INN and db 4,6

[Russ Allbery]

Bill Davidsen <davidsen at tmr.com> writes:

> The main issue with bdb is summed up in the first post in this thread,
> "Just when you thought it safe, db 4.6 bites you in the bullocks!!" It
> would just be nice not to have a moving target. And while SQL
> programmers are pretty thick on the ground, I not only can't name five
> people who could program bdb, I'm not sure I could find five people off
> this list who know what it is, beyond "some database program, right?" I
> just worry that Oracle will suddenly decide to de-support it, or make
> some really bizarre change in the API, or... something.

> I'm not really speaking in favor of something as much as not really 
> likeing what we have.

Well, given the resources we have right now, my preferred solution to this
is to tell people to use tradindexed.  :)  I don't think that you'd be
that happy with the performance of a generic SQL layer, but more to the
point, we don't have one.

Porting to new versions of BerkeleyDB is usually pretty trivial, but it
requires someone to have the time to go do it.  It would be more trivial
if we dropped support for all older versions, particularly anything older
than 4.0.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.