Re: Static snapshot data

Manfred Koizar wrote:
> On Fri, 23 May 2003 13:15:07 -0400, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
> wrote:
> >[good reasons for having SERIALIZABLE subtransactions in READ
> > COMMITTED main transactions]
>
> All I'm saying is if we can have
> (1) a simple version with some restrictions for 7.4 and
> SERIALIZABLE subtransactions for 7.5 or
> (2) nothing for 7.4 and everything for 7.5
> I'd rather have (1); as long as we don't cause incompatibilities, of
> course.

Agreed. Let's get this boat in the water first unless it will be harder
add this functionality later.

> >We already have START TRANSACTION [...]
>
> Great. I was so used to BEGIN that I didn't even think of trying \h.
> :-/

I am just now realizing that because autocommit off is assumed, there
wasn't any being transaction statement in SQL92, and the only standard
one is in SQL99 and it is START TRANSACTION, not BEGIN WORK. I thought
BEGIN WORK was standard, but I guess not.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073