> v2.x being a compilation-based featureset program was due to be a > complete nightmare. DSPAM 3 seems to be much easier. not at all :/ it's still configure driven, and .. well, for packaging, it sucks :/ The major problem is in fact not in the storage backend (since I guess that with some smart modifications of Makefiles, we can be able to link dynamically to libdspam, and then have the same dpsam binady-package for each backend).
The thing I can't well understand is how to be able to let the user choose between in header signature, or in body, virtual-users or not ... without having some 2^n different packages ;) > Keep in mind that not only one, but all four SQLite, MySQL, PostgreSQL > and BDB4 *have to be supported*... and this is a lot of work ( i know > from experience, packaging Bacula ). Yeah, I guessed that btw, I talked to Jonathan today, and he said to me he recomended SQLite (which is not part of 3.0.0 but will be soon in stable branch) and MySQL, but that others backend are not that interesting (even really bad -- like bdb4 that has concurency issues) > If you ask because you intend to "adopt" the ITP, i suggest we > cooperated (some help could possibly be very useful). No, I don't want to adopt at all, if work is already done, I can maybe help, and it would be great. > I don't know how much packaging experience you have, but this does not > seem to be suitable for a "first package" effort... well, It's not my first but my second package ? <grin> well, I think I can learn very very very quick. And IMHO the bad part is not in the debian-policy-compliance part, but in the `find a way to package this damn thing to provide some configurability to the user' one ... > I am lately considering posting a RFH to help with the > packaging/maintenance, well, I can help, but I miss some experience, so I think even if you are interested in my help, a third person cannot be bad. > since Jonathan Zdiarski (upstream) is a real "release machine". yeah, it seems so ;) Well, to be honest, I'm co-admin of a alumni website, and we do email forwarding. So i'm really interested in the dspam webmail features, and other stuff like that. So I really have motivation to make it work :) (in fact I plan to use it for me too) -- Pierre Habouzit http://www.madism.org/
signature.asc
Description: Digital signature
