* Dimitri Fontaine (dimitri@2ndQuadrant.fr) wrote:
> Stephen Frost <sfrost@snowman.net> writes:
> > That's not really 'solved' unless you feel we can depend on that "create
> > extension from URL" to work at pg_restore time... I wouldn't have
> > guessed that people would accept that, but I've already been wrong about
> > such things in this thread once.
>
> Basically, with the extra software I want to build out-of-core, what you
> have is an externally maintained repository and the scripts are
> downloaded at CREATE EXTENSION time.
I should have included above "unless we actually dump the extension
objects out during pg_dump."
> With the Extension Template, you then have a solid cache you can rely on
> at pg_restore time.
Extension templates are not needed for us to be able to dump and restore
extensions. I do not understand why you continue to argue for extension
templates as a solution to that problem when it's utter overkill and far
worse than just dumping the extension objects out at pg_dump time and
having a way to add them back as part of the extension on restore.
I understand that you once proposed that and it was shot down but I
think we need to move past that now that we've seen what the alternative
is.. That isn't to say anything about the code or about you
specifically, but, for my part, I really don't like nor see the value of
sticking script blobs into the catalog as some kind of representation of
database objects.
Thanks,
Stephen