-based authentication is turned on by default, and if
12602 you do not take steps A or B above, the out-of-the-box 1.01 will
12603 not allow you to connect to 1.0 databases.
12611 Compile and install 1.01, but DO NOT do the <command>initdb</command> step.
12616 Before doing anything else, terminate your 1.0 postmaster, and
12617 backup your existing <envar>$PGDATA</envar> directory.
12622 Set your <envar>PGDATA</envar> environment variable to your 1.0 databases, but set up
12623 path up so that 1.01 binaries are being used.
12628 Modify the file <filename><envar>$PGDATA</envar>/PG_VERSION</filename> from 5.0 to 5.1
12633 Start up a new 1.01 postmaster
12638 Add the new built-in functions and operators of 1.01 to 1.0
12639 databases. This is done by running the new 1.01 server against
12640 your own 1.0 database and applying the queries attached and saving
12641 in the file 1.0_to_1.01.sql. This can be done easily through <command>psql</command>.
12642 If your 1.0 database is name <literal>testdb</literal>:
12645 % psql testdb -f 1.0_to_1.01.sql
12648 and then execute the following commands (cut and paste from here):
12651 -- add builtin functions that are new to 1.01
12653 create function int4eqoid (int4, oid) returns bool as 'foo'
12654 language 'internal';
12655 create function oideqint4 (oid, int4) returns bool as 'foo'
12656 language 'internal';
12657 create function char2icregexeq (char2, text) returns bool as 'foo'
12658 language 'internal';
12659 create function char2icregexne (char2, text) returns bool as 'foo'
12660 language 'internal';
12661 create function char4icregexeq (char4, text) returns bool as 'foo'
12662 language 'internal';
12663 create function char4icregexne (char4, text) returns bool as 'foo'
12664 language 'internal';
12665 create function char8icregexeq (char8, text) returns bool as 'foo'
12666 language 'internal';
12667 create function char8icregexne (char8, text) returns bool as 'foo'
12668 language 'internal';
12669 create function char16icregexeq (char16, text) returns bool as 'foo'
12670 language 'internal';
12671 create function char16icregexne (char16, text) returns bool as 'foo'
12672 language 'internal';
12673 create function texticregexeq (text, text) returns bool as 'foo'
12674 language 'internal';
12675 create function texticregexne (text, text) returns bool as 'foo'
12676 language 'internal';
12678 -- add builtin functions that are new to 1.01
12680 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
12681 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
12682 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexeq);
12683 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregexne);
12684 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexeq);
12685 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregexne);
12686 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexeq);
12687 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregexne);
12688 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icregexeq);
12689 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icregexne);
12690 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq);
12691 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne);
12699 <title>Changes</title>
12704 * 1.01 is backwards compatible with 1.0 database provided the user
12705 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
12706 If those steps are not taken, 1.01 is not compatible with 1.0 database.
12709 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
12710 * added NeXT port (requires SysVIPC implementation)
12711 * added CAST .. AS ... syntax
12712 * added ASC and DESC key words
12713 * added 'internal' as a possible language for CREATE FUNCTION
12714 internal functions are C functions which have been statically linked
12715 into the postgres backend.
12716 * a new type "name" has been added for system identifiers (table names,
12717 attribute names, etc.) This replaces the old char16 type. The
12718 of name is set by the NAMEDATALEN #define in src/Makefile.global
12719 * a readable reference manual that describes the query language.
12720 * added host-based access control. A configuration file ($PGDATA/pg_hba)
12721 is used to hold the configuration data. If host-based access control
12722 is not desired, comment out HBA=1 in src/Makefile.global.
12723 * changed regex handling to be uniform use of Henry Spencer's regex code
12724 regardless of platform. The regex code is included in the distribution
12725 * added functions and operators for case-insensitive regular expressions.
12726 The operators are ~* and !~*.
12727 * pg_dump uses COPY instead of SELECT loop for better performance
12730 * fixed an optimizer bug that was causing core dumps when
12731 functions calls were used in comparisons in the WHERE clause
12732 * changed all uses of getuid to geteuid so that effective uids are used
12733 * psql now returns non-zero status on errors when using -c
12734 * applied public patches 1-14
12740 <sect1 id="release-1-0">
12741 <title>Release 1.0</title>
12744 <title>Release date</title>
12745 <simpara>1995-09-05</simpara>
12749 <title>Changes</title>
12754 * The copyright of <productname>Postgres</productname> 1.0 has been loosened to be freely modifiable
12755 and modifiable for any purpose. Please read the COPYRIGHT file.
12756 Thanks to Professor Michael Stonebraker for making this possible.
12759 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
12760 EUROPEAN STYLE). This follows SQL-92 specs.
12761 * "delimiters" is now a key word
12764 * sql LIKE syntax has been added
12765 * copy command now takes an optional USING DELIMITER specification.
12766 delimiters can be any single-character string.
12767 * IRIX 5.3 port has been added.
12768 Thanks to Paul Walmsley and others.
12769 * updated pg_dump to work with new libpq
12770 * \d has been added psql
12771 Thanks to Keith Parks
12772 * regexp performance for architectures that use POSIX regex has been
12773 improved due to caching of precompiled patterns.
12774 Thanks to Alistair Crooks
12775 * a new version of libpq++
12776 Thanks to William Wanders
12779 * arbitrary userids can be specified in the createuser script
12780 * \c to connect to other databases in psql now works.
12781 * bad pg_proc entry for float4inc() is fixed
12782 * users with usecreatedb field set can now create databases without
12783 having to be usesuper
12784 * remove access control entries when the entry no longer has any
12786 * fixed non-portable datetimes implementation
12787 * added kerberos flags to the src/backend/Makefile
12788 * libpq now works with kerberos
12789 * typographic errors in the user manual have been corrected.
12790 * btrees with multiple index never worked, now we tell you they don't
12791 work when you try to use them
12797 <sect1 id="release-0-03">
12798 <title><productname>Postgres95</productname> Release 0.03</title>
12801 <title>Release date</title>
12802 <simpara>1995-07-21</simpara>
12806 <title>Changes</title>
12809 Incompatible changes:
12810 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
12811 (due to system catalog changes and indexing structure changes).
12812 * double-quote (") is deprecated as a quoting character for string literals;
12813 you need to convert them to single quotes ('). <!-- " -->
12814 * name of aggregates (eg. int4sum) are renamed in accordance with the
12815 SQL standard (eg. sum).
12816 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
12817 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
12818 previous releases); you might have to do typecasting if you depend on it
12819 being of type float8. If you neglect to do the typecasting and you assign
12820 a float literal to a field of type float8, you may get incorrect values
12822 * LIBPQ has been totally revamped so that frontend applications
12823 can connect to multiple backends
12824 * the usesysid field in pg_user has been changed from int2 to int4 to
12825 allow wider range of Unix user ids.
12826 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
12827 single BSD44_derived port. (thanks to Alistair Crooks)
12829 SQL standard-compliance (the following details changes that makes postgres95
12830 more compliant to the SQL-92 standard):
12831 * the following SQL types are now built-in: smallint, int(eger), float, real,
12832 char(N), varchar(N), date and time.
12834 The following are aliases to existing postgres types:
12835 smallint -> int2
12836 integer, int -> int4
12837 float, real -> float4
12838 char(N) and varchar(N) are implemented as truncated text types. In
12839 addition, char(N) does blank-padding.
12840 * single-quote (') is used for quoting string literals; '' (in addition to
12841 \') is supported as means of inserting a single quote in a string
12842 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
12843 (Also, aggregates can now be overloaded, i.e. you can define your
12844 own MAX aggregate to take in a user-defined type.)
12845 * CHANGE ACL removed. GRANT/REVOKE syntax added.
12846 - Privileges can be given to a group using the "GROUP" key word.
12848 GRANT SELECT ON foobar TO GROUP my_group;
12849 The key word 'PUBLIC' is also supported to mean all users.
12851 Privileges can only be granted or revoked to one user or group
12854 "WITH GRANT OPTION" is not supported. Only class owners can change
12856 - The default access control is to to grant users readonly access.
12857 You must explicitly grant insert/update access to users. To change
12858 this, modify the line in
12859 src/backend/utils/acl.h
12860 that defines ACL_WORLD_DEFAULT
12863 * the bug where aggregates of empty tables were not run has been fixed. Now,
12864 aggregates run on empty tables will return the initial conditions of the
12865 aggregates. Thus, COUNT of an empty table will now properly return 0.
12866 MAX/MIN of an empty table will return a row of value NULL.
12867 * allow the use of \; inside the monitor
12868 * the LISTEN/NOTIFY asynchronous notification mechanism now work
12869 * NOTIFY in rule action bodies now work
12870 * hash indexes work, and access methods in general should perform better.
12871 creation of large btree indexes should be much faster. (thanks to Paul
12874 Other changes and enhancements:
12875 * addition of an EXPLAIN statement used for explaining the query execution
12876 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
12878 * WARN and NOTICE messages no longer have timestamps on them. To turn on
12879 timestamps of error messages, uncomment the line in
12880 src/backend/utils/elog.h:
12881 /* define ELOG_TIMESTAMPS */
12882 * On an access control violation, the message
12883 "Either no such class or insufficient privilege"
12884 will be given. This is the same message that is returned when
12885 a class is not found. This dissuades non-privileged users from
12886 guessing the existence of privileged classes.
12887 * some additional system catalog changes have been made that are not
12888 visible to the user.
12891 * The -oid option has been added to the "pg_result" tcl command.
12892 pg_result -oid returns oid of the last row inserted. If the
12893 last command was not an INSERT, then pg_result -oid returns "".
12894 * the large object interface is available as pg_lo* tcl commands:
12895 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
12897 Portability enhancements and New Ports:
12898 * flex/lex problems have been cleared up. Now, you should be able to use
12899 flex instead of lex on any platforms. We no longer make assumptions of
12900 what lexer you use based on the platform you use.
12901 * The Linux-ELF port is now supported. Various configuration have been
12902 tested: The following configuration is known to work:
12903 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
12904 with everything in ELF format,
12907 * ipcclean added to the distribution
12908 ipcclean usually does not need to be run, but if your backend crashes
12909 and leaves shared memory segments hanging around, ipcclean will
12910 clean them up for you.
12913 * the user manual has been revised and libpq documentation added.
12919 <sect1 id="release-0-02">
12920 <title><productname>Postgres95</productname> Release 0.02</title>
12923 <title>Release date</title>
12924 <simpara>1995-05-25</simpara>
12928 <title>Changes</title>
12932 Incompatible changes:
12933 * The SQL statement for creating a database is 'CREATE DATABASE' instead
12934 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
12935 of 'DESTROYDB'. However, the names of the executables 'createdb' and
12936 'destroydb' remain the same.
12939 * pgperl - a Perl (4.036) interface to Postgres95
12940 * pg_dump - a utility for dumping out a postgres database into a
12941 script file containing query commands. The script files are in a ASCII
12942 format and can be used to reconstruct the database, even on other
12943 machines and other architectures. (Also good for converting
12944 a Postgres 4.2 database to Postgres95 database.)
12946 The following ports have been incorporated into postgres95-beta-0.02:
12947 * the NetBSD port by Alistair Crooks
12948 * the AIX port by Mike Tung
12949 * the Windows NT port by Jon Forrest (more stuff but not done yet)
12950 * the Linux ELF port by Brian Gallew
12952 The following bugs have been fixed in postgres95-beta-0.02:
12953 * new lines not escaped in COPY OUT and problem with COPY OUT when first
12955 * cannot type return to use the default user id in createuser
12956 * SELECT DISTINCT on big tables crashes
12957 * Linux installation problems
12958 * monitor doesn't allow use of 'localhost' as PGHOST
12959 * psql core dumps when doing \c or \l
12960 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
12961 * libpgtcl has a hard-wired default port number
12962 * SELECT DISTINCT INTO TABLE hangs
12963 * CREATE TYPE doesn't accept 'variable' as the internallength
12964 * wrong result using more than 1 aggregate in a SELECT
12970 <sect1 id="release-0-01">
12971 <title><productname>Postgres95</productname> Release 0.01</title>
12974 <title>Release date</title>
12975 <simpara>1995-05-01</simpara>
12984 <sect1 id="timing-results">
12985 <title>Timing Results</title>
12988 These timing results are from running the regression test with the commands
12991 % cd src/test/regress
12993 % time make runtest
12997 Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
12998 to run, presumably due to the scheduling vagaries of multitasking systems.
13002 <title>Version 6.5</title>
13005 As has been the case for previous releases, timing between
13006 releases is not directly comparable since new regression tests
13007 have been added. In general, 6.5 is faster than previous
13012 Timing with <function>fsync()</function> disabled:
13016 02:00 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
13017 04:38 Sparc Ultra 1 143MHz, 64MB, Solaris 2.6
13022 Timing with <function>fsync()</function> enabled:
13026 04:21 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
13029 For the <systemitem class="osname">Linux</systemitem> system above, using <acronym>UW-SCSI</acronym> disks rather than (older) <acronym>IDE</acronym>
13030 disks leads to a 50% improvement in speed on the regression test.
13035 <title>Version 6.4beta</title>
13038 The times for this release are not directly comparable to those for previous releases
13039 since some additional regression tests have been included.
13040 In general, however, 6.4 should be slightly faster than the previous release (thanks, Bruce!).
13045 02:26 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13051 <title>Version 6.3</title>
13054 The times for this release are not directly comparable to those for previous releases
13055 since some additional regression tests have been included and some obsolete tests involving
13056 time travel have been removed.
13057 In general, however, 6.3 is substantially faster than previous releases (thanks, Bruce!).
13062 02:30 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13063 04:12 Dual Pentium Pro 180, 96MB, EIDE, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13069 <title>Version 6.1</title>
13074 06:12 Pentium Pro 180, 32MB, EIDE, Linux 2.0.30, gcc 2.7.2 -O2 -m486
13075 12:06 P-100, 48MB, Linux 2.0.29, gcc
13076 39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
13084 <!-- Keep this comment at the end of the file
13089 sgml-minimize-attributes:nil
13090 sgml-always-quote-attributes:t
13093 sgml-parent-document:nil
13094 sgml-default-dtd-file:"./reference.ced"
13095 sgml-exposed-tags:nil
13096 sgml-local-catalogs:("/usr/lib/sgml/catalog")
13097 sgml-local-ecat-files:nil