Skip to content

Wöchentlicher PostgreSQL Newsletter - 25. September 2011

Der Originalartikel befindet sich unter:

== Wöchentlicher PostgreSQL Newsletter - 25. September 2011 ==

Security releases 9.1.1, 9.0.5, 8.4.9, 8.3.16 and 8.2.22 are out.
Sicherheitsupdates 9.1.1, 9.0.5, 8.4.9, 8.3.16 und 8.2.22 sind
erschienen. Zeitnah updaten, wenn Sie betroffen sind. Details unter:

== PostgreSQL Produkt Neuigkeiten ==

MicroOLAP Database Designer 1.8.1 für PostgreSQL released, jetzt mir
Support für PostgreSQL 9.1.

== PostgreSQL Jobs im September ==

== PostgreSQL Lokal ==

PostgreSQL Conference West (#PgWest) findet vom 27. bis 30. September
2011 im San Jose Convention Center in Jan Jose, Kalifornen, USA statt.

PostgreSQL Conference Europe 2011 findet vom 18. bis
21. Oktober in Amsterdam statt.

PG-Day Denver 2011 findet am Freitag, dem 21. Oktober 2011 auf dem
Auraria Campus in der Nähe von Downtown Denver, Colorado statt.

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

PGConf.DE 2011 ist die Deutschsprachige PostgreSQL Konferenz
und wird am 11. November 2011 im Rheinischen Industriemuseum
in Oberhausen, Deutschland, stattfinden. Der Call for Papers ist offen.

Die fünfte Edition des Italienischen PostgreSQL
Tags (PGDay.IT 2011) findet am 25. November in
Prato, Italien statt.

Der Call for Papers für die FLOSS UK ist offen, die Veranstaltung
findet in Edinburgh vom 20. bis 22. März 2012 statt. Der
Einsendeschluß für Einreichungen ist der 18. November 2011,
ausgewählte Sprecher werden bis zum 25. November benachrichtigt.
Einreichungen sollen an postgresql2012 AT flossuk DOT org gehen.
Mehr Informationen:

== 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

== Angewandte Patches ==

Tom Lane pushed:

- Improve reporting of newlocale() failures in CREATE COLLATION.  The
  standardized errno code for "no such locale" failures is ENOENT,
  which we were just reporting at face value, viz "No such file or
  directory".  Per gripe from Thom Brown, this might confuse users, so
  add an errdetail message to clarify what it means.  Also, report
  newlocale() failures as ERRCODE_INVALID_PARAMETER_VALUE rather than
  using errcode_for_file_access(), since newlocale()'s errno values
  aren't necessarily tied directly to file access failures.

- Suppress "unused function" warning when not HAVE_LOCALE_T.  Forgot
  to consider this case ...

- Make EXPLAIN ANALYZE report the numbers of rows rejected by filter
  steps.  This provides information about the numbers of tuples that
  were visited but not returned by table scans, as well as the numbers
  of join tuples that were considered and discarded within a join plan
  node.  There is still some discussion going on about the best way to
  report counts for outer-join situations, but I think most of what's
  in the patch would not change if we revise that, so I'm going to go
  ahead and commit it as-is.  Documentation changes to follow (they
  weren't in the submitted patch either).  Marko Tiikkaja, reviewed by
  Marc Cousin, somewhat revised by Tom Lane

- Update release notes for 9.1.1, 9.0.5, 8.4.9, 8.3.16, 8.2.22.  Man,
  we fixed a lotta bugs since April.

- Stamp 8.2.22, 8.3.16, 8.4.9, 9.0.5, 9.1.1.

- Update for the new location of our Windows timezone
  map.  I wasn't aware of this script till Magnus Hagander mentioned
  it just now ...

- Fix our mapping of Windows timezones for Central America.  We were
  mapping "Central America Standard Time" to "CST6CDT", which seems
  entirely wrong, because according to the Olson timezone database
  noplace in Central America observes daylight savings time on any
  regular basis --- and certainly not according to the USA DST rules
  that are implied by "CST6CDT".  (Mexico is an exception, but they
  can be disregarded since they have a separate timezone name in
  Windows.)  So, map this zone name to plain "CST6", which will
  provide a fixed UTC offset.  As written, this patch will also result
  in mapping "Central America Daylight Time" to CST6.  I considered
  hacking things so that would still map to CST6CDT, but it seems it
  would confuse to put those two names in separate
  entries.  Since there's little evidence that any such zone name is
  used in the wild, much less that CST6CDT would be a good match for
  it, I'm not too worried about what we do with it.  Per complaint
  from Pratik Chirania.

- Recognize self-contradictory restriction clauses for non-table
  relations.  The constraint exclusion feature checks for
  contradictions among scan restriction clauses, as well as
  contradictions between those clauses and a table's CHECK
  constraints.  The first aspect of this testing can be useful for
  non-table relations (such as subqueries or functions-in-FROM), but
  the feature was coded with only the CHECK case in mind so we were
  applying it only to plain-table RTEs.  Move the
  relation_excluded_by_constraints call so that it is applied to all
  RTEs not just plain tables.  With the default setting of
  constraint_exclusion this results in no extra work, but with
  constraint_exclusion = ON we will detect optimizations that we
  missed before (at the cost of more planner cycles than we expended
  before).  Per a gripe from Gunnlaugur Þór Briem.  Experimentation
  with his example also showed we were not being very bright about the
  case where constraint exclusion is proven within a subquery within
  UNION ALL, so tweak the code to allow set_append_rel_pathlist to
  recognize such cases.

- Un-break compression of plain-text output format in pg_dump.
  pg_dump has historically understood -Z with no -F switch to mean
  that it should emit a gzip-compressed version of its plain text
  output.  This got broken through a misunderstanding in the 9.1 patch
  that added directory output format.  Restore the former behavior.
  Per complaint from Roger Niederland and diagnosis by Adrian Klaver.

- Avoid unnecessary snapshot-acquisitions in BuildCachedPlan.  I had
  copied-and-pasted a claim that we couldn't reach this point when
  dealing with utility statements, but that was a leftover from when
  the caller was required to supply a plan to start with.  We now will
  go through here at least once when handling a utility statement, so
  it seems worth a check to see whether a snapshot is actually needed.
  (Note that analyze_requires_snapshot is quite a cheap test.) Per
  suggestion from Yamamoto Takashi.  I don't think I believe that this
  resolves his reported assertion failure; but it's worth changing
  anyway, just to save a cycle or two.

- Fully const-ify PQconnectdbParams, PQconnectStartParams, and
  PQpingParams.  The keywords and values arguments of these functions
  are more properly declared "const char * const *" than just "const
  char **".  Lionel Elie Mamane, reviewed by Craig Ringer

Robert Haas pushed:

- Fix another bit of unlogged-table-induced breakage.  Per bug #6205,
  reported by Abel Abraham Camarillo Ojeda.  This isn't a particularly
  elegant fix, but I'm trying to minimize the chances of causing yet
  another round of breakage.  Adjust regression tests to exercise this

- Add --{no-,}replication flags to createuser.  Fujii Masao, reviewed
  by Cédric Villemain, with some doc changes by me.

- Add missing brackets to chkselinuxenv.

- Document some more apparently-harmless error messages.

- Teach sepgsql about database labels.  This is still a bit of a hack,
  but it's better than the old way, for sure.  KaiGai Kohei, with one
  change by me to make it compile

- Memory barrier support for PostgreSQL.  This is not actually used
  anywhere yet, but it gets the basic infrastructure in place.  It is
  fairly likely that there are bugs, and support for some important
  platforms may be missing, so we'll need to refine this as we go

Peter Eisentraut pushed:

- Translation updates

Simon Riggs pushed:

- synchronous_commit is an enum not a boolean.  Jaime Casanova

Magnus Hagander pushed:

- Note that sslmode=require verifies the CA if root cert is present
  This mode still exists for backwards compatibility, making
  sslmode=require the same as sslmode=verify-ca when the file is
  present, but not causing an error when it isn't.  Per bug 6189,
  reported by Srinivas Aji

- Fix typo

Bruce Momjian pushed:

- Expand pgrminclude to exclude use of macros CppAsString and

- Document pgrminclude limitations.

- In pgrminclude, document requirement to use pgcompinclude, and sort
  files so include removal is more predictable.

== Abgelehnte Patches (bis jetzt) ==

No one was disappointed this week :-)

== Eingesandte Patches ==

Jeff Davis sent in another revision of the patch to create range

Peter Geoghegan sent in three revisions of a patch to speed up
comparators by inlining them.

Muhammad Asif sent in two revisions of a patch to fix an issue with
BSD sockets on HP-UX.

Alexander Korotkov sent in two more revisions of the double sorting
split patch.

Fujii Masao sent in another revision of the patch to allow doing a
backup from a hot standby.

Pavel Stehule sent in another revision of the patch to remove
unnecessary ccache search when an array variable is updated in

Yeb Havinga sent in a patch to refine dependency checking in

Magnus Hagander sent in two revisions of a patch to make the TABLE
command tab-complete both tables and views in psql.  Before, it only
tab-completed tables.

Magnus Hagander sent in a patch to add a call to posix_fadvise with
POSIX_FADV_DONTNEED on all the files being read when doing a base
backup, to help the kernel not to trash the filesystem cache.

Oleg Bartunov sent in another revision of the patch to enable
space-partitioned GiST indexes.

Marti Raudsepp sent in another revision of the patch to cache stable
expressions with constant arguments.

ITAGAKI Takahiro sent in a patch to allow COPY to support UTF8 files
with a byte order mark (BOM).


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