<<< Date Index >>>     <<< Thread Index >>>

Re: ANNOUNCE: maildir header cache for current development cvs



Hello together,
I found an answer to the 'why I had that much cache misses with ndbm
with Michael Elkins and my patch':

        The sum of the sizes of a key/content pair must  not  exceed
        the internal block size (currently 1024 bytes). Moreover all
        key/content pairs that hash together must fit  on  a  single
        block.  dbm_store() will return an error in the event that a
        disk block fills with inseparable data.

Michael reseverd 2^14 to store the content which more than 2^10 bytes.
Now the question is. Do we want to suport ndbm and dbm? Than I have to
split up the content in two or more key/content pairs. It shouldn't be a
problem. But we could also just use gdbm and berkley db.

        Thomas