In my new list, I have a user - one of the foundation members, so an important use case - who will not be talked out of using a very old version of Eudora (!) for mail. Replies to threads from that mailer keep sometimes being properly threaded in Hyperkitty and then randomly jumping out into a new thread. This messes up the web view of the threads, and it's not right. Yet the threading appears to be correct in mail programs such as Thunderbird, using the email based interface to the list.
Have others seen such behaviour with ancient mailers? Is it the mailer, or is there a known issue with threading the web interface? Looking carefully at message ID's and how they are threaded to examine the behaviour, I can't see what is causing this. Possibly some interaction with other RFC-822 headers?
Andrew
Am 31. Juli 2019 um 11:42 Uhr -0000 schrieb Andrew Bernard:
Have others seen such behaviour with ancient mailers? Is it the mailer, or is there a known issue with threading the web interface?
Not only with ancient mailers. See:
- https://lists.mailman3.org/archives/list/mailman-users@mailman3.org/thread/Y...
- https://gitlab.com/mailman/hyperkitty/issues/216
It's a bug in Hyperkitty that triggers if the Message-ID is too long and gets folded over multiple lines.
Marvin
-- Blog: https://mg.guelker.eu
Looking at that issue #216, it appears to be fixed in 1.2.3 at least.
However, this reply jumping out problem still persists. Therefore I suppose it must be something else.
How can I raise a bug?
It's quite a bad one, because it renders the web interface as a very messy place, and users don't know which thread to reply to, when there should only be one thread. To the point where I am looking for alternate mailing list software! (but I would really like to stick with GNU Mailman 3).
Andrew
Andrew Bernard writes:
Replies to threads from that mailer keep sometimes being properly threaded in Hyperkitty and then randomly jumping out into a new thread.
I don't understand what you're describing here. Is the archive public so that we can look at it?
Yet the threading appears to be correct in mail programs such as Thunderbird,
Then it's probably a Mailman (HyperKitty) bug. #216 looks like it might be responsible, although in another message you say you think it's fixed in your instance. Again, we need more information to be sure. It's sometimes the case that for various reasons a site has multiple versions installed and the one the site thinks is running isn't actually handling the mail or the archive or whatever.
Possibly some interaction with other RFC-822 headers?
Seems unlikely. Some threading algorithms use only References, others use both References and In-Reply-To. However, in most cases In-Reply-To is just a copy of the most recent message ID in References.
To check hypotheses about a threading bug we really need to see the Message-IDs, References, and In-Reply-To from the problematic messages and the immediate replies, at least.
participants (3)
-
Andrew Bernard -
Marvin Gülker -
Stephen J. Turnbull