3 Current maintainer: Bruce Momjian (bruce@momjian.us)
4 Last updated: Mon Aug 11 23:48:39 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 SSL 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 * Add 'hostgss' pg_hba.conf option to allow GSS link-level encryption
94 http://archives.postgresql.org/pgsql-hackers/2008-07/msg01454.php
96 * Improve server security options
98 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01875.php
99 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00000.php
101 * Configuration files
103 o Allow pg_hba.conf to specify host names along with IP addresses
105 Host name lookup could occur when the postmaster reads the
106 pg_hba.conf file, or when the backend starts. Another
107 solution would be to reverse lookup the connection IP and
108 check that hostname against the host names in pg_hba.conf.
109 We could also then check that the host name maps to the IP
111 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00569.php
113 o %Allow postgresql.conf file values to be changed via an SQL
114 API, perhaps using SET GLOBAL
115 o Allow the server to be stopped/restarted via an SQL API
116 o Issue a warning if a change-on-restart-only postgresql.conf value
117 is modified and the server config files are reloaded
118 o Consider normalizing fractions in postgresql.conf, perhaps
121 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00550.php
123 o Allow Kerberos to disable stripping of realms so we can
124 check the username@realm against multiple realms
126 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00009.php
128 o Add functions to syntax check configuration files
130 o Improve LDAP authentication configuration options
132 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01745.php
134 o Add external tool to auto-tune some postgresql.conf parameters
136 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00000.php
142 o Allow a database in tablespace t1 with tables created in
143 tablespace t2 to be used as a template for a new database created
144 with default tablespace t2
146 Currently all objects in the default database tablespace must
147 have default tablespace specifications. This is because new
148 databases are created by copying directories. If you mix default
149 tablespace tables and tablespace-specified tables in the same
150 directory, creating a new database from such a mixed directory
151 would create a new database with tables that had incorrect
152 explicit tablespaces. To fix this would require modifying
153 pg_class in the newly copied database, which we don't currently
156 o Allow reporting of which objects are in which tablespaces
158 This item is difficult because a tablespace can contain objects
159 from multiple databases. There is a server-side function that
160 returns the databases which use a specific tablespace, so this
161 requires a tool that will call that function and connect to each
162 database to find the objects in each database for that tablespace.
164 o Allow WAL replay of CREATE TABLESPACE to work when the directory
165 structure on the recovery computer is different from the original
167 o Allow per-tablespace quotas
170 * Statistics Collector
172 o Allow statistics collector information to be pulled from the collector
173 process directly, rather than requiring the collector to write a
174 filesystem file twice a second?
175 o Reduce file system activity overhead of statistics file pgstat.stat
177 http://archives.postgresql.org/pgsql-general/2007-12/msg00106.php
179 o Allow statistics last vacuum/analyze execution times to be displayed
180 without requiring stats_row_level to be enabled
182 http://archives.postgresql.org/pgsql-docs/2007-04/msg00028.php
184 o Clear table counters on TRUNCATE
186 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00169.php
189 * Point-In-Time Recovery (PITR)
191 o Allow a warm standby system to also allow read-only statements
193 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00050.php
195 o %Create dump tool for write-ahead logs for use in determining
196 transaction id for point-in-time recovery
198 This is useful for checking PITR recovery.
200 o Allow recovery.conf to support the same syntax as
201 postgresql.conf, including quoting
203 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php
205 o -Fix server restart problem when the server was shutdown during
208 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00800.php
210 o Recreate pg_xlog/archive_status/ if it doesn't exist after
211 restoring from a PITR backup
213 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00487.php
215 o Reduce PITR WAL file size by removing full page writes and
216 by removing trailing bytes to improve compression
222 * Change NUMERIC to enforce the maximum precision
223 * Reduce storage space for small NUMERICs
225 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php
226 http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php
227 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php
229 * Fix data types where equality comparison isn't intuitive, e.g. box
230 * Add support for public SYNONYMs
232 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php
234 * Fix CREATE CAST on DOMAINs
236 http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php
237 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php
239 * Allow domains to be cast
241 http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php
242 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php
244 * Add support for SQL-standard GENERATED/IDENTITY columns
246 http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php
247 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php
248 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php
249 http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php
250 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php
252 * Improve XML support
254 http://developer.postgresql.org/index.php/XML_Support
256 * Consider placing all sequences in a single table, or create a system
259 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
261 * Allow the UUID type to accept non-standard formats
263 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php
265 * Consider a special data type for regular expressions
267 http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php
269 * Reduce BIT data type overhead using short varlena headers
271 http://archives.postgresql.org/pgsql-general/2007-12/msg00273.php
273 * Allow xml arrays to be cast to other data types
275 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00981.php
276 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00231.php
277 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00471.php
279 * Simplify integer cross-data-type operators
281 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00189.php
283 * Allow XML to accept more liberal DOCTYPE specifications
285 http://archives.postgresql.org/pgsql-general/2008-02/msg00347.php
287 * Allow adding/renaming/removing enumerated values to an existing
290 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01718.php
294 o Allow infinite dates and intervals just like infinite timestamps
295 o Merge hardwired timezone names with the TZ database; allow either
296 kind everywhere a TZ name is currently taken
297 o Allow TIMESTAMP WITH TIME ZONE to store the original timezone
298 information, either zone name or offset from UTC
300 If the TIMESTAMP value is stored with a time zone name, interval
301 computations should adjust based on the time zone rules.
302 http://archives.postgresql.org/pgsql-hackers/2004-10/msg00705.php
304 o Fix SELECT '0.01 years'::interval, '0.01 months'::interval
305 o Add a GUC variable to allow output of interval values in ISO8601
307 o Have timestamp subtraction not call justify_hours()?
309 http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php
311 o Improve timestamptz subtraction to be DST-aware
313 Currently subtracting one date from another that crosses a
314 daylight savings time adjustment can return '1 day 1 hour', but
315 adding that back to the first date returns a time one hour in
316 the future. This is caused by the adjustment of '25 hours' to
317 '1 day 1 hour', and '1 day' is the same time the next day, even
318 if daylight savings adjustments are involved.
320 o Fix interval display to support values exceeding 2^31 hours
321 o Add overflow checking to timestamp and interval arithmetic
322 o Extend timezone code to allow 64-bit values so we can
323 represent years beyond 2038
325 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php
327 o -Use LC_TIME for localized weekday/month names, rather than
330 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php
332 o Add ISO INTERVAL handling
334 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php
335 http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php
337 o Support ISO INTERVAL syntax if units cannot be determined from
338 the string, and are supplied after the string
340 The SQL standard states that the units after the string
341 specify the units of the string, e.g. INTERVAL '2' MINUTE
342 should return '00:02:00'. The current behavior has the units
343 restrict the interval value to the specified unit or unit
344 range, INTERVAL '70' SECOND returns '00:00:10'.
346 For syntax that isn't uniquely ISO or PG syntax, like '1' or
347 '1:30', treat as ISO if there is a range specification clause,
348 and as PG if there no clause is present, e.g. interpret '1:30'
349 MINUTE TO SECOND as '1 minute 30 seconds', and interpret
350 '1:30' as '1 hour, 30 minutes'.
352 This makes common cases like SELECT INTERVAL '1' MONTH
353 SQL-standard results. The SQL standard supports a limited
354 number of unit combinations and doesn't support unit names in
355 the string. The PostgreSQL syntax is more flexible in the
356 range of units supported, e.g. PostgreSQL supports '1 year 1
357 hour', while the SQL standard does not.
359 o Add support for year-month syntax, INTERVAL '50-6' YEAR
361 o Interpret INTERVAL '1 year' MONTH as CAST (INTERVAL '1
362 year' AS INTERVAL MONTH), and this should return '12 months'
363 o Round or truncate values to the requested precision, e.g.
364 INTERVAL '11 months' AS YEAR should return one or zero
365 o Support precision, CREATE TABLE foo (a INTERVAL MONTH(3))
370 o Delay resolution of array expression's data type so assignment
371 coercion can be performed on empty array expressions
372 o Add support for arrays of domains
374 http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php
376 o Allow single-byte header storage for array elements
381 o Improve vacuum of large objects, like contrib/vacuumlo?
382 o Add security checking for large objects
383 o Auto-delete large objects when referencing row is deleted
385 contrib/lo offers this functionality.
387 o Allow read/write into TOAST values like large objects
389 This requires the TOAST column to be stored EXTERNAL.
391 o Add API for 64-bit large object access
393 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php
397 * Add locale-aware MONEY type, and support multiple currencies
399 http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
400 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
402 * MONEY dumps in a locale-specific format making it difficult to
403 restore to a system with a different locale
404 * Allow MONEY to be easily cast to/from other numeric data types
408 o Allow dictionaries to change the token that is passed on to
411 http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php
413 o Consider a function-based API for '@@' searches
415 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php
417 o Improve text search error messages
419 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00966.php
420 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01146.php
422 o Consider changing error to warning for strings larger than one
425 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00190.php
426 http://archives.postgresql.org/pgsql-patches/2008-03/msg00062.php
433 * Allow INET subnet tests using non-constants to be indexed
434 * Allow to_date() and to_timestamp() accept localized month names
435 * Fix to_date()-related functions to consistently issue errors
437 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php
439 * Add missing parameter handling in to_char()
441 http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php
443 * Allow substring/replace() to get/set bit values
444 * Allow to_char() on interval values to accumulate the highest unit
447 Some special format flag would be required to request such
448 accumulation. Such functionality could also be added to EXTRACT.
449 Prevent accumulation that crosses the month/day boundary because of
450 the uneven number of days in a month.
452 o to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
453 o to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
454 o to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
455 o to_char(INTERVAL '3 years 5 months','MM') => 41
457 * Implement inlining of set-returning functions defined in SQL
458 * Allow SQL-language functions to return results from RETURNING queries
460 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php
462 * Allow SQL-language functions to reference parameters by parameter name
464 Currently SQL-language functions can only refer to dollar parameters,
467 * Add SPI_gettypmod() to return the typemod for a TupleDesc
468 * Enforce typmod for function inputs, function results and parameters for
469 spi_prepare'd statements called from PLs
471 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php
473 * Allow holdable cursors in SPI
474 * Tighten function permission checks
476 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php
478 * Fix IS OF so it matches the ISO specification, and add documentation
480 http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php
481 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php
483 * Add missing operators for geometric data types
485 Some geometric types do not have the full suite of geometric operators,
488 * Implement Boyer-Moore searching in strpos()
490 http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php
492 * Prevent malicious functions from being executed with the permissions
493 of unsuspecting users
495 Index functions are safe, so VACUUM and ANALYZE are safe too.
496 Triggers, CHECK and DEFAULT expressions, and rules are still vulnerable.
497 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php
499 * Reduce memory usage of aggregates in set returning functions
501 http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php
503 * -Add temporal versions of generate_series()
505 http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php
507 * Add array_accum() and array_to_set() functions for arrays
509 The standards specify array_agg() and UNNEST.
510 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php
512 * Fix /contrib/ltree operator
514 http://archives.postgresql.org/pgsql-bugs/2007-11/msg00044.php
516 * Fix inconsistent precedence of =, >, and < compared to <>, >=, and <=
518 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00145.php
520 * Fix regular expression bug when using complex back-references
522 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00000.php
524 * Have /contrib/dblink reuse unnamed connections
526 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00895.php
528 * Add SQL-standard array_agg() and unnest() array functions
530 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01017.php
532 * Allow calling of a procedure outside a SELECT that can control the
535 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01375.php
539 Multi-Language Support
540 ======================
542 * Add NCHAR (as distinguished from ordinary varchar),
543 * Allow locale to be set at database creation
545 Currently locale can only be set during initdb. No global tables have
546 locale-aware columns. However, the database template used during
547 database creation might have locale-aware indexes. The indexes would
548 need to be reindexed to match the new locale.
550 * Allow encoding on a per-column basis optionally using the ICU library;
553 Right now only one encoding is allowed per database.
555 http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php
556 http://archives.postgresql.org/pgsql-patches/2005-08/msg00039.php
557 http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php
558 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00110.php
559 http://archives.postgresql.org/pgsql-patches/2005-09/msg00020.php
560 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01121.php
561 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00767.php
562 http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php
563 http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php
564 http://wiki.postgresql.org/wiki/Todo:Collate
565 http://wiki.postgresql.org/wiki/Todo:ICU
567 * Support multiple simultaneous character sets, per SQL92
568 * Improve UTF8 combined character handling?
569 * Add octet_length_server() and octet_length_client()
570 * Make octet_length_client() the same as octet_length()?
571 * Fix problems with wrong runtime encoding conversion for NLS message files
572 * Add URL to more complete multi-byte regression tests
574 http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php
576 * Fix ILIKE and regular expressions to handle case insensitivity
577 properly in multibyte encodings
579 http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php
580 http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php
582 * Set client encoding based on the client operating system encoding
584 Currently client_encoding is set in postgresql.conf, which
585 defaults to the server encoding.
586 http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php
588 * Change memory allocation for multi-byte functions so memory is
589 allocated inside conversion functions
591 Currently we preallocate memory based on worst-case usage.
598 * Automatically create rules on views so they are updateable, per SQL99
600 We can only auto-create rules for simple views. For more complex
601 cases users will still have to write rules manually.
603 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php
604 http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php
606 * Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
607 * Allow VIEW/RULE recompilation when the underlying tables change
609 Another issue is whether underlying table changes should be reflected
610 in the view, e.g. should SELECT * show additional columns if they
611 are added after the view is created.
613 * Make it possible to use RETURNING together with conditional DO INSTEAD
614 rules, such as for partitioning setups
616 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php
618 * Add the ability to automatically create materialized views
620 Right now materialized views require the user to create triggers on the
621 main table to keep the summary table current. SQL syntax should be able
622 to manager the triggers and summary table automatically. A more
623 sophisticated implementation would automatically retrieve from the
624 summary table when the main table is referenced, if possible.
626 * Improve ability to modify views via ALTER TABLE
628 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00691.php
635 * Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
636 * Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
637 * %Allow SET CONSTRAINTS to be qualified by schema/table name
638 * %Add a separate TRUNCATE permission
640 Currently only the owner can TRUNCATE a table because triggers are not
641 called, and the table is locked in exclusive mode.
643 * Fix TRUNCATE ... RESTART IDENTITY so its affect on sequences is rolled
644 back on transaction abort
645 * Allow PREPARE of cursors
646 * Allow finer control over the caching of prepared query plans
648 Currently queries prepared via the libpq API are planned on first
649 execute using the supplied parameters --- allow SQL PREPARE to do the
650 same. Also, allow control over replanning prepared queries either
651 manually or automatically when statistics for execute parameters
652 differ dramatically from those used during planning.
654 * Improve logging of prepared transactions recovered during startup
656 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php
658 * Improve failure message when DROP DATABASE is used on a database that
659 has prepared transactions
660 * Allow prepared transactions with temporary tables created and dropped
661 in the same transaction, and when an ON COMMIT DELETE ROWS temporary
664 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00047.php
666 * Add a GUC variable to warn about non-standard SQL usage in queries
667 * Add SQL-standard MERGE/REPLACE/UPSERT command
669 MERGE is typically used to merge two tables. REPLACE or UPSERT
670 command does UPDATE, or on failure, INSERT. This is similar to UPDATE,
671 then for unmatched rows, INSERT. Whether concurrent access allows
672 modifications which could cause row loss is implementation independent.
673 To implement this cleanly requires that the table have a unique index
674 so duplicate checking can be easily performed. It is possible to do it
675 without a unique index if we require the user to LOCK the table before
678 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php
679 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php
680 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01157.php
681 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01475.php
682 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01890.php
684 * Add NOVICE output level for helpful messages like automatic sequence/index
686 * Add GUC to issue notice about statements that use unjoined tables
687 * Allow EXPLAIN to identify tables that were skipped because of
689 * Allow EXPLAIN output to be more easily processed by scripts, perhaps XML
690 * Enable standard_conforming_strings
691 * Make standard_conforming_strings the default in 8.5?
693 When this is done, backslash-quote should be prohibited in non-E''
694 strings because of possible confusion over how such strings treat
695 backslashes. Basically, '' is always safe for a literal single
696 quote, while \' might or might not be based on the backslash
699 * Simplify dropping roles that have objects in several databases
700 * Allow COMMENT ON to accept an expression rather than just a string
701 * Allow the count returned by SELECT, etc to be represented as an int64
702 to allow a higher range of values
703 * Add SQL99 WITH clause to SELECT
704 * Add SQL:2003 WITH RECURSIVE (hierarchical) queries to SELECT
706 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php
707 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php
708 http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php
709 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01334.php
710 http://archives.postgresql.org/pgsql-patches/2008-01/msg00105.php
711 http://archives.postgresql.org/pgsql-patches/2008-03/msg00327.php
713 * Add DEFAULT .. AS OWNER so permission checks are done as the table
716 This would be useful for SERIAL nextval() calls and CHECK constraints.
718 * Allow DISTINCT to work in multiple-argument aggregate calls
719 * Add column to pg_stat_activity that shows the progress of long-running
720 commands like CREATE INDEX and VACUUM
722 http://archives.postgresql.org/pgsql-patches/2008-04/msg00203.php
724 * Implement SQL:2003 window functions
726 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00380.php
727 http://archives.postgresql.org/pgsql-hackers/2008-07/msg00232.php
729 * Allow INSERT/UPDATE ... RETURNING inside a SELECT 'FROM' clause or
732 http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php
733 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php
734 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00124.php
736 * Increase locking when DROPing objects so dependent objects cannot
737 get dropped while the DROP operation is happening
739 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php
741 * -Allow AS in "SELECT col AS label" to be optional in certain cases
743 * Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
744 RETURNING to supply values to the INSERT
745 http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979
747 * Add comments on system tables/columns using the information in
750 Ideally the information would be pulled from the SGML file
753 * Improve reporting of UNION type mismatches
755 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00944.php
756 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00597.php
761 o Allow CREATE TABLE AS to determine column lengths for complex
762 expressions like SELECT col1 || col2
764 o Have WITH CONSTRAINTS also create constraint indexes
766 http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php
768 o Have CONSTRAINT cname NOT NULL record the contraint name
770 Right now pg_attribute.attnotnull records the NOT NULL status
771 of the column, but does not record the contraint name
773 o Prevent concurrent CREATE TABLE table1 from sometimes returning
774 a cryptic error message
776 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00169.php
778 o Add CREATE SCHEMA ... LIKE that copies a schema
783 o Allow UPDATE tab SET ROW (col, ...) = (SELECT...)
785 http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php
786 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php
787 http://archives.postgresql.org/pgsql-patches/2007-04/msg00315.php
788 http://archives.postgresql.org/pgsql-patches/2008-03/msg00237.php
790 o Research self-referential UPDATEs that see inconsistent row versions
791 in read-committed mode
793 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php
794 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php
796 o Allow GLOBAL temporary tables to exist as empty by default in
799 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php
804 o Have ALTER TABLE RENAME rename SERIAL sequence names
806 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
808 o Have ALTER SEQUENCE RENAME rename the sequence name stored
809 in the sequence table
811 http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php
812 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php
813 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
815 o Add ALTER DOMAIN to modify the underlying data type
816 o %Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
818 http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php
820 o %Allow ALTER TABLE to change constraint deferrability and actions
821 o Add missing object types for ALTER ... SET SCHEMA
822 o Allow ALTER TABLESPACE to move to different directories
823 o Allow databases to be moved to different tablespaces
824 o Allow moving system tables to other tablespaces, where possible
826 Currently non-global system tables must be in the default database
827 tablespace. Global system tables can never be moved.
829 o -Prevent parent tables from altering or dropping constraints
830 like CHECK that are inherited by child tables unless CASCADE
832 o -Prevent child tables from altering or dropping constraints
833 like CHECK that were inherited from the parent table
834 o Have ALTER INDEX update the name of a constraint using that index
835 o Add ALTER TABLE RENAME CONSTRAINT, update index name also
836 o Allow column display reordering by recording a display,
837 storage, and permanent id for every column?
839 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php
841 o Allow an existing index to be marked as a table's primary key
843 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00500.php
848 o Automatically maintain clustering on a table
850 This might require some background daemon to maintain clustering
851 during periods of low usage. It might also require tables to be only
852 partially filled for easier reorganization. Another idea would
853 be to create a merged heap/index data file so an index lookup would
854 automatically access the heap data too. A third idea would be to
855 store heap rows in hashed groups, perhaps using a user-supplied
857 http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php
859 o %Add default clustering to system tables
861 To do this, determine the ideal cluster index for each system
862 table and set the cluster setting during initdb.
864 o %Add VERBOSE option to report tables as they are processed,
870 o Allow COPY to report error lines and continue
872 This requires the use of a savepoint before each COPY line is
873 processed, with ROLLBACK on COPY failure.
874 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00572.php
876 o Allow COPY on a newly-created table to skip WAL logging
878 On crash recovery, the table involved in the COPY would
879 be removed or have its heap and index files truncated. One
880 issue is that no other backend should be able to add to
881 the table at the same time, which is something that is
882 currently allowed. This currently is done if the table is
883 created inside the same transaction block as the COPY because
884 no other backends can see the table.
886 o Consider using a ring buffer for COPY FROM
888 http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php
889 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php
891 o Allow COPY FROM to create index entries in bulk
893 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php
895 o Allow COPY in CSV mode to control whether a quoted zero-length
896 string is treated as NULL
898 Currently this is always treated as a zero-length string,
899 which generates an error when loading into an integer column
900 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php
902 o Impove COPY performance
904 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00954.php
906 o Allow COPY to report errors sooner
908 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01169.php
914 o Allow column-level privileges
915 o %Allow GRANT/REVOKE permissions to be applied to all schema objects
918 The proposed syntax is:
919 GRANT SELECT ON ALL TABLES IN public TO phpuser;
920 GRANT SELECT ON NEW TABLES IN public TO phpuser;
922 o Allow GRANT/REVOKE permissions to be inherited by objects based on
925 o Allow SERIAL sequences to inherit permissions from the base table?
930 o Prevent DROP TABLE from dropping a row referenced by its own open
936 o Allow INSERT/UPDATE of the system-generated oid value for a row
937 o In rules, allow VALUES() to contain a mixture of 'old' and 'new'
943 o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
949 o Allow LISTEN/NOTIFY to store info in memory rather than tables?
951 Currently LISTEN/NOTIFY information is stored in pg_listener.
952 Storing such information in memory would improve performance.
954 o Add optional textual message to NOTIFY
956 This would allow an informational message to be added to the notify
957 message, perhaps indicating the row modified or other custom
960 o Allow multiple identical NOTIFY events to always be communicated
961 to the client, rather than sent as a single notification to the
964 http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php
966 o Allow NOTIFY in rules involving conditionals
967 o Improve LISTEN concurrency
969 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01106.php
973 Referential Integrity
974 =====================
976 * Add MATCH PARTIAL referential integrity
977 * Change foreign key constraint for array -> element to mean element
979 * Fix problem when cascading referential triggers make changes on
980 cascaded tables, seeing the tables in an intermediate state
982 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
983 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
985 * Allow DEFERRABLE and end-of-statement UNIQUE constraints?
987 This would allow UPDATE tab SET col = col + 1 to work if col has
988 a unique index. Currently, uniqueness checks are done while the
989 command is being executed, rather than at the end of the statement
991 http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html
992 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php
994 * Optimize referential integrity checks
996 http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php
997 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php
1000 Server-Side Languages
1001 =====================
1004 o Fix RENAME to work on variables other than OLD/NEW
1006 http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php
1007 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php
1008 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php
1010 o Allow function parameters to be passed by name,
1011 get_employee_salary(12345 AS emp_id, 2001 AS tax_year)
1012 o Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
1013 o Allow listing of record column names, and access to
1014 record columns via variables, e.g. columns := r.(*),
1015 tval2 := r.(colname)
1017 http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php
1018 http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php
1019 http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php
1021 o Add support for SCROLL cursors
1022 o Add support for WITH HOLD cursors
1023 o Allow row and record variables to be set to NULL constants,
1024 and allow NULL tests on such variables
1026 Because a row is not scalar, do not allow assignment
1027 from NULL-valued scalars.
1029 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php
1031 o Review handling of MOVE and FETCH
1033 http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php
1035 o Improve logic of determining if an identifier is a a
1036 variable or column name
1038 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php
1040 o Consider keeping seperate cached copies when search_path changes
1042 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01009.php
1044 o -Add CASE capability to language (already in SQL)
1046 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00696.php
1052 o Add table function support to pltcl, plpythonu
1053 o Add support for polymorphic arguments and return types to
1054 languages other than PL/PgSQL
1055 o Add capability to create and call PROCEDURES
1056 o Add support for OUT and INOUT parameters to languages other
1058 o Add PL/PythonU tracebacks
1060 http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php
1062 o Allow data to be passed in native language formats, rather
1065 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php
1067 o Add ability to obfuscate function bodies
1069 http://archives.postgresql.org/pgsql-patches/2008-01/msg00125.php
1076 * Have pg_ctl look at PGHOST in case it is a socket directory?
1077 * Allow pg_ctl to work properly with configuration files located outside
1078 the PGDATA directory
1080 pg_ctl can not read the pid file because it isn't located in the
1081 config directory but in the PGDATA directory. The solution is to
1082 allow pg_ctl to read and understand postgresql.conf to find the
1083 data_directory value.
1085 * Add a function like pg_get_indexdef() that report more detailed index
1088 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php
1091 o -Have psql show current values for a sequence
1092 o Move psql backslash database information into the backend, use
1095 This would allow non-psql clients to pull the same information out
1096 of the database as psql.
1097 http://archives.postgresql.org/pgsql-hackers/2004-01/msg00191.php
1099 o Make psql's \d commands more consistent
1101 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1102 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1104 o Consistently display privilege information for all objects in psql
1105 o Add auto-expanded mode so expanded output is used if the row
1106 length is wider than the screen width.
1108 Consider using auto-expanded mode for backslash commands like \df+.
1110 o Prevent tab completion of SET TRANSACTION from querying the
1111 database and therefore preventing the transaction isolation
1112 level from being set.
1114 Currently SET <tab> causes a database lookup to check all
1115 supported session variables. This query causes problems
1116 because setting the transaction isolation level must be the
1117 first statement of a transaction.
1119 o Add a \set variable to control whether \s displays line numbers
1121 Another option is to add \# which lists line numbers, and
1122 allows command execution.
1124 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php
1126 o Prevent escape string warnings when object names have
1129 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php
1131 o Have \d show foreign keys that reference a table's primary key
1133 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php
1135 o Have \d show child tables that inherit from the specified parent
1136 o -Have \l+ show database size, if permissions allow
1138 Ideally it will not generate an error for invalid permissions
1140 o Include the symbolic SQLSTATE name in verbose error reports
1142 http://archives.postgresql.org/pgsql-general/2007-09/msg00438.php
1144 o -Improve display of enums to show valid enum values
1145 o Add prompt escape to display the client and server versions
1146 o Add option to wrap column values at whitespace boundaries,
1147 rather than chopping them at a fixed width.
1149 Currently, "wrapped" format chops values into fixed
1150 widths. Perhaps the word wrapping could use the same
1151 algorithm documented in the W3C specification.
1152 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00404.php
1153 http://www.w3.org/TR/CSS21/tables.html#auto-table-layout
1154 o Add "auto" expanded mode that outputs in expanded format if
1155 "wrapped" mode can't wrap the output to the screen width
1157 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00417.php
1160 * pg_dump / pg_restore
1161 o %Add dumping of comments on index columns and composite type columns
1162 o %Add full object name to the tag field. eg. for operators we need
1163 '=(integer, integer)', instead of just '='.
1164 o Add pg_dumpall custom format dumps?
1165 o Allow selection of individual object(s) of all types, not just
1167 o In a selective dump, allow dumping of an object and all its
1169 o Add options like pg_restore -l and -L to pg_dump
1170 o Stop dumping CASCADE on DROP TYPE commands in clean mode
1171 o Allow pg_dump --clean to drop roles that own objects or have
1173 o Change pg_dump so that a comment on the dumped database is
1174 applied to the loaded database, even if the database has a
1175 different name. This will require new backend syntax, perhaps
1176 COMMENT ON CURRENT DATABASE.
1177 o Remove unnecessary function pointer abstractions in pg_dump source
1179 o Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
1180 multiple objects simultaneously
1182 The difficulty with this is getting multiple dump processes to
1183 produce a single dump output file. It also would require
1184 several sessions to share the same snapshot.
1185 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1187 o Allow pg_restore to utilize multiple CPUs and I/O channels by
1188 restoring multiple objects simultaneously
1190 This might require a pg_restore flag to indicate how many
1191 simultaneous operations should be performed. Only pg_dump's
1192 -Fc format has the necessary dependency information.
1193 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php
1195 o To better utilize resources, allow pg_restore to check foreign
1196 keys simultaneously, where possible
1197 o Allow pg_restore to create all indexes of a table
1198 concurrently, via a single heap scan
1200 This requires a pg_dump -Fc file because that format contains
1201 the required dependency information.
1202 http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php
1204 o Allow pg_restore to load different parts of the COPY data
1206 o -Prevent pg_dump/pg_restore from being affected by
1209 Using psql to restore a pg_dump dump is also affected.
1211 o Remove pre-7.3 pg_dump code that assumes pg_depend does not exit
1212 o Allow pre/data/post files when schema and data are dumped
1213 separately, for performance reasons
1215 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1222 Document differences between ecpg and the SQL standard and
1223 information about the Informix-compatibility module.
1225 o Solve cardinality > 1 for input descriptors / variables?
1226 o Add a semantic check level, e.g. check if a table really exists
1227 o fix handling of DB attributes that are arrays
1228 o Use backend PREPARE/EXECUTE facility for ecpg where possible
1230 o Fix nested C comments
1231 o %sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified
1232 o Make SET CONNECTION thread-aware, non-standard?
1233 o Allow multidimensional arrays
1234 o Add internationalized message strings
1235 o Implement COPY FROM STDIN
1239 o Add PQescapeIdentifierConn()
1240 o Prevent PQfnumber() from lowercasing unquoted the column name
1242 PQfnumber() should never have been doing lowercasing, but
1243 historically it has so we need a way to prevent it
1245 o Allow statement results to be automatically batched to the client
1247 Currently all statement results are transferred to the libpq
1248 client before libpq makes the results available to the
1249 application. This feature would allow the application to make
1250 use of the first result rows while the rest are transferred, or
1251 held on the server waiting for them to be requested by libpq.
1252 One complexity is that a statement like SELECT 1/col could error
1253 out mid-way through the result set.
1255 o Consider disallowing multiple queries in PQexec() as an
1256 additional barrier to SQL injection attacks
1258 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php
1260 o Add PQexecf() that allows complex parameter substitution
1262 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php
1264 o Add SQLSTATE severity to PGconn return status
1266 http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php
1272 * Add deferred trigger queue file
1274 Right now all deferred trigger information is stored in backend
1275 memory. This could exhaust memory for very large trigger queues.
1276 This item involves dumping large queues into files, or doing some
1277 kind of join to process all the triggers, or some bulk operation.
1278 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00876.php
1280 * Allow triggers to be disabled in only the current session.
1282 This is currently possible by starting a multi-statement transaction,
1283 modifying the system tables, performing the desired SQL, restoring the
1284 system tables, and committing the transaction. ALTER TABLE ...
1285 TRIGGER requires a table lock so it is not ideal for this usage.
1287 * With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
1289 If the dump is known to be valid, allow foreign keys to be added
1290 without revalidating the data.
1292 * Allow statement-level triggers to access modified rows
1293 * Support triggers on columns
1295 http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php
1297 * Allow AFTER triggers on system tables
1299 System tables are modified in many places in the backend without going
1300 through the executor and therefore not causing triggers to fire. To
1301 complete this item, the functions that modify system tables will have
1304 * Tighten trigger permission checks
1306 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
1308 * Allow BEFORE INSERT triggers on views
1310 http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php
1312 * -Add ability to trigger on TRUNCATE
1313 * Add database and transaction-level triggers
1315 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00451.php
1316 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00620.php
1318 * Reduce locking requirements for creating a trigger
1320 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00635.php
1327 * Add UNIQUE capability to non-btree indexes
1328 * Prevent index uniqueness checks when UPDATE does not modify the column
1330 Uniqueness (index) checks are done when updating a column even if the
1331 column is not modified by the UPDATE.
1333 * Allow the creation of on-disk bitmap indexes which can be quickly
1334 combined with other bitmap indexes
1336 Such indexes could be more compact if there are only a few distinct values.
1337 Such indexes can also be compressed. Keeping such indexes updated can be
1340 http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php
1341 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php
1342 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00265.php
1343 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01214.php
1344 http://archives.postgresql.org/pgsql-patches/2007-05/msg00013.php
1345 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00741.php
1347 * Allow accurate statistics to be collected on indexes with more than
1348 one column or expression indexes, perhaps using per-index statistics
1350 http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php
1351 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php
1353 * Consider increasing the default and maximum number of statistics targets,
1354 and reduce statistics target overhead
1356 Also consider having a larger statistics target for indexed columns
1357 and expression indexes.
1358 http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php
1359 http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php
1360 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01066.php
1361 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00188.php
1363 * Consider smaller indexes that record a range of values per heap page,
1364 rather than having one index entry for every heap row
1366 This is useful if the heap is clustered by the indexed values.
1367 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php
1368 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php
1369 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php
1370 http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php
1371 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00014.php
1372 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00487.php
1373 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01589.php
1375 * Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
1377 This is difficult because you must upgrade to an exclusive table lock
1378 to replace the existing index file. CREATE INDEX CONCURRENTLY does not
1379 have this complication. This would allow index compaction without
1381 http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php
1383 * Allow multiple indexes to be created concurrently, ideally via a
1384 single heap scan, and have pg_restore use it
1386 * Consider sorting entries before inserting into btree index
1388 http://archives.postgresql.org/pgsql-general/2008-01/msg01010.php
1390 * Allow index scans to return matching index keys, not just the matching
1393 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01657.php
1399 o Allow inherited tables to inherit indexes, UNIQUE constraints,
1400 and primary/foreign keys
1401 o Honor UNIQUE INDEX on base column in INSERTs/UPDATEs
1402 on inherited table, e.g. INSERT INTO inherit_table
1403 (unique_index_col) VALUES (dup) should fail
1405 The main difficulty with this item is the problem of
1406 creating an index that can span multiple tables.
1408 o Allow SELECT ... FOR UPDATE on inherited tables
1409 o Require all CHECK constraints to be inherited
1411 http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php
1413 o Add checks to prevent a CREATE RULE views on inherited tables
1415 http://archives.postgresql.org/pgsql-general/2008-02/msg01420.php
1416 http://archives.postgresql.org/pgsql-general/2008-03/msg00077.php
1421 o Add more GIST index support for geometric data types
1422 o Allow GIST indexes to create certain complex index types, like
1423 digital trees (see Aoki)
1428 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php
1430 o Pack hash index buckets onto disk pages more efficiently
1432 Currently only one hash bucket can be stored on a page. Ideally
1433 several hash buckets could be stored on a single page and greater
1434 granularity used for the hash algorithm.
1436 http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php
1438 o Consider sorting hash buckets so entries can be found using a
1439 binary search, rather than a linear scan
1440 o In hash indexes, consider storing the hash value with or instead
1442 o Add WAL logging for crash recovery
1443 o Allow multi-column hash indexes
1444 o -During index creation, pre-sort the tuples to improve build speed
1448 * Consider using hash buckets to do DISTINCT, rather than sorting
1450 This would be beneficial when there are few distinct values. This is
1451 already used by GROUP BY.
1453 * Consider whether duplicate keys should be sorted by block/offset
1455 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00558.php
1457 * -Avoid tuple some tuple copying in sort routines
1458 * Consider being smarter about memory and external files used during
1461 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01101.php
1462 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00045.php
1464 * Consider detoasting keys before sorting
1470 * Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
1472 Ideally this requires a separate test program that can be run
1473 at initdb time or optionally later. Consider O_SYNC when
1476 * Add program to test if fsync has a delay compared to non-fsync
1477 * Consider sorting writes during checkpoint
1479 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php
1488 We could use a fixed row count and a +/- count to follow MVCC
1489 visibility rules, or a single cached value could be used and
1490 invalidated if anyone modifies the table. Another idea is to
1491 get a count directly from a unique index, but for this to be
1492 faster than a sequential scan it must avoid access to the heap
1493 to obtain tuple visibility information.
1495 * Provide a way to calculate an "estimated COUNT(*)"
1497 Perhaps by using the optimizer's cardinality estimates or random
1500 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php
1502 * Allow data to be pulled directly from indexes
1504 Currently indexes do not have enough tuple visibility information
1505 to allow data to be pulled from the index without also accessing
1506 the heap. One way to allow this is to set a bit on index tuples
1507 to indicate if a tuple is currently visible to all transactions
1508 when the first valid heap lookup happens. This bit would have to
1509 be cleared when a heap tuple is expired.
1511 Another idea is to maintain a bitmap of heap pages where all rows
1512 are visible to all backends, and allow index lookups to reference
1513 that bitmap to avoid heap lookups, perhaps the same bitmap we might
1514 add someday to determine which heap pages need vacuuming. Frequently
1515 accessed bitmaps would have to be stored in shared memory. One 8k
1516 page of bitmaps could track 512MB of heap pages.
1518 A third idea would be for a heap scan to check if all rows are visible
1519 and if so set a per-table flag which can be checked by index scans.
1520 Any change to the table would have to clear the flag. To detect
1521 changes during the heap scan a counter could be set at the start and
1522 checked at the end --- if it is the same, the table has not been
1523 modified --- any table change would increment the counter.
1525 http://archives.postgresql.org/pgsql-patches/2007-10/msg00166.php
1526 http://archives.postgresql.org/pgsql-patches/2008-01/msg00049.php
1528 * Consider automatic caching of statements at various levels:
1531 o Query execute plan
1533 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00823.php
1535 * Consider increasing internal areas when shared buffers is increased
1537 http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php
1539 * Consider decreasing the amount of memory used by PrivateRefCount
1541 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php
1542 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php
1544 * Consider allowing higher priority queries to have referenced buffer
1545 cache pages stay in memory longer
1547 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php
1554 * Improve speed with indexes
1556 For large table adjustments during VACUUM FULL, it is faster to cluster
1557 or reindex rather than update the index. Also, index updates can bloat
1560 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php
1561 http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php
1562 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php
1564 * Auto-fill the free space map by scanning the buffer cache or by
1565 checking pages written by the background writer
1567 http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php
1568 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php
1570 * Create a bitmap of pages that need vacuuming
1572 Instead of sequentially scanning the entire table, have the background
1573 writer or some other process record pages that have expired rows, then
1574 VACUUM can look at just those pages rather than the entire table. In
1575 the event of a system crash, the bitmap would probably be invalidated.
1576 One complexity is that index entries still have to be vacuumed, and
1577 doing this without an index scan (by using the heap values to find the
1578 index entry) might be slow and unreliable, especially for user-defined
1581 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php
1582 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php
1583 http://archives.postgresql.org/pgsql-patches/2007-03/msg00508.php
1584 http://archives.postgresql.org/pgsql-patches/2007-04/msg00347.php
1585 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00156.php
1586 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00546.php
1587 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00416.php
1589 * Allow FSM to return free space toward the beginning of the heap file,
1590 in hopes that empty pages at the end can be truncated by VACUUM
1591 * Allow FSM page return free space based on table clustering, to assist
1592 in maintaining clustering?
1593 * -Improve dead row detection during multi-statement transactions usage
1595 http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php
1597 * Consider a more compact data representation for dead tuples
1599 http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php
1604 o %Issue log message to suggest VACUUM FULL if a table is nearly
1606 o Improve control of auto-vacuum
1608 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php
1610 o Prevent long-lived temporary tables from causing frozen-xid
1611 advancement starvation
1613 The problem is that autovacuum cannot vacuum them to set frozen xids;
1614 only the session that created them can do that.
1615 http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php
1617 o Store per-table autovacuum settings in pg_class.reloptions.
1619 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php
1620 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php
1622 o Prevent autovacuum from running if an old transaction is still
1623 running from the last vacuum
1625 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php
1632 * Fix priority ordering of read and write light-weight locks (Neil)
1634 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php
1635 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php
1637 * Fix problem when multiple subtransactions of the same outer transaction
1638 hold different types of locks, and one subtransaction aborts
1640 http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php
1641 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php
1642 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php
1643 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php
1645 * Allow UPDATEs on only non-referential integrity columns not to conflict
1646 with referential integrity locks
1648 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php
1650 * Add idle_in_transaction_timeout GUC so locks are not held for long
1653 * Improve deadlock detection when a page cleaning lock conflicts
1654 with a shared buffer that is pinned
1656 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php
1657 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1658 http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php
1660 * Detect deadlocks involving LockBufferForCleanup()
1662 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1666 Startup Time Improvements
1667 =========================
1669 * Experiment with multi-threaded backend for backend creation
1671 This would prevent the overhead associated with process creation. Most
1672 operating systems have trivial process creation time compared to
1673 database startup overhead, but a few operating systems (Win32,
1674 Solaris) might benefit from threading. Also explore the idea of
1675 a single session using multiple threads to execute a statement faster.
1682 * Eliminate need to write full pages to WAL before page modification
1684 Currently, to protect against partial disk page writes, we write
1685 full page images to WAL before they are modified so we can correct any
1686 partial page writes during recovery. These pages can also be
1687 eliminated from point-in-time archive files.
1688 http://archives.postgresql.org/pgsql-hackers/2002-06/msg00655.php
1690 o When off, write CRC to WAL and check file system blocks
1693 If CRC check fails during recovery, remember the page in case
1694 a later CRC for that page properly matches.
1696 o Write full pages during file system write and not when
1697 the page is modified in the buffer cache
1699 This allows most full page writes to happen in the background
1700 writer. It might cause problems for applying WAL on recovery
1701 into a partially-written page, but later the full page will be
1704 * Allow WAL traffic to be streamed to another server for stand-by
1706 * Reduce WAL traffic so only modified values are written rather than
1709 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php
1711 * Allow WAL information to recover corrupted pg_controldata
1713 http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php
1715 * Find a way to reduce rotational delay when repeatedly writing
1718 Currently fsync of WAL requires the disk platter to perform a full
1719 rotation to fsync again. One idea is to write the WAL to different
1720 offsets that might reduce the rotational delay.
1721 http://archives.postgresql.org/pgsql-hackers/2002-11/msg00483.php
1723 * Allow WAL logging to be turned off for a table, but the table
1724 might be dropped or truncated during crash recovery
1726 Allow tables to bypass WAL writes and just fsync() dirty pages on
1727 commit. This should be implemented using ALTER TABLE, e.g. ALTER
1728 TABLE PERSISTENCE [ DROP | TRUNCATE | DEFAULT ]. Tables using
1729 non-default logging should not use referential integrity with
1730 default-logging tables. A table without dirty buffers during a
1731 crash could perhaps avoid the drop/truncate.
1732 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1734 * Allow WAL logging to be turned off for a table, but the table would
1735 avoid being truncated/dropped
1737 To do this, only a single writer can modify the table, and writes
1738 must happen only on new pages so the new pages can be removed during
1739 crash recovery. Readers can continue accessing the table. Such
1740 tables probably cannot have indexes. One complexity is the handling
1741 of indexes on TOAST tables.
1742 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1744 * Speed WAL recovery by allowing more than one page to be prefetched
1746 This should be done utilizing the same infrastructure used for
1747 prefetching in general to avoid introducing complex error-prone code
1749 http://archives.postgresql.org/pgsql-general/2007-12/msg00683.php
1750 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00497.php
1751 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php
1753 * Improve WAL concurrency by increasing lock granularity
1755 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00556.php
1757 * Be more aggressive about creating WAL files
1759 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01325.php
1760 http://archives.postgresql.org/pgsql-hackers/2004-07/msg01075.php
1761 http://archives.postgresql.org/pgsql-hackers/2005-04/msg00556.php
1763 * Have resource managers report the duration of their status changes
1765 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01468.php
1767 * Move pgfoundry's xlogdump to /contrib and have it rely more closely
1768 on the WAL backend code
1770 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00035.php
1775 Optimizer / Executor
1776 ====================
1778 * Improve selectivity functions for geometric operators
1779 * Precompile SQL functions to avoid overhead
1780 * Create utility to compute accurate random_page_cost value
1781 * Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
1782 * Have EXPLAIN ANALYZE issue NOTICE messages when the estimated and
1783 actual row counts differ by a specified percentage
1784 * Improve how ANALYZE computes in-doubt tuples
1786 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00771.php
1788 * Log statements where the optimizer row estimates were dramatically
1789 different from the number of rows actually found?
1790 * Consider compressed annealing to search for query plans
1792 This might replace GEQO, http://sixdemonbag.org/Djinni.
1794 * Improve merge join performance by allowing mark/restore of
1797 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php
1799 * Consider using a hash for joining to a large IN (VALUES ...) list
1801 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php
1808 * Consider having the background writer update the transaction status
1809 hint bits before writing out the page
1811 Implementing this requires the background writer to have access to system
1812 catalogs and the transaction status log.
1814 * Consider adding buffers the background writer finds reusable to the
1817 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1819 * Automatically tune bgwriter_delay based on activity rather then using a
1822 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1824 * Consider wither increasing BM_MAX_USAGE_COUNT improves performance
1826 http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php
1828 * Test to see if calling PreallocXlogFiles() from the background writer
1829 will help with WAL segment creation latency
1831 http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php
1835 Miscellaneous Performance
1836 =========================
1838 * Do async I/O for faster random read-ahead of data
1840 Async I/O allows multiple I/O requests to be sent to the disk with
1841 results coming back asynchronously.
1843 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php
1844 http://archives.postgresql.org/pgsql-performance/2007-09/msg00255.php
1845 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php
1846 http://archives.postgresql.org/pgsql-patches/2008-01/msg00170.php
1848 * Use mmap() rather than SYSV shared memory or to write WAL files?
1850 This would remove the requirement for SYSV SHM but would introduce
1851 portability issues. Anonymous mmap (or mmap to /dev/zero) is required
1852 to prevent I/O overhead.
1854 * Consider mmap()'ing files into a backend?
1856 Doing I/O to large tables would consume a lot of address space or
1857 require frequent mapping/unmapping. Extending the file also causes
1858 mapping problems that might require mapping only individual pages,
1859 leading to thousands of mappings. Another problem is that there is no
1860 way to _prevent_ I/O to disk from the dirty shared buffers so changes
1861 could hit disk before WAL is written.
1863 * Add a script to ask system configuration questions and tune postgresql.conf
1864 * Consider ways of storing rows more compactly on disk
1866 o Reduce the row header size?
1867 o Consider reducing on-disk varlena length from four bytes to
1868 two because a heap row cannot be more than 64k in length
1870 * Consider increasing NUM_CLOG_BUFFERS
1872 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php
1873 http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php
1875 * Consider transaction start/end performance improvements
1877 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00948.php
1878 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1880 * Allow user configuration of TOAST thresholds
1882 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php
1883 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php
1885 * Allow configuration of backend priorities via the operating system
1887 Though backend priorities make priority inversion during lock
1888 waits possible, research shows that this is not a huge problem.
1890 http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php
1892 * Experiment with multi-threaded backend better I/O utilization
1894 This would allow a single query to make use of multiple I/O channels
1895 simultaneously. One idea is to create a background reader that can
1896 pre-fetch sequential and index scan pages needed by other backends.
1897 This could be expanded to allow concurrent reads from multiple devices
1898 in a partitioned table.
1900 * Experiment with multi-threaded backend better CPU utilization
1902 This would allow several CPUs to be used for a single query, such as
1903 for sorting or query execution.
1905 * Consider increasing the minimum allowed number of shared buffers
1907 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php
1909 * Expire published xmin for read-only and idle transactions
1911 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00343.php
1913 * Consider if CommandCounterIncrement() can avoid its
1914 AcceptInvalidationMessages() call
1916 http://archives.postgresql.org/pgsql-committers/2007-11/msg00585.php
1918 * Improve performance of shared invalidation queue for multiple CPUs
1920 http://archives.postgresql.org/pgsql-performance/2008-01/msg00023.php
1922 * Consider Cartesian joins when both relations are needed to form an
1923 indexscan qualification for a third relation
1925 http://archives.postgresql.org/pgsql-performance/2007-12/msg00090.php
1927 * Consider not storing a NULL bitmap on disk if all the NULLs are
1930 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00624.php
1931 http://archives.postgresql.org/pgsql-patches/2007-12/msg00109.php
1933 * Sort large UPDATE/DELETEs so it is done in heap order
1935 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01119.php
1937 * SMP scalability improvements
1939 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00439.php
1940 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00206.php
1941 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1943 * Research reducing deTOASTing in more places
1945 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00895.php
1947 * Allow one transaction to see tuples using the snapshot of another
1950 This would assist multiple backends in working together.
1951 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00400.php
1953 * Consider decreasing the I/O caused by updating tuple hint bits
1955 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00847.php
1956 http://archives.postgresql.org/pgsql-patches/2008-07/msg00199.php
1963 * Add use of 'const' for variables in source tree
1964 * Move some things from contrib into main tree
1965 * %Remove warnings created by -Wcast-align
1966 * Move platform-specific ps status display info from ps_status.c to ports
1967 * Add optional CRC checksum to heap and index pages
1968 * Improve documentation to build only interfaces (Marc)
1969 * Remove or relicense modules that are not under the BSD license, if possible
1970 * Acquire lock on a relation before building a relcache entry for it
1971 * Allow cross-compiling by generating the zic database on the target system
1972 * Improve NLS maintenance of libpgport messages linked onto applications
1973 * Clean up casting in contrib/isn
1975 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php
1977 * Improve the /contrib installation experience
1979 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00132.php
1981 * Use UTF8 encoding for NLS messages so all server encodings can
1983 * Update Bonjour to work with newer cross-platform SDK
1985 http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php
1986 http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php
1988 * Consider GnuTLS if OpenSSL license becomes a problem
1990 http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php
1991 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php
1993 * Consider changing documentation format from SGML to XML
1995 http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php
1997 * Consider making NAMEDATALEN more configurable in future releases
1998 * Update our code to handle 64-bit timezone files to match the zic
1999 source code, which now uses them
2000 * Have configure choose integer datetimes by default
2002 http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php
2004 * Support scoped IPv6 addresses
2006 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php
2008 * Consider allowing 64-bit integers and floats to be passed by value on
2011 Also change 32-bit floats (float4) to be passed by value at the same
2014 * Research use of signals and sleep wake ups
2016 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php
2018 * Add automated check for invalid C++ source code constructs
2020 http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php
2022 * Consider simplifying how memory context resets handle child contexts
2024 http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php
2026 * Remove use of MAKE_PTR and MAKE_OFFSET macros
2028 http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php
2030 * Convert single quotes to apostrophes in the PDF documentation
2032 http://archives.postgresql.org/pgsql-docs/2007-12/msg00059.php
2034 * Create three versions of libpgport to simplify client code
2036 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00154.php
2038 * Remove old-style routines for manipulating tuples
2040 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00851.php
2042 * Improve detection of shared memory segments being used by others
2043 by checking the SysV shared memory field 'nattch'
2045 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00656.php
2046 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00673.php
2048 * Implement the non-threaded Avahi service discovery protocol
2050 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00939.php
2051 http://archives.postgresql.org/pgsql-patches/2008-02/msg00097.php
2052 http://archives.postgresql.org/pgsql-hackers/2008-03/msg01211.php
2053 http://archives.postgresql.org/pgsql-patches/2008-04/msg00001.php
2055 * Implement a module capability for loading /contrib-style extensions
2057 http://archives.postgresql.org/pgsql-patches/2008-04/msg00164.php
2059 * Fix system views like pg_stat_all_tables to use set-returning
2060 functions, rather than views of per-column functions
2065 o Remove configure.in check for link failure when cause is found
2066 o Remove readdir() errno patch when runtime/mingwex/dirent.c rev
2068 o -Remove psql newline patch when we find out why mingw outputs an
2070 o Allow psql to use readline once non-US code pages work with
2072 o Fix problem with shared memory on the Win32 Terminal Server
2073 o Diagnose problem where shared memory can sometimes not be
2074 attached by postmaster children
2076 http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php
2078 o Improve signal handling
2080 http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php
2082 o Convert MSVC build system to remove most batch files
2084 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php
2086 o Prevent SSL from sending network packets to avoid interference
2087 with Win32 signal emulation
2089 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00455.php
2091 o Support pgxs when using MSVC
2093 o Fix MSVC NLS support, like for to_char()
2095 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00485.php
2096 http://archives.postgresql.org/pgsql-patches/2008-02/msg00038.php
2098 o Find a correct rint() substitute on Windows
2100 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00808.php
2104 * Wire Protocol Changes
2106 o Allow dynamic character set handling
2107 o Add decoded type, length, precision
2109 o Update clients to use data types, typmod, schema.table.column names
2110 of result sets using new statement protocol
2116 * Add pre-parsing phase that converts non-ISO syntax to supported
2119 This could allow SQL written for other databases to run without
2122 * Allow plug-in modules to emulate features from other databases
2123 * Add features of Oracle-style packages (Pavel)
2125 A package would be a schema with session-local variables,
2126 public/private functions, and initialization functions. It
2127 is also possible to implement these capabilities
2128 in any schema and not use a separate "packages"
2131 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php
2133 * Consider allowing control of upper/lower case folding of unquoted
2136 http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php
2137 http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php
2138 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00849.php
2139 http://archives.postgresql.org/pgsql-hackers/2008-07/msg00415.php
2141 * Add autonomous transactions
2143 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php
2147 Features We Do _Not_ Want
2148 =========================
2150 * All backends running as threads in a single process (not wanted)
2152 This eliminates the process protection we get from the current setup.
2153 Thread creation is usually the same overhead as process creation on
2154 modern systems, so it seems unwise to use a pure threaded model.
2156 * Optimizer hints (not wanted)
2158 Optimizer hints are used to work around problems in the optimizer. We
2159 would rather have the problems reported and fixed.
2161 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php
2162 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php
2163 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php
2166 Because we support postfix operators, it isn't possible to make AS
2167 optional and continue to use bison.
2168 http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php
2170 http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php
2172 * Embedded server (not wanted)
2174 While PostgreSQL clients runs fine in limited-resource environments, the
2175 server requires multiple processes and a stable pool of resources to
2176 run reliabily and efficiently. Stripping down the PostgreSQL server
2177 to run in the same process address space as the client application
2178 would add too much complexity and failure cases.