Hi,
> > ifplugin Mail::SpamAssassin::Plugin::BodyEval
> > if can(Mail::SpamAssassin::Plugin::BodyEval::has_check_body_length)
>
> > body __BODY_LENGTH_100 eval:check_body_length('100')
>
> This indeed may be a neat substitution to the __RB_LE_nnn and __CHUNK
> rules discussed, to match short message bodies.
>
> However, there's a considerable difference: The rawbody includes all
> textual MIME-parts only, while the check_body_length() eval uses the
> pristine body, covering all MIME-parts.
>
> In retrospect, this actually (though slightly fuzzily) matches the
> desired behavior of the rules originally discussed with Alex back in Oct
> 2013. The goal was, to identify short message bodies with a URI, but to
> exempt messages where the actual payload is an attachment. That eval
> includes the attachments, rawbody doesn't.
>
> > meta BLAH_100 (__HAS_URI && __BODY_LENGTH_100)
> > endif
> > endif
This is the quoted-printable text/html content:
<HTML><HEAD></HEAD>
<BODY dir=3Dltr>
<DIV dir=3Dltr>
<DIV style=3D"FONT-SIZE: 12pt; FONT-FAMILY: 'Calibri'; COLOR: #000000">
<DIV>Hi! <A =
href=3D"http://example.ru/rnj/trouble.php";>http://example.ru/rnj/trouble.php
</A></DIV></DIV></DIV></BODY></HTML>
where example.ru is the real domain (which is now a dead link anyway). The
rendered version would be so much easier to parse, but can that be done
separately, and will there always be a text component?
Thanks,
Alex
