Re: Vote totals for SET in aborted transaction

Marc G. Fournier wrote:
> On Fri, 26 Apr 2002, Jan Wieck wrote:
>
> > Bruce Momjian wrote:
> > > Marc G. Fournier wrote:
> > > > On Thu, 25 Apr 2002, Bruce Momjian wrote:
> > > >
> > > > >
> > > > > Marc is suggesting we may want to match Oracle somehow.
> > > > >
> > > > > I just want to have our SET work on a sane manner.
> > > >
> > > > Myself, I wonder why Oracle went the route they went ... does anyone have
> > > > access to a Sybase / Informix system, to confirm how they do it? Is
> > > > Oracle the 'odd man out', or are we going to be that? *Adding* something
> > > > (ie. DROP TABLE rollbacks) that nobody appears to have is one thing ...
> > > > but changing the behaviour is a totally different ...
> > >
> > > Yes, let's find out what the others do. I don't see DROP TABLE
> > > rollbacking as totally different. How is it different from SET?
> >
> > Man, you should know that our transactions are truly all or
> > nothing. If you discard a transaction, the stamps xmin and
> > xmax are ignored. This is a fundamental feature of Postgres,
> > and if you're half through a utility command when you ERROR
> > out, it guarantees consistency of the catalog. And now you
> > want us to violate this concept for compatibility to Oracle's
> > misbehaviour? No, thanks!
>
> How does SET relate to xmin/xmax? :)
>

SET does not. But Bruce said he doesn't see DROP TABLE beeing
totally different. That is related to xmin/xmax, isn't it?
What I pointed out (or wanted to point out) is, that we
cannot ignore rollback for catalog changes like DROP TABLE.

Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck(at)Yahoo(dot)com #