Skip to content

Wöchentlicher PostgreSQL Newsletter - 25. Oktober 2009


Der Originalartikel befindet sich unter:

http://www.postgresql.org/community/weeklynews/pwn20091025



== Wöchentlicher PostgreSQL Newsletter - 25. Oktober 2009 ==

== PostgreSQL Produkt Neuigkeiten ==

once:fabrik, ein ERP System, das auf PostgreSQL basiert, ist zum Testen
bis zum 31. Oktober 2009 kostenlos verfügbar.
http://www.oncetechnologies.com/

== PostgreSQL Jobs im Oktober ==

http://archives.postgresql.org/pgsql-jobs/2009-10/threads.php

== PostgreSQL Lokal ==

LISA 09, die Large Installation and Systems Administration Konferenz,
findet vom 1. bis 6. November 2009 statt. Es wird einen PostgreSQL
Stand am Mittwoch und Donnerstag geben.
http://www.usenix.org/event/lisa09/

PGDay.EU 2009 wird an der Telecom ParisTech Universität in Paris,
Frankreich, am 6. und 7. November 2009 stattfinden.
http://www.pgday.eu/

OpenSQL Camp in Portland sucht Sponsoren. Bereite deine Reisepläne
jetzt vor! :)
http://www.chesnok.com/daily/2009/07/29/opensql-camp-comes-to-portland-november-14-15-2009/

Die 10. jährliche JPUG Konferenz findet am 20-21. November 2009 in
Tokio, Japan, statt.
http://archives.postgresql.org/pgsql-announce/2009-05/msg00018.php

FOSDEM 2010 findet am 6. und 7. Februar 2010 in Brüssel,
Belgien statt.
http://www.fosdem.org/

Die Chemnitzer Linuxtage finden am 13. und 14. März in
Chemnitz, Deutschland statt.
http://chemnitzer.linux-tage.de/

== PostgreSQL in den News ==

Planet PostgreSQL: http://planet.postgresql.org/

Dieser wöchentliche PostgreSQL Newsletter wurde erstellt von David
Fetter.

Sende Neuigkeiten und Ankündigungen bis Sonntag, 15 Uhr Pazifischer
Zeit. Bitte sende englische Beiträge an david@fetter.org, deutsche an
pwn@pgug.de, italienische an pwn@itpug.org.



== Angewandte Patches ==

Peter Eisentraut committed:

- In pgsql/src/tools/RELEASE_CHANGES, update translation updating
  procedure.  This pertains to some changes I made to automatically
  exclude translations below the 80% minimum.

- Translations update for 8.5alpha2

- In pgsql/doc/src/sgml/release-8.5.sgml, preliminary release notes
  for 8.5alpha2.

- In pgsql/doc/src/sgml/release-8.5.sgml, finalize 8.5alpha2 release
  notes, with updates from Josh Berkus.

- Version stamp 8.5alpha2.

- When querying a table with child tables, do not check permissions on
  the child tables.  This was found to be useless and confusing in
  virtually all cases, and also contrary to the SQL standard.

Tom Lane committed:

- Remove add_missing_from GUC and associated parser support for
  "implicit RTEs".  Per recent discussion, add_missing_from has been
  deprecated for long enough to consider removing, and it's getting in
  the way of planned parser refactoring.  The system now always
  behaves as though add_missing_from were OFF.

- Remove regex_flavor GUC, so that regular expressions are always
  "advanced" style by default.  Per discussion, there seems to be
  hardly anything that really relies on being able to change the regex
  flavor, so the ability to select it via embedded options ought to be
  enough for any stragglers.  Also, if we didn't remove the GUC, we'd
  really be morally obligated to mark the regex functions
  non-immutable, which'd possibly create performance issues.

- Re-implement EvalPlanQual processing to improve its performance and
  eliminate a lot of strange behaviors that occurred in join cases.
  We now identify the "current" row for every joined relation in
  UPDATE, DELETE, and SELECT FOR UPDATE/SHARE queries.  If an
  EvalPlanQual recheck is necessary, we jam the appropriate row into
  each scan node in the rechecking plan, forcing it to emit only that
  one row.  The former behavior could rescan the whole of each joined
  relation for each recheck, which was terrible for performance, and
  what's much worse could result in duplicated output tuples.  Also,
  the original implementation of EvalPlanQual could not re-use the
  recheck execution tree --- it had to go through a full executor init
  and shutdown for every row to be tested.  To avoid this overhead,
  I've associated a special runtime Param with each LockRows or
  ModifyTable plan node, and arranged to make every scan node below
  such a node depend on that Param.  Thus, by signaling a change in
  that Param, the EPQ machinery can just rescan the already-built test
  plan.  This patch also adds a prohibition on set-returning functions
  in the targetlist of SELECT FOR UPDATE/SHARE.  This is needed to
  avoid the duplicate-output-tuple problem.  It seems fairly
  reasonable since the other restrictions on SELECT FOR UPDATE are
  meant to ensure that there is a unique correspondence between source
  tuples and result tuples, which an output SRF destroys as much as
  anything else does.

== Abgelehnte Patches (bis jetzt) ==

No one was disappointed this week :-)

== Eingesandte Patches ==

Dave Page sent in two more revisions of the patch to make the
connecting application name available for display and modification via
a GUC.

Dean Rasheed sent in a WIP patch to scale up deferred UNIQUE
constraints.

ITAGAKI Takahiro sent in a patch to implement a process utility hook.

ITAGAKI Takahiro sent in a patch to allow psql to and use Byte Order
Marks.

ITAGAKI Takahiro sent in another revision of a patch which adds the
option to add an SQL:2008-compliant WHEN clause to trigger
definitions.

Marko (johto) Tiikkaja sent in a WIP patch implementing writeable
CTEs.

ITAGAKI Takahiro sent in a WIP patch to make VACUUM FULL unneeded

Magnus Hagander sent in a patch to parse all configuration files in a
directory.  This is, among other things, infrastructure for tools
which could manipulate such files.

Roger Leigh sent in three more revisions of the Unicode pretty-print
patch for psql.

Jeff Davis sent in another revision of the operator exclusion
constraints patch.

Trackbacks

No Trackbacks

Comments

Display comments as Linear | Threaded

No comments

Add Comment

Enclosing asterisks marks text as bold (*word*), underscore are made via _word_.
E-Mail addresses will not be displayed and will only be used for E-Mail notifications.
To leave a comment you must approve it via e-mail, which will be sent to your address after submission.
Form options