Thomas Lockhart <thomas@fourpalms.org> writes:
>> That's because I already committed the other changes he pointed out ;-).
>> But yeah, we seem to be copy-clean again.
> I had thought that you objected to the guard code in the copy functions
> since nodes should not have had the content they did. And afaik I have
> now fixed the upstream problems with the content.
Right, the SET DEFAULT problem is fixed that way. Fernando had pointed
out a couple of problems in unrelated constructs (GRANT and something
else I forget now) that also needed to be fixed. Those fixes did get
committed.
> Had you changed you mind about the necessity for the guard code?
No. I think Value is fine as-is.
regards, tom lane