Date: Sun, 8 Apr 2007 19:07:27 -0600
From: Michael Fuhr <mike@fuhr.org>
To: Dirk Jagdmann <jagdmann@gmail.com>
Cc: pgsql-sql@postgresql.org
Subject: Re: MD5 sums of large objects
Message-ID: <20070409010727.GA80424@winnie.fuhr.org>
References: <5d0f60990704081003p615530acx86257a70bff33fa4@mail.gmail.com>
	<20070408173757.GA78999@winnie.fuhr.org>
	<5d0f60990704081707m56b29423ob5d63103623305b2@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <5d0f60990704081707m56b29423ob5d63103623305b2@mail.gmail.com>

On Mon, Apr 09, 2007 at 02:07:16AM +0200, Dirk Jagdmann wrote:
> this works like charm. Although I did fix the argument for lo_lseek:

Oops; thanks for fixing that.

>  INV_READ  constant integer := 262144; -- 0x40000 from libpq-fs.h

You could also use a hex constant (the cast to integer is necessary):

  INV_READ  constant integer := x'40000'::integer;

>  IF id is null THEN
>    RETURN NULL;
>  END IF;

The above check is unnecessary since the function is marked STRICT,
aka RETURNS NULL ON NULL INPUT.

> I vote for this function beeing included either somewhere in the
> contrib directories, as you often don't need the full power of
> pgcrypto is md5 suffices for your hashing needs.

You could make a proposal in pgsql-hackers but I think 8.3 is in
feature freeze so don't expect to see it until 8.4, if it's accepted
at all.  There's always PgFoundry :-)

-- 
Michael Fuhr