PostgreSQL Weekly News - June 11th 2003

From: Robert Treat <xzilla(at)users(dot)sourceforge(dot)net>
To: pgsql-announce(at)postgresql(dot)org
Subject: PostgreSQL Weekly News - June 11th 2003
Date: 2003-06-12 00:47:14
Message-ID: 1055378834.7086.263.camel@camel
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-announce

== PostgreSQL Weekly News - June 11th 2003 ==

Last week there were some problems with the PostgreSQL project's back
end servers which disrupted the websites, mailing lists, cvs, and
probably some other stuff too. It should all be fixed at this point so
if you do notice something wrong please either post to one of the lists
or email webmaster(at)postgresql(dot)org(dot)

In all that chaos, last week's weekly news seemed to die somewhere
between here and there, so I'll do my best to catch everyone up on
what's been going on with the project for the last two weeks. The
biggest change is that 7.4 code freeze and beta testing is being pushed
back 2 weeks to account for the cvs downtime. Code freeze will now be
July 1st, with beta testing starting July 15th. This should allow
everyone enough time to get their patches in and get the currently
submitted patches all caught up.

The list of miscellaneous fixes is fairly good sized so lets get to it.
First and foremost, libpq is now able to talk to either 3.0 or 2.0
protocol servers. It first tries protocol 3, then falls back to 2 if
postmaster rejects the startup packet with an old-format error message.
A side benefit of the rewrite is that SSL-encrypted connections should
now be able to be made without blocking. Also all flex lexers that are
part of the backend now ensure that a yy_fatal_error() call results in
elog(ERROR) not exit().

There was a fix so that an error occurring during standalone-backend
processing in initdb will result in exit(1) (allowing the initdb script
to realize that there's something wrong). We no longer explicitly
close() the session socket during backend exit; instead it is left open
for the kernel to do after the process dies. (This allows clients to
wait for the backend to exit if they wish). The latest bison spelling of
'syntax error' (rather than 'parse error') was adopted for
grammar-detected problems, and the configure script was modified to
check for bison version >= 1.875. Some other smaller fixes included
fixing markTargetListOrigin which neglected to handle outer-scope Vars
properly, added support for Intel compiler on Linux, added some
information schema views about functions, a configure check to use
$YACC, an uninitialized memory bug was fixed, added -DFRONTEND for
non-MinGW Win32 ports, and -fPIC was change to -fpic on HPPA.

Two nice performance improvements include one for hash joins and hash
aggregation where, when the plan is being ReScanned, the hash table no
longer requires a rebuild if there is no parameter change for its child
node. Some planner performance problems with large WHERE clauses was
fixed by introducing new 'FastList' list-construction subroutines to use
in hot spots. This avoids the O(N^2) behavior of repeated lappend's by
keeping a tail pointer, while not changing behavior by reversing list
order as the lcons() method would do.

We also saw some enhanced functionality put in; namely the replacing of
the functional-index facility with expressional indexes. Now any column
of an index can now be a computed expression instead of a simple
variable. Restrictions on expressions are the same as for predicates
(only immutable functions, no sub-selects). This fixes problems
recently introduced with in-lining SQL functions, because the in-lining
transformation is applied to both expression trees so the planner can
still match them up. At the same time index predicate handling was made
more efficient (both predicates and index expressions are now cached by
the relcache). The other change was to implement outer-level aggregates
to conform to the SQL spec, with extensions to support our historical
behavior. The implementation involves adding an agglevelsup field to
Aggref, and treating outer aggregates like outer variables at planning
time.

The JDBC protocol should now have integrated support for the new front
end/back end protocol.It still has some issues (SSL support for one) but
is definitely test worthy and the build number was bumped up to reflect
this change. Other recent changes include some cleanups to deal with
work better with the Eclipse IDE, improvement to buffer sizes to avoid
necessary resizing, some small type fixes, and proper support for
removal of autocommit. JDBC users should also be aware that a beta
version of the JDBC driver was included in the 7.3.3 rpm; the latest
driver is available at jdbc.postgresql.org.

ECPG also had a number of fixes put in including forward definition of
structs, fix variable parsing so struct[n].attr works, a prototype
function for describe, and some other minor cleanup.

We've also had a number of FAQ updates, including updates for the
German FAQ (Ian Barwick), the Russian FAQ (Viktor Vislobokov) and the
addition of a Turkish FAQ (Devrim GUNDUZ). And I can't forget to mention
some enhancements to pg_dumpall. It now includes GRANT/REVOKE for
database-level permissions in its output, works with server versions
back to 7.0, and supports the -a, -s, -x options of pg_dump.

And one final note from the "hey that's cool" file, we received a
report on the mailing lists this week of the successful use of
PostgreSQL on a dual K8D-based Opteron box running Debian Linux in
64-bit mode. As an extra bonus, no code changes were required for the
compile, just the addition of a few compiler flags. Way to go!

== PostgreSQL Product News ==

KmusicdB 0.11.0
http://apps.kde.com/uk/0/info/vid/9267?sid=

PostgreSQL 7.3.3 Package for OS X
http://www.macupdate.com/info.php/id/8561

== PostgreSQL In the News ==

Industrial Strength MVC (using PHP and PostgreSQL)
http://newsvac.newsforge.com/newsvac/03/06/09/2020257.shtml?tid=15

The Register signs up for Bricolage
http://www.theregister.co.uk/content/31/30959.html

Beyond Linux
http://www.infoworld.com/article/03/05/23/21OPconnection_1.html?platforms

== Upcoming Events ==

Open Source Conference: Portland, Oregon: July 7-11
A PostgreSQL track is available this year
http://conferences.oreillynet.com/os2003

== PostgreSQL Weekly News - June 11th 2003 ==

Don't forget to read Elein Mustain's Weekly Summary of the PostgreSQL
General Mailing List http://www.varlena.com/GeneralBits/

On the Web:
http://www.postgresql.org
http://advocacy.postgresql.org

Browse pgsql-announce by date

  From Date Subject
Next Message David Wheeler 2003-06-12 17:05:09 ANNOUNCE: Bricolage 1.6.1
Previous Message The Hermit Hacker 2003-06-09 04:45:13 Archives re-generating ...