Re: proposal: session server side variables

On 01/04/2017 04:36 PM, Craig Ringer wrote:
> On 5 January 2017 at 08:35, Craig Ringer <craig(at)2ndquadrant(dot)com> wrote:
>> On 5 January 2017 at 01:49, Fabien COELHO <coelho(at)cri(dot)ensmp(dot)fr> wrote:

>>> Good. So we seem to agree that GUCS are transactional?
>>
>> No. We don't agree. They aren't.
>
> Uh. I take that back.
>
> craig=> SET x.s = 'x';
> SET
> craig=> BEGIN;
> BEGIN
> craig=> SET x.s = 'y';
> SET
> craig=> ROLLBACK;
> ROLLBACK
> craig=> SHOW x.s;
> x.s
> -----
> x
> (1 row)
>
>
> I'm surprised, I never knew this.

(I have not been able to keep up with the shear volume on this thread,
but this caught my eye...)

Yeah -- I found it surprising when I first discovered it too. My opinion
is that the design for variables should not behave this way.

Joe

--
Crunchy Data - http://crunchydata.com
PostgreSQL Support for Secure Enterprises
Consulting, Training, & Open Source Development