PostgreSQL Weekly News - August 6th 2003

== PostgreSQL Weekly News - August 6th 2003 ==

Well, I've been talking about for a few weeks now, and it has finally
happened: 7.4 has gone beta! The tar ball was wrapped up Monday night,
and by now should be available from your local mirror. I'm happy to see
a bzip'd version is available, I used that myself for my test run. If
your wondering what you can do to help, a good start is to download,
compile, install, and then load your most recent dump file from your
production systems. If you encounter any troubles they can be reported
to pgsql-bugs(at)postgresql(dot)org(dot) If you want to follow along, subscribe to
the pgsql-hackers list and keep an eye on
http://developers.postgresql.org.

We've already had a couple of things shake out from the beta release,
and we have a good list of items fixed just before packaging, so we
might as well cover those items first. Array[] constructs now return a
NULL array (rather than error) if any input element is NULL. This was
done as an intermediate step until arrays can have null elements.
Numeric_smaller, numeric_larger, float4smaller, float4larger,
float8smaller, float8larger (and thereby the MIN/MAX aggregates on these
datatypes) to agree with the datatypes' comparison operations in regards
to NaN handling.

A patch from Andreas Pflug added pretty-printing variants of
pg_get_viewdef and related functions. The parsing code for ACLs was
upgraded to be more cognizant of the actual logical structure and
quoting rules being used. We now unset PGUSER and other variables we
might have inherited from the environment that could possibly cause psql
to fail to connect when working with temp installations. We no longer
pull in libbind except on BeOs. The inline_function test for overly
complex parameters was tightened up a bit. The --enable-recode feature
was removed due to its brokenness with the IPv6 changes and an apparent
lack of anyone actually using it. Some minor cleanups in S_LOCK_TEST
code. Thread handling in configure was fixed. Some more issues with
testing error returns from SSL were resolved and fixes were put into the
back end for the problems with SSL_read and SSL_write similar to what
was required by libpq.

We now (properly) print the servers failure message before trying to
reconnect, not after. USE_THREADS was re-added in places where it was
needed by ecpg. Vacuumlo was made properly schema-aware and had its
documentation improved. The HISTORY file was updated to reflect change
for 7.4, and a pgindent run was done to help clean everything up a bit.
Several places where fractional-second inputs were mis-processed in
HAVE_INT64_TIMESTAMP cases were fixed, thanks in part to a bug report
from Philipp Reisner. An ecpg thread testing file was added.

The dbf2pg functionality (which allows user to insert xBase-style
.dbf-files into a PostgreSQL table) got a bit of a face lift thanks to
Thomas Behr. Some improvements to the "-s oldname=newname" functionality
were put in, compatibility with the COPY command was improved, and the
README file was tweaked.

Tsearch2 also received some compatibility adjustments, namely creating
sub-Makefiles for compiling on certain platforms. Also a ts_debug
function was added to help debug configurations, and documentation for
this was added in.

Of course the docs were updated to reflect proper version info and
copyright notices, but several other documentation updates went in
including improvements to ParseDateTime() and a note that AND and OR are
commutative as oddly they are not in some other databases. Devrim Gunduz
also provided an update to the Turkish FAQ, and speaking of
translations, the some more PostgreSQL translations have rolled in.

The new library-preload feature now reports error if the specified
initialization function is not found, and the initialization functions
for the pl libraries all have standard names. Some final error code
assignments were touched upon, and a few mistaken assignments were
fixed. Support for STATEMENT triggers was added for pltcl and plpython,
along with some documentation of just how our statement level triggers
work as there are some wide differences in some of the other major
databases statement level trigger support. Speaking of procedural
languages, as most of you know PostgreSQL already has the most extensive
support for for procedural languages of any dbms out there, but things
just keep on getting better. This week the folks from Command Prompt
have released a beta version of plphp, a PHP based procedural language
for PostgreSQL. More information and mailing lists are available at
http://www.commandprompt.com/entry.lxp?lxpe=260, you are encouraged to
test it out and see how well it works. As for PostgreSQL, this means we
now have more than 10 different procedural languages available for your
database function hacking pleasure, and I hear rumors that some folks
are already working on a plmono, so watch out.

Some final quick notes about things recently added to the TODO list:
* Allow long lines to include session-level information like
database/user
* Allow server log information to be output as inserts
* Add fflush of stdout when outputting a query in psql
* Implement dirty reads or shared row locks and use them in RI
triggers
* Allow plpgsql EXECUTE query_var INTO record_var

If you didn't get in on the fun of 7.4, nows your chance for an early
entry into 7.5 :-)

== PostgreSQL Product News ==

Alpha version of PgSqlClient, an ADO.NET Data Provider for PostgreSQL 7.4+
http://pgsqlclient.sourceforge.net/?q=node/view/3

Command Prompt release beta version of plphp
http://www.commandprompt.com/entry.lxp?lxpe=260

SQL: Visual QuickStart Guide book review
http://books.slashdot.org/books/03/06/19/1651249.shtml

Rogue Wave Software Launches SourcePro(TM) C++ Edition 6
http://biz.yahoo.com/prnews/030805/cgtu024_1.html

== PostgreSQL In the News ==

Vic Greens set to complete Linux switch by year-end
http://www.smh.com.au/articles/2003/08/06/1060064230512.html

== Upcoming Events ==

PostgreSQL training at the Big Nerd Ranch: Atlanta, Georgia: Aug 18-22
Bruce Momjian will provide a one week training class.
http://www.bignerdranch.com/Classes/Postgresql.html

1-day PostgreSQL Workshop: Turkey / Istanbul'da: August 23
FrontSITE & TDM is offering 1-day course on PostgreSQL
http://www.frontsite.com.tr
http://www.tdmsoft.com/egitim_ayrinti.php?id=1

== PostgreSQL Weekly News - August 6th 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