Re: Error on failed COMMIT

On Mon, 24 Feb 2020 at 07:25, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:

> On Mon, Feb 24, 2020 at 7:29 AM Dave Cramer <davecramer(at)postgres(dot)rocks>
> wrote:
> > Well the driver really isn't in the business of changing the semantics
> of the server.
>
> I mean, I just can't agree with that way of characterizing it. It
> seems clear enough that the driver not only should not change the
> semantics of the server, but that it cannot. It can, however, decide
> which of the things that the server might do (or that the application
> connected to it might do) ought to result in it throwing an exception.
> And a slightly different set of decisions here would produce the
> desired behavior instead of behavior which is not desired.
>
> --
>

Fair enough. What I meant to say was that the driver isn't in the business
of providing different semantics than the server provides.

> Dave Cramer
> http://www.postgres.rocks
>