> Yeah, that's the parallel I've been making internally. It's have to be
> interpreter per-thread, due to perl5's threading instabilities
yup. at least the interpreter could have persistence.
there's also security dragons lurking in the perl approach that
would have to be slain. ;-)
> right now MySQL doesn't have enough hooks to make it really useful,
i think that's what i'd like to see the beginnings of a framework for.
> though a subclass of Procedure and a call-into-perl udf abstraction
> would be useful.
heh, that's another idea for my back burner.
> That's all armchair quarterbacking, I'm not persuing a perl-in mysqld
> approach because our datasets are too big to do that, and because it's
> all relatively simple integer based foreign & primary key, so Perl's not
> bringing too much advantage to the party.
i guess i'm wondering about the mysqld interface mostly. then we
could both be free to address the specifics of whatever language
we want to solve the problem at hand.