Re: phasing out pg_pltemplate?

Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com> writes:
> What is the plan for pg_pltemplate? Is there a roadmap to get rid of
> it? (It's not currently blocking anything for me. I'm just wondering.)

I think it's just waiting for someone to put in the effort to make it
unnecessary.

It seems like the extension mechanism could supersede it now, by switching
to a convention where the CREATE LANGUAGE command in the extension script
specifies all the language parameters explicitly. But we would need to do
something extra to replace the functionality of tmpldbacreate --- perhaps
another extension control file flag? Or maybe it'd be good enough to
hard-wire the db-owner-can-create behavior as enabled by TRUSTED, since
tmpltrusted = tmpldbacreate in every existing row.

One thing we'd have to address is how to not choke on old dump scripts
that contain "CREATE LANGUAGE foo" rather than CREATE EXTENSION.
I wonder if we could finesse that by redefining CREATE LANGUAGE with
no parameters as equivalent to CREATE EXTENSION.

pg_upgrade'ing across such a change might pose some challenges too,
not sure.

regards, tom lane