Skip to content

Wöchentlicher PostgreSQL Newsletter - 12. Juli 2009

Der Originalartikel befindet sich unter:

== Wöchentlicher PostgreSQL Newsletter - 12. Juli 2009 ==

Es gibt noch verfügbare Plätze für den PgDay San Jose am Sonntag,
dem 19. Juli. Jetzt anmelden!

Es gibt ein neues Tool für Commitfests. Danke an Robert Haas, Brendan
Jurd und andere.

== PostgreSQL Produkt Neuigkeiten ==

check_postgres 2.9.2, ein Nagios Plugin zum monitoren von PostgreSQL,
ist erschienen.

PostgreSQL RPM Building Projekt hat die PostgreSQL 8.4 Live CD

pgbouncer 1.3.1, ein Connection Pooler für PostgreSQL, ist erschienen.

== PostgreSQL Jobs im Juli ==

== PostgreSQL Lokal ==

pgDay San Jose findet am Sonntag, dem 19. Juli 2009, unmittelbar
vor der OSCON statt. Anmelden unter:

Das OpenSQL Camp findet am 22. und 23. August in St. Augustin,
Deutschland, in der Nähe von Bonn und Köln, statt. Einsendeschluß für
den CfP ist der 19. Juli 2009. Vorträge einreichen!

Die Deutsche PostgreSQL Usergruppe wird einen Dev-Room auf der
FrOSCon am Sonntag, dem 23. August 2009 haben. Der Call for Papers
ist eröffnet:

Der CfP für PyCon Argentina ist bis zum 29. Juni 2009 23:59 ART
geöffnet. Die Konferenz selbst findet in Buenos Aires am 4. und 5.
September 2009 statt.

Andreas (ads) Scherbaum wird einen Kurs "PostgreSQL im
Unternehmenseinsatz" an der Volkshochschule Magdeburg, Deutschland,
vom 7. bis zum 11. Septemper 2009 geben. Details unter:

Es findet eine Konferenz in Athens, Georgia, USA am 19. September
2009 statt. Der CfP ist eröffnet.

Es findet eine Konferenz in Seattle, Washington, USA am 16.-18.
Oktober 2009 statt. Der CfP ist eröffnet:

PGCon Brasilien wird am 23.-24. Oktober 2009 auf dem Unicamp in
Campinas, Sao Paulo, stattfinden. Der CfP ist eröffnet!

PGDay.EU 2009 wird an der Telecom ParisTech Universität in Paris,
Frankreich, am 6. und 7. Nivember 2009 stattfinden. Der CfP ist
eröffnet. Vorträge einreichen!

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, Josh Berkus und Devrim GUNDUZ.

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

== Angewandte Patches ==

Peter Eisentraut committed:

- In pgsql/src/bin/psql/describe.c, show definition of index columns
  in \d on index.  This adds a column called "Definition" to the
  output of psql \d on an index, which shows the full expression
  behind the index column.  For indexes on plain columns, this is
  redundant,  but for expression indexes, this reveals the real
  expression.  Author: Khee Chin.

- More sensible character_octet_length.  For character types with
  typmod, character_octet_length columns in the information schema now
  show the maximum character length times the maximum length of a
  character in the server encoding, instead of some huge value as

- In pgsql/src/bin/psql/describe.c, sort child tables by name instead
  of OID in \d+ display.  This was an oversight in the recent patch.
  Found by Tom Lane.

- In pgsql/src/bin/psql/describe.c, psql backward compatibility fix.
  For servers older than 8.3, sort display of child tables by relname
  instead of oid::regclass::text, because the cast from regclass to
  text did not work back then.  The older display may be slightly
  worse when different schemas are involved, but that should be rare

- In pgsql/src/makefiles/, when calling unsupported "make
  check" with a pgxs module, return a nonzero exit code.

- In pgsql/doc/src/sgml/acronyms.sgml, correct what ISO stands for.

- Alter some gratuitous uses of "ANSI" when "SQL standard" might have
  been meant or the reference to a standard was unnecessary.

Tom Lane committed:

- Fix set_append_rel_pathlist() to deal intelligently with cases where
  substituting a child rel's output expressions into the appendrel's
  restriction clauses yields a pseudoconstant restriction.  We might
  be able to skip scanning that child rel entirely (if we get constant
  FALSE), or generate a one-time filter.  8.3 more or less
  accidentally generated plans that weren't completely stupid in these
  cases, but that was only because an extra recursive level of
  subquery_planner() always occurred and allowed const-simplification
  to happen.  8.4's ability to pull up appendrel members with non-Var
  outputs exposes the fact that we need to work harder here.  Per
  gripe from Sergey Burladyan.

- In pgsql/src/backend/utils/adt/timestamp.c, use floor() not rint()
  when reducing precision of fractional seconds in timestamp_trunc,
  timestamptz_trunc, and interval_trunc().  This change only affects
  the float-datetime case; the integer-datetime case already behaved
  like truncation instead of rounding.  Per gripe from Mario Splivalo.
  This is a pre-existing issue but I'm choosing not to backpatch,
  because it's such a corner case and there have not been prior
  complaints.  The issue is largely moot anyway given the trend
  towards integer datetimes.

- In pgsql/src/bin/psql/describe.c, code review for patch to show
  definition of index columns in \d on index.  Safely schema-qualify
  the pg_get_indexdef call, make the query a bit prettier in -E mode,
  remove useless join to pg_index, make it more obvious that the
  header[] array is not overrun.

- In pgsql/src/include/catalog/pg_proc.h, fix typo in comment.

- In pgsql/src/backend/catalog/information_schema.sql, query in SQL
  function still not schema-safe; add a couple more pg_catalog.

- In pgsql/src/backend/utils/mb/mbutils.c, don't use 'return' where
  you should use 'PG_RETURN_xxx'.

- In pgsql/src/bin/psql/describe.c, just a little more
  schema-qualification ...

- In pgsql/doc/src/sgml/datatype.sgml, add missing HOUR TO SECOND
  option to list of possible INTERVAL field sets, as noted by
  Sebastien Flaesch.  Also update the claim that we simply throw away
  fields outside this set --- that got changed later to only discard
  less-significant fields.

- In pgsql/src/backend/postmaster/postmaster.c, remove
  no-longer-necessary transmission of postmaster's LC_COLLATE and
  LC_CTYPE settings to children via BackendParameters.  Per
  discussion, the postmaster is now just using system defaults anyway,
  so we might as well save a few cycles during backend startup.

- In pgsql/contrib/xml2/xslt_proc.c, fix xslt_process() to ensure that
  it inserts a NULL terminator after the last pair of parameter
  name/value strings, even when there are MAXPARAMS of them.
  Aboriginal bug in contrib/xml2, noted while studying bug #4912
  (though I'm not sure whether there's something else involved in that
  report).  This might be thought a security issue, since it's a
  potential backend crash; but considering that untrustworthy users
  shouldn't be allowed to get their hands on xslt_process() anyway,
  it's probably not worth getting excited about.

- In pgsql/src/backend/optimizer/path/costsize.c, fix set_rel_width()
  to do something reasonable with non-Var items in a RelOptInfo
  targetlist.  It used to be that the only possibility other than a
  Var was a RowExpr representing a whole-row child Var, but as of
  8.4's expanded ability to flatten appendrel members, we can get
  arbitrary expressions in there.  Use the expression's type info and
  get_typavgwidth() to produce an at-least-marginally-sane result.
  Note that get_typavgwidth()'s fallback estimate (32 bytes) is the
  same as what was here before, so there will be no behavioral change
  for RowExprs.  Noted while looking at recent gripe about constant
  quals pushed down to FunctionScan appendrel members ...  not only
  were we failing to recognize the constant qual, we were getting the
  width estimate wrong :-(

- Move some declarations in the raw-parser header files to create a
  clearer distinction between the external API (parser.h) and
  declarations that only need to be visible within the raw parser code
  (gramparse.h, which now is only included by parser.c, gram.y,
  scan.l, and keywords.c).  This is in preparation for the upcoming
  change to a reentrant lexer, which will require referencing YYSTYPE
  in the declarations of base_yylex and filtered_base_yylex, hence
  gram.h will have to be included by gramparse.h.  We don't want any
  more files than absolutely necessary to depend on gram.h, so some
  cleanup is called for.

- Fix up PGDLLIMPORT marking for standard_conforming_strings.  Moving
  it into a header file that plpgsql's scan.l can see broke the
  previous kluge.  Per buildfarm results.

- Advance the minimum required version of "flex" from 2.5.4 to 2.5.31,
  and update documentation accordingly.  This is required in order to
  have support for a reentrant scanner.  I'm committing this bit
  separately in order to have an easy reference if we later decide to
  make the minimum something different (like 2.5.33).

- Convert the core lexer and parser into fully reentrant code, by
  making use of features added to flex and bison since this code was
  originally written.  This change doesn't in itself offer any new
  capability, but it's needed infrastructure for planned improvements
  in plpgsql.  Another feature now available in flex is the ability to
  make it use palloc instead of malloc, so do that to avoid possible
  memory leaks.  (We should at some point change the other lexers
  likewise, but this commit doesn't touch them.)

- In pgsql/src/backend/parser/scan.l, although the flex documentation
  avers that yyalloc and yyrealloc take size_t arguments, the emitted
  scanner actually prototypes them with type yy_size_t, which is
  sometimes not the same thing depending on flex version and platform.
  Easiest fix seems to be to use yy_size_t.  Per buildfarm results.

- Make the configure messages rejecting old bison and flex versions
  include the full path and version of the program being rejected.

Heikki Linnakangas committed:

- In pgsql/src/backend/utils/adt/formatting.c, fix ancient bug in
  handling of to_char modifier 'TH', when used with Hitoshi Harada.
  In what seems like an oversight, we used to treat 'TH' the same as
  lowercase 'th', but only with HH/HH12.

- In pgsql/src/backend/utils/init/postinit.c, need to use
  pg_perm_setlocale when setting LC_CTYPE and LC_COLLATE at startup.
  Otherwise, the LC_CTYPE/COLLATE setting gets reverted when using
  plperl, which leads to incorrect query results and index corruption.
  This was accidentally broken in the per-database locale patch in
  8.4.  Pointed out by Andrew (RhodiumToad) Gierth.

Bruce Momjian committed:

- Stamp minor library version numbers for 8.5; sorry for the delay.

== Abgelehnte Patches (bis jetzt) ==

No one was disappointed this week :-)

== Eingesandte Patches ==

Fujii Masao sent a WIP patch to introduce a new value,
sync_file_range, to wal_sync_method.

Heikki Linnakangas sent in a WIP patch to set client_encoding
automatically from locale in libpq.

Pavel Stehule sent in a patch to fix BUG #4907: stored procedures and
changed tables.

ITAGAKI Takahiro sent in two patches to parallelize pgbench.

Robert Haas sent in three patches to help fix the issue around the
*_collapse_limit GUC parameters.

Josh Tolley sent in a patch against the Commitfest code to keep
Perl::Critic quiet.

Peter Eisentraut sent in two revisions of a patch to allow byteas to
be escaped for input and output as hex.

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

Brendan Jurd sent in another revision of the patch to add scientific
notation to to_char().

Bernd Helmle sent in another revision of a patch to allow for dropping
columns and constraints conditionally.

Tom Lane sent in a patch to increase the minimum flex version in order
to enable a reentrant parser.

Dean Rasheed sent in two WIP patches to allow deferrable unique

Marko (johto) Tiikkaja sent in a patch to allow INSERT...RETURNING
inside a CTE.

Andrew (RhodiumToad) Gierth sent in a patch to update hstore to CVS

Teodor Sigaev sent in a patch to fix wrong algorithms for "contains"
and "overlaps" operations over polygons, including regression tests
for same.

Jaime Casanova sent in an update to Mark Kirkwood's patch which adds
lock wait statistics to the pg_catalog, including a new view.


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