3 Current maintainer: Bruce Momjian (bruce@momjian.us)
4 Last updated: Wed May 7 21:45:08 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 statistics collector information to be pulled from the collector
37 process directly, rather than requiring the collector to write a
38 filesystem file twice a second?
39 * Reduce file system activity overhead of statistics file pgstat.stat
41 http://archives.postgresql.org/pgsql-general/2007-12/msg00106.php
43 * Allow statistics last vacuum/analyze execution times to be displayed
44 without requiring stats_row_level to be enabled
46 http://archives.postgresql.org/pgsql-docs/2007-04/msg00028.php
48 * Allow log_min_messages to be specified on a per-module basis
50 This would allow administrators to see more detailed information from
51 specific sections of the backend, e.g. checkpoints, autovacuum, etc.
52 Another idea is to allow separate configuration files for each module,
53 or allow arbitrary SET commands to be passed to them.
55 * Simplify ability to create partitioned tables
57 This would allow creation of partitioned tables without requiring
58 creation of triggers or rules for INSERT/UPDATE/DELETE, and constraints
59 for rapid partition selection. Options could include range and hash
62 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00375.php
63 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00151.php
64 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00028.php
65 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00248.php
66 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00387.php
67 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00413.php
69 * Allow auto-selection of partitioned tables for min/max() operations
70 * Allow more complex user/database default GUC settings
72 Currently ALTER USER and ALTER DATABASE support per-user and
73 per-database defaults. Consider adding per-user-and-database
74 defaults so things like search_path can be defaulted for a
75 specific user connecting to a specific database.
77 * Allow custom variable classes that can restrict who can set the values
79 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00911.php
81 * Implement the SQL standard mechanism whereby REVOKE ROLE revokes only
82 the privilege granted by the invoking role, and not those granted
85 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00010.php
87 * Allow SSL authentication/encryption over unix domain sockets
89 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00924.php
91 * Allow SSL key file permission checks to be optionally disabled when
92 sharing SSL keys with other applications
94 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00069.php
96 * Allow client certificate names to be checked against the client
99 This is already implemented in
100 libpq/fe-secure.c::verify_peer_name_matches_certificate() but the code
103 * Configuration files
105 o Allow pg_hba.conf to specify host names along with IP addresses
107 Host name lookup could occur when the postmaster reads the
108 pg_hba.conf file, or when the backend starts. Another
109 solution would be to reverse lookup the connection IP and
110 check that hostname against the host names in pg_hba.conf.
111 We could also then check that the host name maps to the IP
114 o %Allow postgresql.conf file values to be changed via an SQL
115 API, perhaps using SET GLOBAL
116 o Allow the server to be stopped/restarted via an SQL API
117 o Issue a warning if a change-on-restart-only postgresql.conf value
118 is modified and the server config files are reloaded
119 o Consider normalizing fractions in postgresql.conf, perhaps
122 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00550.php
124 o Allow Kerberos to disable stripping of realms so we can
125 check the username@realm against multiple realms
127 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00009.php
132 o Allow a database in tablespace t1 with tables created in
133 tablespace t2 to be used as a template for a new database created
134 with default tablespace t2
136 Currently all objects in the default database tablespace must
137 have default tablespace specifications. This is because new
138 databases are created by copying directories. If you mix default
139 tablespace tables and tablespace-specified tables in the same
140 directory, creating a new database from such a mixed directory
141 would create a new database with tables that had incorrect
142 explicit tablespaces. To fix this would require modifying
143 pg_class in the newly copied database, which we don't currently
146 o Allow reporting of which objects are in which tablespaces
148 This item is difficult because a tablespace can contain objects
149 from multiple databases. There is a server-side function that
150 returns the databases which use a specific tablespace, so this
151 requires a tool that will call that function and connect to each
152 database to find the objects in each database for that tablespace.
154 o Allow WAL replay of CREATE TABLESPACE to work when the directory
155 structure on the recovery computer is different from the original
157 o Allow per-tablespace quotas
160 * Point-In-Time Recovery (PITR)
162 o Allow a warm standby system to also allow read-only statements
164 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00050.php
166 o %Create dump tool for write-ahead logs for use in determining
167 transaction id for point-in-time recovery
169 This is useful for checking PITR recovery.
171 o Allow recovery.conf to support the same syntax as
172 postgresql.conf, including quoting
174 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php
176 o Fix server restart problem when the server was shutdown during
179 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00800.php
181 o Recreate pg_xlog/archive_status/ if it doesn't exist after
182 restoring from a PITR backup
184 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00487.php
190 * Change NUMERIC to enforce the maximum precision
191 * Reduce storage space for small NUMERICs
193 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01331.php
194 http://archives.postgresql.org/pgsql-patches/2007-02/msg00505.php
195 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00715.php
197 * Fix data types where equality comparison isn't intuitive, e.g. box
198 * Add support for public SYNONYMs
200 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00519.php
202 * Fix CREATE CAST on DOMAINs
204 http://archives.postgresql.org/pgsql-hackers/2006-05/msg00072.php
205 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01681.php
207 * Allow domains to be cast
209 http://archives.postgresql.org/pgsql-hackers/2003-06/msg01206.php
210 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00289.php
212 * Add support for SQL-standard GENERATED/IDENTITY columns
214 http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php
215 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00038.php
216 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00344.php
217 http://archives.postgresql.org/pgsql-patches/2007-05/msg00076.php
218 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00604.php
220 * Improve XML support
222 http://developer.postgresql.org/index.php/XML_Support
224 * Consider placing all sequences in a single table, or create a system
227 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
229 * Allow the UUID type to accept non-standard formats
231 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php
233 * Consider a special data type for regular expressions
235 http://archives.postgresql.org/pgsql-hackers/2007-08/msg01067.php
237 * Reduce BIT data type overhead using short varlena headers
239 http://archives.postgresql.org/pgsql-general/2007-12/msg00273.php
241 * Allow xml arrays to be cast to other data types
243 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00981.php
244 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00231.php
245 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00471.php
247 * Simplify integer cross-data-type operators
249 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00189.php
251 * Allow XML to accept more liberal DOCTYPE specifications
253 http://archives.postgresql.org/pgsql-general/2008-02/msg00347.php
255 * Allow adding/renaming/removing enumerated values to an existing
258 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01718.php
262 o Allow infinite dates and intervals just like infinite timestamps
263 o Merge hardwired timezone names with the TZ database; allow either
264 kind everywhere a TZ name is currently taken
265 o Allow TIMESTAMP WITH TIME ZONE to store the original timezone
266 information, either zone name or offset from UTC
268 If the TIMESTAMP value is stored with a time zone name, interval
269 computations should adjust based on the time zone rules.
270 http://archives.postgresql.org/pgsql-hackers/2004-10/msg00705.php
272 o Fix SELECT '0.01 years'::interval, '0.01 months'::interval
273 o Add a GUC variable to allow output of interval values in ISO8601
275 o Have timestamp subtraction not call justify_hours()?
277 http://archives.postgresql.org/pgsql-sql/2006-10/msg00059.php
279 o Improve timestamptz subtraction to be DST-aware
281 Currently subtracting one date from another that crosses a
282 daylight savings time adjustment can return '1 day 1 hour', but
283 adding that back to the first date returns a time one hour in
284 the future. This is caused by the adjustment of '25 hours' to
285 '1 day 1 hour', and '1 day' is the same time the next day, even
286 if daylight savings adjustments are involved.
288 o Fix interval display to support values exceeding 2^31 hours
289 o Add overflow checking to timestamp and interval arithmetic
290 o Extend timezone code to allow 64-bit values so we can
291 represent years beyond 2038
293 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01363.php
295 o Use LC_TIME for localized weekday/month names, rather than
298 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00390.php
300 o Add ISO INTERVAL handling
302 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00250.php
303 http://archives.postgresql.org/pgsql-bugs/2006-04/msg00248.php
305 o Support ISO INTERVAL syntax if units cannot be determined from
306 the string, and are supplied after the string
308 The SQL standard states that the units after the string
309 specify the units of the string, e.g. INTERVAL '2' MINUTE
310 should return '00:02:00'. The current behavior has the units
311 restrict the interval value to the specified unit or unit
312 range, INTERVAL '70' SECOND returns '00:00:10'.
314 For syntax that isn't uniquely ISO or PG syntax, like '1' or
315 '1:30', treat as ISO if there is a range specification clause,
316 and as PG if there no clause is present, e.g. interpret '1:30'
317 MINUTE TO SECOND as '1 minute 30 seconds', and interpret
318 '1:30' as '1 hour, 30 minutes'.
320 This makes common cases like SELECT INTERVAL '1' MONTH
321 SQL-standard results. The SQL standard supports a limited
322 number of unit combinations and doesn't support unit names in
323 the string. The PostgreSQL syntax is more flexible in the
324 range of units supported, e.g. PostgreSQL supports '1 year 1
325 hour', while the SQL standard does not.
327 o Add support for year-month syntax, INTERVAL '50-6' YEAR
329 o Interpret INTERVAL '1 year' MONTH as CAST (INTERVAL '1
330 year' AS INTERVAL MONTH), and this should return '12 months'
331 o Round or truncate values to the requested precision, e.g.
332 INTERVAL '11 months' AS YEAR should return one or zero
333 o Support precision, CREATE TABLE foo (a INTERVAL MONTH(3))
338 o Delay resolution of array expression's data type so assignment
339 coercion can be performed on empty array expressions
340 o Add support for arrays of domains
342 http://archives.postgresql.org/pgsql-patches/2007-05/msg00114.php
344 o Allow single-byte header storage for array elements
349 o Improve vacuum of large objects, like contrib/vacuumlo?
350 o Add security checking for large objects
351 o Auto-delete large objects when referencing row is deleted
353 contrib/lo offers this functionality.
355 o Allow read/write into TOAST values like large objects
357 This requires the TOAST column to be stored EXTERNAL.
359 o Add API for 64-bit large object access
361 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00781.php
365 * Add locale-aware MONEY type, and support multiple currencies
367 http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
368 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
370 * MONEY dumps in a locale-specific format making it difficult to
371 restore to a system with a different locale
372 * Allow MONEY to be easily cast to/from other numeric data types
376 o Allow dictionaries to change the token that is passed on to
379 http://archives.postgresql.org/pgsql-patches/2007-11/msg00081.php
381 o Consider a function-based API for '@@' searches
383 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00511.php
385 o Improve text search error messages
387 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00966.php
388 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01146.php
390 o Consider changing error to warning for strings larger than one
393 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00190.php
394 http://archives.postgresql.org/pgsql-patches/2008-03/msg00062.php
401 * Allow INET subnet tests using non-constants to be indexed
402 * Allow to_date() and to_timestamp() accept localized month names
403 * Fix to_date()-related functions to consistently issue errors
405 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00915.php
407 * Add missing parameter handling in to_char()
409 http://archives.postgresql.org/pgsql-hackers/2005-12/msg00948.php
411 * Allow substring/replace() to get/set bit values
412 * Allow to_char() on interval values to accumulate the highest unit
415 Some special format flag would be required to request such
416 accumulation. Such functionality could also be added to EXTRACT.
417 Prevent accumulation that crosses the month/day boundary because of
418 the uneven number of days in a month.
420 o to_char(INTERVAL '1 hour 5 minutes', 'MI') => 65
421 o to_char(INTERVAL '43 hours 20 minutes', 'MI' ) => 2600
422 o to_char(INTERVAL '43 hours 20 minutes', 'WK:DD:HR:MI') => 0:1:19:20
423 o to_char(INTERVAL '3 years 5 months','MM') => 41
425 * Implement inlining of set-returning functions defined in SQL
426 * Allow SQL-language functions to return results from RETURNING queries
428 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00665.php
430 * Allow SQL-language functions to reference parameters by parameter name
432 Currently SQL-language functions can only refer to dollar parameters,
435 * Add SPI_gettypmod() to return the typemod for a TupleDesc
436 * Enforce typmod for function inputs, function results and parameters for
437 spi_prepare'd statements called from PLs
439 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php
441 * Allow holdable cursors in SPI
442 * Tighten function permission checks
444 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php
446 * Fix IS OF so it matches the ISO specification, and add documentation
448 http://archives.postgresql.org/pgsql-patches/2003-08/msg00060.php
449 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00060.php
451 * Add missing operators for geometric data types
453 Some geometric types do not have the full suite of geometric operators,
456 * Implement Boyer-Moore searching in strpos()
458 http://archives.postgresql.org/pgsql-patches/2007-08/msg00012.php
460 * Prevent malicious functions from being executed with the permissions
461 of unsuspecting users
463 Index functions are safe, so VACUUM and ANALYZE are safe too.
464 Triggers, CHECK and DEFAULT expressions, and rules are still vulnerable.
465 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00268.php
467 * Reduce memory usage of aggregates in set returning functions
469 http://archives.postgresql.org/pgsql-performance/2008-01/msg00031.php
471 * Add temporal versions of generate_series()
473 http://archives.postgresql.org/pgsql-hackers/2007-04/msg01180.php
475 * Add array_accum() and array_to_set() functions for arrays
477 The standards specify array_agg() and UNNEST.
478 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00464.php
480 * Fix /contrib/ltree operator
482 http://archives.postgresql.org/pgsql-bugs/2007-11/msg00044.php
484 * Fix inconsistent precedence of =, >, and < compared to <>, >=, and <=
486 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00145.php
488 * Fix regular expression bug when using complex back-references
490 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00000.php
492 * Have /contrib/dblink reuse unnamed connections
494 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00895.php
496 * Add SQL-standard array_agg() and unnest() array functions
498 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01017.php
500 * Allow calling of a procedure outside a SELECT that can control the
503 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01375.php
507 Multi-Language Support
508 ======================
510 * Add NCHAR (as distinguished from ordinary varchar),
511 * Allow locale to be set at database creation
513 Currently locale can only be set during initdb. No global tables have
514 locale-aware columns. However, the database template used during
515 database creation might have locale-aware indexes. The indexes would
516 need to be reindexed to match the new locale.
518 * Allow encoding on a per-column basis optionally using the ICU library;
521 Right now only one encoding is allowed per database.
523 http://archives.postgresql.org/pgsql-hackers/2005-03/msg00932.php
524 http://archives.postgresql.org/pgsql-patches/2005-08/msg00039.php
525 http://archives.postgresql.org/pgsql-patches/2005-08/msg00309.php
526 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00110.php
527 http://archives.postgresql.org/pgsql-patches/2005-09/msg00020.php
528 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01121.php
529 http://archives.postgresql.org/pgsql-hackers/2006-01/msg00767.php
530 http://archives.postgresql.org/pgsql-patches/2006-03/msg00233.php
531 http://archives.postgresql.org/pgsql-hackers/2006-09/msg00662.php
532 http://wiki.postgresql.org/wiki/Todo:Collate
533 http://wiki.postgresql.org/wiki/Todo:ICU
535 * Support multiple simultaneous character sets, per SQL92
536 * Improve UTF8 combined character handling?
537 * Add octet_length_server() and octet_length_client()
538 * Make octet_length_client() the same as octet_length()?
539 * Fix problems with wrong runtime encoding conversion for NLS message files
540 * Add URL to more complete multi-byte regression tests
542 http://archives.postgresql.org/pgsql-hackers/2005-07/msg00272.php
544 * Fix ILIKE and regular expressions to handle case insensitivity
545 properly in multibyte encodings
547 http://archives.postgresql.org/pgsql-bugs/2005-10/msg00001.php
548 http://archives.postgresql.org/pgsql-patches/2005-11/msg00173.php
550 * Set client encoding based on the client operating system encoding
552 Currently client_encoding is set in postgresql.conf, which
553 defaults to the server encoding.
554 http://archives.postgresql.org/pgsql-hackers/2006-08/msg01696.php
556 * Change memory allocation for multi-byte functions so memory is
557 allocated inside conversion functions
559 Currently we preallocate memory based on worst-case usage.
566 * Automatically create rules on views so they are updateable, per SQL99
568 We can only auto-create rules for simple views. For more complex
569 cases users will still have to write rules manually.
571 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00586.php
572 http://archives.postgresql.org/pgsql-patches/2006-08/msg00255.php
574 * Add the functionality for WITH CHECK OPTION clause of CREATE VIEW
575 * Allow VIEW/RULE recompilation when the underlying tables change
577 Another issue is whether underlying table changes should be reflected
578 in the view, e.g. should SELECT * show additional columns if they
579 are added after the view is created.
581 * Make it possible to use RETURNING together with conditional DO INSTEAD
582 rules, such as for partitioning setups
584 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00577.php
586 * Add the ability to automatically create materialized views
588 Right now materialized views require the user to create triggers on the
589 main table to keep the summary table current. SQL syntax should be able
590 to manager the triggers and summary table automatically. A more
591 sophisticated implementation would automatically retrieve from the
592 summary table when the main table is referenced, if possible.
599 * Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
600 * Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
601 * %Allow SET CONSTRAINTS to be qualified by schema/table name
602 * %Add a separate TRUNCATE permission
604 Currently only the owner can TRUNCATE a table because triggers are not
605 called, and the table is locked in exclusive mode.
607 * Allow PREPARE of cursors
608 * Allow finer control over the caching of prepared query plans
610 Currently queries prepared via the libpq API are planned on first
611 execute using the supplied parameters --- allow SQL PREPARE to do the
612 same. Also, allow control over replanning prepared queries either
613 manually or automatically when statistics for execute parameters
614 differ dramatically from those used during planning.
616 * Improve logging of prepared transactions recovered during startup
618 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00092.php
620 * Improve failure message when DROP DATABASE is used on a database that
621 has prepared transactions
622 * Allow prepared transactions with temporary tables created and dropped
623 in the same transaction, and when an ON COMMIT DELETE ROWS temporary
626 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00047.php
628 * Add a GUC variable to warn about non-standard SQL usage in queries
629 * Add SQL-standard MERGE/REPLACE/UPSERT command
631 MERGE is typically used to merge two tables. REPLACE or UPSERT
632 command does UPDATE, or on failure, INSERT. This is similar to UPDATE,
633 then for unmatched rows, INSERT. Whether concurrent access allows
634 modifications which could cause row loss is implementation independent.
635 To implement this cleanly requires that the table have a unique index
636 so duplicate checking can be easily performed. It is possible to do it
637 without a unique index if we require the user to LOCK the table before
640 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00501.php
641 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00536.php
642 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01475.php
644 * Add NOVICE output level for helpful messages like automatic sequence/index
646 * Add GUC to issue notice about statements that use unjoined tables
647 * Allow EXPLAIN to identify tables that were skipped because of
649 * Allow EXPLAIN output to be more easily processed by scripts, perhaps XML
650 * Enable standard_conforming_strings
651 * Make standard_conforming_strings the default in 8.5?
653 When this is done, backslash-quote should be prohibited in non-E''
654 strings because of possible confusion over how such strings treat
655 backslashes. Basically, '' is always safe for a literal single
656 quote, while \' might or might not be based on the backslash
659 * Simplify dropping roles that have objects in several databases
660 * Allow COMMENT ON to accept an expression rather than just a string
661 * Allow the count returned by SELECT, etc to be represented as an int64
662 to allow a higher range of values
663 * Add SQL99 WITH clause to SELECT
664 * Add SQL:2003 WITH RECURSIVE (hierarchical) queries to SELECT
666 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01375.php
667 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00642.php
668 http://archives.postgresql.org/pgsql-patches/2007-03/msg00139.php
669 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01334.php
670 http://archives.postgresql.org/pgsql-patches/2008-01/msg00105.php
671 http://archives.postgresql.org/pgsql-patches/2008-03/msg00327.php
673 * Add DEFAULT .. AS OWNER so permission checks are done as the table
676 This would be useful for SERIAL nextval() calls and CHECK constraints.
678 * Allow DISTINCT to work in multiple-argument aggregate calls
679 * Add column to pg_stat_activity that shows the progress of long-running
680 commands like CREATE INDEX and VACUUM
681 * Implement SQL:2003 window functions
682 * Allow INSERT/UPDATE ... RETURNING inside a SELECT 'FROM' clause
684 http://archives.postgresql.org/pgsql-general/2006-09/msg00803.php
685 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00693.php
687 * Increase locking when DROPing objects so dependent objects cannot
688 get dropped while the DROP operation is happening
690 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00937.php
692 * -Allow AS in "SELECT col AS label" to be optional in certain cases
694 * Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
695 RETURNING to supply values to the INSERT
696 http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979
698 * Add comments on system tables/columns using the information in
701 Ideally the information would be pulled from the SGML file
704 * Improve reporting of UNION type mismatches
706 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00944.php
707 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00597.php
712 o Allow CREATE TABLE AS to determine column lengths for complex
713 expressions like SELECT col1 || col2
715 o Have WITH CONSTRAINTS also create constraint indexes
717 http://archives.postgresql.org/pgsql-patches/2007-04/msg00149.php
719 o Have CONSTRAINT cname NOT NULL record the contraint name
721 Right now pg_attribute.attnotnull records the NOT NULL status
722 of the column, but does not record the contraint name
724 o Prevent concurrent CREATE TABLE table1 from sometimes returning
725 a cryptic error message
727 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00169.php
731 o Allow UPDATE tab SET ROW (col, ...) = (SELECT...)
733 http://archives.postgresql.org/pgsql-hackers/2006-07/msg01306.php
734 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00865.php
735 http://archives.postgresql.org/pgsql-patches/2007-04/msg00315.php
736 http://archives.postgresql.org/pgsql-patches/2008-03/msg00237.php
738 o Research self-referential UPDATEs that see inconsistent row versions
739 in read-committed mode
741 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00507.php
742 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00016.php
744 o Allow GLOBAL temporary tables to exist as empty by default in
747 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00006.php
752 o Have ALTER TABLE RENAME rename SERIAL sequence names
754 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
756 o Have ALTER SEQUENCE RENAME rename the sequence name stored
757 in the sequence table
759 http://archives.postgresql.org/pgsql-bugs/2007-09/msg00092.php
760 http://archives.postgresql.org/pgsql-bugs/2007-10/msg00007.php
761 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
763 o Add ALTER DOMAIN to modify the underlying data type
764 o %Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
766 http://archives.postgresql.org/pgsql-patches/2006-02/msg00168.php
768 o %Allow ALTER TABLE to change constraint deferrability and actions
769 o Add missing object types for ALTER ... SET SCHEMA
770 o Allow ALTER TABLESPACE to move to different directories
771 o Allow databases to be moved to different tablespaces
772 o Allow moving system tables to other tablespaces, where possible
774 Currently non-global system tables must be in the default database
775 tablespace. Global system tables can never be moved.
777 o Prevent parent tables from altering or dropping constraints
778 like CHECK that are inherited by child tables unless CASCADE
780 o %Prevent child tables from altering or dropping constraints
781 like CHECK that were inherited from the parent table
782 o Have ALTER INDEX update the name of a constraint using that index
783 o Add ALTER TABLE RENAME CONSTRAINT, update index name also
784 o Allow column display reordering by recording a display,
785 storage, and permanent id for every column?
787 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php
793 o Automatically maintain clustering on a table
795 This might require some background daemon to maintain clustering
796 during periods of low usage. It might also require tables to be only
797 partially filled for easier reorganization. Another idea would
798 be to create a merged heap/index data file so an index lookup would
799 automatically access the heap data too. A third idea would be to
800 store heap rows in hashed groups, perhaps using a user-supplied
802 http://archives.postgresql.org/pgsql-performance/2004-08/msg00349.php
804 o %Add default clustering to system tables
806 To do this, determine the ideal cluster index for each system
807 table and set the cluster setting during initdb.
809 o %Add VERBOSE option to report tables as they are processed,
815 o Allow COPY to report error lines and continue
817 This requires the use of a savepoint before each COPY line is
818 processed, with ROLLBACK on COPY failure.
819 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00572.php
821 o Allow COPY on a newly-created table to skip WAL logging
823 On crash recovery, the table involved in the COPY would
824 be removed or have its heap and index files truncated. One
825 issue is that no other backend should be able to add to
826 the table at the same time, which is something that is
827 currently allowed. This currently is done if the table is
828 created inside the same transaction block as the COPY because
829 no other backends can see the table.
831 o Consider using a ring buffer for COPY FROM
833 http://archives.postgresql.org/pgsql-patches/2008-02/msg00140.php
834 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php
836 o Allow COPY FROM to create index entries in bulk
838 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00811.php
840 o Allow COPY in CSV mode to control whether a quoted zero-length
841 string is treated as NULL
843 Currently this is always treated as a zero-length string,
844 which generates an error when loading into an integer column
845 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00905.php
847 o Impove COPY performance
849 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00954.php
855 o Allow column-level privileges
856 o %Allow GRANT/REVOKE permissions to be applied to all schema objects
859 The proposed syntax is:
860 GRANT SELECT ON ALL TABLES IN public TO phpuser;
861 GRANT SELECT ON NEW TABLES IN public TO phpuser;
863 o Allow GRANT/REVOKE permissions to be inherited by objects based on
866 o Allow SERIAL sequences to inherit permissions from the base table?
871 o Prevent DROP TABLE from dropping a row referenced by its own open
877 o Allow INSERT/UPDATE of the system-generated oid value for a row
878 o In rules, allow VALUES() to contain a mixture of 'old' and 'new'
884 o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
890 o Allow LISTEN/NOTIFY to store info in memory rather than tables?
892 Currently LISTEN/NOTIFY information is stored in pg_listener.
893 Storing such information in memory would improve performance.
895 o Add optional textual message to NOTIFY
897 This would allow an informational message to be added to the notify
898 message, perhaps indicating the row modified or other custom
901 o Allow multiple identical NOTIFY events to always be communicated
902 to the client, rather than sent as a single notification to the
905 http://archives.postgresql.org/pgsql-general/2008-01/msg00057.php
907 o Allow NOTIFY in rules involving conditionals
908 o Improve LISTEN concurrency
910 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01106.php
914 Referential Integrity
915 =====================
917 * Add MATCH PARTIAL referential integrity
918 * Change foreign key constraint for array -> element to mean element
920 * Fix problem when cascading referential triggers make changes on
921 cascaded tables, seeing the tables in an intermediate state
923 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
924 http://archives.postgresql.org/pgsql-hackers/2005-09/msg00174.php
926 * Allow DEFERRABLE and end-of-statement UNIQUE constraints?
928 This would allow UPDATE tab SET col = col + 1 to work if col has
929 a unique index. Currently, uniqueness checks are done while the
930 command is being executed, rather than at the end of the statement
932 http://people.planetpostgresql.org/greg/index.php?/archives/2006/06/10.html
933 http://archives.postgresql.org/pgsql-hackers/2006-09/msg01458.php
935 * Optimize referential integrity checks
937 http://archives.postgresql.org/pgsql-performance/2005-10/msg00458.php
938 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00744.php
941 Server-Side Languages
942 =====================
945 o Fix RENAME to work on variables other than OLD/NEW
947 http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php
948 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php
949 http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php
951 o Allow function parameters to be passed by name,
952 get_employee_salary(12345 AS emp_id, 2001 AS tax_year)
953 o Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
954 o Allow listing of record column names, and access to
955 record columns via variables, e.g. columns := r.(*),
958 http://archives.postgresql.org/pgsql-patches/2005-07/msg00458.php
959 http://archives.postgresql.org/pgsql-patches/2006-05/msg00302.php
960 http://archives.postgresql.org/pgsql-patches/2006-06/msg00031.php
962 o Add support for SCROLL cursors
963 o Add support for WITH HOLD cursors
964 o Allow row and record variables to be set to NULL constants,
965 and allow NULL tests on such variables
967 Because a row is not scalar, do not allow assignment
968 from NULL-valued scalars.
970 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00070.php
972 o Review handling of MOVE and FETCH
974 http://archives.postgresql.org/pgsql-patches/2007-04/msg00527.php
976 o Improve logic of determining if an identifier is a a
977 variable or column name
979 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00436.php
981 o Consider keeping seperate cached copies when search_path changes
983 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01009.php
985 o Add CASE capability to language (already in SQL)
987 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00696.php
993 o Add table function support to pltcl, plpythonu
994 o Add support for polymorphic arguments and return types to
995 languages other than PL/PgSQL
996 o Add capability to create and call PROCEDURES
997 o Add support for OUT and INOUT parameters to languages other
999 o Add PL/PythonU tracebacks
1001 http://archives.postgresql.org/pgsql-patches/2006-02/msg00288.php
1003 o Allow data to be passed in native language formats, rather
1006 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00289.php
1008 o Add ability to obfuscate function bodies
1010 http://archives.postgresql.org/pgsql-patches/2008-01/msg00125.php
1017 * Have pg_ctl look at PGHOST in case it is a socket directory?
1018 * Allow pg_ctl to work properly with configuration files located outside
1019 the PGDATA directory
1021 pg_ctl can not read the pid file because it isn't located in the
1022 config directory but in the PGDATA directory. The solution is to
1023 allow pg_ctl to read and understand postgresql.conf to find the
1024 data_directory value.
1026 * Add a function like pg_get_indexdef() that report more detailed index
1029 http://archives.postgresql.org/pgsql-bugs/2007-12/msg00166.php
1032 o Have psql show current values for a sequence
1033 o Move psql backslash database information into the backend, use
1036 This would allow non-psql clients to pull the same information out
1037 of the database as psql.
1038 http://archives.postgresql.org/pgsql-hackers/2004-01/msg00191.php
1040 o Make psql's \d commands more consistent
1042 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1043 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00014.php
1045 o Consistently display privilege information for all objects in psql
1046 o Add auto-expanded mode so expanded output is used if the row
1047 length is wider than the screen width.
1049 Consider using auto-expanded mode for backslash commands like \df+.
1051 o Prevent tab completion of SET TRANSACTION from querying the
1052 database and therefore preventing the transaction isolation
1053 level from being set.
1055 Currently SET <tab> causes a database lookup to check all
1056 supported session variables. This query causes problems
1057 because setting the transaction isolation level must be the
1058 first statement of a transaction.
1060 o Add a \set variable to control whether \s displays line numbers
1062 Another option is to add \# which lists line numbers, and
1063 allows command execution.
1065 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00255.php
1067 o Prevent escape string warnings when object names have
1070 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00227.php
1072 o Have \d show foreign keys that reference a table's primary key
1074 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00424.php
1076 o Have \d show child tables that inherit from the specified parent
1077 o Have \l+ show database size, if permissions allow
1079 Ideally it will not generate an error for invalid permissions
1081 o Include the symbolic SQLSTATE name in verbose error reports
1083 http://archives.postgresql.org/pgsql-general/2007-09/msg00438.php
1085 o -Improve display of enums to show valid enum values
1086 o Add prompt escape to display the client and server versions
1089 * pg_dump / pg_restore
1090 o %Add dumping of comments on index columns and composite type columns
1091 o %Add full object name to the tag field. eg. for operators we need
1092 '=(integer, integer)', instead of just '='.
1093 o Add pg_dumpall custom format dumps?
1094 o Allow selection of individual object(s) of all types, not just
1096 o In a selective dump, allow dumping of an object and all its
1098 o Add options like pg_restore -l and -L to pg_dump
1099 o Stop dumping CASCADE on DROP TYPE commands in clean mode
1100 o Allow pg_dump --clean to drop roles that own objects or have
1102 o Change pg_dump so that a comment on the dumped database is
1103 applied to the loaded database, even if the database has a
1104 different name. This will require new backend syntax, perhaps
1105 COMMENT ON CURRENT DATABASE.
1106 o Remove unnecessary function pointer abstractions in pg_dump source
1108 o Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
1109 multiple objects simultaneously
1111 The difficulty with this is getting multiple dump processes to
1112 produce a single dump output file. It also would require
1113 several sessions to share the same snapshot.
1114 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1116 o Allow pg_restore to utilize multiple CPUs and I/O channels by
1117 restoring multiple objects simultaneously
1119 This might require a pg_restore flag to indicate how many
1120 simultaneous operations should be performed. Only pg_dump's
1121 -Fc format has the necessary dependency information.
1122 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php
1124 o To better utilize resources, allow pg_restore to check foreign
1125 keys simultaneously, where possible
1126 o Allow pg_restore to create all indexes of a table
1127 concurrently, via a single heap scan
1129 This requires a pg_dump -Fc file because that format contains
1130 the required dependency information.
1131 http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php
1133 o Allow pg_restore to load different parts of the COPY data
1135 o Prevent pg_dump/pg_restore from being affected by
1138 Using psql to restore a pg_dump dump is also affected.
1140 o Remove pre-7.3 pg_dump code that assumes pg_depend does not exit
1141 o Allow pre/data/post files when schema and data are dumped
1142 separately, for performance reasons
1144 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
1151 Document differences between ecpg and the SQL standard and
1152 information about the Informix-compatibility module.
1154 o Solve cardinality > 1 for input descriptors / variables?
1155 o Add a semantic check level, e.g. check if a table really exists
1156 o fix handling of DB attributes that are arrays
1157 o Use backend PREPARE/EXECUTE facility for ecpg where possible
1159 o Fix nested C comments
1160 o %sqlwarn[6] should be 'W' if the PRECISION or SCALE value specified
1161 o Make SET CONNECTION thread-aware, non-standard?
1162 o Allow multidimensional arrays
1163 o Add internationalized message strings
1164 o Implement COPY FROM STDIN
1168 o Add PQescapeIdentifierConn()
1169 o Prevent PQfnumber() from lowercasing unquoted the column name
1171 PQfnumber() should never have been doing lowercasing, but
1172 historically it has so we need a way to prevent it
1174 o Allow statement results to be automatically batched to the client
1176 Currently all statement results are transferred to the libpq
1177 client before libpq makes the results available to the
1178 application. This feature would allow the application to make
1179 use of the first result rows while the rest are transferred, or
1180 held on the server waiting for them to be requested by libpq.
1181 One complexity is that a statement like SELECT 1/col could error
1182 out mid-way through the result set.
1184 o Consider disallowing multiple queries in PQexec() as an
1185 additional barrier to SQL injection attacks
1187 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00184.php
1189 o Add PQexecf() that allows complex parameter substitution
1191 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01803.php
1193 o Add SQLSTATE severity to PGconn return status
1195 http://archives.postgresql.org/pgsql-interfaces/2007-11/msg00015.php
1201 * Add deferred trigger queue file
1203 Right now all deferred trigger information is stored in backend
1204 memory. This could exhaust memory for very large trigger queues.
1205 This item involves dumping large queues into files.
1207 * Allow triggers to be disabled in only the current session.
1209 This is currently possible by starting a multi-statement transaction,
1210 modifying the system tables, performing the desired SQL, restoring the
1211 system tables, and committing the transaction. ALTER TABLE ...
1212 TRIGGER requires a table lock so it is not ideal for this usage.
1214 * With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
1216 If the dump is known to be valid, allow foreign keys to be added
1217 without revalidating the data.
1219 * Allow statement-level triggers to access modified rows
1220 * Support triggers on columns
1222 http://archives.postgresql.org/pgsql-patches/2005-07/msg00107.php
1224 * Allow AFTER triggers on system tables
1226 System tables are modified in many places in the backend without going
1227 through the executor and therefore not causing triggers to fire. To
1228 complete this item, the functions that modify system tables will have
1231 * Tighten trigger permission checks
1233 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
1235 * Allow BEFORE INSERT triggers on views
1237 http://archives.postgresql.org/pgsql-general/2007-02/msg01466.php
1239 * -Add ability to trigger on TRUNCATE
1240 * Add database and transaction-level triggers
1242 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00451.php
1249 * Add UNIQUE capability to non-btree indexes
1250 * Prevent index uniqueness checks when UPDATE does not modify the column
1252 Uniqueness (index) checks are done when updating a column even if the
1253 column is not modified by the UPDATE.
1255 * Allow the creation of on-disk bitmap indexes which can be quickly
1256 combined with other bitmap indexes
1258 Such indexes could be more compact if there are only a few distinct values.
1259 Such indexes can also be compressed. Keeping such indexes updated can be
1262 http://archives.postgresql.org/pgsql-patches/2005-07/msg00512.php
1263 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01107.php
1264 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00265.php
1265 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01214.php
1266 http://archives.postgresql.org/pgsql-patches/2007-05/msg00013.php
1267 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00741.php
1269 * Allow accurate statistics to be collected on indexes with more than
1270 one column or expression indexes, perhaps using per-index statistics
1272 http://archives.postgresql.org/pgsql-performance/2006-10/msg00222.php
1273 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01131.php
1275 * Consider increasing the default and maximum number of statistics targets,
1276 and reduce statistics target overhead
1278 Also consider having a larger statistics target for indexed columns
1279 and expression indexes.
1280 http://archives.postgresql.org/pgsql-general/2007-05/msg01228.php
1281 http://archives.postgresql.org/pgsql-general/2007-06/msg00542.php
1282 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01066.php
1283 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00188.php
1285 * Consider smaller indexes that record a range of values per heap page,
1286 rather than having one index entry for every heap row
1288 This is useful if the heap is clustered by the indexed values.
1289 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00341.php
1290 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01264.php
1291 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00465.php
1292 http://archives.postgresql.org/pgsql-patches/2007-03/msg00163.php
1293 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00014.php
1294 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00487.php
1295 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01589.php
1297 * Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
1299 This is difficult because you must upgrade to an exclusive table lock
1300 to replace the existing index file. CREATE INDEX CONCURRENTLY does not
1301 have this complication. This would allow index compaction without
1303 http://archives.postgresql.org/pgsql-performance/2007-08/msg00289.php
1305 * Allow multiple indexes to be created concurrently, ideally via a
1306 single heap scan, and have pg_restore use it
1308 * Consider sorting entries before inserting into btree index
1310 http://archives.postgresql.org/pgsql-general/2008-01/msg01010.php
1312 * Allow index scans to return matching index keys, not just the matching
1315 http://archives.postgresql.org/pgsql-hackers/2008-04/msg01657.php
1321 o Allow inherited tables to inherit indexes, UNIQUE constraints,
1322 and primary/foreign keys
1323 o Honor UNIQUE INDEX on base column in INSERTs/UPDATEs
1324 on inherited table, e.g. INSERT INTO inherit_table
1325 (unique_index_col) VALUES (dup) should fail
1327 The main difficulty with this item is the problem of
1328 creating an index that can span multiple tables.
1330 o Allow SELECT ... FOR UPDATE on inherited tables
1331 o Require all CHECK constraints to be inherited
1333 http://archives.postgresql.org/pgsql-bugs/2007-04/msg00026.php
1335 o Add checks to prevent a CREATE RULE views on inherited tables
1337 http://archives.postgresql.org/pgsql-general/2008-02/msg01420.php
1338 http://archives.postgresql.org/pgsql-general/2008-03/msg00077.php
1343 o Add more GIST index support for geometric data types
1344 o Allow GIST indexes to create certain complex index types, like
1345 digital trees (see Aoki)
1350 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00051.php
1352 o Pack hash index buckets onto disk pages more efficiently
1354 Currently only one hash bucket can be stored on a page. Ideally
1355 several hash buckets could be stored on a single page and greater
1356 granularity used for the hash algorithm.
1358 http://archives.postgresql.org/pgsql-hackers/2004-06/msg00168.php
1360 o Consider sorting hash buckets so entries can be found using a
1361 binary search, rather than a linear scan
1362 o In hash indexes, consider storing the hash value with or instead
1364 o Add WAL logging for crash recovery
1365 o Allow multi-column hash indexes
1366 o -During index creation, pre-sort the tuples to improve build speed
1373 * Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
1375 Ideally this requires a separate test program that can be run
1376 at initdb time or optionally later. Consider O_SYNC when
1379 * Add program to test if fsync has a delay compared to non-fsync
1380 * Consider sorting writes during checkpoint
1382 http://archives.postgresql.org/pgsql-hackers/2007-06/msg00541.php
1391 We could use a fixed row count and a +/- count to follow MVCC
1392 visibility rules, or a single cached value could be used and
1393 invalidated if anyone modifies the table. Another idea is to
1394 get a count directly from a unique index, but for this to be
1395 faster than a sequential scan it must avoid access to the heap
1396 to obtain tuple visibility information.
1398 * Provide a way to calculate an "estimated COUNT(*)"
1400 Perhaps by using the optimizer's cardinality estimates or random
1403 http://archives.postgresql.org/pgsql-hackers/2005-11/msg00943.php
1405 * Allow data to be pulled directly from indexes
1407 Currently indexes do not have enough tuple visibility information
1408 to allow data to be pulled from the index without also accessing
1409 the heap. One way to allow this is to set a bit on index tuples
1410 to indicate if a tuple is currently visible to all transactions
1411 when the first valid heap lookup happens. This bit would have to
1412 be cleared when a heap tuple is expired.
1414 Another idea is to maintain a bitmap of heap pages where all rows
1415 are visible to all backends, and allow index lookups to reference
1416 that bitmap to avoid heap lookups, perhaps the same bitmap we might
1417 add someday to determine which heap pages need vacuuming. Frequently
1418 accessed bitmaps would have to be stored in shared memory. One 8k
1419 page of bitmaps could track 512MB of heap pages.
1421 A third idea would be for a heap scan to check if all rows are visible
1422 and if so set a per-table flag which can be checked by index scans.
1423 Any change to the table would have to clear the flag. To detect
1424 changes during the heap scan a counter could be set at the start and
1425 checked at the end --- if it is the same, the table has not been
1426 modified --- any table change would increment the counter.
1428 http://archives.postgresql.org/pgsql-patches/2007-10/msg00166.php
1429 http://archives.postgresql.org/pgsql-patches/2008-01/msg00049.php
1431 * Consider automatic caching of statements at various levels:
1434 o Query execute plan
1436 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00823.php
1438 * Consider increasing internal areas when shared buffers is increased
1440 http://archives.postgresql.org/pgsql-hackers/2005-10/msg01419.php
1442 * Consider decreasing the amount of memory used by PrivateRefCount
1444 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00797.php
1445 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00752.php
1447 * Consider allowing higher priority queries to have referenced buffer
1448 cache pages stay in memory longer
1450 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00562.php
1457 * Improve speed with indexes
1459 For large table adjustments during VACUUM FULL, it is faster to cluster
1460 or reindex rather than update the index. Also, index updates can bloat
1463 http://archives.postgresql.org/pgsql-hackers/2007-03/msg00024.php
1464 http://archives.postgresql.org/pgsql-performance/2007-05/msg00296.php
1465 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00307.php
1467 * Auto-fill the free space map by scanning the buffer cache or by
1468 checking pages written by the background writer
1470 http://archives.postgresql.org/pgsql-hackers/2006-02/msg01125.php
1471 http://archives.postgresql.org/pgsql-hackers/2006-03/msg00011.php
1473 * Create a bitmap of pages that need vacuuming
1475 Instead of sequentially scanning the entire table, have the background
1476 writer or some other process record pages that have expired rows, then
1477 VACUUM can look at just those pages rather than the entire table. In
1478 the event of a system crash, the bitmap would probably be invalidated.
1479 One complexity is that index entries still have to be vacuumed, and
1480 doing this without an index scan (by using the heap values to find the
1481 index entry) might be slow and unreliable, especially for user-defined
1484 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01188.php
1485 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00121.php
1486 http://archives.postgresql.org/pgsql-patches/2007-03/msg00508.php
1487 http://archives.postgresql.org/pgsql-patches/2007-04/msg00347.php
1488 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00156.php
1489 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00546.php
1490 http://archives.postgresql.org/pgsql-hackers/2008-04/msg00416.php
1492 * Allow FSM to return free space toward the beginning of the heap file,
1493 in hopes that empty pages at the end can be truncated by VACUUM
1494 * Allow FSM page return free space based on table clustering, to assist
1495 in maintaining clustering?
1496 * Improve dead row detection during multi-statement transactions usage
1498 http://archives.postgresql.org/pgsql-patches/2007-03/msg00358.php
1500 * Consider a more compact data representation for dead tuples
1502 http://archives.postgresql.org/pgsql-patches/2007-05/msg00143.php
1507 o %Issue log message to suggest VACUUM FULL if a table is nearly
1509 o Improve control of auto-vacuum
1511 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00876.php
1513 o Prevent long-lived temporary tables from causing frozen-xid
1514 advancement starvation
1516 The problem is that autovacuum cannot vacuum them to set frozen xids;
1517 only the session that created them can do that.
1518 http://archives.postgresql.org/pgsql-general/2007-06/msg01645.php
1520 o Store per-table autovacuum settings in pg_class.reloptions.
1522 http://archives.postgresql.org/pgsql-hackers/2007-02/msg01440.php
1523 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00724.php
1525 o Prevent autovacuum from running if an old transaction is still
1526 running from the last vacuum
1528 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00899.php
1535 * Fix priority ordering of read and write light-weight locks (Neil)
1537 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00893.php
1538 http://archives.postgresql.org/pgsql-hackers/2004-11/msg00905.php
1540 * Fix problem when multiple subtransactions of the same outer transaction
1541 hold different types of locks, and one subtransaction aborts
1543 http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php
1544 http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php
1545 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00435.php
1546 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00773.php
1548 * Allow UPDATEs on only non-referential integrity columns not to conflict
1549 with referential integrity locks
1551 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00073.php
1553 * Add idle_in_transaction_timeout GUC so locks are not held for long
1556 * Improve deadlock detection when a page cleaning lock conflicts
1557 with a shared buffer that is pinned
1559 http://archives.postgresql.org/pgsql-bugs/2008-01/msg00138.php
1560 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1561 http://archives.postgresql.org/pgsql-committers/2008-01/msg00365.php
1563 * Detect deadlocks involving LockBufferForCleanup()
1565 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00873.php
1569 Startup Time Improvements
1570 =========================
1572 * Experiment with multi-threaded backend for backend creation
1574 This would prevent the overhead associated with process creation. Most
1575 operating systems have trivial process creation time compared to
1576 database startup overhead, but a few operating systems (Win32,
1577 Solaris) might benefit from threading. Also explore the idea of
1578 a single session using multiple threads to execute a statement faster.
1585 * Eliminate need to write full pages to WAL before page modification
1587 Currently, to protect against partial disk page writes, we write
1588 full page images to WAL before they are modified so we can correct any
1589 partial page writes during recovery. These pages can also be
1590 eliminated from point-in-time archive files.
1591 http://archives.postgresql.org/pgsql-hackers/2002-06/msg00655.php
1593 o When off, write CRC to WAL and check file system blocks
1596 If CRC check fails during recovery, remember the page in case
1597 a later CRC for that page properly matches.
1599 o Write full pages during file system write and not when
1600 the page is modified in the buffer cache
1602 This allows most full page writes to happen in the background
1603 writer. It might cause problems for applying WAL on recovery
1604 into a partially-written page, but later the full page will be
1607 * Allow WAL traffic to be streamed to another server for stand-by
1609 * Reduce WAL traffic so only modified values are written rather than
1612 http://archives.postgresql.org/pgsql-hackers/2007-03/msg01589.php
1614 * Allow WAL information to recover corrupted pg_controldata
1616 http://archives.postgresql.org/pgsql-patches/2006-06/msg00025.php
1618 * Find a way to reduce rotational delay when repeatedly writing
1621 Currently fsync of WAL requires the disk platter to perform a full
1622 rotation to fsync again. One idea is to write the WAL to different
1623 offsets that might reduce the rotational delay.
1624 http://archives.postgresql.org/pgsql-hackers/2002-11/msg00483.php
1626 * Allow WAL logging to be turned off for a table, but the table
1627 might be dropped or truncated during crash recovery
1629 Allow tables to bypass WAL writes and just fsync() dirty pages on
1630 commit. This should be implemented using ALTER TABLE, e.g. ALTER
1631 TABLE PERSISTENCE [ DROP | TRUNCATE | DEFAULT ]. Tables using
1632 non-default logging should not use referential integrity with
1633 default-logging tables. A table without dirty buffers during a
1634 crash could perhaps avoid the drop/truncate.
1635 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1637 * Allow WAL logging to be turned off for a table, but the table would
1638 avoid being truncated/dropped
1640 To do this, only a single writer can modify the table, and writes
1641 must happen only on new pages so the new pages can be removed during
1642 crash recovery. Readers can continue accessing the table. Such
1643 tables probably cannot have indexes. One complexity is the handling
1644 of indexes on TOAST tables.
1645 http://archives.postgresql.org/pgsql-hackers/2005-12/msg01016.php
1647 * Speed WAL recovery by allowing more than one page to be prefetched
1649 This should be done utilizing the same infrastructure used for
1650 prefetching in general to avoid introducing complex error-prone code
1652 http://archives.postgresql.org/pgsql-general/2007-12/msg00683.php
1653 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00497.php
1654 http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php
1656 * Improve WAL concurrency by increasing lock granularity
1658 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00556.php
1660 * Be more aggressive about creating WAL files
1662 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01325.php
1663 http://archives.postgresql.org/pgsql-hackers/2004-07/msg01075.php
1664 http://archives.postgresql.org/pgsql-hackers/2005-04/msg00556.php
1666 * Have resource managers report the duration of their status changes
1668 http://archives.postgresql.org/pgsql-hackers/2007-10/msg01468.php
1670 * Move pgfoundry's xlogdump to /contrib and have it rely more closely
1671 on the WAL backend code
1673 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00035.php
1678 Optimizer / Executor
1679 ====================
1681 * Improve selectivity functions for geometric operators
1682 * Precompile SQL functions to avoid overhead
1683 * Create utility to compute accurate random_page_cost value
1684 * Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
1685 * Have EXPLAIN ANALYZE issue NOTICE messages when the estimated and
1686 actual row counts differ by a specified percentage
1687 * Improve how ANALYZE computes in-doubt tuples
1689 http://archives.postgresql.org/pgsql-hackers/2007-11/msg00771.php
1691 * Consider using hash buckets to do DISTINCT, rather than sorting
1693 This would be beneficial when there are few distinct values. This is
1694 already used by GROUP BY.
1696 * Log statements where the optimizer row estimates were dramatically
1697 different from the number of rows actually found?
1698 * Consider compressed annealing to search for query plans
1700 This might replace GEQO, http://sixdemonbag.org/Djinni.
1702 * Improve merge join performance by allowing mark/restore of
1705 http://archives.postgresql.org/pgsql-hackers/2007-01/msg00096.php
1707 * Consider using a hash for joining to a large IN (VALUES ...) list
1709 http://archives.postgresql.org/pgsql-hackers/2007-05/msg00450.php
1716 * Consider having the background writer update the transaction status
1717 hint bits before writing out the page
1719 Implementing this requires the background writer to have access to system
1720 catalogs and the transaction status log.
1722 * Consider adding buffers the background writer finds reusable to the
1725 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1727 * Automatically tune bgwriter_delay based on activity rather then using a
1730 http://archives.postgresql.org/pgsql-hackers/2007-04/msg00781.php
1732 * Consider wither increasing BM_MAX_USAGE_COUNT improves performance
1734 http://archives.postgresql.org/pgsql-hackers/2007-06/msg01007.php
1736 * Test to see if calling PreallocXlogFiles() from the background writer
1737 will help with WAL segment creation latency
1739 http://archives.postgresql.org/pgsql-patches/2007-06/msg00340.php
1743 Miscellaneous Performance
1744 =========================
1746 * Do async I/O for faster random read-ahead of data
1748 Async I/O allows multiple I/O requests to be sent to the disk with
1749 results coming back asynchronously.
1751 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00820.php
1752 http://archives.postgresql.org/pgsql-performance/2007-09/msg00255.php
1753 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php
1754 http://archives.postgresql.org/pgsql-patches/2008-01/msg00170.php
1756 * Use mmap() rather than SYSV shared memory or to write WAL files?
1758 This would remove the requirement for SYSV SHM but would introduce
1759 portability issues. Anonymous mmap (or mmap to /dev/zero) is required
1760 to prevent I/O overhead.
1762 * Consider mmap()'ing files into a backend?
1764 Doing I/O to large tables would consume a lot of address space or
1765 require frequent mapping/unmapping. Extending the file also causes
1766 mapping problems that might require mapping only individual pages,
1767 leading to thousands of mappings. Another problem is that there is no
1768 way to _prevent_ I/O to disk from the dirty shared buffers so changes
1769 could hit disk before WAL is written.
1771 * Add a script to ask system configuration questions and tune postgresql.conf
1772 * Consider ways of storing rows more compactly on disk
1774 o Reduce the row header size?
1775 o Consider reducing on-disk varlena length from four bytes to
1776 two because a heap row cannot be more than 64k in length
1778 * Consider increasing NUM_CLOG_BUFFERS
1780 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00030.php
1781 http://archives.postgresql.org/pgsql-performance/2007-08/msg00024.php
1783 * Consider transaction start/end performance improvements
1785 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00948.php
1786 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1788 * Allow user configuration of TOAST thresholds
1790 http://archives.postgresql.org/pgsql-hackers/2007-02/msg00213.php
1791 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00082.php
1793 * Allow configuration of backend priorities via the operating system
1795 Though backend priorities make priority inversion during lock
1796 waits possible, research shows that this is not a huge problem.
1798 http://archives.postgresql.org/pgsql-general/2007-02/msg00493.php
1800 * Experiment with multi-threaded backend better I/O utilization
1802 This would allow a single query to make use of multiple I/O channels
1803 simultaneously. One idea is to create a background reader that can
1804 pre-fetch sequential and index scan pages needed by other backends.
1805 This could be expanded to allow concurrent reads from multiple devices
1806 in a partitioned table.
1808 * Experiment with multi-threaded backend better CPU utilization
1810 This would allow several CPUs to be used for a single query, such as
1811 for sorting or query execution.
1813 * Consider increasing the minimum allowed number of shared buffers
1815 http://archives.postgresql.org/pgsql-bugs/2008-02/msg00157.php
1817 * Expire published xmin for read-only and idle transactions
1819 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00343.php
1821 * Consider if CommandCounterIncrement() can avoid its
1822 AcceptInvalidationMessages() call
1824 http://archives.postgresql.org/pgsql-committers/2007-11/msg00585.php
1826 * Improve performance of shared invalidation queue for multiple CPUs
1828 http://archives.postgresql.org/pgsql-performance/2008-01/msg00023.php
1830 * Consider Cartesian joins when both relations are needed to form an
1831 indexscan qualification for a third relation
1833 http://archives.postgresql.org/pgsql-performance/2007-12/msg00090.php
1835 * Consider not storing a NULL bitmap on disk if all the NULLs are
1838 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00624.php
1839 http://archives.postgresql.org/pgsql-patches/2007-12/msg00109.php
1841 * Sort large UPDATE/DELETEs so it is done in heap order
1843 http://archives.postgresql.org/pgsql-hackers/2008-01/msg01119.php
1845 * -Avoid tuple some tuple copying in sort routines
1846 * SMP scalability improvements
1848 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00439.php
1849 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00206.php
1850 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00361.php
1852 * Research reducing deTOASTing in more places
1854 http://archives.postgresql.org/pgsql-hackers/2007-09/msg00895.php
1856 * Consider being smarter about memory and external files used during
1859 http://archives.postgresql.org/pgsql-hackers/2007-11/msg01101.php
1860 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00045.php
1862 * Allow one transaction to see tuples using the snapshot of another
1865 This would assist multiple backends in working together.
1866 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00400.php
1872 * Add use of 'const' for variables in source tree
1873 * Move some things from contrib into main tree
1874 * %Remove warnings created by -Wcast-align
1875 * Move platform-specific ps status display info from ps_status.c to ports
1876 * Add optional CRC checksum to heap and index pages
1877 * Improve documentation to build only interfaces (Marc)
1878 * Remove or relicense modules that are not under the BSD license, if possible
1879 * Acquire lock on a relation before building a relcache entry for it
1880 * Allow cross-compiling by generating the zic database on the target system
1881 * Improve NLS maintenance of libpgport messages linked onto applications
1882 * Clean up casting in contrib/isn
1884 http://archives.postgresql.org/pgsql-hackers/2006-11/msg00245.php
1886 * Use UTF8 encoding for NLS messages so all server encodings can
1888 * Update Bonjour to work with newer cross-platform SDK
1890 http://archives.postgresql.org/pgsql-hackers/2006-09/msg02238.php
1891 http://archives.postgresql.org/pgsql-patches/2006-10/msg00048.php
1893 * Consider detoasting keys before sorting
1894 * Consider GnuTLS if OpenSSL license becomes a problem
1896 http://archives.postgresql.org/pgsql-patches/2006-05/msg00040.php
1897 http://archives.postgresql.org/pgsql-hackers/2006-12/msg01213.php
1899 * Consider changing documentation format from SGML to XML
1901 http://archives.postgresql.org/pgsql-docs/2006-12/msg00152.php
1903 * Consider making NAMEDATALEN more configurable in future releases
1904 * Update our code to handle 64-bit timezone files to match the zic
1905 source code, which now uses them
1906 * Have configure choose integer datetimes by default
1908 http://archives.postgresql.org/pgsql-patches/2007-05/msg00046.php
1910 * Support scoped IPv6 addresses
1912 http://archives.postgresql.org/pgsql-bugs/2007-05/msg00111.php
1914 * Consider allowing 64-bit integers and floats to be passed by value on
1917 Also change 32-bit floats (float4) to be passed by value at the same
1920 * Research use of signals and sleep wake ups
1922 http://archives.postgresql.org/pgsql-hackers/2007-07/msg00003.php
1924 * Add automated check for invalid C++ source code constructs
1926 http://archives.postgresql.org/pgsql-patches/2007-07/msg00056.php
1928 * Consider simplifying how memory context resets handle child contexts
1930 http://archives.postgresql.org/pgsql-patches/2007-08/msg00067.php
1932 * Remove use of MAKE_PTR and MAKE_OFFSET macros
1934 http://archives.postgresql.org/pgsql-general/2007-08/msg01510.php
1936 * Convert single quotes to apostrophes in the PDF documentation
1938 http://archives.postgresql.org/pgsql-docs/2007-12/msg00059.php
1940 * Create three versions of libpgport to simplify client code
1942 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00154.php
1944 * Remove old-style routines for manipulating tuples
1946 http://archives.postgresql.org/pgsql-hackers/2007-10/msg00851.php
1948 * Improve detection of shared memory segments being used by other
1951 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00656.php
1953 * Implement the non-threaded Avahi service discovery protocol
1955 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00939.php
1956 http://archives.postgresql.org/pgsql-patches/2008-02/msg00097.php
1957 http://archives.postgresql.org/pgsql-hackers/2008-03/msg01211.php
1958 http://archives.postgresql.org/pgsql-patches/2008-04/msg00001.php
1962 o Remove configure.in check for link failure when cause is found
1963 o Remove readdir() errno patch when runtime/mingwex/dirent.c rev
1965 o Remove psql newline patch when we find out why mingw outputs an
1967 o Allow psql to use readline once non-US code pages work with
1969 o Fix problem with shared memory on the Win32 Terminal Server
1970 o Diagnose problem where shared memory can sometimes not be
1971 attached by postmaster children
1973 http://archives.postgresql.org/pgsql-general/2007-08/msg01377.php
1975 o Improve signal handling
1977 http://archives.postgresql.org/pgsql-patches/2005-06/msg00027.php
1979 o Convert MSVC build system to remove most batch files
1981 http://archives.postgresql.org/pgsql-hackers/2007-08/msg00961.php
1983 o Prevent SSL from sending network packets to avoid interference
1984 with Win32 signal emulation
1986 http://archives.postgresql.org/pgsql-hackers/2007-12/msg00455.php
1988 o Support pgxs when using MSVC
1990 o Fix MSVC NLS support, like for to_char()
1992 http://archives.postgresql.org/pgsql-hackers/2008-02/msg00485.php
1993 http://archives.postgresql.org/pgsql-patches/2008-02/msg00038.php
1995 o Find a correct rint() substitute on Windows
1997 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00808.php
2001 * Wire Protocol Changes
2003 o Allow dynamic character set handling
2004 o Add decoded type, length, precision
2006 o Update clients to use data types, typmod, schema.table.column names
2007 of result sets using new statement protocol
2013 * Add pre-parsing phase that converts non-ISO syntax to supported
2016 This could allow SQL written for other databases to run without
2019 * Allow plug-in modules to emulate features from other databases
2020 * Add features of Oracle-style packages (Pavel)
2022 A package would be a schema with session-local variables,
2023 public/private functions, and initialization functions. It
2024 is also possible to implement these capabilities
2025 in any schema and not use a separate "packages"
2028 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00384.php
2030 * Consider allowing control of upper/lower case folding of unquoted
2033 http://archives.postgresql.org/pgsql-hackers/2004-04/msg00818.php
2034 http://archives.postgresql.org/pgsql-hackers/2006-10/msg01527.php
2035 http://archives.postgresql.org/pgsql-hackers/2008-03/msg00849.php
2037 * Add autonomous transactions
2039 http://archives.postgresql.org/pgsql-hackers/2008-01/msg00893.php
2043 Features We Do _Not_ Want
2044 =========================
2046 * All backends running as threads in a single process (not wanted)
2048 This eliminates the process protection we get from the current setup.
2049 Thread creation is usually the same overhead as process creation on
2050 modern systems, so it seems unwise to use a pure threaded model.
2052 * Optimizer hints (not wanted)
2054 Optimizer hints are used to work around problems in the optimizer. We
2055 would rather have the problems reported and fixed.
2057 http://archives.postgresql.org/pgsql-hackers/2006-08/msg00506.php
2058 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00517.php
2059 http://archives.postgresql.org/pgsql-hackers/2006-10/msg00663.php
2062 Because we support postfix operators, it isn't possible to make AS
2063 optional and continue to use bison.
2064 http://archives.postgresql.org/pgsql-hackers/2003-04/msg00436.php
2066 http://archives.postgresql.org/pgsql-sql/2006-08/msg00164.php
2068 * Embedded server (not wanted)
2070 While PostgreSQL clients runs fine in limited-resource environments, the
2071 server requires multiple processes and a stable pool of resources to
2072 run reliabily and efficiently. Stripping down the PostgreSQL server
2073 to run in the same process address space as the client application
2074 would add too much complexity and failure cases.