Re: [List] Ignore

On 2026-07-19 04:41:19 (+0800), Frank Leonhardt wrote:

> On 18/07/2026 21:00, Kevin P. Neal wrote:
>> On Sat, Jul 18, 2026 at 10:12:43AM +0100, Frank Leonhardt wrote:
>>> Regular readers of this list will know that for reasons unknown my re=
plies
>>> to questions stop appearing, and this is apparently reset by sending =
a
>>> direct message. This will either trigger that, or prove the list is a=
t least
>>> working for non-replies.
>>>
>>> For anyone new to the problem, they come from Thunderbird on a static=
 Unix
>>> domain via a fixed mail server that's been running on that name for o=
ver 20
>>> years and there should be nothing "odd" about them. And yes, I've bee=
n on
>>> this mailing list from the start on the new server, and a decade or t=
wo
>>> before that. It's not a local/user problem.
>> Are you using SPF+DKIM+DMARC? I've read that not having them may cause=

>> problems, but having only SPF records I haven't seen any problems. So
>> I don't know what not having SPF does.
>
> You have to bear in mind I've been on this list a long time - long befo=
re it was moved to this current mailer even. And my domain name predates =
freebsd.org. Yet I have these intermittent phases of having replies to ce=
rtain posts here disappear into the ether. it's been looked at before.
>
> This is the sendmail log of contact with FreeBSD's server:
>
> Jul 18 15:12:50 bs1 sm-mta[64439]: 66IEClIn064437: to=3D<questions@free=
bsd.org>, delay=3D00:00:03, xdelay=3D00:00:03, mailer=3Desmtp, pri=3D3833=
1, relay=3Dmx1.freebsd.org. [96.47.72.80], dsn=3D2.0.0, stat=3DSent (Ok: =
queued as 4h2TJK16nNz48W2)
>
> It's delivered okay, mx1.freebsd.org has handed back it's queue ID. it =
just doesn't make it to the list. Unfortunately the only way I can try to=
 debug it is by sending a message to the list to test it - and these alwa=
ys get through. There must be something about the message I'm replying to=
 that causes the problem. I'm always using the same Thunderbird mailer. S=
ome thread I can reply to, some I can't. The mail server this end handles=
 thousands of emails a day, and has been there for years.
>
> If someone with access at the FreeBSD end of things could figure out wh=
at happened to this message ID after it was accepted and queued that woul=
d be wonderful. No one has ever come forward to say they can do it.

You could ask [email protected] to take a look. ;)

Sat Jul 18 14:13:00 2026 access - A mail from "[email protected]" wit=
h header "X-Spamd-Bar: +++" was denied by rule #1 "deny ^X-Spamd-Bar: \+\=
+\+"

The questions@ list drops spam scores >=3D 3 because it attracts rather a=
 lot of bots.  I'm tempted to drop it to 2+ but ...

This specific message (4h2TJK16nNz48W2) in our queue scored as follows:

2026-07-18 14:12:50 #42979(rspamd_proxy) <4h2TJK>; proxy; rspamd_task_wri=
te_log: id: <[email protected]>, qid: <4h2TJ=
K16nNz48W2>, ip: 185.174.224.37, from: <[email protected]>, (default:=
 F (no action): [3.21/15.00] [NEURAL_SPAM_LONG(1.00){1.000;},SUBJECT_ENDS=
_QUESTION(1.00){},URI_COUNT_ODD(1.00){7;},NEURAL_SPAM_MEDIUM(0.90){0.903;=
},NEURAL_HAM_SHORT(-0.69){-0.691;},ONCE_RECEIVED(0.20){},R_SPF_ALLOW(-0.2=
0){+ip4:185.174.224.37;},MIME_GOOD(-0.10){multipart/alternative;text/plai=
n;},RCVD_NO_TLS_LAST(0.10){},ARC_NA(0.00){},ASN(0.00){asn:12703, ipnet:18=
5.174.224.0/24, country:GB;},DMARC_NA(0.00){fjl.co.uk;},FROM_EQ_ENVFROM(0=
=2E00){},FROM_HAS_DN(0.00){},MID_RHS_MATCH_FROM(0.00){},MIME_TRACE(0.00){=
0:+;1:+;2:~;},MLMMJ_DEST(0.00){[email protected];},PREVIOUSLY_DELIVER=
ED(0.00){[email protected];},RCPT_COUNT_ONE(0.00){1;},RCVD_COUNT_ONE(=
0.00){1;},R_DKIM_NA(0.00){},TO_DN_NONE(0.00){},TO_MATCH_ENVRCPT_ALL(0.00)=
{}]), len: 8744, time: 1336.404ms, dns req: 27, digest: <bc7f300d4a6b7831=
ba136156f82d0b63>, rcpts: <[email protected]>, mime_rcpts: <questions=
@freebsd.org>

There isn't really much low-hanging fruit here.  You could score points w=
ith SPF or DKIM on fjl.co.uk.  The NEURAL_* tests tend to be very transie=
nt (and they cancelled each other out for this message).  I'm not sure wh=
at knobs we can easily turn for you.

Philip [hat: [email protected]]