Re: sqlsmith: ERROR: XX000: bogus varno: 2

On Mon, Dec 20, 2021 at 4:20 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> pg_get_expr doesn't (or didn't) depend on expression_tree_walker,
> so there wasn't a problem there before. I am worried that there
> might be other code paths, now or in future, that could try to apply
> expression_tree_walker/mutator to relpartbound trees, which is
> why I think it's a reasonable idea to teach them about such trees.

I agree that doing so is totally reasonable. I merely don't think that
previous failure to do so makes anyone a "bozo". It was far from
obvious that it was required.

> It's only a problem if you hold the opinion that there should be
> no user-reachable ERRCODE_INTERNAL_ERROR errors. Which is a fine
> ideal, but I fear we're a pretty long way off from that.

I do hold that opinion, and I think we ought to work in that direction
even if we can't hope to get there quickly.

--
Robert Haas
EDB: http://www.enterprisedb.com