All,
Over the past years (as early as 2014, I think), there have been many times we've seen the need to move off Mailman2: it's implemented in python2.7 which has long seen the end of its life.
The alternative has always been Mailman3, which is the technological successor of Mailman3, but a complete rewrite. As early as its initial release in 2015, I've been eyeballing moving to it. Back then, there was no migration path from 2 to 3, but it was on the roadmap, which made the choice to move or stay easy: wait for the release which makes it possible.
A lot of time has gone by since. The situation gradually worsened, because the migration to Mailman3 became a blocker to upgrade the operating system on common-lisp.net: python2.7 was phased out, which means the mailman2 list software wouldn't run on it anymore (that is, without significant maintenance burden to maintain our own python2.7 install).
Long story short: Yesterday, I installed the containers which allow us to run mailman3 in an isolated environment. There's a lot of tweaking to be done, but I've been able to configure mailman3 to run alongside of mailman2, which is a good start.
As of right now, I've started migrating to mailman3, which I'm doing in steps:
- basic validations - archives of defunct lists - active mailman2 lists with configuration and archives
Before going the route of migrating active lists, I may do an assessment of the activity of mailing lists, asking owners of mailing lists which have gone without activity for 5 years, whether there's still an active interest.
Regards,
Hi Anton,
On Fri, Jul 14, 2023 at 10:03 PM Anton Vodonosov avodonosov@yandex.ru wrote:
Question: will old archive links remain working after the migration?
I mean if somebody in the past shared in some doc, article or in some ticket a link to a mailing list discussion. That can even be a reference from one mailing list message to a thread in another mailing list or in the same mailing list.
Ideal would be to keep such links working.
Yes. I think it's best we leave the existing pages in place. They're static pages, so there's no load on the server to serve them whatsoever. I do think we need to add a header to all the pages saying that there are newer (searchable) archives elsewhere and then pointing to those archives. The pages which used to host the subscriptions should probably be auto-forwarded to the new subscription page.