DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=message-id:date:from:to:subject:cc:in-reply-to:mime-version
	:content-type:content-transfer-encoding:content-disposition
	:references;
	b=xAoSJfeSQAWWqg2W2IizC3BV2EKM+Yrp8aEzlnKpjMySFVqwwzDqu6l3vDGXWzpF92
	0UNCQCtdUyFDhVMe3AY1iwXCiaA4/u/pIKnOX6zkbhkk1GVmg2PhPCUd2oHLvcQq8/Ks
	/4Ychn/UwzUZkdzn9gvIk87tFp4AXiUzPXvZo=
Message-ID: <603c8f070810290617rcdc4da4yc2cf88deae4b5ae3@mail.gmail.com>
Date: Wed, 29 Oct 2008 09:17:18 -0400
From: "Robert Haas" <robertmhaas@gmail.com>
To: "Heikki Linnakangas" <heikki.linnakangas@enterprisedb.com>
Subject: Re: Updating FSM on recovery
Cc: "Tom Lane" <tgl@sss.pgh.pa.us>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
In-Reply-To: <49085327.5010608@enterprisedb.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <49072021.7010801@enterprisedb.com>
	<24334.1225205478@sss.pgh.pa.us> <49074779.2080309@enterprisedb.com>
	<9672.1225215390@sss.pgh.pa.us> <49085327.5010608@enterprisedb.com>

> Buffer ReadBuffer(Relation reln, BlockNumber blockNum);
> Buffer ReadBufferExt(Relation reln, ForkNumber forkNum, BlockNumber
> blockNum, BufferAccessStrategy strategy, ReadBufferMode mode);
> Buffer ReadBufferWithoutRelcache(RelFileNode rnode, bool isTemp, ForkNumber
> forkNum, BlockNumber blockNum, ReadBufferMode mode);
>
> Thoughts?

I'm not sure why we would abbreviate Extended to Ext when nothing else
in here is abbreviated.  Seems needlessly inconsistent.

We may also want to rethink our approach to BufferAccessStrategy a
bit.  Right now, we don't admit that
GetBufferAccessStrategy(BAS_NORMAL) just returns a NULL pointer - we
expect the caller to get that strategy and later call
FreeBufferAccessStrategy it just as if it were a real object.
Particularly in light of this API change, I think we should just give
up on that.  Otherwise, a caller that wants to specify a fork number
or ReadBufferMode has to get and free an access strategy that doesn't
amount to anything.  Perhaps it would be sufficient to do this:

#define NormalBufferAccessStrategy NULL

That way, it would be easy to grep for any place where we used this to
get around a useless pair of get/free calls if we ever need to go back
and make a normal buffer access strategy into a real object.

...Robert