Negative scores (was Re: [Mimedefang] scores displayed in $tests)

[listuser at numbnuts.net]

On Wed, 7 May 2003, David F. Skoll wrote:

> On another topic:
> 
> I had an interesting experience trying to send my earlier message to
> the SA-talk list.  I've configured my mail server to reject mail from
> "<>" to "postmaster at roaringpenguin.com".  The theory here is that I
> never send mail as postmaster, so I should never get any bounces to
> that address.  Turns out Sourceforge's mail server does a check for a
> postmaster@ account using "<>" as the sender address, and it was
> bouncing my mail (because it looked like the postmaster@ account
> didn't exist.)
> 
> To get around this stupid behaviour, I now defer the rejection to the end
> of the DATA statement, keeping Sourceforge's lame checker happy, while
> keeping bounces from forged spam off my system. :-)

Sourceforge has done a lot of lame things lately.  Their download
configuration made cpan roll over and die a horrible death.  Following the
download link would send you to a page that generates a 404.  However that
404 page also has a link to a random mirror and your file.  It also uses a
META refresh to send you to that link.  The 404 it generated caused cpan
to roll over and die.  CPAN wouldn't follow the META refresh, probably
because the page was supposedly a 404.  I reported it to the cpan folks.  
Needless to say SF is no longer a CPAN mirror, or so it appears.

http://download.sourceforge.net/mirrors/

Oddly enough at the same time I also found out that one of perl.org's MXs 
was hosted by a spamming company.  The gentlemen that responded to my 
email, Ask Bjoern Hanson, received a bounce from my MTA while trying to 
reply.  He forwarded me the bounce from another address and I looked into 
it.  One of perl.org's MXs was 64.70.54.95 (onion.valueclick.com) and I of 
course block valueclick's netblocks like they were the plague.

http://groups.google.com/groups?hl=en&lr=&ie=ISO-8859-1&q=valueclick&btnG=Google+Search&meta=group%3Dnews.admin.net-abuse.*

It would appear the perl.org no longer uses Valueclick, Inc either.

Justin