Re: ERROR: "ft1" is of the wrong type.

On Fri, Jul 09, 2021 at 09:00:31PM +0900, Kyotaro Horiguchi wrote:
> Mmm. Ok, I distributed the mother regression test into each version.

Thanks, my apologies for the late reply. It took me some time to
analyze the whole.

> PG11, 12:
>
> - ATT_TABLE | ATT_PARTITIONED_INDEX
>
> This test doesn't detect the "is of the wrong type" issue.
>
> The item is practically a dead one since the combination is caught
> by transformPartitionCmd before visiting ATPrepCmd, which emits a
> bit different error message for the test.

Yes, I was surprised to see this test choke in the utility parsing.
There is a good argument in keeping (ATT_TABLE |
ATT_PARTITIONED_INDEX) though. I analyzed the code and I agree that
it cannot be directly reached, but a future code change on those
branches may expose that. And it does not really cost in keeping it
either.

> PG13:
> Of course this works fine but doesn't seem clean, but it is
> apparently a matter of the master branch.
>
> - ATT_TABLE | ATT_MATVIEW | ATT_INDEX | ATT_PARTITIONED_INDEX | ATT_FOREIGN_TABLE
> Added and works as expected.

HEAD had its own improvements, and what you have here closes some
holes of their own, so applied. Thanks!
--
Michael