Re: BUG #18950: pgsql function that worked in Postgresql 16 does not return in Postgresql 17

I will work on that over the next few days. I really appreciate you and
your team!

Lowell

On Sat, Jun 7, 2025 at 10:26 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Lowell Hought <lowell(dot)hought(at)gmail(dot)com> writes:
> > I can try. I am not sure how to go about that. I did not see on the bug
> > report page where I could upload files, and I am afraid the file size of
> > the tables needed might be too large for email.
>
> No, uploading stuff to that webform doesn't work. But at this point
> we're just conversing on the pgsql-bugs mailing list, so anything you
> can squeeze into email is fine. Having said that, nobody likes
> multi-gigabyte emails.
>
> > The entire database when
> > written to an sql dump file is about 20 GB, so not terribly large. I
> could
> > attempt to dump the schema definition in one file and then the underlying
> > tables in another. Would that work? Or would you also need the files
> for
> > the function and any views the query relies upon?
>
> Yeah, we'd need all the moving parts.
>
> Usually people with this kind of problem don't want to expose their
> data anyway, for privacy and/or legal reasons. So what I'd suggest
> is trying to create some little script that generates fake data
> that's close enough to trigger the problem. Then you just need to
> provide that script and the DDL and function definitions.
>
> regards, tom lane
>