Re: Parallel INSERT (INTO ... SELECT ...)

For cfffe83ba82021a1819a656e7ec5c28fb3a99152, if a bool was written (true |
false), READ_INT_FIELD calls atoi() where atoi("true") returns 0
and atoi("false") returns 0 as well.

I am not sure if the new release containing these commits had a higher cat
version compared to the previous release.
If the new releases did have a higher cat version, I guess there was no
issue, by chance.

Cheers

On Sat, Mar 6, 2021 at 8:12 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:

> On Sun, Mar 7, 2021 at 8:24 AM Zhihong Yu <zyu(at)yugabyte(dot)com> wrote:
> >
> > I was looking at src/backend/nodes/readfuncs.c
> >
> > READ_NODE_FIELD(relationOids);
> > + READ_NODE_FIELD(partitionOids);
> >
> > READ_NODE_FIELD would call nodeRead() for partitionOids. However, such
> field may not exist.
> > Since there is no 'if (strncmp(":partitionOids", token, length) == 0) {'
> check, I was curious whether CATALOG_VERSION_NO should be bumped.
> >
>
> Won't that be true only if we are reading the stored planned stmt from
> disk? But is it possible in any way? The last two commits in this
> function (cfffe83b and 45639a05) also didn't bump the catversion.
>
> --
> With Regards,
> Amit Kapila.
>