PostgreSQL 8.3.1, 8.2.7 Update Release

Updates for versions 8.3 and 8.2 of the PostgreSQL database management
system are available today.  This minor release fixes more than 20 minor
issues uncovered by our community in PostgreSQL 8.3.0 and 8.2.6 over the
last few weeks.  All users of 8.3 and 8.2 are urged to schedule an upgrade
at the earliest reasonable opportunity.

As with other minor releases, users are not required to dump and reload
their database in order to apply this update release; you may simply
upgrade the PostgreSQL binaries.  However, users affected by the Windows
locale issue (see release notes) may need to REINDEX some indexes.

Release Notes: http://www.postgresql.org/8.3/static/release-8-3-1.html
http://www.postgresql.org/8.2/static/release-8-2-7.html
Source Code: http://www.postgresql.org/ftp/source/v8.3.1
http://www.postgresql.org/ftp/source/v8.2.7
Binaries: http://www.postgresql.org/ftp/binary/v8.3.1
http://www.postgresql.org/ftp/binary/v8.2.7

--PostgreSQL Global Development Group
--PostgreSQL Core Team

Josh Berkus wrote:
> Updates for versions 8.3 and 8.2 of the PostgreSQL database management
> system are available today. This minor release fixes more than 20 minor
> issues uncovered by our community in PostgreSQL 8.3.0 and 8.2.6 over the
> last few weeks. All users of 8.3 and 8.2 are urged to schedule an upgrade
> at the earliest reasonable opportunity.
>
> As with other minor releases, users are not required to dump and reload
> their database in order to apply this update release; you may simply
> upgrade the PostgreSQL binaries. However, users affected by the Windows
> locale issue (see release notes) may need to REINDEX some indexes.
>
> Release Notes: http://www.postgresql.org/8.3/static/release-8-3-1.html
> http://www.postgresql.org/8.2/static/release-8-2-7.html

These links are clearly wrong. Please at least try to click on them
before you post them.

I have fixed the broken docs links in the announcement on the website.
Which funny enough had *different* errors, three in total, but were
still wrong. Someone should probably send an update on this announcement...

The binary links look ok AFAICS.

//Magnus

Magnus,

> These links are clearly wrong. Please at least try to click on them
> before you post them.

I did actually run the announcement past 16 people, none of whom caught the
error. However, since they weren't web people, nobody looked at the
links. For some reason, the core web team was booted off -packagers,
something I was unaware of, so I didn't know they'd gone unexamined.

Unfortunately, I don't have a composer for e-mail which allows me to "click
on" the links; so if I forget a "cut & paste" test for some of the links
they don't get fixed. Oddly, I did actually check & fix the ones on
wwwmaster, but those didn't save, apparently?

Anyway, we obviously need the core web team (you, Robert, Stefan) in the
review lineup since nobody else checks things like links. What's the best
way to do that?

--
--Josh Berkus

Josh Berkus
PostgreSQL Project Core Team
www.postgresql.org

(all opinions expressed are my own; I do not speak
for the Project unless specifically noted.)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

>> Release Notes: http://www.postgresql.org/8.3/static/release-8-3-1.html
>> http://www.postgresql.org/8.2/static/release-8-2-7.html
...
> I have fixed the broken docs links in the announcement on the website.

Now that the horse is out of the barn, could we at least put in a
one-time redirect for the links above so they aren't 404?

- --
Greg Sabino Mullane greg(at)turnstep(dot)com
PGP Key: 0x14964AC8 200803181856
http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8
-----BEGIN PGP SIGNATURE-----

iEYEAREDAAYFAkfgSK4ACgkQvJuQZxSWSsh8BQCeM9rvYwXCab9JkffkV54doVlc
B1EAn1/ZbDZInn6T7zbqYwwa4uqwvo1p
=TusQ
-----END PGP SIGNATURE-----

All,

> Now that the horse is out of the barn, could we at least put in a
> one-time redirect for the links above so they aren't 404?

Also, can someone please clear my correction announcement? As far as I can
tell, it still hasn't gone out.

Sigh. I'm still paying for Marc accidentally deleting my e-mail account 2
weeks ago. Several people who I thought had reviewed the announcement
didn't because they never received it ...

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

Josh Berkus wrote:
> All,
>
> > Now that the horse is out of the barn, could we at least put in a
> > one-time redirect for the links above so they aren't 404?
>
> Also, can someone please clear my correction announcement? As far as I can
> tell, it still hasn't gone out.
>
> Sigh. I'm still paying for Marc accidentally deleting my e-mail account 2
> weeks ago. Several people who I thought had reviewed the announcement
> didn't because they never received it ...

Uh, you sent out the announcement. I don't see how you can blame Marc
for the wrong URLs at this point.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://postgres.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +

Bruce,

> Uh, you sent out the announcement. I don't see how you can blame Marc
> for the wrong URLs at this point.

I *can* blame him for not being subscribed to pgsql-packagers, when I
thought I was.

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

All,

Anyway, we clearly need a mailing list to run the announcements through
which includes web people, or this'll keep happening. Should that be
-packagers or -slaves? If the former, can we get Robert and Stefan added
to -packagers? Or at least *me*?

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

On Tue, Mar 18, 2008 at 05:03:16PM -0700, Josh Berkus wrote:
> Bruce,
>
> > Uh, you sent out the announcement. I don't see how you can blame
> > Marc for the wrong URLs at this point.
>
> I *can* blame him for not being subscribed to pgsql-packagers, when
> I thought I was.

I need to be subscribed to that list, too.

Cheers,
David (not a packager, but I do need to know in advance when and why
releases are going out for what should be pretty obvious reasons)
--
David Fetter <david(at)fetter(dot)org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david(dot)fetter(at)gmail(dot)com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, 18 Mar 2008 17:08:41 -0700
Josh Berkus <josh(at)agliodbs(dot)com> wrote:

> All,
>
> Anyway, we clearly need a mailing list to run the announcements
> through which includes web people, or this'll keep happening. Should
> that be -packagers or -slaves? If the former, can we get Robert and
> Stefan added to -packagers? Or at least *me*?
>

Why not press?

Joshua D. Drake

- --
The PostgreSQL Company since 1997: http://www.commandprompt.com/
PostgreSQL Community Conference: http://www.postgresqlconference.org/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL political pundit | Mocker of Dolphins

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFH4GtbATb/zqfZUUQRAsDVAJ9Bk8gdBd6NiHWLeL3DNQfOYjKK8wCfeh/0
UkoVMDMIi81G4ifKxkiN6Uk=
=KZ//
-----END PGP SIGNATURE-----

Josh,

> Why not press?

That's an alias for response to public inqury, not a mailing list.

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, 18 Mar 2008 18:33:46 -0700
Josh Berkus <josh(at)agliodbs(dot)com> wrote:

> Josh,
>
> > Why not press?
>
> That's an alias for response to public inqury, not a mailing list.

Then let's make it a open, no subscribe list. It wouldn't behave any
differently than it does now but would allow us to service the
announcement review etc...

Sincerely,

Joshua D. Drake

- --
The PostgreSQL Company since 1997: http://www.commandprompt.com/
PostgreSQL Community Conference: http://www.postgresqlconference.org/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL political pundit | Mocker of Dolphins

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFH4G3fATb/zqfZUUQRAtpQAKCnzumpoxjWSQ1h0y+0HFXsjgkm+QCfcBnw
gOIPCuU6JkQetfgqYWo+Kkg=
=5MKt
-----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, 18 Mar 2008 18:35:27 -0700
"Joshua D. Drake" <jd(at)commandprompt(dot)com> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Tue, 18 Mar 2008 18:33:46 -0700
> Josh Berkus <josh(at)agliodbs(dot)com> wrote:
>
> > Josh,
> >
> > > Why not press?
> >
> > That's an alias for response to public inqury, not a mailing list.
>
> Then let's make it a open, no subscribe list. It wouldn't behave any
> differently than it does now but would allow us to service the
> announcement review etc...

And actually in thinking about this what needs to happen is that
announcements don't go out until they are publicly verified on X list
by two people. This idea that we can all keep up with the various state
of infrastructure or changes to every nook and cranny is silly.

There is a reason when I posted my URL for web changes, I waited, and
waited and then reposted when nobody responded.

Sincerely,

Joshua D. Drake

- --
The PostgreSQL Company since 1997: http://www.commandprompt.com/
PostgreSQL Community Conference: http://www.postgresqlconference.org/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL political pundit | Mocker of Dolphins

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFH4HFyATb/zqfZUUQRAiT/AJ43HYQqq63xbWJ437M7zL6ErEi+GgCdGXBe
C4GMnRSP34VRMOK/DCoFqTg=
=co5I
-----END PGP SIGNATURE-----

Hi,

On Tue, 2008-03-18 at 17:08 -0700, Josh Berkus wrote:
> Anyway, we clearly need a mailing list to run the announcements
> through which includes web people, or this'll keep happening. Should
> that be -packagers or -slaves? If the former, can we get Robert and
> Stefan added to -packagers? Or at least *me*?

I'd vote for -slaves list. Let's keep packagers list for packaging
discussions only.

Regards,
--
Devrim GÜNDÜZ , RHCE
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Managed Services, Shared and Dedicated Hosting
Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/

Devrim =?ISO-8859-1?Q?G=DCND=DCZ?= <devrim(at)CommandPrompt(dot)com> writes:
> On Tue, 2008-03-18 at 17:08 -0700, Josh Berkus wrote:
>> Anyway, we clearly need a mailing list to run the announcements
>> through which includes web people, or this'll keep happening. Should
>> that be -packagers or -slaves? If the former, can we get Robert and
>> Stefan added to -packagers? Or at least *me*?

> I'd vote for -slaves list. Let's keep packagers list for packaging
> discussions only.

+1. Last-minute checks of announcements seem like material for that
group.

The -packagers list knows about security problems well in advance
of releases, so there are obvious reasons to keep that as tightly
held as possible.

regards, tom lane

Greg Sabino Mullane wrote:
>
>>> Release Notes: http://www.postgresql.org/8.3/static/release-8-3-1.html
>>> http://www.postgresql.org/8.2/static/release-8-2-7.html
> ...
>> I have fixed the broken docs links in the announcement on the website.
>
> Now that the horse is out of the barn, could we at least put in a
> one-time redirect for the links above so they aren't 404?
>
I'd vote against it. If we keep putting in redirects everytime someone
posts incorrect URLs, we'll end up with so many redirects we'll soon
enough be entirely unable to add new pages to the site...

The links from the version list on the frontpage are correct, so just
removing the path from the URL will bring you to the right spot - which
is what I'd expect to be standard practice if you encounter a broken URL
in an announcement.

//Magnus

Josh,

> Then let's make it a open, no subscribe list. It wouldn't behave any
> differently than it does now but would allow us to service the
> announcement review etc...

Absolutely not. Requests made to press@ are often semi-confidential;
companies don't want to share their requests about export certification
with Google, and reporters *definitely* don't want to share their story
ideas with their collegues!

I'm really confused on why you're trying to turn a confidential,
quick-response alias into a public mailing list, when you'd just force us
to create another alias.

It sounds like -slaves is the way to go. Anyone have any other ideas?

--
--Josh

Josh Berkus
PostgreSQL @ Sun
San Francisco

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 19 Mar 2008 11:13:27 -0700
Josh Berkus <josh(at)agliodbs(dot)com> wrote:

> Josh,
>
> > Then let's make it a open, no subscribe list. It wouldn't behave any
> > differently than it does now but would allow us to service the
> > announcement review etc...
>
> Absolutely not. Requests made to press@ are often semi-confidential;
> companies don't want to share their requests about export
> certification with Google, and reporters *definitely* don't want to
> share their story ideas with their collegues!

Huh? Just make the archives private.

>
> I'm really confused on why you're trying to turn a confidential,
> quick-response alias into a public mailing list, when you'd just
> force us to create another alias.
>

I am confused why you are confused :P. You can make it an open list
that has closed archives. This protects confidentiality but allows all
communication on the list itself to be archived for further review by
the list members itself.

Further it solves the problem of who needs to be on the list to check
things. It seems that announcements would be best reviewed by those who
are going to be asked about those announcements, thus a press list.

I was trying to avoid being on yet another list. As I already receive
press@ it made sense to me.

Sincerely,

Joshua D. Drake

- --
The PostgreSQL Company since 1997: http://www.commandprompt.com/
PostgreSQL Community Conference: http://www.postgresqlconference.org/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL political pundit | Mocker of Dolphins

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFH4VrFATb/zqfZUUQRAk0iAJ97m3LO6RYw7ps0KSPekR8lCZJlfgCdFRd9
lM20B6XgI2ASsqN0/cZiNMI=
=1789
-----END PGP SIGNATURE-----

On Tuesday 18 March 2008 22:15, Devrim GÜNDÜZ wrote:
> Hi,
>
> On Tue, 2008-03-18 at 17:08 -0700, Josh Berkus wrote:
> > Anyway, we clearly need a mailing list to run the announcements
> > through which includes web people, or this'll keep happening. Should
> > that be -packagers or -slaves? If the former, can we get Robert and
> > Stefan added to -packagers? Or at least *me*?
>
> I'd vote for -slaves list. Let's keep packagers list for packaging
> discussions only.
>

THe issue I have with the slaves list is that it's very high noise to signal,
and I'd be concerned about missing such an email... can we not just do it on
www ?

--
Robert Treat
Build A Brighter LAMP :: Linux Apache {middleware} PostgreSQL

Robert Treat wrote:
> On Tuesday 18 March 2008 22:15, Devrim GÜNDÜZ wrote:
>> Hi,
>>
>> On Tue, 2008-03-18 at 17:08 -0700, Josh Berkus wrote:
>>> Anyway, we clearly need a mailing list to run the announcements
>>> through which includes web people, or this'll keep happening. Should
>>> that be -packagers or -slaves? If the former, can we get Robert and
>>> Stefan added to -packagers? Or at least *me*?
>> I'd vote for -slaves list. Let's keep packagers list for packaging
>> discussions only.
>>
>
> THe issue I have with the slaves list is that it's very high noise to signal,
> and I'd be concerned about missing such an email... can we not just do it on
> www ?

Most of that noise is from the nagios alerts, right? Perhaps we should
look over how we do those notifies. Since a lot of them seem to simply
be ignored, they should at least not be distributed to all the people
who cannot possibly do anything about them...

//Magnus

Hi,

On Thu, 2008-03-20 at 12:21 +0100, Magnus Hagander wrote:

> Most of that noise is from the nagios alerts, right? Perhaps we should
> look over how we do those notifies. Since a lot of them seem to simply
> be ignored, they should at least not be distributed to all the people
> who cannot possibly do anything about them...

+1.

Regards,
--
Devrim GÜNDÜZ , RHCE
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Managed Services, Shared and Dedicated Hosting
Co-Authors: plPHP, ODBCng - http://www.commandprompt.com/

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 20 Mar 2008 07:08:10 -0400
Robert Treat <xzilla(at)users(dot)sourceforge(dot)net> wrote:

>
> THe issue I have with the slaves list is that it's very high noise to
> signal, and I'd be concerned about missing such an email... can we
> not just do it on www ?
>

We can't use www because sometimes the release is private, literally
until we release. It must be a closed list.

Sincerely,

Joshua D. Drake

- --
The PostgreSQL Company since 1997: http://www.commandprompt.com/
PostgreSQL Community Conference: http://www.postgresqlconference.org/
United States PostgreSQL Association: http://www.postgresql.us/
Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFH4obWATb/zqfZUUQRAvt0AJ4pJB6Xjr3TUWPIXAgX0C2gyVeLOgCghNtp
WjzINvI109VLK0oT1q4IhKI=
=0MYD
-----END PGP SIGNATURE-----

On Tue, 2008-03-18 at 15:31 -0700, Josh Berkus wrote:
> > These links are clearly wrong. Please at least try to click on them
> > before you post them.

Sorry for not getting back to you earlier on this. The email ended up in
the grayzone that is "deal with this later"-emails...

> I did actually run the announcement past 16 people, none of whom caught the
> error. However, since they weren't web people, nobody looked at the
> links. For some reason, the core web team was booted off -packagers,
> something I was unaware of, so I didn't know they'd gone unexamined.

I think that was by mistake really.

> Unfortunately, I don't have a composer for e-mail which allows me to "click
> on" the links; so if I forget a "cut & paste" test for some of the links

Wow, such mail programs even exist? :-) I can even do that through mutt
on the commandline :-P

> they don't get fixed. Oddly, I did actually check & fix the ones on
> wwwmaster, but those didn't save, apparently?

Maybe we fixed them at the very same time? There is no locking for edits
on that page, so if you started editing it and I also started editing it
before you saved, I got the non-fixed version.

> Anyway, we obviously need the core web team (you, Robert, Stefan) in the
> review lineup since nobody else checks things like links. What's the best
> way to do that?

A good start is probably to add "verify all web links" to the release
checklist? Doesn't have to be a core web guy to do that - it just
requires someone with a browser...

//Magnus