So, as a longtime Mailman2 site-admin, I've been poking around Brian's Mailman3 installation, getting to know it better. Mailman3 was a from-scratch rewrite. Among other things, the familiar "pipermail" archiver (in Python 2.7.x) has been replaced with "hyperkitty" in Python 3, and the familiar "listinfo" Web interface for subscribers and site-admins has been replaced by "postorius".
Those of us still running Mailman2 (see for comparsion http://linuxmafia.com/mailman/listinfo/conspire ) have a near-term problem, because Python 2.7.x is now dead, dead, dead, and therefore so is pre 3.x Mailman. OTOH, Mailman3 has been controversial for reasons I don't want to get deep into, to the point that some site admins are looking at ways to move sideways to other MLMs (mailing list managers).
The postorius page for lug-nuts is here: https://bigbrie.com/mailman3/postorius/lists/lug-nuts.bigbrie.com/ The archives ("hyperkitty") page is here: https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ (As with Mailman2, those two pages cross-link to each other.)
Point 1: Brian, maybe you should add one of those URLs (maybe the postorius one) to the standard footer applied to postings by Mailman. IMO, that was a very useful default feature of Mailman2. Anticipating the objection: Yes, that information is provided in a MLM-added header in all postings, which anyone who shows full headers can see. That's in fact how I found those two pages. However, IMO, it's really handy to have one in the footer, where _non_-technical people might notice it. (Like, I'll bet that maybe two other people reading this comment were aware of those normally-hidden headers.)
Point 2: On the postorius page, I notice there's no link for subscribers to view the membership roster, as there was on the corresponding Mailman2 listinfo page. Brian, is that public-facing feature just totally gone? Is there a way to switch it back on?
IMO, there are multiple good reasons why, for most mailing lists, the subscribers should have access to view the subscriber roster. Among them is continuity: It makes a mailing list easier to re-host if any of a variety of bad things happen. (Like, sadly, we're all mortal.)
Point 3: On a hunch that the "Subscription/Unsubscription" link on the postorius page gives access to subscription options, as it did on the Mailman2 listinfo page, I tried that "Log in" button, taking me to https://bigbrie.com/mailman3/accounts/login/?next=/mailman3/postorius/lists/... .
That page has "Username or e-maiL" and "Password" fields, and also a "Forgot Password?" link. Attempting my subscriber e-mail address and what I _historically_ used as my subscription password, I got "The e-mail address and/or password you specified are not correct."
OK, fair enough. Quite probably, Brian let Mailman3 generate random subscriber passwords when he re-hosted the mailing list. So, I go back to https://bigbrie.com/mailman3/accounts/login/?next=/mailman3/postorius/lists/... , and attempt "forgot Password?", which goes to https://bigbrie.com/mailman3/accounts/password/reset/ , where you can re-enter your e-mail address and hit the "Reset My Password" button. When I do that, I see "The e-mail address is not assigned to any user account" -- and cannot do subscriber-password reset.
So, Brian, although I'm self-evidently subscribed, Mailman3 claims I'm not when I ask to reset my password. Stalemate. Something is broken.
Point 4: On the https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ archives page, _yesterday_ when I tried to drill down to individual archived threads and postings, I got HTTP errors. I was just now rechecking, in order to cite specifics, and _now_ all those links appear to work.
So, I guess there's no reproducible bug to investigate, but I swear I wasn't hallucinating. ;-> If I had to guess, I think there were glitches in hyperkitty's back-end database at the time. Be advised.
Point 5: Any chance you can get the pre-Feb. 28, 2022 mailing list archive from Roberto Leibman, and import it into the Mailman3 history? I'm sure hyperkitty has the facility to read in the mbox file from a Mailman2 mailing list, as part of the migration/upgrade facilities.
I wrote:
Point 4: On the https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ archives page, _yesterday_ when I tried to drill down to individual archived threads and postings, I got HTTP errors. I was just now rechecking, in order to cite specifics, and _now_ all those links appear to work.
This glitch is back.
Starting at the archives (hyperkitty) page, https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ , _this_ thread is listed as thread #1:
#1 Poking at Mailman3: some glitches Sat Mar 26, 6:02 p.m.
Selecting that link goes to https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/thread/3VF... which displays only this diagnostic:
Server error An error occurred while processing your request.
I wrote:
#1 Poking at Mailman3: some glitches Sat Mar 26, 6:02 p.m.
^^^^
Also, time stamps for postings in _UTC_? Really? (Perhaps this weird default can be corrected? Or, hey, I suppose we can all pretend we live in Greenwich. The Royal Naval Observatory is worth visiting, completely aside from the opportunity to stand on the Prime Meridian marker.)
On Sat, Mar 26, 2022 at 11:15:36AM -0700, Rick Moen wrote:
I wrote:
#1 Poking at Mailman3: some glitches Sat Mar 26, 6:02 p.m.
^^^^Also, time stamps for postings in _UTC_? Really? (Perhaps this weird default can be corrected? Or, hey, I suppose we can all pretend we live in Greenwich. The Royal Naval Observatory is worth visiting, completely aside from the opportunity to stand on the Prime Meridian marker.)
I will look at this at some point.
Quoting Brian E. Lavender (brian@brie.com):
I will look at this at some point.
And, FWIW, just now, the detailed-postings display on https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ (the "hyperkitty" archives page) is just spining endlessly, not filling in the details of any threads. Again, intuition suggests this "smells like" a database query failure. (Just a surmise from the outside.)
Does anyone still use listserv?
On Sat, Mar 26, 2022 at 11:19 PM Rick Moen rick@linuxmafia.com wrote:
Quoting Brian E. Lavender (brian@brie.com):
I will look at this at some point.
And, FWIW, just now, the detailed-postings display on https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ (the "hyperkitty" archives page) is just spining endlessly, not filling in the details of any threads. Again, intuition suggests this "smells like" a database query failure. (Just a surmise from the outside.)
Lug-nuts mailing list -- lug-nuts@bigbrie.com To unsubscribe send an email to lug-nuts-leave@bigbrie.com
Quoting Linus Sphinx (sphinxtar@gmail.com):
Does anyone still use listserv?
Nobody I know -- but L-Soft still offers it for sale. Got a spare $2,500 - $4,300 for a copy of LISTSERV Lite? http://www.lsoft.se/products/listserv_license.asp
It originated on BITNET in 1986 running on IBM mainframes, kept being maintained by author Eric Thomas while he worked at CERN, and then went with him to Sweden when he got hired by Swedish University Network (SUNET). In the early '90s, he refactored the code to make it easier to port, and released versions for VMS and Unix in 1994 (then for WinNT in '95).
L-Soft lost mindshare and market share to less-expense and less restricted alternatives ListProc and Majordomo during that decade, and eventually to full-featured open source competitors such as GNU Mailman, Sympa, and ezmlm.
The Linux/LUG world mass-abandoned Majordomo around 1998, the moment GNU Mailman proved feasible, having the advantage of an integrated Web-archiver (pipermail) and being open source, which Majordomo has never been.
I need to assemble an online bestiary of all credible open-source alternatives to Mailman3, because with the EOLing of all Python2 code, Mailman 2.x is not going to be feasible much longer -- and I still have doubts about Mailman3.
I'v started to put such a resource together a couple of times, but didn't keep my notes, darn it. So, need to re-research all that.
Quoting Linus Sphinx (sphinxtar@gmail.com):
Does anyone still use listserv?
As a minor correction to my prior post, the _very_ first implmentation of LISTSERV was written _prior_ to 1986 by Ricky Hernandez of EDUCOM, based on a software design by Ira Fuchs of BITNET and Dan Oberst of EDUCOM. Eric Thomas's reimplementation, originally called Revised LISTSERV, replaced it in 1986.
Some more of that history is linked from this mailing list FAQ page on my Web server. http://linuxmafia.com/pub/skeptic/skeptic_faq.html#history
(I note that ListProc, which I alluded to earlier, is actually open source -- but incredibly crufty at this late date.)
I get a number of rest error messages
On March 26, 2022 11:18:34 PM PDT, Rick Moen rick@linuxmafia.com wrote:
Quoting Brian E. Lavender (brian@brie.com):
I will look at this at some point.
And, FWIW, just now, the detailed-postings display on https://bigbrie.com/mailman3/hyperkitty/list/lug-nuts@bigbrie.com/ (the "hyperkitty" archives page) is just spining endlessly, not filling in the details of any threads. Again, intuition suggests this "smells like" a database query failure. (Just a surmise from the outside.)
Lug-nuts mailing list -- lug-nuts@bigbrie.com To unsubscribe send an email to lug-nuts-leave@bigbrie.com
Quoting Brian E. Lavender (brian@brie.com):
I get a number of rest error messages
I wish you successful hunting.
In Mailman 2.1.x, the logfiles (typically in /var/lib/mailman/logs/) are very useful for debugging -- although rarely needed, in my experience.
If you'll pardon my saying so, it seems to me that using Fedora as a server platform was taking a bit of a chance on, well, a perpetual beta. Personally, if I were doing a server deployment of the scarlet chapeau persuasion, I think I'd look first at Rocky Linux. FWIW, I notice that the most recent full release of Rocky Linux (8.5, released 2021-11-15) offers a mailman-2.1.29 package.
https://distrowatch.com/table.php?distribution=rocky https://rockylinux.org/ https://old.reddit.com/r/linux/comments/qlvku7/we_are_rocky_linux_ama/
-
Brian E. Lavender -
Linus Sphinx -
Rick Moen