On Wed, Mar 31, 2021 at 09:37:44AM +0900, Michael Paquier wrote:
> On Tue, Mar 30, 2021 at 12:02:45PM +0900, Michael Paquier wrote:
> > Okay. So I have looked at that stuff in details, and after fixing
> > all the issues reported upthread in the code, docs and tests I am
> > finishing with the attached. The tests have been moved out of
> > src/bin/pg_dump/ to src/test/modules/test_pg_dump/, and include both
> > positive and negative tests (used the trick with plpgsql for the
> > latter to avoid the dump of the extension test_pg_dump or any data
> > related to it).
>
> I have double-checked this stuff this morning, and did not notice any
> issues. So, applied.
I noticed the patch's behavior for relations that are members of non-dumped
extensions and are also registered using pg_extension_config_dump(). It
depends on the schema:
- If extschema='public', "pg_dump -e plpgsql" makes no mention of the
relations.
- If extschema='public', "pg_dump -e plpgsql --schema=public" includes
commands to dump the relation data. This surprised me. (The
--schema=public argument causes selectDumpableNamespace() to set
nsinfo->dobj.dump=DUMP_COMPONENT_ALL instead of DUMP_COMPONENT_ACL.)
- If extschema is not any sort of built-in schema, "pg_dump -e plpgsql"
includes commands to dump the relation data. This surprised me.
I'm attaching a test case patch that demonstrates this. Is this behavior
intentional?
