Tom Lane wrote:
> Evidently the order in which child tables are visited isn't too stable.
> I'm inclined to think that that's fine and this regression test needs
> reconsideration.
Thanks for pointing it out.
I was going to just drop one of the tables, but then thought it is worth
keeping a rewr
A significant fraction of the buildfarm is showing this in HEAD:
***
*** 375,382
create table rewritemetoo1 of rewritetype;
create table rewritemetoo2 of rewritetype;
alter type rewritetype alter attribute a type text cascade;
- NOTICE: Table 'rewritemetoo1' is being rewrit
