[SC-Help] Re: false multipart header, Spamcop misses URL in body

[Graeme Leith]

Mike Easter wrote:

> That's because you are acting like a human bean, which in some capacities
> is a lot smarter than an algorithm.  You are humanly looking at the
> content type and determining that it is text/html and changing the
> headerline accordingly.
> 
> The parser has to deal with content type multipart alternative
> appropriately.
> 
> Altho' I asked it previously, you haven't said how if you were an
> algorithm instead of a human bean, you would go about parsing a spam
> whose headers contained content type multipart alternative.  Would you
> change them all to text/html, or do it some other way?

Ironically, since JT upgraded to SA 3.0, every one of my mails that SC 
has failed to find links in has had an SA tag of URIBL_SC_SURBL.

Perhaps Julian could use the MIME handling and URI detection algorithms 
from SA to improve SC.

-- 
Evidence shows Cyveillance abuse internet resources.
I recommend unchecking their box in SpamCop reports.
Cyveillance are part of the problem.
They are not part of the solution.