== PostgreSQL Weekly News - Feb 4th 2003 ==
7.3.2 should be released either today or tomorrow, barring any
last minute meltdowns. Keep an eye out on the mailing lists or the
PostgreSQL website for the official announcement.
Also worthy of a mention, last Friday 7.2.4 was released. It
contains some important bug fixes for the 7.2.x series. While it is
still recommended that folks upgrade to 7.3.2 asap, if you can't for
some reason 7.2.4 should help tide you over. Again, for those who want
to read the official announcement it is available
http://archives.postgresql.org/pgsql-announce/2003-01/msg00020.php
In other development news, an issue with plpgsql parsing mixed-case
names has been resolved (Neil Conway), and there were some improvements
to plpython error handling (Tom Lane) and in handling of NULL's (Andrew
Bosma). There was also a work around put in place for libpgtcl to
prevent memory leaks and core dumps during connection shutdown that
could sometimes occur due to a bug in tcl. With a couple more changes
this week from Oliver Jowett and Kris Jurkas, the jdbc interface is
really getting a wide base of people pitching in. Keep up the good work
guys.
Another bug was fixed that was causing folks with all numeric user
names. This one made it into 7.3.2, thanks to all who reported it. Other
miscellaneous improvements include pg_dump's handling of large objects
have been improved, and I change in the optimizer that should result in
a 10% increase in select count(*) from foo queries. Rod Taylor also has
improved the ALTER DOMAIN functionality to work with domain constraint
operations stored in rules.
Something that happened late last week that I thought was interesting
was a report that was received on the -bugs list from a group of
Stanford researchers. They had developed a tool to detect potential
out-of-bonds array accesses and buffer overruns. They had been working
on linux for a few years, but in an effort to branch out they decided to
try there tool against the PostgreSQL source code. While not all of
there reported bugs were actual bugs, working with some of the
PostgreSQL developers they were able to fix some problems in the
PostgreSQL sources and also make some improvements in their tool. Thanks
to Yichen Xie and his gang. For those who want to read the thread it
starts at http://archives.postgresql.org/pgsql-bugs/2003-01/msg00175.php
On a more practical level, there was another good thread in the -admin
list about porting a database system from oracle to postgresql. At first
he was seeing dreadful performance compared to what he had gotten from
oracle, in the range of 1/5th the number of transaction per second for
select tests. The problem ended up being that he was trying to run
comparisons of integer based numbers on bigint based indexes, a common
problem for folks new to postgresql. However with a little review and
suggestions from the PostgreSQL faithful, he was able to correct the
problem and his read test now are perform almost *5 times faster* on
PostgreSQL than they did with oracle. If you'd like to read the whole
thread, which has data on a number of different methods used in testing,
plus numbers for write testing, you can check out this thread at
http://archives.postgresql.org/pgsql-performance/2003-01/msg00214.php
== PostgreSQL Product News ==
Incognito Software Introduces Name Commander(tm) with Successful
Deployment at BT
http://www.canadait.com/cfm/index.cfm?It=106&Id=14265&Se=0&Sv=&Lo=2
Npgsql adopted into the Mono Class Library
http://www.postgresql.com/news.php?NewsID=124
== PostgreSQL In the News ==
Making a Living Saving the Government Money
http://newsforge.com/article.pl?sid=03/01/28/1829254&mode=thread&tid=3
== Upcoming Events ==
FOSDEM: Brussels, Belgium: February 8-9
A two-day PostgreSQL training track will be part of the FOSDEM
conference.
http://candle.pha.pa.us/fosdem/
http://www.fosdem.org/index
== PostgreSQL Weekly News - Feb 4th 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