Re: Dump public schema ownership & seclabels

On Sat, May 01, 2021 at 09:43:36AM -0700, Zhihong Yu wrote:
> On Sat, May 1, 2021 at 8:16 AM Asif Rehman <asifr(dot)rehman(at)gmail(dot)com> wrote:
> > The following review has been posted through the commitfest application:
> > make installcheck-world: tested, passed
> > Implements feature: tested, passed
> > Spec compliant: tested, passed
> > Documentation: not tested
> >
> > Hi,
> >
> > I have tested this patch. This patch still applies and it works well.
> >
> > Regards,
> > Asif
> >
> > The new status of this patch is: Ready for Committer

Thanks. Later, I saw that "pg_dump --schema=public" traditionally has yielded
"CREATE SCHEMA public" and "COMMENT ON SCHEMA public". I've updated the
patches to preserve that behavior.

I'll push this when v15 branches. I do think it's a bug fix and could argue
for including it in v14. On the other hand, I mailed three total patch
versions now known to be wrong, so it would be imprudent to count on no
surprises remaining.

> For public-schema-comment-dump-v2.patch :
>
> + if (ncomments == 0)
> + {
> + comments = &empty_comment;
> + ncomments = 1;
> + }
> + else if (strcmp(comments->descr, (fout->remoteVersion >= 80300 ?
> + "standard public schema" :
> + "Standard public schema")) == 0)
> + {
> + ncomments = 0;
>
> Is it possible that, in the case ncomments > 0, there are more than one
> comment ?

Yes, I think that's normal when the search terms include an objsubid (subid !=
InvalidOid).