From: Christopher Browne <cbbrowne@acm.org>
Subject: Re: Nested Sets
Date: Sat, 16 Apr 2005 17:34:28 -0400
Organization: cbbrowne Computing Inc
Lines: 27
Message-ID: <m31x9a2yor.fsf@knuth.cbbrowne.com>
References: <fc63e1af0504160531795a314d@mail.gmail.com>
	<pan.2005.04.16.13.04.47.372088@arvin.dk>
	<fc63e1af05041607286c0447e8@mail.gmail.com>
	<20050416181457.GA14377@phlogiston.dyndns.org>
	<pan.2005.04.16.20.02.24.212579@arvin.dk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
User-Agent: Gnus/5.1007 (Gnus v5.10.7) XEmacs/21.4 (Jumbo Shrimp, linux)
Cancel-Lock: sha1:NkvUyWmAT9goJdci1RDTVaUqMdc=
To: pgsql-sql@postgresql.org

After takin a swig o' Arrakan spice grog, troels@arvin.dk (Troels Arvin) belched out:
> On Sat, 16 Apr 2005 14:14:57 -0400, Andrew Sullivan wrote:
>> You shouldn't need to do anything special around table
>> locking.
>
> - Except of one wants "True Serializability" (see chapter 12.2 in the
> manual). But I don't know if it's possible to handle table locking from
> within a user defined function.

"Inside stored functions" is one of the ultimate examples of places
where you are certain to be honest-to-goodness inside a transaction.

You can't change transactions while inside a function; all the
in-the-function processing is sure to take place in one transaction's
context.

In theory, savepoints may ultimately change that a little bit, in that
you might have portions of processing in different subtransactions.

But nonetheless Andrew's point remains valid: There is no need to do
any special locking surrounding processing that goes on inside a
stored procedure because it is all suitably embedded in a transaction.
-- 
(reverse (concatenate 'string "moc.enworbbc" "@" "enworbbc"))
http://linuxdatabases.info/info/slony.html
"The problem with the current Lisp Machine system is that nothing ever
calls anything anymore."  -- KMP