Peter Geoghegan <pg@bowt.ie> writes:
> Is somebody going to commit this soon? Alexander?
Done.
I thought the point of the test case was to check VALUES folding
on both sides of the join. I'm not sure why that's actually
a useful expenditure of test cycles, but Alena's patch undid that
without adjusting the comment claiming that's what it's doing.
So I kept the shape of the test the same while substituting
hopefully-stabler tables as she did.
If it's still not stable I think the next step is to nuke both
test queries, since I remain of the opinion that they're likely
a waste of cycles.
regards, tom lane
