Re: Dropped and generated columns might cause wrong data on subs when REPLICA IDENTITY FULL

Hi Shi Yu, all

Thanks for updating the patches. It seems you forgot to attach the patches
> of
> dropped columns for HEAD and pg15, I think they are the same as v2.
>
>
Yes, it seems I forgot. And, yes they were the same as v2.

> On HEAD, we can re-use clusters in other test cases, which can save some
> time.
> (see fccaf259f22f4a)
>
>
Thanks for noting.

> In the patches for pg12 and pg11, I am not sure why not add the test at
> end of
> the file 100_bugs.pl. I think it would be better to be consistent with
> other
> versions.
>

I applied the same patch that I created for HEAD, and then the "patch"
command
created this version. Given that we are creating new patches per version, I
think
you are right, we should put them at the end.

> The attached patches modify these two points. Besides, I made some minor
> changes, ran pgindent and pgperltidy.

oh, I didn't know about pgperltidy. And, thanks a lot for making changes and
making it ready for the committer.

> These are patches for dropped columns,
> because I think this would be submitted first, and we can discuss the fix
> for
> generated columns later.
>
>
Makes sense, even now we have 5 different patches, lets work on generated
columns
when this is fixed.

I applied all patches for all the versions, and re-run the subscription
tests,
all looks good to me.

Thanks,
Onder KALACI