3 Current maintainer: Bruce Momjian (bruce@momjian.us)
4 Last updated: Tue Jul 15 11:03:34 EDT 2008
6 The most recent version of this document can be viewed at
7 http://www.postgresql.org/docs/faqs.TODO.html.
9 #A hyphen, "-", marks changes that will appear in the upcoming 8.4 release.#
10 #A percent sign, "%", marks items that are easier to implement.#
12 This list contains all known PostgreSQL bugs and feature requests. If
13 you would like to work on an item, please read the Developer's FAQ
14 first. There is also a developer's wiki at
15 http://developer.postgresql.org.
21 * -Allow administrators to safely terminate individual sessions either
22 via an SQL function or SIGTERM
23 * Check for unreferenced table files created by transactions that were
24 in-progress when the server terminated abruptly
26 http://archives.postgresql.org/pgsql-patches/2006-06/msg00096.php
28 * Set proper permissions on non-system schemas during db creation
30 Currently all schemas are owned by the super-user because they are copied
31 from the template1 database. However, since all objects are inherited
32 from the template database, it is not clear that setting schemas to the db
35 * -Add function to report the time of the most recent server reload
36 * Allow log_min_messages to be specified on a per-module basis
38 This would allow administrators to see more detailed information from
39 specific sections of the backend, e.g. checkpoints, autovacuum, etc.
40 Another idea is to allow separate configuration files for each module,
41 or allow arbitrary SET commands to be passed to them.
43 * Simplify ability to create partitioned tables
45 This would allow creation of partitioned tables without requiring
46 creation of triggers or rules for INSERT/UPDATE/DELETE, and constraints
47 for rapid partition selection. Options could include range and hash
50 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00375.php
51 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00151.php
52 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00028.php
53 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00248.php
54 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00387.php
55 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00413.php
57 * Allow auto-selection of partitioned tables for min/max() operations
58 * Allow more complex user/database default GUC settings
60 Currently ALTER USER and ALTER DATABASE support per-user and
61 per-database defaults. Consider adding per-user-and-database
62 defaults so things like search_path can be defaulted for a
63 specific user connecting to a specific database.
65 * Allow custom variables to appear in pg_settings()
66 * Allow custom variable classes that can restrict who can set the values
68 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00911.php
70 * Implement the SQL standard mechanism whereby REVOKE ROLE revokes only
71 the privilege granted by the invoking role, and not those granted
74 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00010.php
76 * Allow SSL authentication/encryption over unix domain sockets
78 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00924.php
80 * Allow SSL key file permission checks to be optionally disabled when
81 sharing SSL keys with other applications
83 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00069.php
85 * Allow client certificate names to be checked against the client
88 This is already implemented in
89 libpq/fe-secure.c::verify_peer_name_matches_certificate() but the code
92 * Improve server security options
94 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01875.php
95 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00000.php
99 o Allow pg_hba.conf to specify host names along with IP addresses
101 Host name lookup could occur when the postmaster reads the
102 pg_hba.conf file, or when the backend starts. Another
103 solution would be to reverse lookup the connection IP and
104 check that hostname against the host names in pg_hba.conf.
105 We could also then check that the host name maps to the IP
107 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00569.php
109 o %Allow postgresql.conf file values to be changed via an SQL
110 API, perhaps using SET GLOBAL
111 o Allow the server to be stopped/restarted via an SQL API
112 o Issue a warning if a change-on-restart-only postgresql.conf value
113 is modified and the server config files are reloaded
114 o Consider normalizing fractions in postgresql.conf, perhaps
117 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00550.php
119 o Allow Kerberos to disable stripping of realms so we can
120 check the username@realm against multiple realms
122 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00009.php
124 o Add functions to syntax check configuration files
126 o Improve LDAP authentication configuration options
128 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01745.php
133 o Allow a database in tablespace t1 with tables created in
134 tablespace t2 to be used as a template for a new database created
135 with default tablespace t2
137 Currently all objects in the default database tablespace must
138 have default tablespace specifications. This is because new
139 databases are created by copying directories. If you mix default
140 tablespace tables and tablespace-specified tables in the same
141 directory, creating a new database from such a mixed directory
142 would create a new database with tables that had incorrect
143 explicit tablespaces. To fix this would require modifying
144 pg_class in the newly copied database, which we don't currently
147 o Allow reporting of which objects are in which tablespaces
149 This item is difficult because a tablespace can contain objects
150 from multiple databases. There is a server-side function that
151 returns the databases which use a specific tablespace, so this
152 requires a tool that will call that function and connect to each
153 database to find the objects in each database for that tablespace.
155 o Allow WAL replay of CREATE TABLESPACE to work when the directory
156 structure on the recovery computer is different from the original
158 o Allow per-tablespace quotas
161 * Statistics Collector
163 o Allow statistics collector information to be pulled from the collector
164 process directly, rather than requiring the collector to write a
165 filesystem file twice a second?
166 o Reduce file system activity overhead of statistics file pgstat.stat
168 http://archives.postgresql.org/pgsql-general/2007-12/msg00106.php
170 o Allow statistics last vacuum/analyze execution times to be displayed
171 without requiring stats_row_level to be enabled
173 http://archives.postgresql.org/pgsql-docs/2007-04/msg00028.php
175 o Clear table counters on TRUNCATE
177 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00169.php
180 * Point-In-Time Recovery (PITR)
182 o Allow a warm standby system to also allow read-only statements
184 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00050.php
186 o %Create dump tool for write-ahead logs for use in determining
187 transaction id for point-in-time recovery
189 This is useful for checking PITR recovery.
191 o Allow recovery.conf to support the same syntax as
192 postgresql.conf, including quoting
194 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php
196 o -Fix server restart problem when the server was shutdown during
199 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00800.php
201 o Recreate pg_xlog/archive_status/ if it doesn't exist after
202 restoring from a PITR backup
204 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00487.php
206 o Reduce PITR WAL file size by removing full page writes and
207 by removing trailing bytes to improve compression
213 * Change NUMERIC to enforce the maximum precision
214 * Reduce storage space for small NUMERICs
216 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php
217 http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php
218 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php
220 * Fix data types where equality comparison isn't intuitive, e.g. box
221 * Add support for public SYNONYMs
223 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php
225 * Fix CREATE CAST on DOMAINs
227 http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php
228 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php
230 * Allow domains to be cast
232 http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php
233 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php
235 * Add support for SQL-standard GENERATED/IDENTITY columns
237 http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php
238 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php
239 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php
240 http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php
241 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php
243 * Improve XML support
245 http://developer.postgresql.org/index.php/XML_Support
247 * Consider placing all sequences in a single table, or create a system
250 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
252 * Allow the UUID type to accept non-standard formats
254 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php
256 * Consider a special data type for regular expressions
258 http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php
260 * Reduce BIT data type overhead using short varlena headers
262 http://archives.postgresql.org/pgsql-general/2007-12/msg00273.php
264 * Allow xml arrays to be cast to other data types
266 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00981.php
267 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00231.php
268 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00471.php
270 * Simplify integer cross-data-type operators
272 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00189.php
274 * Allow XML to accept more liberal DOCTYPE specifications
276 http://archives.postgresql.org/pgsql-general/2008-02/msg00347.php
278 * Allow adding/renaming/removing enumerated values to an existing
281 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01718.php
285 o Allow infinite dates and intervals just like infinite timestamps
286 o Merge hardwired timezone names with the TZ database; allow either
287 kind everywhere a TZ name is currently taken
288 o Allow TIMESTAMP WITH TIME ZONE to store the original timezone
289 information, either zone name or offset from UTC
291 If the TIMESTAMP value is stored with a time zone name, interval
292 computations should adjust based on the time zone rules.
293 http://archives.postgresql.org/pgsql-hackers/2004-10/msg00705.php
295 o Fix SELECT '0.01 years'::interval, '0.01 months'::interval
296 o Add a GUC variable to allow output of interval values in ISO8601
298 o Have timestamp subtraction not call justify_hours()?
300 http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php
302 o Improve timestamptz subtraction to be DST-aware
304 Currently subtracting one date from another that crosses a
305 daylight savings time adjustment can return '1 day 1 hour', but
306 adding that back to the first date returns a time one hour in
307 the future. This is caused by the adjustment of '25 hours' to
308 '1 day 1 hour', and '1 day' is the same time the next day, even
309 if daylight savings adjustments are involved.
311 o Fix interval display to support values exceeding 2^31 hours
312 o Add overflow checking to timestamp and interval arithmetic
313 o Extend timezone code to allow 64-bit values so we can
314 represent years beyond 2038
316 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php
318 o -Use LC_TIME for localized weekday/month names, rather than
321 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php
323 o Add ISO INTERVAL handling
325 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php
326 http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php
328 o Support ISO INTERVAL syntax if units cannot be determined from
329 the string, and are supplied after the string
331 The SQL standard states that the units after the string
332 specify the units of the string, e.g. INTERVAL '2' MINUTE
333 should return '00:02:00'. The current behavior has the units
334 restrict the interval value to the specified unit or unit
335 range, INTERVAL '70' SECOND returns '00:00:10'.
337 For syntax that isn't uniquely ISO or PG syntax, like '1' or
338 '1:30', treat as ISO if there is a range specification clause,
339 and as PG if there no clause is present, e.g. interpret '1:30'
340 MINUTE TO SECOND as '1 minute 30 seconds', and interpret
341 '1:30' as '1 hour, 30 minutes'.
343 This makes common cases like SELECT INTERVAL '1' MONTH
344 SQL-standard results. The SQL standard supports a limited
345 number of unit combinations and doesn't support unit names in
346 the string. The PostgreSQL syntax is more flexible in the
347 range of units supported, e.g. PostgreSQL supports '1 year 1
348 hour', while the SQL standard does not.
350 o Add support for year-month syntax, INTERVAL '50-6' YEAR
352 o Interpret INTERVAL '1 year' MONTH as CAST (INTERVAL '1
353 year' AS INTERVAL MONTH), and this should return '12 months'
354 o Round or truncate values to the requested precision, e.g.
355 INTERVAL '11 months' AS YEAR should return one or zero
356 o Support precision, CREATE TABLE foo (a INTERVAL MONTH(3))
361 o Delay resolution of array expression's data type so assignment
362 coercion can be performed on empty array expressions
363 o Add support for arrays of domains
365 http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php
367 o Allow single-byte header storage for array elements
372 o Improve vacuum of large objects, like contrib/vacuumlo?
373 o Add security checking for large objects
374 o Auto-delete large objects when referencing row is deleted
376 contrib/lo offers this functionality.
378 o Allow read/write into TOAST values like large objects
380 This requires the TOAST column to be stored EXTERNAL.
382 o Add API for 64-bit large object access
384 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php
388 * Add locale-aware MONEY type, and support multiple currencies
390 http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
391 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
393 * MONEY dumps in a locale-specific format making it difficult to
394 restore to a system with a different locale
395 * Allow MONEY to be easily cast to/from other numeric data types
399 o Allow dictionaries to change the token that is passed on to
402 http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php
404 o Consider a function-based API for '@@' searches
406 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php
408 o Improve text search error messages
410 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00966.php
411 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01146.php
413 o Consider changing error to warning for strings larger than one
416 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00190.php
417 http://archives.postgresql.org/pgsql-patches/2008-03/msg00062.php
424 * Allow INET subnet tests using non-constants to be indexed
425 * Allow to_date() and to_timestamp() accept localized month names
426 * Fix to_date()-related functions to consistently issue errors
428 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php
430 * Add missing parameter handling in to_char()
432 http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php
434 * Allow substring/replace() to get/set bit values
435 * Allow to_char() on interval values to accumulate the highest unit
438 Some special format flag would be required to request such
439 accumulation. Such functionality could also be added to EXTRACT.
440 Prevent accumulation that crosses the month/day boundary because of
441 the uneven number of days in a month.
443 o to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
444 o to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
445 o to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
446 o to_char(INTERVAL '3 years 5 months','MM') => 41
448 * Implement inlining of set-returning functions defined in SQL
449 * Allow SQL-language functions to return results from RETURNING queries
451 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php
453 * Allow SQL-language functions to reference parameters by parameter name
455 Currently SQL-language functions can only refer to dollar parameters,
458 * Add SPI_gettypmod() to return the typemod for a TupleDesc
459 * Enforce typmod for function inputs, function results and parameters for
460 spi_prepare'd statements called from PLs
462 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php
464 * Allow holdable cursors in SPI
465 * Tighten function permission checks
467 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php
469 * Fix IS OF so it matches the ISO specification, and add documentation
471 http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php
472 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php
474 * Add missing operators for geometric data types
476 Some geometric types do not have the full suite of geometric operators,
479 * Implement Boyer-Moore searching in strpos()
481 http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php
483 * Prevent malicious functions from being executed with the permissions
484 of unsuspecting users
486 Index functions are safe, so VACUUM and ANALYZE are safe too.
487 Triggers, CHECK and DEFAULT expressions, and rules are still vulnerable.
488 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php
490 * Reduce memory usage of aggregates in set returning functions
492 http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php
494 * -Add temporal versions of generate_series()
496 http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php
498 * Add array_accum() and array_to_set() functions for arrays
500 The standards specify array_agg() and UNNEST.
501 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php
503 * Fix /contrib/ltree operator
505 http://archives.postgresql.org/pgsql-bugs/2007-11/msg00044.php
507 * Fix inconsistent precedence of =, >, and < compared to <>, >=, and <=
509 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00145.php
511 * Fix regular expression bug when using complex back-references
513 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00000.php
515 * Have /contrib/dblink reuse unnamed connections
517 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00895.php
519 * Add SQL-standard array_agg() and unnest() array functions
521 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01017.php
523 * Allow calling of a procedure outside a SELECT that can control the
526 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01375.php
530 Multi-Language Support
531 ======================
533 * Add NCHAR (as distinguished from ordinary varchar),
534 * Allow locale to be set at database creation
536 Currently locale can only be set during initdb. No global tables have
537 locale-aware columns. However, the database template used during
538 database creation might have locale-aware indexes. The indexes would
539 need to be reindexed to match the new locale.
541 * Allow encoding on a per-column basis optionally using the ICU library;
544 Right now only one encoding is allowed per database.
546 http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php
547 http://archives.postgresql.org/pgsql-patches/2005-08/msg00039.php
548 http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php
549 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00110.php
550 http://archives.postgresql.org/pgsql-patches/2005-09/msg00020.php
551 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01121.php
552 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00767.php
553 http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php
554 http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php
555 http://wiki.postgresql.org/wiki/Todo:Collate
556 http://wiki.postgresql.org/wiki/Todo:ICU
558 * Support multiple simultaneous character sets, per SQL92
559 * Improve UTF8 combined character handling?
560 * Add octet_length_server() and octet_length_client()
561 * Make octet_length_client() the same as octet_length()?
562 * Fix problems with wrong runtime encoding conversion for NLS message files
563 * Add URL to more complete multi-byte regression tests
565 http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php
567 * Fix ILIKE and regular expressions to handle case insensitivity
568 properly in multibyte encodings
570 http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php
571 http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php
573 * Set client encoding based on the client operating system encoding
575 Currently client_encoding is set in postgresql.conf, which
576 defaults to the server encoding.
577 http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php
579 * Change memory allocation for multi-byte functions so memory is
580 allocated inside conversion functions
582 Currently we preallocate memory based on worst-case usage.
589 * Automatically create rules on views so they are updateable, per SQL99
591 We can only auto-create rules for simple views. For more complex
592 cases users will still have to write rules manually.
594 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php
595 http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php
597 * Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
598 * Allow VIEW/RULE recompilation when the underlying tables change
600 Another issue is whether underlying table changes should be reflected
601 in the view, e.g. should SELECT * show additional columns if they
602 are added after the view is created.
604 * Make it possible to use RETURNING together with conditional DO INSTEAD
605 rules, such as for partitioning setups
607 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php
609 * Add the ability to automatically create materialized views
611 Right now materialized views require the user to create triggers on the
612 main table to keep the summary table current. SQL syntax should be able
613 to manager the triggers and summary table automatically. A more
614 sophisticated implementation would automatically retrieve from the
615 summary table when the main table is referenced, if possible.
617 * Improve ability to modify views via ALTER TABLE
619 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00691.php
626 * Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
627 * Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
628 * %Allow SET CONSTRAINTS to be qualified by schema/table name
629 * %Add a separate TRUNCATE permission
631 Currently only the owner can TRUNCATE a table because triggers are not
632 called, and the table is locked in exclusive mode.
634 * Fix TRUNCATE ... RESTART IDENTITY so its affect on sequences is rolled
635 back on transaction abort
636 * Allow PREPARE of cursors
637 * Allow finer control over the caching of prepared query plans
639 Currently queries prepared via the libpq API are planned on first
640 execute using the supplied parameters --- allow SQL PREPARE to do the
641 same. Also, allow control over replanning prepared queries either
642 manually or automatically when statistics for execute parameters
643 differ dramatically from those used during planning.
645 * Improve logging of prepared transactions recovered during startup
647 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php
649 * Improve failure message when DROP DATABASE is used on a database that
650 has prepared transactions
651 * Allow prepared transactions with temporary tables created and dropped
652 in the same transaction, and when an ON COMMIT DELETE ROWS temporary
655 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00047.php
657 * Add a GUC variable to warn about non-standard SQL usage in queries
658 * Add SQL-standard MERGE/REPLACE/UPSERT command
660 MERGE is typically used to merge two tables. REPLACE or UPSERT
661 command does UPDATE, or on failure, INSERT. This is similar to UPDATE,
662 then for unmatched rows, INSERT. Whether concurrent access allows
663 modifications which could cause row loss is implementation independent.
664 To implement this cleanly requires that the table have a unique index
665 so duplicate checking can be easily performed. It is possible to do it
666 without a unique index if we require the user to LOCK the table before
669 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php
670 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php
671 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01157.php
672 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01475.php
673 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01890.php
675 * Add NOVICE output level for helpful messages like automatic sequence/index
677 * Add GUC to issue notice about statements that use unjoined tables
678 * Allow EXPLAIN to identify tables that were skipped because of
680 * Allow EXPLAIN output to be more easily processed by scripts, perhaps XML
681 * Enable standard_conforming_strings
682 * Make standard_conforming_strings the default in 8.5?
684 When this is done, backslash-quote should be prohibited in non-E''
685 strings because of possible confusion over how such strings treat
686 backslashes. Basically, '' is always safe for a literal single
687 quote, while \' might or might not be based on the backslash
690 * Simplify dropping roles that have objects in several databases
691 * Allow COMMENT ON to accept an expression rather than just a string
692 * Allow the count returned by SELECT, etc to be represented as an int64
693 to allow a higher range of values
694 * Add SQL99 WITH clause to SELECT
695 * Add SQL:2003 WITH RECURSIVE (hierarchical) queries to SELECT
697 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php
698 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php
699 http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php
700 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01334.php
701 http://archives.postgresql.org/pgsql-patches/2008-01/msg00105.php
702 http://archives.postgresql.org/pgsql-patches/2008-03/msg00327.php
704 * Add DEFAULT .. AS OWNER so permission checks are done as the table
707 This would be useful for SERIAL nextval() calls and CHECK constraints.
709 * Allow DISTINCT to work in multiple-argument aggregate calls
710 * Add column to pg_stat_activity that shows the progress of long-running
711 commands like CREATE INDEX and VACUUM
713 http://archives.postgresql.org/pgsql-patches/2008-04/msg00203.php
715 * Implement SQL:2003 window functions
716 * Allow INSERT/UPDATE ... RETURNING inside a SELECT 'FROM' clause or
719 http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php
720 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php
721 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00124.php
723 * Increase locking when DROPing objects so dependent objects cannot
724 get dropped while the DROP operation is happening
726 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php
728 * -Allow AS in "SELECT col AS label" to be optional in certain cases
730 * Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
731 RETURNING to supply values to the INSERT
732 http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979
734 * Add comments on system tables/columns using the information in
737 Ideally the information would be pulled from the SGML file
740 * Improve reporting of UNION type mismatches
742 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00944.php
743 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00597.php
748 o Allow CREATE TABLE AS to determine column lengths for complex
749 expressions like SELECT col1 || col2
751 o Have WITH CONSTRAINTS also create constraint indexes
753 http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php
755 o Have CONSTRAINT cname NOT NULL record the contraint name
757 Right now pg_attribute.attnotnull records the NOT NULL status
758 of the column, but does not record the contraint name
760 o Prevent concurrent CREATE TABLE table1 from sometimes returning
761 a cryptic error message
763 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00169.php
765 o Add CREATE SCHEMA ... LIKE that copies a schema
770 o Allow UPDATE tab SET ROW (col, ...) = (SELECT...)
772 http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php
773 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php
774 http://archives.postgresql.org/pgsql-patches/2007-04/msg00315.php
775 http://archives.postgresql.org/pgsql-patches/2008-03/msg00237.php
777 o Research self-referential UPDATEs that see inconsistent row versions
778 in read-committed mode
780 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php
781 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php
783 o Allow GLOBAL temporary tables to exist as empty by default in
786 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php
791 o Have ALTER TABLE RENAME rename SERIAL sequence names
793 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
795 o Have ALTER SEQUENCE RENAME rename the sequence name stored
796 in the sequence table
798 http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php
799 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php
800 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
802 o Add ALTER DOMAIN to modify the underlying data type
803 o %Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
805 http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php
807 o %Allow ALTER TABLE to change constraint deferrability and actions
808 o Add missing object types for ALTER ... SET SCHEMA
809 o Allow ALTER TABLESPACE to move to different directories
810 o Allow databases to be moved to different tablespaces
811 o Allow moving system tables to other tablespaces, where possible
813 Currently non-global system tables must be in the default database
814 tablespace. Global system tables can never be moved.
816 o -Prevent parent tables from altering or dropping constraints
817 like CHECK that are inherited by child tables unless CASCADE
819 o -Prevent child tables from altering or dropping constraints
820 like CHECK that were inherited from the parent table
821 o Have ALTER INDEX update the name of a constraint using that index
822 o Add ALTER TABLE RENAME CONSTRAINT, update index name also
823 o Allow column display reordering by recording a display,
824 storage, and permanent id for every column?
826 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php
828 o Allow an existing index to be marked as a table's primary key
830 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00500.php
835 o Automatically maintain clustering on a table
837 This might require some background daemon to maintain clustering
838 during periods of low usage. It might also require tables to be only
839 partially filled for easier reorganization. Another idea would
840 be to create a merged heap/index data file so an index lookup would
841 automatically access the heap data too. A third idea would be to
842 store heap rows in hashed groups, perhaps using a user-supplied
844 http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php
846 o %Add default clustering to system tables
848 To do this, determine the ideal cluster index for each system
849 table and set the cluster setting during initdb.
851 o %Add VERBOSE option to report tables as they are processed,
857 o Allow COPY to report error lines and continue
859 This requires the use of a savepoint before each COPY line is
860 processed, with ROLLBACK on COPY failure.
861 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00572.php
863 o Allow COPY on a newly-created table to skip WAL logging
865 On crash recovery, the table involved in the COPY would
866 be removed or have its heap and index files truncated. One
867 issue is that no other backend should be able to add to
868 the table at the same time, which is something that is
869 currently allowed. This currently is done if the table is
870 created inside the same transaction block as the COPY because
871 no other backends can see the table.
873 o Consider using a ring buffer for COPY FROM
875 http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php
876 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php
878 o Allow COPY FROM to create index entries in bulk
880 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php
882 o Allow COPY in CSV mode to control whether a quoted zero-length
883 string is treated as NULL
885 Currently this is always treated as a zero-length string,
886 which generates an error when loading into an integer column
887 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php
889 o Impove COPY performance
891 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00954.php
893 o Allow COPY to report errors sooner
895 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01169.php
901 o Allow column-level privileges
902 o %Allow GRANT/REVOKE permissions to be applied to all schema objects
905 The proposed syntax is:
906 GRANT SELECT ON ALL TABLES IN public TO phpuser;
907 GRANT SELECT ON NEW TABLES IN public TO phpuser;
909 o Allow GRANT/REVOKE permissions to be inherited by objects based on
912 o Allow SERIAL sequences to inherit permissions from the base table?
917 o Prevent DROP TABLE from dropping a row referenced by its own open
923 o Allow INSERT/UPDATE of the system-generated oid value for a row
924 o In rules, allow VALUES() to contain a mixture of 'old' and 'new'
930 o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
936 o Allow LISTEN/NOTIFY to store info in memory rather than tables?
938 Currently LISTEN/NOTIFY information is stored in pg_listener.
939 Storing such information in memory would improve performance.
941 o Add optional textual message to NOTIFY
943 This would allow an informational message to be added to the notify
944 message, perhaps indicating the row modified or other custom
947 o Allow multiple identical NOTIFY events to always be communicated
948 to the client, rather than sent as a single notification to the
951 http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php
953 o Allow NOTIFY in rules involving conditionals
954 o Improve LISTEN concurrency
956 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01106.php
960 Referential Integrity
961 =====================
963 * Add MATCH PARTIAL referential integrity
964 * Change foreign key constraint for array -> element to mean element
966 * Fix problem when cascading referential triggers make changes on
967 cascaded tables, seeing the tables in an intermediate state
969 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
970 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
972 * Allow DEFERRABLE and end-of-statement UNIQUE constraints?
974 This would allow UPDATE tab SET col = col + 1 to work if col has
975 a unique index. Currently, uniqueness checks are done while the
976 command is being executed, rather than at the end of the statement
978 http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html
979 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php
981 * Optimize referential integrity checks
983 http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php
984 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php
987 Server-Side Languages
988 =====================
991 o Fix RENAME to work on variables other than OLD/NEW
993 http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php
994 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php
995 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php
997 o Allow function parameters to be passed by name,
998 get_employee_salary(12345 AS emp_id, 2001 AS tax_year)
999 o Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
1000 o Allow listing of record column names, and access to
1001 record columns via variables, e.g. columns := r.(*),
1002 tval2 := r.(colname)
1004 http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php
1005 http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php
1006 http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php
1008 o Add support for SCROLL cursors
1009 o Add support for WITH HOLD cursors
1010 o Allow row and record variables to be set to NULL constants,
1011 and allow NULL tests on such variables
1013 Because a row is not scalar, do not allow assignment
1014 from NULL-valued scalars.
1016 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php
1018 o Review handling of MOVE and FETCH
1020 http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php
1022 o Improve logic of determining if an identifier is a a
1023 variable or column name
1025 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php
1027 o Consider keeping seperate cached copies when search_path changes
1029 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01009.php
1031 o -Add CASE capability to language (already in SQL)
1033 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00696.php
1039 o Add table function support to pltcl, plpythonu
1040 o Add support for polymorphic arguments and return types to
1041 languages other than PL/PgSQL
1042 o Add capability to create and call PROCEDURES
1043 o Add support for OUT and INOUT parameters to languages other
1045 o Add PL/PythonU tracebacks
1047 http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php
1049 o Allow data to be passed in native language formats, rather
1052 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php
1054 o Add ability to obfuscate function bodies
1056 http://archives.postgresql.org/pgsql-patches/2008-01/msg00125.php
1063 * Have pg_ctl look at PGHOST in case it is a socket directory?
1064 * Allow pg_ctl to work properly with configuration files located outside
1065 the PGDATA directory
1067 pg_ctl can not read the pid file because it isn't located in the
1068 config directory but in the PGDATA directory. The solution is to
1069 allow pg_ctl to read and understand postgresql.conf to find the
1070 data_directory value.
1072 * Add a function like pg_get_indexdef() that report more detailed index
1075 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php
1078 o -Have psql show current values for a sequence
1079 o Move psql backslash database information into the backend, use
1082 This would allow non-psql clients to pull the same information out
1083 of the database as psql.
1084 http://archives.postgresql.org/pgsql-hackers/2004-01/msg00191.php
1086 o Make psql's \d commands more consistent
1088 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1089 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1091 o Consistently display privilege information for all objects in psql
1092 o Add auto-expanded mode so expanded output is used if the row
1093 length is wider than the screen width.
1095 Consider using auto-expanded mode for backslash commands like \df+.
1097 o Prevent tab completion of SET TRANSACTION from querying the
1098 database and therefore preventing the transaction isolation
1099 level from being set.
1101 Currently SET <tab> causes a database lookup to check all
1102 supported session variables. This query causes problems
1103 because setting the transaction isolation level must be the
1104 first statement of a transaction.
1106 o Add a \set variable to control whether \s displays line numbers
1108 Another option is to add \# which lists line numbers, and
1109 allows command execution.
1111 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php
1113 o Prevent escape string warnings when object names have
1116 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php
1118 o Have \d show foreign keys that reference a table's primary key
1120 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php
1122 o Have \d show child tables that inherit from the specified parent
1123 o -Have \l+ show database size, if permissions allow
1125 Ideally it will not generate an error for invalid permissions
1127 o Include the symbolic SQLSTATE name in verbose error reports
1129 http://archives.postgresql.org/pgsql-general/2007-09/msg00438.php
1131 o -Improve display of enums to show valid enum values
1132 o Add prompt escape to display the client and server versions
1133 o Add option to wrap column values at whitespace boundaries,
1134 rather than chopping them at a fixed width.
1136 Currently, "wrapped" format chops values into fixed
1137 widths. Perhaps the word wrapping could use the same
1138 algorithm documented in the W3C specification.
1139 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00404.php
1140 http://www.w3.org/TR/CSS21/tables.html#auto-table-layout
1141 o Add "auto" expanded mode that outputs in expanded format if
1142 "wrapped" mode can't wrap the output to the screen width
1144 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00417.php
1147 * pg_dump / pg_restore
1148 o %Add dumping of comments on index columns and composite type columns
1149 o %Add full object name to the tag field. eg. for operators we need
1150 '=(integer, integer)', instead of just '='.
1151 o Add pg_dumpall custom format dumps?
1152 o Allow selection of individual object(s) of all types, not just
1154 o In a selective dump, allow dumping of an object and all its
1156 o Add options like pg_restore -l and -L to pg_dump
1157 o Stop dumping CASCADE on DROP TYPE commands in clean mode
1158 o Allow pg_dump --clean to drop roles that own objects or have
1160 o Change pg_dump so that a comment on the dumped database is
1161 applied to the loaded database, even if the database has a
1162 different name. This will require new backend syntax, perhaps
1163 COMMENT ON CURRENT DATABASE.
1164 o Remove unnecessary function pointer abstractions in pg_dump source
1166 o Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
1167 multiple objects simultaneously
1169 The difficulty with this is getting multiple dump processes to
1170 produce a single dump output file. It also would require
1171 several sessions to share the same snapshot.
1172 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1174 o Allow pg_restore to utilize multiple CPUs and I/O channels by
1175 restoring multiple objects simultaneously
1177 This might require a pg_restore flag to indicate how many
1178 simultaneous operations should be performed. Only pg_dump's
1179 -Fc format has the necessary dependency information.
1180 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php
1182 o To better utilize resources, allow pg_restore to check foreign
1183 keys simultaneously, where possible
1184 o Allow pg_restore to create all indexes of a table
1185 concurrently, via a single heap scan
1187 This requires a pg_dump -Fc file because that format contains
1188 the required dependency information.
1189 http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php
1191 o Allow pg_restore to load different parts of the COPY data
1193 o -Prevent pg_dump/pg_restore from being affected by
1196 Using psql to restore a pg_dump dump is also affected.
1198 o Remove pre-7.3 pg_dump code that assumes pg_depend does not exit
1199 o Allow pre/data/post files when schema and data are dumped
1200 separately, for performance reasons
1202 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1209 Document differences between ecpg and the SQL standard and
1210 information about the Informix-compatibility module.
1212 o Solve cardinality > 1 for input descriptors / variables?
1213 o Add a semantic check level, e.g. check if a table really exists
1214 o fix handling of DB attributes that are arrays
1215 o Use backend PREPARE/EXECUTE facility for ecpg where possible
1217 o Fix nested C comments
1218 o %sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified
1219 o Make SET CONNECTION thread-aware, non-standard?
1220 o Allow multidimensional arrays
1221 o Add internationalized message strings
1222 o Implement COPY FROM STDIN
1226 o Add PQescapeIdentifierConn()
1227 o Prevent PQfnumber() from lowercasing unquoted the column name
1229 PQfnumber() should never have been doing lowercasing, but
1230 historically it has so we need a way to prevent it
1232 o Allow statement results to be automatically batched to the client
1234 Currently all statement results are transferred to the libpq
1235 client before libpq makes the results available to the
1236 application. This feature would allow the application to make
1237 use of the first result rows while the rest are transferred, or
1238 held on the server waiting for them to be requested by libpq.
1239 One complexity is that a statement like SELECT 1/col could error
1240 out mid-way through the result set.
1242 o Consider disallowing multiple queries in PQexec() as an
1243 additional barrier to SQL injection attacks
1245 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php
1247 o Add PQexecf() that allows complex parameter substitution
1249 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php
1251 o Add SQLSTATE severity to PGconn return status
1253 http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php
1259 * Add deferred trigger queue file
1261 Right now all deferred trigger information is stored in backend
1262 memory. This could exhaust memory for very large trigger queues.
1263 This item involves dumping large queues into files, or doing some
1264 kind of join to process all the triggers, or some bulk operation.
1265 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00876.php
1267 * Allow triggers to be disabled in only the current session.
1269 This is currently possible by starting a multi-statement transaction,
1270 modifying the system tables, performing the desired SQL, restoring the
1271 system tables, and committing the transaction. ALTER TABLE ...
1272 TRIGGER requires a table lock so it is not ideal for this usage.
1274 * With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
1276 If the dump is known to be valid, allow foreign keys to be added
1277 without revalidating the data.
1279 * Allow statement-level triggers to access modified rows
1280 * Support triggers on columns
1282 http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php
1284 * Allow AFTER triggers on system tables
1286 System tables are modified in many places in the backend without going
1287 through the executor and therefore not causing triggers to fire. To
1288 complete this item, the functions that modify system tables will have
1291 * Tighten trigger permission checks
1293 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
1295 * Allow BEFORE INSERT triggers on views
1297 http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php
1299 * -Add ability to trigger on TRUNCATE
1300 * Add database and transaction-level triggers
1302 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00451.php
1303 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00620.php
1309 * Add UNIQUE capability to non-btree indexes
1310 * Prevent index uniqueness checks when UPDATE does not modify the column
1312 Uniqueness (index) checks are done when updating a column even if the
1313 column is not modified by the UPDATE.
1315 * Allow the creation of on-disk bitmap indexes which can be quickly
1316 combined with other bitmap indexes
1318 Such indexes could be more compact if there are only a few distinct values.
1319 Such indexes can also be compressed. Keeping such indexes updated can be
1322 http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php
1323 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php
1324 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00265.php
1325 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01214.php
1326 http://archives.postgresql.org/pgsql-patches/2007-05/msg00013.php
1327 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00741.php
1329 * Allow accurate statistics to be collected on indexes with more than
1330 one column or expression indexes, perhaps using per-index statistics
1332 http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php
1333 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php
1335 * Consider increasing the default and maximum number of statistics targets,
1336 and reduce statistics target overhead
1338 Also consider having a larger statistics target for indexed columns
1339 and expression indexes.
1340 http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php
1341 http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php
1342 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01066.php
1343 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00188.php
1345 * Consider smaller indexes that record a range of values per heap page,
1346 rather than having one index entry for every heap row
1348 This is useful if the heap is clustered by the indexed values.
1349 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php
1350 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php
1351 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php
1352 http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php
1353 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00014.php
1354 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00487.php
1355 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01589.php
1357 * Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
1359 This is difficult because you must upgrade to an exclusive table lock
1360 to replace the existing index file. CREATE INDEX CONCURRENTLY does not
1361 have this complication. This would allow index compaction without
1363 http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php
1365 * Allow multiple indexes to be created concurrently, ideally via a
1366 single heap scan, and have pg_restore use it
1368 * Consider sorting entries before inserting into btree index
1370 http://archives.postgresql.org/pgsql-general/2008-01/msg01010.php
1372 * Allow index scans to return matching index keys, not just the matching
1375 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01657.php
1381 o Allow inherited tables to inherit indexes, UNIQUE constraints,
1382 and primary/foreign keys
1383 o Honor UNIQUE INDEX on base column in INSERTs/UPDATEs
1384 on inherited table, e.g. INSERT INTO inherit_table
1385 (unique_index_col) VALUES (dup) should fail
1387 The main difficulty with this item is the problem of
1388 creating an index that can span multiple tables.
1390 o Allow SELECT ... FOR UPDATE on inherited tables
1391 o Require all CHECK constraints to be inherited
1393 http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php
1395 o Add checks to prevent a CREATE RULE views on inherited tables
1397 http://archives.postgresql.org/pgsql-general/2008-02/msg01420.php
1398 http://archives.postgresql.org/pgsql-general/2008-03/msg00077.php
1403 o Add more GIST index support for geometric data types
1404 o Allow GIST indexes to create certain complex index types, like
1405 digital trees (see Aoki)
1410 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php
1412 o Pack hash index buckets onto disk pages more efficiently
1414 Currently only one hash bucket can be stored on a page. Ideally
1415 several hash buckets could be stored on a single page and greater
1416 granularity used for the hash algorithm.
1418 http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php
1420 o Consider sorting hash buckets so entries can be found using a
1421 binary search, rather than a linear scan
1422 o In hash indexes, consider storing the hash value with or instead
1424 o Add WAL logging for crash recovery
1425 o Allow multi-column hash indexes
1426 o -During index creation, pre-sort the tuples to improve build speed
1430 * Consider using hash buckets to do DISTINCT, rather than sorting
1432 This would be beneficial when there are few distinct values. This is
1433 already used by GROUP BY.
1435 * Consider whether duplicate keys should be sorted by block/offset
1437 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00558.php
1439 * -Avoid tuple some tuple copying in sort routines
1440 * Consider being smarter about memory and external files used during
1443 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01101.php
1444 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00045.php
1446 * Consider detoasting keys before sorting
1452 * Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
1454 Ideally this requires a separate test program that can be run
1455 at initdb time or optionally later. Consider O_SYNC when
1458 * Add program to test if fsync has a delay compared to non-fsync
1459 * Consider sorting writes during checkpoint
1461 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php
1470 We could use a fixed row count and a +/- count to follow MVCC
1471 visibility rules, or a single cached value could be used and
1472 invalidated if anyone modifies the table. Another idea is to
1473 get a count directly from a unique index, but for this to be
1474 faster than a sequential scan it must avoid access to the heap
1475 to obtain tuple visibility information.
1477 * Provide a way to calculate an "estimated COUNT(*)"
1479 Perhaps by using the optimizer's cardinality estimates or random
1482 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php
1484 * Allow data to be pulled directly from indexes
1486 Currently indexes do not have enough tuple visibility information
1487 to allow data to be pulled from the index without also accessing
1488 the heap. One way to allow this is to set a bit on index tuples
1489 to indicate if a tuple is currently visible to all transactions
1490 when the first valid heap lookup happens. This bit would have to
1491 be cleared when a heap tuple is expired.
1493 Another idea is to maintain a bitmap of heap pages where all rows
1494 are visible to all backends, and allow index lookups to reference
1495 that bitmap to avoid heap lookups, perhaps the same bitmap we might
1496 add someday to determine which heap pages need vacuuming. Frequently
1497 accessed bitmaps would have to be stored in shared memory. One 8k
1498 page of bitmaps could track 512MB of heap pages.
1500 A third idea would be for a heap scan to check if all rows are visible
1501 and if so set a per-table flag which can be checked by index scans.
1502 Any change to the table would have to clear the flag. To detect
1503 changes during the heap scan a counter could be set at the start and
1504 checked at the end --- if it is the same, the table has not been
1505 modified --- any table change would increment the counter.
1507 http://archives.postgresql.org/pgsql-patches/2007-10/msg00166.php
1508 http://archives.postgresql.org/pgsql-patches/2008-01/msg00049.php
1510 * Consider automatic caching of statements at various levels:
1513 o Query execute plan
1515 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00823.php
1517 * Consider increasing internal areas when shared buffers is increased
1519 http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php
1521 * Consider decreasing the amount of memory used by PrivateRefCount
1523 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php
1524 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php
1526 * Consider allowing higher priority queries to have referenced buffer
1527 cache pages stay in memory longer
1529 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php
1536 * Improve speed with indexes
1538 For large table adjustments during VACUUM FULL, it is faster to cluster
1539 or reindex rather than update the index. Also, index updates can bloat
1542 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php
1543 http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php
1544 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php
1546 * Auto-fill the free space map by scanning the buffer cache or by
1547 checking pages written by the background writer
1549 http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php
1550 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php
1552 * Create a bitmap of pages that need vacuuming
1554 Instead of sequentially scanning the entire table, have the background
1555 writer or some other process record pages that have expired rows, then
1556 VACUUM can look at just those pages rather than the entire table. In
1557 the event of a system crash, the bitmap would probably be invalidated.
1558 One complexity is that index entries still have to be vacuumed, and
1559 doing this without an index scan (by using the heap values to find the
1560 index entry) might be slow and unreliable, especially for user-defined
1563 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php
1564 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php
1565 http://archives.postgresql.org/pgsql-patches/2007-03/msg00508.php
1566 http://archives.postgresql.org/pgsql-patches/2007-04/msg00347.php
1567 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00156.php
1568 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00546.php
1569 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00416.php
1571 * Allow FSM to return free space toward the beginning of the heap file,
1572 in hopes that empty pages at the end can be truncated by VACUUM
1573 * Allow FSM page return free space based on table clustering, to assist
1574 in maintaining clustering?
1575 * -Improve dead row detection during multi-statement transactions usage
1577 http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php
1579 * Consider a more compact data representation for dead tuples
1581 http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php
1586 o %Issue log message to suggest VACUUM FULL if a table is nearly
1588 o Improve control of auto-vacuum
1590 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php
1592 o Prevent long-lived temporary tables from causing frozen-xid
1593 advancement starvation
1595 The problem is that autovacuum cannot vacuum them to set frozen xids;
1596 only the session that created them can do that.
1597 http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php
1599 o Store per-table autovacuum settings in pg_class.reloptions.
1601 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php
1602 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php
1604 o Prevent autovacuum from running if an old transaction is still
1605 running from the last vacuum
1607 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php
1614 * Fix priority ordering of read and write light-weight locks (Neil)
1616 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php
1617 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php
1619 * Fix problem when multiple subtransactions of the same outer transaction
1620 hold different types of locks, and one subtransaction aborts
1622 http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php
1623 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php
1624 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php
1625 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php
1627 * Allow UPDATEs on only non-referential integrity columns not to conflict
1628 with referential integrity locks
1630 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php
1632 * Add idle_in_transaction_timeout GUC so locks are not held for long
1635 * Improve deadlock detection when a page cleaning lock conflicts
1636 with a shared buffer that is pinned
1638 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php
1639 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1640 http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php
1642 * Detect deadlocks involving LockBufferForCleanup()
1644 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1648 Startup Time Improvements
1649 =========================
1651 * Experiment with multi-threaded backend for backend creation
1653 This would prevent the overhead associated with process creation. Most
1654 operating systems have trivial process creation time compared to
1655 database startup overhead, but a few operating systems (Win32,
1656 Solaris) might benefit from threading. Also explore the idea of
1657 a single session using multiple threads to execute a statement faster.
1664 * Eliminate need to write full pages to WAL before page modification
1666 Currently, to protect against partial disk page writes, we write
1667 full page images to WAL before they are modified so we can correct any
1668 partial page writes during recovery. These pages can also be
1669 eliminated from point-in-time archive files.
1670 http://archives.postgresql.org/pgsql-hackers/2002-06/msg00655.php
1672 o When off, write CRC to WAL and check file system blocks
1675 If CRC check fails during recovery, remember the page in case
1676 a later CRC for that page properly matches.
1678 o Write full pages during file system write and not when
1679 the page is modified in the buffer cache
1681 This allows most full page writes to happen in the background
1682 writer. It might cause problems for applying WAL on recovery
1683 into a partially-written page, but later the full page will be
1686 * Allow WAL traffic to be streamed to another server for stand-by
1688 * Reduce WAL traffic so only modified values are written rather than
1691 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php
1693 * Allow WAL information to recover corrupted pg_controldata
1695 http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php
1697 * Find a way to reduce rotational delay when repeatedly writing
1700 Currently fsync of WAL requires the disk platter to perform a full
1701 rotation to fsync again. One idea is to write the WAL to different
1702 offsets that might reduce the rotational delay.
1703 http://archives.postgresql.org/pgsql-hackers/2002-11/msg00483.php
1705 * Allow WAL logging to be turned off for a table, but the table
1706 might be dropped or truncated during crash recovery
1708 Allow tables to bypass WAL writes and just fsync() dirty pages on
1709 commit. This should be implemented using ALTER TABLE, e.g. ALTER
1710 TABLE PERSISTENCE [ DROP | TRUNCATE | DEFAULT ]. Tables using
1711 non-default logging should not use referential integrity with
1712 default-logging tables. A table without dirty buffers during a
1713 crash could perhaps avoid the drop/truncate.
1714 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1716 * Allow WAL logging to be turned off for a table, but the table would
1717 avoid being truncated/dropped
1719 To do this, only a single writer can modify the table, and writes
1720 must happen only on new pages so the new pages can be removed during
1721 crash recovery. Readers can continue accessing the table. Such
1722 tables probably cannot have indexes. One complexity is the handling
1723 of indexes on TOAST tables.
1724 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1726 * Speed WAL recovery by allowing more than one page to be prefetched
1728 This should be done utilizing the same infrastructure used for
1729 prefetching in general to avoid introducing complex error-prone code
1731 http://archives.postgresql.org/pgsql-general/2007-12/msg00683.php
1732 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00497.php
1733 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php
1735 * Improve WAL concurrency by increasing lock granularity
1737 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00556.php
1739 * Be more aggressive about creating WAL files
1741 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01325.php
1742 http://archives.postgresql.org/pgsql-hackers/2004-07/msg01075.php
1743 http://archives.postgresql.org/pgsql-hackers/2005-04/msg00556.php
1745 * Have resource managers report the duration of their status changes
1747 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01468.php
1749 * Move pgfoundry's xlogdump to /contrib and have it rely more closely
1750 on the WAL backend code
1752 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00035.php
1757 Optimizer / Executor
1758 ====================
1760 * Improve selectivity functions for geometric operators
1761 * Precompile SQL functions to avoid overhead
1762 * Create utility to compute accurate random_page_cost value
1763 * Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
1764 * Have EXPLAIN ANALYZE issue NOTICE messages when the estimated and
1765 actual row counts differ by a specified percentage
1766 * Improve how ANALYZE computes in-doubt tuples
1768 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00771.php
1770 * Log statements where the optimizer row estimates were dramatically
1771 different from the number of rows actually found?
1772 * Consider compressed annealing to search for query plans
1774 This might replace GEQO, http://sixdemonbag.org/Djinni.
1776 * Improve merge join performance by allowing mark/restore of
1779 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php
1781 * Consider using a hash for joining to a large IN (VALUES ...) list
1783 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php
1790 * Consider having the background writer update the transaction status
1791 hint bits before writing out the page
1793 Implementing this requires the background writer to have access to system
1794 catalogs and the transaction status log.
1796 * Consider adding buffers the background writer finds reusable to the
1799 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1801 * Automatically tune bgwriter_delay based on activity rather then using a
1804 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1806 * Consider wither increasing BM_MAX_USAGE_COUNT improves performance
1808 http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php
1810 * Test to see if calling PreallocXlogFiles() from the background writer
1811 will help with WAL segment creation latency
1813 http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php
1817 Miscellaneous Performance
1818 =========================
1820 * Do async I/O for faster random read-ahead of data
1822 Async I/O allows multiple I/O requests to be sent to the disk with
1823 results coming back asynchronously.
1825 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php
1826 http://archives.postgresql.org/pgsql-performance/2007-09/msg00255.php
1827 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php
1828 http://archives.postgresql.org/pgsql-patches/2008-01/msg00170.php
1830 * Use mmap() rather than SYSV shared memory or to write WAL files?
1832 This would remove the requirement for SYSV SHM but would introduce
1833 portability issues. Anonymous mmap (or mmap to /dev/zero) is required
1834 to prevent I/O overhead.
1836 * Consider mmap()'ing files into a backend?
1838 Doing I/O to large tables would consume a lot of address space or
1839 require frequent mapping/unmapping. Extending the file also causes
1840 mapping problems that might require mapping only individual pages,
1841 leading to thousands of mappings. Another problem is that there is no
1842 way to _prevent_ I/O to disk from the dirty shared buffers so changes
1843 could hit disk before WAL is written.
1845 * Add a script to ask system configuration questions and tune postgresql.conf
1846 * Consider ways of storing rows more compactly on disk
1848 o Reduce the row header size?
1849 o Consider reducing on-disk varlena length from four bytes to
1850 two because a heap row cannot be more than 64k in length
1852 * Consider increasing NUM_CLOG_BUFFERS
1854 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php
1855 http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php
1857 * Consider transaction start/end performance improvements
1859 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00948.php
1860 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1862 * Allow user configuration of TOAST thresholds
1864 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php
1865 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php
1867 * Allow configuration of backend priorities via the operating system
1869 Though backend priorities make priority inversion during lock
1870 waits possible, research shows that this is not a huge problem.
1872 http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php
1874 * Experiment with multi-threaded backend better I/O utilization
1876 This would allow a single query to make use of multiple I/O channels
1877 simultaneously. One idea is to create a background reader that can
1878 pre-fetch sequential and index scan pages needed by other backends.
1879 This could be expanded to allow concurrent reads from multiple devices
1880 in a partitioned table.
1882 * Experiment with multi-threaded backend better CPU utilization
1884 This would allow several CPUs to be used for a single query, such as
1885 for sorting or query execution.
1887 * Consider increasing the minimum allowed number of shared buffers
1889 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php
1891 * Expire published xmin for read-only and idle transactions
1893 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00343.php
1895 * Consider if CommandCounterIncrement() can avoid its
1896 AcceptInvalidationMessages() call
1898 http://archives.postgresql.org/pgsql-committers/2007-11/msg00585.php
1900 * Improve performance of shared invalidation queue for multiple CPUs
1902 http://archives.postgresql.org/pgsql-performance/2008-01/msg00023.php
1904 * Consider Cartesian joins when both relations are needed to form an
1905 indexscan qualification for a third relation
1907 http://archives.postgresql.org/pgsql-performance/2007-12/msg00090.php
1909 * Consider not storing a NULL bitmap on disk if all the NULLs are
1912 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00624.php
1913 http://archives.postgresql.org/pgsql-patches/2007-12/msg00109.php
1915 * Sort large UPDATE/DELETEs so it is done in heap order
1917 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01119.php
1919 * SMP scalability improvements
1921 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00439.php
1922 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00206.php
1923 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1925 * Research reducing deTOASTing in more places
1927 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00895.php
1929 * Allow one transaction to see tuples using the snapshot of another
1932 This would assist multiple backends in working together.
1933 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00400.php
1935 * Consider decreasing the I/O caused by updating tuple hint bits
1937 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00847.php
1944 * Add use of 'const' for variables in source tree
1945 * Move some things from contrib into main tree
1946 * %Remove warnings created by -Wcast-align
1947 * Move platform-specific ps status display info from ps_status.c to ports
1948 * Add optional CRC checksum to heap and index pages
1949 * Improve documentation to build only interfaces (Marc)
1950 * Remove or relicense modules that are not under the BSD license, if possible
1951 * Acquire lock on a relation before building a relcache entry for it
1952 * Allow cross-compiling by generating the zic database on the target system
1953 * Improve NLS maintenance of libpgport messages linked onto applications
1954 * Clean up casting in contrib/isn
1956 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php
1958 * Improve the /contrib installation experience
1960 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00132.php
1962 * Use UTF8 encoding for NLS messages so all server encodings can
1964 * Update Bonjour to work with newer cross-platform SDK
1966 http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php
1967 http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php
1969 * Consider GnuTLS if OpenSSL license becomes a problem
1971 http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php
1972 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php
1974 * Consider changing documentation format from SGML to XML
1976 http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php
1978 * Consider making NAMEDATALEN more configurable in future releases
1979 * Update our code to handle 64-bit timezone files to match the zic
1980 source code, which now uses them
1981 * Have configure choose integer datetimes by default
1983 http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php
1985 * Support scoped IPv6 addresses
1987 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php
1989 * Consider allowing 64-bit integers and floats to be passed by value on
1992 Also change 32-bit floats (float4) to be passed by value at the same
1995 * Research use of signals and sleep wake ups
1997 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php
1999 * Add automated check for invalid C++ source code constructs
2001 http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php
2003 * Consider simplifying how memory context resets handle child contexts
2005 http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php
2007 * Remove use of MAKE_PTR and MAKE_OFFSET macros
2009 http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php
2011 * Convert single quotes to apostrophes in the PDF documentation
2013 http://archives.postgresql.org/pgsql-docs/2007-12/msg00059.php
2015 * Create three versions of libpgport to simplify client code
2017 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00154.php
2019 * Remove old-style routines for manipulating tuples
2021 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00851.php
2023 * Improve detection of shared memory segments being used by others
2024 by checking the SysV shared memory field 'nattch'
2026 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00656.php
2027 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00673.php
2029 * Implement the non-threaded Avahi service discovery protocol
2031 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00939.php
2032 http://archives.postgresql.org/pgsql-patches/2008-02/msg00097.php
2033 http://archives.postgresql.org/pgsql-hackers/2008-03/msg01211.php
2034 http://archives.postgresql.org/pgsql-patches/2008-04/msg00001.php
2036 * Implement a module capability for loading /contrib-style extensions
2038 http://archives.postgresql.org/pgsql-patches/2008-04/msg00164.php
2040 * Fix system views like pg_stat_all_tables to use set-returning
2041 functions, rather than views of per-column functions
2046 o Remove configure.in check for link failure when cause is found
2047 o Remove readdir() errno patch when runtime/mingwex/dirent.c rev
2049 o -Remove psql newline patch when we find out why mingw outputs an
2051 o Allow psql to use readline once non-US code pages work with
2053 o Fix problem with shared memory on the Win32 Terminal Server
2054 o Diagnose problem where shared memory can sometimes not be
2055 attached by postmaster children
2057 http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php
2059 o Improve signal handling
2061 http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php
2063 o Convert MSVC build system to remove most batch files
2065 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php
2067 o Prevent SSL from sending network packets to avoid interference
2068 with Win32 signal emulation
2070 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00455.php
2072 o Support pgxs when using MSVC
2074 o Fix MSVC NLS support, like for to_char()
2076 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00485.php
2077 http://archives.postgresql.org/pgsql-patches/2008-02/msg00038.php
2079 o Find a correct rint() substitute on Windows
2081 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00808.php
2085 * Wire Protocol Changes
2087 o Allow dynamic character set handling
2088 o Add decoded type, length, precision
2090 o Update clients to use data types, typmod, schema.table.column names
2091 of result sets using new statement protocol
2097 * Add pre-parsing phase that converts non-ISO syntax to supported
2100 This could allow SQL written for other databases to run without
2103 * Allow plug-in modules to emulate features from other databases
2104 * Add features of Oracle-style packages (Pavel)
2106 A package would be a schema with session-local variables,
2107 public/private functions, and initialization functions. It
2108 is also possible to implement these capabilities
2109 in any schema and not use a separate "packages"
2112 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php
2114 * Consider allowing control of upper/lower case folding of unquoted
2117 http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php
2118 http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php
2119 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00849.php
2120 http://archives.postgresql.org/pgsql-hackers/2008-07/msg00415.php
2122 * Add autonomous transactions
2124 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php
2128 Features We Do _Not_ Want
2129 =========================
2131 * All backends running as threads in a single process (not wanted)
2133 This eliminates the process protection we get from the current setup.
2134 Thread creation is usually the same overhead as process creation on
2135 modern systems, so it seems unwise to use a pure threaded model.
2137 * Optimizer hints (not wanted)
2139 Optimizer hints are used to work around problems in the optimizer. We
2140 would rather have the problems reported and fixed.
2142 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php
2143 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php
2144 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php
2147 Because we support postfix operators, it isn't possible to make AS
2148 optional and continue to use bison.
2149 http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php
2151 http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php
2153 * Embedded server (not wanted)
2155 While PostgreSQL clients runs fine in limited-resource environments, the
2156 server requires multiple processes and a stable pool of resources to
2157 run reliabily and efficiently. Stripping down the PostgreSQL server
2158 to run in the same process address space as the client application
2159 would add too much complexity and failure cases.