From: Bruce Momjian <pgman@candle.pha.pa.us>
Message-Id: <200510051952.j95JqiQ16533@candle.pha.pa.us>
Subject: Re: Updated documentation for new sequence binding
In-Reply-To: <20051005185853.GZ40138@pervasive.com>
To: "Jim C. Nasby" <jnasby@pervasive.com>
Date: Wed, 5 Oct 2005 15:52:44 -0400 (EDT)
Cc: PostgreSQL-development <pgsql-hackers@postgresql.org>,
	PostgreSQL-documentation <pgsql-docs@postgresql.org>
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII

Jim C. Nasby wrote:
> On Sun, Oct 02, 2005 at 10:54:10PM -0400, Bruce Momjian wrote:
> > pgman wrote:
> > > I have marged Tom's description of the new sequence binding with text I
> > > was working on.  I modified it to follow the existing "we used to do X,
> > > now we do Y" pattern in the surrounding entries:
> > > 
> > > 	http://candle.pha.pa.us/tmp/pgsql/release.html#RELEASE-8-1
> > 
> > Sorry, this is a better URL:
> > 
> > 	http://candle.pha.pa.us/main/writings/pgsql/sgml/release.html#RELEASE-8-1
> 
> Out of curiosity, how is this file maintained as development is done?
> The reason I'm asking is because it would be nice to have links to more
> information, especially for the 'one-liner' items in 1.3.1 for example,
> and it seems like that would be easier to do along-the-way rather than
> waiting for the end. Even a link to a mailing list discussion would be
> better than nothing...

I go through the CVS commits and make the release notes just before
beta.  During it incrementally is much more work.

> That said, what's
> 
>  Improve the optimizer, including auto-resizing of hash joins (Tom)
> 
> mean?

We would sometimes fail in a query where the allocated memory was larger
than our hash could hold.  This fixed that.

> 
> On full_page_writes, are we certain that all battery-backed disk caches
> ensure that partial-page writes can't happen?

Well, I would think so, but I have no documentation to state that.  The
battery-backed memory is supposed to allow for power failure and keep
writes in the cache until they hit the disk.

> Maybe pg_column_siz should just be included in the item for all the
> other size functions brought into the backend? If not, ISTM they should
> at least be one after the other in the list...

Not really.  It is different in that is does not aggregate values, but
just the storage of a column.

> Finally, weren't more changes made it contrib than what's listed?
> Nothing's said about pg_autovacuum for example.

Moved to the main server.  That is mentioned.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073