On Wed, Jan 6, 2021 at 7:04 PM David Fetter <david@fetter.org> wrote:
>
> On Tue, Jan 05, 2021 at 06:03:54PM +0100, Magnus Hagander wrote:
> > On Tue, Jan 5, 2021 at 5:34 PM David Fetter <david@fetter.org> wrote:
> > >
> > > On Mon, Jan 04, 2021 at 09:49:43AM -0700, Saad Benateigha wrote:
> > > > Please try not to use PWN as a subject it sent to junk maik
> > > >
> > > >
> > > > From: "PWN via PostgreSQL Announce" <announce-noreply@postgresql.org>
> > > > To: "PostgreSQL Announce" <pgsql-announce@lists.postgresql.org>
> > > > Sent: Sunday, January 3, 2021 3:42:56 PM
> > > > Subject: PostgreSQL Weekly News - January 3, 2021
> > >
> > > Saad,
> > >
> > > Thanks for the heads-up.
> > >
> > > Infrastructure team,
> > >
> > > Is it worth expanding this acronym so's not to trigger what are
> > > apparently keywords having to do with script kiddies?
> >
> > But there is no "PWN" in the subject line?
> >
> > In general, the PWN posts trigger a lot of typical spam filter rules.
> > But I*m pretty sure the subject isn't one of them... For example, the
> > number of links, the link-to-text ratio, the large amounts of repeat,
> > etc.
> >
> > The best recommendation for people with delivery issues is to update
> > their subscription to pgsql-announce to not include PWN, and then read
> > PWN on the website or through RSS.
> >
> > It has gotten a *lot* better with the new posting system, where it is
> > now properly DKIMed and things like that, but it's still by far the
> > largest source of spam-bounces we have. And I don't really think there
> > is any further ways around that, while maintaining it the way it is.
> > Some systems will keep learning more over time though, and hopefully
> > they will learn to stop flagging it.
>
> If there's some ritual I can do on my end that might help with this,
> I'm happy to do it.
One thing could be to stop linking to the individual commits, and
instead just have one link to the gitweb page. that would reduce the
number of links quite a bit, and that's a high-scoring trigger. And
also in that not reproducing the direct links to the list discussions,
instead having people go through the gitweb itnerface for it.
I think reducing those links would probably help with at least those
sets of triggers.
--
Magnus Hagander
Me: https://www.hagander.net/
Work: https://www.redpill-linpro.com/