On 2/28/20 6:17 AM, Jim Popovitch via Mailman-Users wrote:
On Fri, 2020-02-28 at 19:52 +0900, Stephen J. Turnbull wrote:
Brian Carpenter writes:
I have hired a professional PHP developer to begin work on a new admin/forum interface for Mailman 3.
Too bad. I really sympathize with your goals but am unlikely to be able to contribute directly to implementation (assuming an eventual open-sourcing). Never learned PHP, not going to do it anytime soon.
Stephen, It's difficult for me to parse your thought process on this. Why do you say "Too bad" about someone wanting to improve something that you admit you want no part of?
Well, Steve channeled me earlier, so I'll return the favor. I think Steve is saying "Too bad" he is only talking about the choice of PHP as a platform rather than Python. We absolutely encourage people to develop alternatives to Postorius and HyperKitty for archiving and web management of Mailman. I think all Steve is saying is he could be more helpful with Python as opposed to PHP.
See this paragraph:
That's OK, the point of REST is so *you* *can* do that. I can only speak for myself, but we can help to some extent on the Python side of the REST interface.
Who is this "we", you referenced them a few times in this email.
See the initial paragraph in the Acknowledgements section at https://www.list.org/.
I'm fairly confident in saying that Mark has said (repeatedly now) that there will never ever ever ever ever be another Mailman2 release beyond v2.1.30. Stephen, why do you say there will be? Do you have the project authority to make that statement? Who do I beleive?
Actually, I have said there will not be another release from the GNU Mailman project. That does not preclude anyone from forking that project from Launchpad and doing whatever with it.
I personally am not interested in porting Mailman 2 to Python 3. That's already been done. The result, with a real backend database and some extensions such as the concept of "user" that doesn't exist in MM 2, is Mailman 3 core.
What I'm reading between the lines is that nothing but Django was considered for MM3 (by "we") and everything else is inferior and not worth the time. I'd love to be wrong on that.
The web based archiving and list management we distribute are based on Django because that's how the people who developed those things wanted to do it.
The whole point of Mailman 3 is all that stuff is separate from the core engine and communicates with core via a REST API, so there can be lots of different web management UIs. We knew if we released Mailman 3 core only without a web UI for list management and archive access, no one would use it, so we needed those things and the people who were willing to implement them built what we have.
We certainly hoped that there would be people like Brian implementing alternatives and we're glad to see it.
Agreed. I didn't even know bounce processing wasn't working until this summer (my production lists are all in-house to personal acquaintances to same-university addresses -- if mail isn't flowing to somebody, it's not going to anybody, even Mailman!)
MM3 has been on python.org for a while, was it not noticed there either?
Of course. We began discussing this almost 3 years ago https://gitlab.com/mailman/mailman/issues/343#note_31870366. The implementation was mostly done last year by a GSOC student.
We are a small project. We have very few people working on Mailman on a regular basis, and everyone is a volunteer, no one is paid. If you want things to happen faster, please contribute.