To: <josh@agliodbs.com>
Cc: <pgsql-hackers@postgresql.org>,
	"Heikki Linnakangas" <heikki@enterprisedb.com>,
	"FAST PostgreSQL" <fastpgs@fast.fujitsu.com.au>
Subject: Re: Updatable cursor doubt
In-Reply-To: <200709041737.09541.josh@agliodbs.com> (Josh Berkus's message of
	"Tue\, 4 Sep 2007 17\:37\:07 -0700")
User-Agent: Gnus/5.11 (Gnus v5.11) Emacs/22.1 (gnu/linux)
References: <46DCB86F.9070102@fast.fujitsu.com.au>
	<46DD226F.7060602@enterprisedb.com>
	<200709041737.09541.josh@agliodbs.com>
From: Gregory Stark <stark@enterprisedb.com>
Organization: EnterpriseDB
Date: Wed, 05 Sep 2007 02:12:40 +0100
Message-ID: <877in56hkn.fsf@oxford.xeocode.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii

"Josh Berkus" <josh@agliodbs.com> writes:

> Who are we, MySQL?   We ought not to accept the syntax if we're not going 
> to enforce it.

I think the thinking is that the syntax doesn't promise anything about
enforcing any restrictions. It's a method for the user to declare what
features he needs. Ie, without that clause (or with a FOR UPDATE?) the
database should signal an error in cases where the cursor won't handle
updates. But with that clause the user is telling us that he's ok with
not being able to update the cursor.

Perhaps a better way to think about this case is "should you raise an 
error if someone opens a file in read-only mode when they actually do
have write permission?"

-- 
  Gregory Stark
  EnterpriseDB          http://www.enterprisedb.com