On Apr 19, 2013, at 02:40 AM, Patrick Ben Koetter wrote: >My original point was that we should use MILTERs because we shouldn't reinvent >what already has been implemented very well and we shouldn't come up with a >proprietary solution, because that will make it harder to expand Mailman and >also less versatile as a tool.
I don't know the milter API, but from a quick look at milter.org, it seems like there's a C libmilter library. It looks like there are several options for integrating this with Python, none of which I've actually looked at yet: https://pypi.python.org/pypi?%3Aaction=search&term=milter&submit=search >From your description, it sounds like you might not even need this for our purposes. Just open a socket, shove the right bunch of bytes down it, then read and parse the response. In any case, I'm not opposed to the *idea* of supporting milters in MM3, but I think that's a long way from "yes, let's do it". A lot more analysis and design is necessary, so it would be cool if someone is really interested in pursing this. >I think we miss out some important customers if we stop at "but you only need >to add code to get what you want". > >Mail systems are created and run by postmasters in most cases. Even though >there's a strong movement at the moment towards 'devops', the majority of >sysops/postmasters can't program. > >If we stop at a programmable interface - which I understand is what you and >Terri suggest - they will not be able to use Mailman in more complex setups, >because integrating additional mail processing components would require them >to program. Well, it would require *someone* to program, but I hope not the sysadmin unless they really want to. One of the design goals of MM3 is to allow for much easier extensibility. For some sense of what I envision, look at the way HyperKitty hooks into MM3. You grab some code someone else has written, stick it in some place on the file system, set a configuration variable or two, and restart. The nice thing is that this all then becomes re-usable with no source hacking required, which makes things much better than in MM2. -Barry _______________________________________________ Mailman-Developers mailing list Mailman-Developers@python.org http://mail.python.org/mailman/listinfo/mailman-developers Mailman FAQ: http://wiki.list.org/x/AgA3 Searchable Archives: http://www.mail-archive.com/mailman-developers%40python.org/ Unsubscribe: http://mail.python.org/mailman/options/mailman-developers/archive%40jab.org Security Policy: http://wiki.list.org/x/QIA9