Skip to content

Wöchentlicher PostgreSQL Newsletter - 03. April 2011

Der Originalartikel befindet sich unter:

== Wöchentlicher PostgreSQL Newsletter - 03. April 2011 ==

PostgreSQL 9.1alpha5 erschienen, mit einer Menge Verbesserungen!
Dies ist die letzte geplante Alpha für 9.1. Du kannst sie
hier bekommen:

Die Türkische PostgreSQL User Group organisiert eine eintägige
Veranstaltung am 30. April 2011. Folge @PgDayTR für Details,
auf Türkisch.

== PostgreSQL Produkt Neuigkeiten ==

DBD::Pg 2.18.0, ein Perl Treiber für
PostgreSQL, ist erschienen.

== PostgreSQL Lokal ==

Es wird eine umfangreiche PostgreSQL Präsenz auf der MySQL Konferenz
und Expo am am 11-14. April 2011 in Santa Clara, Kalifornien, geben.

Das Open Database Camp findet vom 7. bis 9. Mai 2011 in Sardinien,
Italien statt.

PGCon findet am 19. und 20. Mai 2011 an der Universität
von Ottawa statt, vorher gibt es am 17. und 18. Mai
zwei Tage mit Trainings.

Das Open Database Camp findet vom 7. bis 9. Mai 2011 in Sardinien,
Italien statt.

pgbr findet in Sao Paulo, Brazilien, am 3. und 4. November 2011 statt.

== PostgreSQL in den News ==

Planet PostgreSQL:

Dieser wöchentliche PostgreSQL Newsletter wurde erstellt von David Fetter.

Sende Neuigkeiten und Ankündigungen bis Sonntag, 15 Uhr Pazifischer
Zeit. Bitte sende englische Beiträge an, deutsche an, italienische an, spanische an

== Reviews ==

== Angewandte Patches ==

Tom Lane pushed:

- Get rid of links that don't work when building HISTORY.

- Prevent a rowtype from being included in itself.  Eventually we
  might be able to allow that, but it's not clear how many places need
  to be fixed to prevent infinite recursion when there's a direct or
  indirect inclusion of a rowtype in itself.  One such place is
  CheckAttributeType(), which will recurse to stack overflow in cases
  such as those exhibited in bug #5950 from Alex Perepelica.  If we
  were sure it was the only such place, we could easily modify the
  code added by this patch to stop the recursion without a complaint
  ... but it probably isn't the only such place.  Hence, throw error
  until such time as someone is excited enough about this type of
  usage to put work into making it safe.  Back-patch as far as 8.3.
  8.2 doesn't have the recursive call in CheckAttributeType in the
  first place, so I see no need to add code there in the absence of
  clear evidence of a problem elsewhere.

- Fix typo in PQconnectStartParams().  This would lead to leaking the
  PGconn structure after an error detected by conninfo_array_parse(),
  as well as failing to return a useful error message in such cases.
  Backpatch to 9.0 where the error was introduced.  Joseph Adams

Alvaro Herrera pushed:

- Add missing #include

Robert Haas pushed:

- Mark up release notes using <link>.  Unlike <xref>, this actually

- Fix compiler warning.

- Support comments on FOREIGN DATA WRAPPER and SERVER objects.  This
  mostly involves making it work with the objectaddress.c framework,
  which does most of the heavy lifting.  In that vein, change
  GetForeignDataWrapperOidByName to get_foreign_data_wrapper_oid and
  GetForeignServerOidByName to get_foreign_server_oid, to match the
  pattern we use for other object types.  Robert Haas and Shigeru

- pg_dump support for comments on FOREIGN DATA WRAPPER and SERVER
  objects.  Shigeru Hanada, with some corrections.


- Avoid possible hang during smart shutdown.  If a smart shutdown
  occurs just as a child is starting up, and the child subsequently
  becomes a walsender, there is a race condition: the postmaster might
  count the exstant backends, determine that there is one normal
  backend, and wait for it to die off.  Had the walsender transition
  already occurred before the postmaster counted, it would have
  proceeded with the shutdown.  To fix this, have each child that
  transforms into a walsender kick the postmaster just after doing so,
  so that the state machine is certain to advance.  Fujii Masao

- Remove mention of using "man" from the tutorial.  This isn't
  applicable on Windows, and the internal link to the psql
  documentation should be more than sufficient.  Susanne Ebrecht

- Improve documentation on the range of the numeric data type.  Gianni
  Ciolli, reviewed by Noah Misch.

- Rearrange "add column" logic to merge columns at exec time.  The
  previous coding set attinhcount too high in some cases, resulting in
  an undumpable, undroppable column.  Per bug #5856, reported by Naoya
  Anzai.  See also commit 31b6fc06d83c6de3644c8f2921eb7de0eb92fac3,
  which fixes a similar bug in ALTER TABLE .. ADD CONSTRAINT.  Patch
  by Noah Misch.

Bruce Momjian pushed:

- Add 9.1 release note link for synchronous replication.  We already
  had links to the GUC variables that control it.

- In pg_upgrade, add C comment about how autovacuum is disabled.

- Expand comment on how pg_upgrade is turning off autovacuum.

- Mention pg_upgrade sets autovacuum_freeze_max_age to maximum.

Peter Eisentraut pushed:

- Add maintainer-check target.  This can do various source code checks
  that are not appropriate for either the build or the regression
  tests.  Currently: duplicate_oids, SGML syntax and tabs check, NLS
  syntax check.

- Make duplicate_oids return nonzero exit status if duplicates were
  found.  Automatic detection of errors is easier that way.

- Update SQL features list.  Feature F692 "Extended collation support"
  is now also supported.  This refers to allowing the COLLATE clause
  anywhere in a column or domain definition instead of just directly
  after the type.  Also correct the name of the feature in accordance
  with the latest SQL standard.

Heikki Linnakangas pushed:

- Adjust error message, now that we expect other message types than
  connection close at this point.  Fix PQsetnonblocking() comment.
  Fujii Masao

- Automatically terminate replication connections that are idle for
  more than replication_timeout (a new GUC) milliseconds.  The TCP
  timeout is often too long, you want the master to notice a dead
  connection much sooner.  People complained about that in 9.0 too,
  but with synchronous replication it's even more important to notice
  dead connections promptly.  Fujii Masao and Heikki Linnakangas

- Check that we've reached end-of-backup also when we're not
  performing archive recovery.  It's possible to restore an online
  backup without recovery.conf, by simply copying all the necessary
  WAL files to pg_xlog. "pg_basebackup -x" does that too. That's the
  use case where this cross-check is useful.  Backpatch to 9.0. We
  used to do this in earlier versins, but in 9.0 the code was
  inadvertently changed so that the check is only performed after
  archive recovery.  Fujii Masao.

- Improve error message when WAL ends before reaching end of online

- Reword the phrase on zero replication_timeout in the docs.

- Don't leak the temporary PLyProcedure struct we create for inline
  plpython blocks.  Investigation by Jan Urbański, though I didn't use
  his patch.

- Increase SHMEM_INDEX_SIZE from 32 to 64.  We're currently at 40
  entries in ShmemIndex, so 64 leaves some headroom.  Kevin Grittner

- Fix a tiny race condition in predicate locking. Need to hold the
  lock while examining the head of predicate locks list.  Also, fix
  the comment of RemoveTargetIfNoLongerUsed, it was neglected when we
  changed the way update chains are handled.  Kevin Grittner.

- Fix two missing spaces in error messages.  Josh Kupershmidt.

- Escape greater than and less than characters in docs.  Susanne
  Ebrecht and me.

Andrew Dunstan pushed:

- Attempt to unbreak windows builds broken by commit 754baa2.

Magnus Hagander pushed:

- Avoid palloc before CurrentMemoryContext is set up on win32.
  Instead, write the unconverted output - it will be in the wrong
  encoding, but at least we don't crash.  Rushabh Lathia

== Abgelehnte Patches (bis jetzt) ==

No one was disappointed this week :-)

== Eingesandte Patches ==

Heikki Linnakangas sent in a variant patch to prevent memory leaks in
DO blocks.

Heikki Linnakangas sent in another revision of the patch to add a
replication server timeout.

Bernd Helmle sent in another revision of the patch to use a more
reasonable size for \dt in psql.

Simon Riggs sent in a patch which adds 5 more levels of durability to
sync rep.

Joseph Adams sent in five more revisions of the patch to add JSON as
a data type.

Kevin Grittner sent in a patch to prevent large tables from breaking
SSI by overflowing predicate lock hash buckets.

Peter Eisentraut sent in a patch to fix complaints from gcc 4.6.

Peter Eisentraut sent in a patch to fix a case where typed tables
broke pg_upgrade.

Robert Haas and Shigeru HANADA traded patches to allow commenting on
SQL/MED objects.

Peter Eisentraut sent in a patch to fix how exit_nicely() works.

Radoslaw Smogura sent in another revision of the patch to add a second
level buffer cache.

Jan Wieck sent in a patch to fix an issue with autovacuum performance.

Merlin Moncure sent in a patch to fix an issue with the hint bit

Alexey Klyukin sent in a patch to help build the infrastructure for a
validator for configuration files.

Bruce Momjian sent in a patch to adjust a comment in autovacuum.c.

Joseph Adams sent in a patch to allow casting int2, int4 and int8 to

Shigeru HANADA sent in two revisions of a patch to fix the foreign
table docs.

Brendan Jurd sent in a patch to rationalize parsing of pg_hba.conf.

Rushabh Lathia sent in a patch to fix a crash on Windows.

Dan Ports sent in a patch to display the pid of the holding
transaction in SSI.

Dan Ports sent in a patch to fix an issue in SSI where a lock could be
taken but never cleared.


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