Re: the case for machine-readable error fields

On Wed, Aug 05, 2009 at 09:30:02AM -0500, Kevin Grittner wrote:
> Anyway, the upshot is -- I think that it would be beneficial to allow,
> to the extent we can confirm it's not a violation of any applicable
> standard, a user-defined SQLSTATE to be associated with a constraint.
> I also think that it would be valuable to provide a mechanism for
> PostgreSQL-specific application code to be able to pick off one or two
> table names related to a "standard" constraint violation. I'm less
> convinced at the column or data value level, but I can see where it
> might be useful.

Not sure if overloading SQLSTATE is the right way of doing this is it?
It already has things like 23514 for a check violation and any other
client code relying in this would break if it started getting different
things back.

--
Sam http://samason.me.uk/

p.s. I think you were agreeing with everything else I was saying, even
if I didn't explain myself well enough for you to understand me!