PostgreSQL Weekly News - July 3rd 2003

== PostgreSQL Weekly News - July 3rd 2003 ==

As of Tuesday we are now in feature freeze. There are a few patches that have
been submitted that will still be reviewed for possible inclusion, but anything
else new that is submitted will be held over for 7.5. Expect formal beta testing
to start in about two weeks (around July 15th), at which time there will be an
official announcement and beta packages will be available. In the mean time
patches that fix bugs and/or enhance documentation are still more than welcome.

We do have a solid list of fixes for items that went in just before feature
freeze, we'll start with a number of improvements for ecpg. A compile problem
for missing LONG_LONG_MIN on BSD/OS was fixed, LONG_LONG_MIN redefinition was
wrapped around HAVE_LONG_LONG_INT_64, ISO dates in pgtypeslib are now used by
default, and fixed an initialization bug in pgtypeslib, and Philip Yarra supplied
some work to enhance thread support. Informix compatible was further increased;
we now make sure Informix style decimal vars are initialized, all cursors are
"with hold" running under INFORMIX_SE mode with autocommit set, and a new data
type "decimal" (similar to our numeric but uses a fixed length array to store
digits) was added.

We also saw a good batch of JDBC enhancements come in this week. IPv6 parsing of
the jdbc URL is now supported, an ArrayIndexOutOfBounds error when calling
moveToCurrentRow while currentRow is "beforeFirst" was fixed, some bounds
checking was added in setMaxRows(), setQueryTimeout(), and setFetchSize(),
getByte and getSort now work with inputs containing decimals or whitespace,
setObject was fixed for objects that are BITs, and some JDBC doc improvements
were put in. Thanks to Kris Jurka, Kim Ho, Barry Lind, and Dave Cramer for their
efforts.

Recent work on array support continues to be built upon. The less than optimal
array_eq has been replaced with real array comparison functions that use the
element datatype's comparison functions, a btree index opclass was created for
ANYARRAY making it possible to create indexes on array columns, we now cache the
results of catalog lookups across multiple array options rather than repeating
the lookups on every call, Several proof-of-concept functions were removed and
string_to_array and array_to_string functions were added, and there were some
minor adjustments to behavior with empty or zero-dimensional arrays. Along
similar lines, expressions of the form 'scalar op ANY (array) and 'scalar op ALL
(array)' are now supported; the operator is applied between the lefthand scalar
and each element of the array and the result of the construct is the OR or AND of
the per-element results, respectively.

Following the array work up was the new ability for SQL functions can have
arguments and results declared ANYARRAY or ANYELEMENT, the effect being to
postpone typechecking of the function body until runtime. Speaking of SQL
functions, we also now support inlining of polymorphic SQL functions, and now run
the raw parser against those functions to catch syntax errors. Polymorphic
functions are also now supported in plpgsql, and the link-list search of function
cache was replacec with hash-table lookup. Rounding out our plymorphic trifecta,
aggregates can now be polymorphic (by using polymorphic implementation
functions). It also works to create a non-polymorphic aggregate from polymorphic
functions, should you want to do that. One last function language mention,
plpython has been renamed to plpythonu and the docs were updated to reflect its
now-untrusted status. This is a significant change; if you are currently using
plpython you will want to be sure to test your existing functions against the new
plythonu implementation.

Everyone’s favorite database GUI, psql, was updated to take advantage of some
features of the new FE/BE protocol. There is a client-side AUTOCOMMIT mode now:
'\set AUTOCOMMIT off' supports SQL-spec commit behavior. We got rid of the
"LO_TRANSACTION hack" since the LO operations just work now, using libpq's
ability to track the transaction status. A VERBOSE variable to control
verboseness of error message display was added along with a %T prompt-string code
to show current transaction-block status. Superuser state display in the prompt
string correctly follows SET SESSION AUTHORIZATION commands and Control-C works
to get out of COPY IN state.

Rounding out this weeks changes includes a number of compiler warnings being
fixed, restructuring the building of join relation targetlists so that a join
plan node emits only those vars that are actually needed above it in the plan
tree, fix some places that were 0(N^2) in the targetlist length, adding some
missing calls to PQclear in large_obj.c and describe.c, is_superuser parameter
reporting was added, Information schema is now aware of arrays, the view
element_types was fixed, the information schema views for group privileges are
now available, a GUC option to log_error_verbosity was added (it controls which
fields of error reports get put into the postmaster log), the syntax for EXECUTE
INTO was changed to CREATE TABLE AS EXECUTE, and the contrib module isbn_issn,
the ISBN/ISSN definitions are now grouped.

== PostgreSQL Product News ==

Mac OS Guru Shipping SQL4X Manager Update
http://www.macobserver.com/article/2003/06/30.2.shtml

== PostgreSQL In the News ==

Analysis: Open source databases
http://www.computerworld.com.au/index.php?id=787941582&fp=16&fpid=0

== Upcoming Events ==

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

LinuxTag: Karlsruhe Convention Center, Germany: July 10-12
A PostgreSQL Talk will be given on July 12th
http://www.linuxtag.org/2003/en/index.html

== PostgreSQL Weekly News - July 3rd 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