Re: pgsql_fdw, FDW for PostgreSQL server

Hi all,

(2012/03/07 3:39), Tom Lane wrote:
> A bigger issue with postgresql_fdw_validator is that it supposes that
> the core backend is authoritative as to what options libpq supports,
> which is bad design on its face. It would be much more sensible for
> dblink to be asking libpq what options libpq supports, say via
> PQconndefaults().
>
> We might find that we have to leave postgresql_fdw_validator as-is
> for backwards compatibility reasons (in particular, being able to load
> existing FDW definitions) but I think we should migrate away from using
> it.

In the discussion about pgsql_fdw which was proposed for 9.2, some
issues about postgresql_fdw_validator are pointed out.

* The name "postgresql_fdw_validator" conflicts with the name of the FDW
for PostgreSQL which
follows the naming habit of other FDWs.
* dblink depends on postgresql_fdw_validator.
* postgresql_fdw_validator assumes that libpq supports some particular
options.

An idea to resolve these is to add dblink's own validator which doesn't
assume much about libpq, and obsolete postgresql_fdw_validator.

* Add dblink_fdw_validator to contrib/dblink, which is similar to
postgresql_fdw_validator but it assumes less about libpq.
* Add dblink_fdw as default FDW of dblink, which uses
dblink_fdw_validator, and recommend to use it. This would prevent users
from using postgresql_fdw_validator with dblink.
* Mention that postgresql_fdw_validator might be obsolete in future
release in the document of CREATE FOREIGN DATA WRAPPER.

To make the behavior of dblink_fdw_validator similar to that of current
postgresql_fdw_validator, we need to assume that libpq supports "user"
option, and allow it and secret options in only USER MAPPING options,
and allow others in only SERVER options (and reject all debug options).
IMO this is not unreasonable assumption.

Is this proposal reasonable? Any comments and questions are welcome.

Regards,
--
Shigeru HANADA