On Tue, May 12, 2020 at 09:45:45PM -0300, Euler Taveira wrote:
>> In any case, it seems to me that the comment of
>> build_replindex_scan_key needs to be updated.
>>
>> * This is not generic routine, it expects the idxrel to be replication
>> * identity of a rel and meet all limitations associated with that.
>
> It is implicit that a primary key can be a replica identity so I think this
> comment is fine.
Agreed. I don't think either that we need to update this comment. I
was playing with this patch and what you have here looks fine by me.
Two nits: the extra parenthesis in the assert are not necessary, and
the indentation had some diffs. Tom has just reindented the whole
tree, so let's keep things clean.
--
Michael