On Tue, Sep 22, 2020 at 10:57 PM Michael Paquier <
michael@paquier.xyz> wrote:
>
> On Tue, Sep 22, 2020 at 05:35:48PM -0400, Tom Lane wrote:
> > What exactly do you argue has changed since the previous decision
> > that should cause us to change it? In particular, where is the
> > additional data to change our minds about the safety of such a thing?
>
From a technical perspective I really don't know how to solve it, but my intention here is to raise a hand and demonstrate there are real scenarios where Postgres breaks so easily.
And unfortunately for the user perspective it sounds a bit fragile. Ok it's not a very common use case and the solution isn't easy, because if it is I'm sure it was already solved before.
> Not sure that's safe, as we also want to avoid circular dependencies
> similarly for pg_class, pg_index and pg_attribute.
>
Adding a TOAST can cause circular dependencies between those relations?? If you don't mind can you explain more about it?