On Wed, Dec 19, 2018 at 05:34:22PM +0900, Michael Paquier wrote:
> On Wed, Dec 19, 2018 at 04:49:19PM +0900, Etsuro Fujita wrote:
>> BTW I noticed that this error occurs not only for foreign tables but for
>> views with INSTEAD OF INSERT triggers. Here is an example on HEAD:
>>
>> To fix this I think we would also need the same treatment for the view case.
>
> Yes, this means an extra relkind check for views. Would you like to
> update the patch yourself?
While I still have that stuff in mind, for a patch on HEAD we could just
use RELKIND_CAN_HAVE_STORAGE() as additional check and also remove the
check on RELKIND_PARTITIONED_TABLE. And just add a check on
RELKIND_VIEW for back-branches on top of what's already present. If you
would like to write a new patch, please feel free to. Or I am fine to
do it myself.
--
Michael