Re: Quick Extensions Question

"David E. Wheeler" <david(at)kineticode(dot)com> writes:
> On Mar 3, 2011, at 1:31 PM, Tom Lane wrote:
>> This looks like it would be at most a few hours' work to change,
>> and it would enable creation of extensions for the built-in languages
>> that can be loaded with the same permissions as before.

> Would that time include having extension records for the core PLs, created when you CREATE LANGUAGE and removed when you DROP LANGUAGE?

Extensions yes, but not managed with those commands. You'd have to
switch over to saying "CREATE/DROP EXTENSION plpgsql", etc. The LANGUAGE
commands themselves would now only occur within those extension
scripts.

BTW, a fine point I didn't mention in the previous summary is that if
CREATE LANGUAGE within an extension script creates language support
functions pursuant to what it finds in pg_pltemplate, those functions
should be marked as belonging to the extension, so that they'd get
dropped during DROP EXTENSION. I'm not sure whether the path of control
is such that that'd happen today, but if it doesn't we'd need to tweak
things.

> I assume that non-core PLs must be installed by a superuser?

Right. That's true already, because only a superuser can do the
unvetted 'CREATE FUNCTION ... LANGUAGE c' commands that are necessary
for installing a PL not known to pg_pltemplate. I would like to get
rid of pg_pltemplate eventually (in favor of trusting commands coming
from an extension script), but we're not there yet.

> And if so, then they could be distributed as extensions with superuser = true?

Right.

regards, tom lane