On Tue, 11 Mar 2025 at 17:14, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> I wrote:
> > Double ugh. I guess we could get preprocess_function_rtes to
> > insert the appropriate relid ...
>
> OK, that was less painful than I feared. makeWholeRowVar has
> several different special cases for RTE_FUNCTION, but most of them
> don't bear on this problem, because we wouldn't have applied inlining
> when they did. It seems sufficient to fetch pg_type.typrelid for
> the function's nominal return type and store that if it's not 0.
>
Hmm, this introduces a new problem. Testing a case with a composite type:
create table foo (a int, b int);
insert into foo values (1,2);
create type t as (a int, b int);
create or replace function f() returns setof t as
$$ select 1,2 from foo offset 0 $$ language sql stable;
then doing
update foo set b = f.b from f() where f.a = foo.a returning f;
or even just
select f from f();
triggers an Assert() in relation_open() from
get_relation_data_width(), from set_rel_width() because now that the
RTE has a relid, it attempts to open it, without having previously
locked it. Maybe we need to not clear rte->functions, and use that
instead.
Regards,
Dean
