1 TODO list for PostgreSQL
2 ========================
3 Last updated: Tue Dec 24 17:06:43 EST 2002
5 Current maintainer: Bruce Momjian (pgman@candle.pha.pa.us)
7 The most recent version of this document can be viewed at
8 the PostgreSQL web site, http://www.PostgreSQL.org.
10 A dash (-) marks changes that will appear in the upcoming 7.4 release.
12 Bracketed items "[]" have more detailed.
18 * Add replication of distributed databases [replication]
21 o master/slave replication
22 o multi-master replication
23 o partition data across servers
24 o sample implementation in contrib/rserv
25 o queries across databases or servers (two-phase commit)
26 o allow replication over unreliable or non-persistent links
27 o http://gborg.postgresql.org/project/pgreplication/projdisplay.php
28 * Point-in-time data recovery using backup and write-ahead log
29 * Create native Win32 port [win32]
35 * Allow elog() to return error codes, module name, file name, line
36 number, not just messages (Peter E)
37 * Add error codes (Peter E)
38 * Make error messages more consistent [error]
39 * Show location of syntax error in query [yacc]
46 * Remove unreferenced table files and temp tables during database vacuum
47 or postmaster startup (Bruce)
48 * Remove behavior of postmaster -o after making postmaster/postgres
50 * Allow easy display of usernames in a group
51 * Allow configuration files to be specified in a different directory
52 * Add start time to pg_stat_activity
53 * Allow limits on per-db/user connections
54 * Have standalone backend read postgresql.conf
60 * Add IPv6 capability to INET/CIDR types
61 * Remove Money type, add money formatting for decimal type
62 * Change factorial to return a numeric
63 * Change NUMERIC data type to use base 10,000 internally
64 * Change NUMERIC to enforce the maximum precision, and increase it
65 * Add function to return compressed length of TOAST data values (Tom)
66 * Allow INET subnet tests using non-constants
67 * Add now("transaction|statement|clock") functionality
68 * -Add GUC variables to control floating number output digits (Pedro Ferreira)
69 * Have sequence dependency track use of DEFAULT sequences, seqname.nextval
72 o Allow better handling of numeric constants, type conversion
76 o Allow nulls in arrays
77 o Allow arrays to be ORDER'ed
78 o Support construction of array result values in expressions
81 o Improve vacuum of large objects, like /contrib/vacuumlo
82 o Add security checking for large objects
83 o Make file in/out interface for TOAST columns, similar to large object
84 interface (force out-of-line storage and no compression)
85 o Auto-delete large objects when referencing row is deleted
88 Multi-Language Support
89 ======================
91 * Add NCHAR (as distinguished from ordinary varchar),
92 * Allow LOCALE on a per-column basis, default to ASCII
93 * Support multiple simultaneous character sets, per SQL92
94 * Improve Unicode combined character handling
95 * Optimize locale to have minimal performance impact when not used (Peter E)
96 * Add octet_length_server() and octet_length_client() (Thomas, Tatsuo)
97 * Make octet_length_client the same as octet_length() (?)
98 * Prevent mismatch of frontend/backend encodings from converting bytea
99 data from being interpreted as encoded strings
100 * Remove Cyrillic recode support
106 * Automatically create rules on views so they are updateable, per SQL92 [view]
107 * Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
108 * Allow NOTIFY in rules involving conditionals
109 * Have views on temporary tables exist in the temporary namespace
110 * Move psql backslash information into views
111 * Allow RULE recompilation
117 * Allow CREATE INDEX zman_index ON test (date_trunc( 'day', zman ) datetime_ops)
118 fails index can't store constant parameters
119 * Order duplicate index entries by tid for faster heap lookups
120 * Allow inherited tables to inherit index, UNIQUE constraint, and primary
121 key, foreign key [inheritance]
122 * UNIQUE INDEX on base column not honored on inserts from inherited table
123 INSERT INTO inherit_table (unique_index_col) VALUES (dup) should fail
125 * Add UNIQUE capability to non-btree indexes
126 * Add btree index support for reltime, tinterval, regproc
127 * Add rtree index support for line, lseg, path, point
128 * Certain indexes will not shrink, e.g. indexes on ever-increasing
129 columns and indexes with many duplicate keys
130 * Use indexes for min() and max() or convert to SELECT col FROM tab ORDER
131 BY col DESC LIMIT 1 if appropriate index exists and WHERE clause acceptible
132 * Allow LIKE indexing optimization for non-ASCII locales
133 * Use index to restrict rows returned by multi-key index when used with
134 non-consecutive keys or OR clauses, so fewer heap accesses
135 * Be smarter about insertion of already-ordered data into btree index
136 * Prevent index uniqueness checks when UPDATE does not modifying column
137 * Use bitmaps to fetch heap pages in sequential order [performance]
138 * Use bitmaps to combine existing indexes [performance]
139 * Improve handling of index scans for NULL
140 * Allow SELECT * FROM tab WHERE int2col = 4 to use int2col index, int8,
141 float4, numeric/decimal too [optimizer]
142 * Add FILLFACTOR to btree index creation
143 * Add concurrency to GIST
144 * Improve concurrency of hash indexes (Neil)
150 * Add BETWEEN ASYMMETRIC/SYMMETRIC (Christopher)
151 * Allow LIMIT/OFFSET to use expressions
152 * CREATE TABLE AS can not determine column lengths from expressions [atttypmod]
153 * Allow UPDATE to handle complex aggregates [update]
154 * Allow command blocks to ignore certain types of errors
155 * Allow backslash handling in quoted strings to be disabled for portability
156 * Return proper effected tuple count from complex commands [return]
157 * Allow DELETE to handle table aliases for self-joins [delete]
158 * Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
159 * Allow REINDEX to rebuild all indexes, remove /contrib/reindex
160 * Make a transaction-safe TRUNCATE
161 * Fix SELECT ... INTO and CREATE TABLE AS to have appopriate OID column
162 * Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
166 o ALTER TABLE ADD COLUMN does not honor DEFAULT and non-CHECK CONSTRAINT
167 o ALTER TABLE ADD COLUMN column DEFAULT should fill existing
168 rows with DEFAULT value
169 o ALTER TABLE ADD COLUMN column SERIAL doesn't create sequence because
171 o Add ALTER TABLE tab SET WITHOUT OIDS
172 * Add ALTER SEQUENCE to modify min/max/increment/cache/cycle values
175 o Automatically maintain clustering on a table
176 o Allow CLUSTER to cluster all tables, remove clusterdb
179 o Allow dump/load of CSV format
180 o Allow COPY to report error lines and continue; optionally
181 allow error codes to be specified; requires savepoints or can
182 not be run in a multi-statement transaction
183 o Allow copy to understand \x as hex
186 o Allow BINARY option to SELECT, just like DECLARE
187 o -MOVE 0 should not move to end of cursor (Bruce)
188 o Allow UPDATE/DELETE WHERE CURRENT OF cursor using per-cursor tid
189 stored in the backend
190 o Prevent DROP of table being referenced by our own open cursor
191 o Allow cursors outside transactions [cursor]
194 o Allow INSERT/UPDATE of system-generated oid value for a row
195 o Allow INSERT INTO tab (col1, ..) VALUES (val1, ..), (val2, ..)
196 o Allow INSERT/UPDATE ... RETURNING new.col or old.col; handle
200 o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
203 o Allow EXPLAIN EXECUTE to see prepared plans
204 o Allow SHOW of non-modifiable variables, like pg_controldata
205 o Add GUC parameter to control the maximum number of rewrite cycles
207 * SERVER-SIDE LANGUAGES
208 o Allow PL/PgSQL's RAISE function to take expressions
209 o Change PL/PgSQL to use palloc() instead of malloc()
210 o Add untrusted version of plpython
211 o Allow Java server-side programming, http://pljava.sourceforge.net
213 o Fix problems with complex temporary table creation/destruction
214 without using PL/PgSQL EXECUTE, needs cache prevention/invalidation
215 o Fix PL/pgSQL RENAME to work on variables other than OLD/NEW
216 o Improve PL/PgSQL exception handling
217 o Allow parameters to be specified by name and type during
219 o Allow function parameters to be passed by name,
220 get_employee_salary(emp_id => 12345, tax_year => 2001)
221 o Add PL/PgSQL packages
222 o Allow array declarations and other data types in PL/PgSQL DECLARE
223 o Add PL/PgSQL PROCEDURES that can return multiple values
224 o Add table function support to pltcl, plperl, plpython
225 o Make PL/PgSQL %TYPE schema-aware
231 * Allow psql to show transaction status if backend protocol changes made
232 * Add XML interface: psql, pg_dump, COPY, separate server (?)
233 * -Add schema, cast, and conversion backslash commands to psql (Christopher)
234 * Allow pg_dump to dump a specific schema
235 * Allow psql to do table completion for SELECT * FROM schema_part and
236 table completion for SELECT * FROM schema_name.
239 o Comprehensive test suite. This may be available already.
240 o JDBC-standard BLOB support
241 o Error Codes (pending backend implementation)
242 o Support both 'make' and 'ant'
243 o Fix LargeObject API to handle OIDs as unsigned ints
244 o Use cursors implicitly to avoid large results (see setCursorName())
245 o Add LISTEN/NOTIFY support to the JDBC driver (Barry)
248 o Implement set descriptor, using descriptor
249 o Make casts work in variable initializations
251 o Allow multi-threaded use of SQLCA
252 o Solve cardinality > 1 for input descriptors / variables
253 o Understand structure definitions outside a declare section
254 o sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified
255 o Improve error handling
256 o Allow :var[:index] or :var[<integer>] as cvariable for an array var
257 o Add a semantic check level, e.g. check if a table really exists
258 o Fix nested C comments
260 o fix handling of DB attributes that are arrays
263 o Allow users to register their own types with _pg
264 o Allow SELECT to return a dictionary of dictionaries
265 o Allow COPY BINARY FROM
268 Referential Integrity
269 =====================
271 * Add MATCH PARTIAL referential integrity [foreign]
272 * Add deferred trigger queue file (Jan)
273 * Implement dirty reads and use them in RI triggers
274 * Enforce referential integrity for system tables
275 * Change foreign key constraint for array -> element to mean element
277 * Allow DEFERRABLE UNIQUE constraints
278 * Allow triggers to be disabled [trigger]
279 * -Support statement-level triggers (Neil)
280 * Support triggers on columns (Neil)
286 * Flush cached query plans when their underlying catalog data changes
287 * Use dependency information to dump data in proper order
293 * Overhaul bufmgr/lockmgr/transaction manager
294 * Allow savepoints / nested transactions [transactions] (Bruce)
300 * Add SQL99 WITH clause to SELECT (Tom, Fernando)
301 * Add SQL99 WITH RECURSIVE to SELECT (Tom, Fernando)
302 * Add the concept of dataspaces/tablespaces [tablespaces]
303 * Allow queries across multiple databases [crossdb]
304 * Add pre-parsing phase that converts non-ANSI features to supported features
305 * Allow plug-in modules to emulate features from other databases
306 * SQL*Net listener that makes PostgreSQL appear as an Oracle database
317 * Delay fsync() when other backends are about to commit too [fsync]
318 o Determine optimal commit_delay value
319 * Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
320 o Allow multiple blocks to be written to WAL with one write()
325 * Shared catalog cache, reduce lseek()'s by caching table size in shared area
326 * Add free-behind capability for large sequential scans (Bruce)
327 * Allow binding query args over FE/BE protocol
328 * Consider use of open/fcntl(O_DIRECT) to minimize OS caching
329 * Make blind writes go through the file descriptor cache
335 * Improve speed with indexes (perhaps recreate index instead) [vacuum]
336 * Reduce lock time by moving tuples with read lock, then write
337 lock and truncate table [vacuum]
338 * Provide automatic running of vacuum in the background (Tom) [vacuum]
339 * Allow free space map to be auto-sized or warn when it is too small
345 * Make locking of shared data structures more fine-grained
346 * Add code to detect an SMP machine and handle spinlocks accordingly
347 from distributted.net, http://www1.distributed.net/source,
348 in client/common/cpucheck.cpp
349 * Research use of sched_yield() for spinlock acquisition failure
355 * Experiment with multi-threaded backend [thread]
356 * Add connection pooling [pool]
357 * Allow persistent backends [persistent]
358 * Create a transaction processor to aid in persistent connections and
360 * Do listen() in postmaster and accept() in pre-forked backend
361 * Have pre-forked backend pre-connect to last requested database or pass
362 file descriptor to backend pre-forked for matching database
368 * Have after-change WAL write()'s write only modified data to kernel
369 * Reduce number of after-change WAL writes; they exist only to gaurd against
370 partial page writes [wal]
371 * Turn off after-change writes if fsync is disabled (?)
372 * Add WAL index reliability improvement to non-btree indexes
373 * Find proper defaults for postgresql.conf WAL entries
374 * Add checkpoint_min_warning postgresql.conf option to warn about checkpoints
375 that are too frequent
376 * Allow xlog directory location to be specified during initdb, perhaps
378 * Allow pg_xlog to be moved without symlinks
384 * Improve Subplan list handling
385 * Allow Subplans to use efficient joins(hash, merge) with upper variable
386 * -Add hash for evaluating GROUP BY aggregates (Tom)
387 * Allow merge and hash joins on expressions not just simple variables (Tom)
388 * Make IN/NOT IN have similar performance to EXISTS/NOT EXISTS [exists]
389 * Missing optimizer selectivities for date, r-tree, etc. [optimizer]
390 * Allow ORDER BY ... LIMIT to select top values without sort or index
391 using a sequential scan for highest/lowest values (Oleg)
392 * -Inline simple SQL functions to avoid overhead (Tom)
393 * Precompile SQL functions to avoid overhead (Neil)
394 * Add utility to compute accurate random_page_cost value
395 * Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
396 * Use CHECK constraints to improve optimizer decisions
397 * Check GUC geqo_threshold to see if it is still accurate
398 * Allow sorting, temp files, temp tables to use multiple work directories
404 * Do async I/O for faster random read-ahead of data
405 * Get faster regex() code from Henry Spencer <henry@zoo.utoronto.ca>
407 * Use mmap() rather than SYSV shared memory or to write WAL files (?) [mmap]
408 * Improve caching of attribute offsets when NULLs exist in the row
410 * Wire Protocol Changes
411 o Show transaction status in psql
412 o Allow binding of query parameters, support for prepared queries
413 o Add optional textual message to NOTIFY
414 o Remove hard-coded limits on user/db/password names
415 o Remove unused elements of startup packet (unused, tty, passlength)
416 o Fix COPY/fastpath protocol?
417 o Replication support?
419 o Dynamic character set handling
420 o Special passing of binary values in platform-neutral format (bytea?)
422 o Add decoded type, length, precision
429 * Add use of 'const' for variables in source tree
430 * Rename some /contrib modules from pg* to pg_*
431 * Move some things from /contrib into main tree
432 * Remove warnings created by -Wcast-align
433 * Move platform-specific ps status display info from ps_status.c to ports
434 * Modify regression tests to prevent failures do to minor numeric rounding
435 * -Add OpenBSD's getpeereid() call for local socket authentication
436 * Improve access-permissions check on data directory in Cygwin (Tom)
437 * Add --port flag to regression tests
438 * Add documentation for perl, including mention of DBI/DBD perl location
439 * Add optional CRC checksum to heap and index pages
440 * Change representation of whole-tuple parameters to functions
441 * Clarify use of 'application' and 'command' tags in SGML docs
442 * Better document ability to build only certain interfaces (Marc)
443 * Remove or relicense modules that are not under the BSD license, if possible
444 * Remove memory/file descriptor freeing befor elog(ERROR) (Bruce)
445 * Acquire lock on a relation before building a relcache entry for it
446 * Research interaction of setitimer() and sleep() used by statement_timeout
448 ---------------------------------------------------------------------------
451 Developers who have claimed items are:
452 --------------------------------------
453 * Barry is Barry Lind <barry@xythos.com>
454 * Billy is Billy G. Allie <Bill.Allie@mug.org>
455 * Bruce is Bruce Momjian <pgman@candle.pha.pa.us> of Software Research Assoc.
456 * Christopher is Christopher Kings-Lynne <chriskl@familyhealth.com.au> of
457 Family Health Network
458 * D'Arcy is D'Arcy J.M. Cain <darcy@druid.net> of The Cain Gang Ltd.
459 * Dave is Dave Cramer <dave@fastcrypt.com>
460 * Edmund is Edmund Mergl <E.Mergl@bawue.de>
461 * Fernando Nasser <fnasser@redhat.com> of Red Hat
462 * Gavin Sherry <swm@linuxworld.com.au> of Alcove Systems Engineering
463 * Hiroshi is Hiroshi Inoue <Inoue@tpf.co.jp>
464 * Karel is Karel Zak <zakkr@zf.jcu.cz>
465 * Jan is Jan Wieck <JanWieck@Yahoo.com> of PeerDirect Corp.
466 * Liam is Liam Stewart <liams@redhat.com> of Red Hat
467 * Marc is Marc Fournier <scrappy@hub.org> of PostgreSQL, Inc.
468 * Mark is Mark Hollomon <mhh@mindspring.com>
469 * Michael is Michael Meskes <meskes@postgresql.org> of Credativ
470 * Neil is Neil Conway <neilc@samurai.com>
471 * Oleg is Oleg Bartunov <oleg@sai.msu.su>
472 * Peter M is Peter T Mount <peter@retep.org.uk> of Retep Software
473 * Peter E is Peter Eisentraut <peter_e@gmx.net>
474 * Philip is Philip Warner <pjw@rhyme.com.au> of Albatross Consulting Pty. Ltd.
475 * Rod is Rod Taylor <rbt@zort.ca>
476 * Ross is Ross J. Reedstrom <reedstrm@wallace.ece.rice.edu>
477 * Stephan is Stephan Szabo <sszabo@megazone23.bigpanda.com>
478 * Tatsuo is Tatsuo Ishii <t-ishii@sra.co.jp> of Software Research Assoc.
479 * Thomas is Thomas Lockhart <lockhart@fourpalms.org> of Jet Propulsion Labratory
480 * Tom is Tom Lane <tgl@sss.pgh.pa.us> of Red Hat
481 * Vadim is Vadim B. Mikheev <vadim4o@email.com> of Sector Data