Re: Record returning function accept not matched columns declaration

jian he <jian(dot)universality(at)gmail(dot)com> writes:
> On Sat, Mar 2, 2024 at 1:56 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> I'm having mixed feelings about whether to back-patch this. Somebody
>> might complain that we broke a working query in a minor release.

> context: in postgres 9.3.25, dbfiddle[1]
> this query will fail:

> with a(b) as (values (row(1,2,3)))
> select * from a, coalesce(b) as c(d int, e int);

Sure, but in v12 and up (i.e., the supported versions that we might
back-patch into) the query succeeds, and it might not be obvious
to someone that its behavior isn't as-intended.

Admittedly, this seems like a rather contrived case. In cases
where the record-returning function call isn't reducible to a
constant, you'll get the expected error. So maybe I'm worrying
over something that no one is depending on in practice.

> I think by ExecMakeTableFunctionResult you mean `mainly
> ExecMakeTableFunctionResult's function: tupledesc_match`
> since ExecMakeTableFunctionResult is quite long.

Perhaps. I mentioned ExecMakeTableFunctionResult because that's
what's called directly from nodeFunctionscan.c. Referencing a
static function of another module is risky because it's pretty
unlikely the comment would get updated if refactoring in that
module makes it false.

regards, tom lane