[Mimedefang] graphdefang cores with large amounts of data

[Kevin A. McGrail]

> Berkeley DB is blazingly fast, in my experience.
>
> I haven't looked at the graphdefang code at all, so I can't comment
> on your core dumps.

Wow!

How fast is it on something to the equivalent of select sum(anycolumn) where
column=xyz?

I was expecting Berkeley DB to be fast on the creation and straight reading
but not on queries where the data extraction could be optimized. Any
comments about that?

KAM