Re: Returned mail: User unknown

Mail Delivery Subsystem wrote:
>
> The original message was received at Sun, 2 May 1999 00:35:36 +0800 (KRSS)
> from dune.krs.ru [195.161.16.38]
...ops...

> Subject: Re: [INTERFACES] Some needed features in postgresql
> Date: Sun, 02 May 1999 00:35:35 +0800
> From: Vadim Mikheev <vadim(at)krs(dot)ru>
> Organization: OJSC Rostelecom (Krasnoyarsk)
> To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
> CC: Peter Garner <peter_garner(at)yahoo(dot)com>, hackers(at)hub(dot)org
> References: <1464(dot)925575484(at)sss(dot)pgh(dot)pa(dot)us>
>
> Tom Lane wrote:
> >
> > > 2. Is there any work being done on the fact that
> > > transactions are aborted when an error occurs? It
> > > really lessens the usefulness of the product when,
> > > after adding 20,000 records, the transaction aborts
> > > because the 20,001th record is a dup key! :-)
> >
> > I've griped about that myself in the past, but it's not real
> > clear how else it ought to work. Maybe you should be using
> > smaller transactions ;-)
>
> Oracle sets implicit savepoint before processing any
> user' statement.
>
> We still haven't savepoints but live with hope to implement
> them in future -:) Actualy, having MVCC I would like to see
> savepoints implemented as sooner as possible: there is
> high possibility of aborts in serializable mode...
>
> Vadim