3 Current maintainer: Bruce Momjian (bruce@momjian.us)
4 Last updated: Tue Jul 1 14:09:02 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
210 * Change NUMERIC to enforce the maximum precision
211 * Reduce storage space for small NUMERICs
213 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php
214 http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php
215 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php
217 * Fix data types where equality comparison isn't intuitive, e.g. box
218 * Add support for public SYNONYMs
220 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php
222 * Fix CREATE CAST on DOMAINs
224 http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php
225 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php
227 * Allow domains to be cast
229 http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php
230 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php
232 * Add support for SQL-standard GENERATED/IDENTITY columns
234 http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php
235 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php
236 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php
237 http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php
238 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php
240 * Improve XML support
242 http://developer.postgresql.org/index.php/XML_Support
244 * Consider placing all sequences in a single table, or create a system
247 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
249 * Allow the UUID type to accept non-standard formats
251 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php
253 * Consider a special data type for regular expressions
255 http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php
257 * Reduce BIT data type overhead using short varlena headers
259 http://archives.postgresql.org/pgsql-general/2007-12/msg00273.php
261 * Allow xml arrays to be cast to other data types
263 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00981.php
264 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00231.php
265 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00471.php
267 * Simplify integer cross-data-type operators
269 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00189.php
271 * Allow XML to accept more liberal DOCTYPE specifications
273 http://archives.postgresql.org/pgsql-general/2008-02/msg00347.php
275 * Allow adding/renaming/removing enumerated values to an existing
278 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01718.php
282 o Allow infinite dates and intervals just like infinite timestamps
283 o Merge hardwired timezone names with the TZ database; allow either
284 kind everywhere a TZ name is currently taken
285 o Allow TIMESTAMP WITH TIME ZONE to store the original timezone
286 information, either zone name or offset from UTC
288 If the TIMESTAMP value is stored with a time zone name, interval
289 computations should adjust based on the time zone rules.
290 http://archives.postgresql.org/pgsql-hackers/2004-10/msg00705.php
292 o Fix SELECT '0.01 years'::interval, '0.01 months'::interval
293 o Add a GUC variable to allow output of interval values in ISO8601
295 o Have timestamp subtraction not call justify_hours()?
297 http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php
299 o Improve timestamptz subtraction to be DST-aware
301 Currently subtracting one date from another that crosses a
302 daylight savings time adjustment can return '1 day 1 hour', but
303 adding that back to the first date returns a time one hour in
304 the future. This is caused by the adjustment of '25 hours' to
305 '1 day 1 hour', and '1 day' is the same time the next day, even
306 if daylight savings adjustments are involved.
308 o Fix interval display to support values exceeding 2^31 hours
309 o Add overflow checking to timestamp and interval arithmetic
310 o Extend timezone code to allow 64-bit values so we can
311 represent years beyond 2038
313 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php
315 o -Use LC_TIME for localized weekday/month names, rather than
318 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php
320 o Add ISO INTERVAL handling
322 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php
323 http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php
325 o Support ISO INTERVAL syntax if units cannot be determined from
326 the string, and are supplied after the string
328 The SQL standard states that the units after the string
329 specify the units of the string, e.g. INTERVAL '2' MINUTE
330 should return '00:02:00'. The current behavior has the units
331 restrict the interval value to the specified unit or unit
332 range, INTERVAL '70' SECOND returns '00:00:10'.
334 For syntax that isn't uniquely ISO or PG syntax, like '1' or
335 '1:30', treat as ISO if there is a range specification clause,
336 and as PG if there no clause is present, e.g. interpret '1:30'
337 MINUTE TO SECOND as '1 minute 30 seconds', and interpret
338 '1:30' as '1 hour, 30 minutes'.
340 This makes common cases like SELECT INTERVAL '1' MONTH
341 SQL-standard results. The SQL standard supports a limited
342 number of unit combinations and doesn't support unit names in
343 the string. The PostgreSQL syntax is more flexible in the
344 range of units supported, e.g. PostgreSQL supports '1 year 1
345 hour', while the SQL standard does not.
347 o Add support for year-month syntax, INTERVAL '50-6' YEAR
349 o Interpret INTERVAL '1 year' MONTH as CAST (INTERVAL '1
350 year' AS INTERVAL MONTH), and this should return '12 months'
351 o Round or truncate values to the requested precision, e.g.
352 INTERVAL '11 months' AS YEAR should return one or zero
353 o Support precision, CREATE TABLE foo (a INTERVAL MONTH(3))
358 o Delay resolution of array expression's data type so assignment
359 coercion can be performed on empty array expressions
360 o Add support for arrays of domains
362 http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php
364 o Allow single-byte header storage for array elements
369 o Improve vacuum of large objects, like contrib/vacuumlo?
370 o Add security checking for large objects
371 o Auto-delete large objects when referencing row is deleted
373 contrib/lo offers this functionality.
375 o Allow read/write into TOAST values like large objects
377 This requires the TOAST column to be stored EXTERNAL.
379 o Add API for 64-bit large object access
381 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php
385 * Add locale-aware MONEY type, and support multiple currencies
387 http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
388 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
390 * MONEY dumps in a locale-specific format making it difficult to
391 restore to a system with a different locale
392 * Allow MONEY to be easily cast to/from other numeric data types
396 o Allow dictionaries to change the token that is passed on to
399 http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php
401 o Consider a function-based API for '@@' searches
403 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php
405 o Improve text search error messages
407 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00966.php
408 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01146.php
410 o Consider changing error to warning for strings larger than one
413 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00190.php
414 http://archives.postgresql.org/pgsql-patches/2008-03/msg00062.php
421 * Allow INET subnet tests using non-constants to be indexed
422 * Allow to_date() and to_timestamp() accept localized month names
423 * Fix to_date()-related functions to consistently issue errors
425 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php
427 * Add missing parameter handling in to_char()
429 http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php
431 * Allow substring/replace() to get/set bit values
432 * Allow to_char() on interval values to accumulate the highest unit
435 Some special format flag would be required to request such
436 accumulation. Such functionality could also be added to EXTRACT.
437 Prevent accumulation that crosses the month/day boundary because of
438 the uneven number of days in a month.
440 o to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
441 o to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
442 o to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
443 o to_char(INTERVAL '3 years 5 months','MM') => 41
445 * Implement inlining of set-returning functions defined in SQL
446 * Allow SQL-language functions to return results from RETURNING queries
448 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php
450 * Allow SQL-language functions to reference parameters by parameter name
452 Currently SQL-language functions can only refer to dollar parameters,
455 * Add SPI_gettypmod() to return the typemod for a TupleDesc
456 * Enforce typmod for function inputs, function results and parameters for
457 spi_prepare'd statements called from PLs
459 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php
461 * Allow holdable cursors in SPI
462 * Tighten function permission checks
464 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php
466 * Fix IS OF so it matches the ISO specification, and add documentation
468 http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php
469 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php
471 * Add missing operators for geometric data types
473 Some geometric types do not have the full suite of geometric operators,
476 * Implement Boyer-Moore searching in strpos()
478 http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php
480 * Prevent malicious functions from being executed with the permissions
481 of unsuspecting users
483 Index functions are safe, so VACUUM and ANALYZE are safe too.
484 Triggers, CHECK and DEFAULT expressions, and rules are still vulnerable.
485 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php
487 * Reduce memory usage of aggregates in set returning functions
489 http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php
491 * Add temporal versions of generate_series()
493 http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php
495 * Add array_accum() and array_to_set() functions for arrays
497 The standards specify array_agg() and UNNEST.
498 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php
500 * Fix /contrib/ltree operator
502 http://archives.postgresql.org/pgsql-bugs/2007-11/msg00044.php
504 * Fix inconsistent precedence of =, >, and < compared to <>, >=, and <=
506 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00145.php
508 * Fix regular expression bug when using complex back-references
510 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00000.php
512 * Have /contrib/dblink reuse unnamed connections
514 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00895.php
516 * Add SQL-standard array_agg() and unnest() array functions
518 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01017.php
520 * Allow calling of a procedure outside a SELECT that can control the
523 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01375.php
527 Multi-Language Support
528 ======================
530 * Add NCHAR (as distinguished from ordinary varchar),
531 * Allow locale to be set at database creation
533 Currently locale can only be set during initdb. No global tables have
534 locale-aware columns. However, the database template used during
535 database creation might have locale-aware indexes. The indexes would
536 need to be reindexed to match the new locale.
538 * Allow encoding on a per-column basis optionally using the ICU library;
541 Right now only one encoding is allowed per database.
543 http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php
544 http://archives.postgresql.org/pgsql-patches/2005-08/msg00039.php
545 http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php
546 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00110.php
547 http://archives.postgresql.org/pgsql-patches/2005-09/msg00020.php
548 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01121.php
549 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00767.php
550 http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php
551 http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php
552 http://wiki.postgresql.org/wiki/Todo:Collate
553 http://wiki.postgresql.org/wiki/Todo:ICU
555 * Support multiple simultaneous character sets, per SQL92
556 * Improve UTF8 combined character handling?
557 * Add octet_length_server() and octet_length_client()
558 * Make octet_length_client() the same as octet_length()?
559 * Fix problems with wrong runtime encoding conversion for NLS message files
560 * Add URL to more complete multi-byte regression tests
562 http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php
564 * Fix ILIKE and regular expressions to handle case insensitivity
565 properly in multibyte encodings
567 http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php
568 http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php
570 * Set client encoding based on the client operating system encoding
572 Currently client_encoding is set in postgresql.conf, which
573 defaults to the server encoding.
574 http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php
576 * Change memory allocation for multi-byte functions so memory is
577 allocated inside conversion functions
579 Currently we preallocate memory based on worst-case usage.
586 * Automatically create rules on views so they are updateable, per SQL99
588 We can only auto-create rules for simple views. For more complex
589 cases users will still have to write rules manually.
591 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php
592 http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php
594 * Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
595 * Allow VIEW/RULE recompilation when the underlying tables change
597 Another issue is whether underlying table changes should be reflected
598 in the view, e.g. should SELECT * show additional columns if they
599 are added after the view is created.
601 * Make it possible to use RETURNING together with conditional DO INSTEAD
602 rules, such as for partitioning setups
604 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php
606 * Add the ability to automatically create materialized views
608 Right now materialized views require the user to create triggers on the
609 main table to keep the summary table current. SQL syntax should be able
610 to manager the triggers and summary table automatically. A more
611 sophisticated implementation would automatically retrieve from the
612 summary table when the main table is referenced, if possible.
619 * Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
620 * Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
621 * %Allow SET CONSTRAINTS to be qualified by schema/table name
622 * %Add a separate TRUNCATE permission
624 Currently only the owner can TRUNCATE a table because triggers are not
625 called, and the table is locked in exclusive mode.
627 * Fix TRUNCATE ... RESTART IDENTITY so its affect on sequences is rolled
628 back on transaction abort
629 * Allow PREPARE of cursors
630 * Allow finer control over the caching of prepared query plans
632 Currently queries prepared via the libpq API are planned on first
633 execute using the supplied parameters --- allow SQL PREPARE to do the
634 same. Also, allow control over replanning prepared queries either
635 manually or automatically when statistics for execute parameters
636 differ dramatically from those used during planning.
638 * Improve logging of prepared transactions recovered during startup
640 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php
642 * Improve failure message when DROP DATABASE is used on a database that
643 has prepared transactions
644 * Allow prepared transactions with temporary tables created and dropped
645 in the same transaction, and when an ON COMMIT DELETE ROWS temporary
648 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00047.php
650 * Add a GUC variable to warn about non-standard SQL usage in queries
651 * Add SQL-standard MERGE/REPLACE/UPSERT command
653 MERGE is typically used to merge two tables. REPLACE or UPSERT
654 command does UPDATE, or on failure, INSERT. This is similar to UPDATE,
655 then for unmatched rows, INSERT. Whether concurrent access allows
656 modifications which could cause row loss is implementation independent.
657 To implement this cleanly requires that the table have a unique index
658 so duplicate checking can be easily performed. It is possible to do it
659 without a unique index if we require the user to LOCK the table before
662 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php
663 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php
664 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01157.php
665 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01475.php
666 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01890.php
668 * Add NOVICE output level for helpful messages like automatic sequence/index
670 * Add GUC to issue notice about statements that use unjoined tables
671 * Allow EXPLAIN to identify tables that were skipped because of
673 * Allow EXPLAIN output to be more easily processed by scripts, perhaps XML
674 * Enable standard_conforming_strings
675 * Make standard_conforming_strings the default in 8.5?
677 When this is done, backslash-quote should be prohibited in non-E''
678 strings because of possible confusion over how such strings treat
679 backslashes. Basically, '' is always safe for a literal single
680 quote, while \' might or might not be based on the backslash
683 * Simplify dropping roles that have objects in several databases
684 * Allow COMMENT ON to accept an expression rather than just a string
685 * Allow the count returned by SELECT, etc to be represented as an int64
686 to allow a higher range of values
687 * Add SQL99 WITH clause to SELECT
688 * Add SQL:2003 WITH RECURSIVE (hierarchical) queries to SELECT
690 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php
691 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php
692 http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php
693 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01334.php
694 http://archives.postgresql.org/pgsql-patches/2008-01/msg00105.php
695 http://archives.postgresql.org/pgsql-patches/2008-03/msg00327.php
697 * Add DEFAULT .. AS OWNER so permission checks are done as the table
700 This would be useful for SERIAL nextval() calls and CHECK constraints.
702 * Allow DISTINCT to work in multiple-argument aggregate calls
703 * Add column to pg_stat_activity that shows the progress of long-running
704 commands like CREATE INDEX and VACUUM
706 http://archives.postgresql.org/pgsql-patches/2008-04/msg00203.php
708 * Implement SQL:2003 window functions
709 * Allow INSERT/UPDATE ... RETURNING inside a SELECT 'FROM' clause or
712 http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php
713 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php
714 http://archives.postgresql.org/pgsql-hackers/2008-06/msg00124.php
716 * Increase locking when DROPing objects so dependent objects cannot
717 get dropped while the DROP operation is happening
719 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php
721 * -Allow AS in "SELECT col AS label" to be optional in certain cases
723 * Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
724 RETURNING to supply values to the INSERT
725 http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979
727 * Add comments on system tables/columns using the information in
730 Ideally the information would be pulled from the SGML file
733 * Improve reporting of UNION type mismatches
735 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00944.php
736 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00597.php
741 o Allow CREATE TABLE AS to determine column lengths for complex
742 expressions like SELECT col1 || col2
744 o Have WITH CONSTRAINTS also create constraint indexes
746 http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php
748 o Have CONSTRAINT cname NOT NULL record the contraint name
750 Right now pg_attribute.attnotnull records the NOT NULL status
751 of the column, but does not record the contraint name
753 o Prevent concurrent CREATE TABLE table1 from sometimes returning
754 a cryptic error message
756 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00169.php
758 o Add CREATE SCHEMA ... LIKE that copies a schema
763 o Allow UPDATE tab SET ROW (col, ...) = (SELECT...)
765 http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php
766 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php
767 http://archives.postgresql.org/pgsql-patches/2007-04/msg00315.php
768 http://archives.postgresql.org/pgsql-patches/2008-03/msg00237.php
770 o Research self-referential UPDATEs that see inconsistent row versions
771 in read-committed mode
773 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php
774 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php
776 o Allow GLOBAL temporary tables to exist as empty by default in
779 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php
784 o Have ALTER TABLE RENAME rename SERIAL sequence names
786 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
788 o Have ALTER SEQUENCE RENAME rename the sequence name stored
789 in the sequence table
791 http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php
792 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php
793 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
795 o Add ALTER DOMAIN to modify the underlying data type
796 o %Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
798 http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php
800 o %Allow ALTER TABLE to change constraint deferrability and actions
801 o Add missing object types for ALTER ... SET SCHEMA
802 o Allow ALTER TABLESPACE to move to different directories
803 o Allow databases to be moved to different tablespaces
804 o Allow moving system tables to other tablespaces, where possible
806 Currently non-global system tables must be in the default database
807 tablespace. Global system tables can never be moved.
809 o -Prevent parent tables from altering or dropping constraints
810 like CHECK that are inherited by child tables unless CASCADE
812 o -Prevent child tables from altering or dropping constraints
813 like CHECK that were inherited from the parent table
814 o Have ALTER INDEX update the name of a constraint using that index
815 o Add ALTER TABLE RENAME CONSTRAINT, update index name also
816 o Allow column display reordering by recording a display,
817 storage, and permanent id for every column?
819 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php
821 o Allow an existing index to be marked as a table's primary key
823 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00500.php
828 o Automatically maintain clustering on a table
830 This might require some background daemon to maintain clustering
831 during periods of low usage. It might also require tables to be only
832 partially filled for easier reorganization. Another idea would
833 be to create a merged heap/index data file so an index lookup would
834 automatically access the heap data too. A third idea would be to
835 store heap rows in hashed groups, perhaps using a user-supplied
837 http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php
839 o %Add default clustering to system tables
841 To do this, determine the ideal cluster index for each system
842 table and set the cluster setting during initdb.
844 o %Add VERBOSE option to report tables as they are processed,
850 o Allow COPY to report error lines and continue
852 This requires the use of a savepoint before each COPY line is
853 processed, with ROLLBACK on COPY failure.
854 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00572.php
856 o Allow COPY on a newly-created table to skip WAL logging
858 On crash recovery, the table involved in the COPY would
859 be removed or have its heap and index files truncated. One
860 issue is that no other backend should be able to add to
861 the table at the same time, which is something that is
862 currently allowed. This currently is done if the table is
863 created inside the same transaction block as the COPY because
864 no other backends can see the table.
866 o Consider using a ring buffer for COPY FROM
868 http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php
869 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php
871 o Allow COPY FROM to create index entries in bulk
873 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php
875 o Allow COPY in CSV mode to control whether a quoted zero-length
876 string is treated as NULL
878 Currently this is always treated as a zero-length string,
879 which generates an error when loading into an integer column
880 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php
882 o Impove COPY performance
884 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00954.php
886 o Allow COPY to report errors sooner
888 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01169.php
894 o Allow column-level privileges
895 o %Allow GRANT/REVOKE permissions to be applied to all schema objects
898 The proposed syntax is:
899 GRANT SELECT ON ALL TABLES IN public TO phpuser;
900 GRANT SELECT ON NEW TABLES IN public TO phpuser;
902 o Allow GRANT/REVOKE permissions to be inherited by objects based on
905 o Allow SERIAL sequences to inherit permissions from the base table?
910 o Prevent DROP TABLE from dropping a row referenced by its own open
916 o Allow INSERT/UPDATE of the system-generated oid value for a row
917 o In rules, allow VALUES() to contain a mixture of 'old' and 'new'
923 o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
929 o Allow LISTEN/NOTIFY to store info in memory rather than tables?
931 Currently LISTEN/NOTIFY information is stored in pg_listener.
932 Storing such information in memory would improve performance.
934 o Add optional textual message to NOTIFY
936 This would allow an informational message to be added to the notify
937 message, perhaps indicating the row modified or other custom
940 o Allow multiple identical NOTIFY events to always be communicated
941 to the client, rather than sent as a single notification to the
944 http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php
946 o Allow NOTIFY in rules involving conditionals
947 o Improve LISTEN concurrency
949 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01106.php
953 Referential Integrity
954 =====================
956 * Add MATCH PARTIAL referential integrity
957 * Change foreign key constraint for array -> element to mean element
959 * Fix problem when cascading referential triggers make changes on
960 cascaded tables, seeing the tables in an intermediate state
962 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
963 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
965 * Allow DEFERRABLE and end-of-statement UNIQUE constraints?
967 This would allow UPDATE tab SET col = col + 1 to work if col has
968 a unique index. Currently, uniqueness checks are done while the
969 command is being executed, rather than at the end of the statement
971 http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html
972 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php
974 * Optimize referential integrity checks
976 http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php
977 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php
980 Server-Side Languages
981 =====================
984 o Fix RENAME to work on variables other than OLD/NEW
986 http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php
987 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php
988 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php
990 o Allow function parameters to be passed by name,
991 get_employee_salary(12345 AS emp_id, 2001 AS tax_year)
992 o Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
993 o Allow listing of record column names, and access to
994 record columns via variables, e.g. columns := r.(*),
997 http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php
998 http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php
999 http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php
1001 o Add support for SCROLL cursors
1002 o Add support for WITH HOLD cursors
1003 o Allow row and record variables to be set to NULL constants,
1004 and allow NULL tests on such variables
1006 Because a row is not scalar, do not allow assignment
1007 from NULL-valued scalars.
1009 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php
1011 o Review handling of MOVE and FETCH
1013 http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php
1015 o Improve logic of determining if an identifier is a a
1016 variable or column name
1018 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php
1020 o Consider keeping seperate cached copies when search_path changes
1022 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01009.php
1024 o -Add CASE capability to language (already in SQL)
1026 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00696.php
1032 o Add table function support to pltcl, plpythonu
1033 o Add support for polymorphic arguments and return types to
1034 languages other than PL/PgSQL
1035 o Add capability to create and call PROCEDURES
1036 o Add support for OUT and INOUT parameters to languages other
1038 o Add PL/PythonU tracebacks
1040 http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php
1042 o Allow data to be passed in native language formats, rather
1045 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php
1047 o Add ability to obfuscate function bodies
1049 http://archives.postgresql.org/pgsql-patches/2008-01/msg00125.php
1056 * Have pg_ctl look at PGHOST in case it is a socket directory?
1057 * Allow pg_ctl to work properly with configuration files located outside
1058 the PGDATA directory
1060 pg_ctl can not read the pid file because it isn't located in the
1061 config directory but in the PGDATA directory. The solution is to
1062 allow pg_ctl to read and understand postgresql.conf to find the
1063 data_directory value.
1065 * Add a function like pg_get_indexdef() that report more detailed index
1068 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php
1071 o Have psql show current values for a sequence
1072 o Move psql backslash database information into the backend, use
1075 This would allow non-psql clients to pull the same information out
1076 of the database as psql.
1077 http://archives.postgresql.org/pgsql-hackers/2004-01/msg00191.php
1079 o Make psql's \d commands more consistent
1081 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1082 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1084 o Consistently display privilege information for all objects in psql
1085 o Add auto-expanded mode so expanded output is used if the row
1086 length is wider than the screen width.
1088 Consider using auto-expanded mode for backslash commands like \df+.
1090 o Prevent tab completion of SET TRANSACTION from querying the
1091 database and therefore preventing the transaction isolation
1092 level from being set.
1094 Currently SET <tab> causes a database lookup to check all
1095 supported session variables. This query causes problems
1096 because setting the transaction isolation level must be the
1097 first statement of a transaction.
1099 o Add a \set variable to control whether \s displays line numbers
1101 Another option is to add \# which lists line numbers, and
1102 allows command execution.
1104 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php
1106 o Prevent escape string warnings when object names have
1109 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php
1111 o Have \d show foreign keys that reference a table's primary key
1113 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php
1115 o Have \d show child tables that inherit from the specified parent
1116 o -Have \l+ show database size, if permissions allow
1118 Ideally it will not generate an error for invalid permissions
1120 o Include the symbolic SQLSTATE name in verbose error reports
1122 http://archives.postgresql.org/pgsql-general/2007-09/msg00438.php
1124 o -Improve display of enums to show valid enum values
1125 o Add prompt escape to display the client and server versions
1126 o Add option to wrap column values at whitespace boundaries,
1127 rather than chopping them at a fixed width.
1129 Currently, "wrapped" format chops values into fixed
1130 widths. Perhaps the word wrapping could use the same
1131 algorithm documented in the W3C specification.
1132 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00404.php
1133 http://www.w3.org/TR/CSS21/tables.html#auto-table-layout
1134 o Add "auto" expanded mode that outputs in expanded format if
1135 "wrapped" mode can't wrap the output to the screen width
1137 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00417.php
1140 * pg_dump / pg_restore
1141 o %Add dumping of comments on index columns and composite type columns
1142 o %Add full object name to the tag field. eg. for operators we need
1143 '=(integer, integer)', instead of just '='.
1144 o Add pg_dumpall custom format dumps?
1145 o Allow selection of individual object(s) of all types, not just
1147 o In a selective dump, allow dumping of an object and all its
1149 o Add options like pg_restore -l and -L to pg_dump
1150 o Stop dumping CASCADE on DROP TYPE commands in clean mode
1151 o Allow pg_dump --clean to drop roles that own objects or have
1153 o Change pg_dump so that a comment on the dumped database is
1154 applied to the loaded database, even if the database has a
1155 different name. This will require new backend syntax, perhaps
1156 COMMENT ON CURRENT DATABASE.
1157 o Remove unnecessary function pointer abstractions in pg_dump source
1159 o Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
1160 multiple objects simultaneously
1162 The difficulty with this is getting multiple dump processes to
1163 produce a single dump output file. It also would require
1164 several sessions to share the same snapshot.
1165 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1167 o Allow pg_restore to utilize multiple CPUs and I/O channels by
1168 restoring multiple objects simultaneously
1170 This might require a pg_restore flag to indicate how many
1171 simultaneous operations should be performed. Only pg_dump's
1172 -Fc format has the necessary dependency information.
1173 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php
1175 o To better utilize resources, allow pg_restore to check foreign
1176 keys simultaneously, where possible
1177 o Allow pg_restore to create all indexes of a table
1178 concurrently, via a single heap scan
1180 This requires a pg_dump -Fc file because that format contains
1181 the required dependency information.
1182 http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php
1184 o Allow pg_restore to load different parts of the COPY data
1186 o -Prevent pg_dump/pg_restore from being affected by
1189 Using psql to restore a pg_dump dump is also affected.
1191 o Remove pre-7.3 pg_dump code that assumes pg_depend does not exit
1192 o Allow pre/data/post files when schema and data are dumped
1193 separately, for performance reasons
1195 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1202 Document differences between ecpg and the SQL standard and
1203 information about the Informix-compatibility module.
1205 o Solve cardinality > 1 for input descriptors / variables?
1206 o Add a semantic check level, e.g. check if a table really exists
1207 o fix handling of DB attributes that are arrays
1208 o Use backend PREPARE/EXECUTE facility for ecpg where possible
1210 o Fix nested C comments
1211 o %sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified
1212 o Make SET CONNECTION thread-aware, non-standard?
1213 o Allow multidimensional arrays
1214 o Add internationalized message strings
1215 o Implement COPY FROM STDIN
1219 o Add PQescapeIdentifierConn()
1220 o Prevent PQfnumber() from lowercasing unquoted the column name
1222 PQfnumber() should never have been doing lowercasing, but
1223 historically it has so we need a way to prevent it
1225 o Allow statement results to be automatically batched to the client
1227 Currently all statement results are transferred to the libpq
1228 client before libpq makes the results available to the
1229 application. This feature would allow the application to make
1230 use of the first result rows while the rest are transferred, or
1231 held on the server waiting for them to be requested by libpq.
1232 One complexity is that a statement like SELECT 1/col could error
1233 out mid-way through the result set.
1235 o Consider disallowing multiple queries in PQexec() as an
1236 additional barrier to SQL injection attacks
1238 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php
1240 o Add PQexecf() that allows complex parameter substitution
1242 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php
1244 o Add SQLSTATE severity to PGconn return status
1246 http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php
1252 * Add deferred trigger queue file
1254 Right now all deferred trigger information is stored in backend
1255 memory. This could exhaust memory for very large trigger queues.
1256 This item involves dumping large queues into files, or doing some
1257 kind of join to process all the triggers, or some bulk operation.
1258 http://archives.postgresql.org/pgsql-hackers/2008-05/msg00876.php
1260 * Allow triggers to be disabled in only the current session.
1262 This is currently possible by starting a multi-statement transaction,
1263 modifying the system tables, performing the desired SQL, restoring the
1264 system tables, and committing the transaction. ALTER TABLE ...
1265 TRIGGER requires a table lock so it is not ideal for this usage.
1267 * With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
1269 If the dump is known to be valid, allow foreign keys to be added
1270 without revalidating the data.
1272 * Allow statement-level triggers to access modified rows
1273 * Support triggers on columns
1275 http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php
1277 * Allow AFTER triggers on system tables
1279 System tables are modified in many places in the backend without going
1280 through the executor and therefore not causing triggers to fire. To
1281 complete this item, the functions that modify system tables will have
1284 * Tighten trigger permission checks
1286 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
1288 * Allow BEFORE INSERT triggers on views
1290 http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php
1292 * -Add ability to trigger on TRUNCATE
1293 * Add database and transaction-level triggers
1295 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00451.php
1302 * Add UNIQUE capability to non-btree indexes
1303 * Prevent index uniqueness checks when UPDATE does not modify the column
1305 Uniqueness (index) checks are done when updating a column even if the
1306 column is not modified by the UPDATE.
1308 * Allow the creation of on-disk bitmap indexes which can be quickly
1309 combined with other bitmap indexes
1311 Such indexes could be more compact if there are only a few distinct values.
1312 Such indexes can also be compressed. Keeping such indexes updated can be
1315 http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php
1316 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php
1317 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00265.php
1318 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01214.php
1319 http://archives.postgresql.org/pgsql-patches/2007-05/msg00013.php
1320 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00741.php
1322 * Allow accurate statistics to be collected on indexes with more than
1323 one column or expression indexes, perhaps using per-index statistics
1325 http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php
1326 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php
1328 * Consider increasing the default and maximum number of statistics targets,
1329 and reduce statistics target overhead
1331 Also consider having a larger statistics target for indexed columns
1332 and expression indexes.
1333 http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php
1334 http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php
1335 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01066.php
1336 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00188.php
1338 * Consider smaller indexes that record a range of values per heap page,
1339 rather than having one index entry for every heap row
1341 This is useful if the heap is clustered by the indexed values.
1342 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php
1343 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php
1344 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php
1345 http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php
1346 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00014.php
1347 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00487.php
1348 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01589.php
1350 * Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
1352 This is difficult because you must upgrade to an exclusive table lock
1353 to replace the existing index file. CREATE INDEX CONCURRENTLY does not
1354 have this complication. This would allow index compaction without
1356 http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php
1358 * Allow multiple indexes to be created concurrently, ideally via a
1359 single heap scan, and have pg_restore use it
1361 * Consider sorting entries before inserting into btree index
1363 http://archives.postgresql.org/pgsql-general/2008-01/msg01010.php
1365 * Allow index scans to return matching index keys, not just the matching
1368 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01657.php
1374 o Allow inherited tables to inherit indexes, UNIQUE constraints,
1375 and primary/foreign keys
1376 o Honor UNIQUE INDEX on base column in INSERTs/UPDATEs
1377 on inherited table, e.g. INSERT INTO inherit_table
1378 (unique_index_col) VALUES (dup) should fail
1380 The main difficulty with this item is the problem of
1381 creating an index that can span multiple tables.
1383 o Allow SELECT ... FOR UPDATE on inherited tables
1384 o Require all CHECK constraints to be inherited
1386 http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php
1388 o Add checks to prevent a CREATE RULE views on inherited tables
1390 http://archives.postgresql.org/pgsql-general/2008-02/msg01420.php
1391 http://archives.postgresql.org/pgsql-general/2008-03/msg00077.php
1396 o Add more GIST index support for geometric data types
1397 o Allow GIST indexes to create certain complex index types, like
1398 digital trees (see Aoki)
1403 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php
1405 o Pack hash index buckets onto disk pages more efficiently
1407 Currently only one hash bucket can be stored on a page. Ideally
1408 several hash buckets could be stored on a single page and greater
1409 granularity used for the hash algorithm.
1411 http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php
1413 o Consider sorting hash buckets so entries can be found using a
1414 binary search, rather than a linear scan
1415 o In hash indexes, consider storing the hash value with or instead
1417 o Add WAL logging for crash recovery
1418 o Allow multi-column hash indexes
1419 o -During index creation, pre-sort the tuples to improve build speed
1423 * Consider using hash buckets to do DISTINCT, rather than sorting
1425 This would be beneficial when there are few distinct values. This is
1426 already used by GROUP BY.
1428 * Consider whether duplicate keys should be sorted by block/offset
1430 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00558.php
1432 * -Avoid tuple some tuple copying in sort routines
1433 * Consider being smarter about memory and external files used during
1436 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01101.php
1437 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00045.php
1439 * Consider detoasting keys before sorting
1445 * Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
1447 Ideally this requires a separate test program that can be run
1448 at initdb time or optionally later. Consider O_SYNC when
1451 * Add program to test if fsync has a delay compared to non-fsync
1452 * Consider sorting writes during checkpoint
1454 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php
1463 We could use a fixed row count and a +/- count to follow MVCC
1464 visibility rules, or a single cached value could be used and
1465 invalidated if anyone modifies the table. Another idea is to
1466 get a count directly from a unique index, but for this to be
1467 faster than a sequential scan it must avoid access to the heap
1468 to obtain tuple visibility information.
1470 * Provide a way to calculate an "estimated COUNT(*)"
1472 Perhaps by using the optimizer's cardinality estimates or random
1475 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php
1477 * Allow data to be pulled directly from indexes
1479 Currently indexes do not have enough tuple visibility information
1480 to allow data to be pulled from the index without also accessing
1481 the heap. One way to allow this is to set a bit on index tuples
1482 to indicate if a tuple is currently visible to all transactions
1483 when the first valid heap lookup happens. This bit would have to
1484 be cleared when a heap tuple is expired.
1486 Another idea is to maintain a bitmap of heap pages where all rows
1487 are visible to all backends, and allow index lookups to reference
1488 that bitmap to avoid heap lookups, perhaps the same bitmap we might
1489 add someday to determine which heap pages need vacuuming. Frequently
1490 accessed bitmaps would have to be stored in shared memory. One 8k
1491 page of bitmaps could track 512MB of heap pages.
1493 A third idea would be for a heap scan to check if all rows are visible
1494 and if so set a per-table flag which can be checked by index scans.
1495 Any change to the table would have to clear the flag. To detect
1496 changes during the heap scan a counter could be set at the start and
1497 checked at the end --- if it is the same, the table has not been
1498 modified --- any table change would increment the counter.
1500 http://archives.postgresql.org/pgsql-patches/2007-10/msg00166.php
1501 http://archives.postgresql.org/pgsql-patches/2008-01/msg00049.php
1503 * Consider automatic caching of statements at various levels:
1506 o Query execute plan
1508 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00823.php
1510 * Consider increasing internal areas when shared buffers is increased
1512 http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php
1514 * Consider decreasing the amount of memory used by PrivateRefCount
1516 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php
1517 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php
1519 * Consider allowing higher priority queries to have referenced buffer
1520 cache pages stay in memory longer
1522 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php
1529 * Improve speed with indexes
1531 For large table adjustments during VACUUM FULL, it is faster to cluster
1532 or reindex rather than update the index. Also, index updates can bloat
1535 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php
1536 http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php
1537 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php
1539 * Auto-fill the free space map by scanning the buffer cache or by
1540 checking pages written by the background writer
1542 http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php
1543 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php
1545 * Create a bitmap of pages that need vacuuming
1547 Instead of sequentially scanning the entire table, have the background
1548 writer or some other process record pages that have expired rows, then
1549 VACUUM can look at just those pages rather than the entire table. In
1550 the event of a system crash, the bitmap would probably be invalidated.
1551 One complexity is that index entries still have to be vacuumed, and
1552 doing this without an index scan (by using the heap values to find the
1553 index entry) might be slow and unreliable, especially for user-defined
1556 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php
1557 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php
1558 http://archives.postgresql.org/pgsql-patches/2007-03/msg00508.php
1559 http://archives.postgresql.org/pgsql-patches/2007-04/msg00347.php
1560 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00156.php
1561 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00546.php
1562 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00416.php
1564 * Allow FSM to return free space toward the beginning of the heap file,
1565 in hopes that empty pages at the end can be truncated by VACUUM
1566 * Allow FSM page return free space based on table clustering, to assist
1567 in maintaining clustering?
1568 * -Improve dead row detection during multi-statement transactions usage
1570 http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php
1572 * Consider a more compact data representation for dead tuples
1574 http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php
1579 o %Issue log message to suggest VACUUM FULL if a table is nearly
1581 o Improve control of auto-vacuum
1583 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php
1585 o Prevent long-lived temporary tables from causing frozen-xid
1586 advancement starvation
1588 The problem is that autovacuum cannot vacuum them to set frozen xids;
1589 only the session that created them can do that.
1590 http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php
1592 o Store per-table autovacuum settings in pg_class.reloptions.
1594 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php
1595 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php
1597 o Prevent autovacuum from running if an old transaction is still
1598 running from the last vacuum
1600 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php
1607 * Fix priority ordering of read and write light-weight locks (Neil)
1609 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php
1610 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php
1612 * Fix problem when multiple subtransactions of the same outer transaction
1613 hold different types of locks, and one subtransaction aborts
1615 http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php
1616 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php
1617 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php
1618 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php
1620 * Allow UPDATEs on only non-referential integrity columns not to conflict
1621 with referential integrity locks
1623 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php
1625 * Add idle_in_transaction_timeout GUC so locks are not held for long
1628 * Improve deadlock detection when a page cleaning lock conflicts
1629 with a shared buffer that is pinned
1631 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php
1632 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1633 http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php
1635 * Detect deadlocks involving LockBufferForCleanup()
1637 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1641 Startup Time Improvements
1642 =========================
1644 * Experiment with multi-threaded backend for backend creation
1646 This would prevent the overhead associated with process creation. Most
1647 operating systems have trivial process creation time compared to
1648 database startup overhead, but a few operating systems (Win32,
1649 Solaris) might benefit from threading. Also explore the idea of
1650 a single session using multiple threads to execute a statement faster.
1657 * Eliminate need to write full pages to WAL before page modification
1659 Currently, to protect against partial disk page writes, we write
1660 full page images to WAL before they are modified so we can correct any
1661 partial page writes during recovery. These pages can also be
1662 eliminated from point-in-time archive files.
1663 http://archives.postgresql.org/pgsql-hackers/2002-06/msg00655.php
1665 o When off, write CRC to WAL and check file system blocks
1668 If CRC check fails during recovery, remember the page in case
1669 a later CRC for that page properly matches.
1671 o Write full pages during file system write and not when
1672 the page is modified in the buffer cache
1674 This allows most full page writes to happen in the background
1675 writer. It might cause problems for applying WAL on recovery
1676 into a partially-written page, but later the full page will be
1679 * Allow WAL traffic to be streamed to another server for stand-by
1681 * Reduce WAL traffic so only modified values are written rather than
1684 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php
1686 * Allow WAL information to recover corrupted pg_controldata
1688 http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php
1690 * Find a way to reduce rotational delay when repeatedly writing
1693 Currently fsync of WAL requires the disk platter to perform a full
1694 rotation to fsync again. One idea is to write the WAL to different
1695 offsets that might reduce the rotational delay.
1696 http://archives.postgresql.org/pgsql-hackers/2002-11/msg00483.php
1698 * Allow WAL logging to be turned off for a table, but the table
1699 might be dropped or truncated during crash recovery
1701 Allow tables to bypass WAL writes and just fsync() dirty pages on
1702 commit. This should be implemented using ALTER TABLE, e.g. ALTER
1703 TABLE PERSISTENCE [ DROP | TRUNCATE | DEFAULT ]. Tables using
1704 non-default logging should not use referential integrity with
1705 default-logging tables. A table without dirty buffers during a
1706 crash could perhaps avoid the drop/truncate.
1707 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1709 * Allow WAL logging to be turned off for a table, but the table would
1710 avoid being truncated/dropped
1712 To do this, only a single writer can modify the table, and writes
1713 must happen only on new pages so the new pages can be removed during
1714 crash recovery. Readers can continue accessing the table. Such
1715 tables probably cannot have indexes. One complexity is the handling
1716 of indexes on TOAST tables.
1717 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1719 * Speed WAL recovery by allowing more than one page to be prefetched
1721 This should be done utilizing the same infrastructure used for
1722 prefetching in general to avoid introducing complex error-prone code
1724 http://archives.postgresql.org/pgsql-general/2007-12/msg00683.php
1725 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00497.php
1726 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php
1728 * Improve WAL concurrency by increasing lock granularity
1730 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00556.php
1732 * Be more aggressive about creating WAL files
1734 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01325.php
1735 http://archives.postgresql.org/pgsql-hackers/2004-07/msg01075.php
1736 http://archives.postgresql.org/pgsql-hackers/2005-04/msg00556.php
1738 * Have resource managers report the duration of their status changes
1740 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01468.php
1742 * Move pgfoundry's xlogdump to /contrib and have it rely more closely
1743 on the WAL backend code
1745 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00035.php
1750 Optimizer / Executor
1751 ====================
1753 * Improve selectivity functions for geometric operators
1754 * Precompile SQL functions to avoid overhead
1755 * Create utility to compute accurate random_page_cost value
1756 * Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
1757 * Have EXPLAIN ANALYZE issue NOTICE messages when the estimated and
1758 actual row counts differ by a specified percentage
1759 * Improve how ANALYZE computes in-doubt tuples
1761 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00771.php
1763 * Log statements where the optimizer row estimates were dramatically
1764 different from the number of rows actually found?
1765 * Consider compressed annealing to search for query plans
1767 This might replace GEQO, http://sixdemonbag.org/Djinni.
1769 * Improve merge join performance by allowing mark/restore of
1772 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php
1774 * Consider using a hash for joining to a large IN (VALUES ...) list
1776 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php
1783 * Consider having the background writer update the transaction status
1784 hint bits before writing out the page
1786 Implementing this requires the background writer to have access to system
1787 catalogs and the transaction status log.
1789 * Consider adding buffers the background writer finds reusable to the
1792 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1794 * Automatically tune bgwriter_delay based on activity rather then using a
1797 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1799 * Consider wither increasing BM_MAX_USAGE_COUNT improves performance
1801 http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php
1803 * Test to see if calling PreallocXlogFiles() from the background writer
1804 will help with WAL segment creation latency
1806 http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php
1810 Miscellaneous Performance
1811 =========================
1813 * Do async I/O for faster random read-ahead of data
1815 Async I/O allows multiple I/O requests to be sent to the disk with
1816 results coming back asynchronously.
1818 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php
1819 http://archives.postgresql.org/pgsql-performance/2007-09/msg00255.php
1820 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php
1821 http://archives.postgresql.org/pgsql-patches/2008-01/msg00170.php
1823 * Use mmap() rather than SYSV shared memory or to write WAL files?
1825 This would remove the requirement for SYSV SHM but would introduce
1826 portability issues. Anonymous mmap (or mmap to /dev/zero) is required
1827 to prevent I/O overhead.
1829 * Consider mmap()'ing files into a backend?
1831 Doing I/O to large tables would consume a lot of address space or
1832 require frequent mapping/unmapping. Extending the file also causes
1833 mapping problems that might require mapping only individual pages,
1834 leading to thousands of mappings. Another problem is that there is no
1835 way to _prevent_ I/O to disk from the dirty shared buffers so changes
1836 could hit disk before WAL is written.
1838 * Add a script to ask system configuration questions and tune postgresql.conf
1839 * Consider ways of storing rows more compactly on disk
1841 o Reduce the row header size?
1842 o Consider reducing on-disk varlena length from four bytes to
1843 two because a heap row cannot be more than 64k in length
1845 * Consider increasing NUM_CLOG_BUFFERS
1847 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php
1848 http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php
1850 * Consider transaction start/end performance improvements
1852 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00948.php
1853 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1855 * Allow user configuration of TOAST thresholds
1857 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php
1858 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php
1860 * Allow configuration of backend priorities via the operating system
1862 Though backend priorities make priority inversion during lock
1863 waits possible, research shows that this is not a huge problem.
1865 http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php
1867 * Experiment with multi-threaded backend better I/O utilization
1869 This would allow a single query to make use of multiple I/O channels
1870 simultaneously. One idea is to create a background reader that can
1871 pre-fetch sequential and index scan pages needed by other backends.
1872 This could be expanded to allow concurrent reads from multiple devices
1873 in a partitioned table.
1875 * Experiment with multi-threaded backend better CPU utilization
1877 This would allow several CPUs to be used for a single query, such as
1878 for sorting or query execution.
1880 * Consider increasing the minimum allowed number of shared buffers
1882 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php
1884 * Expire published xmin for read-only and idle transactions
1886 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00343.php
1888 * Consider if CommandCounterIncrement() can avoid its
1889 AcceptInvalidationMessages() call
1891 http://archives.postgresql.org/pgsql-committers/2007-11/msg00585.php
1893 * Improve performance of shared invalidation queue for multiple CPUs
1895 http://archives.postgresql.org/pgsql-performance/2008-01/msg00023.php
1897 * Consider Cartesian joins when both relations are needed to form an
1898 indexscan qualification for a third relation
1900 http://archives.postgresql.org/pgsql-performance/2007-12/msg00090.php
1902 * Consider not storing a NULL bitmap on disk if all the NULLs are
1905 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00624.php
1906 http://archives.postgresql.org/pgsql-patches/2007-12/msg00109.php
1908 * Sort large UPDATE/DELETEs so it is done in heap order
1910 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01119.php
1912 * SMP scalability improvements
1914 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00439.php
1915 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00206.php
1916 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1918 * Research reducing deTOASTing in more places
1920 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00895.php
1922 * Allow one transaction to see tuples using the snapshot of another
1925 This would assist multiple backends in working together.
1926 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00400.php
1932 * Add use of 'const' for variables in source tree
1933 * Move some things from contrib into main tree
1934 * %Remove warnings created by -Wcast-align
1935 * Move platform-specific ps status display info from ps_status.c to ports
1936 * Add optional CRC checksum to heap and index pages
1937 * Improve documentation to build only interfaces (Marc)
1938 * Remove or relicense modules that are not under the BSD license, if possible
1939 * Acquire lock on a relation before building a relcache entry for it
1940 * Allow cross-compiling by generating the zic database on the target system
1941 * Improve NLS maintenance of libpgport messages linked onto applications
1942 * Clean up casting in contrib/isn
1944 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php
1946 * Improve the /contrib installation experience
1948 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00132.php
1950 * Use UTF8 encoding for NLS messages so all server encodings can
1952 * Update Bonjour to work with newer cross-platform SDK
1954 http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php
1955 http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php
1957 * Consider GnuTLS if OpenSSL license becomes a problem
1959 http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php
1960 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php
1962 * Consider changing documentation format from SGML to XML
1964 http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php
1966 * Consider making NAMEDATALEN more configurable in future releases
1967 * Update our code to handle 64-bit timezone files to match the zic
1968 source code, which now uses them
1969 * Have configure choose integer datetimes by default
1971 http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php
1973 * Support scoped IPv6 addresses
1975 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php
1977 * Consider allowing 64-bit integers and floats to be passed by value on
1980 Also change 32-bit floats (float4) to be passed by value at the same
1983 * Research use of signals and sleep wake ups
1985 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php
1987 * Add automated check for invalid C++ source code constructs
1989 http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php
1991 * Consider simplifying how memory context resets handle child contexts
1993 http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php
1995 * Remove use of MAKE_PTR and MAKE_OFFSET macros
1997 http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php
1999 * Convert single quotes to apostrophes in the PDF documentation
2001 http://archives.postgresql.org/pgsql-docs/2007-12/msg00059.php
2003 * Create three versions of libpgport to simplify client code
2005 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00154.php
2007 * Remove old-style routines for manipulating tuples
2009 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00851.php
2011 * Improve detection of shared memory segments being used by others
2012 by checking the SysV shared memory field 'nattch'
2014 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00656.php
2015 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00673.php
2017 * Implement the non-threaded Avahi service discovery protocol
2019 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00939.php
2020 http://archives.postgresql.org/pgsql-patches/2008-02/msg00097.php
2021 http://archives.postgresql.org/pgsql-hackers/2008-03/msg01211.php
2022 http://archives.postgresql.org/pgsql-patches/2008-04/msg00001.php
2024 * Implement a module capability for loading /contrib-style extensions
2026 http://archives.postgresql.org/pgsql-patches/2008-04/msg00164.php
2028 * Fix system views like pg_stat_all_tables to use set-returning
2029 functions, rather than views of per-column functions
2034 o Remove configure.in check for link failure when cause is found
2035 o Remove readdir() errno patch when runtime/mingwex/dirent.c rev
2037 o -Remove psql newline patch when we find out why mingw outputs an
2039 o Allow psql to use readline once non-US code pages work with
2041 o Fix problem with shared memory on the Win32 Terminal Server
2042 o Diagnose problem where shared memory can sometimes not be
2043 attached by postmaster children
2045 http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php
2047 o Improve signal handling
2049 http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php
2051 o Convert MSVC build system to remove most batch files
2053 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php
2055 o Prevent SSL from sending network packets to avoid interference
2056 with Win32 signal emulation
2058 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00455.php
2060 o Support pgxs when using MSVC
2062 o Fix MSVC NLS support, like for to_char()
2064 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00485.php
2065 http://archives.postgresql.org/pgsql-patches/2008-02/msg00038.php
2067 o Find a correct rint() substitute on Windows
2069 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00808.php
2073 * Wire Protocol Changes
2075 o Allow dynamic character set handling
2076 o Add decoded type, length, precision
2078 o Update clients to use data types, typmod, schema.table.column names
2079 of result sets using new statement protocol
2085 * Add pre-parsing phase that converts non-ISO syntax to supported
2088 This could allow SQL written for other databases to run without
2091 * Allow plug-in modules to emulate features from other databases
2092 * Add features of Oracle-style packages (Pavel)
2094 A package would be a schema with session-local variables,
2095 public/private functions, and initialization functions. It
2096 is also possible to implement these capabilities
2097 in any schema and not use a separate "packages"
2100 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php
2102 * Consider allowing control of upper/lower case folding of unquoted
2105 http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php
2106 http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php
2107 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00849.php
2109 * Add autonomous transactions
2111 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php
2115 Features We Do _Not_ Want
2116 =========================
2118 * All backends running as threads in a single process (not wanted)
2120 This eliminates the process protection we get from the current setup.
2121 Thread creation is usually the same overhead as process creation on
2122 modern systems, so it seems unwise to use a pure threaded model.
2124 * Optimizer hints (not wanted)
2126 Optimizer hints are used to work around problems in the optimizer. We
2127 would rather have the problems reported and fixed.
2129 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php
2130 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php
2131 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php
2134 Because we support postfix operators, it isn't possible to make AS
2135 optional and continue to use bison.
2136 http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php
2138 http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php
2140 * Embedded server (not wanted)
2142 While PostgreSQL clients runs fine in limited-resource environments, the
2143 server requires multiple processes and a stable pool of resources to
2144 run reliabily and efficiently. Stripping down the PostgreSQL server
2145 to run in the same process address space as the client application
2146 would add too much complexity and failure cases.