gward@??? (Greg Ward) wrote on 19.09.02 in <20020920005503.GA20537@???>:
> On 18 September 2002, Philip Hazel said:
> > Crude regex test in a "condition" condition in a DATA ACL?
> >
> > deny condition = ${if match{$h_to:}{@}{no}{yes}}
> >
> > "Reject if no @ in the To: header line."
>
> Except that will reject messages with
>
> To: undisclosed-recipients :;
>
> which, while a fairly strong spam indicator, is not 100% proof of
> guilt.
In fact it is a fairly strong indicator of mailing lists. This text is
used in two examples in RFC 2822, actually - I'd say it is an extremely
bad idea to filter for it.
MfG Kai
