== PostgreSQL Weekly News - September 11th 2003 ==
We are another week closer to 7.4, though beta2 sure seemed to have
enough issues to keep everyone busy. As more folks get into the testing
progress undoubtedly more things will shake out. Thanks go out to all
those who have already done beta testing, for those who haven't please
do; your best time to have issues addressed is during the beta period.
This week saw more changes to help smooth out IPv6 support. We now will
allow IPv4-format entries in pg_hba.conf to match IPv6 connections with
IPv4-embedded-in-IPv6 addresses. We will also include an entry for ::1
in the default pg_hba.conf based on if configure has set HAVE_IPv6. Also
a check on source address of a statistics packet was removed as it is
redundant with kernel-level check.
Other general fixes this week included fixing some problems in both
pltcl and plperl with dropped column support. A change was made to avoid
consuming unreasonable amounts of memory when GRANT has many grantees.
All cases of using the COLLATE clause we reviewed and modified to be
more appropriate, which generally would be a syntax error. Load_file()
now removes pre-existing shlibs that match the new file on either name
or inode. Some typos in hba.c were fixed. Some unneeded srandom() calls
were removed. The printability test in contrib/dbase was cleaned up. We
now avoid using s_addr as a local variable name, to work around broken
system headers on Unixware and a patch to make PostgreSQL use a full
path in DT_SONAME for UnixWare was put in, thanks to Larry Rosenman.
Knocking off another TODO item (improve concurrency of hash indexes)
Tom Lane finished up his reimplementation of the hash index locking
algorithms, based on his proposal to -hackers. This should fix the
problem with hash bucket splits corrupting the state of scans being done
concurrently, and should also fixes all the known problems with
deadlocks in hash index operations. While this should resolve most of
the issues, there is still work that could be done (like make hash
indexes WAL-logged), but this should be a good step forward. As I
mentioned last week, users of hash indexes are strongly encouraged to
download beta3 once it is ready and send in feedback.
After a bit of cleanup, a patch from Kim Ho of Red Hat was committed to
give SQLState support to the jdbc driver. Some of the fixes that went
into the patch included adding the ability to pass exceptions, making
sure all places where errors were received were caught, and adding full
support for the V3 protocol error message syntax was put in. Still todo
are to add SQLState values to all existing exceptions thrown in the
driver and add support for parsing the V3 protocol format for notices.
Ecpg also saw it's share of fixes, starting with improved prototype
casting for thread calls. Other changes included a patch for Informix
handling of numeric/int conversion, all new datatypes were changed to
lowercase, and a rounding bug in numerical types was fixed. Some
documentation bugs were also fixed as well.
As always there were several doc fixes including updating information
about Linux's over-commit memory behavior, a bunch of
formerly-undocumented date/time operators, including the SQL-spec
OVERLAPS construct, are now documented, a stylesheet misfeature leading
to excessively long cross-reference text when linking to a different
"part" was fixed, <body> attributes were removed (CSS stylesheets should
handle that), bibliography formatting was improved, fast-forward links
were added for more convenient navigation, and some content free
diagnostic information was removed from the doc files as well.
And finally this week, although it was decided that win32 was not going
to be ready for 7.4, there has been a flurry of activity on the win32
front as more people in the community are starting to come forward to
help get their itches scratched. Some of the recent work that have been
put in includes fixes to enable win32 to compile libpq again and to
enable SSL compiles, ifdef out pqGetpwuid() because Win32 doesn't have
struct passwd, allowing configure to see include/port/win32 include
files, adding a matching Win32 accept() prototype, allowing pg_id to
compile with native Win32 API, fixing invalid mbvalidate() function
calls (existing bug), allowing /scripts to compile with native Win32
API, and adding win32.c to Win32 compiles (already in *.mak files).
Along with code changes, there was also some more infrastructure put in
place, including a pgsql-hackers-win32 mailing list and nightly
snapshots of the win32 development branch in the /pub/win32_dev
directory of the ftp site. There is also now a web page available
containing project information at
http://momjian.postgresql.org/main/writings/pgsql/win32.html.
== PostgreSQL Product News ==
ADOdb 3.90 Released
http://www.phparch.com/news/538
SWsoft Releases Plesk 6 Power Pack
http://thewhir.com/marketwatch/sws090903.cfm
== PostgreSQL In the News ==
'Storage' to Replace Traditional Filesystems?
http://developers.slashdot.org/developers/03/09/05/1241235.shtml?tid=131&tid=137&tid=189&tid=198
== Upcoming Events ==
Gnu/Linux Congress: Veracruz, Mexico: September 17-19
Bruce Momjian has been asked to give a keynote address and he will also
be presenting tutorials
http://congreso.gulev.org.mx
LinuxWorld Conference & Expo 2003 : Frankfurt, Germany : October 27-29
PostgreSQL will have a booth at the LinuxWorld Conference
http://www.linuxworldexpo.de/
International PHP Conference: Frankfurt, Germany: November 2-5
Bruce Momjian will be presenting several PostgreSQL talks at this year's
International PHP Conference.
http://www.phpconference.de/2003/
== PostgreSQL Weekly News - September 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
