Re: opendkim/opendkim-db.c restructure

From: Murray S. Kucherawy <msk_at_blackops.org>
Date: Mon, 28 Jun 2010 13:30:03 -0700 (PDT)

I think this is a good idea overall. Another example of someplace this
was done is the BerkeleyDB interface, where you get a handle when you open
the DB and then reference get/put/close/etc. methods through the handle.

This touches a huge amount of stuff, but overall will make the code
cleaner. Good on ya for taking on the work. :-)

One thing to mention: We should have a stub function that just returns
some kind of "function not implemented" error code, and any element of the
handle not supported by that DB method should reference that. So for
example this would be used if someone tries a "put" through a CSL data
set.

Also, I'd suggest doing this in a new library, maybe "libdataset" or such,
as more and more of our tools are making use of opendkim-db.c and it's
becoming more of a pain to integrate it the way I have. There's an open
RFE to tie opendkim-testkey to it, for example, meaning now's a great time
to make this separation and clean up the build accordingly.

I may have some comments as it takes shape but basically I really like
this idea.
Received on Mon Jun 28 2010 - 20:30:24 PST