Re: possible proposal plpgsql GET DIAGNOSTICS oid = PG_ROUTINE_OID

Hi

po 3. 4. 2023 v 19:37 odesílatel Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> napsal:

> Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> writes:
> > po 27. 3. 2023 v 5:36 odesílatel Kirk Wolak <wolakk(at)gmail(dot)com> napsal:
> >> I have marked the item Ready for Commiter...
>
> > Thank you for doc and for review
>
> I'm kind of surprised there was any interest in this proposal at all,
> TBH, but apparently there is some. Still, I think you over-engineered
> it by doing more than the original proposal of making the function OID
> available. The other things can be had by casting the OID to regproc
> or regprocedure, so I'd be inclined to add just one new keyword not
> three. Besides, your implementation is a bit inconsistent: relying
> on fn_signature could return a result that is stale or doesn't conform
> to the current search_path.
>

ok

There is reduced patch + regress tests

Regards

Pavel

> regards, tom lane
>