Skip to content

Wöchentlicher PostgreSQL Newsletter - 29. März 2009


Der Originalartikel befindet sich unter:
http://www.postgresql.org/community/weeklynews/pwn20090329


== Wöchentlicher PostgreSQL Newsletter - 29. März 2009 ==

Die letzten Einträge im Commitfest für 8.4 werden derzeit geschlossen.
Achte auf eine wichtige Ankündigung später diese Woche.

Das git Repository nutzt nun Community Logins und ist jetzt viel
einfacher zu nutzen und zu verwalten.
http://git.postgresql.org/

Neue Umfrage: Was ist dein bevorzugtes Replikationstool für PostgreSQL?
http://www.postgresql.org/community/

Devrim GUNDUZ hat ein weiteres Set mit Snapshot RPMs veröffentlicht.
Gehe hin und teste :)
http://yum.pgsqlrpms.org/news-8.4devel-ready-for-testing.php

== PostgreSQL Produkt Neuigkeiten ==

pgAdmin 1.10 Beta 2, ein GUI Management Tool für PostgreSQL,
ist erschienen.
http://www.postgresql.org/ftp/pgadmin3/release/v1.10.0-beta2/

DBD::Pg 2.12.0, ein Perl Connector für PostgreSQL, ist erschienen.
http://search.cpan.org/dist/DBD-Pg/

== PostgreSQL Tipp der Woche ==

Du kannst die pg_service Datei nutzen, um verschiedene Profile
zu erstellen für die verschiedenen Server, die du nutzt.

== Postgresql 8.4 Feature der Woche ==

SSL aufgeräumt: Magnus Hagander hat den Code für SSL Verbindungen
verbessert, um selbigen mehr intuitiv und zuverlässig zu gestalten.
Jetzt kannst du auch angeben, welche Zertifikate ein Client vorweisen
muss.

== PostgreSQL Jobs im März ==

http://archives.postgresql.org/pgsql-jobs/2009-03/threads.php

== PostgreSQL Lokal ==

PostgreSQLFr wird einen Stand auf der Solutions Linux 2009 in Paris
haben. Anmelden unter sas AT postgresql DOT fr oder im
französischen Wiki.
http://postgresql.fr/sl2009:start

Die Registrierung für die PostgreSQL Konferenz (JDCon) East 2009
ist jetzt möglich:
http://www.postgresql.us/purchase

Kevin Kempter wird über Backup und Wiederherstellen auf dem ersten
Denver PUG Meeting am 8. April 2009 sprechen.
http://www.diapug.org/Meeting.html

PgDay Brasilien wird am 17. April stattfinden. Kontaktiere fernando.ike
AT gmail DOT com oder eduardo DOT santos AT planejamento DOT gov DOT
br, wenn du teilnehmen möchtest.

Die Percona Performance Konferenz wird im Santa Clara Convention Center
in Santa Clara, Kalifornien, USA, stattfinden.
http://conferences.percona.com/

Der CfP für den PgDay in Sao Paulo am 24. April ist eröffnet.
Kontaktiere marins DOT consultoria AT gmail DOT com oder marcelojscosta
AT gmail DOT com wenn du teilnehmen möchtest.

PostgreSQL Conference U.S. (JDCon) wird einen PgDay auf dem Linuxfest
Nordwest am 25. und 26. April veranstalten. Der Call for Papers ist
unter:
http://www.postgresqlconference.org/

Es werden weiterhin PgDays am 29. April in Porto Velho, RO und am 20.
April in Ji-Parana, RO stattfinden. Kontaktiere Luis Fernando Bueno:
proflfbueno AT gmail DOT com wenn du teilnehmen möchtest.

Michael Renner wird einen Workshop über PostgreSQL Replikation auf
der Netways OSDC 2009 in Nürnbern, Deutschland geben.
http://www.netways.de/english/osdc/y2009/programm/w/michael_renner_postgresql_repliziert_ein_ueberblick/

PGCon 2009 wird vom 21. bis 22. Mai 2009 in Ottawa an der Universität
von Ottawa stattfinden. Dem gehen zwei Tage mit Tutorials vom 19. bis
20. Mai 2009 voraus.
http://www.pgcon.org/2009/

Termin vormerken: pgDay San Jose, am Sonntag dem 19. Juli unmittelbar
vor der OSCON. CfP und mehr Informationen bald.

PGCon Brasilien wird am 23.-24. Oktober 2009 auf dem Unicamp in
Campinas, Sao Paulo, stattfinden.

== PostgreSQL in den News ==

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

Dieser wöchentliche PostgreSQL Newsletter wurde erstellt von David
Fetter und Josh Berkus.

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/pl/tcl/expected/pltcl_queries_1.out, add expected file
  for locale-enabled installations.

- In pgsql/src/bin/psql/help.c, fix my typo.

- In pgsql/src/bin/psql/command.c, no period if it's not a sentence.

- Polish SQL/MED terminology

- In pgsql/src/bin/psql/help.c, in the \? output, align the columns
  consistently.

- In pgsql/src/bin/psql/help.c, update line count for \? output
  (obtained from wc -l).

- In pgsql/src/backend/catalog/catalog.c, complete list of valid fork
  names, and use double quotes.

- In pgsql/src/backend/libpq/hba.c, use double quotes instead of
  single quotes.

- Gettext plural support.  In the backend, I changed only a handful of
  exemplary or important-looking instances to make use of the plural
  support; there is probably more work there.  For the rest of the
  source, this should cover all relevant cases.

Bruce Momjian committed:

- In pgsql/doc/src/sgml/release.sgml, reorder 8.4 release note
  sections.

- In pgsql/doc/src/sgml/release.sgml, correction:  ansi-join
  ->anti-join.

- In pgsql/doc/src/sgml/release.sgml, document that Datestyle no
  longer controls interval output, per suggestion from Ron Mayer

- In pgsql/doc/src/sgml/release.sgml, remove duplicate interval
  (fractional seconds) items;  Ron Mayer.

- In pgsql/doc/src/sgml/release.sgml, improved release notes interval
  wording.

- In pgsql/doc/src/sgml/release.sgml, Fix release note attribution:
  "Support the IS0 8601 time interval syntax (Tom, Kevin Grittner)"
  per Ron Mayer.

- In pgsql/doc/src/sgml/release.sgml, first round of release note
  changes from Guillaume Smet.

- In pgsql/doc/src/sgml/release.sgml, mark Greg Sabino Mullane as the
  instigator of the statistics target increase.

- In pgsql/doc/src/sgml/release.sgml, second batch of release note
  fixes by Guillaume Smet.

- In pgsql/doc/src/sgml/release.sgml, updated release wording, per
  Greg Stark.

- In pgsql/doc/src/sgml/release.sgml, document in release notes that
  NOT IN is only for NOT EXIST clauses.  Andrew (RhodiumToad) Gierth.

- In pgsql/doc/src/sgml/monitoring.sgml, add documentation of the fact
  that dtrace probes evaluate their parameters even when not active.
  Explain how to prevent that with an ENABLED() check.

- Better document PQinitSSL(0) behavior in regards to libcrypto.

- In pgsql/doc/src/sgml/ref/alter_role.sgml, better document that SET
  ROLE does not uset ALTER ROLE SET settings; suggested wording from
  Josh Berkus.

- In pgsql/doc/src/sgml/release.sgml, update release notes to say
  citext is multi-byte aware, per suggestion from patch author.

- In pgsql/src/interfaces/libpq/fe-secure.c, clarify variable naming:
  pq_initssllib -> pq_init_ssl_lib.

- In pgsql/doc/src/sgml/release.sgml, consistent 8.4 release note
  naming for Itagaki Takahiro.

- In pgsql/doc/src/sgml/release.sgml, add SGML markup for
  commands/literal/application/etc in release notes; still more work
  to do.

- In pgsql/doc/src/sgml/release.sgml, more release note markup.

- In pgsql/doc/src/sgml/release.sgml, more release note markup.


- In pgsql/doc/src/sgml/release.sgml, re-add release notes for release
  8.3.7.

- In pgsql/doc/src/sgml/config.sgml, update constraint_exclusion docs
  because parameter is now 3-valued, not just enable/disable.

- In pgsql/doc/src/sgml/release.sgml, first version of 8.4 release
  notes;  markup/cleanup/reorganization still required.

- In pgsql/doc/src/sgml/release.sgml, mention release note items
  current as of 2009-03-16.

- In pgsql/doc/src/sgml/release.sgml, adjust AS OF release notes
  markup.

Tom Lane committed:

- Const-ify the parse table passed to fillRelOptions.  The previous
  coding meant it had to be built on-the-fly at each entry to
  default_reloptions.

- Implement "fastupdate" support for GIN indexes, in which we try to
  accumulate multiple index entries in a holding area before adding
  them to the main index structure.  This helps because bulk insert is
  (usually) significantly faster than retail insert for GIN.  This
  patch also removes GIN support for amgettuple-style index scans.
  The API defined for amgettuple is difficult to support with
  fastupdate, and the previously committed partial-match feature
  didn't really work with it either.  We might eventually figure a way
  to put back amgettuple support, but it won't happen for 8.4.
  catversion bumped because of change in GIN's pg_am entry, and
  because the format of GIN indexes changed on-disk (there's a
  metapage now, and possibly a pending list).  Teodor Sigaev.

- In pgsql/src/backend/optimizer/plan/planner.c, fix stupid
  parenthesization mistake.  Per bug #4728 from Bruce Toll.

- Install a search tree depth limit in GIN bulk-insert operations, to
  prevent them from degrading badly when the input is sorted or nearly
  so.  In this scenario the tree is unbalanced to the point of
  becoming a mere linked list, so insertions become O(N^2).  The
  easiest and most safely back-patchable solution is to stop growing
  the tree sooner, ie limit the growth of N.  We might later consider
  a rebalancing tree algorithm, but it's not clear that the benefit
  would be worth the cost and complexity.  Per report from Sergey
  Burladyan and an earlier complaint from Heikki Linnakangas.
  Back-patch to 8.2; older versions didn't have GIN indexes.

- In pgsql/contrib/pgcrypto/pgp.h, fix old thinko in pgp.h: the idea
  is to declare some named enum types, not global variables of
  anonymous enum types.  This didn't actually hurt much because most
  linkers will just merge the duplicated definitions ...  but some
  will complain.  Per bug #4731 from Ceriel Jacobs.  Backpatch to 8.1
  --- the declarations don't exist before that.

- Adjust the APIs for GIN opclass support functions to allow the
  extractQuery() method to pass extra data to the consistent() and
  comparePartial() methods.  This is the core infrastructure needed to
  support the soon-to-appear contrib/btree_gin module.  The APIs are
  still upward compatible with the definitions used in 8.3 and before,
  although not with the previous 8.4devel function definitions.
  catversion bump for changes in pg_proc entries (although these are
  just cosmetic, since GIN doesn't actually look at the function
  signature before calling it...).  Teodor Sigaev and Oleg Bartunov

- Add btree_gin contrib module.  Teodor Sigaev and Oleg Bartunov.

- In pgsql/doc/src/sgml/release.sgml, fix markup so that 'make
  HISTORY' works.  A couple very minor editorial improvements.

- If we expect a hash join to be performed in multiple batches,
  suppress "physical tlist" optimization on the outer relation (ie,
  force a projection step to occur in its scan).  This avoids storing
  useless column values when the outer relation's tuples are written
  to temporary batch files.  Modified version of a patch by Michael
  Henderson and Ramon Lawrence.

- In pgsql/src/backend/commands/copy.c, adjust phrasing of complaints
  about multi-byte COPY delimiter strings.  Per pgsql-hackers
  discussion of 2009-02-17.

- Fix a couple of places that still claimed LC_COLLATE and LC_CTYPE
  can't be changed after initdb.

- In pgsql/contrib/pg_standby/pg_standby.c, Make pg_standby's
  maxretries option do what one would expect.  Fujii Masao.

- In pgsql/src/backend/catalog/index.c, teach reindex_index() to clear
  pg_index.indcheckxmin when possible.  Greg Stark, slightly modified
  by me.

- Fix possible failures when a tuplestore switches from in-memory to
  on-disk mode while callers hold pointers to in-memory tuples.  I
  reported this for the case of nodeWindowAgg's primary scan tuple,
  but inspection of the code shows that all of the calls in
  nodeWindowAgg and nodeCtescan are at risk.  For the moment, fix it
  with a rather brute-force approach of copying whenever one of the
  at-risk callers requests a tuple.  Later we might think of some sort
  of reference-count approach to reduce tuple copying.

- In pgsql/src/backend/utils/adt/xml.c, add an errdetail explaining
  why we reject infinite dates and timestamps while converting to XML.
  Bernd Helmle

- On Solaris, we should only force use of our own getopt(); it's okay
  to use the system's getopt_long().  The previous coding was the
  result of a sloppy discussion that failed to draw this distinction.
  The result was that PG programs don't handle options as users of
  that platform expect.  Per gripe from Chuck McDevitt.  Although this
  is a pre-existing bug, I'm not backpatching since I think we could
  do with a bit of beta testing before concluding this is really OK.

Andrew Dunstan committed:

- Remove munging of xml and xpath params to xpath(). The XML must now
  be a well formed XML document.

Magnus Hagander committed:

- In pgsql/doc/src/sgml/release.sgml, fix release notes about pg_hba
  changes.

- In pgsql/doc/src/sgml/release.sgml, clarify new SSL certificate
  verification in libpq.

- In pgsql/doc/src/sgml/release.sgml, fix markup, per Devrim GUNDUZ.

Heikki Linnakangas committed:

- In pgsql/src/bin/psql/tab-complete.c, fix tab completion of ANALYZE
  VERBOSE <tab>.  It was previously confused with EXPLAIN ANALYZE
  VERBOSE.  Greg Sabino Mullane, reformatted by myself.  Backpatch to
  8.1, where the bug was introduced.

== Abgelehnte Patches (bis jetzt) ==

Srinath K's patch to implement a global index on partitioned tables,
as it came with an IP disclaimer which precluded anyone's looking at
it.

Hitoshi Harada's patches to fix a tuplestore API issue where a
ScanTupleSlot is not always valid as long as the tuplestore-handling
code expects it to be.  Tom Lane used forced copying instead.

== Eingesandte Patches ==

Kedar Potdar sent in a patch to do automatic partioning.

KaiGai Kohei sent in another revision of his SE-PostgreSQL patches.

Zdenek Kotala sent in another patch to fix the regression tests on the
Czech (cs_CZ) locale.

Fujii Masao sent in a patch which adds a new trigger option to
pg_standby to fix a pg_standby failure mode.

Andrew (RhodiumToad) Gierth sent in a patch to change hstore so that
keys and values can be of arbitrary length, so long as the hstore
itself fits in varlena, supports btree and hash opclasses for hstores,
which in turn allow for UNIQUE constraints, and aggregation, adds
binary I/O, a flock of operators which relate hstores to text, text[]
and record, adds GiST and GIN support, converts between hstore and
record, and populates a record from an hstore.

Andrew (RhodiumToad) Gierth sent in a patch which causes GET
DIAGNOSTICS ROW_COUNT to work for MOVE and FETCH in PL/pgSQL.

Pavel Stehule sent in an example contrib module to implement a JSON
type using the WIP transformation hook modules.

Bruce Momjian sent in a patch which makes a more informative error
message when someone changes a GUC parameter which can only take one
value.

Taro Minowa sent in a patch for 8.5 which allows pg_dump to dump
comments on columns which are composite types.

  • Twitter
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at del.icio.us
  • Facebook
  • Google Bookmarks
  • FriendFeed
  • Digg Wöchentlicher PostgreSQL Newsletter - 29. März 2009
  • Bloglines Wöchentlicher PostgreSQL Newsletter - 29. März 2009
  • Technorati Wöchentlicher PostgreSQL Newsletter - 29. März 2009
  • Fark this: Wöchentlicher PostgreSQL Newsletter - 29. März 2009
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at YahooMyWeb
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at Furl.net
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at reddit.com
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at blinklist.com
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at Spurl.net
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at Simpy.com
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 at blogmarks
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 29. März 2009 with wists
  • wong it!
  • Bookmark using any bookmark manager!
  • Stumble It!
  • Identi.ca

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