On Wed, Oct 16, 2019 at 01:04:57PM +0200, Dmitry Dolgov wrote:
> Thanks for the update. Looking at v17 0003 I have one more question. In all the
> places where we have to do systable_endscan, it followed by heap_close,
> although in the rest of the file table_close is used. I guess this logic is not
> heap specific and should also use table_close?
We need to use table_close as we cannot assume that the relation will
always be opened for heap. This patch set is around for so long, so
that was a rotten part still able to compile as we have a macro to
cover the gap. Thanks for noticing that. I am sending an updated
patch set in a but, Alvaro had more comments.
--
Michael