>
> I was wondering if you guys managed to get in user-setable tuples > 8191
> into the 6.4 beta so I could get a look at it.
>
I didn't manage to get this done. In the meantime, a new approach to this
problem has been briefly discussed. In 6.5, someone (I if no one else) will
implement tuple chaining. Tuples will able to span blocks, thus limiting
the
tuple size to the system file size limit, in theory.
Not sure what use a db with one massive tuple would be though. :)
Tuples that span blocks would obviously have a performance penalty, but
would
remove the size limit.
Any opinions or comments?
Darren