Skip to content

Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009

Der Originalartikel befindet sich unter:

== Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 ==

Der Call for Papers für PGCon Brasilien ist eröffnet!

ITPUG wird auf der Italienischen Konferenz für Freie Software, die in
Bologna am 12. und 13. Juni statfindet, vertreten sein. Gabriele
Bartolini wird einen Vortrag "Studenten und Open Source: der PostgreSQL
Fall" am Samstag dem 13. Juni um 10:35 halten.

== PostgreSQL Produkt Neuigkeiten ==

pgtheme, ein speziell für Drupal 6.x entwickeltes Theme System, ist

== PostgreSQL 8.4 Feature der Woche ==

quote_nullable: verhält sich wie quote_literal, aber liefert den String
NULL für ein Null Argument zurück.

== PostgreSQL Tipp der Woche==

Das "ltree" Contribmodul ist eine Implementierung von Indicies, die
Bäume enthalten und eignet sich hervorragend dafür, ein Dateisystem
oder ähnliche Hierarchien in einer Datenbank abzubilden.

== PostgreSQL Lokal ==

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.

PGDay.EU 2009 wird an der Telecom ParisTech Universität in Paris,
Frankreich, am 6. und 7. Nivember 2009 stattfinden.

Die 10. jährliche JPUG Konferenz hat den Aufruf für Vorträge gestartet.
Die Konferenz findet am 20-21. November 2009 in Tokio, Japan, statt.

== PostgreSQL in den News ==

Planet PostgreSQL:

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, deutsche an, italienische an

== Angewandte Patches ==

Tom Lane committed:

- Fix DecodeInterval to report an error for multiple occurrences of
  DAY, WEEK, YEAR, DECADE, CENTURY, or MILLENIUM fields, just as it
  always has done for other types of fields.  The previous behavior
  seems to have been a hack to avoid defining bit-positions for all
  these field types in DTK_M() masks, rather than something that was
  really considered to be desired behavior.  But there is room in the
  masks for these, and we really need to tighten up at least the
  behavior of DAY and YEAR fields to avoid unexpected behavior
  associated with the 8.4 changes to interpret ambiguous fields based
  on the interval qualifier (typmod) value.  Per my example and
  proposed patch.

- Change AdjustIntervalForTypmod to not discard higher-order field
  values on the grounds that they don't fit into the specified
  interval qualifier (typmod).  This behavior, while of long standing,
  is clearly wrong per spec --- for example the value INTERVAL '999'
  SECOND means 999 seconds and should not be reduced to less than 60
  seconds.  In some cases there could be grounds to raise an error if
  higher-order field values are not given as zero; for example '1 year
  1 month'::INTERVAL MONTH should arguably be taken as an error rather
  than equivalent to 13 months.  However our internal representation
  doesn't allow us to do that in a fashion that would consistently
  reject all and only the cases that a strict reading of the spec
  would suggest.  Also, seeing that for example INTERVAL '13' MONTH
  will print out as '1 year 1 mon', we have to be careful not to
  create a situation where valid data will fail to dump and reload.
  The present patch therefore takes the attitude of not throwing an
  error in any such case.  We might want to revisit that in future but
  it would take more redesign than seems prudent in late beta.  Per a
  complaint from Sebastien Flaesch and subsequent discussion.  While
  at other times we might have just postponed such an issue to the
  next development cycle, 8.4 already has changed the parsing of
  interval literals quite a bit in an effort to accept all
  spec-compliant cases correctly.  This seems like a change that
  should be part of that rather than coming along later.

- In pgsql/doc/src/sgml/config.sgml, Remove the old advice to keep
  from_collapse_limit less than geqo_threshold, instead just pointing
  out that a larger value may trigger use of GEQO.  Per Robert Haas.
  In passing, do a bit of wordsmithing on the Genetic Query Optimizer

- In pgsql/src/backend/commands/copy.c, improve comment about 'if (1)'
  hack in copy.c macros.

- In pgsql/src/bin/initdb/initdb.c, change rather bizarre code
  ordering in get_id().  This isn't strictly cosmetic --- I'm
  wondering if geteuid could have side effects on errno, thus possibly
  resulting in a misleading error message after failure of getpwuid.

- In pgsql/src/backend/tsearch/ts_selfuncs.c, fix tsquerysel() to not
  fail on an empty TSQuery.  Per report from Tatsuo Ishii.

- Clean up ecpg's use of mmerror(): const-ify the format argument, add
  an __attribute__() marker so that gcc can validate the format string
  against the actual arguments, get rid of overcomplicated and unsafe
  usage in base_yyerror().

- Improve the recently-added support for properly pluralized error
  messages by extending the ereport() API to cater for pluralization
  directly.  This is better than the original method of calling
  ngettext outside the elog.c code because (1) it avoids double
  translation, which wastes cycles and in the worst case could give a
  wrong result; and (2) it avoids having to use a different coding
  method in PL code than in the core backend.  The client-side uses of
  ngettext are not touched since neither of these concerns is very
  pressing in the client environment.  Per my proposal of yesterday.

- Remove a couple of debugging messages that have been #ifdef'd out
  for ages.  Seems silly to ask translators to expend work on these,
  especially in pluralized variants.

- Trivial code style cleanup around a couple of ngettext calls.

- GIN's ItemPointerIsMin, ItemPointerIsMax, and ItemPointerIsLossyPage
  macros should use GinItemPointerGetBlockNumber/
  GinItemPointerGetOffsetNumber, not ItemPointerGetBlockNumber/
  ItemPointerGetOffsetNumber, because the latter will Assert() on
  ip_posid == 0, ie a "Min" pointer.  (Thus, ItemPointerIsMin has
  never worked at all, but it seems unused at present.)  I'm not
  certain that the case can occur in normal functioning, but it's
  blowing up on me while investigating Tatsuo-san's data corruption
  problem.  In any case it seems like a problem waiting to bite
  someone.  Back-patch just in case this really is a problem for
  somebody in the field.

- Fix a serious bug introduced into GIN in 8.4: now that
  MergeItemPointers() is supposed to remove duplicate heap TIDs, we
  have to be sure to reduce the tuple size and posting-item count
  accordingly in addItemPointersToTuple().  Failing to do so resulted
  in the effective injection of garbage TIDs into the index contents,
  ie, whatever happened to be in the memory palloc'd for the new
  tuple.  I'm not sure that this fully explains the index corruption
  reported by Tatsuo Ishii, but the test case I'm using no longer

- In pgsql/src/pl/plperl/plperl.c, move variable declaration to avoid
  'unused variable' warning when the ifdef doesn't trigger.  Not worth
  back-patching.  Per buildfarm reports.

- Improve the IndexVacuumInfo/IndexBulkDeleteResult API to allow
  somewhat sane behavior in cases where we don't know the heap tuple
  count accurately; in particular partial vacuum, but this also makes
  the API a bit more useful for ANALYZE.  This patch adds
  "estimated_count" flags to both structs so that an approximate count
  can be flagged as such, and adjusts the logic so that approximate
  counts are not used for updating pg_class.reltuples.  This fixes my
  previous complaint that VACUUM was putting ridiculous values into
  pg_class.reltuples for indexes.  The actual impact of that bug is
  limited, because the planner only pays attention to reltuples for an
  index if the index is partial; which probably explains why beta
  testers hadn't noticed a degradation in plan quality from it.  But
  it needs to be fixed.  The whole thing is a bit messy and should be
  redesigned in future, because reltuples now has the potential to
  drift quite far away from reality when a long period elapses with no
  non-partial vacuums.  But this is as good as it's going to get for

- Revert my patch of 2009-04-04 that removed contrib/intarray's
  definitions of the <@ and @> operators.  These are not in fact
  equivalent to the built-in anyarray operators of the same names,
  because they have different behavior for empty arrays, namely they
  don't think empty arrays are contained in anything.  That is
  mathematically wrong, no doubt, but until we can persuade GIN
  indexes to implement the mathematical definition we should probably
  not change this.  Another reason for not changing it now is that we
  can't yet ensure the opclasses will be updated correctly in a
  dump-and-reload upgrade.  Per recent discussions.

Joe Conway committed:

- In pgsql/contrib/dblink/dblink.c, fix dblink_get_result() as
  reported by Oleksiy Shchukin.  Refactor a bit while we're at it per
  request by Tom Lane.  Specifically, don't try to perform
  dblink_send_query() via dblink_record_internal() -- it was
  inappropriate and ugly.

- Add support for using SQL/MED compliant FOREIGN DATA WRAPPER,
  SERVER, and USER MAPPING as method to supply dblink connect
  parameters.  Per mailing list and PGCon discussions.

Heikki Linnakangas committed:

- In pgsql/src/backend/access/transam/xlog.c, only recycle normal
  files in pg_xlog as WAL segments.  pg_standby creates symbolic links
  with the -l option, and as Fujii Masao pointed out we ended up
  overwriting files in the archive directory before this patch.  Patch
  by Aidan Van Dyk, Fujii Masao and me.

Bruce Momjian committed:

- In pgsql/doc/src/sgml/config.sgml, add example of how to generate
  the session identifier from pg_stat_activity.

- In pgsql/src/backend/commands/copy.c, add comment about why
  "((void) 0)" is used in copy macros.

- Wording improvement for recent sesssion identifier SQL query.

- In pgsql/doc/src/sgml/config.sgml, wording improvement for recent
  sesssion identifier SQL query.

- In pgsql/doc/src/sgml/backup.sgml, remove sleep() from backup script
  example; not needed anymore.  Fujii Masao.

Andrew Dunstan committed:

- Initialise perl library as documented in perl API.  Backpatch to
  release 7.4.

- Search for versioned perl library instead of using hardcoded name on
  Windows.  Backpatch to release 8.3

- Adjust recent PERL_SYS_INIT3 call to avoid platforms where it might
  fail, and to remove compilation warning.  Backpatch the release 7.4.

== Abgelehnte Patches (bis jetzt) ==

No one was disappointed this week :-)

== Eingesandte Patches ==

  • Twitter
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at
  • Facebook
  • Google Bookmarks
  • FriendFeed
  • Digg Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009
  • Bloglines Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009
  • Technorati Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009
  • Fark this: Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at YahooMyWeb
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 at blogmarks
  • Bookmark Wöchentlicher PostgreSQL Newsletter - 07. Juni 2009 with wists
  • wong it!
  • Bookmark using any bookmark manager!
  • Stumble It!


No Trackbacks


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