Boost logo

Boost :

From: Gavin Lambert (boost_at_[hidden])
Date: 2023-05-03 23:44:09


On 4/05/2023 03:55, Andrey Semashev wrote:
> The fact that the initial message is not part of the thread now is a
> relatively recent issue. This is the mailman archive from a year ago:
>
> https://lists.boost.org/Archives/boost/2022/04/index.php
>
> Threading is fine there. In fact, threading was working correctly as
> recently as in February 2023:
>
> https://lists.boost.org/Archives/boost/2023/02/index.php
>
> and has been broken in March:
>
> https://lists.boost.org/Archives/boost/2023/03/index.php
>
> If the system works (and worked) the way you describe, mailman was able
> to associate the initial message with the thread before and now it
> isn't. I suspect, this may have something to do with the headers that
> are added by SES to the messages it sends to the subscribers.

Threading only sometimes works in the gmane view as well. Usually it's
not the mailing list itself at fault, but due to the mail client used by
the person posting -- or sometimes that they're replying to a digest
rather than the original mail, so the context headers are lost (even if
they edit it to look like it was a reply to a single post).

I have noticed an apparent increase recently in the number of posts that
appear "flat" or semi-flat (still showing as a reply, but to a
grandparent post rather than the real parent post) -- but other recent
threads still do appear normally nested, so I don't think it's a
systemic issue.

This is what I'm seeing -- note that the current thread appears
semi-flat (even though most of these sibling messages were actually a
reply chain), but the "replacing the ML with a forum" thread appears
mostly normal.
https://imgur.com/a/GfTS6ig


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk