It's been a while since I've had to deal with this, but it feels like
behavior has changed (INN 2.7.0rc1), or I've misconfigured (or
fundamentally forgot some nuance). What I'm seeing is on two different
servers I am sending a peer articles I don't expect based on the
sitename in the etc/newsfeeds appearing in the Path.

Example one:
On a reader machine I have a single peer (my feeder) with this as the
entry in etc/newsfeeds:

feed1.usenet.blueworldhosting.com\
:*\
:Ap,Tm\
:innfeed!

On that box I am using "pullnews -F feed1.usenet.blueworldhosting.com"
to add feed1.usenet.blueworldhosting.com to the Path of injected
articles to prevent them from being propagated, but these articles are
being fed back to feed1.usenet.blueworldhosting.com.

If I change the etc/newsfeed entry to the following that it stops
feeding the articles:

feed1.usenet.blueworldhosting.com/feed1.usenet.blueworldhosting.com\
:*\
:Ap,Tm\
:innfeed!

The man page for newsfeeds states the following:

"If sitename already appears in the article's Path header field body,
then the article will not be sent to the site."

That doesn't seem to be the case for me. I found other examples on my
feeder machine with similar style etc/newsfeed entries. For example, I
have this in my etc/newsfeeds:

news.swapon.de\
:comp.*,de.*,muenster.*,!*bina*,\
:Ajp,Tm,<65536:innfeed!

When I receive an article from that peer I am feeding it back to them:

Jun 2 02:01:40.496 j news.swapon.de <jfr5ijFc6qcU1@mid.individual.net>
2491 inpaths! news.swapon.de
(extra peers removed from output)

The same is happening for other peers who do not have an exclude pattern.

I don't believe this used to happen in the past, but I put down INN for
a few years since 2.5.x. In examples where I'm seeing this are
copy/pastes from peer's examples, and I've verified the sitename appears
as the Path in the articles in question.

Regards,

Jesse

On 02/06/2022 08:11, Jesse Rehmer wrote:

> Example one:
> On a reader machine I have a single peer (my feeder) with this as the
> entry in etc/newsfeeds:
>
> feed1.usenet.blueworldhosting.com\
>         :*\
>         :Ap,Tm\
>         :innfeed!
The Ap flag means only look at the /exclude pattern and ignore the site
name. You don't have an /exclude pattern in this entry.

Cheers,
Tom

On 6/2/22 5:04 AM, Tom Furie wrote:
> The Ap flag means only look at the /exclude pattern and ignore the site
> name. You don't have an /exclude pattern in this entry.
>
> Cheers,
> Tom

Ah, I see, so when the Ap flag is present it ignores sitename (as its
expected to only use the flag with an abbreviated sitename, apparently).
I wasn't aware using the flag required an exclude pattern.

Interesting that I did not notice this years ago, but the peers whom
gave me these config blocks have had the same ones on their websites for
years.

Thanks for the assist, Tom!

Cheers,

Jesse

On 02/06/2022 17:36, Jesse Rehmer wrote:

> Interesting that I did not notice this years ago, but the peers whom
> gave me these config blocks have had the same ones on their websites for
> years.
Many inn installations are subtly misconfigured, most likely even mine
(or maybe that should be /especially/ mine ;) ). Nah, when my install is
misconfigured, it's rarely subtle :D

> Thanks for the assist, Tom!
You're welcome :)

Cheers,
Tom

On 6/2/22 1:35 PM, Tom Furie wrote:
> Many inn installations are subtly misconfigured, most likely even mine
> (or maybe that should be /especially/ mine ;) ). Nah, when my install is
> misconfigured, it's rarely subtle :D

The only reason I would ever say I prefer Diablo to INN is for it's ease
and straightforward configuration model. With INN it is very easy to
make small mistakes that don't completely break things but produce
weird, sometimes unnoticed behavior.

I'm far from an expert newsmaster, but I do try my best to keep things
configured as close to proper standards as I can get them. Though,
every time I feel I have a handle on dealing with PGP signed control
messages a new problem crops up. Thankfully INN still has an active
(and helpful) community.

Cheers,

Jesse