Subject: Re: [SQL] LIMIT 1 FOR UPDATE or FOR UPDATE LIMIT 1?
From: Rod Taylor <rbt@zort.ca>
To: Robert Treat <xzilla@users.sourceforge.net>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Bruce Momjian <pgman@candle.pha.pa.us>,
	Larry Rosenman <ler@lerctr.org>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
In-Reply-To: <1030544652.3216.7.camel@camel>
References: <200208281352.g7SDqeB13438@candle.pha.pa.us>
	<21006.1030543878@sss.pgh.pa.us>  <1030544652.3216.7.camel@camel>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Date: 28 Aug 2002 10:28:51 -0400
Message-Id: <1030544932.83275.58.camel@jester>
Mime-Version: 1.0

On Wed, 2002-08-28 at 10:24, Robert Treat wrote:
> On Wed, 2002-08-28 at 10:11, Tom Lane wrote:
> > Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > > Larry Rosenman wrote:
> > >> Why? If both old and new are acceptable, why not document it? 
> > >> (Just curious, I'm not wedded to it). 
> > 
> > > Well, showing both versions adds confusion for no good reason,
> > 
> > Yes, particularly considering that LIMIT ... FOR UPDATE corresponds
> > to the implementation behavior (LIMIT acts before FOR UPDATE) while
> > FOR UPDATE ... LIMIT does not.
> > 
> > I concur with documenting only the preferred form (though there should
> > be a note in gram.y explaining that we're supporting the old syntax
> > for backward compatibility).
> > 
> 
> Doesn't the need for a note explaining that we're supporting the old
> syntax say to you that the documentation also needs to say we support
> the old syntax? I can see the bug reports now saying "this is clearly
> not what it says in the docs"...


Yes, both should be documented. But mark the non-preferred version as
depreciated and disappearing soon (whether it does or not is another
story) but discourage people from using it.